Chapter 1

Basic
Jeffrey

Cell Culture
W. Pollard

1. Introduction
This article will describe the basic techniques required for successful cell culture. It will also act to introduce some of the other chapters in this volume. It is not intended, as this volume is not, to describe the establishment of a tissue culture laboratory, nor to provide a historical or theoretical survey of cell culture. There are several books that adequately cover these areas, including the now somewhat dated but still valuable volume by Paul (I), the mult r-authored Methods in Enzymology volume edited by Jakoby and Pastan (Z), and the new edition of Freshney (3). Instead, this chapters focus will be on the techniques for establishing primary rodent cell cultures from embryos and adult skin, maintaining and subculturing these fibroblasts and their transformed derivatives, and the isolation of genetically pure strains. The cells described are all derived from Chinese hamsters since, to date, these cells, have proved to be the most useful for somatic cell genetics (4,5). The techniques, however, are generally applicable to most fibroblastic cell types. I will only discuss growing fibroblastic cells in semidefined media. A very detailed consideration of serum-free culture and the maintenance of epithelial cells can be found in Chapter 21. Methods for culturing many other nonfibroblastic cell types are described in Chapters 2 through 31. 1

Pollard

2. Materials
1. Alpha minimum essential medium (a-MEM): for economy, we buy prepared medium as powder. A 44 g quantity of sodium bicarbonate is added, the powder is made up to 20 L in deionized distilled water, the pH adjusted to 7.4, and the media sterile-filtered through a 0.22 w filter using a pressure vessel coupled to a filtration apparatus and driven by a pressurized 95:5% CO, gas mix. This gas mix maintains pH upon storage. The 500 mL bottles are stored at 4OCin the dark until use (seeNotes 1 and 2). 2. Growth medium: a-MEM plus 15 or 7.5% (v/v) fetal calf serum. This is made up as required and stored at 4OC. 3. Ca2+Mg2+-free Phosphate Buffered Saline (Dulbecco, PBS): NaCl, 8 g/ L; KCl, 0.2 g/L; KH,PO, 0.2 g /L; Na,HPO, 7HZ0, 2.31 g /L pH 7.2. 4. PBS citrate: PBS + sodium citrate at 5.88 g/L. 5. Trypsin: One vial of lyophilized Difco Bacto-trypsin in 400 mL of PBS citrate (0.125% trypsin) or 10 x this concentration for the isolation of embryonic fibroblasts (seeNote 3). 6. Counting fluid; PBS + 0.2% (v/v) FCS. 7. Formalin fixative: 10% (v/v) commercial formaldehyde (comes as a 40% v/v solution). 8. Methylene blue stain: 0.1% (w/v> methylene blue in distilled water filtered through a Whatman No. 1 filter. 9. Trypan blue: 0.5% (w/v) in PBS. 10. Colcemid: 10 pg/mL, stored at 4OC. 11. Karyotype fix: Methanol: acetic acid (3:l) made up on the day of use and kept on ice in a tightly stoppered bottle. 12. Giemsa stain: Use commercial Giemsa concentrate diluted 3:47 parts in commercial Gurrs buffer (one tablet to 1 L distilled water). Alternatively, 10 mM potassium phosphate, pH 6.8, can be used as the buffer. The diluted stain is only stable for 2-3 mo.
l

3. Methods 3.1. Establishment Fibroblast

of Primary CuZtures
Culture

Chinese

3.1.1. Embryo

1. Kill a 10-12-d pregnant Chinese hamster with ether. 2. Wash the animal in tap water and then with 70% ethanol.

Basic Cell Culture

3. Make a surgical incision on the dorsal side to expose the uterus using sterile instruments (these can be dipped in ethanol and flamed to maintain sterility during the operation). 4. Remove the uterus in toto, and transfer it to a sterile Petri dish. Dissect out the embryos, and place them in a new sterile Petri dish (seeNote 4). 5. Mince the embryos very finely, and while still in the Petri dish, wash the pieces with 5 mL of 0.125% Bacto-trypsin at 37OC. 6. Tilt the Petri dish so that embryo pieces go to the side. Remove the pieces into a 50 mL centrifuge tube using a wide-bore pipet. 7. Add 40 mL of fresh 1.25% Bacto-trypsin, and incubate at 37OCfor 5 min. 8. Regain the embryo pieces by centrifugation at 100s for 3.5 min, and discard the supernatant. 9. Resuspend the pieces in 40 mL of fresh 0.125% Bacto-trypsin, and incubate at 37OCfor 25 min (this can be performed in a roller apparatus). 10. Neutralize the trypsin with 4 mL of FCS. 11. Deposit the supernatant through a 100 pm sterile mesh into another centrifuge tube. 12. Centrifuge the supernatant for 5 min at 3008 at room temperature. 13. Resuspend the pellet in 10 mL a-MEM plus 15% FCS, and count the cells in a hemocytometer at about l/100 dilution. 14. Lay down 1.5 x 1Ocells in 40 mL of a-MEM plus 15% FCSinto a 75 cm2, and place it in a 37OCtissue culture incubator. 15. The next day, replace the medium with an equal volume of a-MEM plus 15% FCS. 16. Forty-eight to72 h later, the monolayer should be confluent, and at this point, the cells are ready for subculture. This is performed byincubating the monolayer with 4.5 mL of 0.125% Bacto-trypsin at 37OCuntil the cells detach. Cell detachment can be visualized either by observing the cell monolayer in oblique light or directly under the microscope. When the cells have detached (-go%), add 0.5 mL FCS (lo%), pipet up and down five times, and transfer contents to a 15 mL centrifuge tube. 17. Centrifuge the cells at 3008 for 3.5 min at room temperature. 18. Remove the supernatant, resuspend the cell pellet in 5 mL a-MEM plus 15% FCS, and determine the cell concentration. 19. Resuspend the cells at 4 x 106/vial in a-MEM plus 15% FCS plus 10% (v/v) sterile DMSO, and freeze at -120 or -176OC. The cells will remain viable for several years.

Pollard

20. The cells may also be subcultured at one-third to one-tenth dilutions. They have doubling times of approximately 36 h. At this point, start to calculate the number of mean population doublings by keeping careful records of subculture number and split ratio (seeChapter 3 for details). 3.1.2. Skin Fibroblasts (seeChapter 2 of this volume for human explants) Kill and wash an animal as described for the isolation of embryonic fibroblasts (points 1 to 3). In fact, it is often convenient to prepare skin fibroblasts from the same animal as the one from which the embryos were obtained. Cut small pieces (l-2 mm2> of dermis from the exposed skin flaps using sterile instruments and avoiding any fur. Place several (5-10) small pieces (seeNote 4) into a 25 cm2 flask, and allow them to adhere for 5 min in a very thin film of medium (0.5 mL). Once adhered, add 5 mL of growth medium to the opposite surface of the flask. Place the flask in the incubator in the upside-down position for 24 h (the surface tension holds a thin film of medium to the upper surface). Once the explants are firmly stuck, gently invert the flask and return to the incubator. Thenext day, it is often advisable to change the medium to remove any debris and unattached explants. After several days, first epithelial type cells and then fibroblast will grow out of the explants. Let this process continue until most of the surface is covered with fibroblasts or until obvious necrosis is observedin theexplant. Itmaybenecessary tochange themediumevery week until substantial outgrowth is observed. Remove the explanted material (see Note 5) with a Pasteur pipet attached to a vacuum line leaving the adherent fibroblasts. At this stage, depending on the density, the fibroblasts can either be trypsinized (>50% confluent) or allowed to continue to grow to form a monolayer before they are trypsinized, subcultured, and frozen as described above.

1.

2. 3. 4.

5. 6. 7.

8. 9.

3.2. Maintenance and Subculture of Transformed Cell Lines

Many transformed cell lines will grow both as monolayers and in suspension culture. The CHO-S cell line is just such a line having been

Basic Cell Culture

selected for suspension growth by Thompson from the original Kl isolated by Puck (seeNote 6, Ref. 5). Because CHO cells are transformed, they do not require as much serum as normal diploid fibroblasts, and we routinely culture them in 7.5% (v/v> FCS. Despite the relative ease with which transformed cells can be cultured, however, unlike normal diploid fibroblasts, they do not enter a stationary phase of long-term viability (6,7). In this phase, they rapidly lose viability, and therefore must be subcultured during the exponential phase of growth and cannot be maintained as arrested cultures in reduced serum. 1. CHO cells are stored frozen at -4 x lo6 cells/mL at -120 or at -176C (liquid nitrogen) in growth medium containing 7.5% FCS and 10% (v/v) DMSO. A single vial is removed from the frozen stock, rapidly defrosted in a 37C water bath, and the cells regained by centrifugation at 3008 at room temperature. 2. The supernatant is discarded, and the pellet resuspended in 1 mL of prewarmed medium and placed into a 25 cm2 flask or a 60mm diameter dish containing 4 mL of growth medium. 3. Approximately 2 d later, the cells should be almost confluent and ready for subculture (seeNote 7). They are trypsinized as described for the primary diploid fibroblasts. After cell detachment, FCS is added to 10% and the cells resuspended as single cells by pipeting up and down about five times with a 5 mL pipet. An aliquot of this cell suspension (up to a total of 10% of the recipient volume of the medium) can be added directly to a new tissue culture vessel containing growth medium and returned to the incubator until the next subculture. Alternatively, the cells may be regained by centrifugation, resuspended, and the concentration/ml determined (see below). Known concentrations of cells may then be subcultured by appropriate dilution. In a 25 cm2 flask with 5 mL of growth medium, CHO cells should yield about 2.5 x 105/cm2 but yields are variable depending on serum batch and media used. 4. At this stage, cells may be transferred to a magnetically stirred spinner flask (commercially available) containing pregassed (95% sir/5% CO,) growth medium; usually a 250 mL spinner flask is seeded to give a density of -8 x 104cells/mL. These flasks are then placed in a warm room or in a temperature-regulated water bath (Heto), and stirred at 100 rpm (full details of spinner culture and scale up are described in Chapter 6 of this volume; also see Note 8). CHO cells grown in suspension should give -lo6 cells/mL at saturation density, at which
point the medium will be very yellow (acid).

Pollard

3.3. Determination

of CeZZNumber

This can be performed either using an electronic particle counter (e.g., Coulter Electronics Inc.) or a hemocytometer. The former is the more accurate and can be used to count low concentrations of cells (-lo3 cells/mL); the latter requires higher density and is more prone to sampling error, but allows a visual estimation of the health of the cells and, combined with Trypan blue exclusion, can be used to estimate cell viability. 1. Resuspend cells to give a uniform cell suspension by pipeting up and down against the side of the plastic centrifuge tube. 2. If the cells have been trypsinized, as described above, 0.2 mL of the cell suspension to 7.8 mL of counting fluid in a 15-mL Falcon snap-cap tube will give a statistically reliable cell count (1000-14,000 particles/ 0.5 mL counted). Count three aliquots with the Coulter counter set to count 0.5 mL, sum the three counts, divide by 3, and multiply by40 (for dilution) and 2 to calculate the cells/mL (seeNote. 9). 3. The cells can then be appropriately diluted for the experimental setup or subculture. 4. Alternatively, the cells can be counted on a hemocytometer. The cells need to be resuspended at 3-5 x 105cells/mL. A drop of a cell suspension is added to either side of the hemocytometer, taking care not to overfill it and making sure that the coverslip is firmly in place. 5. Each large square on the hemocytometer (improved Neubauer type) gives an area of 1 mm2 and a depth of 0.1 mm, i.e., the volume is 1W mL. Count the cells in the square (usually using the one bounded on each side with triple lines) on either side of the counter, average the counts, and divide by 2 and multiply by lo4 to give the number of cells/ml. If there are too many cells (>lOOO), just count the 5 diagonal squares and multiply by 5 to give the number to be multiplied by 104. If there are too few cells, count more than one complete square on each side of the chamber, and divide the total cell number accordingly. 6. This procedure can also be used to determine cell viability, since prior to placing the cells in the hemocytometer, they can be diluted 1:l with 0.1% Trypan blue. The number of cells that can then exclude the stain (i.e., have intact cell membrane) can be determined by counting the cells as described above.

X4. Isolation

of Genetically

Pure Cell Lines

Theisolationof somaticcell mutants is outside the scope of this article, and thexeader is referred to Thompson (8) for the considerations necessary

Basic Cell Culture

to isolate such mutants successfully All cell lines will genetically alter over time, however, and periodically the parental type will need to be purified from variants or revertants. The easiestway to do this is to isolate a single clone. This gives some potential problems, however, since a clone may itself be a variant, and thus several clones will need to be isolated and tested to ensure the phenotype selected is the required one. To overcome this problem of clonal variability, it is usually better to contract the cell population to about 100 cells and then expand this to the mass culture. This contraction should statistically remove any variants from the population. It is worth remembering, however, that any variant that has a growth advantage over the parental type will soonovergrow the wholeculture. Once a mass culture is obtained, it should be frozen in a large number of vials (20-50) to provide a base for future experiments. This enables the investigator to grow a culture for approximately 3 mo before discarding it, and then to return to the frozen stock for the next set of experiments. This protocol reduces the genetic drift in the culture and avoids the necessity of frequent genetic purification using the methods described below. 1. Trypsinize a culture, recover the cells, and determine the cell number as described above. 2. Dilute to 2.5 cells/mL with 20 mL of growth medium. 3. Plate out 0.2 mL/well into a 96-well tissue culture plate. 4. Incubate plates at 37C in an humidified incubator for IO-12 d. 5. Examine every well with a microscope, and ring those that have a single clone. 6. Trypsinize 2-3 of these individual clones with 0.2 mL of trypsin and, once detached, transfer the wells contents into 4 ml. of growth medium in a snapcap tube. 7. Pipet this up and down to ensure a single ceI.t suspension, and then plateitagainatone-tenthserial dilutions (i.e.,O.4-3.6mE)and0.2mL~ well into a 96-well tissue culture dish (seeNote IO). 8. Return these new plates to the incubators. Add medium fiorn a di& ferent batch to the trypsinized wells of the old plates, and also return this to theincubator. This provides abackupincasethenewpIatesare contaminated. 9, After lo-12 d, select individual clones in the new plates, and expand them up to mass culture (remember to always keep abackup culture). 10. Freeze alarge stock (20-50 vials) as describedabove,sinceaatthisstage, you will have a genetically pure line (except for the mutations that may have occurred during the clones expansion). Split the frozen

Pollard

stock between a liquid N2 store (long-term) and a -70 or-120C store (short-term experimental stock). 11. Alternatively, the mass culture that needs to be genetically cleansed can be plated into 60-mm dishes containing 5 mL growth at 100 cells/ dish. 12. Leave these to grow for approximately 10 d. Trypsinize the clones from each plate and expand each to a mass culture in the normal way. 13. Freeze 20-50 vials of these cultures as described in item 10.

3.5. Karyotyping
It is often desirable to karyotype your cells. Full details for banding and identifying karyotypes are given in Chapter 32 of this volume. This chapter, therefore, will deal with a simple method, derived from Deaven and Petersen (9) for producing karyotypes of Chinese hamster cells. 1. A culture growing in the exponential phase of growth (i.e., having a high mitotic index) in a 10 mL suspension culture (2 x 105cells/mL) or as a monolayer (lo6 tells/60-mm plate) is treated with colcemid at 0.06 pg/mL for 2 h to accumulate cells in mitosis. 2. For the monolayer culture, tap the plate and remove the medium containing detached mitotic cells. Trypsinize the remaining monolayer, pool with the medium, and proceed. 3. Regain cells by centrifugation at 3008 for 3.5 min at room temperature. 4. Resuspend cells in 1 mL of growth medium, add 3 mL of distilled water, and invert to mix (do not pipet because the cells are fragile). 5. Leave for 7 min to allow the cells to swell (this can be altered if satisfactory spreads are not obtained). 6. Add 4 mL of freshly prepared ice-cold fixative (methanohacetic acid, 3:l) directly to the hypertonic solution to avoid clumping. 7. Regain the cells by centrifugation at 300g for 3.5 min. 8. Disperse the pellet gently by agitation (do not pipet) in 10 mL of fixative. 9. Repeat this procedure three times. At this point, the fixed cells can be stored for a week at 4OC or slides can be made immediately. 10. Using a Pasteur pipet, drop 2-3 drops of the resuspended cells onto a chilled slide from about 20 cm. Blow gently onto the surface, and place the slide onto a hot plate at 60-65OC (just too hot to keep the palm of ones hand on the plate). 11. Leave the slide to dry for 5 min, and then place in a staining chamber (a Coplin jar) ensuring that the surfaces do not touch(see Note 11). 12. Stain the karyotypes with Giemsa for 3 min.

Basic Cell Culture

13. Wash the slides by dipping the slides through three additional Coplin jars each containing 50 mL of water. 14. Dry the slides and count the chromosome number under the microscope, or process for banding (seeChapter 32, this vol.).

3.6. Serum and Media

Testing

Before a new batch of serum or media is purchased, it is advisable to obtain a sample from the manufacturer and test its growth-supporting characteristics. This is particularly important for serum. I usually select two of the most used cell types in the lab-currently these are a human diploid fibroblast strain and CHO cells-to test their growth and plating efficiencies (seeNote 12). 1. Make up individual aliquots of growth media, all containing the same media batch, but with the different test sera and including the serum batch currently being used (or vice versa if you are testing media batches). 2. Plate the cells into 15 dishes for each test media at 5 x 105cells/60 mm tissue culture dish and containing 5 mL of the media. 3. Every day for 5 d thereafter, trypsinize the cells from triplicate plates and determine the cell number/plate. 4. Plot a growth curve (log cell number vs time), and calculate the doubling time and saturation density. 5. At the same time as setting up the growth curves, seed in triplicate 60mm dishes containing5 mL of the appropriate media with 100 and 200 cells (6 plates/ test). 6. After lo-12 d fix the culture for 15 min by flooding with formalin. 7. Tip the media and formalin down the drain, and stain the clones with methylene blue. 8. Leave the stain for 15 min, and then wash it away with water. 9. Leave the plates stacked up against each other to dry in a 37OC room. 10. Count the colonies. 11. The three parameters of doubling time, saturation density, and plating efficiency should allow the section of a serum (or media) that gives optimal growth (seeNote 13).

4. Notes
1. The shelf-life of a powdered medium is several years. Once reconstituted, however, this is reduced to 2-3 mo, mainly because glutamine

10

Pollard is unstable. If older medium is used, the glutamine should be replenished (292 pg/mL). The pH of a medium, upon storage, should not be allowed to rise, and to achieve this, good plastic caps with close-fitting rubber inserts should be used. I also find it useful to seal the caps with a strip of Parafilm, since this prevents condensation around the cap rim and, thus, minimizes the risk of fungal contamination. Medium containing Hepes can also be used to avoid bicarbonate buffering. I have never been entirely happy, however, with the cells long-term growth characteristics in this medium. a-MEM is a rich, multipurpose medium developed by Stanners et al. (10) to grow hamster cells. I have not had the experience of any mammalian cell type that will not grow in this medium, including hybridomas. It is expensive, however, and many cells will tolerate less rich and, therefore, cheaper media. Purified trypsin can also be used and is sometimes necessary, e.g., for macrophage cell lines (II), but it is much more expensive and usually not necessary. The citrate chelates Mg2+ and Ca2+and replaces EDTA (Versene) in the buffer. It is advisable to keep fibroblast cultures from individual animals distinct, since it may be required to distinguish between individuals genetically. If the explant is not necrotic, it is possible to remove it with sterile forceps and transfer it to anew cultureflaskfor further outgrowthof cells. The detailed derivation of the various CHO stains is given in Gottesman (5). It should be noted that CHO is a proline auxotroph and should always be maintained in proline-containing medium. CHO cells can maintain viability, providing the medium pH does not become alkali at 4OCfor extended periods of time (7-10 d). Cultures in capped bottles can therefore be moved to the cold room to avoid subculture under desperate circumstances. Primary cell cultures may also be grown on microcarriers in suspension culture. Full details of this technology are given in Chapter 6 of this volume. The Coulter counter should have a 140 yM aperture and the thresholds set as described in the machines Instruction Manual. Serum in the PBS prevents cells from aggregating and giving unreliable counts. The counter sometimes gets partially blocked; only experience of the time taken for each count and for the cells narticular displav on the

2.

3.

4.

5. 6. 7.

8. 9.

Basic Cell Culture

11

10.

11.

12.

13,

spectroscope will indicate problems with counting. Gentle brushing of the orifice with a camel-hair brush will unblock the counter. The Coulter counter can also give a visual display of cell volumes. This, when combined with a pulse height analyzer, can be used quantitatively to measure cell volume or to determine cell viability by estimating the amount of cell debris in a sample. To maintain genetically pure cell lines, it is absolutely essential not to cross-contaminate cultures. To ensure this, fresh pipets must always be used at every step. Do not re-enter a media bottle with a pipet that has been near a culture. Similarly, never pour from a media bottle into a culture. Splash-backs can occur. If you have more than one culture at a time in a tissue culture hood, only one of these should be opened at any one time. Meticulous attention to these small details will prevent the cross-contamination scandals (e.g., HeLa cells in all cultures!) that one so often reads about. It is usual to prepare one slide and check it with phase contrast microscopy so that adjustments can be made on subsequent slides. If there are many nuclei without cytoplasm and a few metaphases, reduce the swelling time. If there are many scattered chromosomes, blow less vigorously. If the metaphase spreads are overlapping, either swell for a longer time (up to 40 min) or blow more vigorously. All these parameters need to be adjusted according to the local environment conditions and cell type (seeChapter 32 of this volume for greater detail and ref. 12). If many cell lines are being used, it is of ten impractical to test the serum out on all the cell types. Usually the most difficult to grow are chosen for the test, but caution needs to be exercised since I once had a batch of serum that supported the cloning and growth of primary diploid fibroblasts but failed to allow cloning of CHO cells! This procedure need only be performed about once every year. Enough serum can then be ordered for the next year, since the serum is stable at -2O*C for at least 2 yr. We used to check our serum using [3H]-thymidine incorporation 1 d after seeding the cells, but given the hazard of using radioactive thymidine, we abandoned this procedure. It is less labor-intensive, however, than measuring growth curves and gives perfectly adequate results. Details of measuring radioactive isotope incorporation into acid insoluble material may be found in Chapter 9, Volume 1 of this series.

12

Pollard

Acknowledgments
I would like to thank Ms. D. Wylie, my tissue culture unit manager, for all the expert help she has given over the years. I would also like to acknowledge the training given to me by Dr. Cliff Stanners, many of whose methods are represented in this chapter. This paper was written while my research was supported by the Medical Research Council UK and the Cancer Research Campaign UK.

References
1. 2. 3. 4. 5. 6. Paul, J. (1975) CeEZ and Tissue Culture (Churchill-Livingstone, Edinburgh and London). Jakoby, W. B. and Pastan, I. H. feds.) (1979) Methods Enzymd. VoZ 58. CeII Culture (Academic, New York). Freshney, R. I. (1987) Culture of Animal CeZZs: A Manual of Basic Techniques (Alan R. Liss, New York). Siminovitch, L. (1976) On the nature of heritable variation in cultured somatic cells. CeZZ 7, l-l 1. Gottesman, M. M. (1985) Molecular Cell Biology (Wiley, New York and London). Pollard, J. W. and Stanners, C. P. (1979) Characterization of cell lines showing growth control isolated from both the wild type and a leucyl-tRNA synthetase mutant of Chinese hamster ovary cells. J. Cell Physiol. g&571-586. Stanners, C. P., Adams, M. E., Harkins, J. L., and Pollard, J. W. (1979) Transformed cells have lost control of ribosome number through the growth cycle. 1. CeZZ Physiol. 100,127-138. Thompson, L. (1979). Mutant isolation. Methods Enzymol. 58,308-322. Deaven, L. L. and Petersen, D. F. (1974) Measurements of mammalian cellular DNA and its localization in chromosomes, inMethods in CeZZ Biology, vol. 8 (Prescott, D. M., ed.), Academic, New York and London, pp. 179-204. Stanners, C. P., Eliceiri, G. L., and Green, H. (1971) Two types of ribosomesin mousehamster cells. Nature (Nm BioZ.) 230,52-54. Morgan,C., Pollard, J. W., and Stanley, E. R. (1987) Isolation and characterization of a cloned growth factor dependent macrophage cell line, BACI.2F5. 1. Cell Physiol.,

7.

8. 9.

10. 12.

130,420-427.
22. Worton, R. G. and Duff, C. (1979) Karotyping. Mefh. Enzpol. 58,321344.

Chapter

Establishment, Maintenance, and Cloning of Human Primary Cell Strains

Gareth E. Jones

1. Introduction
My laboratory is involved in characterizing the behavior of cultured fibroblasts established from skin biopsies of normal boys and those affected with Duchenne muscular dystrophy. We are also interested in the properties of fibroblasts obtained from female carriers of this X-linked disease who are generally clinically unaffected (1). Following the argument of the hypothesis for random X-chromosome inactivation proposed by Mary Lyon as a gene dosage compensation mechanism in all placental mammals, it can be predicted that the dermis of female carriers of Duchenne dystrophy will be populated by fibroblasts mosaic for the Duchenne genotype. By chance, some cells will be expressing the gene products of the unaffected X chromosome, whereas others will have inactivated the normal X chromosome and be expressing the Duchenne gene product. In theory, each carrier of Duchenne dystrophy should contain a roughly 1:l ratio of normal and Duchenne-expressing fibroblasts, and techniques of cell cloning applied to cultures of carrier biopsy material should produce clones of cells exhibiting properties of either normal OYDuchenne cell cultures in a similar ratio (2). In practice, things are not so simple, but the
13

14

Jones

theory outlined above has led to the development of a routine cell-cloning method in my laboratory, which will be described here. Alternative methods do exist, which I have also tried and found to be quite successful, but more time-consuming. One of these alternatives will be briefly outlined, but I shall largely concentrate on the methods we use to establish a culture of human fibroblasts from a biopsy and then to clone those cells. The success of our cloning method is largely the result of the use of adsorbed fibronectin as a seeding substratum; fibroblastic cells bind avidly to fibronectin via a cell surface glycoprotein-receptor complex known as integrin, resulting in high cell seeding efficiencies. There are a few problems unique to collecting human biopsy material. First, the tissue sample is usually taken for sound clinical reasons, such as aiding diagnosis. The needs of the pathologist must be met first, and it is essential that your wishes should have been communicated to the clinician in charge well in advance of the time of biopsy. Secondly, there is the problem of repeat sampling; multiple biopsy of one human is very rare, since it often cannot be clinically justified. Similarly, one cannot always be confident that the site of biopsy is consistent between patients. For these and other reasons, it is virtually obligatory to obtain the active participation of clinical colleagues who are well versed in the requirements of a prolonged research program plus the backing of the hospital ethical committee for your project. Lack of attention to these early considerations will probably render your research efforts useless.

2. Materials
1. Basal salts solution (BSS) such as Hanks saline (seeAppendix).
2. Growth medium containing 10% fetal calf serum (fetal serum can be 3. 4. 5. 6.

7.

replaced by newborn for subcultures). Calcium- and magnesium-free BSS (CMF). 0.25% Crude trypsin in CMF. For cell cloning, choose a rich medium; we use Hams F10 (see Appendix) or MCDB 105 supplemented with 15% fetal calf serum. Stock solution of human plasma fibronectin prepared following instructions of supplier, diluted to 10 pg/mL in phosphate-buffered saline. Stock solutions of cell lysis buffer: 10 mM Tris-HCl buffer (pH 7.6>, containing 1 mM EDTA, 0.5% (v/v) NP-40,lO JJM NADP, and 1 mM E-amino-n-caproic acid.

Human

Primary

Cell Strains

I5

3. Methods
3.1. Collection of Biopsy
1. Provide a labeled container of medium to your clinical collaborator. A

15-mL sterile centrifuge tube with leakproof cap is ideal, about half full of complete culture medium as given in the Materials section. 2. If it will not be possible to transfer the biopsy directly to the tissue culture laboratory, provide a Thermos flask filled with ice. If kept at 4C, biopsy samples will survive for periods up to 3 or 4 d, though tissue degeneration may be expected. 3. It should not be necessary to ensure that the biopsy will be taken under controlled sterile conditions. The site of collection will be sterilized(usually with 70% ethanol), perhaps following shaving to remove hair, and many patients will ask for a local anesthetic (2-3 mL of lidocaine or similar) (seeNote 1).

3.2. Primary

Explant

Culture

1. Transfer the biopsy to fresh sterile BSSin a centrifuge tube, and rinse by gentle hand agitation for a few minutes. Transfer to a Petri dish. 2. Dissect off unwanted fat and necrotic material. Using scalpels, chop up the material into blocks of about 2-mm cubes. 3. Wet the inside of a few wide-bore Pasteur pipets to stop subsequent sticking of biopsy fragments to the glass. Transfer fragments using Pasteur pipets to a 15-mL sterile centrifuge tube, and allow the pieces to settle. 4. Wash twice in fresh BSS,aspirate off the BSS, rinse using a Pasteur pipet and replace with fresh buffer. Allow the pieces to settle under gravity. 5. Rinse the culture surface of a 25-cm2 culture flask with I mL of complete growth medium. There should be just enough to cover the whole surface with a film of liquid after a vigorous shake or two. This will aid the attachment of the biopsy fragments to the culture surface. 6. Wet a clean Pasteur pipet. Transfer pieces of tissue to the culture flask, placing them evenly over the whole available surface. Ideally, you should aim to have some 20 fragments of material per flask, but less is practicable and often inevitable with human material. 7. Cap the flask and invert the culture vessel so that tissue fragments will be in a hanging drop configuration. Surface tension properties will

Jones
retain a pool of medium around and over the explants, while gravity will drain excess medium to the upper surface of the flask. Transfer the flask to a tissue culture incubator set at 37OCand 5% CO,/ air mixture. A 37C hot room is sufficient if media are HEPESbuffered. Leave the flasks in a hanging drop configuration for up to 3 h. The tissue fragments should stick to the culture surface within this time. Return the culture flasks to a sterile cabinet, still inverted. Uncap and remove excess medium (containing any explant material that failed to adhere to the culture surface) using a Pasteur pipet. Reorientate the culture flask, add 1-2 mL of growth medium, cap the flask, and return to the culture incubator. Leave for 18-24 h. Examine the cultures. If theexplants have adhered, makeup themediurn volume to 5 mL, and then change weekly until a substantial outgrowth of cells is seen (seeNote 2). Remove the explants. These may either be picked off with a pair of sterile fine curved (watchmakers) forceps or dislodged with suction pressure generated down the capillary column of a fine-bore Pasteur pipet. Replace the medium in the culture flask, and monitor the outgrowth of cells over the next few days (Fig. 1). Once some 75% of the culture surface has been covered by cells, the culture is ready for passage (see Notes 3 and 4).

8.

9.

10. 11.

12.

3.3. Subculture

and Maintenance

(See Note 5)

1. Remove the medium, placing the culture flask on end and allowing full drainage. 2. Add a wash of 5 mL of CMF to rinse the cells and remove traces of serum. Remove the wash as above. 3. Add 2 mL of trypsin/CMF, and swirl the flask to cover the monolayer with solution. Incubate in trypsin until the cells begin to round up. Do not force the cells to detach by trituration, since this will cause cell clumping. Gentle rocking of the flask is acceptable. It usually takes some lo-20 min to detach 90% of the monolayer. Do not leave the cells in trypsin longer than is necessary (seeNote 6). 4. Add 2 mL of growth medium to inactivate the trypsin. Disperse the cells with gentle and repeated pipetting to provide a single-cell suspension.

Human Primary Cell Strains

17

Fig. 1. A primary culture showing askin explant (dark area), rounded epithelial cells, and stellate fibroblasts. The epithelial cells are the first to appear from the explant, but after 5d, fibroblasts have migrated out to the substratum and are proliferating rapidly. Magnification x 240.

5. Count the cells by hemocytometer or electronic particle counter. Dilute to the appropriate seeding concentration by simply adding the cell suspension to the total volume of medium required for distribution to the culture flasks. This ensures that each flask will contain the same concentration of cells (seeNote 7). 6. Cap the flasks and return to the incubator. Examine after 1 h to check for pH changes. If needed, gas with sterile 5% CO,. Leave in an incubator for 2 h. Open the caps of the flasks slightly to allow the escape

18

Jones

of air from the flasks, which will have expanded with rise in temperature. Recap the flasks and leave in the incubator. 7. To determine when a medium change is required, check the culture every day for a drop in pH. This is usually evidence of depletion of medium by the growing culture. Examine cells under an inverted microscope for signs of cytoplasmic vacuolation, and estimate cell density. 8. If only feeding is required, remove and discard medium. Add an equal volume of fresh medium that has been prewarmed to 37OC. Return the culture to the incubator (seeNote 8). Continue to feed cultures until the flask culture surface is completely covered with cells (a confluent culture). Then repeat the subculture protocol (seeNote 9).

3.4.

CeZZ CZoning

(See Note

10)

1. Clean glass coverslips in an acid wash (O.lN HCl for 10 min), rinse repeatedly in distilled water, and sterilize at 120C for 2 h in a dry oven, or use a microwave oven (2 min) if available. 2. Snap the coverslips to produce small shards of glass. The size will be very variable, and you should select shards of approximately 4 mm2. 3. Place the glass fragments into wells of a 22-mm multiwell plate. Aim to cover about 75% of the bottom surface of each well with glass. 4. Add a solution of plasma fibronectin to jus t cover the glass fragments. Leave for 2 h at room temperature or 30 min at 37OC. 5. Prepare a cell suspension by trypsinization, following the routine methods given for cell subculture. 6. Aspirate off the fibronectin solution (this can be reused for up to 5 h if kept cold). Wash the wells in medium. Add the cells in completemedium, plating at population densities of about 25 tells/22-mm diameter plate. Keep the depth of medium to a minimum. 7. Place the culture in a 37OCincubator, and leave for 2-3 h to allow adhesion and spreading of cells to fibronectin-coated glass. 8. Examine the culture with an inverted microscope. Identify those glass shards with one cell attached. Using forceps, remove selected pieces to 96-well plates. 9. Place a single fragment of glass into each well before adding medium to the well. The glass will stick firmly to the underlying plastic as a result of surface tension (seeNote 11). 10. Add complete medium to the wells, again keeping the total volume to a minimum. Place the multiwell plate in a tissue culture incubator

Human

Primary

Cell Strains

19

11. 12. 13.

14.

15. 16.

after checking again that each well contains only a single cell, and leave for 18-24 h. Examine cultures for signs of colony formation. At this early stage, it is possible to check that each well contains a single colony. Remove glass shards having multiple colonies using forceps (seeNote 12). Attempt colony separation if feasible by snapping the glass and replacing fragments in separate wells. Allow the cells to proliferate until the base of the well is covered with fibroblasts. Use standard trypsin-detachment procedure to isolate cells. Seed clones into standard 25-cm2 tissue culture flasks in complete medium and allow to grow up. Use the standard subculture protocol to passage cell clones (seeNotes 13 and 14). Test clones for genetic contamination, i.e., a multiple-cell origin. We clone cells from carriers of Duchenne dystrophy who are also heterozygous for a second X-chromosome gene, glucose-6-phosphate dehydrogenase Mediterranean variant (G6PD Med) (3). GGPD phenotype is determined according to the method of Ferraris et al. (4). Trypsinize the cells, and pellet them in a conical centrifuge tube. Lyse and sonicate the cells in cell lysis buffer. Precipitate protein with acetone. Dissolve the protein precipitate with 1M NaOH containing 0.4% SDS. Assay for GGPD activity according to Ferraris et al. (4) (seeNote 12).

4. Notes
1. This chapter deals with the culture of cells from dermal biopsy, but in practice, fibroblastic cells can be grown from virtually any biopsy material following this protocol. 2. You are likely to be provided with very small amounts of biopsy with which to begin your culture. Explant culture is particularly suited to this restriction, since there is a great risk of losing cells during enzymic disaggregation. There are two major disadvantages to this technique. Some tissue has very poor adhesiveness of explant to culture surface. This is not a problem for skin or muscle biopsy. Two variations can be tried to overcome this problem if need be. First, try to make your explant size smaller; this is harder than it sounds and may call for the use of a good binocular dissecting microscope. Secondly, increase the percentage of serum in the growth medium from 10 to 20% or more to aid the attachment of the explants to the culture surface.

Jones The second disadvantage relates to cell selection. The technique will only allow culture of those cell populations with good migratory powers; it does not provide a culture that represents themixture of cell types to be found in the biopsy. Skin biopsy explants will show an initial migration of epithelial cells from the explants over the first few days, these cells are then overgrown by fibroblasts (Fig. 1). 3. Always use fetal bovine serum in the culture medium of primary explant cultures. This gives better survival figures than serum from nonfetal sources. Some material proliferate better on culture plasticware that has been precoated with denatured collagen. This can be made in your own laboratory using rats tails as a source to purify gelatin. However, I find that more consistent substrata can be made using a solution of calf skin gelatin (such as Vitrogen), which can be purchased from various sources. Avoid this added complication if at all possible. 4. Despite the rarity value of good biopsy material, I always discard explants that have been removed from the established primary culture. Many sources suggest that these explants can be reused to seed a second round of cultures. Although this is true in theory, I feel that problems over toxins released as a consequence of tissue necrosis within the explant do not justify the effort. 5. Once a primary culture is subcultured, it becomes known as a cell line, partially reflecting the heterogeneous lineages of the explant. As the cell lines proliferate and are subcultured, a selection process will usually occur that narrows the range of cell phenotype within the line. For consistency between experiments and/or between biopsied samples, it is essential to give a code to each cell line established in your laboratory and a record of the number of subcultures performed. For finite cell lines that are derived from biopsies, it has been a routine to reduce the cell concentration at each subculture by two- or fourfold. Under these circumstances, each passage is accompanied by either a 1 or 2x population doubling prior to the next subculture. A record of the number of passages and population doublings must be kept with each cell culture. For example, a cell line will be designated by NHSF (normal human skin fibroblast) upon the first subculture, together with a number to indicate the source of the cell line (NHSF5 will indicate the fifth biopsy taken). We will also add a note of the population doublings, so this will become NHSF5/1. This last number will increase by one for a split ratio of 1:2 (NHSF5/2, NHSF5/3,

Human

Primary

Cell Strains

21

6.

7.

8.

9.

10.

11.

and so on) or by a two for a split ratio of 1:4 (NHSF5/2,NHSF5/4, and so on). Many proteolytic enzymes are available in varying degrees of purity. Crude trypsin, e.g., Difco 1:250, contain other proteases that may be toxic to some cells. Start with the crude preparation, and move on to purer grades only if necessary. You will also need much lower concentrations for shorter exposure periods than those given here for crude trypsin. Once a routine of medium change and subculture is underway, you may wish to replace the 25-cm2 culture flasks with 75-cm2 versions. Do this if the cells proliferate rapidly, and do it early in the life of the cell line. It is good practice to pretest a batch of serum for the capacity to support your cell cultures and, once satisfied, to use the same batch in all your subculture work for a particular experimental program. In this way, you will avoid problems associated with the considerable batch variation betweeen seemingly identical serum. The subculture method given here will eventually provide a series of replicate samples for setting up your experiment. You should not necessarily adhere to this method when preparing a cell population for an actual experimental assay. Cloning of most primary cultures is a difficult and tedious procedure that is only to be undertaken for strong scientific reasons. Normal tissue cells only have a limited number of generations, and by the time that sufficient cells have been produced from a clone, they may be already near to senescence (seeChapter 3, this volume). Low plating densities are needed to ensure successful isolation of clones, but this results in low survival levels in all but a few cell lines. This method attempts to overcome these difficulties by choosing a rich medium, by seeding cells at high densities (some 5-6 times the recommended concentrations) onto small fragments of easily manipulated glass, and by pretreatment of the seeding surface with fibronectin to improve the adhesiveness of the substratum. It is possible to use feeder layers of irradiated cell monolayers, but I find the extra work involved offers no improvement in the success rate over fibronectin coating. I normally do not take much care over checking the orientation of the glass fragment when it is taken to the well of the 96-well dish; the cell can be either face up to the medium or sandwiched against the surface of the well. Sometimes it seems that cells caught in the latter orien-

22

Jones

tation survive better than those facing up. The confines of the microenvironment between the glass and plastic surfaces may allow the cell to create a locally enriched environment that mimics a higher cell density state. If you have problems with plating efficiency, it would be worth experimenting with this observation, perhaps taking care to place the cloning medium into the well before adding the glass fragment in this case. 12. The visual identification of an isolated cell mass should not be taken as conclusive evidence for its origin from a single progenitor cell. With females, it is possible to use individuals who are heterozygous for some X-chromosome-linked gene such as G6PD, clone their cells, and then test individual clones for GGPD phenotype. You should either follow this suggestion to check against genetic contamination of a clone or find another gene that you can assay for in some way. If al2 you need is a clonal population of fibroblasts, you should use females with a well-known allelic polymorphism of an X-chromosome gene. 13. If no cultures appear in the isolates by3 wk, it is very unlikely that they will do so. I suggest that, if you get this result more than once, you should consider using conditioned media in order to establish the clones, Collect the medium from cultures of fibroblast cell lines, centrifuge at 10,OOOg for 30 min at 4C to clarify, and filter the supernatant through a 0.2 pm sterilizing filter. Add this conditioning medium as a 40% v/v fraction to your cloning medium. 14. We have used cloning rings in combination with Petri dishes to generate clones (3). Isolated cells are plated at densities of between 25 and 50 tells/60-mm diameter Petri dish in Hams FlO medium and 15% fetal calf serum. Individual clones were isolated using glass cylinders that are dipped in silicone grease prior to pressing onto the base of the Petri dish to encompass a growing clone and isolate it. Trypsin solution is then added by pipet to the clone, and the dissociated cells are removed to a culture flask. Production of clones was less efficient using this technique, and this led to the development of the method I have described here.

References
I. Jones, G. E. and Witkowski, J. A. (1983) Membrane abnormalities in Duchenne muscular dystrophy. 1. Neuro2.Sci.S&159-174. 2. Jones, G. E. and Witkowski, J. A. (1983) A cell surface abnormality in Duchenne muscular dystrophy: intercellular adhesiveness of skin fibroblasts from patients and carriers. Hum. Genet. 63,232-237.

Human Primary

Cell Strains

23

3. Hillier, J.,Jones,G. E.,Statham,H. E.,Witkowski, J. A., and Dubowitz,V. (1985) Cell surface abnormality in clones of skin fibroblasts from a carrier of Duchenne muscular dystrophy. J. Med. Genet. 22,100-103. 4. Ferraris, A. M., Giuntini, P., Galiano, S.,and Gaetani, G. F. (1981) 2-deoxyglucose6-phosphate utilization in the study of glucosed-phosphate dehydrogenase mosaicism. Am. J. Hum. Genet. 33,307-313.

Chapter Aging

of Cultured Human Skin Fibroblasts

Calvin B. Harley

1. Introduction
There is substantial evidence that aging is related to the finite ability of somatic cells to divide and repair damaged tissue. Since the seminal observation of Hayflick andMoorhead (1) that cultured human fibroblasts have a finite replicative lifespan, a great deal of basic biological research on aging has been based on the model of cellular senescence in vitro (2). There are three ways in which cultured cells can be used in gerontological research. First, cells can be isolated from donors of various ages or, in longitudinal studies, from one donor at various times. Molecular and cellular features of the cells can then be compared at corresponding early times in culture. Differences are related to the age of the donor. Similarly, a second method of investigation involves comparison of cell cultures from individuals with accelerated aging syndromes such as progeria or Werners syndrome to cultures at corresponding passages in vitro from agedmatched normal donors. These types of comparisons do not make use of the in vitro senescence of cells. Thus, in the third model, cultures initiated from normal donors are followed as a function of cell generations in vitro, and comparisons are made between early and late passages of the culture lifespan. Differences seen as a function of in vitro aging may be related to cellular aging within the organism. Each of these methods has advantages and limitations; it is probably best to use more than one (3). 25

26

Harley

Studies on cellular senescence in vitro demand careful attention to the establishment of the primary culture and ongoing estimation of population doublings. These are the points that will be emphasized in this chapter. For example, periodically throughout the lifespanof a culture, the number of viable cells used as the inoculum (NJ, the fraction that attaches (f,>, the fraction of attached cells that divide (f,), and the cell number at confluence (NC) should be measured (4). These numbers determine the incre mental cell doublings within one passage, provided there are not dramatic changes in other parameters such as cell size and interdivision time:

Doublings = logJ@( - (~-fd)f,N&)/f~dNol

(1)

In practice, however, the fraction of normal skin fibroblasts that attach and divide is close to 1 for most of the lifespan of the culture, and these effects are ignored or assumed in the term MPD (mean population doublings):
MPD = 10g,(N,lNJ = log,0 /split ratio)
(2)

Thus, if cells are split at confluence at a 1:4 or 1:8 ratio, for example, the number of mean population doublings to refill the dish is simply 2 or 3, respectively. It should be realized that this estimate of cell generations can be substantially in error, especially toward the end of the culture when the fraction of cells that attaches and divides becomes significantly less than one, and the number of cells at confluence declines. Other chapters in this volume provide greater detail on maintenance of cell strains and special techniques relevant to specific cell types. Cells other than fibroblasts have in fact been used in studies of aging in culture (2). Themeth o d s outlined below for careful determination of mean population doublings can be applied to any cell growing on a solid matrix.

2. Materials
1. Regular growthmedium (RGM): a-minimal essential medium (GIBCO) supplemented with 15% fetal calf serum. Store the components as indicated by suppliers. Prepare bottles of RGM as needed, and keep at 4C (up to 2 mo) except during use, when it should be prewarmed to 37OC. It is important not to change lots of serum. Therefore, order a small amount of two or three lots, asking for an appropriate number of bottles of each to be kept on hold. Test for optimal growth of cells at high and low cell density, and then place a large order for the best batch (Notes 1,2).

Cellular Aging In Vitro

27

2. Ca2+-and Me-free phosphate buffered saline (PBS): 0.14M NaCl, 2.7 mM KCl, 1.5 mM KHJ?O,, 8.1 mM Na,Hl?O,. 3. Trypsin: Crude Difco or Sigma trypsin at 0.125% (w/v> in PBS. Prepare 5-10 mL aliquots, and store at -2OOC. Thaw as needed. Unused portions of the aliquot may be refrozen and used once or twice more (Note 3).

3. Methods
3.1. Primary Culture
1. A standard 2-4 mm punch biopsy of epidermis plus dermis from the abdominal or inner forearm surface is taken sterily by a qualified person (Note 4). 2. Place the tissue piece into a loo-mm Petri dish containing a small amount of RGM. Cut it in two, and place one-half in a vial containing 10 mL of RGM. Place this vial at 4OCas a backup sample. It will remain viable for many days. 3. Dice the remaining half into pieces about 1 mm3 or less in size, and place three pieces dermis-side down (if possible) in each of several 35mm Petri dishes. Use a siliconized Pasteur pipet to transfer the tissue. Place a sterile 25-mm coverslip over the skin pieces as shown in Fig. 1, add 2 mL RGM, and incubate at 37OC. 4. Monitor the explants for cell growth from around the edges of the tissue every day or two. Refeed weekly. The first cells to appear are the irregular, closely packed keratinocytes, which terminally differentiate and die inRGM. Spindle-shaped fibroblasts appear within several days. 5. When a dish has a total of about 1 cm20f fibroblast outgrowth (2-4 wk), loosen and invert the coverslip, aspirate the RGM, rinse well twice with PBS, and add 0.3 mL trypsin. Tilt the dish several times until trypsin covers all cells. Set at 37OCfor 10 min or until cells start to detach from the dish and/or coverslip. It may be necessary to give the dish one or two gentle shakes during the digestion period. 6. Add 1.5 mL RGM, pipet up and down to suspend cells, remove the coverslip and tissue debris, and count an aliquot of the cell suspension. Pool cells from several dishes. 7. Estimate the mean population doublings (MPD) in the primary culture as shown in Table 1.

28

Harley

25mm coverslip

silicone grease

tissue

Fig, 1. Establishing a primary culture. Tissue fragments are transferred with a siliconized pipet containing a drop of medium into a dry 35mm dish. A dab of sterile silicone grease (e.g., high vacuum lubricant) holds a 25-mm coverslip over the tissue fragments. Medium is then carefully added to the dish. The tissue fragments anchor themselves to the dish and/or coverslip, and cells begin to grow out onto these surfaces usually within l-2 wk.

8. Place the remaining cells into fresh 35- or 60-mm dishes. Record on the lid the date, strain designation, and expected MPD at confluence. The latter is obtained by adding to the MPD of the primary culture 3.3*log (NC/No), where Ncis th e expected number of cells at confluence (Table 2) and IV0 is the initial number of cells (Note 5).

3.2. Secondary

Culture

and Subsequent

Passages

1. When the culture just reaches confluence, aspirate the medium, rinse once with PBS, add trypsin (Table 2), and incubate at 37C for 10 min

Cellular Aging In Vitro

Table 1 Estimated MPD During Primary Culture* Cell number/dish MPD 6 30,000 60,000 7 120,000 8 250,000 9 500,000 10 *This estimation assumesthat about 500 cells initiate outgrowths from the tissue fragments in one 35-mm dish. Some heterogeneity in population doubling is introduced into the culture because of density-dependent inhibiton of growth (4). Table 2 Tissue Culture Dishes 35 mm 60 mm 8 21 5x105 1.3 x lo6 0.3 0.6 1.7 3.4

Surface area (cm21 Cells at confluence (NJ Trypsin (mL) RGM (mL)

100 mm 55 3x106 1.0 9.0

or until cells loosen from the plate. It is often necessary to shake or rap

the dish gently. 2. Add RGM (Table 2) and pipet up and down until all cells are freed from the dish and a single-cell suspension is achieved. Count an aliquot of this suspension, and inoculate a fraction of it into a fresh dish (or dishes) containing an appropriate volume of RGM. The inoculum should represent l/8 to l/2 the number of cells required to form a monolayer of cells in the new dish. Shake the dishes gently in an irregular fashion to ensure a homogeneous dispersion of cells. Record date, strain designation, and expected MPD at confluence. 3. Refeed the culture at least once per week until confluence is again reached.

3.3. Senescence

(Phase II.

1. Visually monitor cells at each passage and estimate growth rate by days to confluence from a standard inoculum size (e.g. l/8 of final confluent cell number). The culture is approaching its terminal passage when growth rate drops and cells become larger and irregular in shape (Fig. 2).

30

Harley

Fig. 2. Cultured human fibroblasts from an adult forearm skin biopsy. The clean, spindle appearance of cells aligned in parallel or spiral arrays at confluence during early passage (A) is gradually replaced in late passageby nonaligned, irregular-shaped cells containing opaque degradation products (B). Early-passage cultures contain numerous mitotic cells (A, arrow) during log phase and early confluence, whereas terminal passage cultures have essentially no dividing cells and, thus, fail to reach confluence even after numerous weekly refeedings.

Cellular Aging In Vitro

31

2. Senescence, also known as MI&, Phase III, or terminal passage, is the point at which one MPD takes longer than one week. For example, if a culture split at a 1:8 ratio is not confluent in 3 wk, it may be considered terminal.

4. Notes
1. Do not use dialyzed or heat-treated serum for routine culturing of normal human fibroblasts. However, fibroblasts are fairly tolerant of media changes for short-term culture, and it is often necessary to use defined or special media lacking certain components for biochemical studies. To ensure that cells are not adversely affected, conduct a growth curve in regular vs experimental medium, or, if less than one population doubling is involved, assay the rate of protein or DNA synthesis. For incorporation assays, it is important that the specific activity of the precursor be the same in each medium tested. 2. Avoid the use of antibiotics and antimycotics, since they may have unexpected effects. They are not needed if sterile technique is diligently practiced. 3. Crude preparations of trypsin may contain other proteases useful in dissociating cells. However, they also vary in activity. Therefore, adjust the concentration up (or down) by a factor of 2 if cells take more than 15 min (or less than 5 min) to detach from the dish. 4. Fibroblast cultures can be initiated from biopsies taken almost any where on the body. Considerations include sun exposure, environmental variation, nature of the study, and compliance from donor to donor. Abdominal and the non-sun-exposed forearm surfaces are commonly used, as are foreskins and other surgical pieces of skin. 5. If cultures or ampules of cells are supplied, one must estimate the previous mean population doubling of the culture based on data from the supplier. Assume 9 MPD for the primary culture and 1-3 MPD for each passage depending on whether cells were previously split at 1:2, 1:4, or 1:8 ratios. Upon arrival, subculture the cells at a split ratio of 1:2 into fresh dishes and add an appropriate number of MPD to this estimate based on how many cells appear to have attached after 6-12 h. For example, if about 1/2of the cells attached, count this initial 1:2 split as 2 MPD. 6. Biopsy fragments can be disaggregated with proteases, and cell suspensions rather than outgrowths are then used to establish the primary culture. However, this often requires extensive proteolytic

32

Harley treatment to liberate cells, and it is difficult to estimate the number of viable founder cells of the fibroblast culture. The advantage of this is that it eliminates the heterogeneity in doublings within the primary cell population, which arises from density-dependent arrest of cells in the outgrowth technique (4). However, this effect is reduced in the technique described here by harvesting the primary culture before outgrowths become large and heterogeneous. Values of NC (Table 2) reflect an average value for human forearm skin fibroblasts. Each laboratory should determine N, and N, periodically throughout the lifespan of each strain studied and use these values for MPD calculations, which are not based on valid split ratio methods. In reporting comparisons of early and late passage cells, it is useful to report culture age as percent lifespan completed, i.e., MPD/MPDmax , together with the value of MPDmax for the culture. Various criteria have been used to define senescence, or Phase III. It is wise to monitor the fate of the terminal passage cells to determine if additional MPD occur. However, we have not been able to further passage cells that fail to reach confluence at a 1:8 split ratio within 3 wk. Other methodology relevant to the culture of human skin fibroblasts has recently been described (5).

7.

8. 9.

10.

Acknowledgments
This article was prepared while the authors work was supported by MRC (Canada) and the Natural Sciences and Engineering Research Council (Canada). I would like to thank Sam Goldstein and Elena Moerman for introducing me to tissue culture techniques.

References
1. Hayflick,L.andMoorhead,P.S.(1961)Thelimitedinvitrolifetimeofhumandiploid cell strains. EXQ. Cell Res. 25,585-621. 2. Stanulis-Praeger, B. M. (1987) Cellular senescence revisited: A review. Mech. Aging Devel. 38,1-48. 3. Harley, C. B., Pollard, J. W., Chamberlain, J. W., Stanners, C. I., and Goldstein, S. (1980)Proteinsyntheticerrorsdonot increaseduringagingof culturedhumanfibroblasts. Proc. Natl. Acad. Sci. 77,1885-1889. 4. Harley, C. B. and Goldstein, S. (1978) Cultured human fibroblasts: Distribution of cell generations and a critical limit. 1. CeEI. Physiol. 97,509-515. 5. Moerman, E. J. and Goldstein, S. (1986) Culture of human skin fibroblasts, Methods in Diabetes Research, Vol. II (Clarke, W. L., Lamer, J., and Pohl, S., eds.), Wiley and Sons, New York, pp. 283-312.

Chapter

Separation and Maintenance of Primary T and B Lymphocytes

Derek Kinchington and Eleanor Berrie
I. Introduction
Two distinct populations of lymphocytes have been identified: T lymphocytes, which are thymus-dependent, and B cells, first observed in the Bursa Fabricus of birds. Mammals do not have an equivalent structure, and there are varying opinions as to the similarity of these cells between species. In humans, current theories are that B lymphocytes differentiate in the fetal liver and in the bone marrow of adults. Human T and B cells are most easily obtained either from peripheral blood or from biopsy of lymphoid tissues (lymph nodes, spleen, Peyers patches from gut, tonsils, and adenoids). To establish T and B lymphocytes from clinical material requires three steps: separation from blood or other tissues, enrichment for either B or T cells, and finally the maintenance of the primary cultures. It may be necessary to carry out several enrichment steps to get greater than 90% purity. These methods will be described in turn.

33

34

Kinchington

and Berrie

2. Materials

and Equipment

1. 2-aminotrylisotriouronium bromide hydrobromide (AET) treatment of Sheep Red Blood Cells (SRBC): Dissolve 402 mg AET in 10 mL distilled water and adjust the pH to 9 with 1M NaOH. Do not overshoot. Mix one packed volume washed SRBC in 4 vol AET, incubate at 37C for 25 min, and mix regularly. Wash five times in saline, and then resuspend in RPM1 and Hepes buffer. This will keep for 1 wk at 4C. 2. B Cell Growth Factor (BCGF). 3. Clostridium perfringens neuraminidase. 4. EBV suspension, 5 x lo4 Transforming Units (TFU). 5. F(ab), fragment of goat anti-human IgM (p-chain specific). 6. Ficoll-paque. 7. RPM1 growth medium contains 5,10, or 15% fetal calf serum, 2 mM Lglutamine, 100 pg/mL penicillin, and 100 mg/mL streptomycin. 8. Growth medium containing 10% FCS, and 2-Mercaptoethanol (2 x lo4 M). 9. Hanks balanced salt solution pH 4 (seeAppendix). 10. Helix pomatia Lectin-Sepharose 6MB. 11. Heparin. 12. IgG: antihuman. 13. IgG: Pan T or Pan B. 14. Interleukin 2,10-40 U/mL final cont. (seeNote 3). 15. Nylon wool: sterile. 16. Pharmacia Column Kg/15 fitted with an BO-pm mesh net. 17. Phosphate Buffered Saline (Dulbecco A) (PBSA). 18. PBSA containing 0.2% human serum albumin and 0.02% sodium azide. 19. Phytohemagglutinin (PHA) (1 mg/mL stock). 20. Protein A-Sepharose 6MB( Pharmacia). 21. Trypan blue 0.1%.

3. Methods 3.1. Separation

from Blood of hmphocytes Using Ficoll

Commercial products such as Ficoll-paque mixtures are available to separate human mononuclear cell populations contaminated with RBC and granulocytes. 1. Collect blood into tubes containing 10 l.tg/mL preservative free heparin.

Primary

T and B Lymphocytes

35

2. Dilute blood 1:2 in serum-free medium. 3. Layer the diluted blood carefully onto the Ficoll-paque using a Pasteur pipet to produce a clean interface between the two layers. To obtain the maximum yield, it is advisable to keep the proportion of blood to Ficoll in a ratio of 1:3. If small volumes of blood are being separated, an ll-mL centrifuge tube is used with 3 mL blood layered onto 7 mL Ficoll-paque. 4. Centrifuge at 2000 rpm for 25 min at room temperature. 5. Collect the white opaque mononuclear fraction from the interface between the diluent and the Ficoll-paque, and add at least 5 vol of serumfree medium. Centrifuge at 1500 rpm for 10 min. Repeat for two more times to remove the Ficoll-paque, which can be toxic to cultured cells. 6. Count the lymphocytes using Trypan blue exclusion and culture in RPMI containing 10% FCS (Fig. 1).

3.2. Separation of Lymphocyte Suspension fkom Biopsied Material

1. Collect biopsy material in a sterile manner. 2. Wash off any contaminating blood using sterile phosphate-buffered saline (PBS). If the material is grossly contaminated, antibiotics can be added at this point. 3. Remove any surface fat, and then chop the tissue into small pieces about 1 mm diameter. Two sterile scalpels are used for this operation, and make sure that the tissue is cut and not torn. 4. To prepare a cell suspension, two methods can be used: (a) place the tissue into sterile stomacher bags with a small volume of medium. Agitate for 10 s and then collect the suspension through a sterile gauze filter, (b) press the soft tissue with a pipet tip against a sterile sieve, and collect the resultant cell suspension. 5. If the cell suspension is grossly contaminated with blood, this can be separated from the lymphocytes by density centrifugation using Ficoll-paque (seesection 2.1.). 6. Count the lymphocytes using Trypan blue exclusion and culture in RPM1 containing 10% FCS.

3.3. T and B Cell Enrichment Methods Using Nonimmunological Substrates

Having obtained a separate mononuclear cell fraction, it is necessary to distinguish and separate the T and B lymphocyte populations. Techniques based on different adherence properties are used.

36

Kinchington

and Berrie

Fig. 1. Phase contrast photomicrograph of human lymphocytes prepared using a Ficoll-paque gradient. The preparation also contains some cell debris and a small number of monocytes. (Mag: 440x).

3.3.1. Rosetting
T lymphocytes spontaneously form rosettes when added to sheep erythrocytes, which are then separated from B cells by centrifugation. 1. Count viable lymphocytes using Trypan blue exclusion, and adjust to 5 x 10 lymphocytes/ml. 2. Adjust freshly prepared ART-treated sheep erythrocytes (seeMaterials section) to a 2.5% v/v suspension in PBS. 3. Mix lymphocyte suspension, fetal calf serum, and sheep erythrocytes in a 1:l:l ratio, and centrifuge at 1000 rpm for 5 min at room temperature. 4. Incubate for 1 h at 4OC. 5. Resuspend this cell mixture by gently tapping the tube; mix 50 j.tL of Trypan blue with 50 PL of the cell suspension and pipet sample into a 0.2~mm deep hemocytometer.

Primary T and B Lymphocytes

37

6. Count 200 lymphocytes and determine the percentage of rosettes, i.e. cells with three or more attached erythrocytes. 7. Layer the rosetted suspension onto Ficoll-paque (3 mL:7 mL, respectively, see section 2.1.) and centrifuge at 2000 rpm for 10 min. 8. Harvest the nonrosetted B cells at the interface. Rosetted T cells will be spun down to the base of the tube. These may be obtained by removing and washing the rosettes; the erythrocytes are lysed with sterile distilled water or high salt.

3.3.2. Lymphocyte Fractionation Lectin-Sepharose

Using Helix Pomatia 6 MB

T cells possess a receptor for Helix pomatia, and this property is used to bind T lymphocytes to a Sepharose column (1). 1. Treat human peripheral lymphocytes (12.5 x lo6 cells/mL) with neuraminidase; 5 bg/mL in Tris buffered Hanks solution pH 7.4 for 45 min at 37OC (2). 2. Wash the lymphocytes free of neuraminidase with PBS containing 0.2% human serum albumin and 0.02% sodium azide. 3. Pipet lo8 cells in 1.5 mL of the PBS buffer onto a 3-mL column of Helix pomatia lectin-Sepharose 6MB equilibrated with the same PBS buffer, and leave for 15 min at room temperature. 4. Add 80 mL of the PBS buffer to the column at a flow rate of 6-10 mL/ min; this elutes unbound cells (enriched for B cells). 5. Then add 80 mL of PBS buffer containing 0.1 mg/mL N-ace+-A-Dgalactosamine; this releases a population of weakly bound cells (mixture of T and B cells). 6. Finally, add 80 mL of buffer containing 1.0 mg/mL N-acetyl-a-D-galactosamine, which elutes the enriched T-lymphocyte fraction. Total cell recovery is about 70%.

3.4. Nylon-Wool

Columns

At 37C in the presence of serum, B cells will bind preferentially to a nylon-wool column. Most of the T cells and null cells remain unbound and pass through the column. This method is rapid and is suitable for the separation of large numbers of cells (3). 1. Pack 600 mg of sterile nylon wool into a 20-mL plastic syringe barrel (a lOO-mL syringe can be used for large-scale preparation), and wash with growth medium containing 5% FCS. 2. Seal the column and incubate at 37OC for 1 h.

38

Kinchington

and Berrie

3. Rinse the column with 5 mL of warm growth medium containing 5% FCS. 4. Add 2 mL of the cell suspension dropwise to the top of the column. Then, add 1 mL of the warm growth medium, and ensure that there are no air bubbles in the column. 5. Seal the column and incubate at 37OC for 45 min. 6. Wash the column with 25 mL of warm growth medium. Collect the eluent, which will consist predominantly of T lymphocytes and null cells. The lymphocytes are concentrated by centrifugation (1000 rpm for 10 min at 4C), and cell viability may be determined by Trypan blue exclusion. Some of the B-cells adhering to the column may be recovered by mechanical elution: 1. Wash the column with 100 mL of warm growth medium and discard this volume. 2. Seal the syringe tip with a plastic cap. 3. Add 2 mL of warm growth medium, and squeeze the nylon wool with blunt stainless-steel forceps. 4. Remove the plastic cap, wash with 10 mL of warm growth medium, and retain. Replace the syringe piston and expel the remaining growth medium. 5. Concentrate these cells by centrifugation (1000 rpm for 10 min at 4OC). 6. The number of viable lymphocytes is assessedby Trypan blue exclusion. These cells will be mostly B lymphocytes, but contaminated by some T lymphocytes and null cells.

3.5. T and B Cell Enrichment Immunological Binding

Lymphoid

Using Specific Methods

differ-

cells are separated from each other by exploiting

ences in the molecules expressed on their surfaces; for example, immunoglobulin molecules, histocompatibility or blood group antigens, and cell surface receptors. Immunoabsorbents consisting of a specific antibody coupled to some form of matrix, either macrobeads (column) or plastic, are used. 3.5.1. Enrichment Using Antihuman IgG Coated Petri Dishes 1. Pour4 mL of 1% antihuman IgG in to a sterile Petri dish, and leave for 1 h at 4C. 2. Rinse the coated Petri dish twice with PBS.

Primary T and B Lymphocytes

39

3. Add IO8 cells (maximum) in 4 mL of growth medium containing 5% FCS. 4. Incubate for 1 h at 4C. 5. Gently swirl the dish several times during the incubation, but keep it level. 6. Remove the medium containing thenonadhering Tcells using a sterile Pasteur pipet. 7. Gently rinse the dish with growth medium containing 5% FCS and discard. 8. Add 4 mL fresh medium, and vigorously aspirate to obtain adherent B cells. 9. Count thedifferent lymphocytesuspensionsusingTrypanblueexclusion, and prepare them for culture.

3.5.2. Separation

Using a Protein A-Sepharose

6MB Column

Protein A binds to the Fc portion of IgG and, when coupled to Sepharose 6MB, will be adsorbed by cells that have been coated with a specific IgG antibody. The immobilized cells are then released by adding excess soluble IgG, which displaces the antibody-coated cells from theProtein A-Sepharose 6MB column (4). 1. Take lo8 cells/mL from a separated lymphocyte culture (seesections 2.1 and 2.2) and treat either with pan B or pan T antibody (20 pg/107 cells/mL) at 4OC for 30 min. 2. Wash the cells three times with RPM1 containing 10% fetal calf serum. Spin down and resuspend in 0.25 mL of the same medium. 3. Pipet the treated cells in to a small plastic column (0.5 cm in diameter) containing 1.5 mL Protein A-Sepharose 6 MB, close the column, and incubate at room temperature for 20 min. 4. Add 20 mL medium at a flow rate of about 2.5 mL/min to remove the unbound cells. 5. Then add 2 mL of dog IgG (20 mg/mL), close the column, and incubate at 37C for 60 min. 6. Elute the cells by adding 3 mL of dog IgG (20 mg/mL) followed by 15 mL of buffer, both at 37C. Approximately 60% of the bound cells are recovered.

3.6.

Establishment

of Primary

Lymphocyte

Cultures

Primary lymphocytes can be maintained by mitogens, e.g., plant lectins phytohemaglutinin (PHA), Concanavalin A (Con A), and allogenic transplantation antigens and feeder layers; or by the addition of specific growth factors to culture media, e.g., IL2 (seeNote 3).

40

Kinchington

and Berrie

1.

2. 3. 4. 5. 6. 7. 8.

3.6.1. Maintenance of T-Lymphocytes Using IL2 Dilute the enriched T-cell suspension (with growth medium containing 10% FCS) to give a density of 5 x 1O5 -1 O6 cells /mL. Depending on the volume, place either 200 PL aliquots in a microtiter plate, 1 mL/ well in a 24-well plate, or 3-7 mL in a 25-cm* flask. Add PHA to give a final concentration of 2 pg/mL. Incubate at 37OC in a humidified 5% CO,/95% air incubator. Observe cell cultures daily for cell growth and evidence of cell death. After 3-4 d, count the cells, using Trypan blue exclusion, harvest cells, and wash (centrifuge at 1000 rpm for 5 min). Gently resuspend the pellet in fresh growth medium (RPM1 plus 15% FCS) supplemented with IL2 (lo-40 U/mL; see Note 3) at a cell concentration of 5 x l@-lo6 cells/mL, and culture but do not add PHA. Continue to adjust the cell concentration every three days, adding fresh medium (plus IL2) each time. After 2 wk, the individual T cells may be cloned and separated for functional tests.

3.7. Maintenance to lkansform

of B-Lymphocytes Using EBV CelZs (See also Chapter 5)

1. Take 2 x 1O6 of enriched B-lymphocytes and pellet (1100 rpm for 10 min at room temperature). 2. Resuspend the cells in 1 mL of EBV suspension (5 x lo4 TFU), and incubate at 37OC for 1 h in a humidified 5% CO,/95% air incubator. 3. Wash three times with PBSA to remove unadsorbed virus, and resuspend in growth medium containing 10% FCS. 4. Culture in 24-well plates at a cell concentration of lo6 cells/mL of growth medium at 37OC in a humidified 5%CO,/95% air incubator. 5. Check for proliferation after 7 d. 6. Remove medium and cells from wells, spin down, and gently resuspend transformed cells in fresh medium. Continue to culture in 24well plates. 7. These cells can be grown inlimiting dilution for B-cell characterization with monoclonal or polyclonal antibodies or functional assays.

3.8. Maintenance
Using B-Cell Growth

of B-CeZZs
Factor

Recently, B cells from peripheral blood from normal donors grown in the presence of B-cell growth factor (BCGF) and anti-IgM has led to the establishment of a long-term cultured human B cell line (5).

Primary

T and B Lymphocytes

41

I. Pipet out 1 mL enriched B cells (106cells/mL) into 24-well plates with growth medium (plus 10% FCS) containing 20 &mL F(ab), fragment and 10% BCGF. 2. Culture cells at 37C in a humidified 5% CO,/95% air incubator. 3. Feed cells every 3-4 d by replacement of 50% of culture-well contents with an equal volume of fresh medium containing 10% BCGF and 20 pg/mL anti-IgM. 4. Check for cell growth and viability using Trypan blue exclusion until cell culture is established.

4. Notes
1. Optimal growth conditions should be established with T cell lines before embarking on experiments using the valuable material just prepared, i. e., check that the sterile lx medium will maintain T cell lines. 2. Fetal calf serum may vary from batch to batch: set up established T or B cell lines in culture, count cells on a daily basis by the Trypan blue exclusion method, and check which serum gives maximal cell growth, or rH]-thymidine may be used to measure cell division rates. 3. IL2 (T-cell growth factor) also shows variation and ideally should be assayed on established cell lines: increased growth is monitored by increased [3H] thymidine uptake.

References
I. 2. 3. 4. 5. Cell Affinity Chromatography: Principles and Methods, Pharmacia Booklet (1984) Wharmacia Ltd., Midsummer Blvd., Milton Keynes, MK9 3HP, England). Hellstrom, U., Dillner, M. L., Hammarstrom, S., and Perlmann, I. (1976) The interaction of nonmitogenic and mitogenic lectins with T-lymphocytes: association of cellular receptor sites. Scf272d. J. Immunol. 5,45-55. Hudson, L. and Hay, F. C. (1980) Practical Immunology (Blackwell Scientific Publications), pp. 212,213. Ghetie, V., Mota, G., and Sjoquist, J, I. (1978) Separation of cells by affinity chromatography on SPA-Sepharose 6MB. J. Immunol. Methods 21,133-141. Hayao, A., Rossio, J. L., Ruscetti, F. W., Matsushima, K., and Oppenheim, J. J. (1986) Establishment of a human B cell line that proliferates in response to B cell growth factor. J. Immunol. Methods 90,111-123.

Chapter 5 Establishment of Lymphoblastoid Cell Lines

A. Doyle
1. Introduction
The ability to establish long-term B lymphocyte cultures from patients carrying particular genetic characteristics or with the ability to secrete specific antibodies (I) is an extremely valuable technique. However, there are several basic principles to follow in the approach to this technology. It is the purpose of this chapter to provide all the information necessary to run an efficient and safe Epstein-Barr Virus (EBV) transformation system. More detailed information regarding the use of this technique for thepreparation of human monoclonal antibodies is given elsewhere (2).

2. Materials
1. B95-8 marmoset cell line. 2. RPM1 1640 culture medium containing lo%, 5%, and 2% FBS. 3. Culture medium (without FBS) RPM1 1640 containing lOU/mL hepaarin (preservative-free), plus penicillin, streptomycin, and Polymixin B sulfate. 4. 1 mL B95.8 cell culture supernatant. 43

44

Lymphoblastoid

Cell Lines

5. 10% FBS RPM1 1640 culture medium, plus 0.5% PHA Wellcome. 6. Ficoll-hypaque or lymphroprep.

3.1. Production

3. Methods of Epstein-Barr

Virus Stock (3)

1. The B95.8 cell line (Note 1) is cultured routinely in 5% FBS RPM1 1640. Routine passage consists of diluting the cells between 1 in 2 and 1 in 5 at weekly intervals. A culture in the logarithmic growth phase would usually yield lo6 cells/ml at 37C. 2. For preparation of virus stocks, dilute from 106/mL to 0.2 x 106/mL in 2% FBS RPM1 1640. 3. Incubate at 33C for 2 w without any medium additions or changes. 4. Allow the cells to settle out at 4OC. 5. Centrifuge the supernatant at low speed (1508 for 5 min) in order to clarify. 6. Pass the supernatant through a 0.2 ym filter. 7. Aliquot the filtered supernatant in 2 mL sterile plastic ampules, and store either short-term (1 mo) at -20C, medium-term (6mo) at -7OC, or preferably in a gaseous phase liquid nitrogen container at -13OOC. 8. A sample of each batch should be examined for microbiological purity (bacteria, fungi,mycoplasma) and also tested for transforming ability.

3.2. EB-Virus Transformation: Blood Sample Preparation

1. The heparinized blood is mixed 1:l with RPM1 1640/heparin (see Note 2). 2. Layer the diluted blood sample onto Ficoll-hypaque at a ratio of 21, i.e., 8 mL Ficoll-hypaque/4 mL blood, in a 15-mL conical centrifuge tube. Care must be taken to prevent the two solutions from mixing. 3. Centrifuge at room temperature at 300g for 20 min. 4. Carefully remove the buffy coat interphase with a Pasteur pipet. This contains the mononuclear cell fraction. 5. Add RPM1 1640/heparin to the cells to make thevolume up to 15 mL. 6. Centrifuge at 1508 for 4 min. 7. Pour off the supernatant, and resuspend the lymphocyte pellet in RPM1 1640/heparin. 8. Repeat this procedure twice more. IMPORTANT: It is essential to ensure that the lymphocyte pellet forms after each centrifugation step. If not, dilute the sample further (step 5 or recentrifuge). Avoid centrifuging too hard; otherwise platelets will sediment.

Doyle

45

9. Resuspend the mononuclear cells in 5 mL of RPM1 1640/heparin. 10. Mix 0.2 mL of cell suspension with 0.2 mL Trypan blue (0.4% in PBS), and count the viable cells on an improved Neubauer counting chamber. 11. At this point, cells can be set up immediately for transformation or stored in liquid nitrogen (seeNote 3).

3.3. Cell-Freezing

Procedure

1. Centrifuge the cell suspension at 1508 for 4 min. 2. Resuspend the pellet in the correct vol of 91% newborn calf serum (NBCS) + 9% DMSO to achieve 5 x lo6 cells mL. 3. Freeze in a programmable freezer at -3C/min. Alternatively, freeze overnight in a polystyrene container in a -8OOCrefrigerator and then remove ampules to -196OC.

3.4. Virus
Either:

Transformation

1. Carefully thaw the ampule in a 37OCwater bath, and transfer contents to 10 mL of 10% FBS RPM1 1640 to remove DMSO. Centrifuge at 1OOg for 5 min, or 2. Centrifuge cells in RPM1 1640/heparin at 1508 for 4 min. Then: 3. Resuspend cell pellet in 1 mL EBV (B95.8) supernatant per 5 x lo6 cells. 4. Incubate at 37OC for l-1.5h. Agitate the suspension at least once during the incubation period. 5. Centrifuge at 1OOg for 5 min. 6. Discard supernatant. 7. Resuspend lymphocytes in an appropriate vol of 10% FBS RPM1 1640 + penicillin/streptomycin + 0.5% PHA (Wellcome) to give lo6 cells/ mL (seeNote 4). 8. Transfer 1 mL of cell suspension to each well of a 24-well tissue culture tray (seeNote 5). Incubate at 37OCin 95% sir/5% CO, for a minimum of 5 d before changing the medium. 9. With smaller cell numbers, 0.2 mL of cell suspension can be transferred to each well in a 96-well plate (seeNote 6). 10. Cells can be transferred sequentially from a 96-well tray to a 24-well tray to a 25-cm2 flask over the next 14-21 d. This depends upon the rate of appearance of cell growth in the wells. The technique can be amended if particular problems are anticipated with cells. SeeNotes 4-7.

Lyrnphoblastoid

Cell Lines

4. Notes
1. A major consideration is the source of EBV. Most laboratories currently use the persistently infected Marmoset cell line B95.8 (4) as a source of virus, although in some cases cocultivation with the cell line QIMR-WIL has been employed. Unfortunately, many samples of the B95.8 cell lines distributed informally between research laboratories are mycoplasma contaminated. It is therefore of the utmost importance that the basic starting material for the technique, the virus source, is obtained from a reliable source, such as an established culture collection. 2. Blood samples (5 mL min.) should be transferred from the syringe immediately into preservative-free heparin (final concentration, 10 U/ mL). If the samples are to be stored prior to preparation, this must be done at room temperature. Storage at 4OC leads to rapid cell death. The maximum period of storage is 4 d. 3. There are two approaches to EBV-transformation: either cells can be transformed immediately or, if this is not convenient, the cells, once separated, can be stored in liquid nitrogen until transformation. It is recommended that all procedures should be carried out in a Class II containment cabinet. 4. The mononuclear cell suspension prepared by Ficoll-hypaque separation contains both B + T cells. There are several approaches to the removal of T cells necessary to prevent specific cytotoxic T cell killing of B cells carrying EBV antigens (seeChapter 4, this volume for details). The alternatives are: a. Rosetting with sheep red blood cells to deplete the T cells from the suspension. This has the disadvantage of reducing the B cell yield and contaminating T cells may remain. b, Cyclosporin-A can be added to kill T cells. This has the disadvantages of requiring assay before use, has carcinogenic properties, and needs to be dissolved in ethanol for use. It may be difficult for some laboratories to obtain supplies. It can have certain advantages in microtechniques. c. PHA can be added. This results in the proliferation of T cells and has a mitogenic effect on B cells. There is no evidence of specific T cell killing following PHA stimulation. In our laboratory, this has proved to be the most efficient technique available: hence, it is reproduced here. (One possible drawback is that mitogenesis leads to differentiation of committed B cells to

plasma cells and then the loss of the C3 receptor, and thus, the cells are not transformed.) 5. Mouse peritoneal macrophages can be prepared as feeder-layers (at 2 x lo4 cells/well) in a 24-well plate 24 h prior to cell transformation. 6. FCS concentration can be increased to 20% with small cell vol. 7. It is important to test the cell line for mycoplasma following establishment in culture. It has been noted inour laboratory that, even with precautions taken, i.e., pretested virus and media components, mycoplasma can be introduced with the original blood sample. Finally, it is advisable to test staff for antibody to EBV (VCA test) prior to commencement of work with the virus. If immunity is not present or is at a very low level, additional caution should be exercised in the handling of this agent. However, over 98% of adults aged 23+ have antibody to VCA.

References
1.
2. 3. 4.
Stein&, M., Klein, G., Koskimies, S., and Makel, 0. (1977) EB virus-induced B lymphocyte cell lines producing specific antibody. Nature 269,420. Crawford, D. H. (1986) Use of the virus to prepare human-derived monoclonal antibodies, in The Epstein-Barr Virus: RecentAdz~ances (Epstein, M. A. and Achong, B. G., eds.), William Heinemann Medical Books, London, p. 249. Adams, A. (1975) EBV Production, Concentration and Purification (Ablushi, D. V., Aalesed, H. G., and De The, G., eds.), IARC, Lyon, France, p. 129. Miller, G. and Lipman, M. (1973) Release of infectious Epstein-Barr virus by transformed marmoset leucocytes. Proc.Nutl. Acud. Sci.USA 70,190.

Chapter 6 Scale-Up of Suspension and Anchorage-Dependent Animal Cells

J, Bryan Griffiths

1. Introduction
In this chapter, scale-up is described in a laboratory context (lo-20 L), but the principles and techniques employed have been successfully adapted so that cells are now grown industrially in unit volumes of up to 8000 L for vaccine, interferon, and monoclonal antibody production. The need to scale-up cell cultures has been expanded from the historical requirement for vaccine manufacture to include not only interferon and antibodies, but many important medical products such as tissue plasminogen activator and a range of hormones and blood factors. The low productivity of animal cells, resulting from their slow growth rate and low expression of product, plus the complexity of the growth conditions and media, led to attempts to use recombinant bacteria to express mammalian cell and virus proteins. However, this has proven unsuitable for many products, mainly because of incomplete expression and contamination with bacterial toxins, and more importance is now being put on expression of recombinant proteins from mammalian cells. This has allowed the use of faster growing and less fastidious cell lines, such as CHO, and amplification of product expression by multiple copies of the gene. 49

50

Griffiths

2. Principles

of Scale-Up

Animals cells are grown in two completely separate systems. For those cells that grow individually in suspension, the range of fermentation equipment developed for bacteria can be readily modified. This is a great advantage, since these culture vessels are economic in terms of space, the environment is homogeneous and can be critically controlled, and scale-up is relatively straightforward. Many cell types, however, will only grow when attached to a substrate or, in some cases, will only produce significant levels of a product when grown in this mode. Scale-up of substrate attached cells is far more difficult to achieve and has given rise to a wide range of alternative culture systems. Two approaches to scale-up can be taken. The first is volumetric-a simple increase in volume while retaining the sam&cell density or process intensity. The second method is to increase the cell density/unit vol lo-loo-fold by means of medium perfusion techniques. Cell densities of over 108/mL can be achieved in a variety of systems, but they are difficult to volumetrically scale-up because of difficulties in supplying media in great enough, and constant enough, volumes. Compromise is possible with large-scale (100-500 L) cultures operating at just 10-20 times above the conventional cell densities (l-3 x 106/mL) by means of special perfusion devices, such as the spin filter. The environmental factors that can most readily be controlled are pH (and redox) and oxygen. The limiting factor in scale-up, particularly in cell density, is usually oxygen. Surface aeration used in small cultures soon becomes inadequate, since the volume (and therefore depth) of medium increases. Bubbling of air/oxygen mixtures into cultures, with turbulent stirring/agitation, is the most efficient means of effecting mass transfer. Unfortunately, cells are fragile, compared to bacteria and only slow stirring and bubbling rates can be used, which are often inadequate for maintaining a sufficient oxygen supply. To overcome this problem, most cultures rely on several oxygenation methods, and many ingeneous methods have been developed for this purpose (1). Two further points should be taken into account during scale-up. The first of these is the increased risk of contamination and the proportionally higher costs of culture failure. The second is a question of logistics in the preparation of medium and particularly cell inocula. It is a small matter to harvest 10Scells and inoculate them in a good physiological state into a new culture. An inoculum of lOlo cells takes a long time to prepare, cells can be left for long periods in damaging conditions, and media can lose its tem-

Scale- Up of Animal

Cells

perature and set pH while these handling procedures are carried out. The objective is to keep both the process and the culture system as simple as possible, having everything well prepared and ready, and not to be overambitious with regard to scale. This will ensure that cultures are initiated with cells in good physiological condition and reduce the risks of microbial contamination.

3. Methods 3.1. Suspension

3.1.1. Culture

Culture

Vessels (Fig. 1)

The simplest means of growing cells in agitated suspension is the spinner culture vessel. The culture pot has a magnetic bar, usually placed a few millimeters from the bottom of the vessel, and is placed on a magnetic stirrer. As long as the bar is able to rotate freely and the stirrer is of sufficient quality to maintain constant stirring speeds, and not overheat, this methodology works extremely well for growing most cells up to densities of l-2 x 106/mL. These glass spinner vessels are available from a wide rangeof suppliers (e.g., Bellco, Wheaton) in sizes from 0.2-20 L (2). Amodification of this principal for shear-sensitive cells is the spinner vessel using surface agitation as exemplified by the BR-06 Bioreactor (Techne). Spinner vessels are only satisfactory up to a certain size-between 2 and 10 L depending upon the cell line and its required use. Above 10 L, glass vessels become inconvenient to handle and the progression to in situ stainless steel vessels should be considered. The other reason for change is that, with scale-up, the need to control the culture environment and carry out specialized manipulations (e.g., perfusion, media changes, and so on) increases. For this level of sophistication, a fermentation system needs to be used. The main differences are that (a) stirring is by a direct-drive mechanism with a motor, and (b) the vessel has a complex top that allows the inclusion of a range of sensors, probes, feed supply lines, and sampling devices for contamination-free monitoring and control. Fermenter kits (laboratory scale) are available in the range from lto 40 L and cost in the region of $450037,500 (E3,000-25,000) (with control of stirring speed, temperature, pH, and oxygen). Culture vessels for animal cells should have the following features: 1. Curved or domed bottom to increase mixing efficiency at the low stirring speeds that are used (loo-350 rpm).

52

Griffiths

I 05;
0 0 0 0 0 0 0

Fig. 1. Range of culture apparatus for suspension cells (A) magnetic bar spinner culture; (B) Techne MCS stirrer; (C) surface stirrer (Techne BR-05/06); (D) small scalefermenter with marine impeller; (E) airlift fermenter; (F) vibro-fermenter (Champ).

2. Water jacket temperature control to avoid the use of immersion heaters that give localized high temperatures. Electrically operated silicone pads are also suitable at volumes up to 5 L-the only disadvantage is the reduction in visibility into the vessel. 3. Absence of baffles and other sharp protrusions that cause turbulence. The interior is finished to a high grade of smoothness to minimize mechanical damage and for cleanliness. 4. An aspect ratio (height to diameter) of 2:l maximum, and preferably no more than X5:1.

Scale- Up of Animal Cells

53

5. Suitable impeller to achieve nondamaging bulk flow patterns (e.g., modified marine or pitched blade impellers) with top drive, so that there are no combinations of moving parts that would grind up the cells. Some animal cells, including many hybridoma lines, are very sensitive to the mechanical effects of stirring. For such cells, there are two alternative means of mixing besides stirring. 1. Vibromixer-this is a nonrotary device using a plate that vibrates in the vertical plane a distance of 0.1-3 mm. Conical perforations in the plate affect the mixing (Vibro-fermenter, Chemap). 2. Airlift-medium is circulated in a low velocity bulkflow pattern by being lifted up a central draft tube by rising air bubbles, and recirculated downwards in the outer ring formed by the draft tube. This system forms a near ideal mixing pattern and allows near-linear scale-up to at least 1000 L. Unfortunately, the apparatus, with a 121 aspect ratio, is very high and a 30-L fermenter needs 3-m ceiling height. Airlift fermenters are available commercially either as complete systems (e.g., LH Fermentation ) or as disposable 570 mL U (CelliftFisher Scientific).

3.1.2.

Culture

Procedure for Suspension

Cells

1. Inoculum should be prepared from a growing suspension of cells (i.e., in mid to late logarithmic phase). Stationary phase cells are either slow to start in a fresh culture or do not grow. 2. Prewarm to 37OC,and equilibrate the pH of the culture medium with the COJair gas mix, before inoculating the cells. 3. Inoculate cells at over 1 x 105/mL. Recommended level is 2-3 x 105/ mL for many cell (hybridoma) lines. 4. Stir the culture within the range 100-300 rpm. This speed depends upon the individual type of culture reactor. Stir at a speed sufficient to keep the cells in homogeneous suspension. Do not use speeds that allow cells to settle out at the bottom of the culture, 5. Monitor cell growth at least daily by taking a small sample, either through a special sampling device or removing the vessel to a laminar flow cabinet and using a pipet, and carry out a viable cell count (Trypan blue stain and a hemocytometer). 6. pH-if the culture is closed (i.e., all ports stoppered with no filters), then the pH will fall. You should remove the culture to a cabinet and gas the head space with air. If the culture is very acid, sterile sodium

54

Griffiths

bicarbonate (5.5%) can be added or, when the cells have settled out, remove 50% of the medium and replace with fresh (prewarmed) medium. Return culture to stirrer. It is preferable to have inlet and outlet filters, so that there is continuous head-space gassing, initially (i. e., first 24 h) with 5% CO, in air, followed by air only. Suitable filters are nonwettable with a 0.22 pm rating. 7. After 3-4 d, the saturation density of 1-2 x 106/cells/mL should be attained. Most suspension cells will then die at a rapid rate unless harvested or maintained with medium changes. 3.1.3. Special Procedures 1. Airlift-follow the above protocol except, instead of stirring, a gas flow rate of approximately 5-20 cc air/L/min is used for mixing. 2. Increase cell density by perfusion. To perfuse a culture (i.e., the continuous or semicontinuous addition of fresh medium and withdrawal of an equal volume of spent medium) means that methods of separating the cells from the medium are needed. There are basically two techniques that can be used, spin filter and hollow fibers. 3.L3.1. Spin FiZter. The problem with most filtration techniques is that the filter rapidly becomes blocked with cells. A spin filter, so called because it is attached to the stirrer shaft, reduces the problem of blockage, because as it spins it produces a boundary effect on its surface that reduces cell contact. Also, they normally have a large surface area and, thus, have a low flow rate at any one point. A porosity of about 6-10 pm is needed, and stainless steel mesh can be used. This is thin enough to be cut to form a cone in which the join can be double-folded and machine-pressed (Fig. 2). This allows a perfusion rate in the order of l-2 vol/d. This device will allow cell densities of over 107cells/mL to be achieved-higher if the culture system has full pH, and oxygen monitoring and control (3). 3.1.3.2. Hollow-Fiber Cartridges. Hollow-fiber units are available at both ultrafiltration and filtration grade. For the purposes of withdrawing medium and returning the cells to the culture in a loop outside the vessel filtration grade (0.22 pm) is sufficient. The scheme is shown diagrammatically in Fig. 3. Many fiber cartridges are not steam sterilizable, but polysulfone and teflon fibers are. The quantity of medium withdrawn must be balanced by adding fresh medium from a reservoir. Using flow rates of up to 1 vol/h, cell densities in the region of 2-5 x 107/mL can be achieved, but oxygen is a limiting factor and an additional filtration cartridge should be put in the external loop as an oxygenator.

Scale- Up of Animal

Cells

55

machine press

direction A

Fig. 2. Spinfilters: (A) construction using folded interleaving edges; (B) diagrammatic representation of a spin filter; (C) open perfusion systems; (D) closed (recirculating perfusion system).

3.1.4. Notes-Suspension

Culture

1. Perfusion-media usually becomes limiting because of acid buildup and low oxygen, and this leaves a surplus of the other nutrients, such as glucose and amino acids. In the open perfusion systems described above, the medium probably only yields 3-5 x 105cells/mL. If, however, the medium is circulated in a closed loop after passing through a reservoir in which violent agitation/aeration and addition of NaOH occurs (possible because it is a cell-free environment), then yields of 1-2 x 106/mL are achieved in media such as Eagles MEM (4). 2. Media-serum is a very high cost addition to culture media, but alternatives such as specialized serum-free media that include a range of growth factors usually work out as expensive, or more so. Although cultures may have to be initiated in a complex media, as cell density

56

Griffith

Fig. 3. Hollow-fiber cartridges in perfusion loop for (A) removing and returning cells to the culture and (B) oxygenating the medium.

spent medium

(H)

increases, so cells become less dependent upon serum and growth factors. Thus, at densities above 5 x 106-lO/mL, serum concentration can be drastically reduced (to 2%) or even excluded. If serum-free medium is used, cells are more susceptible to damage by stirring and air bubbles, but this can be offset to a certain extent by adding pluronic F68 (polyglycol) at 0.1%. (Seereview (4) for other means of cutting the costs of media.) 3. Contamination/sterility-the larger the scale, the more expensive a culture failure becomes. Carry out stringent quality-control procedures by testing growth media several days before it is to be used for bacterial contamination. Do not take shortcuts on the support equipment, but use specialized tubing connectors and sampling devices supplied by fermenter equipment companies. Also, do not overuse the air filter (6-10 sterilization cycles maximum), and do not allow them to get wet (either during autoclaving or with media or condensation). 4. Suspension culture-many cells either attach to the surfaces of the vessel or form unwanted clumps. Media for suspension culture should have a reduced calcium and magnesium-ion concentration (special formulations are commercially available) because of the role of these ions in cell attachment. Attachment to the vessel can be discouraged with a pretreatment of a proprietary silicone solution (e.g., Repelcote, Hopkins & Williams).

Scale- Up

of Animal Cells 3.2. Anchorage-Dependent 3.2.1. Materials Culture

57

Anchorage-dependent culture systems are far more difficult to scaleup than suspension cultures because of the additional requirement of providing the extra surface area in an economical (in terms of space) way and still maintain homogeneity throughout the system. For this reason, suspension culture, in which a 1-L stirred vessel is conceptually similar to a 1000-L vessel, is always the preferred culture method. The first step in scale-up (Fig. 4) usually involves the change from stationary flasks (available in sizes up to 200 cm2) to roller bottles (sizes up to 1750 cm). The larger size of roller bottle will yield in the range of 2-5 x lo* cells, and therefore, for most purposes, a mu1 tiplici ty of rollers has to be used. The next step in scale-up is to use roller bottles that have an increased surface area resulting from the inclusion of glass tubing (Belco-Corbeil, Chemap Gyrogen) or plastic spiral films (Sterlin). By this means, the surface area within a roller bottle can be increased to 8500 cm2 (spiral film) and 15,000 cm2 (glass tubing). An alternative to investing in specialized, and costly, roller culture equipment is to use plastic multi-tray units (Nunc). Each tray has a surface area of 600 cm2 and units of 6,10, and 40 (24,000 cm2) plates can be obtained. Two systems that allow a huge unit scale-up of substrate attached cells are immobilized beds (e.g., constructed of glass spheres) and microcarrier culture (cells growing on 200 ym spheres that are stirred in suspension culture apparatus). Microcarriers can provide 5000 to 50,000 cm2/L and are currently being used in commercial production systems at the 1000-L scale (a total of 15 x lo6 cm2 surface area, which has the potential of supporting 15 x 10 cells). 3.2.1.1. Roller Culture. Reutilizable glass or disposable plastic roller cultures are used. The most commonly used sizes are 750-850 cm2 and 1500- 1700 cm2. Complete modular systems holding up to 48 large bottles can be purchased either free-standing, for use in hot rooms, or within an incubator cabinet. 3.2.1 .l .1. PROCEDURE. The following procedure is based on a 1500 cm2 (24 x 12 cm) roller bottle. 1. Add 200-300 mL of medium. 2. Add 1.5 x 10 cells (observe previously listed advice on preparation of inocula and medium). 3. Revolve the culture at 15 rph.

58

Griffiths

(i&i v

750

1750

cm* 3 II=

Lf

I
d

w
6-250X103

9
Cm* 104cm21L

Y
25X103 Cld L

e
Fig. 4. Scale-up of anchorage-dependent cultures: (a) flask; (b) roller bottle; (c) plastic spiral (Sterlin); (d) glass tubing (Belco-Corbeil); (e) stack plates; (0 multi-tray units (NUNC); (g) immobilized bed (glass spheres); (h) microcarrier culture. The figures define the surface area for each culture vessel.

4. Cell growth can be observed under an inverted microscope with a

long-distance objective. 5. After 3-5 d, the cell sheet will be confluent yielding from 1.5 x 10s (human diploid) to 5 x lo5 (heteroploid cells, e.g., HeLa) cells/cm*. 6. Pour off the medium, wash the cell sheet with prewarmed phosphate buffered saline, and add 50 mL trypsin (0.25%). Place culture back on roller, and allow to revolve for lo-20 min. The cells will detach and can be harvested, diluted in fresh medium and serum, and passaged on. This outline protocol can be considerably modified. An advantage of this method is that the medium volume:surface area ratio can be altered easily. Thus, after a growth phase, and when a product is to be harvested, the medium volume can be reduced to 100 mL in order to obtain higher product concentration. 3.2.1.2. Glass Bead Immobilized Beds (3,5). Apparatus-this type of culture system is easily fabricated in the laboratory (Fig. 5). A suitable

Scale- Up of Animal Cells

59
aiir I

D
Fig. 5. Glass-sphere immobilize&bed culture: (A) immobilized

stirrer, pH and oxygen probes, air spar&g; (Cl pump; (D) inlet for inoculum/trypsin; water at 37OC.

bed; (B) reservoir with (E)

cylindrical or funnel-shaped glass container is packed with borosilicate glass spheres (minimum diameter 3 mm, optimum diameter 4 or 5 mm). Medium is perfused by means of a peristaltic pump from a reservoir (which ideally is a spinner flask or small fermenter), which can be monitored and controlled for pH, oxygen, and so on. The productive capabilities of the system are give in Table 1. 3.2.1.2.1. PROCEDURE 1. Prepare growth medium, and add to reservoir (2 mL/cm2 culture surface area). 2. Equilibriate the system for temperature and pH, and circulate the media through the packed bed (allow sufficient time for the solid glass spheres to reach 37C). 3. Inoculate cells (l-2 x 104/cm2) into a volume of medium equal to the void volume of the bed (250 mL/kg 5 mm spheres). 4. Allow the cells to attach (3-8 h depending upon cell type), but the culture can safely be left overnight (16 h) at this stage.

Griffiths Table 1 Phvsical Characteristics of Glass Sphere Beds Bead diameter 3mm 5mln Surface area (cm? Total 7400 4600 Available (70%) 5200 3200 Void medium vol (cc) 295 250 Total vol (cc) 675 625 Cell count (x 105/cm2) 0.78 2.50 (x W/kg> 4.0 8.0

5. Start medium perfusion, initially at a rate of 1 linear cm of column/l0 min, but as the culture progresses, this rate is increased to a maximum of 5 cm/min. 6. Cell growth can only be monitored by indirect measurement, and the glucose utilization rate is the most convenient. An alternative is oxygen utilization rate. Growth yields should be determined for the particular cell line and nutrient, so that an approximation of cell numbers can be made (e.g., glucose utilization is usually within the range of 2-5 x lo8 cells produced/g glucose). 7. When the culture is estimated to be confluent (after 5-7 d), drain the medium, give the bed a phosphate-buffered saline wash, and add trypsin/versene to harvest the cells. The efficiency of cell detachment can be increased by intermittently draining and pumping back the trypsin solution. The bed acts as a depth filter, and to recover a high percentage of the cells, the bed should be washed through several times with medium after the trypsin has been drained off. (NB., 5 mm beads allow a better drainage and cell recovery than 3 mm.) 8. Wash out the culture bed immediately with a detergent, so that cell debris does not become fixed onto the glass beads. This culture method is basically very simple, and the apparatus is It has a large potential scale-up and has been cheap and reutilizable. proven at the 100 L vol scale (30 L bed of 3 mm beads) (5). It is perhaps most suitable for harvesting a secreted cell product over a long period of time, rather than acting as a source of cells resulting from the difficulties of removing cells from the bed. 3.2.1.3. Microcarrier Czdture. The advantage of this methodology is that the cells, when growing on small carriers, can be treated as a suspen-

Scale- Up of Animal

Cells

61

sion culture with all the advantages of large unit scale-up, homogeneity, and easily controlled environmental conditions. The range of microcarriers commercially available is extensive (Table 2) (2,6), and at least one type is suitable for all cell types, however demanding. The decision of which one to use is influenced by whether a dried powder or alreadyprepared sterile solution is preferred, the cost/cm2, whether a special derivitized surface is needed for a particular cell, or whether one wishes to harvest cells by dissolving the carrier (gelatin, collagen) and thus producing a higher quality cell suspension. Experience has shown that it is worthwhile to evaluate several types in small-scale cultures for each particular cell line, since significant differences in cell yield and longevity of culture (before cell detachment) are seen. 3.2.1.3.1. CULTURE APPARATUS. Modifications of suspension culture vessels are used. Spinner flasks with a magnetic bar are unsuitable, but versions are available with large paddle-type impellers (Bellco) or specially modified stirring actions (e.g., Techne MCS). Stirring rates are much slower (20-70 rpm) than for suspension cells, and thus more efficient mixing at low speeds is required. Scale-up in laboratory fermenters can also use the large-bladed paddles, but there are several modifications of the marine impeller available that are very efficient (e.g., SGI ascenseur). 3.2.1.3.2. PROCEDURE. Microcarrier culture is not a difficult technique, but it does require more critical attention to experimental detail than most methods and the use of the correct culture vessels. The following procedure is based on using Cytodex 3 (Pharmacia) or Dormacell 2.3 (Pfeifer & Langen) at 3 g/L. Prepare the microcarriers according to manufacturers recommendations. 1. It is essential that the medium with microcarriers be prewarmed and stabilized before inoculating the cells. Cell attachment to moving spheres requires conditions to be just right. It is even more important with this method than with previously described ones to initiate the culture with growing (logarithmic cells) and not stationary-phase cells, and cells that have been rapidly prepared and are in good physiological condition (i.e., have not been standing in trypsin, and so on, for extended periods). 2. Inoculate at 2 x 104/cm2 into 30-50% of the final volume. Stir at the minimum speed to maintain homogeneity (20-30 rpm) for 4-8 h. 3. When the cells have attached (expect 70-90% plating efficiency), the volume can be increased to the full working volume. (Stirring speed may also have to be increased to give complete mixing.)

Griffiths Table 2
Commercially Name Acrobead Available Microcarriers

Manufacturer
Galil Solo Hill Eng. Solo Hill Eng. Nunc Pharmacia

Type Derivitized Glass/latex

Polystyrene

Bioglas
Bioplas

Biosilon Cytodex 1,2 Cytodex 3

Cytosphere Dormacell

Pharmacia Lux
Pfeifer & Langen

Polystyrene Dextran Collagen

Polystyrene Dextran

Gelibead Mica Micarcel G Microdex

Superbeads Ventregel Ventreglas

Hazelton Lab.
Muller-Lierheim Reactifs 1BF

Gelatin
Polyacrylamide

5000 350 350 255 6000 4600 250 7000 3800

b

Dextran Prod. Flow Labs.

Ventrex Lab. Ventrex Lab.

Polyacrylamide Dextran Dextran

Gelatin Glass/polystyrene

5000 250 6000 4000 300

A guideline only as different types vary Z-5-fold in cell yield/cm2. bPorous matrix beads.

4. A great advantage of the microcarrier system is that samples can be readily removed and microscopically examined. Unstained preparations will show whether or not cells have attached, spread out, and then begun to grow. Cell counts can be made by the standard nucleicounting procedure, which releases the stained nuclei from the attached cells. 5. As the culture progresses, the stirring rate can be increased to prevent cell-to-cell attachment, bridging microcarriers and causing clumps to form. A maximum of 75 rpm should be arrived at. 6. In nonenvironmental controlled cultures, the media will turn acid after 3-4 d and a partial (50-70%) media change should be carried out. Stop stirring, allow beads to settle (10 min), siphon off spent medium, add fresh (prewarmed) medium, and start stirring, gradually increasing the rate. 7. Cells can be harvested when confluent by allowing the carriers to settle out, giving a serum-free wash, allowing the carriers to settle out again, decanting off as much of the free fluid as possible, adding trypsin, and restarting stirring, but at slightly faster speed (75-125 rpm). After 20 min, allow the beads to settle out for 2 min. Cells can either

Scale- Up of Animal

Cells

be removed by decanting, or the mixture can be filtered through a coarse sintered glass filter that allows passage of the cells, not the microcarriers. If gelatin or collagen carriers are being used, then cells can be released by treatment with trypsin/EDTA (which solubilizes gelatin) or collagenase.

3.2.2.

Scaling- Up

This can be achieved by increasing the culture volume, and increasing the microcarrier concentration from the suggested 3 to 5-15 g/L. If higher concentrations are used, then it is imperative to have a perfused system with full environmental control. The easiest means of perfusing is the spin filter (as described for suspension cells), but a much larger pore size can be used (60-100 pm). This allows much faster perfusion rates to be attained (1-2 vol/ h). Perfusion from a reservoir that is adequately gassed is an efficient means of oxygenating the culture. Spin filter systems are commercially available (LH Fermentation, New Brunswick).

4. Conclusion
Microcarrier is undoubtably the most efficient scale-up method for anchorage-dependent cells currently available. It allows volumetric (to 1000 L) and density scale-up with a spin filter (to 5 x lo7 cells/ml). The equipment mus t be specially designed for the procedure, but it can also be used for suspension cells, or as reservoir vessels for other high process intensity systems (e.g., hollow fiber). A disadvantage is the high substrate cost $1800-225O/kg (E120041500/kg) and the fastidious nature of the technique.

References
I. 2. 3. 4. 5. Spier, R. E. and Griffiths, J. B. (1983) An examination of the data and concepts germane to the oxygenation of cultured animal cells. Develop0Bid. Sfandayd.55,81-92. Griffiths, J.B. (1986) Scaling-up of animal cells, in Animal Cell Culture: AP~ucficuZ Approach (Freshney, I., ed.), IRL Press,Oxford, pp. 33-70. Griffiths, J. B., Cameron, D. R., and Looby, D. (1987) A comparison of unit process systems for anchorage dependent cells. Develop. Bid. Sfunduyd,66,331-338. Griffiths, J. B. (1986) Can cell culture medium costsbe reduced? Strategies and possibilities. Trendsin Biofechnol.4,268-272. Whiteside, J.P. and Spier, R. E. (1981) The scale-up fromO.l to lOOliter of aunit processsystembased on 3 mm diameter glass spheres for the production of four strains of FMDV from BHK monolayer cells. Biofechnol.Bioeng.23,551-565.

Chapter

Mycoplasma

Detection

J. M. Mowles
1. Introduction
Mycoplasma is the generic term used by cell biologists to denote organisms belonging to the Order Mycoplasmatales, which can infect cell cultures. Of particular interest are those organisms that belong to the Families Mycoplasmataceae (Mycoplasma) and Acholeplasmataceae (Acholeplasma). Mycoplasmas differ from other prokaryotes by their lack of a cell wall. Unlike L-forms of bacteria, which under the right environmental conditions are capable of producing a cell wall, mycoplasmas are unable to produce even precursors of bacterial cell wall polymers. Another distinguishing feature is their size; mycoplasmas are the smallest self-replicating prokaryotes with coccoid forms of only 0.3 pm diameter. Their genome size is approximately one-sixth that of Eschevichia coli. Mycoplasma infection of cell cultures was first observed by Robinson et al. in 1956 (I). Since then, the incidence of mycoplasma-infected cell cultures has been found to vary from laboratory to laboratory. Currently, 12% of cell lines received by the ECACC are mycoplasma infected, which may be an uncharacteristically low figure since many lines prior to deposition are mycoplasma screened.

65

Mowles The importance of detecting mycoplasmas in cell cultures should not be underestimated. In an infected cell culture, the concentration of mycoplasmas in the supernatant can be typically in the region of 106-lo8 mycoplasmas/ml. In addition, mycoplasmas will cytoadsorb to the host cells. Mycoplasmas do not necessarily manifest themselves in the manner of most bacterial or fungal contaminants, e.g., culture turbidity or pH change. It is therefore important that an active routine detection procedure be adopted. However, among the effects mycoplasmas have been shown to elicit are interference in the growth rate of cells (2), induction of morphological alterations (including cytopathology) (3), induction of chromosome aberrations (4), influencing amino acid (5) and nucleic acid (6) metabolism, causing membrane alterations (7), and even cell transformation (8). Just as importantly, the various regulatory bodies (9,lO) demand that cell cultures used for the production of reagents for diagnostic kits or cell cultures used to produce therapeutic agents are free from mycoplasma infection. The majority of cell culture mycoplasma infections are caused by five species. The organisms and their natural hosts areM. arginini (bovine), M. fmmentuns (human), M. hyorhinis (porcine), M. orale (human), and A. laidlawii (bovine). The genus Acholeplasma differ from other members of the order Mycoplasmatales by having no requirement for sterol for growth. A wide range of assay techniques are available to detect mycoplasma infection. Where possible, at least two methods should be used to remove the risk of false-positives and false-negatives. At the ECACC, the two methods of choice are DNA staining and culture.

2. Materials
Use analytical grade (Analar) reagents and distilled water throughout. 1. Carnoys fixative: For each preparation, mix methanol (3 mL) and acetic acid (glacial) (1 mL). Prepare fresh as required. 2. Hoechst stainstocksolution (100mL): (Bisbenzimide,Hoechst33258). Add 10 mg Hoechs t 33258 to 100 mL water and filter sterilize through a 0.2 ym filter. The container should be wrapped in aluminum foil and stored in the dark at 4C. 3. Hoechst stain working solution (50 mL): Add 50 PL stock solution to 50 mL water. Prepare fresh as required. 4. Mountant: Citric acid (O.lM) 22.2 mL, disodium phosphate (0.2M) 27.8 mL. Autoclave the above and mix with 50 mL of glycerol. Adjust the pH to 5.5, filter sterilize, and store at 4C.

Mycoplasma Detection
5. Fluorescencemicroscopeequipped

67
for epi-fluorescence; excitation filter

340-380nm, suppression filter 430nm.

6. Sterile tissue culture dishes (33-mm) to which sterile coverslips (22 x 22 mm No 1.5) have been added. 7. Agar Media for mycoplasma culture: Prepare the following stock solutions: a. Mycoplasma agar base* (80 mL): Add 2.8 g of agar base to 80 mL of water, mix, and autoclave at 15 lb/in2 for 15 min. Prepare fresh as required. b. Pig serum (10 mL): Dispense in IO mL aliquots in universal containers, and then heat to 56OC for 45 min. Store at -30C. c. Yeast extract* (10 mL): Add 7 g to 100 mL water, mix, and autoclave at 15 lb/in2 for 15 min. Dispense in 10 mL aliquots in universal containers. Store at 4OC. To prepare the agar media, autoclave the agar base, cool to 50C and then mix with the other two constituents (warmed to 5OOC). Dispense 8 mL/5-cm diameter Petri dish, and store at 4OCin sealed plastic bags. Complete medium should be used within 10 d. 8. Brothmedia for mycoplasma culture: Prepare the following stock solutions: a. Mycoplasma broth base* (70 mL): Add 2 g broth base to 70 mL water, mix, and autoclave at 15 lb/in2 for 15 min. b. Horse serum (20 mL): Dispense in 20 mL aliquots in universal containers. Store at -30C. Do not heat inactivate. c. Yeast extract* (10 mL): As for agar media. To prepare the broth media, mix the three constituents, dispense in 1.8 mL aliquots in glass vials, and store at 4OC. Complete medium may be stored for several weeks without deterioration. 9. Dienes stain stock solution: Methylene blue 2.05 g Azure II 1.25 g Maltose 10.00 g Sodium carbonate (Na,CO,) 0.25 g Distilled water 100 mL To prepare the Dienes stain working solution, filter sterilize (0.2 pm> a 3% dilution of the stock solution in distilled water.
Oxford Ltd., Wade Rd, Basingsfoke, Hauts, RG24 OPW UK.

3. Method
The following general principles are common to both methods employed for the detection of mycoplasma:

68

Mowles

1. Cells for testing should have completed at least two passages in antibiotic-free medium prior to testing, because the presence of antibiotics may mask infection. 2. Since cryoprotectants may also mask infection, previously frozen cell cultures should not be tested until at least two passages are complete.

3.1. DNA Stain

1. Harvest adherent cells using a routine method of subculture, e.g., trypsinization, and resuspend in the original culture medium to a concentration of approximately 5 x 105 cells/mL. Test cell lines growing in suspensions directly from culture at approximately 5 x 105cells/ mL. Experience in working with any given cell line should eventually remove the necessity for an accurate cell count. Sufficient cells should be added to dishes so that, at the time of observation (at 1 and 3 d postincubation), a semiconfluent spread of cells on the coverslip is obtained. Similarly, if it is considered that after a further 3 d incubation the media would be exhausted, cells should be resuspended in a mixture of the media in which they were growing and fresh media. 2. Add 2-3 mL of test cells to each of two tissue culture dishes containing coverslips. 3. Incubate at 36 + 1C in a humidified 5% CO,, 95% air atmosphere for 12-24 h. 4. Remove one dish, leaving the remaining dish for a further 48 h. 5. Prior to fixing the cells, examine on an inverted microscope for bacterial and fungal contamination. 6. Cells are fixed by adding approximately 2 mL Carnoys fixative dropwise to the edge of the dish, which is then left for 3 min at room temperature. The addition of fixative in this way is particularly important for suspension cultures to avoid cells being swept to one side of the dish. 7. Decant the fixative to a waste bottle (avoiding fumes), add a further 2 mL aliquot of fixative to the dish, and leave for 3 min at room temperature. 8. Decant the fixative to a waste bottle. 9. Allow the coverslip to air dry. Invert the lid of the dish and, using forceps, rest the coverslip against the lid for approximately 30 min. 10. Add 2 mL of Hoechs t stain (gloves must be worn) to the coverslip, and leave for 5 min at room temperature. At this point, shield thecoverslip from direct light. 11. Decant the stain to a waste bottle. 12. Add 1 drop of mountant to a labeled slide, and place the coverslip cellside down onto slide.

Mycoplasma

Detection

69

13. Examine the slide under UV epifluorescence at 100 x magnification with oil immersion.

3.2. Microbiological

Culture

In addition to passagein antibiotic-free and cryoprotectant-freemedia, it is important that cultures for mycoplasma detection by microbial culture should not receive a media renewal within 3 d of testing. 1. Harvest adherent cells using a sterile swab, and resuspend in the culture medium to a concentration of approximately 5 x 105cells/mL. Test suspension cells direct at approximately 5 x lo5 cells/mL. 2. Inoculate an agar plate with 0.1 mL of the test sample. 3. Inoculate broth with 0.2 mL of the test sample. 4. Incubate the test plate under anaerobic conditions (Gaspack) at 36 + lC for 21 d and the test broth anaerobically at 36 & lC. 5. Subculture the test broth (0.1 mL) onto agar plates at approximately 7 and 14 d postinoculation, and incubate anaerobically at 36 + lC. 6. Examine all agar plates after 7,14, and 21 d incubation under 40 x or 100 x magnification using an inverted microscope for detection of mycoplasma colonies. 3.2.1. Dienes Stain 1. Add sufficient working solution to the agar plate to cover the suspect colonies, and observe using an inverted microscope at 40 and 100 x magnification. 2. Alternatively, if the agar plate is to be reincubated, remove a portion of agar containing the suspect colonies using a sterile bacteriological loop and place on a microscope slide, with colonies facing up. Add Dienes stain and observe. Mycoplasma colonies stain blue (seeSection 4.2., Fig. 3).

4. Notes 4.1. DNA Stain

1. The fluorochrome dye, Hoechs t 33258, binds specifically to DNA. Uninfected cell cultures observed under fluorescent microscopy are seen as fluorescing nuclei against a negative background (Fig. l), whereas cultures infected with mycoplasma contain fluorescing nuclei plus extranuclear mycoplasmal DNA (Fig. 2).
GasPack (BBL Microbiology Systems, Cockeysville, MD).

Mowles

Fig. 1. Myoplasma-negative cell line.

2. Mycoplasma may appear as filamentous forms, some of which may be branching,indicatingacultureinlogarithmicgrowth,orascocci,which is typical of an older mycoplasma culture. Some slide preparations may contain extranuclear fluorescence caused by disintegrating nuclei, and this should not be confused with mycoplasma. Such fluorescence is usually not of a uniform size and is often too large for mycoplasma. Contaminating bacteria or fungi will also stain using this technique, but will, however, appear much larger and brighter than mycoplasmas. 3. The main advantages of the staining method are the speed at which resultsareobtainedand that theas-yet-noncultivableM. hy~~!zinisstrains observed in cell cultures can be detected. 4. An alternative to the method described is to use an indicator cell line onto which is inoculated the test cells. The advantage of this system is its standardization, which enables the mycoplasma screening of serum andother cell culturereagents that may beinoculatedonto theindicator line. Also, positive and negative controls may be included in each assay. However, the growth of mycoplasma cultures to use as

Mvcoda-smu

Detection

71

Fig. 2. Myoplasma-infected cell line.

positive controls may prove impracticable in many laboratories. The disadvantages of using an indicator line are the extra time and effort required in its preparation. ArecommendedindicatorcelllineistheVeroGreenMonkeyKidney line, which has a high cytoplasm/nucleus ratio. The indicator cells are added to sterile coverslips in tissue culture dishes 10-24 h prior to inoculation of test sample at a concentration at which a semiconfluent monolayer is formed at the time of observation (at 1 and 3 d postinoculation of test sample). Slides are prepared in the same manner as those using the direct method. 5. Positive and negative slides may be kept in the dark for demonstration purposes without deterioration.

4.2. Mycoplasma

Culture

1. Quality control of each new batch of media ingredients should be performed prior to agar or broth preparation. It is especially important that new batches of pig and horse serum are shown to be capable of supporting the growth of a representative sample of species found in-

72

Mowles

fecting cell culture, e.g., any two of A. Eaidlawii,M. auginini,M.feumentans, M. hominis, M. hyouhinis, or M. ouaEe.Type strains may be obtained from the National Collection of Type Cultures (11), or the American Type Culture Collection (12), or wild-type strains may be used. Stock positive control cultures may be kept frozen at -7OOC in mycoplasma broth. 2, Each batch of complete broth media should be quality controlled prior to use by the addition of at least two different strains of positive control mycoplasmas. An uninoculated broth should also be incubated as a negative control. 3. Mycoplasma agar medium should also be tested for its ability to support mycoplasma growth. This can be most conveniently performed at the time of test sample inoculation. An uninoculated plate should also be included as a negative control. 4. It is necessary to be able to distinguish pseudocolonies and cell aggregates frommycoplasma colonies on agar. Pseudocolonies are caused by crystal formation and may even increase in size. They can be distinguished from genuine mycoplasma colonies by using Dienes stain, which stains mycoplasma colonies blue, but does not stain pseudocolonies. Most bacterial and fungal colonies also appear colorless. Cell aggregates, however, do not increase in size and so are more easily distinguished. Also, by using a sterile bacteriological loop, cells may be disrupted leaving the agar surface free of aggregates, but mycoplasma colonies, because of the nature of their growth, will leave a central core that is embedded in the agar. Another good indicator of genuine mycoplasma colonies is their typical fried-egg appearance on agar (Fig. 3); however, this is not always seen in primary isolates.

4.3.

Other

Detection

Methods

1. Two commercially available detection methods recently introduced are the Myco Tect system (13) and Gen-Probe (14), both of which are indirect methods of detection. The Myco Tect system requires cocultivation of tes t cells with 6-me thylpurine deoxyriboside (6 MPDR), which is nontoxic to mammalian cells, but which mycoplasmal phosphorylase converts to6-methylpurine and6-methylpurineriboside,bothof which are toxic to animal cells and, therefore, destroys them. Results are available within 34 d (for a comparison of this method with DNA staining and culture, seeref. 15). 2. Gen-Probe is a DNA hybridization assay that requires nocultureprocedure and claims to aive results within an hour. The protocol reauires

Mycoplasma

Detection

73

Fig. 3. Myoplasma colonies on agar.

incubation of 3H-labeled DNA probe with either cell culture supernatant or the cells themselves. Hydroxyapatite is utilized to separate bound from unbound probe prior to scintillation counting. The DNA probe used is homologous to mycoplasmal r-RNA anD, therefore, hybridizes with different species of mycoplasma, but not with mammalian cellular or mitochondrial r-RNAs.

References
2. Robinson, L. B., Wichelhausen, R. B., and Roizman, B. (1956) Contamination of hu-

man cell cultures by pleuropneumonia-like organisms. Science 124,1147,1148. 2. McGarrity, G. J., Phillips, D., and Vaidya, A. (1980) Mycoplasmal infection of lymphocyte cultures: Infection with M. sdivurium. In Vitro 16,346-356. 3. Butler,M.andLeach,R.H.(1964)Amycoplasma thatinducesacidityandcytopathic effect in tissue culture. J. Gen. Microbial. 34,285-294. 4. Aula, P. and Nichols, W. W. (1967) The cytogenctic effects of mycoplasma in human leucocyte cutures. 1. CeII Physiol. 70,281-290. 5. Stanbridge, E. J., Hayflick, L., and Perkins, F. T. (1971) Modification of amino acid concentrationsinduced bymycoplasmasincellculturemedium. Nafure232,242-244.

Mozules
6. 7. 8. 9. 10. 11. 12. 13. :;: Levine, E. M., Thomas, L., McGregor, D., Hayflick, L. V., and Eagle, H. (1968) Altered nucleic acid metabolism in human cell cultures infected with mycoplasma. Proc. Natl. Acad. Sci. USA 60,583-589. Wise, K. S., Cassell, G. H., and Action, R. T. (1978) Selective association of murine T lymphoblastoid cell surface alloantigens with M. hyorhinis. Proc. Natl. Acad. Sci. USA 7544794483. MacPherson, I. and Russel, W. (19661 Transformations in hampster cells mediated by mycoplasmas. Nature 210,1343-1345. EuropeanPharmacopoeia (1980)BiologicalTests. ZndEd.,Part l,V.2.1.3Mycoplasmas. France: Maisonneuve, S. A. Code of Federal Regulations (1986) Animals and animal products. 9. Ch. 1. Part 113.28 p. 379. Detection of mycoplasma contamination. Office of the Federal Register National Archives and Records Administration, Washington. National Collection of Type Cul tures, Central Public Heal th Laboratory, 61, Colindale Avenue, Colindale, London NW9 5 HT. American Type Culture Collection, 12301 Parklawn Drive, Rockville, Maryland 20852, USA. Myco-Tect, Bethesda Research Laboratories, Inc., Gaithersburg, Maryland 20877, USA. Gen-Probe, Inc., 9620 Chesapeake Drive, San Diego, California 92123, USA. McGarrity, G. J., Kotani, H.,and Carson, D. (1986) Comparativestudies to determine the efficiency of 6 methylpurine deoxyriboside to detect cell culture mycoplasmas. In Vitro Cellular and Developmental Biology 22,301-304.

Chapter

Short-Term and Amniotic

Chorionic Villi Fluid Cultures

M. L. Murer-Orlando and V. M. McGuire

1. Introduction
Prenatal diagnosis of genetic disorders associated with specific biochemical, chromosomal, or molecular characteristics can be achieved from amniotic fluid (AF) or placenta (chorionic villus: CV) samples. Chorion material is usually obtained by sampling the placenta at the implantation site, during the first trimester (i. e., 9-12 wk), using either the transcervical or transabdominal route. The first method entails access to the placenta via the cervical canal, with aspiration through a metal cannula or a flexible catheter. Alternatively, the chorionic villi may be aspirated using a needle, which is passed through the abdominal wall. In both these methods, the positioning of the aspirating device must be made with the help of ultrasound scanning. Later pregnancies can only be sampled by the transabdominal route. Amniocentesis is usually performed at 16-18 wk of gestation by the transabdominal method. In this chapter, we will describe the techniques we currently use to obtain chromosome preparations from CV and AF samples. Chromosome analysis from CV samples can be performed either on spontaneously dividing cells of the cytotrophoblast layer (direct preparation) or on cells 75

76

Murer-Orlando

and McGuire

from cultured villi (culture preparation). The origin of cells cultured and analysed in amniotic fluid samples is difficult to establish. Avariety of cells can be isolated from uncultured amniotic fluid, and they are considered to derive from fetal epidermis or periderm, amnion, trophoblast, mucosa of digestive and urinary tract, and urogenital epithelia. Most of these cells are enucleate or dead, and even those that are viable may not adhere to the culture surface and proliferate.

2. Materials
2.1. Chorionic
1. 2.

ViUus Samples

3. 4. 5. 6.

2.1.1. Direct Preparation A biopsy of at least 10 mg is required for chromosome analysis. Hepes buffered RPM1 1640 medium is used for tissue collection and direct preparation, with appropriate supplements as follows: a. Wash medium: 150 pg/mL streptomycin sulfate and 150 U/ mL penicillin (standard antibiotics) and 5 U/mL heparin. Can be stored at 4OCfor 1 wk. This is used during sampling, in the collection container, and for washing blood-stained tissue. b. Transport medium: Standard antibiotics, 10% fetal bovine serum (FBS) and 20 mM L-glutamine. Can be stored at 4C for 1 wk. This medium is used for transporting washed villi to the laboratory for culture. c. Colcemid medium: Standard antibiotics, 10% FBS, 20 mM Lglutamine and 0.5 pg/mL Colcemid. This is used for direct chromosome preparations and is prepared as required. Colcemid solutions: Stocksolution: 1 mg/mL in sterile distilled water, stored at 4OC. Working sohtion: 50 pg/mL in sterile distilled water stored at 4C. Hypotonic solution: 1% sodium citrate in distilled water. Can be stored at room temperature for one week. Fixative: 1:3 glacial acetic acid ARabsolute methanol AR. Made up as required, but not stored. Acetic acid in distilled water (60%): Made as required, but not stored.

2.1.2. Culture Preparation 1. A biopsy of at least 5 mg is required to set up chorionic villi cultures. Accurate selection of villi is essential to avoid maternal contamination in cultures.

CVS and Amniotic Fluid

2. Wash and transport media: As in protocols of direct preparation (see

above).
3. Incubation medium: Transport medium with an addition of 2.5 pg/ mL Amphoterycin B (Fungizone). This is used for the overnight incubation of tissue, prior to culture, to prevent yeast infection. 4. Medium for primary cultures: Changs medium (Hana Biologic Inc. Both the basal liquid medium and the lyophilized supplement are marked with expiration dates). Standard antibiotics, 20 mM L-Glutamine, and 20 mM Hepes buffer are added to the basal medium before the addition of reconstituted supplement. This is a costly medium, the reconstituted supplement may be stored in appropriate aliquots at -2OOCup to the expiration date, and thawed for dilution with advised volumes of basal medium before use. The complete medium may be stored at 4OC for up to 10 d. 5. Medium for secondary cultures: Secondary cultures can be maintained in RPM1 1640 with standard antibiotics, 20 mM L-glutamine, and 20% FBS. 6. Colcemid solution: Stock and working solutions as in protocols of direct preparations (seeabove). 7. Trypsin solutions: Stock solution: 2.5% trypsin to be aliquoted and stored at -2OOC. Working solution: 0.16% trypsin (diluted from stock solution, with 0.02% EDTA, in PBS without calcium and magnesium). Made up as required, but not stored. 8. Hypotonic solution. 1:6 calf serum in sterile distilled water. Made as required. 9. Fixative. 1:3 glacial acetic acid AR: absolute methanol AR. Made as required, but not stored.

2.2. Amniotic

Fluid

Samples

1. Amniotic fluid samples can appear: clear, cloudy (mostly from the more advanced pregnancies), brown (where there has been some bleeding or fetal distress prior to sampling), or blood stained (because of bleeding during sampling). Approximately lo-20 mL of fluid will be required for culture. 2. Culture medium. Hams FlO Hepes buffered with standard antibiotics and t-Glutamine, supplemented with 5% FBS and 1% Ultroser G. serum substitute (Life Science Labs. Cat. No. 50902). 3. Colcemid, trypsin, hypotonic solutions, and fixative, prepared as in the chorion culture protocols.

78

Murer-Orlando

and McGuire

3. Methods
3.1. Chorionic
1. 2. 3. 4. 5. 6. 7.

Villus

Samples

3.1.1. Direct Preparations (See Notes 1-5) Collect sample in wash medium. Wash at least once, or more if sample is blood stained. Examine villi under a stereomicroscope, and remove maternal decidua and blood clots. Leave selected material at room temperature in 5 mL of colcemid medium for at least 2 h (max. of 4 h). Take off the colcemid medium, and add 4 mL of hypotonic solution; leave for 20 min at room temperature. Take off the hypotonic solution, and add 4 mL of fixative. Two further changes of fixative are recommended to cover a total of 20-min fixation time. Store samples overnight at -20C in fixative. Individual villi are selected for chromosome preparations. Rehydrate eachvillus by passage through decreasing concentrations of methanol to water. Transfer to a clean slide, and add 2-3 drops of 60% acetic acid. Tease the tissue gently with fine forceps to hasten disaggregation, and using a Pasteur pipet, transfer the resulting cell suspension onto two or more slides. Allow to dry on a hot plate (4OOC). 3.1.2. Culture Preparations (See Notes 6-9)

3.1.2.1. Setting Up Cultures 1. Collect, wash, and select villi as described in protocols for direct preparations. Leave overnight at 37OC in incubation medium. 2. Divide the sample into 30-mm Petri dishes (=5 mg/dish), add one drop of Changs medium to moisten the tissue, and mince the villi with a scalpel. 3. Transfer the fragments to a flat-sided Nunc tube. Add approximately 0.1 mL of medium to the edge of the dish, and to the base of the tube to maintain humidity. Leave at room temperature for 2-3 h to allow the tissue to adhere to the plastic. 4. Slowly add approximately 2 mL of Changs medium to each culture. Transfer to a 2% CO, incubator, and leave undisturbed for 3-5 d. 5. Change medium, subsequently, every 2-3 d. Harvesting is usually possible in 8-10 d.

CVS and Amniotic 3.1.2.2. Harvesting

Fluid

79

1. Add colcemid to a final concentration of 0.5 pg/mL (0.01 mL of colcemid working solution for mL of medium), and incubate for 2 h at 37OC. 2. Discard the medium, and gently wash the cell monolayer with 1 mL of 0.02% EDTA at 37C. 3. Replace with 1 mL of trypsin working solution (0.16%) at 37C. When the cells start to detach (30-60 s), shake the container to aid even cell suspension. 4. Add 2 mL of serum hypotonic at 37OC. Incubate for 15 min at 37OC. 5. Transfer the cell suspension to a centrifuge tube containing 0.5 mL of fixative, and centrifuge for 5 min at 25Og. (This prefixation step ensures an even cell suspension.) 6. Discard supernatant, and resuspend the cells in 4 mL of fixative. Leave for 5 min at room temperature. 7. Centrifuge, discard supernatant, and resuspend in 2 mL of fixative. Leave for at least 30 min at room temperature. 8. Centrifuge, discard supernatant, and resuspend pellet in approximately 0.5 mL of fixative. Make slides from the cell suspension. 3.1.2.3. Subculturing 1. Discard the medium, and gently rinse the cultures with 1 mL of 0.02% EDTA. 2. Replace with 0.5 mL of trypsin working solution (0.16%). When the cells are detached, add 4 mL of culture medium, and pipet vigorously to ensure even cell suspension. 3. Transfer half the cell suspension to a new tube. Return the cultures to the CO, incubator. 4. If necessary, change the medium the following day. The highest mitotic activity is usually observed in 48 h.

3.2. Amniotic Fluid Samples (See Notes 10-14)

3.2.1. Setting Up Cultures 1. Mix the sample by inverting the container several times, and pipet equal vols (5-7 mL) into two or more flat-sided Nunc tubes. 2. Centrifuge the tubes for 15 min at 25Og, and pipet off the supernatant fluid, whichis retained for a-fetoprotein or other enzymatic, and metabolic assays should they be requested.

80

Murer-Orlando

and McGuire

3. Tap the tube sharply to resuspend the cell pellet, and add 1.5 mL of cul-

ture medium to each tube. 4. Transfer to a CO, incubator, and leave undisturbed for 7 d. 5. Change medium. Subsequent medium changes should be made every 2-3 d, with the last change on the day before harvesting. Harvesting is usually possible in 10 d. 6. Harvesting and subculturing are the same as the chorion culture protocols.

4. Notes
1. Direct preparations have some technical disadvantages.

2.

3.

4.

5. 6.

7.

The mitotic rates vary widely between samples, there may be a high proportion of incompletemetaphases, and chromosomes are difficult to band. However, following the outlined procedure, preparations will show, on average, 10 complete metaphases per slide of sufficient quality to allow analysis of 200-300 band stage karyotype (1) in at least two cells. Semidirect protocols may also be used to obtain chromosome preparations. Samples can be incubated at 37C in transport RPMI for 1 or 2 d; however, mitoses have been recovered from intact villi incubated for as long as 4 d. To improve the chromosome morphology of direct preparations, blocking agents, such as FudR, can be introduced during incubation in an attempt to synchronize the dividing cells (2). Semidirect protocols have been used to study the cell cycle of the cytotrophoblasts by BrdU incorporation. These studies have shown that the cell cycle is long k36 h) with an extended late S phase of more than 10 h (3). The processing of individual villi allows a more accurate assessment in cases where the chromosome mosaicism is confined to the placenta. Chromosome preparations obtained following the described culture protocolareusually of good quality, withgoodmitoticrate andchromosome morphology, allowing analysis of 600-800 band-stage karyotypes (1). Other methods have been reported, mainly in situ techniques using cells growing on coverslips (4) or monolayer cultures obtained from enzymatically dissociated villi (5). If CVS material is not carefully selected, growth of cells from maternal decidua can present a serious problem in chorion cultures, especially when working with monolayer preparations from cell suspensions. It

CVS and Amniotic Fluid

is, therefore, advisable to set up separate cultures from discrete portions of the sample whenever possible. In order to ensure that chromosome analysis may be repeated if necessary, it is advisable to harvest the cells from the tube only if there is growth in the corresponding dish. If not, the tube must be subcultured. Use of Changs medium is recommended to establish chorion cultures. Even so, the initial cell growth will still be slow (3-4 d) in comparison to cultures from fetal tissues. Secondary cultures may be maintained in RPM1 with 20% FBS, but they deteriorate after a few passages. Following thedescribedprotocol,harvestingof AFculturesforchromosome analysis can be performed in 8-15 d. Preparations have an average of 10 cells/slide, analyzable at the 600-800 stage karyotype (I). Culture methods used for AF cells have been reviewed in detail by Martin (6). Cells are harvested on the growth surface, or the monolayer is trypsinized before being processed as a cell suspension. Results from in situ procedures are usually faster, but the quality of the chromosome preparation is less controllable. In situ procedures have also been described where poly-L-lysine has been used to coat the growth surface (7). Amniotic fluid contains very few cells that are viable in culture. Exceptionally slow cell growth may be related to extraneous factors such as toxic syringes or culture plastic ware, long delay in transport, or to grossly blood-stained samples. The use of Ultroser G supplement is recommended in cultures from blood stained fluids. Cultures from cloudy samples do not vary from those of clear AE. In AF cultures, different colonies are observed, and there is a tendency to classify the cells in three major groups: epithelioid, amniotic fluid, and fibroblas tic cells. Maternal cell growth can occur in AF cultures. It is advisable to use a simple protein test (Albym test-Boehring Catalog number 12 62 50) in all AF samples, to exclude the possibility that the fluid is urine. In blood stained samples, the proportion of maternal to fetal blood contamination can be determined using an histochemical test to differentiate cells carrying adult or fetal hemoglobin (Boeringer Mannheim, Tes t combination-Catalog Number 124-249). This test is recommended when a-fetoprotein analysis needs to be carried out on the same AF sample, since fetal serum will give a falsely raised aFP level.

8.

9.

10. 11.

12.

13.

14.

82

Murer-Orlando

and McGuire

References
1. Harnden,D. G.andKlinger,H.P.,eds.(1985)AnInternationalSysfemfovHumanCytogeneticNomenclature.Karger: Basel,Munchen, Paris, London, New York. Published in collaboration with Cyfogenet. Cell Genet. 2. Gibas, L. M., Grujic, S.,Barr, M. A., and Jackson,L. G. (1987) A single technique for obtaining higher quality chromosome preparations from chorionic villus samples using FdU synchronization. Prenat. Diag. 7,323-327. 3. Zahed, L., Murer-Orlando, M. L., and Bobrow. M. (1988) Cell cycle studies in chorionic villi. Hum. Genet.80,127-134. 4. Heaton, D. E.,Czepulkowski, B.H., Horwell, D. H., and Coleman, D. V. (1984) Chromosome analysis of first trimester chorionic villus biopsies prepared by maceration techniques. Prenat. Diag. 4,279-287. 5. Blakemore, K. J,,Watson,M. S.,Samuelson,J.,Breg, W. R., andMahoney,M. J. (1984) A method for processing first trimester chorionic villus biopsies for cytogenetic analysis. Am. 1, Hum. Genef. 36,1386-1393. 6. Martin, A. D. (1980) Characteristics of amniotic fluid cells in vitro and attempts to improve culture techniques. Clinic. Obsfet. Gynaecol. 7,143-173. 7. Rajendra, B. R., Sciorra, L. J., and Lee, M. (1981) A rapid culture harvest protocol for amniotic cell cultures. Hum. Genet. 59,416-418.

Chapter 9
Keratinocyte
Yvonne Barlow and

Culture
Richard J. Pye

1. Introduction
Thein vitro growth of keratinocytes has proved to be an important tool in the study of the normal biology and disease processes involving the skin, e.g., the influence of extrinsic regulators of growth and differentiation, effects of pharmacological agents, dermo-epidermal interactions, tissue antigenicity, and models of carcinogenesis. Fibroblasts adapt well to culture conditions. However, keratinocyte culture has been hampered by the inadequacy of media, which had hitherto been optimized for fibroblasts, and the overgrowth of cultured keratinocytes by more vigorous stromal elements. These difficulties were initially overcome by the use of lethally treated 3T3 feeder layers, which supported keratinocyte growth and inhibited overgrowth by fibroblasts (2). This tech nique, with or without minor variations, is still widely used today. It led to the supposition however, that undefined elements from the dermal components were essential to keratinocyte growth. Later, Eisinger et al. (2) demonstrated that, by altering seeding density, pH, and incubating conditions, keratinocytes could be grown in the absence of specialized substrates, conditioned medium, or media supplements. These procedures still required large seeding densities, and passage either on plastic or 3T3 feeder cells was limited. More recently, the development of the MCDB series of media (3,4) has permitted serum-free, clonal growth of keratinocytes with as few as 400 cells/cm2 of substrate. The optimization of nutrients, particularly the reduction in calcium concentration and the omission of serum, favors proliferation rather than differentiation, extending the life
83

Barlow and Pye

of the culture and permitting serial propagation. This model is particularly useful because of the absence of undefined supplements in the medium. The design or selection of the tissue culture system to be used does, however, depend on the question to be explored. Indeed, central to our understanding of epidermal-stromal interactions has been the development of the dermal equivalent model (5,6) as a useful working alternative to organ culture of skin. In this way, collagen lattices, seeded with living fibroblasts, are placed on rafts. Keratinocytes are then either incorporated into the lattice as an explant that grows over the surface, or seeded as a single-cell suspension on top of the lattice. These simplified skin models are cultured at an air liquid interface that promotes differentiation of keratinocytes. In this chapter, the techniques of keratinocyte culture on tissue culture plastics, extracellular matrices, 3T3 feeder layers, and in the dermal equivalent system are described.

2. Materials
Media and supplements may be purchased from any company supplying tissue culture products and must be of tissue culture reagent grade. 1. Keratinocyte medium: Keratinocyte medium is a mixture of DulbeccoVogts modification of Eagles Minimum Essential Medium and Hams F12 (3/l, V/V)-3DMEM:F12. (a) Basal medium-3DMEM:F12 DMEM powdered medium 10.03 g 4.77 g Hepes 0.80 g Sodium bicarbonate Hams F12 powdered medium 2.65 g Constituents are made up to 1 L with deionized, double distilled water, brought to pH 7.2-7.4 with 5M sodium hydroxide and sterile filtered. Sterility checks should be made (seeChapter 1, this volume). Alternatively, three parts single strength liquid DMEM containing 2OmM Hepes buffer are mixed with one part Hams F12 liquid medium and sterile sodium bicarbonate (0.8 g/L) and L-glutamine (2 mM> added before use. Glutamine is unstable at 4C and should be kept frozen. Its half-life in medium stored at 4C is approximately 3 wk, and medium stored for long periods should be supplemented with fresh glutamine. (b) Supplements to 3DMEM:F12 keratinocyte medium. Stock solutions.

Keratinocyte

Culture

85

(1) Epidermal growth factor (EGF). Dissolve 0.1 mg EGF from mouse submaxillary glands in 10 mL of distilled water (10 pg/ mL). Sterile-filter and store frozen in I-mL aliquots at -2OOC. (2) Insulin. Dissolve 100 mg of bovine insulin in 20 mL of 0.005N HCl (5 mg/mL). Sterile-filter and store in 1-mL aliquots at -20C. (3) Hydrocortisone. Dissolve 5 mg in I mL of 95% ethanol (5 mg/ mL) and store this concentrated stock at 4C. Add 0.4 mL of concentrated stock to 9.6 mL of serum-free medium, sterile-filter, and store in 1-mL aliquots at -2OOC. (4) Cholera toxin. Dissolve 1.0 mg of cholera toxin in 1.18 mL of distilled water (lOsM), Store this concentrate at 4OC. Add 0.1 mL of concentrate to 9.9 mL of medium containing 10% serum. Sterile-filter into I-mL aliquots and store at -2OOC. (5) Transferrin/Triiodo-L-thyronine (T3). Dissolve 100 mg of human transferrin in 12 mL of phosphate buffered saline. Dissolve 13.6 mg of T3 in the minimum volume of 0.2N sodium hydroxide, and make up to 100 mL in sterile distilled water (2 x WM). Add 0.2 mL of the T3 concentrate to 12 mL of transferrin, and make the total volume up to 20 mL with distilled water. Sterile-filter into I-mL aliquots, and store at -2OOC. (6) Adenine. Dissolve 243 mg in the minimum volume of lNHC1, and make up to 100 mL with sterile distilled water (1.8 x 10-2M). Sterile-filter and store in lo-mL aliquots at -2OOC. (7) Serum. Heat-inactivated fetal calf serum (FCS; 10%) is used to culture epithelial cells. All batches of serum must be tested for their ability to support keratinocyte growth at clonal densities (seeChapter 1, this volume) (Table 1). 2. Feeder cell medium: a. Basal medium-Dulbecco-Vogts modification of Eagles Minimum Essential Medium (DMEM). DMEM powdered medium N-2 Hydroxye thy1 piperazine N-2 ethane 13.37 g 4.77 g sulfonic acid (HEPES) Sodium bicarbonate 0.80 g Constituents are made up to 1 L with deionized, double-distilled water, brought to pH 7.2-7.4 with 5M sodium hydroxide and sterilefiltered. Sterile single-strength, liquid DMEM containing 20 mM Hepes can be purchased ready for use. Sterile sodium bicarbonate (0.8 g/L) and L-glutamine (2 mM) are added separately. b. Serum. Heat-inactivated newborn calf serum (lO%)is added to DMEM before use.

86

Barlow

and Pye

Substrate

Table 1 Keratinocyte Seeding Density on Plastic, Extracellular Matrix, and 3T3 Feeder Cells in Vitro Keratinocyte seeding density/80 cm2 flasks
5-10x106 3-5x106 1-2x106

Tissue culture plastic Extracellular matrix 3T3 Feeder cells

3. Antibiotics: Antibiotic x 10 solution x 1 solution Penicillin 1000 U/mL 100 U/mL Streptomycin 1000 &mL 100 pg/mL 2.5 pg/mL 2.5 pg/mL Fungizone 4. Dispase (Protease type ix from Bacillus polymyxu): Dissolve 250 mg dispase in 100 mL serum-free DMEM. Filter sterilize and store in lomL aliquots at -2OOC. 5. Trypsin: a. Routine subculture of cells: Trypsin (1:250) 0.25% solution containing 0.02% ethylene-diamine-tetra-acetic acid (EDTA) is made up in Hepes buffered calcium and magnesium-free Hanks balanced salt solution. b. Keratinocyte suspensions. Trypsin (1:300) 0.25% containing penicillin 200 U/mL, streptomycin 100 yg/mL, and 0.5 g/L sodium bicarbonate in calcium and magnesium-free Hanks balanced salt solution are stored in 5-mL aliquots at 20C. Before use, each 5-mL aliquot is made up to 20 mL with calcium and magnesium-free Hanks balanced salt solution. 6. Trypan blue solution for cell viability counts. Dissolve 400 mg Trypan blue, 810 mg sodium chloride, and 60 mg potassium dihydrogen orthophosphate. Adjust the pH to 7.2-7.3 with 1M sodium hydroxide, and make up to 100 mL with distilled water. 7. Dermal equivalents:
a. Medium for preparation of collagen rafts.

Eagles minimum essential medium powder Hepes Sodium bicarbonate Distilled water

9.78 g 5.0 g
2.0 g 535 mL

b. Collagen. Collagen can be purchased from a number of commercial sources. Pure Type I calf skin collagen (1 mg/mL) is

dialyzed against several changes of distilled water until the

pH is no lower than pH 4.5. Dialyzed collagen is stored in 5mL aliquots at -2OOC until required.

Keratinocyte Culture
c. Reconstitution of dermal equivaltents. Eagles MEM Collagen Sodium hydroxide (0.W) Fetal calf serum Fibroblast cell suspension (8 x lo5 cells/mL) 2.3 mL 1.5 mL 0.25 mL 0.45 mL 0.5 mL

87

3. Methods
All procedures using human tissue are carried out in a Class II laminar flow hood using aseptic technique.

3.1. Preparation

of Keratinocytes

1. Human skin may be obtained from surgical specimens removed at circumcision, mastectomy, apronetomy, or during cosmetic surgery. Not all skin grows equally well in culture. In our experience, skin from ears, hands, and feet are less successful than skin from breast, abdomen, or preputial skin. Growth rate is also affected by donors health and age. Keratinocytes from children and young healthy adults undergo more population doublings. The skin sample should be placed in a dry sterile container and stored at 4OC until collection can be arranged. For maximum viability, skin should be processed as soon as possible after excision. However, where necessary, skin can be stored in this way for several days with a small loss of viability. If prolonged storage is contemplated, a small quantity of phosphate buffered saline containing penicillin and streptomycin should be added, to prevent drying of the sample. 2. Fatty tissue should be removed from abdomen and breast samples; otherwise tissues should be directly immersed in 10 x strength antibiotic solution for 30 min. After several washes in serum-free medium, tissues are left to soak in single-strength antibiotic solution for a further 30 min (seeNote 1). 3. The skin is transferred to a sterile 90-mm Petri dish with the sample placed dermal side down. Using sterile scissors and forceps, the skin is raised in the forceps, and small pieces of skin (0.5 x 0.5 cm) are cut from the surface leaving behind as much of the connective tissue as possible. 4. The pieces of tissues are placed in Dispase solution at 4OCovernight with approximately 4-5 cm2 of tissue/ml of Dispase. For convenience, tissues may be left in Dispase solution for 48 h without significant loss of viability.

88

Barlow and Pye

5. After 24 h, the tissue incubated in Dispase solution is transferred to a sterile go-mm Petri dish. Using sterile forceps, the epidermis is peeled very easily from the dermis. If the epidermis does not peel easily, the tissue can be incubated for several hours at 37OC, or placed in fresh Dispase and incubated for a further 24 h at 4C. The pieces of peeled epidermis should be kept moist in calcium and magnesium-free Hanks balanced salt solution during this procedure. 6. The pieces of epidermis are transferred to a sterile universal containing 10 mL of prewarmed trypsin solution and incubated at 37C for 30-45 min. 7. After this time, the contents of the universal container are shaken once or twice toloosen the trypsinized cells, which are released into suspension. The solution is allowed to settle briefly, after which the pieces of epidermis float to the surface while cellular debris falls to the bottom of the universal. 8. The keratinocyte suspension is carefully pipeted into a centrifuge tube. To each 5 mL of keratinocyte cell suspension, 5-mL of DMEM containing 20% serum is added and the suspension centrifuged at 2008 for 5 min. 9. The supernatants are aspirated and 1 mL of fresh medium containing 10% serum added to each cell pellet. Using a plugged sterile Pasteur pipet, the cell pellets are very gently resuspended to form a single-cell suspension. IO. More fresh, prewarmed trypsin solution is added to the pieces of epidermis, and the process is repeated. Three trypsinizations may be performed to maximize cell yield. The cells harvested in this way are pooled and counted. 11. An equal volume of cell suspension and Trypan blue solution are mixed and left at room temperature for 5 min. Using the improved Nebauer hemocytometer, both chambers are filled by touching the edge of the coverslip with the pipet and allowing the suspension to be drawn up by capillary action. Avoid flooding the chamber. 12. Each hemocytometer chamber is divided into nine large squares delineated by triple white lines. The center square is further subdivided into 25 squares, which are again subdivided into 16 squares. Viable cells, i. e., those not stained blue, are counted in the two central squares of each chamber and the mean value for one large square calculated. The number of cells in one large square is multiplied by the dilution factor for Trypan blue, i. e., x 2 = n cells. Multiply the number of cells nby104 = cells/mL. The number of cells/mL multiplied by the number of mLs of cell suspension = total number of cells harvested.

Keratinocyte Culture

89

Fig. 1. Human foreskin keratinocyte culture 3 d after plating (200x1.

3.1.1. Seeding of Keratinoeytes

Keratinocytes may be seeded into tissue culture flasks, onto 3T3 feeder layers, or onto an extracelluar matrix.
3.1.1.1. Tissue Culture Flasks

1. The suspension of cells is diluted in an appropriate volume and seeded into plastic tissue culture flasks at a density of 6.25 x plastic 104 -1.25 x 10s cells/cm*. 2. After the cells have attached (1-2 d), cultures should be refed with complete keratinocyte medium every 2-3 d. Cultures should be fed with keratinocyte medium without EGF when first seeded into flasks. When cells approach confluence, they may require to be fed every day. Cultures are incubated at 37C. 3. Cells may appear flattened a few days after plating (Fig. l), but as the culture becomes denser, cells assume the typical pavement-like appearance of epithelium (Figs. 2,3). Cultures should approach confluence within 10-15 d and should be subcultured (seebelow) before they reach confluence. Cultures held at confluence differentiate, and blistering may occur.

90

Barlow and Pye

Fig. 2. Human foreskin keratinocyte culture 7 d after plating (200x).

Fig. 3. Human foreskin keratinocyte culture 12d after plating. Keratinocytes assume typical pavement-like structure as cultures approach confluence (200x mag).

Keratinocyte Culture

91

Fig. 4. Swiss 3T3 cells 2 d after seeding at 1.5 x 106 cells/80 cm2 flask (100x).

3.1.1.2. 3T3 Feeder Layers

1. 3T3 Swiss albino transformed mouse embryo fibroblasts (ref. no. CCL92) may be obtained from the European Collection of Animal Cell Cultures, PHLS Centre for Applied Microbiology, Porton Down, Salisbury. 3T3 cells are grown in DMEM plus 10% newborn calf serum. Cultures are routinely passaged two or three time/wk. 2. Prior to seeding 3T3 cells with keratinocytes, cells are passaged and grown to SO-70% confluence (Fig. 4). The 3T3 cells are then lethally irradiated with 6000 rads using any appropriate source or treated with mitomycin C (seeNote 2). Cultures SO-70% confluent are incubated with medium containing mitomycin C at 5-10 pg/mL (see Note 2) for 1 h at 37OC. 3. The cells are then thoroughly washed with serum-free medium before keratinocytes are seeded onto the 3T3 cells. 4. Keratinocytes are seeded at a concentration of 2.5 x lo cells/cm2, fed with keratinocyte medium without EGF, and incubated at 37OC. 5. After cells have attached, cultures should be fed with complete keratinocyte medium. Clones of epithelial cells may not be apparent

92

Barlow

and Pye

among the 3T3 feeder cells for several days, but with increasing incubation, small islands of 20-50 cells should be easily visible (Fig. 5). The cultures should approach confluence in lo-14 d and must be subcultured before cells reach confluence. Keratinocytes may be subcultured at seeding densities as low as 1.25-2.5 x 104cells/cm2 onto new lethally treated 3T3 cells. 3.1.1.3. Extracellular Matrix 1. ECM made in this laboratory from 3T3 cells or human foreskin fibroblasts is less successful in promoting growth of keratinocytes than commercially prepared bovine cornea1 endothelial cell matrix (7). Once confluent, these cultured cells were lysed with 0.5% Triton-X100 in 0.0125M ammonium hydroxide for 5-10 min and washed in balanced salt solution before use (Fig. 6). 2. Keratinocytes seeded onto bovine cornea ECM have up to two to threefold greater plating efficiency than on tissue culture plastic, and confluence may be achieved in up to 2 wk with seeding densities of 3.5-6.0 x lo4 cels/cm2 in 80 cm2 flasks. 3. The cultures are fed with keratinocyte medium without EGF for the first day, after which cultures are refed as required with complete keratinocyte medium. Cells should be subcultured before confluence is reached.

3.2.

Subculture

1. Nearly confluent flasks of keratinocytes are washed twice in calcium and magnesium-free Hanks balanced salt solution. 2. To each 80 cm2 flask, 5 mL of trypsin-EDTA solution is added, and the flasks incubated for 5-10 min at 37OC. 3. The cells are dispersed into solution by tapping the flasks. 4. Trypsinization is stopped by the addition of 5 mL of 3DMYEM:F12 + 20% FCS to each flask. 5. The cell suspensions are centrifuged at 200gfor 5 min and the supernatants aspirated. 6. The cell pellet is gently resuspended in 1 mL of 3DMEM:F12 + 20% FCS and the cells counted in a hemocytometer. The cells are diluted to an appropriate volume for subculturing onto various substrates (Table 2). 7. Cells are fed with keratinocyte medium 3DMEM:F12 + 10% FCS plus additives, but without EGF (Table 2). EGF is added to the cultures at the first feeding (7). Proliferation rate is dependent on culture conditions, age, health of donor, and the site of donor skin. Table 2 is a guide to seeding densities used in this laboratory on different substrates.

Keratinocyte

Culture

93

Fig. 5. Colonies of keratinocytes (K) growing on lethally after seeding at 2 x 106 keratinocytes/80 an2 flask (200x).

treated 3T3 feeder cells, 7 d

Fig. 6. Keratinocytes from 60-yr-old female after 3rd passage on extracellular from bovine cornea1 cells. The ghosts of the cornea1 cells can clearly be seen.

matrix

94 Table 2 Complete Keratinocyte Medium Supplement 3DMEM:F12 Serum (HIFCS) Adenine EGF Insulin Hydrocortisone Cholera toxin TransferrWT3 Vol 90 mL 10 mL 1.0 mL 0.1 mL 0.1 mL 0.2 mL 0.1 mL 0.1 mL

Barlow and Pye

Final concentration 10%

1.8 x 10-4M

10 ng/mL
5 pg/mL 0.4 yg/mL

10-OM
5 pg/mL/2

x 10-gM

3.3. Dermal

Equivalents

Keratinocytes attach to the surface of the dermal equivalent, divide, and grow to cover the collagen lattice. Differentiation into a multilayered epidermis occurs, and keratinization is much more complete in these cultures incubated at an air-liquid interface (Fig. 7). Dermal equivalents are made in this laboratory in 60-mm diameter Petri dishes, but can be scaled UD or down to the required size.
;.

2. 3.

4.

5.

6.

7.

Sufficient collagen, medium, serum, and human skin fibroblasts (as described in the Materials section) are mixed and poured into 60-mm diameter Petri dishes. The dishes are incubated in humidified boxes at 37OC, and within a few h, the collagen gels solidify. At this stage, a skin plug approximately 2-3 mm in diameter can be inserted into the gel, which upon contraction will hold the skin plug tightly in place. Keratinocytes migrate from the explanted tissue across the surface of the collagen/fibroblast matrix. The gels will normally contract within 2-3 d (Fig. B), but the degree and speed of contraction of the gel depend partially on the seeding density of the fibroblasts. The contracted rafts are raised onto transwell plates (Fig. 9). Rafts without epithelial plugs may be seeded with keratinocytes at this stage (3 x lo5 cells/ml). This procedure can be standardized using stainless-s tee10 rings. The cultures are then fed with enough keratinocyte medium to reach just below the level of the raft. The medium is drawn up by capillary action. Cultures are refed every 2-3 d.

Keratinocyte Culture

95

Fig. 7. Keratinocytes

cultured on dermal equivalent

(x100).

Fig. 8. Contracted collagen-fibroblast the gel was 60 mm (100x).

matrix 3 d after seeding. Original

diameter

of

96

Barlow and Pye

Fig. 9. Collagen-fibroblast matrix raised to air-liquid interface on stainless-steelgrid

cmoxh

4. Notes
1. Several other antibiotics have been tested in this laboratory for use with keratinocyte cultures. Vancomycin was found to be very effective when tested against ten methicillin-resistant strains of Staphyloccocus and has very low toxicity to cultured keratinocytes. Washing in medium containing 1 mg/mL vancomycin for 30 min is recommended for infected cultures. After several washes in serum-free medium, cultures can be incubated with 100-200 pg/mL of vancomycin. Vancomycin is only recommended in special circumstances and should not be used routinely. 2. Mitomycin C, which disrupts microtubule formation in cells and prevents completion of cell division, is often used as an alternative to

Keratinocyte

Culture

97

lethal irradiation. Potency and toxicity of different batches of mitomytin C varies, and each batch should be tested. In this laboratory, mitomycin C has been found to be effective at concentration of 5-10 pg/ mL. If none of the available flasks of 3T3 cells are at a suitable stage, a fully confluent culture of 3T3 cells can be lethally treated, and subcultured into a flask freshly seeded wi th keratinocytes the same day or the following day with beneficial results.

Acknowledgments
Our thanks to Ms. Debbie Coakley for excellent technical assistance. Our grateful thanks to Dr. J.K. Wright for assistance with photography and to Dr. T. E. Cawston, Rheumatology Research Unit, Addenbrookes Hospital, Cambridge for the collagen used in the dermal equivalent.

References
1. 2. Rheinwalg, J. and Green, H. (1975) Serial cultivation of strains of human epidermal keratinocytes: the formation of keratinizing colonies from single cells. Cell 6,331339. Eisinger, M., Lee, J. S., Hefton, J. M., Darzynkiewicz, Z., Chaio, J. W., and Deharven, E. (1979) Human epidermal cell cultures: Growth and differentiation in the absence of dermal components or medium supplements. Proc. N&l. Acud., USA 76,53405344. Tsao, M. C., Walthall, B. J., and Ham, R. G. (1982) Clonal growth of normal human keratinocytes in a defined medium. J. Cell Physiol. 110,219-229. Boyce, S. and Ham, R. G. (1983) Calcium-regulated differentiation of normal human epidermal keratinocytes in chemically defined clonal culture and serum free serial culture. J. Invest. Dermatd. 81,533-540. Bell, E., Ehrlich, H. P., Sher, S., Merrill, C., Sarber, R., Hull, B., Nakatsuji, T., Church, D., and Buttle, D. J. (1981) Development and use of a living skin equivalent. J. Plastic and Reconstructive Surgery 67,3&S-392. Bell, E., Sher, S., Hull, B., Merrill, C., Rosen, S., Chamson, A., Asselineau, D., Dubertret, L., Coulomb, B., Lapiere, C., Nusgens, B., and Neveux, Y. (1983) The reconstruction of a living skin. J. lnvesf. Dermafol. 81,52-510. Gospadarowicz, D., Mescher, A. L., and Birdwell, C. R. (1977) Stimulation of cornea1 endothelial cell proliferation by fibroblast and epidermal growth factors. Experimental Eye Res. 25,75-89.

3. 4. 5. 6. 7.

Chapter

10

Establishment of Mouse Epithelial Lung Cell Strains and Cell Lines

Garry J. Smith

1. Introduction
Mesenchymal cells have been cultured successfully for many years owing to the ease of their culturability. With the advent of more complex growth media, and some changes to methods, it has become possible to culture epithelial cell types from a wide variety of rodent and human tissue sources successfully (1,2,3). We have used medium CMRL 1066 with serum to establish and cultivate lung cell lines that are related to the type II pneumocyte of lung alveolus (4-7). Other laboratories have successfully cultured different types of lung cells, including tracheal and bronchial cell types (1,2,3,8). A critically important factor in satisfactory culture of any cell type is the ability to observe markers that will identify the cell type of origin of the culture. In some cell types, ultrastructural markers at the electron microscope level can be very useful (for example, presence of microvilli and properly structured desmosomes can indicate epithelial origin), and some cells may exhibit specific structures (such as intracellular surfactant lamellar bodies in type II pneumocytes, or Weibel-Palade bodies in endothelial 99

100

Smith

cells) that denote cell type of origin. However, ultrastructural markers can be misleading, for example, desmosome-like bodies occur in a variety of cultured cell types (1). Therefore, antibodies to specific ultra-structural features can be more reliable. Identification by antibodies to keratinor prekeratins is an extremely valuable marker denoting epithelial origin, as these intermediate filaments tend not to be expressed in culture bynonepithelial cell types. Furthermore, it may be possible to prepare antisera to cell-specific proteins (9,ZO). The major problem with identification of cells in culture is the general observation that differentiated cell types do not culture well. Secondly, because of the similarity of culturing procedures and media formulations, cells from different tissue types tend to be placed under the same selection pressures in culture. Therefore, cells of any tissue type in culture tend to grow towards a similar end point with respect to differentiated characteristics. Cell culturing can be a particularly useful means for isolating purified cell types for further study, provided that these limitations in ability to culture cells of a desired type are taken into consideration. This present chapter will use culturing of type II pneumocyte-related cells of the lung alveolus as an example of epithelial culturing procedures. Several textbooks have very good overviews of cell culturing methods (11,12), We will consider in this chapter the culturing of type II pneumocyte-related cells from both normal lung and from chemically induced lung adenomas. This chapter does not contain information on basic cell culture techniques, such as media preparation, cell cloning, cryopreservation, cell passaging, prevention of cell-line cross-contamination, laboratory design, and laboratory safety, which are dealt with elsewhere in this volume.

2. Materials
The powdered medium CMRL 1066 is made up freshly or at most within several days of use. Serum is kept frozen and added to the medium as required together with antibiotics. The procedure in our laboratory is to have the stocks of liquid medium, serum, antibiotics, and so on, made up in a laminar flow hood as independent stocks that have been filtered through 0.2 micron filters in order to eliminate prokaryotic contamination. Upon full formulation of the medium with serum and antibiotics, the culture formulation is again filtered through 0.2 micron filters in order to provide a second stage of sterilization. Enzymes such as trypsin are kept as concentrated frozen stocks to be subsequently diluted with independ-

Lung Cell Strains and Cell Lines

101

ently sterilized phosphate buffered saline (II), and this formulation is subsequently refiltered and kept at 4C. Media constituents glutamine and fungizone both have a life of only several days in unfrozen liquid formulations. Formulations for use in explanting and culture of type II pneumocyte related cell lines from mouse lung are as follows: 1. Phosphate buffered saline (PBS): Made as a 10x concentrate without calcium and magnesium consisting of (as w/v) 0.2%KCl, 0.2% KH,J?O~ 8% NaCl, 1.15% Na,HPO,, pH 7.0-7.4 (with NaOH). 2. The complete formulation of culture medium, designated medium A, consists of medium CMRL 1066 (seeAppendix), 10% (v/v> fetal calf serum, 2.5 mg/mL Fungizone, 100 IU/mL penicillin, 100 mg/mL streptomycin, and 2 mM L-glutamine. 3. Trypsin is used at 0.25% (w/v) in calcium and magnesium-free PBS.

3. Methods
The animals of choice for theseexperiments are specific-pathogen-free animals. The initial dissection and explanting is performed under semisterile conditions in a room that has preferably been presterilized by ultraviolet light. The possibility of the operator contaminating the preparation tissue material is minimized by the wearing of a mask or performance of the explanting of tissue underneath a plastic screen standing in front of the operator.
1. One adult female Balb / c mouse is killed in an ether chamber. The ven-

tral surface of the animal is swabbed well with alcohol, the neck region dissected, and the trachea tied off tightly with fine cotton gauze in order to prevent lung collapse upon opening the chest. The chest region is then opened by a ventral incision from just above the diaphragm through the rib cage and up to the neck area. The rib cage on either side is removed with scissors and forceps to expose the inflated lungs. 2. An optional procedure at this stage is removal of blood cells by perfusion of the lungs with medium A via the right ventricle of the heart. However, this procedure does not enhance the culturing method. A 1 mL syringe is used to infuse the lungs gently via the trachea below the thread tie with 1 mL of calcium and magnesium-free PBScontaining 0.25% (v/v) trypsin. The trachea is then tied off below this needle

102

Smith incision. This lung infusion procedure increases the yield of alveolar cells in culture, but is optional for other epithelial cell types (I). The lungs and heart are removed by dissecting through the trachea with scissors above the tie and gently pulling upwards on the trachea, while dissecting connective tissue away from the dorsal part of the heart and lungs. The inflated lungs come free of the body wall, and may be placed into a solution of medium A containing antibiotics and Fungizone in order to wash off excess blood and to minimize contamination from airborne prokaryotes. The tissue is then taken to a laminar flow hood where it is put into a glass Petri dish and a few drops of medium A are placed upon the tissue. The lungs are dissected free of the other tissues and isolated in the Petri dish. The moist tissue is chopped into very fine pieces of maximum size 1 mm3 using two sharp, sterile scalpels drawn across the tissue in a cross-chopping fashion. The mashed small tissue fragments may then be transferred to a sterile, capped plastic centrifuge tube, and 4 mL of calcium and magnesium-free PBS containing 0.25% (v/v> trypsin added. The lung tissue is swirled to allow complete distribution within the trypsin solution. This tissue suspension is incubated at 37OC for 15 min in a water bath with occasional gentle hand mixing. The lung tissue generally distributes between small tissue pieces floating on the top surface still associated with air or small tissue pieces that sink in the solution. Following the 37OCincubation, the tissue suspension is placed on ice for several minutes to cool. Overtrypsinization can lead to cell lysis and viscous DNA release. Centrifugation at 2000 rpm for 5 min in a bench centrifuge results in a small pellet of tissue pieces that is resuspended in 15 mL of medium A. The upper floating fraction of lung, which represents lung not in close contact with the trypsin solution during infusion, is discarded. The resuspended small tissue pieces are equally distributed into three 25 cm2 plastic culture flasks and the primary tissue explants incubated at 37OC in a 5% (v/v> CO,-air atmosphere. The small size of the chopped tissue fragments is crucial to the explant, since only these small pieces are accessible to the medium and the atmosphere, and so will remain viable. At twice weekly intervals, the 25 cm2 flask is briefly placed on its side at an angle so that nonattached tissue fragments may settle to the

3.

4.

5.

6.

7,

8.

9.

Lung Cell Strains and Cell Lines

103

Fig. 1. A dark tissue explant with spindloid fibroblasticoutgrowth followed by cobblestone-like epithelioid cell outgrowth (arrowed). Six weeks after primary culture. Phase contrast 165x.

bottom corner of the flask, half the medium A is removed with a Pasteur pipet, and is replaced with fresh medium A. This is referred to as a twice weekly half-medium change. 10. After 3-7 d, the small tissue pieces should attach to the plastic bottom surface of the flasks, and fibroblastoid cells emerge onto the plastic from the tissue piece followed by more epithelial-like cells (Fig. I). The easily identified outgrowths reach a static phase or grow very slowly after 1-2 wk in culture. The most desirable situation is when cobblestone-like epithelioid cells establish themselves in the explant and can be easily distinguished from the spindle-shaped fibroblastoid cells (Fig. 2). In the case of cells explanted from chemically induced lung adenomas, an alternative primary culture behavior may occur. Here the adenoma cells may appear as small, discrete foci of polygonal cells emerging over wide areas of the plastic culture surface and not deriving directly from small explanted tissue islands (Fig. 3). 11. In either the tissue focus or emerging island situation, the epithelioid

Smith

Fig. 2. Mixed epithelioid (arrowed) and fibroblastoid in primary culture. Phase contrast 330x.

cells early after cell outgrowth

Fig. 3. Foci of multinuclear epithelioid cells appearing tissue after two weeks. Phase contrast 330x.

in primary

culture of adenoma

Lung Cell Strains and Cell Lines

105

12.

13.

14.

15.

cells must be preserved free from overgrowth by fibroblastoid cells. This may be efficiently achieved in a semis terile room by scraping the cultures under an inverted microscope using a drawn-out, pointed Pasteur pipet that has been sterilized by flaming. With practice, such a fine glass probe can be used to detach encroaching fibroblas toid cells without damaging the epithelioid cells nearby. The detached cells may then be washed free of the epithelioid cells and removed from culture with a Pasteur pipet, and the normal conditioned or halfchange medium of the cultures replaced onto the epithelioid cells. Cultures are maintained on twice weekly half-medium changes. The factors that will determine the success of the epithelioid cell culture are a. the ability to identify discreet epithelioid colonies separated from fibroblastoid cells; b. the ability to prevent fibroblastoid overgrowth of cultures successfully; and c. the transition of the epithelioid cells to growth in culture. Transition to growth in culture, which has occurred in some 20% of culture experiments in our procedures, is observed as a proliferation of the epithelioid cells a month or so after primary culture. It is a phenomenon that appears to be caused by minor rearrangement of the mouse cell genome as the cells persist or slowly divide in culture. Only with such transition to growth in culture can the epithelioid cells be successfully established as growing cell lines and cell strains. In the event that the cells do demonstrate continued growth in culture, the epithelioid cell colonies, when they have grown to some 50 cells or more and are showing continuousgrowth in culture, may be selectively trypsinized away from other cell types. Having removed the normal culture medium, a local application of 1 mL of 0.25% (v/ v) trypsin for one or two min using a Pasteur pipet lifts the cells off the plastic. The cells, usually accompanied by some fibroblasts, may then be transferred to a new culture flask and fed with normal conditioned medium or half-change medium (Fig. 4). However, it is advisable to attempt to maintain the cell density in culture at its highest practical level during these procedures. This may be done by culturing the passaged cells into small-well plastic culture dishes. In the event that a primary culture remains static and epithelioid transition to growth does not occur, trypsinization, after some 4 wk in primary culture of the whole culture, and transfer to a new flask can encourage epithelioid growth.

106

Smith

Fig. 4. A focus of epithelioid cells that have been selectively trypsinized from primary culture.

16. The most desirable endpoint for passaged explants is where the epithelioid cells continue to grow, with elimination of fibroblasts using a glass rod as necessary, such that theepithelioid cells dominate the culture and approach confluence within several months. During this phase, the epithelioid cells should maintain a healthy morphology, with clearly polygonal cell shape, and clear nucleus and nuclei under phase contrast microscopy, and a slow but definite growth rate in culture (Fig. 5). Tumor-derived cells generally exhibit a high nuclear to cytoplasmic ratio (Fig. 3). 17. The next stage in cell culture is a verification of the derivation of the growing cells. Some indication of the epithelial nature of the cells can be obtained from overall morphology using the inverted light microscope. Phase contrast microscopy should show polygonal cell types and, in the case of nonneoplastic cells, a relatively low nuclear to cytoplasmic ratio. The cells should be quite visible, and show discreet nuclei and vacuoles using phase contrast microscopy (Fig. 5). In general, epithelial cell cultures from nonmalignant cells will grow in closely associated islands of cells forming a polygonal cobblestone-

Lung Cell Strains and Cell Lines

107

Fig. 5. Proliferating culture of epithelioid cells that have transformed culture and are approaching confluence. Phase contrast 330x.

to growth in

like pattern. At the earliest time, as cell numbers increase, a sample of cells should be taken for electron microscopy using a transmission electron microscope. The electron microscopy should again indicate polygonal cell structure on a trypsinized population with cells showing microvilli or other epithelial-type ultrastructural marker (Fig. 6). Particularly useful, especially on cells early in culture, is the ability to see spot-desmosomes between cells that have not been trypsinized apart. However, it is critical that the desmosome be of a correct ultrastructural type (2) with a central lamella and the presence of lonm filaments moving away on either side of the central lamella. One of the best indications of epithelial nature is the ability to identify prekeratin or keratin using specific antisera to this intermediate filament protein family. Immunological identification (seeChapter 35) of keratin on cell pellets that have been sectioned in the usual histopathology fashion or on cells that have been grown on glass coverslips and permeabilized with 0.3% Triton X-100 is a good indication of their epithelial nature (6,7). Finally, it may be possible to identify cells as coming from a particular epithelial cell lineage by

108

Smith

micrographof of epithelial cells in culture exhibiting surface microvilli Fig. 6. Electron micrograph (closed arrows), osmiophilic lamellar bodies (open arrows), and polygonal cell shape 123lOx.
identification with cell-specific antisera. In the case of type II pneumocyte-related cell strains cultured in our laboratory, we have identified them as epithelial based upon ultrastructural ultrastructural observation of microvilli and identification with an antiserum antiserum to prekeratin prekeratin Q&-7), Q&-7), and as type II pneumocyte-related by electron-microscopic observation of osmiophilic lamellar bodies and by identification with a polyclonal antiserum that reacted with only type II pneumocytes and pneumocyte-related cell types (4,ZO). It has been frequently observed that some differentiated features of the primary culture, notably ultrastructural lamellar bodies, are not well retained beyond passages of around passage number 30.

4. Notes
1. An alternative culture procedure that has been found suitable for type II pneumocyte-related cell strain culture, and may be more suitable for

Lung Cell Strains and Cell Lines

109

some epithelial cell types, is as follows: The minced tissue may be directly diluted into 3 mL of the PBS trypsin solution for digestion and, after digestion, the tissue solution gravity-sedimented for 1 min at room temperature by standing in the capped, sterile centrifuge tube on the bench. This allows large tissue pieces, which may subsequently necrose in culture, to go to the bottom of the tube and the smaller tissue fragments, which do not sediment under gravity after 1 min, to be equally distributed between three 25-cm2 plastic culture flasks each containing 4 mL of culture medium A. The floating lung fragments again are discarded in this preparation procedure. 2. Prevention of overgrowth of cultures by fibroblasts can be one of the most difficult procedures in establishing epithelial cell strains. Some inhibitors can be used for prevention of fibroblast overgrowth, for example cis-hydroxy-proline (13) or dexamethasone (14). However, the results of using fibroblast inhibitors can be variable, and we have found that mechanical elimination of the flat, stringy fibroblastoid cells using sterile glass rods can be very effective provided that the epithelioid cells transform to growth in culture. 3. At early stages of passaging from primary culture, it is advisable to minimize the dilutionof cells and to passage them at l-2or at most 1-5 dilutions for rapidly growing cells. It is extremely important that the cells, as they are passaged and with increasing passage number in culture, are subjected to a constant and consistent culturing regime. Therefore, it is of advantage to passage the cells at the same or similar dilution ranges from passage to passage and, furthermore, never to allow the cells to become too confluent prior to passaging. These procedures minimize the possibility of karotypic changes in culture, which may occur if cells are lef t at confluence, thereby minimizing loss of differentiated features of a given cell strain. 4. The fundamental needs for setting up epithelial cell culture, and in this case culture of type II pneumocytes, is a good laboratory setup in which sterile conditions may be maintained together with thorough and consistent culture protocols. Patience in waiting for cells to transform to growth in culture, often required for nonmalignant cell types, and the ability to keep strict and thorough cell records are very important. A good characterization of cells for their general morphology, their ultrastructure by electron microscopy, and their specific differentiated characteristics, such as expression of intermediate filament proteins or special differentiated cell functions, should be per-

110

Smith

formed early in the life of the culture and routinely thereafter during the life of the cell lines. Surveillance preventing contamination by prokaryotes and for any possibility of cross-contamination of cell types is most important in maintaining the integrity of the cell lines. 5. The main pointers that can prevent time wastage and ineffectual cell culture are a. do epithelioid cells appear in culture after l-2 wk of primary tissue explant? b. do fibroblasts overgrow the epithelioid cell types within a month of culture? c. do epithelioid cells transform to growth in culture either in their own right after a month or so in culture or within a few weeks of a trypsin passaging of the primary explant and d.can epithelioid cells be confirmed as epithelial either ultrastructurally or by expression of epithelial specific markers, such as keratins, and can specific cell type markers be identified?

Acknowledgments
This work has been supported by The National Health and Medical Research Council and The New South Wales State Cancer Council.

References
1. Franks,L.M.andWigley,C.B. (eds.)(1979)NeoplasficTrunsformutioninDilferentiuted EpifheliaI Cell Systems in Vitro (Academic Press,London). 2. Harris, C. C. (1987) Human tissuesand cells in carcinogenesis research. Cancer&s. 47,1-10. 3. Smith, G. J, and Stewart, B. W. (eds.) (1982) In Vitro EpitheZiuICeILDifferentiation and Neoplusiu (TheAustralian CancerSot., Sydney). 4. Smith, G. J.,LeMesurier, S.M., DeMontfort, M. L., and Lykke, A. W. J. (1984) Establishment of epithelial cell strains from normal adult mouse lung resembling a urethane-induced lung adenoma cell strain and a metastasing mouse lung carcinoma cell line. Cell Biol. International Reportst&161-169. 5. Smith G. J.,Le Mesurier, S.M., DeMontfort, M. L., and Lykke, A. W. J. (1984) Development and characterization of type II pneumocyte related cell lines from normal mouse lung. PuthoIogy16,401-405. 6. Smith, G. J. and Lykke, A. W. J. (1985) Characterization of a neoplastic epithelial cell strain derived by dexamethasone treatment of cultured normal mouse type II pneumocytes. J. of Pathology147,165-172.

Lung Cell Strains and Cell Lines

111

7. Smith, G. J., Bennett, F. A., Steele,J. G., and Bentel, J. M. (3986) Onset of neoplastic phenotype in an epithelial cell strain from adult BALB/c mouse lung alveolus. J,
Natl. Cancer Inst. 7673-79.

8. Nettesheim, P.,Terzaghi, M., and Klein-Szanto, A. J.P. (1982)Development andprogression of neoplastic disease. Morphologic and cell culture studies with airway epithelium, in Mechanisms of Chemical Carcinogenesis (Harris, C. C., ed.1, Liss, New York, pp. 473-489. 9. Ward, J. M., Singh, G., Katyal, S.L., Anderson, L. M., and Kovatch, R. M. (1985) Immunocytochemical localization of the surfactant apoproteinand Claracell antigen in chemically induced and naturally occurring pulmonary neoplasms of mice. Amer. I: Pathol. 118,493-499. 10. Smith, G. J.,Kumar, R. K., Hristoforidis, C. P., and Lykke, A. W. J. (1985) Expression of a type II pneumocyte-specific antigen by a cell strain from normal adult mouse lung. Cell Biol. International Reports 9,1115-1122. Il. Paul, J. (1975) CeZZ and Tissue Culture, 5th Ed. (Churchill-Livingstone, London). 12. Kuchler, R. J. (1977) Biochemical Methods in Cell Culture and Virology. (Dowden, Hutchinson and Ross,Stroudsburg, Pennsylvania). 23. Kao, W. W.-Y. and Prockop, D. J. (1977) Proline analogue removes fibroblasts from cultured mixed cell populations. Nature 266,63-64. 14. Edwards, A. M. and Lucas, C. M. (1982) Gamma-glutamyltranspeptidase as a preneoplastic marker in hepato carcinogenesis: Expression in hepatocytes isolated from normal and carcinogen-treated rats, in In Vitro Epithelial Cell Differentiation and Neoplasia (Smith, G. J. and Stewart, B.W., eds.) The Australian Cancer Sot., Sydney.

Chapter 11 Culture of Human Brain Tumors on an Extracellular Matrix Derived from Bovine Cornea1 Endothelial Cells and Cultured Human Glioma Cells
Manfred Westphal, Marianne H&nsel, Hildegard Nausch, Eva Rohde, and Hans-Dietrich Herrmann
1. Introduction
Cellculture has becomeanintegralpart of thedailyroutineof most oncology laboratories. It has enabled researchers to investigate a wide range of cellular parameters in a defined system in which the experimental conditions can be controlled and repeated. Although the manufacturers of tissue culture materials are continually improving their products, cell attachment and initial survival of primary cultures from tumor cells are still problems in many laboratories. Many different approaches have been taken to circumvent this problem, and coating of tissue culture dishes with 113

114

Westphal et al.

attachment enhancers, such as polyamino acids (I), fibronectin (Z), laminin (3), and collagen (4), has been found helpful. For a long time, it was known that endothelial cells produce abasement membrane, and this led to the use of bovine cornea1 basement membrane by Gospodarowicz et al. (5,6), in their research into the phenomenon of regeneration and nonregeneration of cornea1 endothelium in different species. The application of this bovine cornea1 extracellular matrix (bECM) has sincebeen greatly expanded (5,6). bECM has found broad approval, and has been used for mammary carcinoma (7), urological tumors (a), and different kinds of pituitary adenomas (9JO) as well as CNS tumors (II). It is generally recognized that interaction of cells, normal or neoplastic, with ECMs may interfere with the cellular biology (12,13) and that matrices from different sources may affect cells in a variety of ways. It could be shown, for example, that an ECM derived from human arachnoid cells induces differentiation in a glioma cell line (14). It should thus be mentioned in this introduction that many different cells will be able to produce matrices, including tumor cells themselves, and that the use of ECMs in oncology will not only facilitate initiation of cell cultures, but is an interesting research field of its own as will be illustrated by some observations obtained with an ECM derived from a cell line that was derived from glioblastoma.

2. Materials 2.1. ECM from Bovine CorneaL Endothelium,

bECM

1. Bovine eyes: following the instructions from protocols published earlier and elsewhere (6,X), fresh bovine eyes have to be obtained from the local slaughterhouse (seeNote 1). 2. Isolation of the cornea (seeNote 2): it requires a. no. 1 cannulae b. angled microscissors c. foreceps d. a grooved director e. phosphate buffered saline (PBS). All instruments should be sterile. 3. Cell culture medium: Dulbeccos modification of Eagles medium with high glucose content supplemented with 10% fetal calf serum, glutamine, (1 mM), pyruvate (2 n&l), gentamycin (25 pg/mL), and fungizone (2.5 pg/mL). 4. ECM production: a. dextran (MW 40,000)

Human Brain Tumors on ECM b. basic bovine fibroblast growth factor (FGF) c. sterile ammonium hydroxide solution (20 mM) d. sterile PBS e. sterile PBS containing gentamycin (25 bg/mL). 5. FGF isolation (see16): a. 4 kg of bovine brain (seeNote 3) b. 8 L of 0.15M ammonium sulfate pH 5.6 c. 6N HCl d. Ammonium sulfate e. Dialysis tubing (MW cutoff approx. 10,000) f. CM-Sephadex C-50 g. Buffers: O.lM sodium phosphate pH 6.0, O.lM sodium phosphate pH 6.00,15M NaCl, O.lM sodium phosphate pH 6.00,6M NaCl. 6. Tumor cell isolation (seeNote 4): a. Hanks balanced salt solution, Ca*+and Mg2+ free b. angled microscissors c. 5 mL tissue culture pipets with smoothened openings (best from Falcon or Costar) d. Enzyme cocktail: Pronase (0.05%), Collagenase (0.03%), DNase (O.Ol%), in Hanks filtered sterile buffer.

115

3. Methods
3.1. Preparation of bECM 1. Only eyes that are uninjured, and show no signs of inflammation or ulceration should be used. They are taken with the left hand,rinsed thoroughly with alcohol from a spray bottle, and then held down firmly on a paper towel that is also soaked in alcohol. The eye is then punctured tangentially with a cannula, so the chamber water can drain and the cornea can shrink after the pressure is taken off the eyeball. Starting from the needle insertion, the cornea is carefully cut out without contaminating it with other tissues from within the eyeball (seeNote 2). 2. Once the cornea is removed,it is put external surface down onto a Petri dish on which it will stick because of adhesive forces. It is then rinsed with PBS. Then the exposed inner surface is scraped once with the grooved director, and by this maneuver, sheets of endothelial cells will detach. The grooved director is then dipped into a tissue culture dish already containing the complete cell culture medium. Use 1 dish/eye. Process ten eyes in one session.

116

Westphal et al.

3. The dishes are then placed in an incubator and should not be disturbed

for 5 d. At this point, the first colonies will have formed. The different stages in the production of bECM are illustrated in Fig. 1. 4. After the primary cultures have become about 50% confluent with colonies, these are detached with trypsin, which sometimes takes up to 20 min because the cells attach firmly to their support. 5. The total cells are next distributed to 25 dishes that have been precoated with gelatin. To achieve this, 0.2% gelatin is dissolved in PBS and autoclaved. It is advisable to filter this solution because it may still contain granular matter that disturbs the cell culture. The IO-cm diameter cell culture dishes are covered with 5 mL of the PBS-gelatin and left in the refrigerator overnight. Prior to use, the excess PBSgelatin is aspirated. The advantage of this gelatin coating is that, at the next step, the cells will be very easy to detach. 6. In the next step, the cells from the precoated dishes are passaged at the point when they are just subconfluent. The dispersed cells harvested from eight to ten lo-cm dishes are added to a 550-mL bottle of complete medium, which in addition is supplemented with 5% dextran (MW 40,000, dissolved by heating in a small volume of medium without additives and sterile filtered). To this suspension, basic bovine fibroblast growth factor (bFGF) is added (seeNote 5). From this suspension, the cells are seeded onto the culture dishes to be covered with ECM. Six mL are added to a T25 flask, 25 mL to a T75, and 1 mL/well in a 24-mL multiwell tissue culture dish. At this density and in the presence of bFGF, the cells should reach confluence within 2-3 days. After confluence, the cells are left for one more week to ten d and then they are lysed. 7. The cells are lysed by aspirating the culture media and replacing it with distilled water containing ammonium hydroxide (20 mM). The cells burst rapidly by osmotic shock, and within 3 min, the gelatinous debris can be removed and rinsed off. The ECM is left behind and should be washed at least once more with PBS. The ECM-coated dishes can be stored at 4OCwhile being covered with PBS containing gentamycin (25 pg/mL). At that time, the matrix can be seen with a phase contrast microscope (Fig. 1, and seeNote 6).

3.2. ECM from a Human Brain WC2M)

Tumor

CeZZ Line

Cells from a human glioblastoma cell line that is strongly positive for cell surface fibronectin are used in early passages when the cells still grow

Human

Brain Tumors on ECM

117

Fig. 1. Different stages in the production of ECM. (A) A sheet of endothelial cells as it is scraped off the cornea. (B) A colony of endothelial cells after 6 d. (C) Subconfluent layer of bovine cornea1 endothelial cellsin the stage of bECM production. (D) Appearance of bECM in the phase contrast microscope after lysis of the cells.

in an orderly arrangement. The cells cannot be used for tECM production after passage 20 (Fig. 2, andse.eNote 7). Cell culture conditions are the same as for the preparation of ECM except that Earles modified MEM is used as a basal medium.

3.3. Isolation

of FGF

Of great importance to the production of any ECM appears to be the addition of FGF to the cultures, since this speeds up the rate at which the cultures become confluent and also influences the quality of the matrix. The isolation of basic FGF (bFGF) has been described by Gospodarowicz, (26) but is is not necessary for the production of ECM to purify bFGF to homogeneity. 1. Eight frozen brains are homogenized sulfate pH 5.6. in 8 L of 0.15M ammonium-

118

Westphd et al.

gliom in two different stagesof culture. (A) In passage tECM and grow in streams of parallel arranged cells in monolayer. (B) At a later stage,passage 46, in which the cells grow on top of each other and tECM is produced but comesoff the plastic support. Fig. 2. 7he KM-producing
15, in which they produce

2. The homogenate is adjusted to pH 4.5 with 6N HCl and stirred for at

least 1 h at 4C. 3. The homogenate is centrifuged at 20,000 x g, and the burgundy colored supematants collected. CAUTION: If the color has turned brown because the pH was too low, even if this was only for a short moment, the procedure should be stopped and started anew because bFGF is very intolerant to acidic treatment. is now added slowly to the extract, with stirring, 4. Ammonium-sulfate to give a concentration of 200 g/L and the suspension again cenfrifuged at 20,000 x g after having been stirred for 60 min. 5. Again the supernatants are collected, and again ammonium-sulfate is

Human

Brain Tumors on ECM

119

added to give a final concentration of 450 g/L. After centrifugation, the supernatants are discarded and the pellets collected. 6. The collected pellets are dissolved in a small volume of water (approx. 250 mL), adjusted to pH 6.0 with formic acid, and dialyzed against water. (Use dialysis tubing with mol wt cutoff between 10,000 and
12,000.) 7. The dialysate is then loaded onto a CM Sephadex C50 column (3.5 x 20

cm) that has been preequilibrated with O.lM sodium phosphate pH 6.0. The ionic strength of the dialysate should be equal to or less than that of the running buffer. 8. The material retained on the column is eluted with a stepwise increase of NaCl concentration in the sodium phosphate buffer, first to 0.15M and then to 0.6M. 9. The material eluting at the high salt concentration is collected. It is less than 0.1% pure, but this is sufficiently pure for use in the production of ECM. It can be stored frozen as well as lyophylized until reconstitution. The concentration at which the extract has to be used should be tested in a bioassay on endothelial cells. It is variable between extractions.

3.4. Isolation

and Culture

of Brain

Tumor

Cells

Brain tumors for this study will be divided into meningeomas, gliomas, and other tumors (such as acoustic neuromas, pinealomas, pituitary adenomas, meduloblastomas, or gangliogliomas). (Seealso notes on individual tumor types.) First the cell isolation and cultivation is described. 1. Representative portions of tumors are obtained at surgery (seeNote8). The selected tissue fragments are placed immediately into 15 mL of sterile Hanks balanced salt solution without calcium and magnesium (CMF). NOTE: In this solution, the material can be maintained at least for 2 h at room temperature. A yellow decoloration of the Phenol red indicator in the CMF around the tissue fragment will at that point confirm its metabolic activity and thereby its viability. Usually, 1 cm3is plenty of material. 2. The tissue is freed from coagulated blood and minced with angled scissors. The fragments are transferred into the enzyme cocktail of pronase, collagenase, and DNase in Hanks buffer. After 20 min incubation at 37OCand shaking of the incubation mixture, the fragments are again agitated by repeated trituration. The undispersed fibrous fragments should be allowed to settle and the turbid supernatant transferred to a centrifuge tube.

120

Westphal et al.

3. Alternatively, mechanical dispersion can be used after mincing. The fragments are agitated by repeated trituration until a turbid solution is obtained. Allow large pieces or undispersed fibrous fragments to settle, and remove the supernatant into a centrifuge tube. 4. The cells and small-cell aggregates obtained from the digestion procedure or from the mechanical dispersion are pelleted at 80 x g for 10 min, and thereafter redispersed in cell culture medium containing 10% FCS. In the case of enzymatic dispersion, great care should be taken in carefully removing the enzymes completely, because theleftover enzymes will digest the ECM. Otherwise, the cells can be washed once in medium. 5. At this point, the suspension should be microscopically examined to rule out the possibility that either, because of the nature of the selected tissue or as a consequence of the isolation procedure, it consists of cell debris or broken up fibers. (Trypan blue exclusion assay is optional at this point, but with growing experience, the experimentor will be able to do without.) 6. The suspension is then plated onto ECM-coated dishes and left to attach for at least 6 h, but preferably overnight. 7. Cells are maintained at 37C in a humidified atmosphere supplemented with 5% CO, for the meningeomas and 8% CO, for the gliomas. Meningeomas are kept in Dulbeccos Modified Eagle medium (DME-H21) containing 10% fetal calf serum, 2 mM glutamine, 1 mM pyruvate, gentamycin (25 pg/mL), and amphotericin B (2.5 pg/ mL). Gliomas are maintained in MEM-Earle with the same supplements. Splitting and passaging of the cells is performed with a 0.05% [;iv.) trypsin/0.04% (w/v) EDTA solution in PBS/Ca2+, Mg2+ free . 8. Careful inspection of the cells on the next day is mandatory. In extreme cases, the cultures are confluent, and the debris from red blood cells and unattached fragments needs to be rigorously washed off. In any case, it is advisable to change the medium as early as possible (see Note 9).

4. Notes
4.1. Materials and Methods
1. Eyes can be conveniently transported in a plastic bag and do not need special prerequisites for transportation, If they are obtained by slaughterhouse workers at inconveniently early hours, they should be

Human Brain Tumors on ECM

121

2. 3.

4.

5.

kept in a refrigerator until the experimentor comes to pick them up. Although cornea1 endothelial cells are rather sturdy, it is advisable to use the eyes within 6-8 h after the animals have been killed. The cornea is a very tough tissue, and tightly gripping forceps should be used to lift the cornea after it is excised halfway. Also, the scissors should be sharp. After the brains are obtained from the slaughterhouse, they should rapidly be frozen. If they are kept at -20C, they should be put to 4C the evening before the isolation to turn them gradually into a waxy consistency. It is necessary for the brains to be at least frozen once before they are used because that facilitates the homogenization. For gliomas and meningeomas alike, the decision has to be made as to whether the tissue is to be digested with enzymes or only mechanically dispersed. Enzymatic dispersion has the advantage that a more homogeneous suspension is obtained, which can immediately be seeded for experiments in which comparable aliqots are to be used. Purely mechanical dispersion has the advantage that only the loosely connected tumor cells is obtained, vessels are left intact, and endothelial contamination is minimized. Even from small fragments, only those cells that will attach and migrate are mostly tumor cells. IMPORTANT NOTE: If the decision is made to use mechanical dispersion by trituration with a tissue culture pipet, these pipets should have a smoothened mouth. Some manufacturers have the tips cut, resulting in a very rough mouth that will break up the cells. The decision has to be made, however, on a case-to-case basis depending on the softness of the tissue and the experimental intentions. Medulloblastomas as well as pituitary adenomas and pineocytomas are usually very soft and will readily disperse by repeated trituration through a 5mL tissue culture pipet alone. Acoustic neuromas are often very tough and fibrous, so that enzymes help in releasing the tumor cells. In any case, the tissue is first minced with scissors and then triturated. Even in the face of abundant tumor material, one should refrain from using too much tissue. The addition of FGF is not quantified because the amount to be added will depend on the preparation. The protein peak eluting from the CM-Sephadex C-50 contains both acidic and basic FGF. The shape of the peak will depend on the geometry of the column used and the exact running conditions. Also, the final FGF concentration will depend on the width of the peak and when it is decided to stop the collection. An aliquot should be taken and tested at different dilutions to find out the maximally effective dilution.

122

Westphal et al.

6. bECM will have a different appearance from batch to batch. It may look homogeneous and bubbly as in Fig. 1, but it may also look more like a webb or show pronounced ridges, depending on how long the cornea1 endothelial cells had been propagated in culture, what the initial density was, and how long it took for the cells to become confluent. 7. It has to be noted, however, that the cells from tumors undergo changes in vitro and may alter their growth pattern. In this cell line, there was a gradual change in culture morphology towards a more disorderly arrangement (Fig. 2). Before that event, the cells would lyse just like the bovine cornea1 endothelial cells and leave a matrixbehind. Thereafter, it appeared as if the adhesive forces between the cells were stronger than the adhesion of the cells to their support. After the cells had lysed and the tECM was to be separated from the cellular debris, the whole matrix detached from the plastic in one piece. 8. It is advisable to obtain a preliminary histological diagnosis by frozen section. Viable tissue is selected on the basis of surgical experience. This means that, in those cases where the experimentor is not the surgeon, a close collaboration with the surgeon is advisable. Frozen sections from the immediate vicinity of the removed specimen are helpful. 9. One should also note whether the ECM is intact or lysed. In 116 cases of initiated gliomas, we noted that, in some of the initial cultures of high-grade malignant tumor cells, a slow lysis of the matrix occured. This could have been because of the release of proteolytic enzymes after cell death or because of fibronectin degrading enzymes, which are known to be present in membranes of transformed cells (17). In these cases, the ECM was barely visible in the phase contrast microscope 1 d after plating. Nevertheless, its beneficial effect was already visible at that point.

4.2. Notes on the Applications

of ECM

4.2.1. Meningeomas Meningeomas are usually benign epithelial lesions that are regularly homogeneous and of rather tough consistency. They are often highly vascularized. Meningeomas usually plate very well even without bECM. In our hands, there were three cases from a total number of 60, however, that did not plate on plastic but did so on ECM. Among them was the only meningeoma in our series that was derived from a patient with v. Recklinghausens disease. In general, there was very little difference in the cellular

Human

Brain Tumors on ECM

123

morphology between the coated and uncoated dishes in the primary culture of meningeomas. Nevertheless, the cultures seeded on ECM were confluent several days prior to the cultures platedonuncoated dishes. This is because of a better initial plating efficiency. Based on our experience, there are two major applications for ECM in the culture of meningeomas: First, those laboratories concerned with cytogenetic investigations of tumor cells prefer to do karyotype analysis as early as possible. If the cells are initiated on bECM, several days may be gained in reaching a point at which experiments are possible. Second, the bECM is very useful in reducing the requirements for serum and, therefore, can be used in those experiments that require low serum or serum-free culture conditions. 4.2.2. Gliomas Gliomas have to be divided into low-grade gliomas and malignant gliomas, and those cases of either that are operated upon as recurrencies after prior surgery, radiation, or other treatment. Gliomas, especially the more malignant lesions, are often very heterogeneous with numerous different cell lines within one tumor (18). It has to be assumed that this heterogeneity is the primary basis for the therapeutic difficulties that are intrinsic to glioma management. Only the understanding of this heterogeneity and the investigation of the different growth-control mechanisms present in one glioma will lead to an advancement of specific therapies. 4.2.2.1. Pvimavy cultures. In contrast to meningeomas, we found great differences in the behavior of primary glioma cultures on plastic and bECM. Three recent examples from our series of 116 cases may illustrate the different advantages, which are also documented in Fig. 3. Case 1 is derived from a tumor that, in frozen section, seemed to be an astrocytoma grade 2, but the definitive histological diagnosis classified it as a mixed oligodendroglioma/astrocytoma B according to Ludwig. Equal aliquots of mechanically dispersed tumor tissue were seeded onto either plastic or bECM-coated flasks. There were clearly more single cells attached on bECM. In addition, these cells had extensive communications with each other by means of an elaborate fiber network. After several days in culture, it was obvious, however, that there was very little mitotic activity and the cells came almost to a rest. In this case, early experimentation, i.e., assessment of cell lineage by immunocytochemistry with glial cell markers or chromosome analysis while the cells were still proliferating, would have been impossible without bECM coating. Case 2 is derived from a tumor in which histologically a glioblastoma with beginning sarcomatous changes was diagnosed. This highly malignant lesion did much better on bECM, allowing rapid expansion and early

124

Westphd et al.

Fig. 3. Phasecontrast micrographs from three different examples of cells after initiation on bECM (upper row) and on plastic (bottom row). (A) Primary cultures from NCEG109x (mixed oligodendroglioma/astrocytoma grade B according to Ludwig). (B) Primary cultures from NCE-Clll, a glioblastoma with sarcomatous changes. (C) Primary cultures from NCE-112, a gemistocytic astrocytoma grade 3. experiments such as growth-factor receptor assays, immunostaining, and karyotyping. Case 3, which is morphologically distinct from the other two cases, is derived from an astrocytoma in malignant transition. There is a good amount of attachment also on plastic, but the amount of attached single cells is much higher on bECM, thus permitting rapid change of medium, for example, the day after plating. This increases the content of tumor cells and minimizes the possibility of stromal contamination, which would be derived from normal cells growing out of small clumps of undispersed tissue. 4.2.2.2. Established cell Eines. If growth factors are to be investigated under defined conditions or if conditioned media from cells with SUSpetted autocrine growth factor secretion are to be collected, the useof ECM greatly facilitates these experiments. Any kind of ECM, because of its fibronectin content, lowers the requirements for serum in the culture media. bECM influences the proliferation of established cell lines. However, the effects of bECM canbebothpositive andnegativewithrespect to prolifera-

Human Brain Tumors on ECM

125

tion. Figure 4 describes two examples from an experiment in which cells from established glioma cell lines were seeded onto uncoated and bECMcoated multiwell dishes. One day after plating, the cells were changed to serum-free conditions and counted. They were then counted on the days indicated. It is obvious that, in one of the cases, the cells grew better on the bECM and better on plastic in the other case. It has to be noted, however, that, despite a better rate of proliferation in some cases, the cells look more spherical, and it can be observed that, after longer times in culture, the cells will detach and start floating. In some cases in which cells fromestablished lines were used, the cells growing on bECM had a more differentiated appearance and grew more strictly in monolayer than those on plastic. The comparative growth rates from such experiments indicate that there are cell lines in which thereis rather a proliferation-inhibitingeffectexerted by the bECM. 4.2.2.3. Applications of tECM. Production of an ECM by tumor cells is rather the rule than the exception. In our laboratory, we were interested in the effects of this tECM on primary cultures of gliomas. It was interesting to note that, in contrast to the nongeometrically arranged bECM, there was a profound effect of the tECM on the culture morphology (Fig. 5). Initial plating in proliferation experiments as well as growth rates was also affected (Fig. 6). In experiments in which the response of cultured glioma cells to epidermal growth factor, platelet-derived growth factor, and bFGF were evaluated, the quality of the responses were not different from plastic, but the overall cell number was increased (Fig. 7). For the reasons mentioned above, the experience with tECM is limited to the use of few batches, but is included because it illustrates the broadness of matrix biology. One aspect that is very well demonstrated in the comparison of the two ECMs is the drastically different geometrical arrangement. Using immunostaining of bECM with an antiserum obtained by immunizing rabbits against whole bECM, the disorderly structure can be visualized. Using antisera against fibronectin, one can see the structure of tECM that in itself has some kind of direction (Fig. 8).

4.2.3.

Other Neurosurgically

Obtained

Tumors

4.2.3.2. Pituitary adenomas. The vast majority of these are soft and should always be dispersed mechanically. Adenomas that are tough are rare. Even with mechanical dispersion, fibroblast overgrowth will be a problem after 3-6 wk of continuous culture because pituitary adenoma cells do not proliferate. Cells usually attach within a few hours and can then be used in endocrinological experiments in which test substances are added. For this the medium needs to be completely changed for the hor-

126

Westphal et al.

I
i.i,

NE-GU

NUgER

OFlJ*-fS

Fig. 4. Two examples of established glioma cell lines showing the diversity of effects when growing in serum-free medium on plastic (b and a) and on bECM (a and c). NCEG62 was derived from a glioblastoma in a child andNCFA84 from an adult glioblastoma. The cell counts were obtained at the days indicated from quadruplicates detached from multiwell plates with trypsin.

65L-

NCE-G22

6-

NCE-GL7

I 1

4 NUMBER OF DAYS IN CULTURE

Fig. 5. Comparative growth curves of cells growing on tECM (closed squares) and plastic (open squares). The cells were seeded at a density of 10,000 cells/well in tissue culture multiwell dishes and counted on the indicated days after tryptic detachment.

Human

Brain Tumors on ECM

127

Fig. 6. Phase contrast micrographs from primary cultures from two different gliomas that were initiated on bECM, tECM, and untreated pIastic dishes. Upper row, NCE-G74B, derived from an oligodendroglioma grade D; bottom row, cells from NCE-G70, a gemistrocytic astrocytoma grade 3.

mone assay. The adenoma cells are firmly attached and will not be removed with the medium. NOTE THE EXCEPTION: Adenomas derived from ACTH-secreting adenomasinpatientswithCushingsdiseaseattachmuchmoreslowlyand sometimes it takes up to two days. The time it takes for the cells to attach is correlated with the degree of the patients cortisol excess (9). In addition to the rapid attachment, the most striking effect of bECM in comparison to plastic is the rapid and obvious changes in cellular morphology after exposure to hypothalamic releasing factors, such as GRF for HGH-secreting cells and CRF or vasopressin for ACTH-secreting cells WO). 4.2.3.2. Acoustic Neuromas. Acoustic neuromas are difficult to culture even on ECM. If they are soft and cellular, they do attach well to plastic support, too. 4.2.3.3. Pined Tumors. Pineal tumors are very heterogeneous. Those that belong to the glial cell lineage behave as elaborated above. Two ger-

128

Westphal et al.

NCE-

G66

tECM PLASTIC

x
0

AX 0

-0 CONTROL

0 FGF

EGF

PDGF

Fig. 7. Comparison of the effects of maximally proliferation-stimulating doses of epidermal growth factor (10 nM), platelet derived growth factor (35 ng/mL), and bFGF (80 pg/mL), which were added every second day to serum-free cultures derived from a human glioma on plastic and tECM. minomas, however, showed very good attachment to bECM and did very

poorly on plastic. Morphologically, they had great resemblance to pituitary adenomas. 4.2.3.4. Medulloblastomas: This particular type of tumor has been difficult to maintain in culture, even on bECM. In three cases in which the tissue was very soft and only mechanically dispersed, there was little attachment of thevery small tumor cells that presented as little clusters. Only one case, which was a very malignant recurrence (presenting 6 wk after the macroscopic total removal of the tumor), attached very well to bECM, allowing immunocytochemical analysis, but it failed to continue to proliferate in vitro. 4.2.3.5. Gangliogliomus: Of these very rare tumors, only one was taken into culture, showing a remarkable heterogeneity. This was much more visible on bECM. In this case, the influence of bECM on the expression of cell-type specific markers with and without addition of B-Br-CAMP was studied, showing that bECM did not change the immunostaining pattern.

Human

Brain Tumors on ECM

129

Fig. 8. (A) Immunostaining of bECM with an&bECM antiserum and (B) immunostaining of tECM with antifibronectin antiserum. The staining was visualized by FITC labeled second antibody.

4.3.

General

Notes

1. Cornea1 endothelial cells can be frozen and stored in liquid nitrogen. They can be cultured for many in vitro passages and will continue to produce matrix material. It has to be noted, however, that the bECM gets thinner and less homogeneous with increasing age of the cornea1 endothelial cells. ECM can be stored at 4OCfor up to 1 yr. It can also be dried after all salt has been washed off. 2. ECMs are different in their chemical composition, depending on the tissue from which they are derived. One major aspect that has to be taken into account if autocrine mechanisms of cell proliferation are investigated is the sequestration of growth factors such asbFGF into tie

130

Westphal et al. ECM (29). They are then slowly released and act as an intrinsic mitogen. Furthermore, heparin binding growth factors appear to be stabilized in the presence of ECM that contains heparin sulfate.

Acknowledgments
The authors are grateful to D. Gospodarowicz, in whoselaboratory the introduction to ECM biology and the methodological training was obtained. This study was supported by the Heinrich Bauer Stiftung fur Hirntumorbiologie and the Deutsche Forschungsgemeinschaft. We express our thanks to the departmental operating room staff for their collaboration and appreciate the photographical work of S. Freist.

References
1. Bottenstein, J. E. and Sato, G. H. (1980) Fibronectin and polylysine requirement for proliferation of neuroblastoma cells in defined medium. Exp. Cell Res. 129,361-366. 2. Terranova, V. I., Aumalley, M., Sultan, L. H., Martin, G. R., and Kleinman, H. K. (1986) Regulation of cell attachment and cell number by fibronectin and laminin. I. Cell. Physiol. 127, 473-479. 3. Couchman, J.R.,Hook,M.,Rees, D.,andTimpl,R. (1983)Adhesion, growth,and matrix production by fibroblasts on laminin substrates. J. Cell. Biol. 96,177-X% 4. Varani, J., Carey, T. E., Fligiel, S. E. G., McKeever, P. E., and Dixit, V. (1987) Tumor type-specific differences in cell-substrate adhesion among human tumor cell lines. lnt. J. Cancer 39,397-403. 5. Gospodarowicz, D., Vlodavsky, I., and Savion, N. (1981) The role of fibroblast growth factor and the extracellular matrix in the control of proliferation and differentiation of cornea1 endothelial cells. Vision Res. 21,87-103. 6. Gospodarowicz, D., Cohen, D., and Fujii, D. K. (1982) Regulation of cell growth by the basal lamina and plasma factors: Relevance to embryonic control of cell proliferation and differentiation, in Cold Spring Harbor Conference on Cell Proliferation Vol. 9 (Cold Spring Harbor Laboratories, Cold Spring Harbor, New York), pp. 95-124. 7. Wichia, M. S., Lowrie, G., KohnE., Bagavandoss,P., and Mahn,T. (1982) Extracellular matrix promotes mammary epithelial growth and differentiation in vitro. Proc. Natl. Acad. Sci. 79,3213-3217. M. 8. Pavelic, K., Bulbul, M. A., Slocum, H. K., Pavelic, Z. I., Rustum,Y. M.,Niedbala, J., and Bernacki, R. J. (1986) Growth of human urological tumors on extracellular Matrix as a model for the in vitro cultivation of primary tumor explants. Cancer Res. 46, 3653-3662. 9. Westphal, M., Jaquet, P., and Wilson, C. B. (1986) Long-term culture of human corticotropin-secreting adenomas on extracellular matrix and evaluation of serum-free conditions. Acta Neuropath. 71,142-149. 10. Westphal, M., Hahn, H., and Liidecke, D. K. (1987) Culture of dispersed cells from human pituitary adenomas from acromegalic patients on extracellular matrix. In Growth Hormone, Growth Factors, and Akromegaly (Liidecke, D. K. and Tolis, eds.), Raven Press, New York, pp. 125-133.

Human Brain Tumors on ECM

11.

131

12. 13. 14. 15. 16. 17. 18. 19.

Westphal, M., Hansel, M., Brunken, M., Kijnig, A., Kdppen, J. A., and Herrman, H. D. (1987) Initiation of primary cell cultures from human intracranial tumors on extracellular matrix from bovine cornea1 endothelial cells. Exp. Cell. Bid 55,152-163. Ill, C. R. and Gospodarowicz, D. (1982) Factors involved in supporting the growth and steroidogenic functions of bovine adrenal cortical cells maintained on extracellular matrix and exposed to a serum-free medium. J. Cell. Physiol. 113,373-384. Vlodavsky, I., Lui, G. M., and Gospodarowicz, D. J. (1980) Morphological appearance, growth, behavior, and migratory activity of human tumor cells maintained on extracellular matrix versus plastic. Cell 19,607-616. Rutka, J.T. (1986)Effects of extracellular matrix proteinson the growth and differentiation of an anaplastic glioma cell line. Can. J. Neurd. Sci. 13,301306. Weiner, R. I., Bethea, C. L., Jaquet, I., Ramsdell, J.S.,and Gospodarowicz, D. J. (1983) Culture of dispersed anterior pituitary cells on extracellular matrix. Methods in Enzymol. 103,287. Gospodarowicz, D., Cheng, J., Lui, G. M., Baird, A., and Bohlen, P. (1984) Isolation of brain fibroblast growth factor by heparin sepharose affinity chromatography: Identity with pituitary fibroblast growth factor. Pm. NatI. Ad. Sci. 81,6963-6967. Chen, J. -M. and Chen, W. T. (1987) Fibronectin degrading proteases from the membranes of transformed cells. Cell 48,193-203. Shapiro, J. R. (1986) Biology of gliomas: Chromosomes, growth factors, and oncogenes. Seminars in Oncology 13,4-15. Vlodavsky, I., Folkman, J.,Sullivan, R., Fridman, R., Ishai-Michaeli, R., Sasse, J.,and Klagsbrun, M. (1987) Endothelial cell-derived basic fibroblast growth factor: Synthesis and deposition into subendothelial extracellular matrix. Pm. Naf2. Acud. Sci.
84,2292-2296.

Chapter 12
Embryonic Brain Cells in Culture
Narayan R. Bhat

Rodent

1. Introduction
Because of its cellular complexity and regional heterogeneity, the mammalian central nervous system is not easily amenable for experimental analysis. The study of the developing brain becomes even more complicated because of the differential growth rates of different parts of the brain. Primary culture techniques involving dissociation of discrete regions of the developing brain into component cells offer an excellent opportunity to study the regulation of growth and differentiation of neural cells and to investigate their biochemical, morphological, and physiological behavior under well-defined conditions. Methods are available now to isolate and grow individual classes of neural cells (i.e., neurons, astrocytes, oligodendrocytes), thereby enabling one to study the cellular behavior at the individual level and to uncover the nature of cell-cell interactions that presumably govern cell differentiation. The procedure widely used to prepare primary cultures of a mixed neuronal-glial population involves dissociation of embryonic rat/mouse brain. The cultures can be maintained either as suspension cultures or as surface cultures. Both systems recapitulate normal developmental events, and thus, have proved highly useful for the study of neuronal and glial differentiation (1,Z). Described here is a version of the method to set up surface-adhering primary cultures of embryonic rodent brain cells.
133

134

Bhat

2. Materials
1. Instruments: Surgicalinstruments such asmicroforceps (straight and curved; sharp and blunt-pointed), dissecting scissors, nylon mesh (20 and 78 pm sieve size), glass pestle (Belco). 2. Animals: Pregnant (15 d) rats/mice. 3. Dulbeccos Modified Eagles Medium (DMEM): To prepare DMEM, dissolve the powdered medium, sodium bicarbonate (3.7g/L), glucose i io t ic-antimycotic mixture (10 mL/L) representing (5g/L) an d an tb 100 U/mL penicillin, 100 pg/mL streptomycin, and 250 ng/mL amphotericin B in an appropriate vol of distilled, deionized water. After adjusting its pH to 7.4, the medium is filter-sterilized.

3. Method
Before the start of the experiment, make sure all the instruments, nylon mesh, paper towels, and small pieces (-1 in*) of Whatman No. 1 paper are sterilized. Warm up the medium to room temperature. Turn on the laminar flow hood and light the burner. Place the sterile surgical instruments in a beaker containing 70% alcohol. Also, keep 100 x 20-mm dishes containing DMEM ready. 1. Removal of the embryos: Outside the hood, sacrifice the pregnant rats by CO,-gassing. Wipe the abdominal area with 70% alcohol. Remove the embryos by Caesarean section and place them in a culture dish containing DMEM. The rest of the procedureis to be carried out under the hood. 2. Dissection: Force out the fetuses from the sacs with a gentle cut with the scissors. Individual embryos are wiped off by gently rolling on a sterile paper towel. Dissect the head and place it upright in an empty dish. Then, holding the front end down with a forceps, remove the entire brain with another forceps through a midline incision in the skull made with scissors. The brains thus taken out are collected in a dish containing DMEM. The cerebral hemispheres may then be separated from the rest of the brain with a cut across the mid brain. 3. Removal of meninges: This is done under a dissecting microscope kept in the hood. Place the cerebral tissue on a piece of sterile paper (Whatman No. 1) kept in a dish (60 x 15 mm) containing4 mL DMEM. Looking through the microscope, one can visualize the blood vessels embedded in the meningeal membrane covering the tissue. Remove the meningeal membrane by probing through it with pointed forceps.

Rodent Brain Cells

135

If done correctly, the entire membrane should peel off in one piece. 4. Dissociation: Cerebra are then dissociated mechanically. The tissue is first smoothed in a culture dish with the aid of a glass pestle, then suspended in medium containing 15% FCS, and dispersed by passing the suspension through a sterile pipet several times. Further dissociation is carried out by sequential passages through nylon meshes of pore sizes 78 pm and 20 pm. The resulting suspension of dissociated cells is centrifuged at a low speed (150 x g, 5 min) to obtain the pellet. The pelleted cells are washed twice with the medium and suspended at a density representing 2.5 x 106cells/mL. Final plating density: 3 x 105/cm2 surface area. 5. Maintenance of cultures: The initial change of medium is after 34 d of culture (when most cells have attached to the surface of the dish/ flask) and twice a week thereafter. DMEM containing 10% FCS is used for maintaining the cultures.

4. Notes
1. The dissociated embryonic brain cells have a tendency to reaggregate. The aggregates of cells thus formed attach to the surface of the dish and grow fibrous processes. By about a week in culture, an extensive network of neuritic processes can be seen covering the surface of a layer of proliferating astroglial precursors (Fig. 1). Cell proliferation, mostly accounted for by glial cells, peaks around 6 d in culture (3). This proliferative period is followed by cell differentiation, as indicated by an increase in neuronal and glial properties (e.g., 4-7). The time course of cell differentiation corresponds closely to the in vivo pattern, and thus, the culture system serves as an excellent model of the developing brain. 2. It should be pointed out that, under the culture conditions described here, neurons start degenerating after the second week. Thus, if the experiments call for prolonged neuronal survival, the following modifications may be incorporated. a. Use of antimitotic agents: An overgrowth of nonneuronal (glial) cells in these cultures seems to be the main reason for neuronal decay. In order to prevent this, investigators have used antimitotic agents such as fluorodeoxyuridine and cytosine arabinoside (8). Thus, for example, the cultures are set up as described (the seeding cell density may be reduced to one-half). The cultures grown for 6-8 d are treated with the inhibitor at a concentration of 1 x 10-5M for 24 h.

136

Bhat

Fig. 1. Phase contrast micrograph embryonic rat cerebral tissue.

of a 7-d-old primary

culture established

from 15-d

b. It has been noticed that supplementation of the growth medium with a combination of fetal calf serum and horse serum results in a reduced @al growth, thereby promoting neuronal survival. better still, Honegger et al. (9) have described a serum free defined medium [based on the N-2 medium originally described by Bottenstein and Sato (ZO)] that seems to support the long-term growth of mixed neuronal-glial cell populations. A similar growth medium has been used to study neuronal differentiation in surface-adhering cultures of fetal mouse hypothalamic cells (22). 3. Pettman et al. (22) have described a technique to grow pure neuronal cultures from embryonic chick brain. In this procedure, dissociated cells from 8-d embryonic brain are seeded onto polylysine-coated culture dishes. Glial cells fail to proliferate under these conditions for two reasons: (a) the early embryonic age chosen when gliogenesis is minimal and ( b) the growth of few glial cells present in the initial cell suspension is inhibited by the polylysine substrate. Alternatively, theproceduredescribed by Hansonet al. (13) separates neurons and glial cells based on (a) the differential adhesion properties of neurons and nonneuronal cells to a collagen substrate and (b) the capacity

Rodent Brain Cells

137

of neurons to form homotypic aggregates. Briefly, culture flasks are coated with specially treated rat tail collagen. The cell suspension is plated on collagen-coated flasks and placed in an incubator on an intermittantly agitating platform. The nonneuronal cells progressively attach to the substrate while the neurons form homotypic neuronal aggregates in the supernatant. These aggregates are then triturated with a Pasteur pipet and replated on a fresh flask coated with collagen. The original flasks are rinsed free of the remaining neurons with fresh medium to obtain glial enriched cultures. The above procedures originally used to prepare culture of embryonic chick brain cells can easily be adapted for embryonic rat/mouse brain cells. It is a fact that the cellular composition of primary brain cell cultures varies with the age of the donor tissue. Thus, cultures established from embryonic cortex generally results in a mixed neuronal-glial population, whereas cultures prepared from neonates produce nearly pure glial populations. The latter system can also be used as a source of pure cultures of astrocytes, oligodendrocytes (14,15), and bipotential glial progenitors (16).

References
1, Seeds, N. W. (1973) Differentiation of aggregating brain cell cultures, in Tissue Culfure of the Nervous System (Sato, G., ed.), Plenum Press, New York, London, pp. 3553. 2. Sensenbrenner, M. (1977) Dissociated Brain Cells in Primary Cultures, in Cell, Tissue, and OrganCuIturesinNeurobioIogy(Fedoroff,S., andHertz,L.,eds.), AcademicPress, New York, San Francisco, London, pp. 191-213. 3. Bhat, N. R., Shanker, G., and Peiringer, R. A. (1983) Cell proliferation ingrowingcultures of dissociated embryonic mouse brain: Macromolecule and ornithine decarboxylase synthesis and regulation by hormones and drugs. J. Neurosci. Res. 10,221230.

4. 5. 6. 7. 8.

Yavin, A. and Yavin, E. (1977) Synaptogenesis and myelinogenesis in dissociated cerebral cells from rat embryo on polylysine-coated surfaces. Exp. Bruin Res. 29, 137-147. Bhat, N. R., Subbarao, G., and Pieringer, R. A. (1981) Investigations on myelination in vitro: Regulation of sulfolipid synthesis by thyroid hormone in cultures of dissociated brain cells from embryonic mice. J. Biol. Chem. 256,1167-1171. Bhat, N. R., Shanker, G., and Peiringer, R. A. (1981) Investigations on myelination in vitro: Regulation of 2,3-cyclic nucleotide 3 phosphohydrolase by thryoid hormone inculturesof dissociated braincellsfromembryonic mice. J. Neurochenz. 37,695701. Bologa, L., Bisconte, J. C., Joubert, R., Marangos, P. J., Derbin, C., Rioux, F., and Herschkowitz,N. (1982)Accelerateddifferentiationofoligodendrocytesinneuronal rich embryonic mouse brain cell culture. Brain Res. 252,129-136. Godfrey, E. W., Nelson, P. G., Schrier, B. K., Breuer, A. C., and Ransom, B. R. (1975) Neurons from fetal rat brain in a new cell culture system: a multidisciplinary analysis. Bruin Res. 90,1-21.

138

Bhat

9. Honegger, I., Lenoir, D., and Favrod, I. (1979) Growth and differentiation of fetal brain cells in a serum-free defined medium. Nature 282,305308. 20. Bottenstein, J. E. and Sato, G. H. (1979) Growth of a rat neuroblastoma cell line in a serum-free supplemented medium. Proc. Natl. Acad. Sci USA 76,514517. 11. Faivre-Bauman, A., Rosenbaum, E., Puymirat, J., Grouselle, D., and Tixier-Vidal, A. (1981) Differentiation of fetal mouse hypothalamic cells in serum-free medium. Dev. Neurosci. 4,118-129. 12. Pettman, B., Louis, J. C., and Sensenbrenner,M. (1979) Morphologicaland biochemical maturation of neurones cultured in the absence of glial cells. Nature (Lord .) 281,

378-380. 13. Hanson, G. R., Iversen, P. L., and Partlow, L. M. (1982) Preparation

and partial characterization of highly purified primary cultures of neuronal and non-neuronal (glial) cells from embryonic chick cerebral hemispheres and several other regions of the nervous system. Deu. Brain Res. 3,529~545. 14. McCarthy, D. K. and devellis, J. (1980) Preparation of separate astroglial and oligodendroglial cell cultures from rat cerebral tissue. I. CeEZ Bio2.85,890-902. 15. Saneto, R. I. and devellis, J. (1985) Characterization of cultured rat oligodendrocytes proliferating in a serum-free, chemically defined medium. Proc. Nafl. Acad. Sci. USA 82,3509-3513. 26. Behar, T., McMorris, F. A., Novotny, E. A., Barker, J. L., and Dubois-Dalq, M. (1988) Growth and differentiation properties of O-2A progenitors purified from rat cerebral hemispheres. J. Neurosci. Res. 21,168-180.

Chapter 13 Human Epithelial

Thyroid

Cells

and

D. W. Williams David Wynford-Thomas

1. Introduction

The thyroid gland contains two populations of epithelial cells,of quite different embryological origin and function. Only the major component the follicular cells-will be considered here. The minor C cell population, which forms only a few percent, can be ignored for the purposes of primary culture. The follicular cells are organized in the intact gland into discrete functional units or follicles that consist of spheres lined by a single layer of epithelium and filled with the secretory product of the lining cells-thyroglobulin. The follicles are embedded in a vascular stroma. As with other epithelial tissues, almost all published methods for obtaining primary cultures have employed some form of proteolytic digestion combined with mechanical disaggregation. Earlier workers used trypsin in protocols similar to those used for embryo fibroblast preparation (1,2,3). These yield mainly single-cell suspensions for which it is difficult to obtain epithelial cells free of fibroblast contamination (4). A major im139

140

Williams

and Wynford-Thomas

provement can be achieved by the use of collagenase (and/or dispase), which unlike trypsin does not digest the intercellular junctional complexes that hold together the apical poles of the follicular cells. As a result, provided that digestion and disaggregation are not excessive, the majority of the follicular cells can be released in the form of follicles rather than single cells (5), which can be separated from the single-cell fibroblast component by differential sedimentation (6) or filtration (7). In addition, it has been shown that, as expected, avoidance of trypsin improves the viability of the resulting cultures and their sensitivity to hormones that act through cellsurface receptors, notably thyroid stimulating hormone (TSH) (8). Finally, this method of preparation allows for culture of follicular cells as intact follicles in suspension as well as in monolayer. The follicular cell is highly polarized both ultrastructurally and functionally (for example with respect to hormone and growth factor receptor distribution), and it has been shown that in monolayer its polarity is reversed, the apical rather than the basal surface being in contact with the environment (9). This can result, particularly in confluent monolayers, in a loss or diminution of many responses to extracellular agents, for example, the stimulation of adenylate cyclase and iodide uptake by TSH (IO), and of proliferation by epidermal growth factor (II). Culture in suspension as intact follicles avoids such artifacts, since the normal polarity is maintained. Indeed, the thyroid offers a rare opportunity to culture a glandular epithelium without disrupting its in vivo organizational unit, a condition that can never be entirely met with branched ductular-acinar structures such as breast or pancreas. All digestion protocols require that the tissue be first minced into small enough pieces to permit rapid diffusion of enzyme to all areas. Various strategies have been subsequently employed for digestion and disruption. Most simply, the tissue fragments can be incubated for several hours at 37OC and then disrupted in one step, for example by pipeting, followed by harvesting of the released cells (12). This inevitably results, however, in some cells being exposed to enzyme for too long and hence to an excessive proportion of single cells and poor viability. It is preferable to fractionate the procedure by harvesting at multiple intervals during the digestion (gentle mechanical disruption being applied throughout), so that cells (in the form of follicles) are removed from the enzyme as soon as possible after their release. One variation of this approach that we have found (13) to give reproducibly high yields of human follicles with minimal fibroblast contamination (~0.1%) is detailed here. Similar methods have been applied to cultures of dog (5,14), sheep (E), pig (I 6), rat (6), and human (27) thyroid, although not all including a fractionated harvesting protocol.

Thyroid Epithelial

Cells

141

2. Materials
1. Hanks Balanced Salt Solution (HBSS): This is a simple inorganic salt solution buffered for use in atmospheric CO,. We use the modified, calcium- and magnesium-free formulation supplied by Flow Laboratories. 2. Collagenase: 200 U/mg. Store powder at 4OC. 3. Dispase: neutral protease from Bacillus polymyza, Grade II. Store powder at 4C. 4, Enzyme mixture: Dissolve 40 mg of collagenase plus 60 mg of dispase in 60 mL of HBSS. Filter through 0.45 pm nitrocellulose filter to sterilize. Prepare fresh each time; keep on ice until use. 5. Nylon mesh: 200 pm nominal pore size. Sterilize by autoclaving. 6. RPM1 1640 medium: The cells have also been grown successfully in Dulbeccos Modified Eagles Medium (DMEM). We routinely add penicillin (50 U/mL) and streptomycin (50 pg/mL). 7. FCS (Fetal Calf Serum) 8. NBS (Newborn Bovine Serum) 9. Acridine orange and ethidium bromide: Prepare solution containing O.lpg/mL of each in HBSS. Store in dark bottle at 4C. These are mutagenic and gloves should be worn when handling them. 10. Agar: high gel strength. Make 100 mL of a 2% solution in double-distilled water. Boil in a bottle with loosened lid to dissolve and sterilize. Store at 4C. Reboil before each use. 11. Trypsin/EDTA solution: 0.05% (w/v> trypsin plus 0.02% (w/v> EDTA in Dulbeccos calcium- and magnesium-free phosphate-buffered saline. Dilute from 10x stock (stored at -20C). Store working solution at 4OC (up to 1 mo). 12. Dimethylsulphoxide (DMSO): Tissue-culture tested product, supplied sterile. Store at room temperature in dark. Store freezing mixture (DMSO:NBS/ 1:4) at 4OCin dark.

3. Methods
The following procedure is summarized in Fig. 1. 1. Human thyroid tissue is best obtained as freshly excised surgical material-most of our samples come from thyroid lobectomies performed for removal of cold nodules (seeNote l), which usually provide at least 1 g of histologically normal gland. The tissue sample is transported to the tissue culture laboratory in HBSS on ice (but speed

142

Williams

and Wynford-Thomas

Follicles single ceils & debris

Add to enzyme soiution 2oog 2 min

Remove enzyme solution Resuspend pellet in RPM

Allow to sediment

Single ceils

debris

foiiici&

plus debris

Fig. 1. Protocol for preparation

of human thyroid

follicles.

Thyroid Epithelial

Cells

143

2.

3.

4.

5.

6.

7.

8.

at this stage is not particularly vital, since we have found no deleterious effect from storage at 4OCfor up to 1 h). Prior to enzyme digestion, the tissue is first minced with sterile scalpel blades as finely as possible (to around 2 mm cubes). The pieces are then washed four times in 15 mL of HBSS at 4OCin a universal container to remove as much blood as possible. (Each wash consists of manual shaking followed by sedimentation under gravity- there is no need to centrifuge.) The HBSS is removed and replaced by 10 mL of enzyme mixture consisting of collagenase (130 U/mL) and dispase (1 mg/mL) in HBSS. The container is incubated in a static water bath at 37C with intermittent gentle agitation by manual shaking for 20 s every 15 min. After the first hour of incubation, the first fraction is harvested (see Note 2). The enzyme mixture (containing suspended follicles and single cells) is removed (leaving the undigested tissue fragments at the bottom of the universal container-see Note 3), and transferred to a sterile 15 mL centrifuge tube. FCS is added to a final concentration of 0.5% to neutralize proteases, and the tube placed on ice. Fresh enzyme mixture (prewarmed to 37OC) is added to the undigested tissue fragments and the incubation continued. After a further 30 min, the next fraction is harvested exactly as above. The process is repeated until no more tissue remains (usually after 3 h). While digestion is continuing, the fractions already collected are centrifuged at 200g for 2 min in a swinging bucket rotor to recover the cells and follicles. The supernatant (enzyme mixture plus FCS) is removed and the pelleted cells thoroughly resuspended by pipeting in 10 mL RPMI medium at 4C. They are then allowed to resediment under gravity for 0.5-l h (seeNote 4). The supernatant (containing single cells) is then discarded and the sedimented follicles from all fractions pooled in fresh RPMI. After the final fraction has been processed, the pooled follicles are filtered through a 200 pm nylon mesh to remove large debris such as partially digested tissue fragments. Finally, the follicles are washed twice by centrifugation (2OOg,2 min) and resuspended in l-4 mL of RPMI. To obtain a crude cell count and viability index, 15 PL of follicle suspension is mixed with 15 PL of acridine orange/ethidium bromide solution and viewed in a hemocytometer under phase contrast (for cell/follicle counting > and UV light for assessment of viability (see Note 5).

144

Williams

and Wynford-Thomas

Fig. 2. Phase contrast of human thyroid follicular cells in monolayer culture, showing an island of cells derived from a single follicle 3 d after attachment (350x).

9. Prior to plating out or freezing (seeNote 6), the follicles are first incubated overnight (in a standard humidified 5% CO, atmosphere at 37OC)at a density of around lo5/cm2in RPMI (still serum-free) on Petri dishes coated with 2% agar (seeNote 7) to prevent cell attachment. This maneuver allows most of the follicles that will have been ruptured during extraction to reform and leads to death of most of the small proportion of single cells (particularly fibroblasts) that may still be present. 10. Monolayer Culture (Fig. 2): Follicles areplatedon standard tissue-culture grade plastic Petri dishes in RPMI medium supplemented with 10% FCS. Plating efficiency is poor at low density; densities are best kept above 104/cm2. Initial doubling times average 2-3 d. At confluence, the cultures can be passaged (split ratio up to 1 in 8) by standard trypsinization schedules using trypsin/EDTA solution (see Materials). As expected for a primary culture, senescence occurs after a finite number of divisions-in 10% FCS between 4 and 8. The cells flatten, mitoses are no longer observed, and after several weeks, the culture eventually dies. Escape from senescence has never been observed in these human cultures.

Thyroid Epithelial

Cells

145

Fig. 3. Phase contrast of thyroid follicles in suspension culture (350x).

The cells retain the tissuqxxific control of growth by thyroidstimulating hormone (TSl3, which can be demonstrated by stimulation of PI-II-thymidine labeling index in serum-free medium by TSH, in the presence of insulin-like growth factor-l (IGF-1) (18). 11. Suspension (Follicle) Culture (Figs. 3 and4): Cells (follicles) are plated at 5 x W-5 x 105/mL in RPMI supplemented with the desired serum concentraton. The cells will survive even in serum-free medium for many days (-95% viability after 1 wk). In 10% FCS, proliferation is very limited in comparison to monolayer cultures, [3Hl-thymidine uptake falling rapidly during the first few days. This is probably partly the result of a simple physical restriction on cell spreading in the follicle, but there is evidence also that the loss of proliferative capacity correlates with the development of follicular inversion, which occurs over the first week of culture (seeNote 8). In the absence of serum, inversion is delayed, and clear proliferative responses to addition of pure growth factors (e.g., TSH plus IGF-1) can be observed for at least a week. This culture system was the first to demonstrate the growthstimulatory effect of TSH on human follicular cells in vitro (23).

146

Williams and Wynford-Thomas

Fig. 4. Electron micrograph of follicles in suspension. The follicles were collected by centrifugation, fixed in 1% glutaraldehyde, and embedded in L.RWhite resin. Sections were stained with uranyl acetate/lead citrate. O3OOOx).

4. Notes
1. Cold nodules are most often a benign follicular adenoma; the surrounding unaffected tissue is usually clearly distinguishable macroscopically from the tumor, but the normality of the samples used for

Thyroid Epithelial

Cells

147

2. 3. 4.

5.

6.

7.

tissue culture is checked retrospectively by histological examination of tissue sections. Laryngectomies provide an alternative source of surgical samples. Some workers have successfully employed postmortem material, which is of course available in much larger quantity w. Very little release of cells occurs during the first hour of enzymic digestion. Most follicles are recovered in the next four successive 30min incubations. The whole procedure usually lasts 3-4 h. The undigested tissue pieces sink rapidly compared to the released follicles, so that on standing after a brief shake the fragments will all have settled in less than 1 min, leaving the follicles still in suspension. At step 6, the tube is allowed to stand for a much longer period than in Note 3 (longer than 30 min). In this time, the follicles will sediment even under unit gravity in a medium of sp. gr. 1, permitting adequate separation from the single cells (including fibroblasts), which remain in suspension. This is made possible by the large size of human follicles; with rat follicles, for example, a Percoll separation medium is needed (6). Viable cells exclude the ethidium bromide and are stained green by the a&dine orange. Dead cells take up the ethidium bromide and consequently show an orange fluorescence. The quantitation of cell number is subject to observer error, because of the difficulty of distinguishing individual cells in some follicles. This has not proven a problem provided absolute figures are not important and that all observations for a given experiment are made by the same operator. If higher accuracy is needed, a follicle sample can be treated with trypsin to produce a single-cell preparation. Freezing: The follicles are first incubated for 24 h in suspension to allow reformation of closed follicles, since this has been found to improve subsequent viability after thawing. The follicles are harvested from the agar-coated dish by washing with RPMI medium into a 15mL centrifuge tube, and then centrifuged at 200g for 2 min. The pellet is resuspended in ice-cold RPM1 containing 10% NBS at a maximum cell concentration of lO/mL. An equal vol of a 1:4 (v/v> mixture of DMSO and NBS is added, and the cells thoroughly dispersed by gentle pipeting. The suspension is aliquoted into freezing ampules and allowed to cool slowly to 70C inside a tightly closed polystyrene box. The next day, the ampules are transferred to liquid nitrogen for prolonged storage. We have routinely obtained viability of >80% after 3 yr in store. The agar is prepared by boiling a 2% solution in double-distilled

148

Williams

and Wynford-Thomas

water. The hot solution is dispensed rapidly to avoid cooling, using a disposable plastic pipet. An excess is applied (around 5 mL for a 9cm Petri dish) and allowed to cool for 1 min. Most of the agar is then removed, leaving a thin coating. 8. Follicle inversion is a well-recognized process in serum containing media (20). The mechanism is unclear, but the result is an inside-out structure in which the polarity of the cell is reversed, the microvillous border coming to lie on the outside in contact with the medium, accompanied by corresponding changes in position of the Golgi apparatus, nucleus, and intercellular junctions. Perhaps as a result of inappropriate fluid and electrolyte transfer, the lumen of the follicle becomes greatly distended, and the lining epithelium very attenuated. Such cells show a marked alteration in both functional and proliferative responses (13,21).

Acknowledgments
We are grateful to the Cancer Research Campaign and to the Welsh Scheme for Development of Health and Social Research for grant support.

References
1. Pastan,I. (1961)Certain functions of isolated thyroid cells.Endocrinology 68,924-931. 2. Tong, W., Kerkof, P. R., and Chaikoff, I. L. (1962) Iodine metabolism in dispersed thyroid cells obtained by trypsinization of sheep thyroid glands. Biochim.Biophys. Ada. 60,1-19. 3. Fayet, G., Pacheco,H., and Tixier, R. (19701Sur la reassociation in vitro des cellules isolees de thyroide de port et la biosynthese de la thyroglobuline. 1. Conditions pour Iinduction des reassociations cellulaires par la thyreostimuline. Bull. Sot. Chim. Biol. 52,299-306. 4. Murphy, A., Mothersill, C., OConnor, M. K., Malone, J.F.,Cullen, M. J.,and Taaffe, J. K. (1983) An investigation of the optimum culture conditions for a differentiated culture of sheep thyroid cells. Ada Endocrinol. 104,431-436. 5. Rapoport, B. (1976) Dog thyroid cells in monolayer tissue culture: Adenosine 3, Scyclic monophosphate response to thyrotropic hormone. Endocrinology 98,11891197. 6. Smith, I., Williams, E. D., and Wynford-Thomas, D. (1987) In vitro demonstration of a TSH-specific growth desensitizing mechanism in rat thyroid epithelium. Mol. Cell. Endocrinol. 51,51-58. 7. Nitsch, L. and Wollman, S. H. (1980) Suspension culture of separated follicles consisting of differentiated thyroid epithelial cells. Proc.Nat. Ad. Sci.USA 77,472476.

8. Stockle,G., Wahl, R., and Seif, F. J.(1981) Micromethod of human thyrocyte cultures for detection of thyroid-stimulating antibodies and thyrotrophin. Ada Endocrinologica 97,369-375.

Thyroid Epithelial

Cells

149

9. Chambard, M., Gabrion, J., and Mauchamp, J. (1981) Influence of collagen gel on the orientation of epithelial cell polarity: Follicle formation from isolated thyroid cells and from preformed monolayers. J. Cell. Bid. 91,157-X6. 10. Chambard, M., Verrier, B., Gabrion, J., and Mauchamp, J. (1983) Polarization of thyroid cells in culture: Evidence for the basolateral localization of the iodide pump and of the thyroid-stimulating hormone receptor-adenyl cyclase complex. J. Cell. Bid. 96,1172-1177. 11. Westermark, K., Westermark, B., Karlsson, F. A., and Ericson, L. E. (1986) Location of epidermal growth factor receptors on porcine thyroid follicle cells and receptor regulation by thyrotropin. EndocrinoZogy 118,1040-1046. 12. Stringer, B. M. J., Wynford-Thomas, D., and Williams, E. D. (1985) In vitro evidence for an intracellular mechanism limiting the thyroid follicular cell growth response to thyrotropin. EndocrinoIogy 116,611-615. 13. Williams, D. W., Wynford-Thomas, D., and Williams, E. D. (1987) Control of human thyroid follicular cell proliferation in suspension and monolayer culture. Mol. CeZZ. Endocrinol. 51,33-40. 14. Roger, P. P., Hotimsky, A., Moreau, C., and Dumont, J. E. (1982) Stimulation by thyrotropin, cholera toxin and dibutyryl cyclic AMP of the multiplication of differentiated thyroid cells in vitro. MoZ. Cell. Endocrinol. 26,165-176. 15. Westennark, K. and Westermark, B. (1982) Mitogenic effect of epidermal growth factor on sheep thyroid cells in culture. Exp CeZlRes.138,47-55. 26. Westermark,K., Karlsson, F. A.,and Westermark,B. (1983) Epidermal growth factor modulates thyroid growth and function in culture. Endocrinology112,1680-1686. 17. Davies, T. F., Platzer, M., Schwartz, A., and Friedman, E. (1983) Functionality of thyroid-stimulating antibodies assessed by cryopreserved human thyroid cell bioassay. 1. CZin. Endocrinol. Metab. 57,1021-1027. 18. Williams, D. W., Wynford-Thomas, D., and Williams, E. D. (1987) Human thyroid adenomas show escape from IGF-1 dependence for growth. Annales dEndocrinoZogie 48,82A. 19. Roger, P. I. and Taton, M. (1987) TSH is a direct growth factor for normal human thyrocytes. Annales dEndocrinoZogie 48,158A. intermediate stagesin polarity 20. Nitsch, L. and Wollman, S. H. (1980) Ultrastructureof of thyroid epithelium in follicles in suspension culture. 1. Cell BioZ.86,875-880. 21. Gartner, R., Greil, W., Stubner, D., Permanetter, W., Horn, K., and Pickardt, C. R. (1985) Preparation of porcine thyroid follicles with preserved polarity: functional and morphological properties in comparison to inside-out follicles. Mol. Cell Endo-

crinol. 40,9-16.

Chapter 14 Preparation of Isolated Rat Liver Hepatocytes

Bjern Quistorfg and Niels John Grunnet Dich,

1. Introduction
The bulk volume (about 85%) of the mammalian liver parenchyma is contributed by the hepatocytes, whereas at least four other types of cells constitute the remainder (1). Procedures for the isolation of all five cell types have been described, although not from the same liver. However, the isolation of hepatocytes has clearly been the most widely used preparation and has proved extremely valuable for a wide variety of experiments, spanning fields like pharmacokinetics, drug metabolism, and metabolic functions of the liver. Historically two classes of methods have been used for the preparation of the isolated hepatocytes, either mechanical/chemical methods or enzymatic methods. The enzymatic techniques in which collagenase perfusion is applied as the principle of disintegration of the liver as first described by Berry and Friend (2) is completely dominating the field now and forms the basis of the method described in detail in this chapter. A number of symposia proceedings and reviews on the preparation and use of isolated hepatocytes have appeared (3-B). 151

152

Quistorfi

Dich, and Grunnet

An added complication in the use of isolated hepatocytes is the demonstration in recent years of functional metabolic differences of hepatocytes of periportal and perivenous origin (for review, see 7). This chapter does not take such differences into account; however, a following chapter describes a method by which preparation of isolated hepatocytes, enriched in either periportal or perivenous cells, may be accomplished (this book, Chapter 16).

2. Materials
1. Adult Wistar rats weighing about 200 g were housed at 21OC with alternating 12-h cycles of light (7 AM - 7 PM) and darkness. 2. Solution I: 118.0 mM NaCl, 4.7 mM KCl, 1.2 mM KH.J?O, and 25 mM NaHCO,. 3. Solution II: As solution I plus 3 mM CaCl,, 1.2 mM MgSO,, and collagenase (0.2-0.4 mg/mL). 4. Solution III: As Solution II without collagenase. In order to avoid micro-floral contamination, all solutions were prepared from individual stock solutions 5 times more concentrated. Solutions were equilibrated with oxygen/carbon dioxide (19:1, vol/vol) for 1 h before use and kept in a thermostated waterbath at 39OCduring the cell isolation procedure (Fig. 1). Because of heat loss in the tubing system, the perfusate temperature at the liver outlet is 34-35C. 5. Trypan blue solution: 0.5 g Trypan blue in 100 mL of buffer containing 140 mM NaCl, 4.7 mM KCl, 1.2 mM KHJ?O, 0.6 mM MgSO,, and 10 mM HEPES, pH 7.4. The solution is stirred for 2 h and filtered. Aliquots can be preserved at -18OC. 6. Water bath: A conventional thermostated shaking water bath with an accuracy of & 0.2C and a stroke amplitude of about 2 cm is used. 7. Oxygenation: It is sufficient to bubble solutions with mounted Pasteur pipets connected to the oxygen/carbon dioxide tank (19:1, vol/ vol). 8. A two-channel roller pump (e.g., LKB 2115, LKB, Bromma, Sweden) is preferable. The pump should be able to give a variable flow rate of 20-40 mL/min. 9. Bubble trap: Since infusion of only small amounts of air is deleterious to the perfusion, it is highly advisable to mount a simple bubble trap on the infusion line. 10. Cannula: A glass cannula or a double cannula (19-gage x l-3/4 in [needle], 16-gage x l-l /4 in [catheter]) can be used. The latter is easier to handle by unexperienced persons.

Rat Liver Hepatocytes

153

BT

V. PORTA

V. CAVA -------w

Fig. 1. Perfusion setup for preparation of isolated rat liver hepatocytes. The three solutions I, II, and III are kept in a thermostated water bath. Pl, P2 is a two-channel roller pump. BT bubble-trap. B is a beaker for collection of effluent for recirculation during collagenase perfusion with perfusate II.

11. Filtration device: A simple but efficient filtration device consisting of a 250-mL beaker, a Biichner funnel rubber gasket, and a nylon net (100 mesh [160 pm]) is shown in Fig. 2. 12. Centrifuge: A conventional laboratory centrifuge equipped with a swing-out rotor can be used, provided that low g-values can be obtained reproducibly. 13. Tubing: Silastic tubing 4 mm/2 mm o/i diameter was used. For cannulation of the superior caval vein, a 2-mm/1.5-mm o/i diameter polyethylene catheter is used.

154

Quistorff, Did,

and Grunnet

lib MESH NYLON NET

RUBBER HEPATOCYTE SUSPENSION u > GASKET

250

ml

BEAKER

Fig. 2. Simple filtration device. The rubber gasket with the nylon net is pushed toward the bottom before the cell suspension is loaded on the net. The filtration is carried out by moving the gasket up and down a couple of times, if necessary,with addition of extra buffer.

14. Counting chamber: A conventional counting chamber with avolume of 0.1 PL. 15. Collagenase (type II) is used.

3. Methods
The essential features of the collagenase cell isolation procedure from rat liver involve five steps as listed in Table 1. Quantitative details of the procedure described below and in Table 1 refer to a rat weighing 200 g. The procedure may, of course, be scaled to larger or smaller rats as required. 1. The liver of the anesthetized rat is perfused in situ via a cannulation of the portal vein. The abdomen is opened widely by a long midline incision and two perpendicular cuts at the level of the kidneys. The guts are moved to the left, out of the abdominal cavity, giving free access to the portal vein and the inf. v. cava.

Rat Liver Hepatocytes Table 1 Schematic Presentation of the Preparation of Isolated Rat Liver Hepatocytes
Perfusion/ Flow

155

Time
min

Phase Phase 1

incubation Calcium removal

mL/min 35-40

Buffer/ comment Solution I Flowthrough perfusion Solution II Recirculation Solution III Flowthrough perfusion
Incubation with shaking
(60-70

8-12

Phase 2

Collagenase perfusion Rinse-cycle perfusion

20-30

15-25

Phase 3

35-40

2-4

Phase 4

Disintegration of the liver,

10

followed by incubation

strokes/
mid in solution III

Phase 5

Filtration and cell separation

Centrifugation 3x5og x2min

2. Prior to cannulation, ligatures are placed loosely around the v. porta (ca. 5 mm from the bifurcation to the different lobes), and around the v. cava inf. above the right kidney. 3. As soon as the cannula is in place in the portal vein and retrograde bleeding is observed, the ligature is tightened, the perfusion is started with perfusate I (seeFig. I), and the inf. v. cava is cut open below the right kidney. The chest cavity is cut widely open, the v. cava sup. cannulated through the left atrium, and the ligature around v. cava inf. in the abdominal cavity is tightened. 4. Establishing the flow-through perfusion to this point usually takes a trained technician about 44 min. The liver must be homogenously

Quistorff, Dich, and Grunnet bleached and absolutely without spots of unperfused areas. If that is
not the case, the liver should be discarded. An additional interval of 8-12 min of perfusion with the calcium-free buffer is required to accomplish efficient calcium removal. Inefficient calcium removal will decrease cell yield drastically, since the cells will not separate properly. The perfusate is switched to perfusate II (collagenase), and recirculation is established (seeFig. 1). The total perfusate volume at this time amounts to approximately 150 mL. The collagenase perfusion is usually continued for 15-25 min depending upon the appearance of the liver. With experience, it is possible to judge quite accurately from the appearance of the liver surface the degree of disintegration. One semi-objective indicator of sufficient collagenase perfusion is when a slight impression made (with the handle of a pair of tweezers) on the liver surface tends to remain. During collagenase perfusion, the perfusate starts to leak out through the liver surface, and in most experiments, 75-100% of the perfusate will be lost by the time when the collagenase treatment is sufficient. In a few experiments, it is necessary to prepare additional collagenase perfusate in order to complete the treatment. The liver is now flushed for 3 min with perfusate III in a flow-through perfusion. The perfusion is stopped, and the liver cut out and gently dispersed in a Petri dish with two forks in perfusate III. The dispersed liver is incubated in a conical 1-L flask with a O,/CO, atmosphere (95/5 ~01%) at 37OCfor 10 min with gentle shaking (60-70 strokes/min). The next step is filtering of the disintegrated liver through a nylon net. This may be accomplished as shown in Fig. 2. One should not try to force the liver cell suspension through the filter by mashing with a spoon or a spatula. Instead, the filter on the rubber ring may be moved up and down a couple of times and some extra medium may be added (see Table 1). A successful cell preparation at this stage is usually characterized by a remainder on the filter that constitutes mostly the fibrous vascular tree of the liver. The final step consists of a separation of hepatocytes from cell debris and other smaller cells of the liver. This is accomplished by three lowspeed centrifugations (2 min at 508) in two 30-mL centrifuge tubes. The hepatocytes will sediment as a dark brown layer. Only these cells

5. 6.

7. 8. 9. 10.

11.

Rat Liver Hepatocytes

157

are useful, whereas debris and other cells, which will remain in the supernatant, are discarded. The centrifugation is repeated with two resuspensions of the cell pellet; the last time cells are pooled in a graduated tube in a medium to which is usually added l-3% bovine serum albumin. After the last centrifugation, the cells are suspended in the final buffer to be used in the particular experiment. The cell concentration here ranges from 2-10% in most experiments (seeNotes l-6).

4. Notes
1. With the procedure described above one routinely obtains a yield of 2-3 mL of packed cells corresponding to 2-4 x lo8 cells, which is about 50% of the total hepatocyte population of the liver. 2. The most frequently used test of cell viability is Trypan blue staining. The test is carried out immediately after the final centrifugation by mixing 50 PL of cell suspension with 50 PL Trypan blue solution. The cell suspension is allowed to stand for 1 min prior to a lo-fold dilution with 0.9% NaCl, followed by examination in an hemocytometer (vol 0.1 PL). Several factors are observed in the light microscope at a magnification power of 40: a. The number of cells/O.1 PL b. The number of cells smaller than 10 pm c. The appearance of the cells, rounded and with or without blebs d.The number of hepatocytes that do not have blue stained nucleus. This number in percentages is referred to as the percentage of viable cells. More than 200 cells should be counted. Usually >90% of the cells obtained by the procedure described above exclude Trypan blue and show normal morphology, i.e., rounded cells without blebs of a size of 15-25 ym. Less than 5% nonhepatocytes are found in the preparation. The criterion for positive Trypan blue staining is a visible blue nucleus. Unfortunately, there is no correlation between the percentage of stained cells and the metabolic capability of the isolated hepatocytes (9). Therefore, a number of other viability criteria have been suggested. Among these, the content of ATP is probably the easiest. ATI? should be > 2 ~mol/108 cells. Also the capacity for glucose synthesis, the leak of cytosolic enzymes (LDH, ALAT), or the

Quistorff, Dich, and Grunnet response in terms of oxygen consumption upon incubation with various substrates may be used (3,7-g). In the practical experiment, however, it is difficult to perform an extensive testing of the isolated cells prior to the experiment itself, since the cells tend to deteriorate in terms of metabolic capacity over about 2-3 h. In terms of the viability test, the bottom line is that anyone who proposes to work on isolated hepatocytes should satisfy himself that he can reproduce the maximum rates reported in the literature before studying new aspects. Merely looking at cells- at their shape and at dye exclusion-is not good enough. Nor is reproducibility of data, because this does not reveal systematic errors... as stated by Krebs et al. (9). 3. Inclusion of hyaluronidase in the collagenase-containing perfusate (2), addition of EDTA to the calcium-free perfusate (IO), the use of buffers other than Krebs-Henseleit bicarbonate, e.g., Hanks solutions (12), higher concentrations of collagenase, inclusion of erythrocytes (12), and ex situ perfusion in thermostated perfusion cabinet are variations of the described technique, which have been used successfully, but without significant improvement over the simpler technique described here. Likewise, oxygenators are superfluous as long as the perfusion solutions are bubbled efficiently with 95% OJ 5% CO, (vol/vol>. Thus, it appears that many details of the cell preparation procedure are uncritical and only a few essential factors may be identified. In our opinion, these are the quality of the liver perfusion, efficient depletion of calcium, and the presence of calcium in the collagenase solution. Some authors recommend inclusion of DNase in the buffer used for incubation of the digested liver (phase 4, Table I) to prevent clumping of the cells (23). Normally, however, clumping of the cells is not a problem. For the purpose of comparison, it may be desirable to obtain a liver sample before the preparation of isolated cells. After the perfusion has been established, the small caudate lobe (next to the right kidney) can be ligated and cut off without affecting the remaining part of the procedure (24). 4. Isolated hepatocytes are usually incubated in conical flasks at 37OC at a concentration of 0.5-10 x lo6 cells/mL, corresponding to approximately 4-70 mg wet wt./mL (15). Shaking of the flasks, 60-80 strokes/min, is necessary. With a gas phase of 95% 0,/5% CO,, the depth of the suspension should not under these conditions exceed 6 mm in order to ensure proper oxygenation. Flasks equipped with a

Rat Liver Hepatocytes

159

center well improve stirring of the cell suspension. Albumin (l-3%) is often included in the final suspension medium and in the incubation medium. Unless fatty acids are included in the medium, there appears to be no beneficial effect of albumin on metabolic rates or on cell survival. The incubation time should be kept below 2 h. Isolated hepatocytes may be used for primary cultures (this book, Chapter 15). 5. Isolated hepatocytes in suspension are able to carry out most intracellular processes at rates close to those attainable in the perfused liver. In some respects, however, isolated hepatocytes appear less competent. This is the casefor glycogen synthesis, which is nonexistent in isolated hepatocytes unless special conditions prevail, i.e. high glucose concentrations (>15 mM) or lactate (5 mM), pyruvate (1 mM> and glutamine (10 mM> (26,27). Isolated hepatocytes are in a state of marked, negative nitrogen balance, which may be alleviated by the addition of amino acids to the incubation medium (18). During preparation, the hepatocytes are depleted of low mol wt intermediates (3). Consequently, they are in an oxidized state (a state 2) (29) withsubstratesupplybeingrate-limiting for therespiratory chain and synthetic processes. Addition of lactate (5 mM) plus pyruvate (I mM> appears adequate to establish a physiological NAD-redox level in the cytosolic and mitochondrial cell compartment and to supply precursors for, e.g., gluconeogenesis and fatty acid synthesis. Exogenous fatty acids results in a rather reduced NAD-redox state, unless lactate and pyruvate are added (20). Secretory processes, e.g., the secretion of albumin, occur at a rate comparable to that in perfused liver, but lower than in vivo (21). 6. The enzyme assays and metabolite assays employed in the evaluation of cell viability and functional capacity may be used directly asdescribed, e.g., in the handbooks (22,23).

References
2.
Greengaard, O., Fedex-man, M., and Knox, W. E. (1972) Cytomorphometry of developing rat liver and its application to enzyme differentiation. J Cell Biol. 52,261-X2. 2. Berry, N. M. and Friend, D. S. (1969) High yield preparation of isolated rat liver parenchymal cells. 1. Cell Bid. 43,506-520. 3. Krebs, H. A., Cornell, N. W., Lund, P., and Hems, R. (1974) Isolated liver cells as experimentalmaterial,inRegulationofHepaticMefuboZism. (Lundquist,F.andTygstrup, N., eds.), Munksgaard, Copenhagen, pp. 726-750. 4. Seglen, P. 0. (1976) Preparation of isolated rat liver cells, in Methods in Cell Biology, vol. 19 (Prescott, D. M., ed.), Academic, New York, pp. 29-83.

160

Quistorfi

Dich, and Grunnet

5. Tager, J. M.,SGlling,H. D., and Williamson, J. R. (eds.) (1976) UseoflsoIated Liver Cells and Kidney Tubules in Metabolic Studies (Elsevier, Amsterdam). 6. Seglen, P. 0. (1979) Disaggregation and separation of rat liver cells, in Cell Pop&zfions (Reid, E., ed.), Wiley, New York, pp. 25-46. 7. Thurman, R. G., Kauffman, F. C., and Jungermann, K. (eds.) (1986) Regulation of hepatic metabolism, in Infer- and Infru-Cellular Compartmenfafion (Plenum, New York). 8. Harris, R. A. and Cornell, N. W., eds. (1983) Isolation, Characterization, and Use of H~tocytes (Elsevier Biomedical, Amsterdam). 9. Krebs, H. A., Lund, I., and Edwards, M. (1979) Criteria of metabolic competence of isolated hepatocytes, in Cell PopuIafions (Reid, E., ed.), Wiley, New York, pp. l-6. 10. Seglen, P. 0. (1972) Preparation of rat liver cells. I. Effect of calcium on enzymatic dispersion of isolated perfused liver. Exp. Cell. Res. 74,450-454. II. Berg, T. and Msrland, J. (1975) Induction of tryptophan oxygenase by dexamethasone in isolated hepatocytes. Dependence oncomposition of medium and pH. Biochim. Biophys. Acfu. 392,233-241. 12. Quistorff, B., Bondesen, S., and Grunnet, N. (1973) Preparation and biochemical characterization of parenchymal cells from rat liver. Biochim. Biophys. Acfu. 320, 503-516. 13. Bellemann, P., Gebhardt, R., and Mecke, D. (1977) An improved method for the isolation of hepatocytes from liver slices. And. Biochem. 81,408-415. 24. Quistorff, B. (1986) Gluconeogenesisinperiportal and perivenoushepatocytesof rat liver, isolated by a new high-yield digitonin/collagenase perfusion technique. Biothem. J 229,221-226. 15. Jurin, R. R. and McCune, S. A. (1985) Effect of cell density on metabolism in isolated rat hepatocytes. 7. Cell. Physiol. 123,442448. 16. Katz, J., Golden, S., and Wals, P. (1976) Stimulation of hepatic glycogen synthesis by amino acids. PYOC.Nufl. Acud. Sci. 733433-3437. 17. Boyd, M. E., Albright, E. B., Foster, D. W., and McGarry, J. D. (1981) In vitro reversal of the fasting state of liver metabolism in the rat. 1. Clin. Invest. 68,142-152. 18. Seglen, P. O., Gordon, P. B., and Poli, A. (1980) Amino acid inhibition of autophagic/ lysosomal pathway of protein degradation in isolated rat hepatocytes. Biochim. Biophys. Acfu. 630,103-118. 29. Chance, B. and Williams, G. R. (1955) Respiratory enzymes in oxidative phosphorylation. III. The steady state. 1. Biol. Chem. 217,409-427. 20. Berry, N. M., Grivell, A. R., and Wallace, P. G. (1980) Energy dependent regulation of the steady state concentration of the components of the lactate dehydrogenase reaction in the liver. FEBS Left. 119,317-322. 21. Dich, J. A. and Glud, C. N. (1976) Effect of glucagon on cyclic AMP, albumin metabolism and incorporation of l4c-leucine into proteins in isolated parenchymal rat liver cells. Acfu. Physiol. Scud. 97,457469. 22. Bergmeyer, H. U. (1974)Mefhods of Enzymatic Analysis. 2nd English Ed. (Academic, New York). 23. Lowry, 0. H. and Passonneau, J. V. (1972) A Flexible System of Enzymatic Analysis (Academic, New York).

Chapter 15
Primary Cultures of Rat Hepatocytes
John Dich and Niels
1. Introduction
With the advent in 1969 of the collagenase-perfusion technique for the high-yield preparation of isolated, differentiated hepatocytes (I), easy establishment of primary cultures of hepatocytes was made possible. Since then, this experimental system has been increasingly used in many research fields. Hepatocytes are anchorage-dependent cells, and cultures can be established as confluent, nonproliferating monolayers. This experimental system is superior to other in vitro liver preparations because of a longevity of several weeks, as compared to a few hours for perfused liver or suspensions of isolated hepatocytes, making studies of cellular processes that occur within a time scale of hours or days feasible. Another advantage of primary hepatocyte culture, as compared to freshly isolated suspensions, may be that deficiencies of the latter, resulting from damage of the cell membrane by the collagenase treatment, are resolved during the early stages of culture (2-4). Examples of topics that have been studied in primary cultures of hepatocytes are DNA-synthesis (5,6), transcriptional and translational events including hormone effects (7-g), the regulation of hepatic metabolism, (ZUJZ), secretory processes, (12-14), metabolism and effects of xenobiotics 161

Grunnet

162

Dich and Grunnet

such as ethanol and drugs (25-17), and hepatotoxicity, including carcinogenic and mutagenic effects (18). Hepatocyte cultures have also been used as an activating system in the Salmanellu mutagenicity test (19). One important consideration when using primary cultures of hepatocytes is the phenotypic stability of the cells. Early work with hepatocyte cultures demonstrated changes in characteristic functions of differentiated hepatocytes, such as declining activity of liver-specific enzymes (24) and cytochrome P-450 content (3,15,20,22), increased production of a-fetoprotein (22), and enhanced response to @-adrenergic agents (4,23,24). Although much work has been devoted to the establishment of phenotypic stable cultures, many reports appear without adequate characterization of the culture system actually used. This aspect will be discussed in more detail in section 4. In the following sections, a standard procedure for the preparation of primary monolayer cultures of hepatocytes from adult rats will be described, together with some methods for the preparation of samples for analysis. Methods for preparation of hepatocyte cultures from neonatal and postnatal rats and from other species, and conditions for proliferating hepatocyte cultures will not be described in detail, but will be discussed briefly in section 4.

2. Materials
1. Adult Wistar rats, weighing 200-230 g and starved for 16 h. Animals are housed at 21C with alternating 12-h cycles of light (HAM-~IJM) and darkness. 2. A solution of collagen prepared from rat tails. Six to seven tails are immersed in ethanol (70%) to ensure sterility. Under sterile conditions, the tails are skinned, and the four ligaments are pulled off and transferred to a sterile flask with 1 L of acetic acid (0.1%). After stirring for 48 h at 4OC,the suspension is centrifuged (10,OOOg for 30 min) and the supernatant is decanted. The collagen solution is kept at 4OCand can be used for at least 1mo. Coating of Petri dishes is normally performed the day before isolation of the cells. A collagen solution of 1.0-1.5 mL is evenly distributed on the dish. Immediately before plating, the dish is tilted, the solution is carefully sucked off, and the dish is used without further treatment. 3. The composition of the standard hepatocyte culture medium is given in Table 1. Stock solutions, 50x concentrated, of solutions 1,3a, 3b, 3c, and 4 can be prepared in water. Preparation of stock solution of folic

Cultures of Rat Hepatocytes Table 1 Hepatocyte Culture Medium, HCM, Standard Vitamins and glutathione mg/L w L-Ascorbic acid D-Biotin Choline. HCl Inositol Nicotinamide Panthothenate, Ca Pyridoxine. HCl Thiamine. HCl Vitamin B,, Glutathione, reduced Folic acid and riboflavin Folk acid Riboflavin Inorganic salts CaC& 2H,O MgSO,. 7I-JO NaCl KC1 Na,HPO,. 2H,O NaHCO, Carbohydrates D-Glucose, H,O Amino acids L-Alanine L-Asparagine L-Arginine L-Aspartic acid L-Cysteine L-Glutamic acid Glycine L-Histidine L-Isoleucine L-Leucine L-Lysine, HCl
1.80 1.50 3.00 0.46 0.57 0.80 4.00 0.51 0.81 1.50 2.45 160.4 225.2 522.6 61.2 69.1 117.7 300.3 79.2 106.3 99.4 4.1 1790.4 10.0 8.2 2.1 4.9 29.6 0.15 48.8 17.5 1.0 250.0 1.8 1.0 1.0 1.0 10.0 0.2 15.0

163

2.3 2.7

1.0 1.0

mM
0.82

mg/L
120.0

1.99
72.33 2.01 2.11 0.59 25.00

490.0
4225.2 150.0 376.0 80.0 2100.0

wpo4

196.8
447.4

(continued)

164 Table 1 (continued) Amino Acids L-Methionine L-Phenylalanine L-Proline L-Serine L-Threonine L-Tyrosine L-Tryptophan L-Valine L-Glutamine Antibiotics, pH-indicator
Penicillin G

Dich and Grunnet

mg/L
0.50 1.50 3.00 74.6 247.8 345.3 105.1 117.9 217.4 140.9 117.2 584.5

1.00 0.99
1.20

0.69 1.00
4.00

Streptomycin sulfate Phenol red

60.0 100.0

10.0

acid and riboflavin requires the addition of NaOH. Preparation of stock solution of amino acids (minus glutamine) requires addition of 150-mL cont. HCl. Portions of 20 mL are stored at -18OC and can be used for at least half a year. Preparation of 1L of HCM: a. Dissolve 20 mL each of solutions 1,2, and 5 in 500 mL of water. Adjust pH to 7.4 with NaOH (about 25 mmol). b. Add 20 mL of solutions 3a, 3b, 3c, and 4. c. Add NaHCO,, L-glutamine, phenol red, and antibiotics as indicated in Table 1. d. The solution is bubbled with OJCO, (19:1, vol/vol) for 30 min and pH is adjusted to pH 7.4 if necessary. e. Add water to 1 L. 4. Stock solutions of oleate (250 mM), palmitate (175 mM), plus linoleate (75 mM) are prepared by dissolving the fatty acids in dimethyl sulfoxide. Appropriate volumes are filled in ampules, sealed under N2, and stored at -18OC. 5. HCM with albumin and fatty acids: Prepare 1 L of HCM as described above, and then heat the medium to 37OC. Add, with stirring, 10 g of fatty acid-free bovine albumin, obtained commercially or prepared according to ref. 25. Complete solubilization takes 1-2 h. To the warm albumin-containing medium add slowly, with stirring, 1.0 mL of the stock solution of fatty acids. Continue stirring at 37OC for 2-3 h and afterwards at 4OCuntil the next day. Centrifuge the medium in sterile tubes (10,OOOg for 30 min).

Cultures of Rat Hepatocytes

165

6.

7. 8. 9. 10. Il. 12.

13.

Other hydrophobic compounds may be solubilized in dimethyl sulfoxide. Care should be taken to avoid high concentrations of dimethyl sulfoxide in the medium, as concentrations of l-2% have been reported to have their own effects on cultures of hepatocytes (26,27). Pentothal-Natrium: 0.5 g is dissolved in 10 mL of sterile water before use. Hibitane: A 0.5% solution in water is prepared before use. Iodide solution containing 50 mg of iodine and 35 mg of potassium iodide/ml of ethanol. Trypan blue: A solution is prepared as described in Chapter 14 (28). Portions are stored at -18OC. Penicillin and streptomycin. Dexamethasone is obtained as Decadron (4 mg/mL) from Merck, Sharp & Dome, Haarlem, Nederlands and kept at 4C. Crystalline insulin and glucagon. Solutions are prepared in 40 mM phosphate buffer, pH7.4 with 1% human albumin after dissolving the desired amounts of insulin and glucagon in a small amount of diluted HCl and NaOH, respectively. Aliquots can be stored at -18OC for at least 1 yr. The water used for preparation of all solutions should be of high quality as, e.g., triple distilled water. During the isolation of the hepatocytes, sterile technique is necessary. Glassware, instruments for surgery, tubings, and so on, should be autoclaved. Alternatively, washing with ethanol (70%) is advisable. All solutions and media are filtered through 0.45 pm filters. Before use, media are brought to room temperature and bubbled with sterile OJCO, (19:1, vol/vol) if pH is too high. After isolation of the cells, all work is performed in a laminar flow bench.

3. Methods
1. Rats,usually starved for 16h, areanesthesizedwithPentothal-Natrium (10 mg/lOO g rat), administered intraperitoneally. Care is taken to avoid noise and harmful distress. The rat is immobilized on the dorsal side. Before surgical procedures, the ventral side of the animal is washed thoroughly with Hibitane solution. A midline incision through the skin of the abdominal cavity is performed. The skin is pulled aside and the underlying tissue is washed with iodine solution before the abdominal cavity is opened. Cannulation of portal and caval veins, perfusion of the liver, incubation, and filtration of dispersed cells are performed as described in Chapter 14.

Dich and Grunnet

2. After filtration of the dispersed liver, the cells obtained are transferred to two 50-mL centrifuge tubes with caps, and centrifuged at ZOg,, for 1 min. 3. The supernatant is discharged, and the cell pellets are gently suspended in a few mL of HCM. Additional medium is added to bring the volume in the tubes to about 30 mL, and the cells are thoroughly mixed by inverting the tubes several times. The cells are centrifuged as described above. 4. The cell pellets are resuspended in a few mL of HCM, and the suspension is transferred to a graduated 15-mL centrifuge tube. Sufficient medium is added to bring the volume to about 12 mL, and the cells are mixed again before centrifugation at 30gm, for 2 min. The amount of packed cells varies from preparation to preparation and is between 2 and 5 mL. 5. Immediately after centrifugation, the packed cells are suspended and diluted to 20x the packed volume with HCM. 6. To 0.1 mL of the diluted cell suspension, 0.1 mL of Trypan blue solution and 0.8 mL of Krebs-Henseleit buffer are added. After mixing, an aliquot is immediately transferred to a counting chamber, and the number of stained and unstained cells counted. Only cell preparations in which more than 85% of the cells exclude vital dye should be used for culturing. Normally, the hepatocytes from starved rats at this stage of preparation appear spherical with a distinct cell border in the light microscope. 7. The stock solution of cells is diluted with HCM and horse serum. Dexamethasone and insulinare added to final concentrations of 10% (vol/ vol), 1 w and 0.1 @I, respectively. The diluted cell suspension should contain 0.55 million cells /mL. Normally, the yield from a 200 g starved rat is between 250 and 350 million hepatocytes, which is sufficient for 100-150 culture dishes (60 mm). 8. Cell suspension of 4.5 or 3.0 mL is pipeted into 60- and 35-mm Petri dishes, respectively, equivalent to application of 120,000 cells/cm2. If multidishes with 24 chambers are used, the cell suspension is further diluted to O.Emillion cells/mL. To each chamber is added 1.5 mL cell suspension, equivalent to 120,000 cells/cm2. Since hepatocytes sediment rapidly, it is necessary to swirl the suspended cells frequently during pipeting. 9. After plating, the cells are incubated at 37OC with atmospheric air/ CO, (19:1, vol/vol> for 3 h. 10. At that time, the medium is sucked off in order to eliminate loose cells. 3.0 or 2.0 mL HCM with albumin. fattv acids. dexamethasone (0.1

Cultures of Rat Hepatocytes

167

or lw), insulin (0.1 u.M or 0.01 @I), and glucagon (0.1 nM) are added to 60- and 35-mm Petri dishes, respectively. Tomultidishes are added 0.5 mL of the same medium to each well. The cultures are normally changed every second day. With the medium used, the interval between changes can be prolonged to 3 d. When medium is changed, the dishes are tilted slightly. The medium is sucked off by using, e.g., a sterile Pasteur pipete connected to a suction device. Replacement of medium should be done cautiously to avoid mechanical damage of the monolayer. 11. Preparation of samples for analyses: a. Dishes to be used for analyses are initially placed on ice. The medium is collected and stored at -18OC or -70C until analysis. The dishes are afterward tilted and residual medium is sucked off. b. For determination of enzyme activities and DNA, the cells are homogenized in 1.5 mL (60-mm dish) glycyl-glycine buffer (glycyl-glycine, 25 mM, KCl, 150 mM, MgSO, 5 mM, EDTA, 5 mM, dithiothreitol, 1 mA4, and defatted albumin, 0.2%,pH7.5). Homogenization by ultrasonication (40 W for 10 s) is performed in the dishes without previous scraping. Unless analysis is performed immediately, homogenates should be frozen at -7OOC. c. For determination of the content of cytochromeP-450, medium is removed as described and 0.8 mL (60-mm Petri dish) phosphate buffer (phosphate, O.lM, EDTA, 1 mM, dithiothreitol, 1 mM, glycerol, 20%, and Lubrol PX, 2%) is added to the dishes. The cells are scraped with a rubber policeman, transferred to vials, and frozen at -70C. d. For determination of metabolites, 0.2 mL perchloric acid (70%) is added to a 60-mm dish with 3 mL medium. The dishes are swirled and allowed to stand on ice for 10 min. All material is transferred to a centrifuge tube and centrifuged. An aliquot of the supernatant is neutralized and stored at -7OOCuntil analysis. 4.

Notes

1. Results should preferably be related to DNA, and the DNAcontent/ culture dish should be reported. The common practice of using protein content as a measure of cell number may lead toerroneousresul ts, since the protein content/cell may vary with culture conditions. Fur-

168

Dich and Grunnet

thermore, accurate determination of the cellular protein content is incompatible with albumin- and serum-containing media. 2. Cell death may be evaluated by determination of the DNA-content of the monolayer, since dead cells detach from the substratum and are removed by medium changes. Cell disintegration (cell death or membrane damage) may be quantitated by determination of enzyme leakage to the culture medium. Lactate dehydrogenase is well suited for this purpose because of its high activity and cytosolic localization, but other enzymes may also beused, e.g., transaminases and the liver-specific argininosuccinate lyase. Uptake of Trypan blue by cells in culture is a poor criterion for cell survival, since dead cells detach from the substratum. Light microscopic observations are difficult to quantitate. Counting of cells is possible but cumbersome. 3. Several parameters can and should be applied to evaluate the metabolic competence and the maintenance of hepatocyte-specific properties during the culture period. It is worth mentioning that the majority of hepatocyte cultureexperiments have been carried out for a few days only, and that the choice of parameters of course has to depend on the purpose of the actual experiments. Some useful parameters are described in the following: a. Hepatocyte-specific enzymes (seealso Chapter 14, this volume). Glucokinase, pyruvate kinase (L-form), and urea-cycle enzymes are hepatocyte specific and should ideally be maintained at least at the same activity/cell as in vivo. b. ATP-level. The level of ATP reflects cell integrity and the metabolic condition of the cells, and ought to be no less than 2.5 pmol/g cells (29,30), corresponding to 1 pmol ATP/mg DNA. c. Rate of metabolic pathways. Gluconeogenesis (10,21,32,32), glycogen synthesis (33,34), urea synthesis (35,36), fatty acid oxidation (16), and esterification (12,16,37), ethanol metabolism (38,39,40), and protein synthesis (4) are examples of integrated metabolic pathways that require the interaction of intracellular compartments, and that have been shown to proceed at acceptable rates in hepatocyte cultures. d. Rate of secretory processes. Hepatocytes in vivo secrete anumber of proteins, which are not produced by other cell types. Examples are albumin, acute-phase proteins, and some lipoproteins. Albumin secretion is often used as a criterion for the maintenance of specific characteristics. It is, however, not the most sensitive liver-specific parameter, since other proper ties

Cultures of Rat Hepatocytes

169

are lost before albumin secretion starts to decrease (36). Cultured hepatocytes produce and secrete bile acids (23), however, there are some indications that these are secreted at a much lower rate than in vivo as the cultures age, and that primary cultures of hepatocytes may represent a sort of cholestatic state (40. e. Response to hormones. Primary cultures of hepatocytes have been shown to respond to glucocorticoids (4,11,36,40,42,43), insulin (4,36,40,44), glucagon (12,35,36), a-adrenergic agonists (4,23), triiodothyronine (45,46), and growth hormone (45,47-50), indicating that receptors for these hormones are present and that a functional receptor-coupling exist in the cells. Conversely, the estrogen receptor has been reported almost to disappear within 24 h of culture (49). 4. Hepatocytes in primary culture survive better, when plated on some sort of substratum rather than directly on tissue culture polystyrene plastic. On tissue culture plastic, the cells do not spread, and the typical monolayer of polygonal cells with sharply delineated cell borders is not established (52). Furthermore, the cells detach rapidly from naked plastic (52). Primaria plastic (Falcon, Becton Dickinson) has been claimed suitable for hepatocyte cultures without further treatment (22). Satisfactory results have been obtained with plastic or glass surfaces coated with rat tail collagen (36,52), with floating collagen gels (53) or with collagen-gels supported by nylon mesh (25). Pure substances, e.g., Con A, collagen I, collagen IV, fibronectin, and laminin have also been successfully used as substrata (9,51,52,54,55), although they in some respects, e.g., transcription of liver-specific genes, appear inferior to rat tail collagen (9). Extracts of extracellular components, so-called biomatrix, have been reported to have advantages over other substrata (52,56); however, little comparative work has been carried out. Cocultures of hepatocytes and a rat liver epithelial cell line (57), the latter of which may be considered as a special case of substratum for the hepatocytes, appear to be superior to hepatocytes cultured on plain plastic. No systematic comparison of such cocultures and hepatocyte cultures on collagen-coated plastic has been reported. 5. Commercially available tissue culture media (seeAppendix) may be used for hepatocyte cultures, the most commonly used being Leibovitz L15, Dulbeccos Modified Eagles Medium, Williams E, Waymouth 752/l, RPM1 1640, and Ham F-12, or mixtures thereof. A num-

170

Dich and Grunnet ber of modifications havebeen claimed to improve the performance of hepatocyte cultures. In the following, the various media components will be discussed shortly: a. Serum. l-10% is beneficial during the initial adhesion period of Z-4 h (58). At later time-points, serum addition has been reported to improve cell survival (59) but also to increase the level of liver-unspecific mRNAs (60). Satisfactory culture conditions can be established without serum addition, and since serum is not chemically defined, it should be avoided, except for the initial adhesion period. b. Substrates. A high concentration of glucose (25 mM) has proven superior to a lower concentration (5 mM) as judged by the activity of alcohol dehydrogenase (40). By the same criterion, high amino acid concentrations (Table 1) are superior to lower concentrations (40). This effect of amino acids may be related to the depletion of several amino acids from the usual standard media within 24 h (24,55,61). A high concentration of amino acids may enable medium change every second or third day, only. Substitution of arginine by ornithine may retard growth of nonparenchymal cells without affecting hepatocytes (62). However, growth of nonparenchymal cells is normally not a problem. c. Fatty acids are not an essential component of media for hepatocyte cultures, at least not for short time cultures. They are, however, a quantitatively important substrate for the liver cells in vivo, and may be added to culture media as an albumin-fatty acid complex to attain physiological concentrations. d. Micronutrients. Although a requirement for metal ions, vitamins, or other usual media components (inositol, choline, purine, and pyrimidine bases) has not been documented for hepatocyte cultures, these substances are often included as a matter of precaution. Ascorbate (63) and selenium (64) have been reported to improve the maintenance of cytochrome P-450. e. Hormones. There is a general agreement that glucocorticoids significantly improve survival and performance of primary hepatocyte cultures. Dexamethasone is normally preferred over naturally occurring glucocorticoids because of its resistance to degradation (65). Insulin is also in many respects beneficial to the cultures (4,7,12,14,36,44), and the combined action of glucocorticoid and insulin is necessary for the expression of some hepatocyte-specific functions (7,11,50,66). It is therefore

Cultures of Rat Hepatocytes

171

advisable to add both hormones to cultures of hepatocytes. Insulin at a concentration of 10-8M is rapidly degraded by hepatocytes with a half-life of about 8-16 h in primary cultures (unpublished results, 67). A general effect of other hormones on confluent hepatocyte cultures has not been reported, although they respond to several hormones (seeNote 3). f. Other medium additions. Addition of 2% dimethyl sulfoxide to the medium has been reported to maintain albumin secretion in cultures for up to 40 d (26). Addition of glycosaminoglycans or proteoglycans to the medium has been shown to enhance the expression of liver-specific genes and to suppress the expression of tissue-unspecific genes (9). A number of medium additions have been recommended for maintaining the content of cytochrome P-450 in hepatocyte cultures, e.g., nicotinamide (68) metyrapone (69), heme (64), ascorbate (63), selenium (64), or high concentrations of dexamethasone (3). However, addition of these substances only partially prevents the decrease of the content of cytochrome P-450. 6. Hepatocytes prepared from fetal rat liver divide in primary culture, respond to several hormones including androgens and estrogens, and mature during culture to express adult hepatocyte characteristics (70). Hepatocytes from adult rat livers proliferate in primary culture, if the cells are plated at a low density (50,000 cells/cm*), and if the medium contains glucocorticoid, insulin, and epidermal growth factor (5,60). Once confluency is reached, the cells stop dividing. Expression of hepatocyte characteristics in this preparation appears inversely proportional to the rate of cell division (71). 7. The preparation of primary cultures has been described using livers from a number of other species including humans (72,73), dogs (73), monkeys (73), mice (74), hamsters (74), guinea pigs (73,74), rabbits (73,74), chicken embryos (75), and fish (76).

References
2. Berry, M. N. and Friend, D. S.(1969) High-yield preparation of isolated rat liver parenchymal cells. J Cell Bid. 43,506-520. 2. Bissell, D. M., Hammaker, L. E., and Meyer, U. A. (1973) Parenchymal cells from adult rat liver in nonproliferating monolayer culture. J Cell Bid. 59,722-734. 3. B&sell, D. M. and Guzelian, P.S.(1980) Phenotypic stability of adult rat hepatocytes in primary monolayer culture. Ann. N.Y. Acud. Sci. 349,85-98. 4. Ichihara, A.,Nakamura, T., and Tanaka, K. (1982)Use of heptocytes in primary culture for biochemical studies on liver functions. MoI. Cell. Biochem. 43,145-X0.

172
5. 6.

Dich and Grunnet

Nakamura, T. and Ichihara, A. (1985) Control of growth and expression of differential functions of mature hepatocytes in primary culture. Cell Structure and Function lO, l-16. Tomura, A., Sawada, N., Sattler, G. L., Kleinman, H. K., and Pitot, H. C. (1987) The control of DNA synthesis in primary cultures of hepatocytes from adult and young rats: Interaction of extracellular matrix components, epidermal growth factor, and the cell cycle. J. CeII. Physiol. 130,221-227. Jefferson, D. M., Reid, L. M., Giambrone, M.-A., Schafritz, D. A., and Zern, M. A. (1985) Effects of dexamethasone on albumin and collagen expression in primary cultures of adult rat hepatocytes. HepatoIogy 5,14-20. Steward, A. R., Wrighton, S. A., Pasco, D. S., Fagan, J. B., Li, D., and Guzelian, P. S. (1985) Synthesis and degradation of 3-methylcholanthrene-inducible cytochrome P-450 and their mRNAs in primary monolayer cultures of adult rat hepatocytes. Arch. Biochem. Biophys. 241,494-508. Fujita, M., Spray, D. C., Choi, H., Saez, J. C., Watanabe, T., Rosenberg, L. C., Hertzberg, E. L., and Reid, L. M. (1987) Glycosaminoglycans and proteoglycans induce gap function expression and restore transcription of tissue-specific mRNAs in primary liver culture. Hqxltology 7,1S-9s. Probst, I., Schwartz, P., and Jungermann, K. (1982) Induction in primary culture of gluconeogenic and glycolytic hepatocytes resembling periportal and perivenous cells. Eur. 1. Biochem. 126,271-278. Kletzien, R. F., Stumpo, D. J., Kelley, D. S., and Todderud, C. G. (1983) Primary cultures of hepatocytes as a model system for studies on the chronic effects of hormones on hepatic carbohydrate metabolism, in Isolation, Characterization, and Use of Hep&cytes (Harris, R. A. and Cornell, N. W., eds.), Elsevier, New York, pp. 77-86. Mangiapane, E. H. and Brindley, D. N. (1986) Effects of dexamethasone and insulin on the synthesis of triacylglycerols and phosphatidylcholine and the secretion of very-low-density lipoproteins and lysophosphatidylcholine by monolayer cultures of rat hepatocytes. Biochem. ].233,151-160. Gebhardt, R. (1986) Use of isolated and cultured hepatocytes in studies of bile formation, in Isolated and Cultured Hep~focytes (Guillouzo, A. and Guguen-Guillouzo, C., eds.), John Libbey, London and Paris, pp. 353-376. Hutson, S. M., Stinton-Fischer, C., Shiman, R., and Jefferson, L. S. (1987) Regulation of albumin synthesis by hormones and amino acids in primary cultures of rat hepatocytes. Am. J Physiol. 252, E291-E298. Sirica,A.E.andPitot,H.C.(1980)Drugmetabolismandeffectsofcarcinogensincultured hepatic cells. Phurmucol. Reu. 31,205-228. Grunnet, N., Kondrup, J., and Dich, J. (1985) Effect of ethanol on lipid metabolism in cultured hepatocytes. Biochem. J. 228,673-681. Guillouzo, A. (1986) Use of isolated and cultured hepatocytes for xenobiotic metabolism and cytotoxicity studies, inIsolated and Culfured Heputocyfes (Guillouzo, A. and Guguen-Guillouzo, C., eds.), John Libbey, London and Paris, pp. 313-332. Michalopoulos, G. K., Strom, S. C., and Jirtle, R. L. (1986) Use of hepatocytes for studies of mutagenesis and carcinogenesis, in IsoZuted and Cultured Hepatocytes (Guillouzo, A. and Guguen-Guillouzo, C., eds.), John Libbey, London and Paris, pp. monolayer culture: A tool for investigating the regulation of carcinogen metabolism, in Microsomes, Drug Oxidations, and Chemical Carcinogenesk (Coon, M. J., Conney, A. H., Estabrook,

7. 8.

9.

10. 21,

12,

13. 14. 25. 16. 27.

18.

333-352. 19. Gayda, D. P. and Pa&a, M. W. (1980) Rat hepatocytesinprimary

Cultures of Rat Hepatocytes

173

R. W., Gelboin,H. V., Gilette, J. R., and OBrien, P. J., eds.), Academic,New York, pp. 1165-1168. 20. Guzelian, P. S., Bissell, D. M., and Meyer, U. A. (1977) Drug metabolism in adult rat hepatocytes in primary culture. Gust~oente~oZogy 72,1232-1239. 21. Fry, J. R. and Bridges, J. W. (1979) Use of primary hepatocyte cultures in biochemical toxicology. Rev. Biochem. Toxicol. 1,201-247. 22. Sirica, A. E., Richards, W., Tsukada,Y.,Sattler, C. A.,and Pitot,H. C. (1979) Fetal phenotypic expression by adult rat hepatocytes on collagen gel/nylon meshes. Proc. Natl. Acad. Sci. 76,283-287. 23. Tsujimoto, A., Tsujimoto, G., Azhar, S., and Hoffman, B. 8. (1986) Altered responsiveness to a- and j3-adreno-receptor stimulation in hepatocytes cultured in defined medium. Biochem. Phamacol. 35,1400-1404. 24. Refsnes, M. Sandnes, D., and Christoffersen, T. (1987) The relationship between j3adrenoreceptor regulation and beta-adrenergic responsiveness in hepatocytes. Eur. I. B&hem. 163,457-466. 25. Chen, R. F. (1967) Removal of fatty acids from serumalbuminby charcoal treatment. 1. Bid. Chem. 242,173-181. 26. Isom, H. C., Secott, T., Georgoff, I., Woodworth, C., and Mummaw, J. (1985) Maintenance of differentiated rat hepatocytes in primary culture. Proc. Natl. Acud. Sci. USA 82,3252-3256. 27. Baribault, H. and Marceau,N. (1986) Dexamethasone and dimethylsulfoxide as distinct regulators of growth and differentiation of cultured suckling rat hepatocytes. I. Cell. Physiol. 129,77&k 28. Seglen, P. 0. (1972) Preparation of liver cells. ExptI. CeZI Res. 74,450X%. 29. Krebs, H. A., Cornell, N. W., Lund, I., and Hems, R. (1974) Isolated liver cells as experimental material, in Regulation of HepdicMetuboZism (Lundquist, F. and Tygstrup, N., eds.), Munksgaard, Copenhagen, pp. 726-750. 30. Cornell, N. W. (1983) Evaluation of hepatocyte quality: Cell integrity and metabolic rates, in Isolation, Characterization, and Use of Heptocytes (Harris, R. A. and Cornell, N. W., eds.), Elsevier, New York, pp. 11-20. 31. Quistorff, B., Dich, J., and Grunnet, N. (19861 Periportal andperivenous hepatocytes retain their zonal characteristics in primary culture. Biochem. Biuphys. Res. Commun. 139,1055-1061. 32. Fleig,W. E.,Noether-Fleig,G.,Roeben,H.,andDitschuneit,H. (1984)Hormonalregulation of key gluconeogenic enzymes and glucose release in cultured hepatocytes: Effects of dexamethasone and gastrointestinal hormones in glucagon action. Arch. Biochem. Biophys. 229,368-378. of tri33. Dich, J., Bro, B., Grunnet, N., Jensen, F., and Kondrup, J. (1983) Accumulation acylglycerol in cultured hepatocytes is increased by ethanol and by insulin and dexamethasone. Biochem. J 212,617-623. 34. Fleig, W. E., Enderle, D., Steudter, S., Niither-Fleig, G., and Ditschuneit, H. (1987) Regulation of basal and insulin-stimulated glycogen synthesis in cultured hepatocytes. J Biol. Chem. 262,1155-1160. 35. Gebhardt, R., Bellemann, P., and Mecke, D. (1978) Metabolic and enzymatic characteristics of adult rat liver parenchymal cells in non-proliferating primary monolayer cultures. Exp. Cell Res. 112,4311141. 36. Dich, J., Vind, C., and Grunnet, N. (1987) Long-term culture of hepatocytes. Effect of hormones on enzyme activities and metabolic capacity. Heptology 8,3945. 37. Chambaz, J., Guillouzo, A., Cardot, P., Pepin, D., and Bereziat, G. (19861 Essential

174

Dich and Grunnet

fatty acid uptake and esterification in primary culture of rat hepatocytes. Biochim. Biophys. A& 878,310-319. Dich, J., Vind, C., and Grunnet, N. (1987) Long-term culture of hepatocytes. Ethanol oxidation and effect of ethanol on enzyme activities and albumin secretion. Alcohol and Alcoholism Suppl. 1,271-275. Wood, C. K. and Lamb, R. G. (1979)Theeffect of ethanol onglycerolipid biosynthesis by primary monolayer cultures of adult rat hepatocytes. Biochim. Biophys. A&. 572, 121-131. Grunnet, N., Vind, C, and Dich, J. (1989) Maintenance of alcohol dehydrogenase activity in long-term culture of hepatocytes from female rat. Alcohol. Clin. Exp. Res. 13, in press. Buscher, H.-P., Friedrich, H., Thorn, H., Schneider, S., Gerok, W., and Kurz, G. (1985) Hepatocyte primary cultures represent a cholestatic state. Eur. Assoc. for the Study of the Liver, abstract no. 51. Miyamaga, O., Evans, C., and Cottam, G. L. (1983) The effect of dexamethasone on pyruvate kinase activity in primary cultures of hepatocytes. Biochim. Biophys. Acfu. 758,42-48. Marceau, N., Baribault, H., and Leroux-Nicollet, I. (1984) Dexamethasone can modulate the synthesis and organization of cytokeratins in cultured differentiating hepatocytes. Can. J. Biochem. Cell Biol. 63,448-457. Katz, N. R., Nauck, M. A., and Wilson, P. T. (1979) Induction of glucokinase by insulin under the permissive action of dexamethasone in primary rat hepatocyte cultures. Biochem. Biophys. Res. Commun. SS,23-29. Gebhardt, R. and Mecke, D. (1979) The role of growth hormone, dexamethasone and triiodothyronine in the regulation of glutamine synthetase in primary cultures of rat hepatocytes. Eur. J. Biochem. 100,519-525. Wilson, E. J. and McMurray, W. C. (1983) Effects of hormones on the maintenance and mitochondrial functions of rat hepatocytes cultured in serum-free medium. Can. J. Biochem. Cell Biol. 61,636-643. Schaffer, W. T. (1985) Effects of growth hormone on lipogenic enzyme activities in cultured rat hepatocytes. Am. J. Physiol. 248, E719-E725. Mezey, E., Potter, J. J., and Rhodes, D. L. (1986) Effect of growth hormone on alcohol dehydrogenase activity in hepatocyte culture. Hepatology 6,1386-1390. Crabb, D. W. and Roepke, J. (1987) Loss of growth hormone-dependent characteristics of rat hepatocytes in culture. In Vitro 23,303-307. Grunnet, N., Vind, C., and Dich, J. (1988) Synergistic effect of insulin and glucocorticoid on albumin secretion in cultured hepatocytes. Acta Physiol. Scund. 133,593, 594. Gjessing, R. and Seglen, P. 0. (1980) Adsorption, simple binding and complex binding of rat hepatocytes to various in vitro substrata. Exp. Cell Res. 129,239-249. Reid, L. M.,Narita,M., Fujita,M.,Murray,Z.,Liverpool, C.,andRosenberg,L. (1986) Matrix and hormonal regulation of differentiation in liver cultures, in IsoZafed and Cultured Hepufocyfes (Guillouzo, A. and Guguen-Guillouzo, C., eds.), John Libbey, London and Paris, pp. 225-258. Michalopoulos, G. and Pitot, H. C. (1975) Primary cultureof parenchymal liver cells on collagen membranes. Exp. Cell. Res. 94,70-78. Bissell, D. M., Stamatoglou, S. C., Nermut, M. V., and Hughes, R. C. (1986) Interaction of rat hepatocytes with type IV collagen, fibronectin, and laminin matrices. Distinct matrix-controlled modes of attachment and spreading. Eur. J. Cell Biol. 40, 72- 78.

38. 39. 40. 41. 42. 43. 44. 45. 46. 47. 48. 49. 50. 51. 52.

53. 54.

Cultures of Rat Hepatocytes

175

55. Sawada, N., Tomomura, A., Sattler, C. A., Sattler, G. L., Kleinman, H. K., and Pitot, H. C. (1987) Effects of extracellular matrix components on the growth and differentiation of cultured rat hepatocytes. In Vitro 23,267-273. 56. Rojkind, M., Gatmaitan, Z., Mackensen, S.,Giambrone, M.-A., Ponce, P., and Reid, L. M. (1980) Connective tissue biomatrix: Its isolation and utilization for long-term cultures of normal rat hepatocytes. J. Cell Bid. 87,255-263. 57. Guguen-Guillouzo, C. and Guillouzo, A. (1983) Modulation of functional activities in cultured rat hepatocytes, Mol. CeZZ.,BioZ. 53/54,3!?-56. 58. Rubin, K,, Oldberg, A,, Hook, M., and Obrink, B. (1978) Adhesion of rat hepatocytes to collagen. Exp. Cell Res.117,165-177. 59. Williams, G, M., Bermudez, E.,and Scaramuzzino, D. (1977) Rat hepatocyte primary cell cultures. III. Improved dissociation and attachment techniques and the enhancement of survival by culture medium. In Vitro 13,809-817. 60. Enat, R.,Jefferson, D. M., Ruiz-Opaza, N., Gatmaitan, Z., Leinwand, L. A., and Reid, L, M. (1984) Heptocyteproliferationintifro: Itsdependence on theuseof serum-free hormonally defined medium and substrata of extracellular matrix. Proc.N&l. Ad. Sci. USA 81,1411-1415. 61. Schwarze, P, E., Solheim, A. E., and Seglen, P. 0. (1982) Amino acid and energy requirements for rat hepatocytes in primary culture. In Vitro 18,43-54. 62. Acosta, D., Anufuro, D. C., and Smith, R. V. (1978) Primary monolayer cultures of postnatal rat liver cells withextended differentiated functions. In Vitro 14,428-436. 63. Bissell, D. M. and Guzelian, P, S. (1979) Ascorbic acid deficiency and cytochrome P-450 in adult rat hepatocytes in primary culture. Arch. Biochern.Biophys. 192,
569-576. 64. 65.

Engelmann, G. L., Richardson, A. G., and Fierer, J. A. (1985) Maintenance and inductionofcytochromeP-45Oinculturedrathepatocytes. Arch. Biochem. Biophys.238,
359-367.

Lovell-Smith, C, J. and Garcia-Webb, P. (1986) Glucocorticoids and the isolated rat hepatocyte. Biochem.Biophys.Res.Commun. 135,160-165. 66. Christoffersen,T.,Refsnes, M., Brsnstad,G.O., Ostby,E., Huse, J.,Haffner, F.,Sand, T.-E., Hunt, N. H., and Sonne, 0. (1984) Changes in hormone responsiveness and cyclicAMPmetabolisminrathepatocytesduringprimarycultureandeffectsofsupplementing the medium with insulin and dexamethasone. Eur. I+Biochem. 138,217226. 67. 68. 69.

70. 71.

Fleig, W. E., N&her-Fleig, G., Stendter, S., Enderle, D., and Ditschuneit, H. (1986) Effect of down-regulation and return of insulin receptors on glycogen synthesis in cultured hepatocytes. Biochim.Biophys.Actu 888,191-198. Paine, A. J.,Williams, L. T., and Legg, R. F. (1979) Apparent maintenance of cytochrome P-450 by nicotinamide in primary cultures of rat hepatocytes. Life Sci. 24, 2185-2192. Paine, A, J.,Villa, P.,and Hockin, L. J. (1980)Evidence that ligand formation is a mechanism underlying the maintenance of cytochrome P-450 in rat liver cell culture. Biochem. J, 188,937-939. Kremers, P. (1986) Drug metabolism in cultured fetal hepatocytes, Is&fed and Cultured Hepafocyfes (Guillouzo, A. and Guguen-Guillouzo, C., eds.), John Libbey, London and Paris, pp. 285312. Nakamura, T., Yoshimoto, K., Nakayama, Y., Tomita, Y., and Ichihara, A. (1983) Reciprocal modulation of growth and differentiated functions of mature rat hepatocytes in primary culture by cell-cell contact and cell membranes. Proc.NufZ.Ad. Sci. USA N&7229-7233.

176
72.

Dich and Grunnet

Clement, B., Guguen-Guillouzo, C., Campion, J.-P., Glaise, D., Bourel, M., and Guillouzo, A. (1984) Long-term co-cultures of adult human hepatocytes with rat liver epithelial cells: Modulationof albumin secretion and accumulation of extracellular material. Heputology4,373-380. Byard, J. L., Reese,J.A., and Knadle, S.A. (1983) Isolation and culture of hepatocytes from liver biopsies, in Isolation, Churucterizdion, and Use ofHeputocytes (Harris, R. A. and Cornell, N. W., eds.), Elsevier, New York, pp. 69-76. Maslamsky, C. J. and Williams, G. M. (19831Isolation and culture of primary hepatocytes derived from rat, hamster, guinea pig, and rabbit, in Isolation, Churucterization, and Use of Hqutocytes (Harris, R. A. and Cornell, N. W., eds.), Elsevier, New York, pp. 87-92. Sinclair, J. F., Smith, L., Bement, W. J., Sinclair, P. R., and Bonkowsky, H. L. (1982) Increases in cytochrome P-450 in cultured hepatocytes mediated by 3- and 4-carbon alcohols. Biochem. Pharmucol. 31,2811-2815. Campbell, J.W., Aster,P. L., Casey,C. A., and Vorhaben, J.E. (1983) Preparation and use of fish hepatocytes, i&o&ion, Characterization, and Useofkkputocytes (Harris, R. A. and Cornell, N. W., eds.), Elsevier, New York, pp. 31-40.

73. 74.

75. 76.

Chapter 16
Preparation of Isolated Periportal or Perivenous Hepatocytes from Rat Liver
Bjern Quistorff

1. Introduction
There are a number of functional, metabolic differences between periportal and perivenous hepatocytes in the mammalian liver resulting from zonal differences in the activity of several enzymes, and possibly from morphological differences as well (1). Examples of key enzymes where the periportal/perivenous activity ratio is >1 are glucose-6-phosphatase (Z), phosphoenolpyruvate carboxykinase (3), alanine aminotransferase (4), ornithine carbamoyltransferase (5), and carbamoylphosphate synthetase (6), whereas enzymes such as glucokinase (7), P-450-dependent hydroxylation reactions (8), and glutamine synthetase (9) show the opposite zonation. Recent studies on lipogenic enzymes (acetyl-CoA carboxylase, citrate lyase, and fatty acid synthase) show that, although there is a periportal predominance in both the fed and fasted male rat, the ca. threefold enzyme induction observed after a fasting-refeeding transition takes place almost entirely in the perivenous zone (IO). It is clear that this metabolic heterogeneity that develops a few weeks after birth is closely related to the microcirculatory pattern of the liver (1). However, the signals responsible for the heterogenous expression of the 177

178

Quistorff

hepatocte genome as a function of location in the microcirculation are unknown at present, although the gradients of oxygen (II) and hormones (12) are implicated. The physiological significance of zonation is largely unknown. In case of ammonia fixation it has been suggested that urea synthesis located in the inlet part of the microcirculation serves as the main process of ammonia fixation, whereas the high affinity system of glutamine synthesis located at the very outlet of the liver is considered the backup system for ammonia removal (13). Thus, it is of considerable interest to study isolated periportal and perivenous hepatocytes separately. Several attempts have been made to separate isolated hepatocytes according to various physical and chemical characteristics, but generally with limited success. Recently, however, it was observed that digitonin perfusion allowed selective destruction of the periportal or perivenous part of the microcirculation (24), and by applying this principle in combination with traditional collagenase cell isolation (see Chapter 14), a method was developed for high-yield preparation of isolated hepatocytes enriched in either periportal or perivenous cells (15,16). Applying this technique, periportal predominance of the pathways of gluconeogenesis (16,17) and urea synthesis (17,M) in freshly islolated cells as well as in primary culture has now been demonstrated. A number of review papers on the topic of metabolic zonation of the liver have appeared (1,19,20,21).

2. Materials
1. Rats weighing 170-190 g. Animals are anesthetized with pentobarbital ip; 400 pL, 50 mg/kg. 2. Perfusion system: Switching between forward and reverse perfusion as well as switching between different perfusates may be done manually. We find it advantageous, however, to use a valve system as the one shown schematically in Fig. 1, e.g., Hamilton miniature valves (Hamilton Bonaduz AG, CH-7402 Switzerland) type HV 86729 for the switching of low direction and two type HV 86727 valves for perfusate switching. Alternatively miniature pinch-solenoid valves may be used. Our current perfusion system is based on such 3-way valves (model S30501 EG; Sirai Inc., Italy). The perfusion system is convenient to operate and ensures rapid and precise switching with avery small dead-space in the tubing system. Furthermore, the perfusion system is readily computer controlled, e.g., for more complicated perfusion sequences (22). The system is shown schematically in Fig. 2.

Isolated Hepatocytes

179

Fig. 1. Schematic perfusion setup for preparation of isolated periportal and perivenous hepatocytes. The liver is connected to the perfusion system via the portal vein (PP) and the v. cava sup. (PV). A 90 turn of valve A switches the flow direction through the liver from P+ C to C+P. A 90 turn of valve B switches the perfusate from I to II.

3. Solution I: 118 mM NaCl, 4.7 mM KCl, 1.2 mM KH,PO, 1.2 mM MgSO, and 25 mM NaHCO,. 4. Solution II: As solution I plus 3 mM CaCl, and 0.2-0.4 mg/mL collagenase. 5. Solution III: As solution I plus 1.2 mM CaCl,. 6. Solution IV: As solution I plus 4 mg/mL digitonin. All solutions (I-IV) are kept in a thermostated water bath at 39OC during the cell isolation procedure, constantly equilibrated with oxygen/carbon dioxide (19:1, vol/vol), except for solution IV, which is gassed only until the addition of the digitonin, a few minutes before use. 7. Equipment as described in Chapter 14, except that two roller type pumps (LKB MultiPerpex 2115, Bromma, Sweden) are used. A bubble trap is mounted on the infusion line between valve A and B, Fig. 1. Digitonin: Commercial preparation of digitonin vary considerably in 8. terms of purity and in terms of water solubility. We usually use the SIGMA digitonin. In order to obtain good water solubility, the following purification steps are necessary. Five g of digitonin is dissolved in 150 mL boiling methanol (65OC). This solution is cooled to 45OC and

180

Quistorff

RI
Pl

R2

Pl

8j
I !

~
I I
i I

1 4 4
v2

PD

v7

Fig. 2. Schematic perfusion setup based on three-way micro-pinch-solenoid valves. The liver is connected to the perfusion system as in Fig. 1. The valves V5 and V6 are operated synchroneously, switching the perfusion direction through the liver. 01 and 02 are two Clarck type of oxygen electrodes mounted on the infusion line and on the out flow line from the liver, respectively. PD is a pressure difference gage. V7 regulates whether the eluate from the liver is collected or discarded. Vl-V4 regulate which of four different perfusates are being infused. When one of these valves is not selected, the perfusate is automaticallly recycled to the proper reservoir, Rl-R4. N-P4 are different decks on the multiperpex pump or, alternatively, different pumps, e.g., used for fast switching of flow rate (27,221.

precipitated by addition of 450 mL room temperature diethyl ether. The suspension is filtered on a Buckner funnel. After three washings with diethyl ether, the remnant is collected from the funnel, resuspended in water, and freeze-dried for 2-3 d in order to remove all diethyl ether. The freeze-dried digitonin is weighed and redissolved in water (5 mL) in portions of 150 mg digitonin, which is the amount used for each experiment. After centrifugation (20000 g x min), the super-

Isolated Hepatocytes

181

natant is freeze-dried overnight and kept in the powder form in the same tube until use. Note that digitonin is toxic and easily absorbed through the skin. Handling should be done carefully, wearing gloves, and weighing of the fluffy, freeze-dried powder should be done in a hood. 9. Collagenase (type II) is used.

3. Method
The principle of the technique is to destroy selectively cells of either the periportal or perivenous part of the liver microcirculation with digitonin and then perform a standard collagenase cell isolation on the remaining part. Thus, only periportal or perivenous hepatocytes may be prepared from one liver. Table 1 lists the essential steps of the procedure. 1. A method for establishing the initial perfusion of the liver with KrebsHenseleit buffer (solution III) is as described in Chapter 14. After 5-10 min of perfusion, the small caudate liver lobe is routinely ligated, and cut off (approximately 0.3 g of tissue) and homogenized in ice-cold buffer (solution III with 5 mM dithiotreitol). This sample (the start biopsy) is used for analysis of marker enzymes in order to quantitiate the zonal purity of the prepared cells (seeNote 2). 2. The selective destruction of perivenous cells in order to isolate periportal cells is described first (see also Table 1). Flow direction is switched to cava+ porta, and flow rate is decreased to 7 ml/min. 3. Thirty seconds later, the digitonin containing perfusate (solution IV) is switched on and the liver surface is observed. Because of complexing with cholesterol, the digitonin will progressively destroy the plasma membrane of the cells along the path of the microcirculation. Experience has shown that under the present conditions it is possible to achieve a largely synchroneous destruction of the sinusoids, which advance slowly along the path of the perfusate, 50-loo-fold slower than the linear flow rate (14). This is demonstrated very clearly on the liver surface by the development over 30-50 s of a pattern of pale, white spots in the perivenous areas (the dark-brown spots on the normal liver). The goal is to continue the digitonin perfusion until a regularly scattered perivenous decoloration pattern on the liver surface is obtained as demonstrated in Fig. 3B. Under the present conditions, this involves 3045 s of perfusion. 4. The perfusion is then switched to the calcium-free perfusate (solution I), and a few seconds later (when the digitonin perfusate in the tubing

182

Quistorff

Table 1 Essential Steps in the Preparation of Periportal or Perivenous Hepatocytes Flow rate, mL/min flow direction 1. Initial perfusion 2. Selective destruction of PV cells 3. Calcium removal 2a.Selective destruction of PV cells 3a.Calcium removal 4. Collagenase perfusion 5. Rinse cycle 6. Incubation 7. Cell separation, washing 30 P+C 7 C+P 25 P-C
_------------------

Time min lo-15 0.5-l .o 8-12

Samples Biopsy

Buffer Solution III Solution IV

Eluate first 20 s

Solution I

7 P+C 25 C+,P
-------------------

0.5-l .o 8-12 Eluate first 20 s

Solution IV Solution I

25 P+C 25 P+C

10-20

l-2 10

Solution II (recirculation) Solution III Solution III Solution III

Steps 2 and 3 are replaced by steps 2a and 3a in order to prepare perivenous stead of periportal cells.

cells in-

between the liver and valve B, Fig. 1, is cleared), the flow direction is switched to porta+ cava and flow rate increased to ca. 25 mL/min. Initial eluate (15 s) during this interval is collected in 200 PL 40 mM dithiotreitol for marker enzyme analysis (seeNote 2). 5. The procedure for selective destruction of periportal cells in order to prepare perivenous cells is as described above, except that flow directions were opposite with periportal infusion of digitonin at a rate of

Isolated Hepatocytes

183

Fig. 3. Zonqxcific decolorization of the liver after digitonin perfusion. tal decoloration pattern, obtained after45 s of digitonin perfusion 4 mg/mL, B: The complementary perivenous pattern obtained after 45 s of digitonin mg/mL, 5.0 mL/min.

A: Peripor7 mL/min. perfusion 4

only 5 mL/min (see Table 1). The decoloration pattern obtained is complementary to the pattern described above, now with pale, periportal spots rapidly developing into a network-like pattern, (seeFig. 3A). 6. The switch to the calcium-free buffer (solution I) initiates the standard hepatocyte preparation, which is described in detail in Chapter
14.

184

Quistorff

4. Notes
1. Cell yields are usually higher for periportal than for perivenous cells, i.e., 1.3-2.0 x lo6 and 0.8-1.2 x lo6 cells, respectively (X,16). It is essential to keep the flow rate low during digitonin perfusion. With a flow rate of 20 mL/min, the cell yield was ca. fivefold lower, even though the amount of digitonin infused under these conditions was the same as with the low flow (26). It may improve the yield of cells to include deoxyribonuclease 20 pg/mL during the incubation of the isolated cells, step 6 in scheme 1 (15). In terms of viability, it is our experience that both periportal and perivenous cell preparations show the same viability of 85-95% unstained cells as judged by Trypan blue staining, which is not different from control cell preparations. Occasionally preparations of low viability occur. We recommend that such preparations be discarded. Alternatively, some separation of stained and unstained cells may be obtained (at the expense of cell yield) by metrisamide gradient centrifugation (15). (For other viability tests and metabolic competence cells, see Chapter 15.) 2. The purity of the cell preparation in terms of contamination of periportal cells to a preparation of perivenous cells and vice versa is determined by the zonal specificity of the digitonin destruction. It is clear from the pattern of marker enzymes eluted during continuous digitonin perfusion that ideal conditions of equal destruction of all sinusoids do not prevail (14), except for the initial eluation, which has a very high zonal specificity [< 1% contamination (22)]. Thus, a quantitative evaluation of the zonal purity of the final cell preparation is required. Figure 4 illustrates the problem schematically in a simplified model of the microcirculatory unit of the liver. Using this model, which assumes only two populations of hepatocytes, a quantitative evaluation may be attempted as explained below, based on the activity of alanine aminotransferase (ALAT), which is known to have a periportal/perivenous activity gradient of approximately IO-fold
(42).

In cell preparation obtained according to the method described above, we have measurements of ALAT specific activity: a. in the start biopsy, b. in the eluate obtained after digitonin perfusion and c. in the final cell suspension.

Isolated Hepatocytes

185

Control

conditions

PP- Digitonin B I I

pulse

A
PP Schematic microcirculatory unit of liver PV

I
\

PV-Digttonin BIOPSY.
6P*6V 12= 1 TP*;v

pulse

PV- CELLS :
3P+6V ---~P+.$v 9 I

PP- CELLS :

Fig. 4. Model of the liver microcirculation: effect of digitonin. The model assumes only two populations of hepatocytes, periportal (PI3 and perivenous (PV), each of which is homogenous with respect to the parameter in question, but not necessarily of the same size. The figure shows two examples of selectivedestruction with a pulse of digitonin. B shows periportal destruction, which, with ensuing collagenase cell isolation will give a cell preparation enriched in perivenous cells by 67% (see numeric example below figure). Conversly, C shows the case of perivenous cell destruction.

Assuming the two-population model of Fig. 4, the volume fraction v of periportal cells in the intact liver may be calculated to 0.52 (23). This value for v may now be substituted in the equation 0) where Abi is the specific activity of ALAT in the start biopsy, expressed as units/mg of cytosolic protein (see23), while App and Apv is the specific activity in periportal and perivenous cells respectively (expressed as units/mg of cytosolic protein). Since in one particular experiment either App or Apv is known (measured in the initial eluate collected during digitonin perfusion), the other may be calculated. The values obtained for App and Apv may now be substituted in the equation Abi = (v)App + (1-v)Apv

Quistorff Ace11 = 0) App + (l-f, Apv

(2)

where Ace11 is the specific activity in the actual preparation of periportal or perivenous cells, f the fraction of periportal cells, and conversely (l-f> the fraction of perivenous cells. This relation thus allows calculation of the purity of the cell preparation (for l-f> under the assumption of a two-population model for the distribution of the enzyme in question. Typical results using this calculationshowcontaminationofperiportalcellpreparationwith25-35% perivenous cells and vice versa 20-30% contamination of perivenous cells with periportal cells. It has proven difficult to improve these figures significantly. More extensive digitonin destruction does not seem to be the way to go, since in our experience this only decreases cell yield without increasing purity significantly.

References
2.

2. 3. 4, 5, 6. 7. 8. 9.
10. Il.

Thurman, R. G., Kaufman, F. C., and Jungermann, K. (eds.) (1986) Regulation of Hqafic Metabolism. Inter- and Infra Cellular Comparfmentation (Plenum, New York). Katz, N., Teutsch, H. F., Sasse,D., and Jungermann, K. (1977) Heterogeneous distribution of glucose-6-phosphatase in micro dissectedperiportal and perivenous rat liver tissue. FEBS Left. 76,226-230. Guder, W. G. and Schmidt, U. (1976) Liver cell heterogeneity. The distribution of pyruvate kinase and phosphoenolpyruvate carboxykinase in the lobule of fed and starved rats. Hoppe-Seylers Z. Physiol. Chem. 357,1793-1800. Welsh, F. A. (1972) Changes in distribution of enzymes within the liver lobule during adaptive increases. 1. Hisfochem. Cytochem. 20,107-111. Mizutani, A. (1968) Cytochemical demonstrationof ornithine carbamoyltransferase activity in liver mitochondria of rat and mouse. 1. Hisfochem. Cyfochem. 16,172- 180. Gaasbeek-Janzen,J. W., Lamers. W. H., Morman, A. F., de Graf, A., Los, J. A., and Charles, R. (1984) Immunohistochemical localization of carbamoylphosphate synthetas@ in adult rat liver. J. Histochem. Cytochem. 32,557-564. Trus, M., Zawalich, J.,Gaynor, D., and Matschinsky, F. M. (19801 Hexokinase and glucokinase distribution in the liver lobule. 1.Histochem. Cyfochem. 28,579581. Smith, M. T. and Wills, E. D. (19811 Effects of dietary lipid and phenobarbitone on the distribution of cytochrome P-450 in the liver studied by quantitative histochemistry. FEBS Left. 127,33-36. Gebhardt, R. and Mecke, D. (1983) Heterogeneous distribution of glutamine synthase among rat liver parenchymal cells in situ and in primary culture. Embo 1.2, 567-570. Evans, J. L., Quistorff, B.,and Witters, L. A. (1988) Zonation of hepatic lipogenic enzymes identified by dual-digitonin-pulse perfusion. Biochem, J., in press. Wolfe,D. and Jungermann,K. (1985) Long-termeffectsof physiologicaloxygenconcentrations on glycolysis and gluconeogenesis in hepatocyte culture. Eur. J. Biothem. 151,299-303.

Isolated Hepatocytes

187

12. Probst, I. and Jungermann, K. (1983) The glucagon-insulin antagonism and glucagon-dexamethasone synergism in the induction of PEP-carboxy kinase in cultured rat hepatocytes. Hoppe-Seylers Z. Physiol. Chem. 364,1739-1746. 23. H&rssinger, D. (1983) Hepatocyte heterogeneity in glutamine and ammonia metabolism and the role of an intracellular glutamine cycle during ureagenesis in perfused rat liver. Eur. J Biochem. 133,418-422. 24. Quistorff, B.,Grunnet, N., and Cornell, N. W. (1985)Digitonin perfusion of rat liver. A new approach in the study of intra-acinar and intra-cellular compartmentation in the liver. Biochem. J. 226,289-297. 35. Lindros,K. 0. and Penttila, K. E. (1985) Digitonin-collagenaseperfusion for efficient separation of periportal or perivenous hepatocytes. Biochem. J 228,575-560. 16. Quistorff, B. (1985)Gluconeogenesis in periportal and perivenous hepatocytesof rat liver, isolated by a new high-yield digitonin/collagenase perfusion technique. Biochem. J 229,221-226. 17. Quistorff, B., Dich, J., and Grunnet, N. (1986) Periportal and perivenous hepatocytes retain their zonal characteristics in primary culture. Biochem. Biophys. Res. Commun. 139,1055-1061. 18. P&so,R. A., Penttila, K. E., Suolinna, E. M., and Lindros, K. E. (1986) Urea synthesis in freshly isolated and in cultured periportal and perivenous hepatocytes. Biochem.
J. 239,263-267.

19. Jungermann, K. and Katz, N. (1982) Functional hepatocellular heterogeneity. Hepatology 5385-395. 20. Jungermann, K. (1986)Functional heterogeneity of periportal and perivenous hepatocytes. Enzyme 35,161-180. 21. Gumucio, J.J. and Chianale, J. (1988) Liver cell heterogeneity and liver function, in The Liver, Biology and Pathology (Arias, I. M., Jakoby, W. B.,Popper, H., andschachter, D., eds.), Raven, New York, pp. 931-948. 22. Quistorff, B. and Grunnet, N. (1987) Dualdigitonin-pulse perfusion. Concurrent sampling or periportal and perivenous cytosol of rat liver for determination of metabolites and enzyme activities. Biochem. J 243,87-95. 23. Quistorff, B. (1987) Digitonin perfusion in the study of metablic zonation of the rat liver: Potassium as an intracellular concentration reference. Biochem. Sac. Transact. 15,361363.

Chapter 17
Primary Kidney
Mary Taub

Cells

1. Introduction
Hormonally defined serum-free media havebeen developed for growth and functional studies with kidney epithelial cell cultures. Not only can several established kidney epithelial cell lines (MDCK and LLC-PKl) be grown in a serum-free environment (1,2), but in addition, primary cultures of kidney epithelial cells can also be grown serum free (l-4). Investigations with primary kidney cell cultures are advantageous for several reasons. First of all, kidney cells can be grown in vitro from the animal of choice. Thus, the results of tissue culture studies can be more closely correlated with animal studies. Secondly, new tissue culture systems can be developed that more closely resemble kidney cells in vivo than those obtained with established kidney cell lines. This report describes the use of serum-free medium to grow primary rabbit kidney proximal tubule cell cultures that express proximal tubule functions. Rabbit kidney proximal tubules are first purified from the renal cortex by a modification (4-6) of the method of Brendel and Meezan (7,B). The purified proximal tubules are then put into tissue culture dishes containing serum-free medium supplemented with insulin, transferrin, and hydrocortisone. Within the first day of culture, the tubules attach to the dish, and subsequently, epithelial cells grow out from the tubules. After one week confluent monolayers are obtained that express a number of proximal tubule functions (Table 1, Fig. 1). These monolayer cultures can be used for a wide range of purposes, including viral production, 189

Taub Table 1 Properties of Primary Proximal Tubule Cell Cultures Morphology (4,9,2(I) Domes Form polarized monolayers Adjacent cells form tight junctions Brush border (although not as elaborated as in vivo) Transport properties Na+-dependent sugar transport (4,9) Na+-dependent phosphate transport (6) Probenacid-sensitive p-aminohipurate transport (21) Amiloride-sensitive sodium transport (12) Hormone responses Parathyroid hormone-sensitive CAMP production (4) Enzymes; metabolic properties Leucine aminopeptidase (4) Alkaline phosphatase (4) y glutamyl transpeptidase (4) Glutathione S-transferase (16) Angiotensin converting enzyme (10) Phosphoenolpyruvate carboxykinase (13) Aerobic metabolism (9) Hexose monophosphate shunt (9) Growth properties Cell growth in serum-free medium (4) Growth in response to insulin, transferrin, and hydrocortisone (4) Growth in response to laminin, collagen, and fibronectin (14) Can undergo two passages

transfection and immortalization, biochemical and physiologic studies, as well as molecular biologic studies concerning control of expression of differentiated function.

2. Materials
1. The basal medium used in these studies is a 50/50 mixture of Dulbeccos Modified Eagles Medium (with 4.5 g/L D-glucose and L-glutamine, and without Na+ pyruvate or Na+ bicarbonate), and Hams F12

Primary

Kidney Cells

I91

Fig. 1. Culture

of primary

proximal

kidney tubule cells.

medium supplemented with 15 mM HEPES buffer (pH 7.4) and sodium bicarbonate (brought to a final concentration of 20 mM) (DME/ F12). The medium used for tissue culture studies generally is not supplemented with penicillin and streptomycin (unless specifically indicated), since we have observed that theinclusion of these two antibiotics in the medium was inhibitory to primary proximal tubule cell growth. The water to be used for medium preparation is purified first with a Millipore Reverse Osmosis System, and then with a Milli-Q reagent grade water system as previously described (4,5). The medium is sterilized by pumping it through a 0.22 J.&I Millistak filter. 2. Sterile stock solutions of 5 mg/mL bovine insulin in O.OlN HCl, 5 mg/ mL human transferrin in 30, and 103M hydrocortisone in 100% ethanol are used. Insulin and transferrin stock solutions are sterilized using a 100 mL Nalgene filter apparatus (0.22 micron pore size), and distributed into sterile plastic tubes in I-mL aliquots. Insulin and hydrocortisone are kept at 4C for up to 2mo, whereas transferrin is kept frozen and thawed no more than four times. 3. Sterile 0.25% trypsin 1 mM EDTA in Phosphate Buffered Saline (PBS) (Grand Island Biological Corporation) is used in the trypsinization of proximal tubule cell cultures. 4. During the perfusion procedure, a 0.5% iron oxide solution (wt/vol) is used. The iron oxide solution is prepared as described by Cook and

192 Pickering (14). Sodium hydroxide (2.6 g) and (20 g) potassium nitrate are dissolved in 100 mL of oxygen-saturated water. Ferrous sulfate (9 g) is dissolved in 100 mL of oxygen-saturated water. These two loomL solutions are mixed, and the mixture is boiled for 20 min. The resulting precipitate of magnetic iron oxide is washed 5-10 times. In each wash, the precipitate is resuspended in water, and is then attracted to the bottom of the wash flask with a strong magnet. The wash water is then decanted. The iron oxide is resuspended in several liters of 0.9% NaCl. The concentrated iron oxide solution is then autoclaved for future use. Immediately prior to use, the iron oxide solution is diluted in PBS. Collagenase IV. Prior to its routine use, a sample of a particular lot of collagenase is first tested to evaluate if it may have deleterious effects on the outgrowth of cells from the proximal tubules. A stock solution of 10 mg/mL soybean trypsin inhibitor in PBS is filter sterilized and stored frozen for future use. Kidneys are obtained from male New Zealand white rabbits (2-2.5 kg). The kidney is prepared and sliced for tissue culture using a sterile 50-mL beaker, a sterile curved nose scissors, 2-3 sterile forceps, and 2-3 sterile glass loo-mm diameter Petri dishes. Heavy black suture thread and a Kelly hemostat (5-l/2 in straight end) are used in suturing the renal artery over a sterile 20-gage needle (blunt ended, using a file). The needle is attached to a 50-mL glass syringe containing perfusate. A sterile 15 mL Dounce tissue homogenizer (Type A pestle) is used for homogenization of the renal cortical slices. The preparation of proximal tubules for tissue culture also entails the use of Nylon nitrex screening fabric both 253 m and 85 m (TETCO, Inc., Depew,NY 14043), a lOOO-mL beaker, a metal spatula, and magnetic stirring bars. The nylon mesh is held in place in embroidery hoops and sterilized either in a container filled with 95% ethanol or in an autoclave.

5.

6. 7.

8.

3. Method
1. Insulin (5 pg/mL), transferrin (5 pg/mL), and hydrocortisone (5 x lOaM) are added to the culture medium (DME/F12) on the day of preparation of proximal tubules for tissue culture. 2. A rabbit is killed by cervical dislocation. The kidneys are removed (renal artery and vein intact) using sterile scissors, and placed in a sterile 50-mL beaker with ice-cold DME/F12 containing 192 IU/mL penicil-

Primary Kidney Cells

lin and 200 pg/mL streptomycin. The kidney is kept ice cold throughout the procedure. The kidney is perfused as follows. The kidney is placed in a 1OOLmm diameter glass Petri dish containing ice-cold DME/Fl2 with penicillin and streptomycin, and washed (using thesame medium). A sterile needle is inserted into the renal artery8 which is sutured. The kidney is thenperfused,firstwith DME/F12 (toremoveblood), andthenwith iron oxide (the kidney turns gray-black in color). The renal capsule is removed using sterile forceps, and the kidney is immediately transferred into another sterile 100-mm glass Petri dish containing ice-cold DME/F12. (After this point, greater care must be taken with regard to sterility.) Renal cortex slices are then removed from the kidney using a sterile, curved nosed scissors. The nephron segments are separated using the nylon mesh sieves. Toward this end, the 85 pm sieve is placed directlyon top of a sterile l,OOO-mL beaker; then the 253-p sieve is placed over the&S-p sieve. The sieves are washed with DME/Fl2 containing 192 IU/mL penicillin and 200 pg/mL streptomycin. The tubules and glomeruli are removed from the top of the 85-pm sieve using a sterile metal spatula and transferred into a sterile 50-mL plastic tube containingDME$F12. In order to remove the glomeruli in the tubule suspension, a sterile magnetic stir bar is placed into the tube. Glomeruli (covered with iron oxide) adhere to the stir bar. The stir bar is carefully removed. This process is repeated. First soybean trypsin inhibitor, and then collagenase are added to the tubule suspension (the final concentration of both reagents is 0.050 mg/mL). The tubules are incubated with the collagenase for 2 min at 23OC. In order to stop the collagenase treatment, the tube containing the tubules is placed in a desktop centrifuge and spun for 5 min. The tubules are resuspended in DME/F12 and again washed by certtrifugation. Empirically, we have observed that the use of collagenase is required if the outgrowth of epithelial cells from the tubules is to occur in vitro. After collagenase treatment, the tubules are suspended in DME/F12 supplemented with 5 pg/mL insulin, 5 l.tg/mL transferrin, and 5 x lOaM hydrocortisone but lacking antibiotics. We have used up to 400 mL of the medium to suspend proximal tubules obtained from a single rabbit kidney. The tubule suspension is then inoculated into 35-mm diameter tissue culture dishes (2 ml/dish). The culture dishes are placed in a37C incubator with a humidified 5% CO,/95% air environment.

3.

4.

5.

6.

7.

8.

9.

194

Taub

10. The medium is changed the day after plating the tubules, and routine ly every 3-4 d thereafter. Confluent monolayers can be obtained after 5-6 d in culture. 11. Primary proximal tubule monolayers on plastic dishes can readily be subcultured by trypsin treatment on a 1:2 basis. Confluent monolayers will again be obtained if care is taken to minimize the trypsinization period. Trypsin action is stopped by the addition of an equimolar concentration of soybean trypsininhibitor. The cells are then removed from the dish, suspended in a tube containing DME/F12, and spun in a desktop centrifuge at 400 rpm. The cells are resuspended in DME/ F12 containing the three growth supplements, and inoculated into plastic dishes. The proximal tubule cells can also achieve confluency following a second subculturing into plastic dishes.

4. Notes
If the cell cultures do not grow to confluency, a number of problems may have occurred. 1. Enough tubules must be added to the culture dishes to obtain confluent monolayers (0.5 mg protein/ml tubule suspension). The tubules may easily be lost if they are not carefully harvested at the end of the sieving procedure. 2. Cell cultures obtained from kidneys of young adults (as opposed to older animals) are the most successful. 3. A third point of concern is the tissue culture medium. The purity of the water is critical in defined medium studies. Loss of purity because of contamination from a dirty pH probe for example may result in medium that does not support cell growth. Our laboratory determines pH using samples of the medium rather than placing the probe in the medium to be used for tissue culture. In addition, a set of glassware is used in medium preparation that is specifically designated for that purpose. 4. Another point of caution with regard to medium is the hormone supplements. Improper preparation or storage of the growth supplements may be deleterious to cell growth. The growth stimulatory effect of insulin may be lost if the stock solution is frozen. Furthermore, the medium supplements should be added to the medium immediately prior to its use for tissue culture, since (unlike the case with serum-supplemented medium) their stability in the medium is uncertain.

Primary

Kidney Cells

195

5. Care should be taken that the incubator hold a constant 37C temperature and a 5% CO,/95% air humidified environment. Animal cell growth in the absence of serum is more sensitive to shifts in temperature and changes in the medium pH than with serum. The addition of HEPES buffer to the medium alleviates this latter problem. Finally, the primary rabbit kidney proximal tubule cells are less adherent to plastic dishes than many other cell types. Thus, the cells may detach during their manipulation for cell growth studies or transport studies. This problem may be alleviated by growing the cultures on tissue culture dishes coated with laminin, or with laminin and type IV collagen (25). Proximal tubule cell cultures on plastic cell culture dishes can readily be passaged once. Additional subculturings can be obtained, but confluent monolayers are obtained with greater difficulty. However, proximal tubule cell cultures grow more rapidly, achieve a higher saturation density, and a higher passage number when passaged into laminin-coated dishes.

References
1. Taub, M., Chuman, L., Saier, M. H., and Sato, G. (1979) Growth of Madin Darby canine kidney epithelial cell (MDCK) line in hormone-supplemented serum-free medium. Proc. Natl. Acad. Sci. USA, 76,3338-3342. 2.

Chuman, L., Fine, L. G., Cohen, A. I., and Saier, M. H. (1982) Continuous growth of proximal tubular epithelial cells in hormone-supplemented serum-free medium. J,

Cell. Biol. 94,506-510. 3. Taub, M. and Sato, G. (1979) Growth of functional primary cultures of kidney epithelial cells in defined medium. J. Cell. Physiol. 105,369-378. 4. Chung, S. D., Alavi, N., Livingston, D., Hiller, S.,and Taub, M. (1982) Characteriza-

tion of primary rabbit kidney cultures that express proximal tubule functions in a hormonally defined medium. J. Cell. Biol. 95,118-126. 5. Taub, M. (1985) Primary culture of proximal tubule cells in defined medium J. Tissue Cult. Methods 9,67-72. 6. Waqar, M. A., Seto,J.,Chung, S.D., Hiller-Grohol, S.,andTaub, M. (1985) Phosphate 7. 8. 9. 10.

uptake by primary renal proximal tubule cells grown in hormonally defined medium. J. Cell. Physiol. X4,411-423. Brendel, K. and Meezen, E. (1975) Isolation and properties of a pure preparation of proximal kidney tubules obtained without collagenase treatment. Fed. Proc. 34,803. Meezan, E. K., Brendel, J., Ulreich, J., and Carlson, E. C. (1973) Properties of a pure metabolicallyactiveglomerularpreparationfromratkidneys. I. Isolation. J.Phamacol. Exp. Ther. 187,332-341. Sakhrani, L. M., Badie-Dezfooly, B., Trizna, W., Mikhail, N., Lowe, A. C., Taub, M., and Fine, L. G. (1987) Transport and metabolism of glucose by renal proximal tubular cells in primary culture. Amer. 1. PhysioI. 246, F757-F764. Matsuo, S.,Fukatsu, A., Taub, M. L., Caldwell, P. R. B., Brentjens, J. R., and Andres, G. (1987) Nephrotoxic glomerulonepritis induced in the rabbit by antiendothelial antibodies. j. C&z. Invest. 79,1798-1811.

196
11. 12. ii: 15. 26.

Taub
Yang, I. S., Gokiinger,J M, Hong, S. K, and Taub, M. (1988) Preparation of basolatera1 membranes that transport paminohippurate from primary cultures of rabbit kidney proximal tubule cells J. CeU. Physiol. 135,481-487. Fine, I.. GandSakhmni, L. M. (19861 Proximal tubular cells in primary culture. Mineral Eke. Met& 12,X-57. Wang, Y. and Taub, M. (1987) Unpublished observations. Cook, W. F. and Picketi~ G. W. (1958) A rapid method for separating glomeruli from rabbit kidney. N&E @, 1103,1104. Taub, M. and Wang, Y. (1987) Control of rabbit kidney proximal tubule cell growth andfunction~~~~u~~matrixcomponents,inBiologyofG~owfhFucto~,Triennisi Symposium, program and abstracts, U. Toronto, Ontario, Canada. Aleo, M. D, Taub# &!I. L, O&song J R., Nickerson, P. A., and Kostyniak, P. J. (1987) Primary culturesof rabbit renal proximal tubule cells as an in vitro model of nephrotiti&yz &e&of 2 mercurials, in In Viti Totiolagy: Approaches to Validation, vol. 5 G3Aibew Aim M.., ed3, Liebert, New York, pp. 211-225.

Chapter 18
Adipocytes
Shinobu Gamou, Yoshiko Shim&u, and Nobuyoshi Shim&u
1. Introduction
There are several preadipocyte cell lines reported, but in this chapter, we will describe mainly the 3T3-Ll cells, which are best characterized and widely used for molecular biological studies. The 3T3-Ll cells were clonally isolated by Green and Kehinde from 3T3-Swiss albino. When these cells are growing exponentially, they appear indistinguishable from their parental Swiss/3T3 cells. 3T3-Ll cells, however, undergo a differentiation to adipocytes when they enter a confluent and contact-inhibited resting stage (2). Many clones isolated from the original 3T3 stock are able to convert to adipocytes, in most cases, with a much lower frequency than that of Ll cells. Subclones can be generated by serial cloning, which differentiate to adipocytes at a high frequency. 3T3-F422A is such a clone isolated from the same 3T3-Swiss albino stock culture as Ll cells (2). A number of morphological, biochemical, and genetic changes occur during the differentiation of 3T3-Ll cells to adipocytes. Ll cells start to accumulate triglyceride in their cytoplasm, which can be seen as Oil Red 0 staining droplets as they change from a spindle-like fibroblastic cell to a spherical cell (3). Thus, the differentiation can be easily identified under the microscope. The differentiation is also accompanied by a large increase in the enzyme activities involved in de ~OVO fatty acid and triglyceride synthesis

197

198

Gamou, Shimizu, and Shimizu

such as glycerol-3-phosphate dehydrogenase (4). The differentiated cells acquire a sensitivity to physiological concentrations of insulin via an increased number of insulin receptors that have a higher binding capacity than those in preadipocyte cells (5). Although the differentiation of Ll cells can occur spontaneously over a period of 2-4 wk following confluency, the differentiation can be synchronously induced at a high frequency by treating growth-arrested Ll cells with dexamethasone (DEX) and l-methyl-3-isobutylxanthine (MIX) for 2 d, followed by incubation in normal medium for 2-5 d (6). The combinations of DEX and MIX with insulin and biotin can accelerate the differentiation. Growth arrest at a specific stage of the cell cycle plays an important role in the induction of the adipocyte differentiation, and various factors appear to be involved in this process (7). The relationship between cell division in growth-arrested Ll cells and their initiation of differentiation can be closely examined under serum-free, hormone-supplemented conditions (8). We can observe the complex differentiation phenomenon under the microscope within a week by rather simple induction procedures. Thus, Ll cells differentiation is useful as a model of in vivo differentiation. The differentiated adipocytes are also useful as a model hormone-responsive cell sys tern.

2. Materials
2.1. Culture Medium
1. Medium: DMEM/FCSlO. Dulbeccos Modified Eagles Medium with a high concentration of glucose supplemented with fetal calf serum (FCS) at 10% and kanamycin (100 pg/mL). FCS concentration can be reduced to 5% by replacing with heat-inactivated calf serum. 2. PBS(-): Ca2+-and Mg2+-free Dulbeccos phosphate buffered saline (PBS; 2.7 mM KCl, 1.5 mM KH.J?O, 8.1 mA4 Na,HPO,, 137 mM NaCl). 3. Trypsin/EDTA solution: 0.1% trypsin and 0.01% EDTA in PBS(-).

2.2. Inducers

for Adipocyte

Differentiation

1. Dexamethasone (DEX) 1,000 x stock solution: 0.25 mM (98 pg/mL). Dissolve in DMSO and store at -2OOC in small aliquots. 2. l-methyl-3-isobutylxanthine (MIX) 1,000 x stock solution: 0.5M (111 mg/mL). Dissolve in DMSO and store at -2OOC in small aliquots.

Adipocytes

199

3. Insulin 1,000x stock solution: 10 mg/mL. Dissolve in O.OlNHCl, sterilize by filtration through a sterile Millipore filter (0.22 pm> and store at -2OOC. 4. Biotin 1,000 x stock solution: 100 pg/mL. Dissolve in distilled water, sterilize by filtration through a sterile Millipore filter (0.22 pm) and store at -2OOC.

2.3. Oil Red 0 Staining

1. PBS(+): Dulbeccos PBS with Ca2+(0.9 mM) and Mg2 (0.5 mM). 2. 10% Formalin/PBS(+): Mixture of 10 mL formalin solution and 90 mL of PBS(+). 3. Oil Red 0 solution: Dissolve Oil Red 0 (700 mg) in isopropanol(200 mL), stir at 4OCovernight, and then filter through Whatman 3 MM paper. Mix 3 volumes of Oil Red 0 in isopropanol with 2 volumes of distilled water, allow to stand at 4C overnight and filter through Whatman 3 MM. Store at 4OC. 4. 50% (v/v) glycerol solution. 5. Mayers Hematoxylin solution.

2.4. Detection of Glycerot3-Phosphate Dehydrogenase Activity

1. Extraction buffer: 0.5% lYiton X-100,10 mM Tris-HCl, pH 7.5,l mM MgCl,, 20 mM KCl, 0.2 mM Dithiothreitol, 10% Glycerol. 2. Reaction mixture: 125 mM Triethanolamine-HCl, pH 7.5, 2.5 mM EDTA, 0.16 mMNADH, 0.36 mM Dihydroxyacetonephosphate,0.125 mM P-mercaptoethanol. 3. Glycerol-3-phosphate dehydrogenase from rabbit muscle.

2.5. Insulin

Binding

Assay

1. Insulin binding buffer: 120 mM NaCl, 1.2 mM MgSO,, 5 mM KCl, 10 mM Glucose, 1 mM EDTA, 15 mM Sodium acetate, 50 mM HepesNaOH, pH 7.8,1% Bovine serum albumin. Store at -2OOC. 2. Cold insulin solution: Dilute the insulin stock solution described above with the insulin binding buffer at a final concentration of 1 pg/ mL and 10 pg/mL. Make fresh before use. 3. ?insulin: Insulin iodination can be carried out as described in Chapter 36, Vol. 1 of this series. Some technical skill is required to iodinate insulin at a high specific activity without compromising its biological activity. Thus, lZI-insulin obtained from commercial sources

200

Gamou, Shim&

and Shimizu

may be more convenient. Approximately 50,000 cpm/mL (-1 ng/ mL) of y251-insulin in insulin binding buffer is used for binding assay.

3. Methods
3.1. Maintenance of 3T3-Ll Cells

1. Ll cells can be purchased from American Type Culture Collection (12301 Parklawn Drive, Rockville, MD 20852-1776, USA). Ll cells are routinely maintained in DMEM/FCSlO at 37C in 5% CO, and 100% humidity. Ll cells actively grow with a population doubling time of 20-30 h before becoming contact-inhibited at a saturation density of -5 x lo5 tells/3-cm dish. 2. The morphology of the parental 3T3-Swiss albino is quite uniform and shows typical 3T3 shape at confluency. Ll cells, however, show rather heterogeneous morphology, occasionally overlapping each other at confluency, but not piling up like transformed cells. 3. Contact-inhibited Ll cells will differentiate into the adipocytes spontaneously after 2-3 w when refed regularly. 4. Ll cells are easily detached after washing 2-3 times with PBS(-) and treating with trypsin/EDTA at 37C. Transfer cells every4-5 d by 1:20 dilution to maintain the preadipocyte state.

3.2. Induction

of AcZipocyte Differentiation

1. The induction medium contains DEX (0.25 w), MIX (0.5 mM), insulin (10 pg/mL) and biotin (100 ng/mL) in DMEM/FCSlO. Add l/ 1000 vol of DEX and MIX stock solution into DMEM/FCSlO, mix immediately and vigorously, and then add insulin and biotin at the indicated concentration. Prepare the induction medium fresh before use. 2. Replace the medium with induction medium when Ll cells reach confluency, and incubate for 48 h at 37OC. 3. During the 48-h induction incubation, Ll cells start growing to approximately double in cell number, and the medium becomes slightly viscous because of triglyceride secretion. 4. Replace the induction medium with DMEM/FCSlO containing biotin (100 ng/mL) and insulin (10 pg/mL), and incubate for 5 d to allow differentiation into mature adipocytes.

Adipocytes

201

3.3. Oil Red 0 Staining to Identifjt the Morphological Differentiation

1. After the five days of incubation, the medium becomes viscous and spherical adipocytes are seen under the microscope (Fig. 1). The adipocytes often appear in clusters of varying sizes. 2. Wash the differentiated cultures twice with PBS(+), and then fix with 10% formalin/PBS(+) for 30 min. Stain the fixed culture dishes with Oil Red 0 solution for 10 min, wash under tap water for a few minutes, and then cover with 50% glycerol. The adipocyte colonies can be readily identified. 3. For microscopic observation or photography, the cells can be stained with Mayers Hematoxylin for 2-3 min, washed for about 10 min under tap water, and covered with 50% glycerol. 4. To determine the differentiation rate, wash the adipocytes and undifferentiated cells 2-3 times with PBS(-) and detach them by trypsin/ EDTA treatment at 37OC. 5. Resuspend the cells with an appropriate volume of DMEM/FCSlO, and count the number of adipocyte and undifferentiated cells in the hemocytometer. The adipocytes are identified by the presence of large lipid droplets, 6. When the lipid droplets are not large enough to identify under the microscope, add half a volume of Oil Red 0 solution to the cell suspension and incubate for a minute at room temperature. The adipocytes can be easily identified by the presence of red droplets in their cytoplasm.

3.4. Detection of Glycerol-S-Phosphate Dehydrogenase Activity as a Biochemical Differentiation Marker

1. Ll cells grown in 6-cm dish are washed twice with ice-cold PBS(-) and scraped with a rubber policeman. Transfer the Ll cells to a 1.5-mL centrifuge tube, and wash the cells twice with ice-cold PBS(-) by brief centrifugation. 2. Add 100 PLof ice-cold extraction buffer, and incubate for 30 min at 4OC with gentle and continuous agitation. Centrifuge at 16,000 rpm for 60 min at 4C in a Sorval SS34 rotor with a 1.5-mL tube adapter. 3. The resulting supernatant is kept frozen at -7OOC in small aliquots until use. The amount of protein is determined by Lowrys method as described in Chapter 1, Vol. 1 of this series.

202

Gamou, Shimizu, and Shimizu

Fig. 1. Morphology of 3T3-Ll cells. (A) 3T3-Ll cells at high density. 09 3T3-Ll cells after adipocyte differentiation. The cells were treated for 2 d with inducers (DEX, MDC, and insulin) and allowed to differentiate for 5 d.

4. Put 0.4 mL of reaction mixture into a narrow width cuvette and set it in a spectrophotometer. 5. Dilute the cell lysate to 100 PL (W-200 pg protein) with the extraction buffer. Add 100 PL of diluted cell extract into the cuvette and mix rapidly with the reaction mixture to initiate the reaction.

Adipocytes 6. Record the absorbance at 340 nm 30 s after initiation and at 30-s intervals. The absorbance will decrease. 7. The reaction is linear with time (for several minutes) and cell extract concentration. Calculate the difference in absorbance/min/pg protein. 8. To determine the absolute enzyme activity, purified glycerol-3-phosphate dehydrogenase from rabbit muscle (Sigma) can be used as a standard. 9. The enzyme activity in adipocytes increased lO-loo-fold over the activity in preadipocytes. The increase in enzyme activity can be detected just after the induction incubation.

3.5. Insulin

Binding

Assay

I. Ll cells are grown in three 3-cm dishes. One of these is used to determine cell number and differentiation rate as described above. 2. Place the remaining two dishes on ice, and wash twice with ice-cold insulin-binding buffer. 3. Prepare 3 mL of *251-insulin solution, and save 0.5 mL to determine input radioactivity (cpm/mL) with a gamma counter (50,000 cpm/ mL, -1 ng/mL). 4. Add 1 mL of 1251-insulinsolution to one of the dishes, and incubate at 15OCfor 90 min with gentle agitation. 5. Wash the cells four times with ice-cold insulin binding buffer, and then dissolve in 1 mL of 1NNaOH by incubation at 37C for 1 h. Cellasso-ciated radioactivity (TB) is determined with a gamma counter. 6. To determine nonspecific binding (NSB), the remaining dish is preincubated with cold-insulin solution containing 1 pg/mL insulin at 4C for 30 min and then incubated with 1251-insulinsolution containing one-tenth the volume of the 10 pg/mL cold-insulin solution (final insulin cont. 1 pg/mL) at 15OCfor 90 min. Cell-associated radioactivity is determined as above. NSB is usually one-tenth to onefourth of TB, 7. Insulin specific binding (SB) is calculated as follow: SB = (TB-NSB)/ input/lo cells x 100%. Specific binding in the undifferentiated Ll culture ranges from l-3% of input/l06 cells, whereas that in Ll adipocytes is 5-15% depending on the differentiation rate. Insulin binding capacity gradually increases after the induction incubation. 8. Duplicate or triplicate dishes of Ll cells are preferable for the TB and NSB determinations.

204

Gamou, Shimizu,

and Shimizu

4. Notes
1. Although the differentiation potential of Ll cells is heritable, they have a tendency to lose this capacity (2,2). The cells that have differentiated at 90% frequency may lose half of their differentiation ability after lo-20 transfers. Thus, a series of experiments must be carried out using cells with same differentiation potential. Numerous frozen stocks should be stored as early as possible. 2. If all of the frozen cell stocks with high differentiation potential have been depleted, subclones may be isolated using the same method by which the original Ll cells were established (1,2). Isolate lo-20 areas in which the cells have small lipid droplets in their cytoplasm at confluency by the standard cloning method. Measure the differentiation ability and subclone if necessary. 3. The timing for induction is an important factor in order to achieve a high differentiation rate. Sparse cultures can be induced, resulting in a high frequency of adipocyte differentiation. Induction during late confluency appears to be less effective in achieving the high differentiation rate than that induced during early confluency. Ll cells must traverse the cell cycle once before expressing the adipocyte phenotype (8).

4. Incubation of more than five days for the expression of the adipocyte phenotype does not yield higher differentiation rates; however, the adipocytes become more fatted and anchorage independent with increased incubation time. 5. The differentiation rate may be measured quantitatively from the fixed, Oil Red 0 stained culture. Elute Oil Red 0 from the stained culture with isopropanol and measure the absorbance at 510 nm. Cultures containing no adipocytes occasionally bind some dye and give a high background. This, however, does not interfere the quantification of the differentiation rate. 6. When the cell lysates are analyzed by two-dimensional electrophoresis as described in Chapter 10, Vol. 1 of this series, an alteration in the biosynthesis of more than 60 species of differentiation-specific proteins is detected. This includes a decrease in the biosynthesic rates of p and y actins as well as a and p tubulins by 90% that reflect the changes in cell shape. These observations have been confirmed using cDNA probes (9). 7. The biochemical nature of the reaction of glycerol-3-phosphate dehydrogenase is: Dihydroxyaceton phosphate + NADI&&lycerol-3phosphate + NAD+. The decrease in the absorbance at 340 nm is the result of the conversion of NADH to NAD.

Adipocytes
8. Since glycerol-3-phosphate dehydrogenase activity is very low in predifferentiated Ll cells, the cell extract without dilution or the extract from more than two dishes may be required to measure the activity accurately. 9. An automated spectrophotometer with a recorder is preferable, but not necessary, for the measurement of the enzyme activity because the reaction will be linear 2-3 min after initiation. 10. cDNAs for glycerol-3-phosphate dehydrogenase and other major differentiation-related proteins have been cloned from cDNA libraries constructed from the differentiated 3T3-F422A cells (10) and Ll cells (11). 11. The increase in insulin binding capacity during adipocyte differentiation involves an increase in both receptor number and affinity. Thus, Scatchard plot analysis must be used to measure changes in insulin binding capacity during the differentiation process. Prepare 1251-insulin solutions with different specific activities (50,000 cpm/mL of *251-insulin with 1,2,4,10,20,40, and 100 ng/mL of cold-insulin, and 1,2,4,10,20,40, and 100 pg/mL of cold-insulin for nonspecific binding using duplicate or triplicate dishes for each specific activity). The 1251-insulinbinding assay is carried out as described above (3.5.). Calculate how much insulin is bound as follows: Sp. AC. = Input cpm/ [Input insulin concentration]. [Bound] = SB/Sp. AC. [Free] = [Input insulin concentration] - [Bound]. Plot [Bound]/ [Free] on the ordinate and [Bound] on the abscissa. The negative reciprocal of the slope indicates the affinity of the insulin receptor for insulin (Kd: dissociation constant). The intercept on the abscissa indicates the number for insulin receptor. Both predifferentiated Ll cells and adipocytes yield curvilinear plots, suggesting a heterogeneity in receptor affinity types. 12. Serum-free hormone-supplemented medium has been developed for 3T3-Ll cells (8,X?). The basal medium is a 3:l mixture of DMEM and Hams F12 supplemented with several amino acids, vitamins, and trace elements. Hormone-supplemented medium contains the following components in the basal medium: transferrin (2 pg/mL), insulin (10 pg/mL), epidermal growth factor (20 ng/mL), biotin (100 ng/mL), bovine serum albumin (O.l%, Fr. V), and ethanolamine (0.1 mM). Under these conditions, the cells commitment to adipocyte differentiation is separated from the expression of the adipocyte phenotype. 13. There are several drugs that can inhibit the differentiation through several distinct mechanisms. Nicotinamide possibly inhibits the differentiation through the inhibition of poly (ADP-ribose) synthetase (23). Dihydroteleocidin B (DHTB), a potent tumor promoter,

206

Gamou, Shimizu, and Shimizu

almost completely inhibits the differentiation whether added before, during, or after the inducer treatment. Similar inhibition is observed by 12-0-tetradecanoylphorbol-13-acetate (TPA), but with over 90% less efficiency (14). Thus, these inhibitions are useful for dissecting the complex differentiation process. 14. Using insulin-diphtheria toxin A fragment conjugate, variant cells with altered insulin receptors have been isolated from Ll cells (25). Also, DHTB-nonresponsive variants have been isolated by mitotic shake-off method (16). Using these variants, the mechanism of mitogenic signal transduction is being investigated. 15. Proto-oncogene c-fos expression in Ll cells and their derived adipocyte have been studied in response to a variety of growth-promoting agents (17). The intracellular location of protein kinase C has been also studied in the Ll cell during growth and after exposure to phorbol esters (18). Vesicles containing insulin-responsive glucose transporters have been isolated from Ll cells (29). 16. 3T3-F442A cells isolated from 3T3-Swiss albino are able to differentiate into adipocytes with high frequency by growth hormone treatment (20).

References
that accumulate lipid. Cell 1,113-l 16. 2. Green, H. and Kehinde, 0. (1976) Spontaneous heritable changes leading to increased adipose conversion in 3T3 cells. Cell 7,205-113. 3. Novikoff, A. B., Novikoff, I. M., Rosen, 0. M., and Rubin, C. S. (1980) Organelle relationships in cultured 3T3-Ll preadipocytes. J. Cell Bid. 87,180-196. 4. Wise, L. S. and Green, H. (1979) Participation of one isozyme of cytosolic glycerophosphate dehydrogenase in the adipose conversion of 3T3 cells. J. Biol. Chem. 254,

1. Green, H. and Kehinde, 0. (1974) Sublines of mouse 3T3 cells

273-275. 5. Rosen, 0. M., Smith, C. J.,Fung, C., and Rubin, C. S. (2978) Development

of hormone receptors and hormone responsiveness in vitro: Effect of prolonged insulin treatment on hexose uptake in 3T3-Ll adipocytes. J. Bid. Chem. 253,7579-7583. 6. Rubin, C. S., Hirsch, A., Fung, C., and Rosen, 0. M. (1978) Development of hormone receptors and hormonal responsiveness in vitro: Insulin receptors and insulin sensitivity in the preadipocyte and adipocyte forms of 3T3-Ll cells. J. Biol Chem. 253, of growth arrest and differentiation at a distinct state in the G, phase of the cell cycle: GD. Proc. Natl. Ad. Sci. USA 79,845-849. of 3T3-Ll cells in 8. Gamou, S. and Shimizu, N. (1986) Adipocyte differentiation serum-free hormone-supplemented media: Effects of insulin and dihydroteleocidin B. Cell Struct. Fun& 11,21-30.

7570-7578. 7. Scott, R. E., Florine, D. L., Wille, J. J., Jr., and Yun, K. (1982) Coupling

Adipocytes
9. Spiegelman, B. M. and Farmer, S. R. (1982) Decreases in tubulin and actin gene expression prior to morphological differentiation of 3T3 adipocytes. Cell 29,53-60. 10. Spiegelman, B. M., Frank, M., and Green, H. (1983) Molecular cloning of mRNA from 3T3 adipocytes: Regulation of mRNA content for glycerophosphate dehydrogenase and other differentiation-dependent proteins during adipocyte development. J. BioI. Chern.258,10083-10089. Il. Bernlohr, D. A., Angus, C. W., Lane, M. D., Bolanowski, M. A., and Kelly, T. J., Jr. (1984) Expression of specific mRNAs during adipose differentiation: Identification of an mRNA encoding a homologue of myelin P2 protein. Proc.N&Z. Acud. Sci. USA 81,5468-5472. 12. Shimizu, N. (1984) Use of hormone-toxin conjugates and serum-free media for the isolation and study of cell variants in hormone responses,inMethods for Serum-Free Culture ofEpitheliu1 and FibrobZustic CelZs.(Barnes, D. W., Sirubasku, D. A., and Sato, G. H., eds.), Alan R. Liss, New York, pp. 233-248. 13. Lewis, J. E., Shimizu, Y., and Shimizu, N. (1982) Nicotinamide inhibits adipocyte differentiation of 3T3-Ll cells. FEBS Lett. 146,37-41. 14. Shimizu, Y., Shimizu, N., Fujiki, H., and Sugimura, T. (1983) Distinct inhibitory effects of dihydroteleocidin B and the phorbol ester tumor promoters on the adipocyte differentiation of 3T3-Ll cells. Cancer Res. 43,49744979. 15. Shimizu, N., Miskimins, W. K., Gamou, S., and Shin&u, Y. (1982) A genetic approach to the mechanisms of insulin action: Use of the insulin-Diphtheria toxin fragment A conjugates for isolation of genetic variants. Cold Spring Harbor Conferences on Cell Proliferation 9,397-402. 16. Shimizu, Y., Fujiki, H., Sugimura, T., and Shimizu, N. (1986)Mouse 3T3-Ll cell var-

iants unable to respond to mitogenic stimulation of dihydroteleocidin B: Genetic evidence for the synergism of tumor promoters with growth factors. CancerRes. 46,
4027-4031. 17. Stumpo, D. J. and Blackshear, P. J. (1986) Insulin and growth factor effects on c-fos

expression in normal and protein kinase C-deficient 3T3-Ll fibroblasts and adipocytes. Proc. Nutl. Acud. Sci. USA 83,9453-9457. 18. Halsey, 0. L., Girard, I. R., Kuo, J, F., and Blacksher, P. J. (1987) Protein kinase C in fibroblasts: Characteristics of its intracellular location during growth and after exposure to phorbol esters and other mitogens. J. Biol. Chem. 262,2234-2243. 19. Biber, J. W. and Lienhard, G. E. (1986) Isolation of vesicles containing insulin responsive, intracellular glucose transporters from 3T3-Ll adipocytes. 1. Bid. Chem. 261,16180-16184. 20. Morikawa, M., Nixon, T., and Green, H. (1982) Growth hormone and the adipose conversion of 3T3 cells. Cell 29,783-789.

Chapter 19
Tissue Culture of Skeletal Muscle
Terry A. Partridge

1. Introduction
In its simple forms, no special difficulty attaches to the tissue culture of skeletal muscle. Indeed, it is one of the easiest tissues to culture in large amounts because the starting material, skeletal muscle, is plentiful and readily obtainable from a wide variety of species, including humans. Moreover, the stem cells from which muscle develops in tissue culture seem to be very resilient; they will survive anoxic conditions within muscle at temperatures between 4 and 37OC, for up to a few days, and when obtained in suspension, are easy to freeze (with 10% DMSO as a cryoprotectant) and store in liquid nitrogen. Mature muscle fibers, being sensitive to physical damage and anoxia, cannot be maintained in good functional condition out of the body for more than a few hours. In tissue culture, therefore, muscle fibers must be grown afresh from their mononucleate precursor cells (mpc). These latter cells may readily be obtained in large numbers from growing or regenerating muscle and, in smaller but still usable numbers from mature normal muscle. It is of biological interest, but not of great practical value, that the thymus is also a source of such cells (1). 209

210

Partridge

Here, the method we use to grow skeletal muscle from neonatal mouse and a method for partially separating myogenic from nonmyoge nit cells from this species are described in detail (2). The disaggregation method gives satisfactory results with fetal human (15 wk) and neonatal rabbit muscle. A simple method we have used for growing avian muscle in culture is also described, and the use of some mouse muscle cell lines is commented upon. Complications, problems, and refinements are indicated in the Notes.

2. Materials
1. All glassware is washed in 1.5% RBS(Chemical Concentrates), rinsed ten times in tap water and twice in double-glass distilled water, and then air-dried in a warm oven. 2. Muscle cells are usually grown in vented polystyrene tissue culture Petri dishes, 35 or 60 mm diameter. Where we need to make permanent microscopic preparations, immunostained preparations, or microscopic observations at high resolution on the cultures, we grow the cells on a sterile coverslip, placed in the Petri dish prior to addition of the cell suspension and removed when it is required. We routinely coat tissue culture surfaces with sterile rat-tail collagen in distilled water and allow it to air dry. This is not strictly necessary for the mixed cell cultures described here. 3. For filtration of the single-cell suspension, three layers of 45 pm pore size nylon cloth are placed in sterile Swinnex filter holders. 4. Phosphate Buffered Saline (PBS): 8.0 g NaCl, 0.2 g KCl, 1.12 g Na,Hl?O,, and 0.2 g KHJ?O, in 1 L distilled water, adjusted to pH 7.2 with NaOH. The solution is dispensed in lOO-mL aliquots and sterilized by autoclaving. 5. Antibiotic solution: 100 mL sterile PBS, 10 mL Fungizone (amphotercin B, 250 pg/mL), and 2 mL penicillin/streptomycin (I? 5000 IU/mL, S 5000 yg/mL). Can be stored at -2OOC. 6. Hepes-buffered calcium and magnesium-free Hanks balanced salt solution, pH7.4 (Ca2+and Me-free BSS). Purchased as a 10 x solution and diluted l/10 with double-glass distilled water. This solution is buffered by the addition of 1M Hepes solution to give a final concentration of 25 mM Hepes, sterilized by filtration through a 0.45 pm Millipore filter and stored at -2OOC. 7. Trypsin stock solution: Trypsin (1%) (1:250; Difco) in Ca2+-and Mg2+free BSS. Trypsin and BSSare mixed for 2 h at room temperature and filtered through a Whatman No. 1 filter paper. The pH is adjusted to

Tissue Culture of Skeletal Muscle

211

8.

9.

10. 11,

7.4 with lNNaOH, sterilized by filtration through a 0.45-pm Millipore filter, and stored at -20C in 20-mL aliquots. Pangestin stock solution: Pangestin (1:75; Difco) mixed in a 0.5% ratio (w/v) in Ca2+-and Mg2+-free BSSfor 2 h at 37OC,centrifuged at 26008 for 15 min at 4OC,and filtered through a Whatman No. 1 filter paper. The pH is adjusted to 7.4 with 1NNaOH. The solution is sterilized by filtration through a 0.45-pm Millipore filter and stored in 20-mL aliquots at -2OOC. Enzyme solution: On the day prior to use, 60 mL of trypsin stock solution, 20 mL pangestin stock solution, and 20 mL Ca2+-and Mg2+free BSSare mixed under aseptic conditions. The solution is dispensed in 5-mL aliquots in sterile universal bottles, stored at 4OCovernight and brought to room temperature immediately before use. Inhibition medium: 100 mL Hanks balanced salt solution + 20 mL Fetal Calf Serum (FCS) and 2 mL penicillin/streptomycin (I?5000 IU/ mL, S 5000 pg/mL). Stored at 4OCuntil use. Culture media: Basic medium constituents: Eagles Minimum Essential Medium (EMEM), Dulbeccos Minimum Essential Medium (DMEM), Fetal Calf Serum (FCS), Horse Serum (HS), Chick Embryo Extract (CEE), Penicillin and Streptomycin stock solutions, and Fungizone solution are all obtainable from a number of commercial sources (seeNote 1). We make our own CEE by a standard method (3). Fertile chicken eggs, 10-12 d incubation, are held blunt end uppermost, swabbed with 70% ethanol, and given a sharp tap, delivered horizontally, with a sterile scalpel handle at about the lower edge of the air space (this can be seen by holding the egg against a bright light). The cap of the shell and adherent shell membranes are then carefully lifted off with sterile coarse forceps. One prong of a second pair of sterile coarse forceps is plunged through the embryonic membranes and beneath the neck of the embryonic chick, which can then be lifted carefully from the egg and transferred to a sterile beaker. When all the embryos have been removed, they are washed free of all traces of blood and yolk with several changes of BSS.These embryos are then crudely homogenized by removing the plunger from a 20-mL syringe, loading the embryos, a few at a time, into the barrel, carefully reinserting the plunger, and pushing it home to express the contents, via the needle fitting, into sterile plastic universal containers. To the mashed embryo in each container an equal quantity of Hanks BSSis added, and the two are mixed thoroughly by stirring with a sterile glass rod. After 30 min incubation at room temperature, the containers are centrifuged at

212

Partridge

20008 for 20 min and the supernatant CEE is stored in small aliquots at -20C. On thawing, the CEE is recentrifuged before use. 12. Media for mouse primary cultures: 100 mL Dulbeccos Minimum Essential Medium (DMEM), 10 mL FCS, 2 mL CEE. 13. Cell lines and their growth media: a. GB-1: For growth: 100 mL DMEM, 10 mL FCS. For differentiation: 100 mL DMEM, 2% HS. b. C2: For growth: 100 mL DMEM, 20 mL FCS. For differentiation: 100 mL DMEM, 2% H S. c. MM14: 50 mL DMEM, 50 mL Hams FlO, 15 mL HS, 3 mL CEE. d. Human primary cultures: 100 mL DMEM, 20 mL FCS. e. Avian cells: Following the methods described by Konigsberg (4), we have used high growth media consisting of: For chick, 79 mL Eagles MEM (with Earles salts), 15 mL HS, 5 mL CEE, 1 mL Penicillin/Streptomycin, 0.25 mL Fungizone. For Quail, 74mL Eagles MEM (with Earles salts), 15 mL HS, 10 mL CEE, 1 mL Penicillin/Streptomycin, 0.25 mL Fungizone. Primary cultures, given one change of high growth medium on the day after they were set up, and left in this medium, initially proliferate and subsequently, as they exhaust the nutrients, differentiate to form large numbers of myotubes. However, Konigsberg (4) recommends reducing the concentration of CEE to produce a low growth medium that, on addition to the culture, limits proliferation and encourages differentiation. 14. Culture substrata: a. Collagen: This is made by a standard method (3). Rat tails are sterilized by steeping them in 95% ethanol for a few hours. Each tail is then broken into short segments, starting at the tip. To do this, the tail is grasped crossways at the distal end with two pairs of sterile Spencer-Wells forceps, held side by side, the tip is snapped off by a sharp rotation of the more distal of the two forceps, and carefully drawn away to pull out the long white tendons that insert into the tip of the tail. These are then cut off with a pair of sterile scissors and allowed to fall into a beaker containing sterile 1% acetic acid in distilled water (150 mL/ tail). This process is repeated, working progressively toward the base of the tail. When all of the accessible tendon has been removed, the beaker is covered to maintain sterility and kept for a few days at 4OC, after which the solution is centrifuged for 1 h at 25008. The supernatant is then dialyzed, in autoclave-sterilized dialysis tubing, against daily changes of distilled water until it begins to become viscous. This solution

Tissue Culture of Skeletal Muscle

is removed, after sterilizing theoutside of the dialysis tube with a wash of 70% ethanol, and can be stored for several months in aliquots under sterile conditions at 4OC. Should the solution become too gelatinous, its viscosity can be reduced by addition of a little sterile dilute acetic acid. A drop of collagen solution is placed on the culture surface (glass or plastic), spread thinly with a glass spreader (made by using a fine flame to seal the tip of a Pasteur pipet and to bend the terminal cm or so at right angles to themain stem), and left to air-dry in asterile flow hood. Such collagen coated surfaces can be stored dry for several weeks. b. Gelatin: A 0.01% solution of gelatin (Porcine skin, Type I) in distilled water is distributed into I-2-mL aliquots and sterilized by autoclaving. A drop of the solution is placed on the culture surface, spread, and air-dried as for collagen (above).

213

3. Methods 3.1. Tissue Culture of Cells thorn Neonatal Mouse Muscle

1. Newborn mice, (l-2 d, seeNote 2) are killed by decapitation, the tail and paws are cut off, and the body washed in 70% ethanol, then soaked in antibiotic solution. By slitting the skin along the length of the spine, it can be peeled off in one operation by means of two pairs of coarse forceps: one pair is used to grasp the mouse by the exposed end of the spine at the neck and the second pair to pull the free edge of the skin at the sternum back toward the tail. 2. The skinned carcass is then placed in a sterile Petri dish where as much as possible of the skeletal muscle is removed from the limbs and trunk, with the aid of watchmakers forceps and microscissors, and placed in a second sterile Petri dish where it is kept damp with a drop of BSS. Care must be taken not to include fragments of bone or of the brown fat that is present in large amounts in these young animals, especially between the shoulder blades. It is also important, particularly if the animal has suckled, not to puncture the gut. Depending on the age and size of the mice and the skill of the operator, approximately 0.5 g of muscle can be obtained in this way from 3-5 mice in about 2 h: this is the optimal amount to be taken through the remainder of the procedure by one person. 3. Themuscle is minced into small fragments with a pair of sharp, curved fine scissors (if well dampened with saline, the fragments tend to be kept together by surface tension) and divided into ten equal portions,

214

Partridge

4.

5. 6,

7. 8. 9.

10.

each of which is placed in 5 mL of enzyme solution in a sterile plastic universal bottle. These are placed in a 37C waterbath and shaken at 2 Hz for 10 min. Each aliquot is gently sucked into and expelled from a Pasteur pipet continuously for one min to disperse the cells. Remaining muscle fragments are allowed to settle, and the supernatant containing the released cells is decanted into 5 mL of inhibition medium. Fresh enzyme solution (5 mL) is then added to the remaining muscle fragments in each universal bottle and the above disaggregation procedure repeated. The inhibition medium containing all the released cells from both incubations is poured into a large, sterile plastic syringe fitted with the Millepore filter holder containing the 45-pm nylon cloth, and slowly expressed, thus removing any remaining clumps of cells and muscle fragments to give a suspension consisting almost entirely of separated mononucleate cells. This is centrifuged at 3508 for 10 min at 4OC, and the resultant cell pellets are resuspended in a known volume of growth medium. A 100~PLsample of this suspension is added to 50 u.Lof a 0.5% solution of Trypan Blue in PBS, and a count of viable cells (i.e., those that exclude the dye) is made in a hemocytometer. For culture in 35-mm plastic Petri dishes, cells are suspended in culture medium at concentrations of between lo4 - 2 x 105cells/mL, and plated out at 2 mL of cell suspension/dish. It is common to coat dishes with collagen or gelatin. This is important when cloning myogenic cells, but with cultures prepared as above, we can detect no advantage in doing so. The Petri dishes are placed in a 37OCincubator in an atmosphere of 10% CO, and examined daily under an inverted microscope fitted with long-working distance phase contrast objectives. By 5-7 d, numerous myotubes will have begun to form (Fig. 1, seeNote 3). If longer term cultures are required, the medium must be replaced by one containing less FCS and no CEE, so as to discourage overgrowth of the culture by nonmyogenic cells.

3.2. Partial Separation of Myogenic and Nonmyogenic Cells on Discontinuous PercoZZ Gradients
Cell suspensions obtained by enzymatic disaggregation of neonatal mouse muscle contain a majority of nonmyogenic cells that in long-term cultures, tend to overgrow the developing muscle fibers. The method

Tissue Culture of Skeletal Muscle

215

Fig. 1. 7-d primary culture of cells obtained by enzymic disaggregation of neonatal mouse muscle, stained with hematoxylin and eosin, 100x. Note small multinucleated myotubes and the numerous unfused cells. Of the latter, some of the small spindleshaped cells are probably myoblasts.

described below, modified from that described by Turner (5) for the separation of chick muscle cells on Ficoll gradients, can be used to obtain a degree of separation of myogenic from nonmyogenic populations in mouse muscle cell suspensions (seeNote 3 and 4). 1. Percoll dilutions of 26% and 34% are made up in sterile HEPESbuffered Ca2+- M g%?ee BSS, one day before use, stored at 4OC overnight, and brought to room temperature immediately before use. 2. Cells obtained from the disaggregation procedure are suspended in Ca2+- Mg!+-free BSS and placed in 5-mL aliquots in sterile Universal bottles. 3. The 26% Percoll solution is taken up into a syringe equipped with a 21gage needle, and layered beneath the cell suspension by placing the tip of the needle on the bottom of the Universal bottle and gently expelling 5 mL of the solution. In the same way, 5 mL of the 34% Percoll solution is layered beneath the 26% layer. 4. The gradients are then centrifuged at 3508 for 15-20 min at 15OC. Cells that accumulate as cloudy layers at the interfaces and the cell pellet at the tip of the bottle are then carefully removed with a Pasteur pipet. 5. Each fraction is placed in 5 mL of inhibition medium, pelleted at 3508, and suspended in growth medium. Samples of each fraction are

216

Partridge

Fig. 2. Cells obtained from the pellet beneath 34% Percoll after separation of cells obtained from neonatal mouse muscle on a discontinuous density gradient, stained with hematoxylin and eosin. (a) After4 d in culture. Note the high proportion of small strongly stained rounded or bipolar cells, characterized by a high nuclear/cytoplasmic ratio, 250x. (b) Cells fusing after 7 d in culture. Note the conspicuous myotubes. In addition, the majority of nonfused cells have a myoblastic appearance, 100x.

counted, and the remainder plated out in 35-mm plastic Petri dishes and cultured as before. The pellet from the tip of the tube contains a high proportion of small rounded or bipolar spindle-shaped basophilic cells (Fig. 2a). 6. After 5-7 d in culture, large numbers of myotubes can be seen to form by fusion of such cells together (Fig. 2b). Cells obtained from the 26/ 34% Percoll interface do not include the small basophilic type and, on culture, do not produce myotubes (Fig. 3a,b). Very few cells can be harvested from the BSS/26% Percoll interface, but they do include some small basophilic cells and do form a few myotubes when cultured. How the low buoyant density and high buoyant density myogenic cells differ from one another is uncertain; they may represent cells in G, and G, stages, respectively, of the cell cycle.

3.3. Tissue Culture

of Avian

Muscle

Chick and quail embryonic muscles are among the easiest to prepare and culture. No sterilization of the carcass is required. At 9 d for the quail and 11 d for the chick, the muscles can be disaggregated enzymically as described for the mouse, but are also sufficiently soft that the mpc can be released by mechanical disruption of the tissue. Further advantages of the avian muscle culture systems, especially quail, are that a greater proportion of the cells differentiate into myotubes and that there is less experiment to experiment variation than with mouse muscle. For both chick and

Tissue Culture of Skeletal Muscle

217

Fig. 3. Cells obtained from the 26/34% Percoll interface after separation of cells obtained from neonatal mouse muscle on a discontinuous density gradient, stained with hematoxylin and eosin. (a) After 4 d in culture. The majority of cells are large and of fibroblastic appearance, haying a low nuclear/cytoplasmic ratio and pale-staining cytoplasm, 250x. (b) Cells after 7 d in culture. Note absence of myotubes. Only the occasional cell has the bipolar spindle-shape characteristic of myoblasts, 100x.

quail muscle, we have found the following method, based on that described by Caplan (6) and by Konigsberg (4) to be a very quick and satisfactory means of obtaining large numbers of mpc. 1. Eggs of the correct incubation age are swabbed with 70% ethanol and opened at the blunt end. The method normally recommended for doing this is to cut around the end of the egg with sterile scissors, but this generates egg shell dust, which may fall onto the embryo. We find it more satisfactory to use the method described for the preparation of CEE (above): breaching the shell with a blow of a sterilized scalpel handle, lifting off the cap of the egg shell, still adherent to the underlying membrane, with sterile coarse forceps, then using a second pair of forceps to lift out the embryo by the neck, without squeezing it, and transfer it to a sterile Petri dish. 2. The skin overlying the breas t muscle is dissected away, and the muscle is carefully removed and placed in a drop of sterile PBS in a second Petri dish. 3. The pooled muscle is then finely minced with curved scissors (as for the mouse), and the fragments are suspended in sterile growth medium; approximately 0.5 ml/quail breast, 1 ml/chick breast. 4. This suspension is distributed in 2-mL aliquots into large (20-25 mL) conical centrifuge tubes, and each tube is mixed for 30 s on a laboratory vortex mixer at high speed.

218

Partridge

5. After allowing the large fragments to settle, the supernatant is drawn off. The remaining fragments can be subjected to a second agitation in growth medium. 6. The combined supernatants are then filtered through3 layersof ZO-pm pore nitex mesh mounted in a Millepore filter holder and samples counted for viable cells as with the mouse cell preparations. 7. Depending on the desired timing and degree of myoblast fusion, the cells can then be seeded out in growth medium at 104-106cells/35mm Petri dish. Cell debris and nonviable cells initially contaminate cultures prepared in this way, but they can be removed by giving the cultures a brief wash and change of medium on the following day.

3.4. Cell Lines

It is possible to trypsinize primary cultures of muscle cells, using standard tissue-culture methods (see Chapter 1, this volume), and to propagate them for several generations. With an increasing number of passages, there is a tendency to lose the myogenic phenotype. This loss of myogenicity can be offset by seeding the cells at low densities, such that clones develop from individual cells, and subcloning from the nonfused cells present in myogenic clones (7). Stable continuous myogenic lines sometimes appear after several such serial passages. Continuous lines are available from the rat (L6, L8) and mouse (MM14, G8-1, C2), but not from human or avian origins. These lines provide the most experimentally reliable material; in most of them, the majority of cells fuse into myotubes in the right conditions and they present little difficulty in tissue culture. Of the lines of which we have experience, the C2 line (8) is the most satisfactory, being almost compulsively myogenic. It should be kept in mind, however, that these cells are to some extent neoplastic (all lines we have tested form tumors as well as muscle when implanted into nude mice) and should not be accepted uncritically as models of normal myogenesis.

4. Notes
1. Culture media: Sera and commercially available Chick Embryo extract are extremely variable and must be batch tested. As a general rule, all procedures, especially those concerning biologically derived components, must be rigidly standardized. 2. Ageof muscle: In general, the older the starting muscle, the longer the disaggregation period required, and the lower the yield of mpc. We obtain lo-lo8 total cells/g from neonatal muscle; 1-5 x lo5 cells/g from muscle of 6-7 wk-old mice. This problem can be partially over-

Tissue Culture of Skeletal Muscle

219

come by causing the muscle to start to regenerate in vivo prior to taking it for extraction of mpc: this has been done by ligaturing the vascular supply of the muscle, by injecting myotoxic agents such as marcain or notexin, or as we have sometimes done with good results, removing, mincing, and reimplanting the muscle as an autograft. This latter procedure, applied to 7-wk mouse muscle, 3-5 d after grafting yields some lo8 cells/g. About half of these cells are macrophages, but these can be removed by virtue of their rapid adherence to glass (see below), and the yield of mpc is still much improved. 3. Methods for enhancing myogenicity of cell preparations: Apart from the discontinuous density gradient centrifugation method described above for obtaining mouse cell populations enriched and impoverished in myogenic cells, a number of other methods have been employed for this purpose. Below are two of the more commonly used, and what looks to be a promising new method. a. Methods based on differences in the rate at which myogenic and nonmyogenic cells adhere to a solid substratum were originally developed on chick cells where they are reported to be very successful. They involve plating out the mixed cell suspension in a Petri dish for a short time (30-60 min), during which the majority of fibroblasts adhere to the plastic: the supernatant, now enriched in myogenic cells, is then removed and replated in another Petri dish. Such methods have also been applied to mouse skeletal muscle cells, apparently with success (e.g., 9,10), but except possibly for the removal of macrophages, we have not found them to be at all effective with muscle cells from this species. b. A very similar principle is the differential rate of de-adhesion of myogenic and nonmyogenic cells from a sustratum on mild trypsinization: the supernatant of a brief trypsinization is enriched in myogenic cells (10). c. Antigenic differences between myogenic and nonmyogenic cells offer one of the most promising routes to mass separation of the two populations. A monoclonal antibody to human myogenic cells, 5.1 Hll, has been used to separate large numbers of myogenic cells on the fluorescence-activated cell sorter (II). In the chick, a similar monoclonal antibody has been shown to cause rounding and detachment of myogenic cells from the substratum, while leaving nonmyogenic cells attached
(12).

4. Enhancement of differentiation:

In vivo, motor innervation and the

220

Partridge physiological workload greatly influence the state of differentiation of skeletal muscle fibers. Addition of extracts of peripheral nerve has been reported to produce better differentiated and longer-lived cultures than media lacking such extracts (23). By far the most effective in vitro model of skeletal muscle, however, is the nerve muscle coculture system developed by Peterson and Crain (14). In such cultures, slices of fetal mouse spinal chord are placed in culture together with fragments of skeletal muscle. Outgrowths from the spinal chord of nonneural cells, when they contact the muscle fragment, stimulate the proliferation and early differentiation of myogenie cells within it. Subsequently, axonal outgrowths of nerve cells

make synaptic contacts with the developing muscle fibers, causing

them to twitch and to express many of the phenotypic characteristics of normal skeletal muscle. The chief disadvantages of this method of culture are that some experience and a meticulous technique are required to set up and maintain the cultures, and that the amount of well-differentiated muscle tissue formed within them is too small to be of use for many biochemical techniques. 5. Problems: Two major problems of tissue culture in general seem to be exemplified in extreme forms by skeletal muscle in tissue culture. a. It is difficult to set up tissue cultures in which a high proportion of cells become differentiated and form muscle fibers. The principal reason for this is the difficulty of ridding cultures of cells with no myogenic potential. However, even with cloned myogenic cells, not all, and sometimes only a minor proportion, develop into muscle cells. b. The degree of differentiation attained by muscle fibers grown in tissue culture falls far short of what is found in vivo. The innervated cocultures described above are the most effective attempt so far to remedy this. 6. Two practical difficulties we have encountered are: a. Poor reproducibility. Even with standardized techniques and batched media components, variation between individual primary culture preparations is often too great to permit comparisons to be made between them. It is safest, therefore, to study the effects of variables within particular preparations. Cultures of muscle of the Japanese quail seem less susceptible to this problem. b. Long-term cultures. As the contractile myofibrils assemble within developing muscle fibers, these cells respond to minor disturbances, such as illumination under a microscope, change

Tissue Culture of Skeletal Muscle

221

of medium, or addition of more cells, by contracting violently. This may cause irreversible detachment of the entire monolayer of cells from their substratum. In general, it is best to use the simplest culture system for skeletal muscle that will critically test the hypothesis under consideration. If it is necessary to use one of the more complicated methods, it can save much time and effort to acquire experience of it in a laboratory where it is used routinely, because written accounts of a technique do not record all of the subliminal practices that are frequently crucial to its success.

Acknowledgments
I wish to thank Drs. J. Morgan and D. Watt for their advice and criticism. The work on mouse muscle culture is supported by the Muscular Dystrophy Group of Great Britain and Northern Ireland.

References
1. Wekerle, H., Paterson, B., Ketelson, U.-P., and Feldman, M. (1975) Striated muscle fibers differentiate in monolayer cultures of adult thymus reticulum. Nature 256, 493,494. 2. Watt, D. J., Lambert, K., Morgan, J. E., Partridge, T. A., and Sloper, J. C. (1982) Incorporation of donor muscle precursor cells into an area of muscle regeneration in the host mouse. J. Neural. Sci. 57,319-331. 3. Paul, J. (1975) Cell and Tissue CuIture (Churchill Livingstone, Edinburgh). 4. Konigsberg, I. R. (1979) Skeletal myoblasts in culture. Methods in Enzymology LVIII, 51 l-527. 5. Turner, D. C. (1978) Differentiation in cultures derived from embryonic chicken muscle. The post-mitotic fusion-capable myoblast as a distinct cell type. Differentiation 10,81-93. 6. Caplan, A. I. (1976) A simplified procedure for preparing myogenic cells for culture. J. Embyol. Exp. Morphol. 36,175-181. 7. Hauschka, S. D., Linkhart, T. A., Clegg, C., and Merrill, G. (1979) Clonal studies of human and mouse muscle, in Muscle Regeneration (Mauro, A., ed.), Raven, New York, pp. 311322. 8. Yaffe, D. and Saxel, 0. (1977) Serial passaging and differentiation of myogenic cells isolated from dystrophic mouse muscle. Nature 270,725-727. 9. Cossu, G., Pacifici, M., Marino, M., Zani, B., Coletta, M., and Molinaro, M. (1981) Developmental changes in glycopeptides synthesized by normal and dystrophic satellite cells in culture. Exp. Cell. Res. 132,349357. 10. Kiihl, U., Timpl, R., and Von Der Mark, K. (1982) Synthesis of type IV collagen and laminin in cultures of skeletal muscle cells and their assembly on the surface of myotubes. Dev. Sol. 93,344-354. Il. Webster, C., Pavlath, G. K., Walsh, F. S., Parks, D. R., and Blau, H. M. (1988) Isolation of human myoblasts with the fluorescence activated cell sorter. Exp. Cell Res. 174,252-265.

222

Partridge

22. Sasse,J., Horwitz, A., Pacifici, M., and Holtzer, H. (1984) Separation of precursor myogenic and chondrogentc cells in early limb bud mesenchyme by a monoclonal antibody. J. Cell Bid. 99,1856-X366. 13. Oh, T. H. and Markelonis, G. J. (1979) Neurotrophic effects of a protein fraction isolated from peripheral nerves on skeletal muscle in culture, inMuscle Regeneration (Mauro, A., ed.), Raven, New York, pp. 417-427. 24. Peterson, E. R. and Cram, S. M. (1979) Maturation of human muscle after innervation by fetal mouse spinal chord explants in long term cultures, inMuscIe Regenemtion (Mauro, A., ed.), Raven, New York, pp. 429-441.

Chapter

20

Derivation and Maintenance of Embryonic Stem Cell Cultures

Elizabeth J. Robertson

1. Introduction
The ability to derive permanent tissue culture lines (I) of pluripotential stem cell lines (ES cells) from mouse embryos has provided a valuable model system for fundamental research into the cellular differentiation processes occuring in the normal embryo. Perhaps the most attractive feature of embryonic stem cell lines is that they can be manipulated to differentiate into a diversity of cell types either directly in the tissue culture dish or within the context of the normal developing embryo following their return to the embryonic environment. ES cells can be maintained in the undifferentiated state for periods of several months without loss of their developmental capacity. When ES cells are reintroduced into host carrier embryos by the technique of blastocyst injection (2), they resume a normal regulated pattern of proliferation and differentiation to form chimeric conceptuses. ES cells are unrestricted in their pattern of functional differentiation contributing extensively both to the somatic and germ cell lineages of adult chimeras (3). Transmission of ES cells in the germ cells provides a route to deriving transgenic mouse strains, i.e., animals that harbor exogenous introduced DNA sequences. ES cells offer several attractive advantages over the 223

224

Robertson

alternative techniques that are currently available for the production of transgenic mice, such as direct injection of the one-cell mouse egg or retroviral infection of embryos (reviewed 4) as the cells are accessible for experimental manipulation and screening in culture prior to their incorporation into the animal (5-7). This chapter describes the technique for the derivation of ES cells from cultured blastocyst-stage embryos and the protocols for the routine culture of established ES cell lines. For techniques involved in the production and analysis of ES chimeras, the reader is referred elsewhere (2).

2. Materials
1. Complete growth medium: Dulbeccos Modified Eagles Medium (high glucose formulation) plus 10% (v/v) fetal calf serum, 10% (v/v

newborn calf serum, seeNote l), 5 x 1094 Z-mercaptoethanol, antibiotics (50 U/mL penicillin, 50 pg/mL streptomycin), and 1% non-

essential amino acids.

2. Trypsin-EDTA; 0.25% (w/v> powdered porcine trypsin in 0.04% (w/v) EDTA in Phosphate buffered saline (pH 7.6). 3. Phosphate buffered saline (calcium and magnesium free): 0.17M NaCl, 3.4 mM KCl, 4 mM Na,HPO,*12H,O, 2.4 mM KHJYO,. 4. Mitomycin C medium; Dulbeccos Modified Eagles Medium supplemented with 10% (v/v) newborn calf serum and 10 pg/mL mitomycin C (prepare freshly at each use). 5. Lightweight paraffin oil.

3. Methods
3.1. Isolation
of ES CeZZLines

(See Note 2)

3.1.1. Preparation of Embryonic Outgrowths 1. Prepare feeder layers in a 24-well tissue culture plate. Seed 5-7 x lo* mitomycin C-treated ST0 fibroblasts in 1 mL complete growth medium/well. Feeder layers should be prepared 1 d prior to collecting embryos. 2. Set up timed matings using the mouse strain of choice. The blastocyst stage of embryonic development is reached 3.5 d after fertilization of the oocyte. Inspect females daily for copulation plugs (day of plug is day 0). On day 3, kill the females by cervical dislocation, swab the abdomen with liberal quantities of 70% ethanol, and dissect out

Embryonic Stem Cell Cultures

225

individual uterine horns (cut below the uterotubal junction and above the cervix). Remove the horns into a sterile Petri dish. Flush the
uterine lumen by introducing a 26-gage needle attached to a l-mL syringe filled with complete growth medium (seeNotes 3 and 4). 3. Remove the dish to a dissecting binocular microscope and locate the

blastocysts. Using a finely drawn out Pasteur pipet attached to a mouth-controlled tube, collect the blastocysts. Transfer the embryos into a small drop of complete growth medium (overlay the drop with lightweight paraffin oil to prevent evaporation). 4. Transfer the embryos individually into the feeder wells taking care to place the embryo in the center of the well (seeNote 5). Return the plate to the incubator and culture undisturbed for 4 d. After this time,
inspect the embryos. The blastocysts will have attached by the outgrowth of the trophectoderm cells. The inner cell mass (ICM) of the embryo will have proliferated to form a small discrete mass of cells (Fig. 1).

3.1.2. Disaggregation

of Inner Cell Mass (ICM) Clumps

1. Prepare fresh feeder layers in a 24-well plate (as above). 2. Aspirate the medium from the wells containing the embryos and add 1 mL of PBS. 3. Prepare a tissue culture dish containing an array of small drops of trypsin-EDTA solution (approx. 100 pL). Overlay the drops with lightweight paraffin oil. 4. Place the 24-well plate under a dissecting microscope, and using a finely drawn out Pasteur pipet, carefully dislodge the ICM clumps from the trophectoderm cells. Transfer each ICM into a separate trypsin drop. Return the dish to the incubator and leave for 5 min. 5. Pull out a very fine diameter capillary in a bunsen flame from a Pasteur pipet. Attach the pipet to a mouth-controlled tube. Using gentle suction, suck a small quantity of complete growth medium into the

pipet.
6. Return the dish of trypsin drops to the binocular microscope. Dissociate each ICM as follows. Blow a small quantity of medium from the pipet into the trypsin drop (to neutralize the trypsin action). Using mouth control, draw the ICM clump in and out of the end of the pipet (the end of the pipet should be approximately a quarter the diameter of the ICM clump). Fragment the ICM into several small cellular clumps, and transfer the fragments into a fresh feeder well. Repeat for all embryos, and return the 24-well plate to the incubator.

226

Robertson

Fig. 1 Attachment and growth of a single blastocyst on a gelatinized tissue culture surface. (a) 1.5-d culture, blastocyst attaching by trophoblast cell outgrowth; (b) 2.5-d culture, ICM component is revealed by the continued spreading of the trophoblast cells; (c) 4-d culture, cells of the ICM proliferating rapidly; (d) 5-d culture, ICM has formed a mass of cells.

3.1.3.

Identification

and Expansion

of ES Cell Colonies

1. Incubate the ICM-derived fragments for 4 d. During this period, the fragments will attach and proliferate to form small primary colonies of cells.

Embryonic

Stem Cell Cultures

227

Fig. 2 Morphologies of primary colonies that result from the disaggregated ICM. (a) patch of spread trophoblast cells; (b) fibroblast-like cells; (c) epithelioid-like cells; (d) epithelioid cells, single colony of cells at high power; (e) appearance at high power of trophoblast-like cells before cell flattening has occurred; (fl colony of stem cells.

2. Inspect each well under high-power phase contrast (e.g., 200x). With reference to Fig. 2, classify the individual primary colonies according to morphology. A variety of differentiated phenotypes will be noted.

.:d

Fig. 3 Progressive alteration to the morphology of a putative stem cell colony over a 3d culture period. (a) When first located 2 d after disaggregation of the ICM, the constituent cells superficially resemble stem cells. (b) Following a further 3 d culture, the colony has spread and flattened to form a patch of giant trophoblast-like cells.

Mark potential ES cell colonies by circling the posi tion of the colony on the underside of the well with a fine-tipped, indelible marking pen. Return the plate to the incubator. Inspect the marked colonies daily for the following 3 d. Eliminate colonies that differentiate to form large flat cells (seeFig. 3). Genuine colonies of ES cells will increase in size in the absence of overt differentiation (seeFigure 4, seeNote 6).

Embryonic

Stem Cell Cultures

229

Fig. 4 Progressive growth of a colony of stem cells. (a) d 3; (b) d 4; (c) d 5. Note that the colony increases in size in the absence of overt differentiation. The cells pile up and maintain a discrete colony on top of the feeder cells. The constituent cells adhere tightly to one another making it difficult to see individual cells. This growth pattern is characteristic of stem cells, although the colonies may appear slightly flatter and less regularly shaped than the colony illustrated here.

3. Using the following procedure, subculture colonies of stem cells approximately 6-7 d after the primary disaggregation event. To avoid the possibility of cross-contamination with nonstem cells, each putative ES cell colony should be dissociated individually.

230

Robertson

4. Prepare the feeder layers in l-cm wells. Aspirate the well containing primary colonies and add 1 mL of PBS. 5. Using a drawn out Pasteur pipet dislodge the marked colony, transfer it to a small drop of trypsin-EDTA, and incubate for 2-3 min in the incubator. 6. Using a finely drawn out Pasteur pipet dissociate the colony into a single-cell suspension. This should be achieved easily, since stem cells are more readily dissociated than the original ICM-derived clump. 7. Transfer the cell suspension to a l-cm feeder well containing 1 mL of complete growth medium and incubate. 8. Inspect the wells after 2 d for the presence of stem cells. These are readily identifiable as having a very characteristic morphology (see Fig. 5). 9. After 2 d of further culture, passage the wells containing ES cell colonies. Wash the well with 1 mL of PBS and trypsinize the cells by the addition of 100 PL trypsin-EDTA solution. Allow a 3-4 min incubation in trypsin. Then using a Pasteur pipet and bulb, add 100 PL complete growth medium to each well. Vigorously disperse the cells to ensure that a single-cell suspension is obtained. 10. Seed the single-cell suspension from each well into a 3-cm feeder plate containing 2 mL of complete growth medium. 11. Refeed the plates daily. After 3 d, sufficient stem cells should be present to necessitate subculture. Stem cells can also be frozen down at this stage (seeNote 7).

3.2. Routine

Culture

of ES Cell Lines

It is recommended that ES cells be cultured exclusively on feeder layers. However, for some experimental procedures, it is desireable to omit feeder cells. This can be achieved by culturing the cells on gelatinized plates in complete growth medium supplemented with 50% BRL conditioned medium (8). Stem cells grow rapidly and divide approximately every 15-18 h. The ES cell line should be maintained at relatively high densities to ensure maximal growth rate in order to minimize spontaneous differentiation. 1. Cultures of ES cells should be inspected and refed daily. The cultures should be passaged when the stem cells have reached an approximately 80% confluency. 2. Refeed the cultures 2-3 h prior to subculture to maximize the cell viability.

Embryonic

Stem Cell Cultures

231

Fig. 5 Examples of typical areas of cells found within a subconfluent culture of established stem cells, to illustrate the cellular morphology exhibited by pluripotential cells. The culture was seeded initially as single cells. These divide and the sister cells stay together to form small nests. With time the nests increase in size and merge to form a confluent layer.

3. Aspirate the growthmedium andwashwith2-3mLof PBS- Add1 mL of trypsin-EDTA solution, and return the plates to the incubator for 4-5 min.

232

Robertson

4. Gently rock the plates to dislodge the cells. Add 1 mL of complete growth medium, and using a Pasteur pipet and bulb, dissociate the cells to a single-cell suspension. 5. Centrifuge the cell suspension at 1000 rpm for 5 min. Aspirate the medium and resuspend the cells in 5 mL of complete growth medium. Determine the cell density, and replate the cells into fresh feeder plates at a density of IO6 cells/&cm plate. 6. Refeed andinspect the plates daily. An 80% confluent dishof cells will be obtained after 34 d of culture, and at this time, a 6-cm plate will typically yield 2 x 10 cells.

4. Notes
1. Both fetal serum and newborn calf serum are used as medium supplements. For the successful growth of most tissue culture cells, serum quality is critical, and this is especially true for blastocysts and embryonic cells. The use of unsuitable serum is undoubtedly one of the reasons for a failure to isolate stem cells from embryos. All sera should be tested for their ability to support the growth of pluripotential stem cells: serum testing is laborious, and the best strategy is to simultaneously request the maximum number of available samples from all the commercial suppliers. The samples can then be tested in parallel and a bulk order placed for the most suitable sera. Serum can be stored for periods of up to 2 yr at -2OOC. The most sensitive test of the ability of a given serum to support the growth of stem cells is by a comparison of the plating efficiency of single cells from either an established embryo-derived stem cell line or, alternatively, feederdependent EC cell line (e.g., PSA-4). The stem cells are coplated with feeder ST0 cells. Control samples of fetal newborn calf sera are included in the test. Only sera that are comparable to, or better than, the control sera should be bought. The technique is to set up duplicate plates containing a suspension of mitomycin C-treated ST0 fibroblast cells and a small number of s tern cells in serum-free medium. The plates are placed in sets with each set being allocated a specific serum batch. Serum is added to a final concentration of 10%. To a single plate in each set, serum is added to a final concentration of 30%. At this concentration, any toxic component will readily be detected. The plates are incubated for 7-10 d, and the stem cell colonies stained and counted. A good quality serum gives a plating efficiency of 20% or higher. Toxicity effects will be evidenced by a lower plating efficiency and/or the formation of smaller colonies at the 30% concentration Serum toxicity may be the result in part of high levels of

Embryonic Stem Cell Cultures

233

complement. Heat treating serum (56OC for 30 min) to inactivate complement may dramatically decrease toxicity. 2. Karyotype analysis procedures should be used to ascertain the chromosome complement and sex chromosome constitution of a cell line as soon as possible after it has been established. The majority of ES cell lines derived, to date, have a normal euploid chromosome complement when established cultures are analyzed. However, it is inevitable that cell populations will drift away from the normal genotype and subsets of aneuploid cells will be selected over a period of continuous culture. For this reason, it is important to routinely check the chromosome complement of stem cells in culture. Simple chromosome counts are used to determine the model number, and estimate the range and variability of the population. Ideally, this should be complemented by G-banding analysis, which enables the exact chromosomal constitution to be determined. 3. The starting material of choice for the isolation of stem cells is the blastocyst stage embryo. Since the time course of mouse embryogenesis is well documented, and the time of fertilization can be readily determined, the recommended method for obtaining the maximum number of fully viable embryos is to leave the embryos to develop within the reproductive tract until they have reached the appropriate blastocyst stage. Blastocysts are amenable to culture in standard tissue culture media and do not require special requirements. This is not the case for embryos at preblas tocys t stages that have to be grown in a simpler serum-free medium. The mouse strain from which a cell line is to be made is another important consideration. The majority of in-bred strains will, for example, yield relatively few embryos (normally between five and eight). If approaching this technique for the first time, it is recommended that more fecund outbred stocks of mice be used to generate embryos. Outbred mouse stacks are readily available from commercial breeders, and matings provide large numbers of embryos on which to practice the various techniques before turning to the particular strain, or mutations, of interest. The rate of recovery of stem cell lines is rarely higher than 30%, and may be lower than 10%. Therefore, a large percentage of the embryos are wasted. If the mouse strain of interest is available in restricted numbers, or if the desired genotype is carried by only a fraction of the embryos from a given mating, attempts to isolate a stem cell line can be frustrating. 4. Blastocysts that have undergone a period of implantational delay prior to flushing may be more suitable material from which to recover stem cells; the success rate can be significantly improved by the use of

Robertson
delay is brought about by alterThe sourceof estrogen production is removed by surgical ovariectomy 2.5 d postcoitum. Lowering the level of estrogen, together with the postoperative administration of a synthetic progesterone, prevents the embryos from implanting. The embryos develop normally to the blastocyst stage, but since the tissues of the uterine wall have been artificially rendered nonreceptive, the embryos arrest and remain free floating within the uterine lumen. In the delayed state, the embryos remain viable for up to 10 d. For the purposes of isolating stem cells, a 4-5-d delay period is normally used. Embryos are flushed from the uterine horns and treated exactly as described for normal blastocysts. 5. There are a variety of technical considerations to be made regarding the best strategy for the culture of blastocysts. These are as follows: a. Embryos collected on a single day can be kept together and delayed blastocysts. Implantational

ing the hormonalstatus of thepregnantfemale.

cultured

as one group,

or split up into individual

embryo

cultures. For most purposes, it is recommended that embryos be cultured singly. This has the advantage of avoiding the necessity of recloning the stem cell cultures obtained, since a cell line derived from a single embryo is effectively a clonal population in terms of genotype and sex chromosome complement. An additional advantage is that the monitoring and recording of embryos on an individual basis provides information about the overall viability of cultures and the efficiency of recovering stem cells. The major disadvantage to this approach is that it takes considerably more time and materials to maintain large numbers of individual cultures. b. Blastocysts can either be grown in drop cultures under a layer of liquid paraffin oil or in larger tissue culture dishes. The advantage of drop cultures is that embryos are easily located for observation because they are physically confined to a small area. In our experience, however, embryos grown in drop cultures for the necessary 5-6 d fare less well. This is probably because of depletion of an essential growth requirement in the medium, or more likely results from the gradual accumulation into the medium of traces of toxic residues from the overlying oil. The most satisfactory method is to use small tissue culture wells (10 mm) that hold about 1 mL of medium. c. Embryos may either be cultured directly on tissue culture plastic or on feeder layers. The use of feeder layers during
initial culture appears to maximize the viability of blastocyst

Embryonic

Stem Cell Cultures

cultures. The stage at which an individual ICM-derived component is selected for disaggregation is fairly critical. Figure 1 shows examples of the morphologies of blastocyst outgrowths. Under the conditions of culture described above, embryos will normally attain a sui tablemorphology following a 5-d culture period. A degree of variability in the size of the ICM-derived cell clump between embryos of the same batch is a common observation. As a result, it may be necessary to disaggregate ICM clumps over a 2-d period according to the morphology of individual cultures. It is important to note that there is by no means an absolute correlation between the phenotype of the outgrowth and the successful isolation of a stem cell line; stem cells can be ob tained from both vigorously growing ICMs and from small outgrowths. In our experience, however, ICM clumps in which there is extensive endoderm formation, or a relatively rapid progression to an overtly multilayered egg cylinderlike structure, tend to have a reduced chance of retaining pluripotential cells. 6. It is important to be able to accurately identify colonies of pluripotential stem cells. Positive identification is difficult unless one is familiar with the appropriate cellular morphology. Pluripotential cells are typically small, have a large nucleus and minimal cytoplasm, and the nuclei contain one or more prominent dark nucleoli structures, The cells pack tightly together in small nests in which it is difficult to discern the individual component cells. Individual cells aremost easily seen at the edges of the colony. A helpful exercise, prior to attempting to isolate stem cells from embryos, is to plate out single cells from a feeder-dependent Embryonal Carcinoma cell line onto a feeder plate (e.g., 5 x lo2 cells on a IO-cm plate). Carefully observe the growth of the resulting colonies over a period of 5-8 d. 7. It is recommended that embryo-derived stem cell lines should be grown exclusively on feeder layers. This, together with careful culture in high-quality medium, acts to prolong the embryonic phenotype of the cell line specifically in terms of maintaining a high differentiation ability and euploid chromosome complement. As with all permanently established cell lines, long-term tissue culture will, however, select for abnormal cells within the population. For this reason, it is advised that many samples be frozen as soon as possible after founding a specific cell line. Cultures can be replaced as necessary or recloned to establish euploid cultures. Recloning can be achieved by picking single cells into feeder wells. The resulting colonies of cells are

236

Robertson

reexpanded by the method described. For experimental procedures in which it is desirable to minimize the number of contaminating fibroblasts from the feeder layer, it is possible to passage the stem cells once or twice on gelatinized plates. If cultures are kept under these conditions for longer periods of time, extensive cell death occurs.

References
2. Evans, M. J. and Kaufman, M. H. (1981) Establishment in culture of pluripotential cells from mouse embryos. Nature 292,154. 2. Bradley, A. (1987) Production and analysis of chimeric mice, in Terufocarcinomas and Embryonic Stem Cells: A Prucficd Approach (Robertson, E. J., ed.), IRL Press Ltd., Oxford, p. 113. 3. Bradley, A., Evans, M., Kaufman, M. H., and Robertson, E. (1984) Formation of germ line chimeras from embryo-derived teratocarcinoma cell lines. Nature 309,255. 4. Wagner, E. F. and Stewart, C. L. (1987) Integration and expression of genes introduced into mouse embryos, in Experimental Approaches to Mammalian Emb yonic Development, (Rossant, J. and Pedersen, R. A., eds.), Cambridge University Press, UK, p. 509. 5. Robertson, E., Bradley, A., Kuehn, M.,and Evans, M. (1986) Germ-line transmission of genes introduced into cultured pluripotential cells by retroviral vector. Nature 323,154. 6. Kuehn, M., Bradley, A., Robertson, E., and Evans, M. (1987) A potential animal model for Lesch-Nyhan syndrome through introduction of I-IPRT mutations into mice. Nature 326,292. 7. Hooper, M., Hardy, K., Handyside, A., Hunter, S., and Monk, M. (1987) HPRTdeficient (Lesch-Nyhan) mouse embryos derived from germ line colonization by cultured cells. Nature 326,292. 8. Smith, A. and Hooper, M. (1987) Buffalo Rat Liver cells produce a diffusible activity which inhibits the differentiation of Murine Embryonal Carcinoma and Embryonic Stem Cells. Develop. Biol. 121,l.

Chapter 21
Defining Hormone and Matrix Requirements for Differentiated Epithelia
Lola M. Reid

1. Introduction
The culture of differentiated cells requires conditions that acknowledge the complicated cell-cell interactions that both occur in vivo and are responsible for affecting and maintaining the differentiated states of cells. In brief, one must use conditions that mimic the epithelial-mesenchymal relationship that is universal and constitutes theorganizational basis for all metazoan tissues. This relationship is sustained by a set of soluble signals (autocrine, paracrine, and endocrine) and by a set of insoluble signals (the extracellular matrix). Since this is a technical and methodological article, neither the scientific evidence for the importance of the epithelial-mesenchymal relationship nor the evidence forming the basis for the culture conditions will be described. Recent reviews have discussed this background in considerable detail (I-12).

1.1. Classical

Cell Culture

Conditions

As background and for the sake of comparison to the new forms of culture, the response of cells under classical cell culture conditions will be noted. For a detailed description and discussion of classical cell culture methods, see any of a number of recently published books (13-25, and Chapter 1 of this volume). 237

238

Reid

All methods of preparing cells for culture start with the disruption of the tissue and its dispersal into chunks or single-cell suspensions. In classical cell culture, the dispersed tissue or cells are plated onto an inflexible, charged plastic substratum, and suspended in or covered with a liquid medium. The plastic substratum, in the form of culture dishes, consists of polystyrene dishes or containers exposed to ionizing plasma gas that produces a negatively charged surface that will permit the cells to adhere by charge binding. The cells adhere and spread via the negative charges, and then are fed with a liquid basal medium that typically consists of a specific mixture of salts, amino acids, trace elements, carbohydrates, and so on, and that is supplemented from 5 to 25% with a biological fluid, usually serum. The basal media that are commonly used (e.g., RPMI, DME, BME, Waymouths) wereoriginally developed for cultures of fibroblasts (23-U). Although most of the constitutents in these basal media are requirements for both epithelial cells and fibroblasts, some aspects of the basal media must be redefined for some epithelial cell types (26-18). Two classic examples are trace elements and calcium levels. Highly differentiated epithelia require various trace elements or other factors that act as cofactors for the enzymes associated with their tissue-specific functions (16-22). The calcium levels in many of the commercially available media are above 1 mM, concentrations that are permissive for growth of fibroblasts, but that in recent years have been shown to be inhibitory to most epithelial cell types (16-22). This problem is exacerbated by culturing the cells in serum-supplemented media, since serum also contributes significantly to calcium levels. Most normal epithelial cells can grow in calcium concentrations of approximately 0.4 mM (19-22). An excellent review of these aspects of basal media construction has been written by Ham and McKeehan (22). Someinvestigators utilize serum autologous to the cell types to be cultured. However, it is more common that the serum derives from animals that are routinely slaughtered for commercial usage, such as cows, horses, sheep, or pigs (13-25). Fibroblasts do well in serum-supplemented media (SSM). By contrast, epithelial cell types cultured in SSM dedifferentiate rapidly and then die, usually within a week (23,24). Tumor cell lines derived from epithelia are resistant to some of the adverse effects of serum, and will grow and even express some of their tissuespecific functions in SSM (25). Still, if these cell lines are switched into serum-free, defined conditions, their differentiated functions are markedly improved (26).
SSM, serum supplemented medium; HDM, serum-free, hormonally defined medium; EGF, epidermal growth factor; GAGS, glycosaminoglycans, PGs, proteoglycans.

Requirements for Differentiated

Epithelia

239

1.2. Realistic Expectations for Cultures of Distinct Classes of Epithelial Cells

In order to culture epithelial cells, one must define the soluble signals (hormones, growth factors) and the matrix signals needed to sustain the cells either in a growth state or in a differentiated state. Using the defined conditions results in cells that can approximate the behavior of their in vivo counterparts. However, even with optimal defined conditions, one must be realistic in ones expectations for cultures of particular cell types. The expectations differ with several general classes of cells. Three classesof cells have been described in adult tissues that require distinct conditions, and require distinct expectations and assumptions for what the cells will do in culture. All involve stem cells and the lineage of cells derived from them, and generally support the original hypothesis of Pierce and Potter that all adult tissues have stem cells that can become the targets for carcinogenic events (27). A schematic diagram of these classes is given in Fig. 1. These classes are: 1. Stem cells that give rise to constantly regenerating tissues-eg., epidermal cells (skin), bone marrow stem cells (hemopoietic cells), intestinal stem cells (2832). The stem cells require attachment to a form of extracellular matrix, and the presence of specific hormones and growth factors for viability and for sustained growth potential. The adherent cells grow, but do not express their tissue-specific functions. Detachment from the extracellular matrix substratum (mimicking the loss of contact with the stromal layer) results in terminal differentiation occurring in parallel with the loss of viability and restriction of growth potential. The kinetics in the terminal differentiation process and the loss of viability and growth potential vary with the cell type (28-32). Cell cultures of this class of stem cells can be routinely cloned and subcultured. However, cultures of the differentiated derivatives are inherently short-lived; the cultures survive for a matter of days to, at most, a few weeks. Obviously, the life span of the differentiated derivatives is dictated by the inherent life span of the cell in vivo. It is this in vivo life span that sets the limits for what is possible in culture. 2, Stem cells that give rise to quiescent tissues having some regenerative potential-e.g., liver, lung, pancreas, prostate, mammary glands. Recent studies (33-38) have provided strong evidence that all of these tissues contain small numbers of stem cells that give rise to a cell lineage gradually developing into a cell type forming the bulk of the cells in the quiescent tissues. The cell type predominating in the quiescent tissue is stabilized into a G, state (or a very long G, state) by extracellu-

240
Figure 1. Classes of Epithelia Class 1: Tissues that are undergoing haemopoietic cells STEM CELLS --) --) + --) rapid renewal: skin, intestine,

Reid

TERMINALLY

DIFFERENTIATED CELLS

Intermediates in the differentiation

pathway

Class 2: Tissues that exist for long periods in a quiescent state but retain some regenerative potential: liver, prostate, pancreas, lung STEM CELLS + (the number of stem cells declines with
age)

INTERMEDIATE CELL(S) STABILIZED BY MATRIX IN QUIESCENT TISSUE L 1

~ Regeneration Signals

1 (limited cell division)

TERMINALLY

DIFFERENTIATED

CELLS

Class 3: Tissues in which there is little to no regeneration possible: muscle, neurons in the CNS STEM CELLS + + -+ MATRIX STABILIZED TERMINALLY DIFFERENTIATED CELLS

Requirements

for Differentiated

Epithelia

241

lar matrix and hormonal conditions. When regenerative conditions occur, the signals, again both hormones and matrix, trigger these quiescent cells to undergo a few rounds of division, typically 2-4 divisions, accompanied by a terminal differentiation process. The signals for regeneration also recruit the stem cells to send more cells through the cell lineage. Generally speaking, fetal tissues have greater numbers of these stem cells than do adult tissues, and the older the animal, the fewer the stem cells in these tissues. However, studies to date have not defined well either the kinetics of aging of the stem cells or the real replicative ability of these cells for any tissue in this class. Defined culture conditions for the predominant cell in the tissue can result either in amitotic cultures that are fully differentiated and able to be maintained for a long-term (months) (39), or in cultures that can undergo 2-4 divisions and then go into a state of growth arrest (40). In this state, they can survive for months. However, no conditions are known that will enable any of the predominant cell types from these tissues to be routinely subcultured or cloned. It is hypothesized that cultures of the stem cells from these tissues will be able to be cloned or subcultured. Although a number of investigators are making considerable progress in culturing stem cells (41,42), the conditions for routinely maintaining the stem cell populations have not yet been fully defined for any stem cell from this class of tissues. 3. Stem cells that give rise to cells stabilized in the terminally differentiated state-e.g., CNS neurons and muscle cells. Muscle and brain do not regenerate, or do so in a most limited fashion (43,44). The tissues are assumed, in adult state, to consist of terminally differentiated cells stabilized by the matrix and hormones. Whether or not there are stem cells in the adult muscle and brain tissues is unknown. All of the known matrix and hormonal conditions for cultures of these tissues result in cell populations that survive and function, but do not grow.

1.3. Defining the&equirements for a Specific Epithelial Cell 5?)pe

1.3.1. Soluble Signals: Autocrine, Paracrine, and Endocrine Factors An approach to defining the soluble signals from cell-cell interactions has been to replace the serum supplements in medium with known and purified hormones and growth factors (16-18). This results in the development of serum-free, hormonally defined media (HDM). Table 1 provides some representative defined media for several epithelial cell types. There

242
Table 1

Reid

Constituents of Hormonally Defined Media for Several Representative Epithelial Cell9 Proximal kidney tubule cells ms~lm (5 &mL), transferrin (5 &mL), hydrocortisone (5 x WW), T3 (5 x l(PM), prostaglandin El (25 ng/mL), selenium (5 x 10m8M) (4% Mammary epithdial ceh Dexamethasone (10-8M), epidermal growth factor (10 ng/mL), insulin (s-10 pg/mL),transferrin (5-10 pg/mL), prolactm (1 pg/mL), T, (IO-OM), 17P-estradiol (10-8M), linoleic acid/E&I (5 kg/mL) (50) SertoIi cells Epidermal growth factor (1 ng/mL), insulin (5 pg/mL), transferAn (5 pg/mL), growth hormone (0.1 pg/mL), follicle stimulating hormone (0.5 ng/mL,), hydrocortisone (10-8M), adrenocorticotropm (5 mU/mL), vitamin A (50 ng/mL), and vitamin E (0.001%) (18)
*For recipes of other epithelial cell types (or for fibroblasts) see Methods in Cell Biology, Barnes and Sato (27) and Mather (18).

are several books and review articles giving recipes for defined media for other cell types (26-18). Use of HDM results in selection of the epithelial cell type of interest from primary cultures containing multiple cell types, Almost all of the published HDM are optimized for cell growth. To observe optimal expression of differentiated functions, the HDM must be retailored (22). Each tissue-specific function requires a discrete set of hormones and growth factors, often at concentrations that differ from those required for cell growth. Furthermore, some of the hormones conducive to growth can markedly inhibit tissue-specific functions. Thus, a rule of thumb is to develop an HDM for growth of cells and then retailor this medium for the conditions needed for differentiation of those cells (22). The development of an HDM for a cell type begins by defining the hormonal and growth factor requirements for growth of a minimally deviant neoplastic cell line (16-18). Since a cell line is already adapted to culture, it can readily be used in clonal growth assays to define growth requirements under serum-free conditions. The HDM for the tumor cells usually contains a subset of the requirements and is, therefore, a starting point in defining the requirements for the normal cellular counterparts to these tumor cells (45-48; seeTable 2 for examples). Using the minimally deviant tumor cell line, one determines its clonal growth efficiency (percent of cell colonies that grow at low seeding den-

Requirements for Differentiated

Epithelia

243

Table 2 Comparison of the Composition of a Hormonally Defined Medium for a Normal CelI (Hepatocytes)
vs Its Neoplastic Counterpart (Hepatomasp

Hepatocytes EGF Insulin 50 ng/mL 1-5 pg/mL

Hepatomas
Not required 5-10 pg/mL (or more if

insulin-degrading enzymes are produced by the tumor cells) Glucagon

Glucocorticoidb Transferrin 10-8M Not required

10 pg/mL

10 pg/mL 10M 5-10 pg/mL (depends on whether the hepatoma cells have, and to what

Linoleic acid HlzL Triiodothyronine Prolactin Growth hormone Trace elements zinc Copper Selenium

pg/mL

degree, the capacity to produce transfer&$ 10 pg/mL

1 x 10-9M 2 mU/mL 10 pU/mL

n. t.

Not required

10 pg/mL

20 mU/mL 10 pU/mL
1 x lO-OM 1 x 10-7M

3 x lo-OM

1 x lO-OM 1 x 10-7M 3 x lo-OM

These culture conditions (21,12,45,48)permit clonal growth of many hepatoma cell lines (human, rat, and mouse cell lines have been tried). However, normal hepatocytes will grow at high densities (>105cells/60mm dishes), but not at clonal seeding densities (10*-l@ cells/6Omm dishes) when maintained in the above hormonally defined medium. Furthermore, unless protease inhibitors and/or specific matrix substrata are used, the cells must be seeded in a mixture of 5-10% serum plus the hormones for a few hours (4-6 h) and then transferred into the serum-free, hormonally defined medium. bGlucocorticoid can be hydrocortisone or dexamethasone, depending on the species. Linoleic acid must be present in combination with fatty acid free bovine serum albumin (Pentax, Inc.) at a l/l molar ratio. If added alone, it is quite toxic to the cells. Some species of hepatomas (e.g., the human hepatoma cell line, HepG2) have more complex lipid requirements, especially those cell lines that are poor in making cholesterol or one or another of the lipoproteins present in high density lipoprotein (HDL). These cells must be given a more complex lipid source such as HDL.

244

Reid

sities, such as 100-1000 cells/bO-mm dish) on tissue culture plastic plates, and under serum-supplemented medium (SSM) conditions or under serum-free medium (SFM) conditions. This will establish the positive and negative control conditions. Since most differentiated epithelial cells are quite dependent for survival on substrata of extracellular matrix, one should determine the cells clonal growth efficiency in SSM and SFM using a simple collagen substratum (45). Usually, the cells will attach and survive on collagen substrata and in SFM for a week or more. Thus, one can use this condition as a base in which to test the growth effects of soluble signals, one by one. For epithelial cell types, type IV collagen is preferred. However, some epithelial cell types will also accept type 1 collagen. Since type 1 collagen substrata are easier and cheaper to prepare, one can try to use them if the cells permit. The clonal growth efficiency of the minimally deviant tumor cells under these various conditions establishes the critical controls:
1.

Negative control-a condition (cells on tissue culture plastic and in SFM) in which the cells have minimal survival and growth. This control will determine the extent to which the components in the basal medium are able to facilitate cell survival and growth. For most epithelia, the SFM condition permits no survival at all. 2. Survival state-a condition (collagen substratum + SFM) in which the cells can survive but show limited growth or no growth. 3. Growthstate/positivecontrols-acondition (SSM +/-collagen)under which the cells will show the highest clonal growth efficiency. It is likely that the minimally deviant tumor cells will grow best on collagen and in SSM.

In all future experiments, replicate plates of cells are plated under the experimental conditions and under the control conditions, respectively. To test for hormone or growth factor requirements, the cells are plated on collagen substrata and in SSM, and then switched after they have attached (3-4 h) to SFM. The cells are seeded at the lowest density at which they will attach and survive when on collagen and in SFM. Usually, purified hormones and growth factors are prepared individually (see Compendium 1) and aliquoted as 1000 x stocks. The initial concentration tried is based on any information available from in vivo or in vitro studies on the relevance of the factor to the cell type of interest. The factor is added to cultures that are under survival conditions: SFM and on a collagen substratum. There should be 3-4 plates for each control or experimental condition resulting in a total of 12-16 control plates plus 3-4 x the number of test conditions. Medium changes are done every

Requirements

for Differentiated

Epithelia

245

3-4 d. The cells are allowed to grow for 1-2 wk. The number of cell colonies (number to attach and survive) and the diameter of the cell colonies (epithelial cells) or cell density (fibroblasts) are measured. Any hormones or growth factors that increase the number of cell colonies or the colony diameter are studied further. Dosage studies are done, again using clonal growth assays, over a 2-3 log range to determine the optimal concentration of the factor in its effects on growth. Once active factors are determined, they form the basis for a new control condition: SFM, collagen substratum, plus any biologically active factors at their optimal concentrations. This condition is then used as the base condition to screen again for hormones and growth factors. All the factors are tested again, since a factor found negative in the first screen can prove positive in subsequent screens because of synergies with another soluble signal. Any new growth factor requirements that are identified are again assessed for their optimal concentrations. The cycle of: 1. Screening for factors on cells under the minimal survival or under the defined base conditions 2. Determining the active factors 3. Determining their optimal concentrations and 4. Adding them at that concentration to the base condition for the next cycle is repeated until no new factors emerge in the screens. Once the active factors are identified, the optimal concentrations of each one are reassessed in the completed defined medium. A serum-free HDM has been made for many cell lines with the known repertoire of hormones and growth factors. However, if a specific cell line still grows better in SSM than in the HDM, then one must purify from serum or screen tissues for unidentified factors (52). One uses the HDM developed for the minimally deviant tumor cells as a starting condition for their normal cellular counterparts. There is one important caveat: normal cells will not survive at the cell densities that tumor cells do. Empirically, one develops an HDM for the normal cells with a starting seeding density of approximately 105/60-mm dish or higher. By contrast, most tumor cells will survive seeding densities of 102-1@/60-mm dish.

1.4. General Rules That Have Emerged porn Studies Developing Hormonally Defined

Media

From the studies in which HDM were developed for a variety of cell types, several rules have been realized:

246

Reid

1. Since the primary determinant of cell attachment and survival is a ma-

2. 3. 4. 5.

6.

7.

8.

trix substratum, development of an HDM for a cell line or for normal cells is more rapidly accomplished with cells cultured on matrix substrata permissive for growth. All cells require transferrin, an iron-carrying protein. Iron is a requirement for the cells polymerases. All cells require insulin or an insulin-like growth factor (insulin, IGFI, IGFII, NGF). All cells require an antioxidant when they are plated in culture. A cheap antioxidant that is very potent is selenium. All cells require a lipid source. For many cell types, linoleic acid (bound to bovine serum albumin as a carrier) is sufficient as a primary lipid source. For other cell types, more complex lipids, such as high density lipoprotein (HDL), must be given. If cells are on tissue culture plastic, most cells attach better and survive longer, if a glucocorticoid (dexamethasone or hydrocortisone) is given. The glucocorticoid is not usually required if the cells are on collagenous or matrix substrata. The hormone requirements for the minimally deviant tumor cells are a subset of the requirements for their normal cellular counterparts (45-47,53). EGF is a common growth factor requirement of normal epithelial cells and not of their neoplastic derivatives. Platelet-derived growth factor (PDGF) is a common requirement that is normal for stromal cells but not for their neoplastic derivatives. Inboth cases,the autonomy of the growth factor is the result of constitutive secretion of the factor or activation of its hormonal pathway intracellularly because of an overexpression or a mutated oncogene (53). Tumor cells will surive much lower seeding densities than the normal cells.

1.5.

Influence

of HDM on Epithelial

Cells in Culture

HDM has been found to select for parenchymal cells even when the cells are on tissue culture plastic (11,12,16-19,24,45,48). This results, within a few days, in cultures that are predominantly the cell type for which the HDM was developed. However, if the cultures are plated onto tissue culture plastic and in HDM, the life span of the primary cultures has been found to be approximately 1 wk, at which time, the cells peel off the plates in sheets. Achievement of longer culture life spans has been found tobe dependent upon using collagenous or matrix substrata in combination with HDM (11,12,39,48).

Requirements

for Differentiated

Epithelia

247

1.6. Insoluble

Signals:

Extracellular

Matrix

For many years, it has been apparent that extracellular matrix is necessary, especially for normal cells to survive and function (12,12,17,28). In recent years, the knowledge of matrix chemistry and biology and of its complexity has become better understood (54-65). It is now recognized that, in order to optimize cell attachment and survival, and to enable the cells to optimally respond to hormones and growth factors, it is essential for the cells to be plated onto substrata of extracellular matrix (21,22,26,66-70). The extracellular matrix is a complex mixture of molecules between and around cells made insoluble by crosslinking. Excellent reviews are available on its known chemistry and functions (55-65). It is important to note that there are many types of matrices (see Tables 3 and 4) with distinct chemical composition. Each cell type secretes and is associated with a specific type of matrix. Furthermore, the matrix chemistry changes when the cell is growing vs quiescent or when it is in some pathological condition (11,22,55-57,71-73). Thus, if you want the cell type of interest to mimic its in vivo counterpart in a specific physiological state, you must identify the matrix chemistry associated with the cell in that physiological state. One is helped by the understanding that there is a paradigm to how all types of extracellular matrix are made, and by the numerous available studies identifying the matrix chemistry in various tissues.

1.7. The Paradigm

in the Construction of Extracellular Matrix

All cells produce an extracellular matrix, and the extracellular matrix in between any given set of cells contains components derived from all the cell types in contact with the matrix. There are many chemically distinct types of extracellular matrices produced by cells (55-65). Although no type of matrix has been studied sufficiently to know all of its components, nevertheless, all matrices consist of collagen scaffoldings (65) to which cells are bound by adhesion proteins (60-64). In association with the collagens, adhesion proteins, and the cell surface are polymers of sulfated amino sugars, glycosaminoglycans (GAGS) that are bound to a protein core to form proteoglycans (PGs) (58,59). The chemically distinct extracellular matrices are readily grouped into types dictated by the type of collagen used as the scaffolding (65,67). Each collagen type is associated with specific adhesion proteins and proteoglycans that are coordinately synthesized and assembled with the collagen. For example, a type I matrix consists of type I collagen, fibro-

248
Table 3 Representative Matrix Types Produced by Cells In Vivo Associated Associated cell Surface Cell(s) anchorage protein proteoglycan(s) receptor producing Fibronectin Fibronectin Fibronectin Dermatan sulfate, chondroitin sulfate Chondroitin sulfate Heparan sulfate, heparin Integrin

Reid

Collagen type Type * Type II Type III

Type IV

Laminin

Heparan sulfate, heparin

Type V Type VI

Fibronectin Fibronectin

Heparan sulfate Heparan sulfate, heparin

All physiological states: fibroblasts Integrin Chondrocytes (cartilage) Integrin Hepatocytes in quiescent state; fibroblasts in association with epithelia Laminin All epithelial receptor cells; endothelial cells; regenerating hepatocytes Integrin Hepatocytes in quiescent state Integrin Hepatocytes in quiescent state

Table 4 Families of Matrix with No Known Collagen Scaffolding Anchorage protein(s) Adhesion molecules (N-CAMs) Fibronectin Fibronectin Fibronectin Complement C3bi Associated proteoglycams) Heparan sulfate, heparin Heparan sulfate, heparin Heparan sulfate, heparin Heparan sulfate, heparin Heparan sulfate, heparin Cell surface receptor
??

Tissue localization (cells producing) Central nervous system (neurons) Bone marrow, lymphatic tissue (platelets) Bone marrow, lymphatic tissue (lymphocytes: other hemopoietic cells) (Myeloid cell lines) Bone marrow, lymphatic tissue, (monocytes, granulocytes, large granular lymphocytes)

Integrin Integrin (VLA-l,VLA-2) (VLA-3,VLA-4) Integrin (VLA-5) Integrin (Mac-l)

(GP/GP)

Requirements

for Differentiated

Epithelia

249

nectin (plus any other anchorage proteins), and chondroitin sulfate proteoglycan or dermatan sulfate proteoglycan. By contrast, a type IV matrix consists of type IV collagen, laminin, and heparan sulfate proteoglycan. There are two tissues for which collagen scaffoldings have not been identified. Hemopoietic cells produce an extracellular matrix that includes a collagen scaffolding when they are part of the bone marrow. During their differentiation into the various hemopoietic cells in the blood, the collagen synthesis is lost. However, they retain the capacity to produce other matrix components: the anchorage proteins and the proteoglycans. Increasingly, these matrix molecules are being found to be of importance to the functions of these cell types (31). Similarly, embryonic neurons of thecentral nervous system produce a matrix that includes a collagen scaffolding (7,74). However, differentiation into adult, mature neurons results in the loss of collagen synthesis. However, these cells are associated, both intracellularly and extracellularly, with proteoglycans and glycosaminoglycans and extracellularly with adhesion molecules (e.g., the N-CAMS). The proteoglycans have been found to be important for regulating various aspects of neuronal functions (7).

1.8.

Rules

for Cultures

on Matrix

Substrata

1. Cells attach in seconds to minutes on an appropriate matrix substratum. Therefore, when plating cells on any matrix substrata, add the cell in sufficient volume of medium to allow equal distribution over the plates. 2. Since matrix is a mixture of proteins and carbohydrates, use DNA or RNA stains rather than protein or carbohydrate stains for assessing plating efficiency or clonal growth efficiency of cells on matrix. 3. Cells on complex matrices (matrigel, biomatrix) do not detach readily. To do growth curves of cells on complex matrices, you may have to scrape the plates, isolate the DNA, and determine DNA content. 4. Antibodies (and many other reagents) may nonspecifically stick to matrices. For antibody staining of cells on matrices, expose the cells to a control antiserum first and then to the specific antiserum. 5. Cells can achieve a very high density if cultured on matrix substrata, especially if proteoglycans or glycosaminoglycans are used. For example, typical saturation densities for normal cells on a loo-mm tissue culture dish are 7-10 x 106;on appropriate matrix substrata, they can be greater than lOa cells/lOO-mm dish. 6. Extraction of RNA or DNA from cell cultures on collagens, proteoglycans, or complex matrices should avoid first detaching the cells enzymatically unless (1) the enzymes are guaranteed to be free of nucleases

250

Reid
and (2) detachment does not result in alteration of the RNA levels. Rather, use guanidine or guanidinum solutions on the cultures and extract everything on the plate, both cells + matrix. Then use a purification protocol that will accommodate the increase in proteins and carbohydrates from the matrix (e.g., centrifuging through cesium chloride).

2. Methods
2.1. Protocols for Culturing Epithelial Cells
Defined

2.1.1. In a Growth 2.1.1.1. Normal Cells

State with Completely Conditions

1. Isolate the cells by standard protocols for the cell type you are using (seeChapter 1, this volume). 2. Use a basal medium that has a low calcium concentration, approximately 0.04 mM. 3. Plate the cells at densities of 3-4 x W/60 mm or at least l@/lOO mm. Use dishes coated with type IV collagen, a medium that contains both the hormones and growth factors used in the HDM, and some serum (2-10%) to permit them to attach and to inactivate any enzymes that might have been used in isolating the cells. You can avoid the serum altogether if you use laminin with the collagen and if you use no enzymes in isolating the cells (or have effective ways of inactivating those enzymes without providing a condition that is toxic to cells). 4. Incubate the cells for 4 h at 37OC or until they are firmly attached. 5. Rinse the plates, removing debris and floating cells, and feed with the serum-free HDM. Change the medium every day, or at the very least, every other day (prepare the HDM fresh). 6. The cells will grow and will survive up to a month under these conditions. 7. If the cells are stem cells from class I (seeFig. l), you can expect to be able to subculture and clone the cells under these conditions. If the cells are from class II, you can expect several rounds of division and then growth arrest. You will not be able to clone or subculture them, or you will have limited ability to do so. Cells from class III will not grow under these conditions, but may survive for some weeks.

2.1.1.2 Tumor Cells

1. Tumor cells are plated onto tissue culture plastic (for more anaplastic

Requirements for Differentiated

Epithelia

251

2. 3. 4. 5.

tumor cells) or on type IV collagen (for more differentiated tumor cells) at densities of lo4 cells/60 or lo5 cells/lOO-mm dishes and into HDM/SSM. After 4 h, the cultures are rinsed and switched into the HDM designed for the tumor cells. The cells will survive and grow under these conditions for 7-10 d. The tumor cells can be subcultured and cloned under the defined conditions. To date, no HDM plus matrix condition is sufficiently defined to permit indefinite maintenance of stock flasks of cells.

2.1.2. In a Fully with Completely

2.1.2.1. Normal Cells

Differentiated State Defined Conditions

1. Isolate the cells as a single-cell suspension by standard methods. 2. Plate the cells at high density,6-8 x 106/100-mm dish, into HDM/SSM and onto one of the following substrata: type IV collagen mixed with laminin for most epithelial and endothelial cells or a fibrillar collagen, ideally type III collagen, mixed with fibronectin for hepatocytes. 3. After 4 h (or whatever length of time it takes to allow the cells to attach), rinse the cells and give them the serum-free HDM retailored to optimize tissue-specific gene expression. Tailor the medium (as described earlier) to suit whichever genes are to be optimally expressed. 4. The medium fed to the cells after 4 h (not the plating medium) must be supplemented with an appropriate species of glycosaminoglycan (at 10-50 pg/mL) or ideally the corresponding proteoglycan (at l-5 pg/ mL). One can mix the GAG or proteoglycan in with the mixture of collagen and adhesion protein; however, there may be some problems in attachment of the cells. For epithelial and endothelial cells, the glycosaminoglycan/proteoglycan will be a species of heparin or ahighly sulfated species of heparan sulfate proteoglycan. For cartilage cells, the glycosaminoglycan/proteoglycan will be chondroitin sulfate or chondroitin sulfate proteoglycan. 5. Note that the cells will not grow, but will survive. They will usually be three-dimensional and highly differentiated. 6. The cultures will survive for perhaps 1 mo and will retain most of their differentiated functions. If you are able to add the proteoglycan, you will have normal transcription rates for most of the tissue-specific genes. If you use the glycosaminoglycan, you will also seenormal or near normal transcription rates of some tissue-specific genes, but not

252

Reid as many genes or to the same extent as that seen with the proteoglycan. The proteoglycans and glycosaminoglycans can be tissue-specific. So, ideally use the proteoglycan or GAG from the tissue being cultured. With the limited availability of GAGS (and even more so for proteoglycans) you will probably have to screen the available ones and use the one that is most active on your cells.

2.1.2.2. Tumor Cells 1. Prepare stock flasks of tumor cells as single-cell suspensions. 2. Plate the cells at high density (6-S x lo6 cells/lOO-mm dish) under the same conditions as those used for the normal cells. 3. The cells will go into growth arrest and, simultaneously, will differentiate to the extent that they are capable of differentiating. Note that, if a tumor cell line is completely silent for a gene, the known culture conditions for differentiation have yet tobe shown to activate such a gene (45). However, if the tumor cells express even tiny amounts of the gene product, the cells can usually be made to express very high levels of that product. Thus far, this effect has been found to be predominantly by posttranscriptional mechanisms such as mRNA stability. However, some limited improvements in transcriptional signals for tissue-specific mRNAs have been observed. 4. The cells will stay in a state of growth arrest for a matter of days (if they are anaplastic tumors) to a few weeks (if they are minimally deviant tumor cells). The well-known ability of tumor cells to secrete enzymes that degrade matrix and/or to overexpress oncogene products (usually products that are the same as a hormone active on the cell or products that activate the hormonal pathways intracellularly) ultimately overrides the growth-arrested/differentiated state caused by the defined matrix and hormonal conditions. The tumor cells will begin to grow in foci, similar to transformation foci, piling up on top of one another. The cells in these foci will be much less differentiated than the growth-arrested cells. If you remove the growing foci of cells and plate them on fresh matrix plates and into fresh medium, they will again go into growth arrest and into a differentiated state.

2.2. Culturing Cells in Defined Medium but with Substrata of Tissue Extracts in Extracellular Matrix

Conditions Enriched

One can use the tailored HDM in combination with a substratum of tissue extract enriched in extracellular matrix. This will replace the defined

Requirements

for Diferentiated

Epithelia

253 There are

collagen, adhesion protein, and proteoglycan combination. several published protocols for preparation of such extracts.

1. Matrigel-a urea extract (75) from EHS tumors, a transplantable mouse embryonal carcinoma that constitutively produces basement membrane components (76). It has remarkable effects on the differentiation of cells (77-79). Commercially, it is sold at great cost. However, you can readily (and much more cheaply) prepare it yourself from the tumor. Tumor stocks are maintained either in C57BL/6 mice or in athymic nude mice (seethe protocol for matrigel). 2. Biomatrix-a tissue extract prepared by extracting the tissue with NaCl solutions, nucleases, and detergents (26,39). The extract can be made from any tissue. It is the only matrix that has been shown to result in very stable, long-term, quiescent cultures of normal epithelial cells (39). It is partially depleted of proteoglycans and glycosaminoglycans, which are restored during the first few days of cell culture (restored by the cells). However, if you need maximal differentiation within a few days of plating the cells, you must add the proteoglycans or glycosaminoglycans to the medium (seethe protocol for biomatrix). 3. Matrix from amnions-amnions are extracted with dilute alkali or detergent and rinsed (80). The matrix on the side of the epithelial surface is a type IV matrix. On the other side (the stromal surface) is a type III matrix. You can choose which surface is needed for your cell type. Studies using this matrix indicate that a given cell type will respond dramatically differently if plated on one side vs the other (82). 4. ECM (extracellular matrix)-cell cultures are extracted with dilute alkali or detergent and rinsed (82). The matrix chemistry depends on the cell type used in culture and the conditions under which the cells were grown. Cells attach readily, grow well, and show improved differentiation (albeit not to the same extent on matrigel or on long-term cultures as on biomatrix) (83).

2.3. Protocols for Preparing and Storing Hormones and Extracellular Matrix Components
2.3.1. Methods of Preparation and Storage Protocols are described in Compendia 1 and 2 for sources and preparation of commercially available hormones and matrix components. They can be used for all cell types, For the specific requirements for each cell type, consult the literature. Since the costs of the hormones and other factors vary with time, they will not be given.

254

Reid

2.3.2. Preparation of l)pe I and lV Collagens Protocols for the preparation of two collagen types, type I and IV, will be given. For other collagen types, see the methods of Miller (65).
2.3.2.1.

Protocol for Preparation

of Qpe I Collagen Substrata

[modified from protocols published by Pitot and associates (54)] 1. Materials a. Several rat tails (preferably about 250 g body wt). b. One large hemostat or common pliers. c. One medium pair of scissors. d. One medium forceps. e. Several loo-mm Petri dishes, or 4-6 in. weighing boats. f. One filtration apparatus consisting of a Buchner funnel linked with two layers of cheesecloth, set in a 2-L sidearm flask. Autoclave entire apparatus. g. One large-mouthed bottle (such as a roller culture bottle) sterilized with a magnetic stir bar. h. Acetic acid (l/1000 v/v) solution. 2. Procedures a. Freeze tails after cutting from rats and keep frozen until use. b. When ready to use, wash the tails slightly in 70% ethanol. c. After tails are thawed, skin them. d. Grasp the base of the tail in one hand, and with the hemostat or pliers in the other hand, grasp the tip of the tail, break at a joint, and slowly pull out the tendons. Place them on a paper towel to dry. e. Repeat step 4 one or two joints at a time until no more tendons can be pulled out. f. The tendons dry rapidly. When dry, cut off the fleshy vertebrae and cut the remaining fibers into 5-4 cm lengths. g. Then with scrubbed hands, rub bundles of fibers between thumb and forefinger to break into individual fibers. Spread in dishes or boats and place under ultraviolet light for 24 h. h, Using the ratio 1 g fibers/300 mL of acetic acid solution, determine a proper amount of solution, and add this to a sterile bottle together with a magnetic stir bar. i. With a flamed forcep, put the sterilized fibers into the bottle with the acetic acid solution. j. Stir moderately for 48 h or more. k. Filter the collagen solution through the cheesecloth using

Requirements

for Differentiated

Epithelia

255

house line vacuum. This process gets rid of any remaining undissolved fibers. 1. Using sterile technique, pour into bottles andlabel as Solution A. 2.3.2.2. Preparation ofPlates Coated with Type I Collagen. Gels are made by combining two solutions: Solution A: collagen in 1 /lOOO acetic acid (prepared earlier); Solution B: 2 parts serum-free medium (10x) to 1 part 0.34N NaOH (with DME, use 6x strength). 1. Set out on trays the desired number of 60-mm (or loo-mm) Falcon plastic tissue culture dishes Put 1.7 mL of Solution A into each dish with a 10 mL pipet. 2. Then add 0.4 mL of Solution B to each dish, a few dishes at a time (with a 2-mL pipet), and swirl until the color (dark pink or lavender) is consistent throughout the dish. 3. The collagen solution gels in a matter of seconds (usually less than 1 min). (If 150-mm dishes are used, add 1.3 mL (2.6 mL) Solution B to 5 mL (10 mL) of Solution A. If 150-mm dishes are used, add 3.0 mL Solution B to 11.5 mL of Solution A. 4. After all gels are made, add 3 mL (7 mL for large dishes) of phosphate buffered saline at pH 6.8 to each dish. This is usually allowed to sit overnight, although this is not absolutely necessary. 2.3.3. Type IV Collagen Based on the protocol by Kleinman and associates (75), which in turn is based on the original procedures by Miller (65). 2.3.3.1. EHS Tumors (Starting Material) 1. EHS tumor cells are grown as transplantable tumors in C57B1/6 mice or in athymic nude mice of any strain. The EHS tumor is known for its unusual property of constitutively secreting a matrix with a chemistry appropriate for basal lamina (type IV matrix). It is now the primary source for most basal lamina components. The protocol given is the one used with transplantable tumors maintained in athymic nude mice. Tumors should be grown subcutaneously on both sides of the animal and allowed to grow to l-3 cm in diameter. For a starting preparation of EHS tumor-derived type IV collagen, use 250-350 g of tumor. Since each tumor removed from the mice weighs from l-2 g (bigger tumors become necrotic), the number of mice needed, carrying tumors on both sides, will be a minimum of 125 and, on average, about 200. From approximately 300 g of tumor, you can get approxi-

256

Reid

mately 20 mg of pure type IV collagen. To get such yields, the animals must be treated with lathyrogen, a reagent that will block cross-linking of the matrix components. The lathyrogen typically used is paminoproprionitrile (BAPN). For reasons not yet fully understood, BAPN alone results in rather limited lathyritic condition with this tumor. Kleinman and associates have found that other pharmacological treatment of the mice, in addition to the BAPN, is necessary. Their method is: mix 100 g of BAPN, 2 g of Iproniozid (Sigma), 0.2 g of pargyline (Sigma) into 300 mL of distilled water. Sterilize by filtration. Then mix the solution with 5 kg of autoclaved chow. Let it sit for 15 min. Serve it cold. 2. The mice should be killed, and the tumors dissected free under sterile conditions. The tumors should then be placed immediately into a cellector (Bellco Glass, Inc.) with a 20 mesh (860 @4) grid and the tumors pressed through the grid into ice-cold, sterile, serum-free medium (e.g., RPM1 1640) containing antibiotics. This step will eliminate the capsule that surrounds the tumor; the capsulecontains type1 collagen and fibrin. 3. Collect the tumor cells (each one is surrounded by a thick band of basal lamina); put into a 50-mL, sterile centrifuge tube and centrifuge for 5 min at 1500 rpm. Obviously, if you have dissected free many tumors, you can pool them into a large container and centrifuge all of them together. 4. Throw away the supernatant, and use the pellet for preparing type IV collagen. You can store the pellets in liquid nitrogen or in a deep freezer at -7OOCuntil you are ready to extract and purify type IV collagen or any other basal lamina component. 2.3.3.2. Extraction and Purification of Type IVCollagen from EHS Tumors. All solutions and centrifugations should be at 4OCunless otherwise indicated. 1. Buffer A g/L 3.4M NaCl 198.7 0.05M Tris-HCl, pH 7.4 6.0 2 mM N-ethyl maleimide (NEM) 0.25 8mM EDTA 2.69 2. Thaw the EHS tumor cells (or if you have just isolated them, use them immediately) and put them in ice-cold Buffer A at a ratio of 1 mL of buffer/g of tumor. 3. Homogenize with a polytron homogenizer for about 2 min and spin

Requirements for Differentiated

Epithelia

257

4.

5.

6.

7.

8.

9.

at 12,000 rpm for 30 min. Throw away supernatant. (This eliminates blood from the tumors.) Homogenize the pellet in Buffer A with the same vol of buffer used in step2,and spin again at 12,000 rpm for 30 min. Again, throw away the supernatant. Resuspend pellets in Buffer B R/2 L 58.4 0.5M NaCl 12.1 0.05M Tris-HCl, pH 7.4 0.5 2mMNEM 5.4 8 mM EDTA Use 3 mL of Buffer B/g of tumor wt (original tumor wt in step 1) and stir overnight. Spin at 12,000 rpm for 30 min. Save supernatants for laminin preparation; save pellets for type IV collagen. Resuspend pellets in Buffer B again. Stir for several hours, and then spin at 12,000 rpm for 30 min. Again save supernatants for laminin preparation and pellets for type IV collagen. (Supernatants can be frozen at -7OOC until ready for purification of laminin.) Extract pellets with Buffer C. Buffer C g/L 2.OM GuHCl (Guanidine hydrochloride) 190 6 0.05M Tris-HCl, pH 7.4 0.25 2mMNEM 2.69 8 mM EDTA Buffer C should be used at ratio of 2 mL/g of tumor (original tumor wt). Stir overnight. Spin at 12,000 rpm for 30 min. Save both supernatants and pellets. Reextract pellet again in Buffer C for several hours. Spin at 12,000 rpm for 30 min. Pool supernatants from first and second extractions. Pool pellets from first and second extraction. Steps 7 and 8 result in extraction of the type IV collagen, but further extractions in increasingly harsher conditions increases the yields. Freeze supernatants at -70C until ready for them (supernatants: go to step 10). Suspend pellet (from second extraction with Buffer C) in Buffer D Buffer D g/L

2.OM GuHCl 0.05M Tris-HCl, pH 7.4

2mMNEM

190
6 0.25

258

Reid

8 mM EDTA
2 mM Dithiothreitol

2.69

(DIT) 0.31 Use 1.5-2 mL of Buffer D/g of tumor (original tumor wt-step 1) and stir overnight. Centrifuge at 12,000rpm for 30 min. Save supernatants and pellets. Extract pellet again overnight with Buffer D, centrifuge and save supernatant and pellet. Extract pellet a third time over up to 48 h, centrifuge and save supernatant. Extract pellet a fourth time. Then discard pellet. Pool supernatants from all extractions with Buffer D. (You will treat the supernatants from extraction with Buffer D the same as those resulting from extraction with Buffer C. However, keep those from Buffer C and Buffer D separate.) Supernatants: go to step 10. 10. Dialyze supernatants (separately) resulting from extraction with Buffer C or D against Buffer E (ratio of at least 1:lO v/v>. Change Buffer E several times. Rotate bags. Buffer E g/l0 L 1.7M NaCl 993.5 0.05M Tris-HCl, pH 7.4 60.5 2mMDTT 3.1 1mM EDTA 3.4 0.1 mM Phenylmethylsulfonylfluoride (PMSF) 10 mL of stock solution (100 mM in 100% ethanol) 11. Spin at 12,000 rpm for 90 min to collect the precipitates from the dialysis. Combine pelleted precipitates from extractions with Buffers C and D. 12. Resuspend pelleted, precipitated material in Buffer D, about 1/lO the original volume, and stir overnight in cold room to get the precipitated material into solution. 13. Spin down any undissolved pellet (usually small) at 16,000 rpm for 30 min. Throw away pellet. Save the supernatant. 14. Determine protein concentration using Bradfords reagent and run a 5-16% reducing gradient SDS gel to check purity. 15. Store supernatant in this solution at 4C (do not freeze). 16. To make gels, dialyze supernatant from 13 vs Buffer F at 4OC. Use l/ 10 ratio (v/v). Change at least twice and rotate bags. Buffer F g/2 L 480 4M Urea 29.2 0.25M NaCl 12 O.,O5MTris-HCl, pH 8.6

Requirements

for Differentiated

Epithelia

17. 18.

19. 20.

2mM DTT 0.32 2mL 0.1 mM PMSF NOTE: If the resulting gels are not washed thoroughly, they are very toxic to cells; we have obtained more consistent results by dialyzing the supernatant against any basal medium, such as RPM1 1640 or DME/12. Determine protein concentration, as above (step 14). Remove collagen solution from dialysis bags and add to plates to give thin coating on culture plates. Put plates at room temperature (or at 37OC). Gel will form within an hour. Sterilize gels on the plates using gamma irradiaton (10,000 rads). Rinse plates repeatedly with sterile, double-distilled water to eliminate the salts, urea, and so on. Culture the cells on the plates as usual.
2.3.4. Tissue Extracts Enriched in Extracellular Matrix 2.3.4.1. Matrigel. Matrigel _is prepared from EHS tumors according .~~ ~.

to the protocol of Kleinman et al. (75,78). 1. Preparation of EHS tumors (starting material). a. EHS tumor cells are grown as transplantable tumors in C57Bl/6 mice or in athymic nude mice of any strain (seesection 2.3.3.1.). b. The mice should be killed and the tumors dissected free under sterile conditions. The tumors should then be placed immediately into a cellector (Bellco Glass, Inc.) with a 20 mesh (860 l..tM>grid and the tumors pressed through the grid into icecold, sterile, serum-free medium (e.g., RPM1 1640) containing antibiotics. This step will eliminate the capsule that surrounds the tumor; the capsule contains type I collagen and fibrin. c. Collect the pooled tumor cells (eachone is surrounded by a thick band of basement membrane); put into a 50-mL, sterile centrifuge tube and centrifuge for 5 min at 1500 rpm. d. Throw away the supernatant and use the pellet for preparing the matrigel. You can store the pellets in liquid nitrogen or in a deep freezer at -7OOCuntil you are ready to extract and prepare the matrigel. For the preparation of matrigel from EHS tumors, all solutions and centrifugations should be at 4OC unless otherwise indicated. 2. Preparation of matrigel from the tumors

260

Reid

a. Thaw the EHS tumor cells (or if you have just isolated them, use them immediately) and put them in ice-cold Buffer A (see extraction of Type IV collagen) at a ratio of 10 mL of buffer/ g of tumor. b. Homogenize with a polytron homogenizer for about 2 min and spin at 12,000 rpm for 30 min. Throw away supernatant. (This eliminates blood from the tumors). c. Suspend the pellets in Buffer A and stir for 1 h at 4C. d. Centrifuge for 20 min at 12,000 rpm at 4OC, and again save the pellets. e. Extract the pellets overnight with 2 M urea in 0.5M Tris-HCl, pH 7.4. f. Centrifuge at 10,000 rpm at 4C for 30 min and save supernatant. g. Extract pellet again, centrifuge, and pool supernatant with that from step 6. h. Dialyze the urea extract against 0.15M NaCl in 0.05M TrisHCI, pH 7.4. i. Centrifuge at 15,000 rpm in a Sorvall for 20 min to remove any insoluble material. j. Store aliquoted supernatant at -20C. 3. To coat dishes, slides, or any substratum a. Add the viscous solution (4OC)onto plates or dishes. Use approximately 1 mL/60-mm dish. The thickness of the gel is dictated by the length of time you want to culture the cells. For unknown reasons, the gels with cells on them gradually thin over time. Once the gel is gone, the cultures die rapidly. b. Bring the plates to room temperature. c. Sterilize the dishes with 10,000 rads of gamma irradiation. The dishes can be stored at 4OCuntil used. d. Once the matrigel is on the plates, you can store the plates at 4OCuntil use. 2.3.4.2. Biomatrix 1. Isolation of biomatrix: biomatrix can be prepared from any tissue, from any animal, and from tissue that is normal or diseased. This protocol is based on the procedures of Reid and associates (26,39). a. Mince tissue and homogenize in a Waring Blender or with a Polytron homogenizer using 10 vol (to 1 vol of mince) of

Requirements

for Differentiated

Epithelia

261

b. c. d. e.

f.

g.

h. i. j.

k. 1. m. n. o.

Buffer G : 34MNaCl+ 0.1 mg/mL soybean trypsin + antibiotics (e.g., penicillin/streptomycin) at concentrations standard for cell culture (all at 4OC; seeAppendix, this vol.). Homogenize thoroughly with 10-15 s pulses. Centrifuge for 20 min at 10,000 rpm at 4OC in a preparative centrifuge. Save the pellets and resuspend in Buffer G. Stir for 1 h at 4C. Centrifuge for 20 min at 10,000 rpm at 4OCand again save the pellets. Repeat the extractions of the pellets with Buffer G until the supernatants after centrifugation are clear and are negative for proteins by Lowry or Bradford assays or by optical density at 280 nm. Rinse pellets twice, each time for 1/2-h, in serum-free basal medium (e.g., RPM1 1640) containing the soybean trypsin inhibitor at 4OC. Suspend pellets in a serum-free basal medium (must contain calcium and appropriate salts for the nucleases; e.g., RPM1 1640) containing nucleases: For each 10 mL vol of pellet, suspend it in 100 mL of solution containing 1.0 mg DNase and 5 mg RNase. Stir at 37OC for 1 h. Centrifuge for 20 min at 10,000 rpm at 4C and save the pellet. Repeat steps 7 and 8 until the supernatants (after centrifugation) give spectrophotometric readings of less than 0.1 at 260 nm. Suspend pellets in 1% sodium deoxycholate in basal medium for 1 h at room temperature. You must prepare the detergent as a 10x stock in distilled water first-it does not dissolve easily; then, disperse it into any basal medium-e.g., RPM1 1640, DME, F12. Centrifuge samples as above and save the pellets. Rinse the pellets (now considered biomatrix) three times in 1/2-h rinses with serum-free basal medium (e.g., RPM1 1640). Suspend the pellets in serum-free basal medium (v/v of l/5). Sterilize with 10,000 rads of gamma irradiation. Store at -7OOCfor long-term storage or at 4C for samples to be used within a month. (Sterilization is imperative for all samples stored at 4C.)

262

Reid

2. To coat dishes, slides or any substratum a. Freeze matrix pellets at liquid nitrogen temperature. b. Pulverize matrix pellets-with a SpexMill Freezer Mill (Metuchen, New Jersey) into powder, keeping the sample at liquid nitrogen temperature. c. When the sample is completely pulverized, allow it to come to room temperature. It will become like paint. Use a simple, Camel hair brush (small artists brush) to paint the pulverized biomatrix onto dishes, slides, and so on. d. Sterilize the dishes with 10,000 rads of gamma irradiation. The plates, dishes, and so on, can be stored at 4OCuntil used. Just prior to use, soak the plates with serum-free basal medium to be used with the cells (e.g., RPM1 1640) all at 37C in a regular CO, incubator.

Acknowledgments
I wish to thank Elaine Halay for helping to collect the information on the commercial sources and preparations of the hormones and matrix molecules. I also wish to thank Rosina Passela for excellent secretarial assistance, and Isabel Zvibel, Elaine Halay, and Maria Lourdes Ponce for helping to edit and proofread the manuscript. This research was supported by a grant from the American Cancer Society (BC-439) and by grants from the National Institutes of Health (CA30117, P3O-CA13330, AM17702). Lola Reid receives salary support through a Career Development Award (NIH CA00783).

References
I. 2. 3. 4. 5. 6. 7. Cunha, G. R. (1976) Epithelial-stromal interactions in development of the urogenital tract. Int. Rev. CytoZ. 47,137-194. Yamada, K. M., ed. (19831 Molecular Mechanisms, Cell Interactions and Deueloprnent (Wiley Interscience, NY). Protero, J. (1980) Control of stem cell proliferation: A density-dependent commitment model. J. Theor. Bicl. 84,725-736. Reid, L. M. and Jefferson, D. M. (1984) Culturing hepatocytes and other differentiated cells. Hqxltology 4,548-559. Fleischmajer, R. and Billingham, R. E., eds. (1968) Epitheliul-Mesenchymalln teructions (Williams and Wilkins, Baltimore). Stoker, M. and Gherardi, E. (1987) Factors affecting epithelial interactions. Ciba Found. Symp. IU, 217-239. Edelman, G. M. (1987) Epigenetic rules for expression of cell adhesion molecules during morphogenesis. Cibu Found. Symp. 125,192-216.

Requirements for Differentiated

Epithelia

8. Clayton, D. F., Harrelson, A. L., and Darnell, J.E.,Jr. (1985lDependenceof liver-specific transcription on tissue organization. Mo2.and Cell. Biol. 5,2623-2632. 9. Guillouzo, A. and Guguen-Guillouzo, C. (1986) Isolated and Cultured Hep&cytes. (John Libbey, Eurotext Ltd. INSERM, Paris). 10. Fraslin, J. M., Kneip, B.,Vaulont, S.,Glaise, D., Munnich, A., and Guguen-Guillouzo, C. (1985) Dependence of hepatocyte-specific gene expression on cell-cell interactions in primary culture. EMBO ].4,2487-2491. 11. Reid, L. M. and Jefferson, D. (1984) Cell culture studies using extractsof extracellular matrix to study growth anddifferentiationinmamma lian cells. Mammalian CeuCulture (Mather, J.,ed.), Plenum, NY, pp. 239-280. 12. Reid, L. M., Abreu, S., and Montgomery, K. (1988) ExtracelIular matrix and hormonal regulation of gene expression. Liver BioZogy andPathology(Arias, I. M., Jakoby, W. B., Pepper, H., Schachter, D., and Shafritz, D. A., eds.), Plenum, NY, chapter39, pp. 717-737. 23. Freshney, R. I., ed. (1985) Animal Cell Culture: A Practical Approach ORI..,NY). 14. Paul, J. (1975) Cell and Tissue Culture (Churchill Livingstone, NY). 25. Jacoby, W. B. and Pastan, I. (1979) Cell culture. Methods in Enzymology, vol. 58. 16. Barnes, D. and Sato,G. (1980)Serum-freeculture: aunifyingapproach. CeIf 22,649655. 17. Barnes, D. and Sato, G. (19841Methods for serum-free culture of cells. Cell Culture Methodsfor Molecular and Cellular Biology ~01s.l-&(Liss, NY). 28. Mather, J.,ed. (1984) Mammalian Cell Culture (Plenum, NY). 29. Boyce, S.T. and Ham, R. G. (1983) Calcium regulated differentiation of normal human epidennal keratinocytes in chemicalIy defined clonal culture and serum-free serial culture. J. Invest. Dermafol. 81,33s-44s. 20. Eckl,P.M.,Whitcomb,W.R.,Michalopoulos,G.,andJirtle,R.L.(1987)EffectsofEGF and calcium on adult parenchymal hepatocyte ProIiferation. 1. CeII Pht@. 132, 363-366. 21. Stanley, J. R. and Yuspa, S.H. (1983) Specific epidermal protein markers modulated during calcium-induced terminal differentiation. J. Cell Biol. 96,1809-1814. 22. Ham, R. and McKeehan, W. (1979) Media and growth requirements. Methods in Enzymology55,44-93. 23. Clay ton, D. F.and Darnell, J.E.,Jr. (1983)Changesinliver-specificcompared to common gene transcription during primary culture of mouse hepatocytes. Mol. andCell. Biol. 3,1552-1561. 24. Jefferson, D. M., Clayton, D. F., Darnell, J. E., Jr., and Reid, L. M. (1984) Posttranscriptional modulation of gene expression in cultured rat hepatocytes. Mol. Cell. Biol. 4,1929-1934. 25. Clayton, D. F., Weiss,M., and Darnell, J. E., Jr. (1985) Liver-specific RNA metabolism in hepatoma cells: Variationsin transcription ratesand mRNAlevels. Mol. Cell. Biol. $2633-2641. 26. Muschel, R., Khoury, G., and Reid, L. M. (1986) Regulation of insulin mRNA abundance and adenylation: Dependence on hormones and matrix substrata. Mol. Cell. Biol. 6,337-341. 27. Pierce, G. B.,Arechaga, J.,and Wells, R. S.(1986) Embryonic control of cancer. Pmg. Clin. Biol. Res.226,67-77. 28. Potter, C. S.,Chadwick, C., Ijiri, K., Tsubouchi, S.,and Hanson, W. R.(19841Therecruitabiltiy and cell cycle state of intestinal stem cells. J. Cell Cloting 2,12t%I40. 29. Dudouet, B., Robine, S., Huet, C., Sahuquillo, M., Blair, L., Coudrier, E., and Lou-

264

Reid
vard, D. (1987) Changes in villin synthesis and subcellular distribution during intestinal differentiation of HT29-18 clones. J. Cell Biol. 105,359-369.

30. 32. 32. 33. 34. 35. 36. 37. 38. 39. 40. 41. 42.

Dexter, T. M., Whetton, A. D., Spooncer, E., Heyworth, C., and Simmons,P. (1985) The role of stromal cells and growth factors in hemopoiesis and modulation of their effects by the src oncogene. J. Cell Sci.3,83-95. Gallagher, J. T., Spooncer, E.,and Dexter, T. M. (19831Role of the cellular matrix in hemopoiesis. I. Synthesis of glycosaminolglycans by mouse bone marrow cell cultures. J. Cell Sci. 63,155-171. Barrandon, Y. and Green, H. (1985) Cell sizeas a determinant of the clone-forming ability of human keratinocytes. Pm. Natl. Acud. Sci. USA 82,5390-5394. Haynes,N. T., Braun, L.,Yaswen, P.,Brooks, M., and Fausto,N. (1984) Isozyme profiles of oval cells, parenchymal cells and biliary cells isolated by centrifugal elutriation from normal and preneoplastic livers. CancerRes.44,332-338. Hixson, D. C. and Allison, J. P. (1985) Monoclonal antibodies recognizing oval cells induced in the liver of rats by N-2 fluorenylacetamide or ethionine in a choline deficient diet. CancerRes.45,3750-3760. Tatematsu, M., Kaku, T., Ekem, J. K., and Farber, E. (1984) Studies on the proliferation and fate of oval cells in the liver of rats treated with Zacetylaminofluorene and partial hepatectomy. Amer. J. Puthol. 114,418430. Hanahan, D. (1985)Heritable formation of pancreatic beta-cell tumors in transgenic mice expressing recombinant insulin/simian virus 40 oncogens. Nature 315, 115122. Power, R. F., Holm, R., Bishop, A. E., Varndell, I. M., Alpert, S.,Hanahan, D., and Polak, J. M. (1987) Transgenic mouse model: A new approach for the investigation of endocrine pancreatic B-cell growth. Gut. 28,121-129. Cunha, G. R. and Donjacour, A. (1987) Strom&epithelial interactions in normal and abnormal prostatic development. Prog. Clin. Biol. Res. 239,251-272. Rojkind, M., Gatmaitan, Z., Mackensen, S.,Giambrone, M. -A., Ponce, P., and Reid, L. M. (1980) Connective tissue biomatrix: Its isolation and utilization for long-term cultures of normal rat hepatocytes. J. Cell Biol. 87,255-263. Michapoulos, G., Cianciulli, H. D., Novotny, A. R., Kligerman, A. D., Strom, S. C., and Jirtle, R. L. (1982)Liver regeneration studies with rat hepatocytes in primary culture. Cancer Res. 42,4673-4682. Germain, L., Noel, M., Gourdeau, H., and Marceau, N. (1988) Promotion of growth and differentiation of ductular oval cells in primary cultures. Cancer Res. 48,
368-378.

Hammond, S. L., Ham, R. G., and Stampfer, M. R. (1984) Serum-free growth of human mammary epithelial cells: rapid clonal growth in defined medium and extended serial passagewith pituitary extract. Proc. Nutl. Acud. Sci. USA 81,54355439. 43. Morrison, R. S.,Sharma, A., devellis, J., and Bradshaw, R. A. (19861Basic fibroblast growth factor supports the survival of cerebral cortical neurons in primary culture. Pm. Nutl. Acud. Sci. USA83,7537-7541. 44. Dodson, M. V., Allen, R. E.,and Hossner, K. L. (1985) Ovine somatomnedin, multiplication stimulating activity, and insulin promote skeletal muscle satelite cell proliferation in vitro. Endocrinology 117,2357-2363. 45. Gatmaitan, Z., Jefferson, D., Ruiz-Opazo, N., Leinwand, L., and Reid, L. M. (1983) Regulation of growth and differentiation of a rat hepatoma cell line by the synergistic interactions of hormones and collagenous substrata. I. Cell. Biol. 97,1179-l 190.

46. Reid, L. M.,Stiles, C., Saier, M., Jr., and Rindler, M. (1979) Growthof nontumorigenic

Requirements

for Differentiated

Epithelia

47. 48.

49. 50.

51. 52. 53. 54. 55. 56. 57. 58. 59. 60. 61. 62. 63. 64. 65. 66. 67. 68.

cells in millipore diffusion chambers implanted in mice: Implications for in viva growth regulation. Cancer Res. 20,1467-1473. Cherington, P. V., Smith, B. L., and Pardee, A. B. (1980) Loss of epidermal growth factor requirement and malignant transformation. Proc. N&l. Acud. Sci. USA 76, 3937-3941. Enat,R., Jefferson, D. M., Ruiz-Opazo,N., Gatmaitan, Z., Leinwand, L. A., and Reid, L. M. (1984) Hepatocyte proliferation in vitro: Its dependence on the use of serumfree hormonally defined medium and substrata of extracellular matrix. Pm. Nufl. Acud. Sci. USA 81,1411-1415. Taub, M. (1984) Growth of primary and established kidney cell cultures in serumfree media. Cell Culture Methods for Molecular and Cell Biology, vol. 3 (Barnes, D., Sirbasku, D., and Sato, G., eds.) Liss, NY, pp. 3-24. Lippman, M. E. (1984) Definition of hormones and growth factors required for optimal proliferation and expression of phenotypic responses in human breast cancer cells. Ceil Culture Methods for Moleculur and Cell Biology, vol. 2 (Barnes, D., Sirbasku, D., and Sato, G., eds.) Liss, NY, pp. 183-200. Deleted in proof. Hoshi, H. and McKeehan, W. L. (1984) Brain and liver cell-derived factors are required for growth of human endothelial cells in serum-free culture. Proc. N&Z. Acud. Sci. USA 81,6413-6417. Waterfield, M. D. (1983) Platelet-derived growth factor is structurally related to the putative transforming protein ~28~ of simian sarcoma virus. Nature 304,35-39. Michalopoulos, G. and Pitot, H. C. (1975) Primary culture of parenchymal liver cells on collagen membranes. Exp. Cell Res. 94,70-78. Piez, K. A. and Reddi, A. H., eds. (1984) Extracellular Matrix Biochemistry (Elsevier, NW. Ruohslahti, E. and Pierschbacher, M. D. (1988) Molecular basis of cell-extracellular matrix interactions. Liver Biology and Puthobiology, 2nd Ed. (Arias, I. M., Jakoby, W. B., Popper, H., Schachter, D., and Shafritz, D. A., eds.) Raven, NY, pp. 739-751. Mecham, R. I. ed. (1986) Biology of Exfrucellulur Matrix, vol. 1, Academic, NY. Evered, D., Hascall, V. C., and Whelan, J. (eds.) (1986) Functions of the Proteoglycuns (Ciba Foundation Symposium, Wiley, NY). Fransson, L. (1987) Structure and functionof cell-associated proteoglycans. TIBS 12, 406-411. Hynes, R. 0. (1987) Integrins: A family of cell surface receptors. CeZI 48,549-554. Ruoshlahti, L. E. and Pierschbacher, M. D. (1986) Arg-Gly-Asp: A versatile cell recognition signal. Cell 44,517,518. Buck, C. A. and Horwitz, A. F. (1987) Cell surface receptors for extracellular matrix molecules. Ann. Rev. Cell Biol. 3,179-205. Hynes, R. 0 (1986) Fibronectins Sci. Amer. 254,42-51. von-der-Mark, -K., and Kuhl, U. (1985) Laminin and its receptor. Biochim Biophys. Actu. 823,147-X0. Miller, E. J. (1986) Collagen types: Structure, distribution and functions, in CoElagen: Biochemistry, vol. 1 (CRC Press), pp. 134-156. Martin, G. R. and Kleinman, H. K. (1981) Extracellular matrix gives new life to cell culture. Hepufology 1,264-266. Yamada, K. M. (1983) Cell surface interactions with extracellular materials. Ann. Rev. Biochem. 52,761-799. Narita, M., Jefferson, D. M., Miller, E. J., Clayton, D. F., Rosenberg, L. C., and Reid, L. M. (1985) Hormonal and matrix regulation of differentiation in primary liver cul-

266
tures, in Gmwth and Diffkentiution

Reid
of Cells in Defined Environments (Murakami, H., Yamane, I., Hayashi, I., Mather, J. P., Barnes, D. B.,and Sato, G. H., eds.) Springer Verlag, NY, pp. 89-96. Spray, D. C., Fujita, M.,Saez,J.C., Choi, H., Watanabe, T., Hertzberg, E.,Rosenberg, L. C., and Reid, L. M. (1987) Proteoglycans and glycosaminoglycans induce gap junction synthesis in primary liver cultures. 1. Cell BioI. 105,541-551. Fujita, M, Spray, D. C., Choi, H., Saez,J. C., Watanabe, T., Rosenberg, L. C., Hertzberg, E. L., and Reid, L. M. (1987) Glycosaminoglycans and proteoglycans induce gap junction expression and restore transcription of tissue-specific mRNAs in primary liver cultures. Hepatology1, Is-9s. Sell, S.and Ruoslahti, E. (1982) Expression of fibronectin and laminin in the rat liver after partial hepatectomy,during carcinogenesis,and in transplantable hepatocellular carcinomas. J. Natl. Can. Inst. 69,1005-1014. Carlsson, R., Engvall, E., Freeman, A., and Ruoshlahti, E. (1981) Laminin and fibronectin in cell adhesion: Enhanced adhesion of cells form regenerating liver to laminin. Proc. Natl. Acad. Sci. USA 78,2403-2406. Fedarko, N. S. and Conrad, H. E. (1986) A unique heparan sulfate in the nuclei of hepatocytes; structural changes with the growth state of the cells. J. Cell Biol. 102, 587-599.

69.
70.

71.

72.

73.

Chuong, C. M., Crossin, K. L., and Edelman, G. M. (1987) Sequential expression and differential function of multiple adhesion molecules during the formation of cerebellar cortical layers. J. Cell BioZ. 104,331~342. 75. Kleinman, H. K., McGarvey, M. L., Hassell, J. R., and Martin, G. R. (1983) Formation of a supramolecular complex is involved in the reconstitution of basement membrane components. Biochemistry 22,4969-4974. 76. Orkin, R. W., Gehron, P.,McGoodwin, E. B.,Martin, G. R.,Valentine, T., and Swarm, R. (1977) A murine tumor producing a matrix of basement membrane. J. Exp. Med.
74. 145,204-220. 77.

78.

79.

80.

81, 82.

83.

Hadley, M. A., Byers, S. W., Suarez-Quian, C. A., Kleinman, H. K., and Dym, M. (1985) Extracellular matrix regulates Sertoli cell differentiation, testicular cord formation, and germ cell development in vitro. J. CeZZ Biol. 101,1511-1522. Kleinman, H. K., McGarvey, M. L., Hassell, J. R., Star, V. L., Cannon, F. B., Laurie, G. W, and Martin, G. R. (1986) Basement membrane complexes with biological activity. Biochemistry 25,312-318. Martin, G. R., Kleinman, H. K., Terranova, V. P., Ledbetter, S., and Hassell, J. R. (1984) The regulation of basement membrane formation cell-matrix interactions by defined supramolecular complexes. Ciba Found. Symp. 108,97-212. Liotta, L. A., Lee, C. W., and Morakis, D. J. (1981) New method for preparing large surfaces of intact basement membrane for tumor invasion studies. CancerLeft. 11, 141-152. Madri, J. A. and Williams, S.K. (1983) Capillary endothelial cells cultures: Phenotypic modulation by matix components. 1. Cell Biol. 97,153-X5. Gospodarowicz, D. (1984) Preparation of extracellular matrices produced by cultured bovine cornea1 endothelial cells and PF-HR-9 endodermal cells: their use in cell culture. CeZZ CultureMethods for Molecular and CeZZ BioZogy (Barnes, D., Sirbasku, D., Sato, G., eds.) Liss, NY, vol. 1, pp. 275-293. Vlodavsky, I., Levi, A., Lax, I., Fuko, Z., and Schlessinger, J. (1983) Induction of cell attachment and morphological differentiation in a pheochromocytoma cell line and embryonal sensory cells by the extracellular matrix. Dev. BioZ. 93,285-300.

Requirements for Differentiated

Epithelia

267

Compendia of Commercially Available Hormones and Matrix Components

Sterilization Protocols: There are four protocols listed in these compendia and used for either hormones or matrices. In the description of preparation of the individual hormones, growth factors, or matrix molecules, the sterilization protocol is indicated by a number correlating with the following: 1. Sterilize by filtration through a 0.22~pm filter. 2. Sterilize by filtration through a filter with low protein binding properties (e.g., Millex-GV). 3. Sterilize by gamma irradiation, 10,000 rads. 4. Sterile as prepared in the reconstituted form. 5. For protoglycans and glycosaminoglycans, sterility is achieved by ethanol precipitation. Prepare samples in sterile 0.15M sodium acetate in a sterile tube. For best results, stock solutions should be at concentrations of 10-20 mg/mL. However, lower concentrations will also precipitate, but there will be some loss in yields. Add 100% ethanol at a ratio of 21 (v/v>, which results in precipitation of the proteoglycan or glycosaminoglycan. Leave the sample at 4C overnight. Centrifuge sample at 5000 rpm. Wash pellet 2x with 100% ethanol to eliminate the sodium acetate. Allow the pellet to air dry under vacuum (or more slowly in air) under sterile conditions. Weigh the sample, again under sterile conditions. Dissolve the sample at the desired concentration in sterile aqueous solutions. Compendium 1 Commercially Available Hormones and Growth Factors ACTH (adrenocorticotropic hormone) Sources: Armour Pharmaceuticals, NM, Sigma, Calbiochem, Serva, ICN, Biomedicals, Inc. Storage: Freezer at -20C Reconstitution: Dissolve in PBS Sterilization protocol: 2 Typical concentration: 5 pU/mL Ceruloplasmin Sources: Sigma, Calbiochem Serva Reconstitution: Dilute in PBS Storage: Freezer at -20C
(continued)

268 Compendium 1 (confinued)

Reid

Sterilization protocol: 2 Typical concentration: 0.05 U/mL Cl apric sulfate Sources: Sigma, AESAR (Johnson Matthey), Serva Storage: Dry-room temperature Solution: Freezer at -20C Sterilization protocol: 1 Typical concentration: 10-6-10-2M DHT-Dihydrotestosterone Sources: Sigma, Collaborative Research Storage: Dry-room temperature Solution: Refrigerator (40 Reconstitution: Seeprotocol for glucocorticoids Sterilization protocol : 4 Typical concentration: 10-7-10-1*M EGF (Epidermal Growth Factor) Sources: Collaborative Research, Boehringer Mannheim Biochemicals, Gibco, Calbiochem, Chemicon, BTI, IBT, Serva, Upstate Biotechnology, Inc., ICN Reconstitution: Bring vialed product to room temperature. Add 1 mL of sterile, distilled H,O and mix. Solution: Contains 100 yg/mL EGF in 0.55M NaCl Storage: Lyophilized EGF should be stored at 4C, reconstituted products store at -20C (stable for 3 mo) Sterilization protocol : 4 Note: Do not freeze and thaw more than 4 times; therefore, aliquot accordw$Y Estradiol-17B Sources: Sigma, Calbiochem, BTI Storage: Dry-room temperature or refrigerator Solution: Room temperature or refrigerator Reconstitution: Seeprotocol for glucocorticoids Sterilization protocol : 4 Typical concentration: 10-7-10-10M Endothelial cell growth factor Sources: Gibco, Chemicon, Serva, ICN, Collaborative Research, Sigma Storage: Lyophilized powder (refrigerator 4C), solubilized (freezer -20C) Reconstitution: Bring to room temperature and add l-5 mL of sterile serumfree media. Aliquot in plastic tubes. Solution (commercial preparations): may contain streptomycin sulfate and sodium chloride Sterilization protocol : 2 Typical concentration: 50-100 Fg/mL
(continued)

Requirements

for Differentiated

Epithelia

269

Compendium 1 (continued) FGF (Fibroblast Growth Factor) Sources: Collaborative Research, Boehringer Mannheim Biochemicals, R & D Systems, Inc., Gibco, Calbiochem, Chemicon, BTI, Serva, ICN Reconstitution: Bring vialed product to room temperature. Add 1 mL sterile, distilled H,O and mix Solution (commercial preparations): Contains 10 pg/mL FGF, 100 pg/mL albumin in 0.006M NaCl, 0.002M Na,HPO, Sterilization protocol : 2 Storage: Lyophilized 4C, reconstituted product should be stored at -20C Typical concentration: 0.2-2 ng/mL Note: Do not freeze and thaw more than 4 times FSH (Follicle Stimulating Hormone) Sources: Sigma, Calbiochem, Serva Storage: Dry-refrigerator Solution: Freezer at -20C Reconstitution: In protein containing media, e.g., 1% bovine serum albumin solution in PBS Sterilization protocol : 2 Typical concentration: 0.4 -10 pg/mL GHL (Glycyl-histidyl-lysine) Sources: Sigma, Calbiochem, BTI, Serva, ICN Storage: Lyophilized 4C, Solubilized (-20C) Reconstitution: Dissolve in PBS or tissue culture medium Sterilization protocol : 2 Typical concentration: 10-200 ng/mL Glucagon Sources: Sigma, Calbiochem, Serva Storage: Dry-refrigerator Solution: Freezer at -20C Reconstitution: Dissolve in basic H,O (approximately pH 9) at low concentration Sterilization protocol : 2 Typical concentration: 0.5 pg/mL Glucocorticoids (hydrocortisone, dexamethasone) Sources: Sigma, Calbiochem Reconstitution: Make a solution of 10w3M in 95% ethanol, place in air-tight container (a glass scintillation vial is best), and mark the level of the liquid. No evaporation should occur. Store at room temperature. Make a fresh dilution to 10m4M in 95% ETOH every 3 wk. This is a 1000x solution and is the immediate stock. It can be added to media in the bottle of media itself or to the individual plate. When adding to plate, add after the media has been placed and before the cells are added. If cells are already attached to the plate, then

270 Compendium 1 (continued)

Reid

tilt the plate to increase the depth of the media and drop the stock solution onto the top of the deepest part of the liquid. In this way, no cells are damaged by the ethanol. A fresh stock of 1WM should be made up every 3-4 mo. Storage: Dry-room temperature; solution-refrigerator (4OC) Sterilization protocol: 4 Typical concentration: 104-10-7M High Density Lipoprotein (HDL) Source: Sigma Reconstitution: Dissolve in aqueous solutions Sterlization: 2 Storage: Refrigerator Typical concentration: 10 pg/mL Insulin 500 mg/bottle Sources: Sigma, Boehringer Mannheim Biochemicals, Gibco, Calbiochem, BTI Serva, ICN, Collaborative Research Reconstitution: Make a solution of 2 mg/mL in O.OlM HCl. Store in refrigerator (up to 6 wk). Sterilize by filtration through a swinney filter that minimizes protein sticking. Just before use, dilute to 500pg/mL (100x) with PBS. If any precipitate occurs, discard Sterilization protocol: 2 Typical concentration: l-5 pg/mL IGF I (Insulin-like Growth Factor or Somatomedin C) Sources: Sigma, Boehringer Mannheim Biochemicals, Iemcera, Chemicon, Incell, Amgen Storage: Freezer at -20C Reconstitution: Make up in buffer at pH 7.6 (stable for 4-6 mo if frozen) Sterilization protocol: 2 Typical concentration: l-50 ng/mL IGF II Source: Bachem Storage: Freezer at -20C Reconstitution: Dissolve in a aqueous buffer at pH 8.0 Sterilization: 2 Typical concentration: 5 ng/mL Linoleic acid sodium salt Sources: Sigma, Serva, Collaborative Research Storage: Freezer at -20C Reconstitution: Dissolve 1 g of linoleic acid in a solution containing 1% bovine serum albumin (fattv acid free)
(continued)

Requirements

for Differentiated

Epithelia

271

Compendium 1 (continued) Sterilization protocol: 2 (necessary because of the BSA) Typical concentration: 5-10 &mL LRF or LHRF (Luteinizing Hormone Releasing Factor) Sources: Calbiochem, Boehringer Mannheim Biochemicals, Serva, Collaborative Research, Accurate, Chemicon Storage: Freezer at -20C Reconstitution: Dissolve in PBS Sterilization protocol: 2 Typical concentration: 10 ng/mL LH (Luteinizing Hormone) Sources: NIH, Calbiochem Storage: Freezer at -20C Reconstitution: Dissolve in PBS Sterilization protocol: 2 Typical concentration: l-5 pg/mL MSA (Multiplication Stimulating Activity) (an extract containing both insulin-like growth factor I and II) Sources: Collaborative Research, Sigma, Serva Storage: Lyophilizedmaterial-refrigerator (4C);reconstituted-freezer-20C Reconstitution: Bring to room temperature; add 1 mL of 1M acetic acid or sterile basal (serum-free) medium containing 0.25-l mg of bovine serum albumin

Sterilization protocol: 4 (if commercial preparation); 2 (if lyophilized material not sterile) Typical concentration: 8-250 ng/mL NGF (Nerve Growth Factor) 10 pg/mL Sources: Collaborative Research, Boehringer Mannheim Biochemicals, Calbiochem, Chemicon, Biomedical Technology, Inc., Serva, ICN, Sigma Storage: 4C for lyophilized NGF and -2OOCfor the solubilized form Reconstitution: Dissolve the vialed product with 1 mL H,O. Sterilize by filtration Sterilization protocol: 2 Typical concentration: 100 ng/mL Note: NGF absorbs to glass and plastic surfaces. Preparing the NGF in a BSA solution of at least 1 mg/mL will minimize nonspecific absorption. Do not freeze and thaw more than 4 times PDGF ( Platelet Derived Growth Factor) Sources: Imcera, Gibco, Boehringer Mannheim Biochemicals, R & D Systems, Chemicon, PDGF, Inc., ICN, Collaborative Research, Sigma
(continued)

272 Compendium 1 (continued)

Reid

Storage: Lyophilized (4C); solubilized (-20C) Reconstitution: Dilute in serum-free culture medium Sterilization protocol: 2 Typical concentration: 2-8 ng/mL P ogesterone 5 g/bottle Sources: Sigma, Calbiochem, Serva, Collaborative Research Storage: Room temperature or refrigerator Reconstitution: Seeprotocol for glucocorticoids Sterilization protocol: 4 Typical concentration: 10-7-10-10M Prolactin (Luteotropic Hormone) Sources: Sigma, Calbiochem, Accurate Storage: Freezer at -20C Solution: Refrigerate Reconstitution: Dissolve in 100% ethanol. Use glass scintillation vial if possible (it seals very well and prevents evaporation) Sterilization protocol: 2 Typical concentration: 10 ng/mL PTH (Parathyroid Hormone) Sources: Calbiochem, Collaborative Research, York Biological International Storage: Freezer at -20C Reconstitution: Dissolve in PBS Sterilization protocol: 2 Typical concentration: 0.5 ng/mL Selenous Acid Sources: AESAR (Johnson Matthey Chemicals Ltd., England), BTI Storage: Dry-room temperature; Solution: Refrigerate (4C) Reconstitution: Dissolve in sterile, distilled water Sterilization protocol: 1 Typical concentration: 10-8-10-7M Somatostatin (GH Releasing Inhibiting Factor) 0.25 mg/bottle Sources: Sigma, Boehringer Mannheim Biochemcical, Calbiochem, Serva, Collaborative Research Storage: Freezer at -20C Reconstitution: Dissolve in PBS Sterilization protocol: 2 Typical concentration: 10 ng/mL Somatotropin (human growth hormones) Sources: Sigma, Calbiochem, Chemicon, Boehringer Mannheim
(continued)

Requirements

for Differentiated

Epithelia

273

Compendium 1 (continued) Storage: Dry-refrigerate (4C) Solution: Freezer (-20C) Reconstitution: Put in PBS and titrate with O.OlN HCl to dissolve Sterilization protocol: 2 Typical concentration: ng/mL T3 (Triiodo-Thyronine) 100 mg/bottle Sources: Sigma, Calbiochem, Collaborative Research Storage: Freezer at -20C (up to 6 mo) Reconstitution: Make up a 10-2-104M solution in 0.02N NaOH Sterilization protocol: 2 Typical concentration: 10-g-lO-*lM Transferrin 100 mg/bottle Sources: Sigma, Boehringer Mannheim Biochemicals, Gibco, Calbiochem, Chemicon, BTI, Serva, ICN, Collaborative Research Reconstitution: Make solution of 5 mg/mL in PBS. Store this in 0.5-l mL aliquots in freezer. When needed, dilute to 500 pg/mL (100x). Do not thaw and freeze more than 4 times Sterilization protocol: 2 Typical concentration: 5 pg/mL Zinc sulfate Sources: AESAR (Johnson Matthey Chemicals Ltd., England), Serva, Sigma Storage: Dry-room temperature Solution: Refrigerate Reconstitution: Make solution in sterile, distilled water Sterilization protocol: 1 Typical concentration: 10-g-lO-loM

274

Reid

Compendium 2: Commercially Available Matrix Components Adhesion proteins Fibronectin (human) Sources: Collaborative, Gibco, Upstate Biotechnology, Accurate, Biomedical Technologies Storage: Lyophilized-refrigerator (4OC) Solution: Freezer (-20C) Reconstitution: Allow to come to room temperature, add sterile, distilled water and allow to solubilize for 30 min. Use plastic pipet to transfer to plastic tubes for storage or to plastic plates Sterilization protocol: 4 Typical concentration: 1-2 pg/cm2 Laminin (mouse) Sources: Gibco, Iemcera, Serva, ICN, Biomedical, Sigma, Collaborative Research Storage: Freezer at -20C Reconstitution: Thaw the solution; make a lo-20 pg/mL dilution in sterile culture medium. Add to plates at 1-2 pg/cm2. Allow to attach to plates for 45 min at room temperatue. Then add cells Sterilization protocol: 4 Typical concentration: 1-2 pg/cm2 Human Vitronectin Source: Iemcera Storage: Freezer at -20C Reconstitution: As for fibronectin Sterilization protocol: 4 Typical concentration: l-2 pg/cm2 Collagens Type I Collagen (rat tail; bovine achilles tendon) Sources: BTI, Serva, ICN, Collagen Corp., Gibco, Seikagaku America, Inc. Storage: Refrigerator (4C) Preparation for plates: Seeprotocol section 2.3.2. Sterilization protocol: 3 Typical concentration: 1 mg/60-mm plate for gels Type IV Collagen-Mouse, human placenta Sources: Gibco, Collaborative, Seikagaku America, Inc. Storage: Refrigerator (4OC) Preparation for plates: Seeprotocol section 2.3.2. Sterilization protocol: 3

Requirements

for Differentiated

Epithelia

275

Compendium 2 (continued) Typical concentration: 10-100 pg/60-mm dish Note: amount required can be quite variable. A thin coating can be achieved with 10-20 pg/plate. However, for some cells, thicker coatings are necessary. Increase the thickness of the layer as desired Glycosaminoglycans Heparins-bovine, porcine, ovine; from intestine or lung Sources: Made by Hepar Industries (Madison, Wisconsin) and distributed by various companies including Sigma, Serva, Accurate Chemical Storage: Room temperature for powder Solution: Refrigerator (4C) Reconstitution: PBS Sterilization protocol: 5 Typical concentration: 20-50 pg/mL Chondroitin Sulfates-whale, shark; from cartilage Sources: ICN Biologicals, Serva, Sigma Storage: Refrigeration for solution, room temperature for powder Reconstitution: PBS Sterilization protocol: 5 Typical concentration: 20-50 pg/mL Dermatan Sulfates-bovine, porcine; from cartilage Sources: ICN Biologicals, Serva, Sigma Storage: Refrigeration for solution, room temperature for powder Reconstitution: PBS Sterilization protocol: 5 Typical concentration: 20-50 pg/mL Heparan Sulfates-bovine, kidney and intestine Sources: ICN Biologicals, Serva, Sigma Storage: Refrigeration for solution, room temperature for powder Reconstitution: PBS Sterilization protocol: 5 Typical concentration: 20-50 pg/mL Hyaluronates -human, bovine, porcine; from umbilical cords Sources: ICN Biologicals, Serva, Sigma Storage: Refrigeration for solution, room temperature for powder Reconstitution: PBS Sterilization protocol: 5 Typical concentration: 20-50 pg/mL
(cmti?iuf?d)

276 Compendium 2 (continued) Matrigel

Reid

Sources: Serva, Collaborative Research Storage: Freezer at -20C; just prior to usage, thaw and maintain at 4C. Preparation for plates: add viscous solution (4C) to plates. Bring to 37C. At this temperature gels will form. Sterilization Protocol (seeTable 4 for detail): 4 (if bought); 3 (if made yourself). ECM Sources: International Biotechnology, Accurate Chemicals Storage: Refrigerate (4C). Preparation for plates: sold as coated plates. Sterilization Protocol (Table 4): 4 (if bought); 3 (if made yourself). Matrigel is prepared from EHS tumors according to the protocol of Kleinman et al.
(75,78X

Chapter 22 Culturing Primitive Hemopoietic Cells

Long-Term Mouse Marrow Cultures and the Establishment of Factor-Dependent (FDCP-Mix) Hemopoietic Cell Lines
Elaine Spooncer and T. Michael Dexter
1. Introduction
The maintenance of normal primitive hemopoietic cell types in culture for long periods can be achieved either by culture of hemopoietic cells in association with bone marrow-derived stromal cells, as in long-term cultures, or by establishing factor-dependent primitive hemopoietic cell lines (FDCP-Mix cell lines). The long-term culture technique and a reproducible method for establishing FDCP-Mix cell lines will be described here (1,Z). In the stromal cell-dependent long-term cultures a wide range of hemopoietic cell types are maintained for several months, including the stem cells (as recognized by the CFU-S assay (3) and the ability of the cells 277

278

Spooncer and Dexter

to rescue potentially lethally irradiated mice), mature cell types of the myeloid lineage, and all the committed progenitor cell types. Cloned factor-dependent cell lines can be derived from long-term cultures and the method described here produces cells that maintain a homogeneous primitive nature under certain culture conditions and yet can be induced to mature in response to normal regulatory influences. Clearly, long-term in vitro growth of primitive hemopoietic cells is valuable for studying the maintenance of homeostasis in the hemopoietic system in which both stromal cells and hemopoietic growth factors (e.g., interleukin 3, IL3) are thought to play important roles. These cultures are also useful for investigating the biochemistry of the response of hemopoietic cells to biological regulators, the molecular biology of the nature of stem cells and for testing the response of hemopoietic cells to drugs, chemicals, and other insults.

2. Materials
1. Fischers medium supplemented with sodium bicarbonate (1.32g/L), penicillin (500 p/m), streptomycin (50 pg/mL) and L-glutamine (7 mL/L of 200 mM). Store at 4OCand use within 2 wk after the addition of the L-glutamine. 2. Horse serum: each batch must be pretested for its quality to support long-term bone marrow cultures. We have used batches from Northumbria Biologicals Ltd. and Flow Laboratories. Store at -2OOC until use and then keep the aliquot in use at 4OCfor up to 10 d. 3. Hydrocortisone sodium succinate. Make a stock solution in Fischers medium and sterilize through an 0.22 p filter. For use dilute to lOAM and store at -20C. 4. Complete growth medium made up in 100 mL aliquots containing 20-25% v/v horse serum (depending on batch quality) and 1% v/v lOAM hydrocortisone stock (final concentration WM) in Fischers medium. 5. Donor mice, 8 wk or older.

2.1. Additional Materials for Establishing CZonaZ, Factor-Dependent (FDCP-Mix) Cell Lines
3I. NIH-3T3 producer cell line for the 2-1 recombinant syc virus grown as a Moloney murine leukemia virus pseudotype (41, [SK (MoMuLV) producer cells]. The SYC (MoMuLV) has been categorized in this institute to be used under the good microbiological practice Code of Practice by the local Advisory Committee on Genetic Manipulation.

Culturing

Primitive

Hemopoietic

Cells

2. Dulbeccos Modification of Eagles Medium supplemented with 10% v/v newborn calf serum (DMEM 10% NBCS): culture medium for SIC (MoMuLV) producer cells. 3. Radiation source to irradiate producer cell cultures at a dose to kill all producer cells (for example, 20 Gy delivered at 84 Gy/min by linear accelerator). 4. Source of IL3: filtered, conditioned medium (CM) from Wehi-3bleukemic cell line established at the Walter and Eliza Hall Institute, cultures chemically purified IL3 from Wehi-3b CM, or recombinant IL3. 5. Semisolid media for plating cells in the granulocyte-macrophage colony-forming cell type assay, e.g., Noble Agar final concentration 0.3% or methyl cellulose. For more details about working with semisolid media seeChapters 26-28 in this volume.

3. Methods 3.1. Long-Term Mouse Bone Marrow Cultures

1. Assemble the tissue culture equipment and reagents in a microbiological safety cabinet. Prepare the complete growth medium, 10 mL per long-term culture, in 100 mL aliquots. 2. Kill donor mice by cervical dislocation. One femur (approximately 1.5 -2 x 10 nucleated cells) is used for each long-term culture. Remove both femora from each donor, removing as much muscle tissue as possible and cutting first below the knee joint and then at the hip joint. Place the bones in a Petri dish containing a small volume of Fischers medium on ice. If mice are in short supply, then tibiae can also be removed. Use two tibiae to make one long-term culture. The time between killing donor mice and flushing the cells into growth medium should not exceed 15-20 min. Therefore, the speed at which you can work dictates how many donor mice to kill at once. A competent operator should be able to work in batches of five mice (i.e., 10 cultures). 3. Take the femora to the microbiological safety cabinet. Using sterile gauze swabs, clean off any remaining muscle tissue from each bone. Carefully cut off the ends of the femora and avoid splintering the bone (Fig. 1). Insert the tip of a 21 gage needle and 1 mL syringe into the knee end of the bone. For an average sized mouse (approximately 25 g) the needle tip should fit firmly into the bone cavity. If the donor mice are very large or small a slightly larger or smaller needle may be required to give a good fit. Hold the syringe and needle so the bone is just below

280

Spooncer and Dexter

1 OOml. growth Femur

medium

33C

5%

CO,

in air

10 ml marrow

suspension

Fig. 1. Preparationof long-term bone marrow culture.

the level of the medium and flush medium through the bone 5-10 times until the bone shaft is empty of cells and looks white. Repeat with nine more bones into the same 100 mL aliquot of medium. It is not necessary to make a single cell suspension, but do ensure that whole marrow plugs are broken up into fragments. Dispense 10 mL aliquots of the marrow cell suspension into ten tissue culture flasks. Connect a plugged sterile pipet to the 5% CO, in air supply line and turn on the gas at a low pressure. Gas the cultures by inserting the tip of the pipet into the air space of the flask for a few seconds, cap the flask firmly and place the cultures horizontally in a 33C incubator (preferably nonhumidified). 4. The cultures are routinely fed at weekly intervals by removing half the growth medium (5 mL) and replacing it with fresh medium. Gently agitate the flasks to suspend non-adherent cells uniformly in the growth medium and carefully remove 5 mL, taking care not to disturb the adherent cells with the tip of the pipet. Tip the flask sideways and add 5 mL fresh medium down the side of the flask so it does not flow directly onto the adherent cells. Gas the flask gently and replace the top. The feeding routine can be altered if required and the cultures will tolerate a complete media change weekly or more frequent feeding.

Culturing

Primitive Hemopoieti

281

Fig. 2. (a) Inverted phase-contrast micrograph of hemopoietic long-term culture: f arrows a small focus of hemopoietic cells; F arrows an area of fat cells. Note the heterogeneous arrangement of the adherent cell layer and the association of the hemopoietic cells (small round cells) with the fat cellsand underlying stromal cells. (b) Inverted phasecontrast micrograph of long-term culture 15 wk after infection withsrc (MoMuLV). There is an absence of fat cells and extremely reduced numbers of hemopoietic cells. The highly organized structure of the adherent layer is lost and the predominant cell type is a macrophage, many of which have abnormal morphology.

Cells harvested during feeding comprise CFU-S, progenitor cells, and mature cells. 5. During the first 1-3 wk of culture a highly organized adherent layer develops on the base of the flask containing stromal cells of bone marrow origin. The establishment of the adherent layer is essential to the subsequent onset of hemopoietic activity. The development of the adherent layer can be observed using inverted phase-contrast or bright field microscopy. The major cell types in this cellular environment are macrophages, fat cells, and the large, well spread blanket cells that can only be easily observed in the electron microscope. Within the adherent layer, foci of hemopoietic cells develop and grow (Fig. 2a). These are known as cobblestone areas and from these foci hemopoietic cells are released into the growth medium. Hemopoietic cells retained within the adherent layer are the most primitive population in the culture. They include a high concentration of CFU-S and these

282

Spooncer and Dexter

CFU-S have higher self-renewal capacity than that of the CFU-S released into the growth medium. The cells released into the growth medium include the bulk of the mature granulocytes, GM-CFC and CFU-S with lower self-renewal capacity (as measured by serial transplantation). Between 34 wk after establishing a long-term culture, it will have achieved a relatively steady-state of hemopoietic cell production. This level of hemopoietic activity should persist for at least another 8 wk and often longer. Signs that the hemopoietic activity is declining are a predominance of macrophages in the non-adherent cell population and a decline of hemopoietic cell foci in the adherent layer. Table 1 describes some of the characteris tics of active long-term cultures and of nonhemopoietic long-term cultures.

3.2. Establishing CZoned Factor-Dependent (FDCP-Mix) Cell Lines

1. Establish long-term bone marrow cultures. When active hemopoiesis is established (after 4-6 wk) infect the cultures with thesrc (MoMuLV) as described below in steps 3-6 (5). 2. The NIH-3T3 (SYC) MoMuLV producer cells are cultured in DMEM 10% NBCS at 37OCin 5% CO, in air. Subculture twice weekly at a dilution of approximately 1:3 using 0.25% trypsin to detach the cells, according to standard procedure for the culture of adherent cell lines. 3. The day before infecting long-term cultures with src (MoMuLV) feed the long-term cultures by removing 5 mL of growth medium and replacing with 5 mL fresh medium. 4. Prepare NIH-3T3 SYC (MoMuLV) cultures so they will be subconfluent on the day of infection. The day before infection remove all the growth medium and replace with half the usual volume of fresh growth medium. 5. On the day of infection irradiate the producer cell cultures with a dose of irradiation (for example, 20 Gy delivered at 84 Gy/min) sufficient to kill all the cells. Harvest the growth medium from the cultures and centrifuge at lOOO-15OOg on a bench centrifuge for 5 min. Remove all growth medium from the long-term cultures and add 24 mL of the supernatant growth medium from the irradiated producer cells. This should contain 2 x 104-2 x lo5 focus-forming units of sarcoma virus when titered on Rat-l fibroblasts (6). Thesrc (MoMuLV) is labile so the supernatant growth medium should be added to the long-term cultures within 10 min of harvesting the medium from the producer cells. Add 24 mL DMEM 10% NBCS to control long-term cultures.

Characteristics Total cell production

Table 1 of Hemopoietic Long-Term

Cultures@ morphology of nonadherent cells

Progenitor cell production (nonadherent cells)

Differential

Per 105cells Nonadherent cells/culture/wk

3-10 x 10 (406)

Per culture/wk 4!-xxmo (c200)

Early Late Blast granule- granule-Macmcells cytes cytes @ages

CFU-s GM-CFC

15-30 150300

Adherent
3-6X106

4500-30,000 (~2000) Activeculture

Nonhemopoietic
Cllhl??

lo-20 -3

l&20 -3

55-7.5 2040

<5 50-70

cells/culture

This table shows some example figures for cell production in active long-term cultures. The figures shown are ap proximate ranges to illustrate the extent of hemopoietic activity that should be attained in long-term cultures. Figures in brackets describe the characteristics of nonhemopoietic cultures that may be observed when cultures are very old or where culture conditions are not adequate to promote hemopoietic activity.

284

Spooncer and Dexter

6. At 4-6 hr later, make the volume of media in the long-term cultures up to 10 mL with complete long-term culture growth medium. 7. Feed the long-term cultures weekly by removing 5 mL and replacing 5 mL fresh medium. Assess the cultures weekly for signs of viral transformation (5), which is indicated by the following events: a. A fall in the nonadherent cell count b. A rise in the concentration of CFU-S and GM-CFC c. An increase in the self-renewal capacity of CFU-S (measured by serial transplantation in vivo) and GM-CFC (measured by replating in semi-solid media) d. Changes in the adherent cell morphology, including decrease in the number of hemopoietic foci and cobblestone areas, decrease in fat cells, loss of the cellular organization and appearance of transformed stromal cells (Fig. Zb). The changes in hemopoietic activity may first be observed 5-10 wk after infecting the cultures withsrc (MoMuLV). At this stage, factor-dependent cell lines may be derived from the cultures. 8. Two methods have been used to establish FDCP-Mix cell lines from SE (MoMuLV) infected long-term cultures. a. Method A (1) Harvest nonadherent cells fromsrc (MoMuLV) infected cultures and plate them in semisolid medium with Fischers medium and horse serum (10% v/v). Supplement the medium with IL3 at a concentration that will promote optimal development of GM-CFC colonies from normal bone marrow in semisolid media. The number of cells plated should be such that when colonies have developed on the plate they are well spaced and individual clones can easily be harvested. (2) Incubate the plates at 37C in 5% CO, in air (humidified) for 7-10 d. At this stage inspect the plates using, e.g., an Olympus zoom stereomicroscope, and pick out individual colonies from the semisolid medium with a sterile Pasteur pipet. (3) Disperse single colonies in a 2 mL vol of Fischers medium, horse serum 10% (v/v>, and IL3 (FHS/IL3) in 24-well cluster plates. Inspect the individual clones daily using an inverted microscope. Until the cells begin to proliferate, feed the cultures every 34 d by gently removing 1.5 mL FHS/IL3 and replacing it with fresh medium. The cells lie at the bottom of the well and should not be disturbed during a medium change.

Culturing

Primitive

Hemopoietic

Cells

(4) When the cells begin to proliferate, monitor the cell concentration and gradually expand the culture volume, maintaining the cell count at between 105/mL and 5 x 105/mL. Not every clone will survive, but in a successful experiment (seeNotes) >80% of the colonies will establish a cell line. (5) When th e c1ones are established and growing rapidly, freeze some in liquid nitrogen (using standard procedures, see Chapter 1, this vol.) and determine the cell line characteristics. b. Method B (1) Harvest nonadherent cells from the long-term cultures. Wash the cells and suspend them in FHS/IL-3 at approximately 2 x IO5 cells/mL. (2) Observe the cultures for growth of cells and when the cells are growing well, clone cell lines by plating in semi-solid medium with FHS/IL3 exactly as described above in Method A. 9. The cloned cell lines die in the absence of IL3. In liquid culture in FHS/ IL3 they have a doubling time of 15-24 h. The log phase of growth occurs between about 8 x 104-8 x 105cells/mL. The plating efficiency of the cells in semisolid medium with FHS/IL3 is between 5-250/o. The plating efficiency tends to increase as the cell lines age. Cells may not continue to grow if they are diluted below 5 x lo4 cells,/mL in liquid medium and they begin to die when the cell density is above 106cells/ mL. Maintain the cell lines in active growth by subculturing to approximately 6 x lo4 cells/mL twice weekly. If the cells are subcultured only once a week and are therefore in the plateau phase of growth for long periods, the plating efficiency (CFC content) of the cell population declines and the incidence of macrophages increases. Culturing cells in fetal calf serum causes a decline in the self-renewing cells and leads to extinction of the cell lines. 10. The FDCFMix cell lines have the following characteristics (Z), many of which are in common with hemopoietic stem cells. a. Survive and self-renew in response to IL3 (some cell lines may also respond to granulocyte-macrophage colony-stimulating factor). b. Survive and undergo myeloid differentiation in response to inductive hemopoietic environment in vitro (e.g., irradiated long-term culture adherent layer) in the absence of added IL3. c. The cells can be induced to differentiate into mature macrophages, neutrophils, and erythrocytes in the presence of fetal calf serum and erythropoietin in semisolid medium (the CFC-

286

Spooncer and Dexter

Fig. 3. Typical morphology

of FDCP-Mix

cells grown in liquid

culture with FHS/IL3.

mix assay). Occasional megakaryocytes, eosinophils, and mast cells are also seen. d.The cells are not infected with the src (MoMuLV). However, they do produce the ecotropic MoMuLV (helper virus). e. The cells are nonleukemic. f. Early isolates of the cell lines (within3 moof establishment) can form spleen colonies in potentially lethally irradiated mice. The ability to grow in vivo is lost as the cell lines age. The early isolates also establish long-term hemopoiesis in the presence of an inductive hemopoietic environment, i.e.; self-renewal of the cells as well as differentiation is maintained for many months. 11. Figure 3 shows the morphology of FDCP-Mix cell lines grown in liquid culture with FHS/lL3. The majority of the cells have a primitive appearance and contain many basophilic granules. The occasional mature macrophage may be observed (~3%).

4. Notes 4.1. Long-Term Bone Marrow

Cultures

Establishing long-term cultures is technically straightforward, but achieving good growth in the cultures can be very difficult for many reasons (some of which are not defined!). Possible problems are listed below.

Culturing

Primitive

Hemopoietic

Cells

287

1. The most likely problemis that the batch of horse serum is suboptimal. The best solution to this is to obtain a sample of horse serum from a laboratory that has established the technique and is willing to provide about 100 mL of their serum. Compare samples against the good batch by establishing cultures and measuring cell production, CFU-S, and GM-CFC for at least 6 wk. 2. Some strains of mice do not generate hemopoietically active cultures for a long duration. For example, C3H, CBA mice and C57B1/6 mice. (7). 3. Growing cultures with loose caps in a humidified gassing incubator leads to a high level of fungal contamination because of the accumulation of spore-carrying moisture around the neck of the flask. These incubators are best for short-term cultures. If contamination of cultures is a problem, then the cultures can be established individually by flushing a single femur into 10 mL of growth medium in the flask, instead of by pooling 10 femurs into 100 mL of medium. 4. If the adherent layer is not well established within the first 2-3 wk of culture, hemopoietic activity will not occur. These cultures may sometimes be rescued by giving another inoculum of l-2 x 10 fresh bone
marrow cells at 3-4 wk.

5. Rough handling and frequent disturbance of cultures can inhibit the development of the adherent layer or even cause detachment of adherent cells. Handle the cultures gently and resist the temptation to inspect them on the microscope frequently during the first 3 wk.

4.2. Modifications

to the Standard

Culture

Method

1. The technique can be scaled up if large numbers of cells are required

e.g., flush three femora into 30 mL growth medium for culture in a T75 flask (75 cm2 surface area). 2. Terminalerythroiddifferentiationinlong-termculturescanbeachieved by supplementing the growth medium with anemic mouse serum (AMS). The AMS needs to be tested to find an effective batch and is usually effective at l-2% v/v (8). Higher concentrations of AMS are toxic to long-term cultures.

4.3. Establishing Cloned, Factor-Dependent FDCP-Mix CeZZLines

1. In our experience the transformation of long-term cultures by the src (MoMuLV) does not always follow the pattern described above. We do not know the reasons for this, but they may be that the long-term

288

Spooncer and Dexter

cultures are suboptimal at the stage of infection or that the producer cell supernatant does not provide adequate infectious virus. In 12 separate long-term culture infections performed over the last 4 yr, 50% of the experiments have led to hematological transformation and the emergence of FDCP-Mix cell lines. 2. Within a single experiment infecting long-term cultures with SK (MoMuLV) the responses of individual cultures may vary. It is useful, therefore, to assess cultures individually and to select single cultures with particularly high concentrations of CFU-S and CFC from which to derive FDCP-Mix cell lines. 3. When establishing FDCP-Mix cell lines by plating non-adherent cells from SYC (MoMuLV)-infected cultures in semisolid medium we have, in some experiments, serially replated the colonies in semisolid medium 2-3 times before selecting clones for expansion in liquid culture. This procedure may enhance the concentration of CFC with high self-renewal capacity (which can establish FDCP-Mix cell lines) compared with the selection of clones from primary semisolid media cultures.

References
1. Dexter, T.M.,Spooncer,E.,Simmons, P., and Allen,T. D. (1984) Long-termmarrow
culture: An overview of technique and experience, in Long-Term Bone Marrow CuItureKrocFoundation Series 18 (Wright, D. G. and Greenberger, J. S. eds.), Liss, New York, pp. 57-96. Spooncer, E., Heyworth, C. M., Dunn, A., and Dexter, T. M. (1986) Self-renewal and differentiation of Interleukin-3-dependent multipotent stemcells are modulated by stromal cells and serum factors. Differentiation 31,111-118. Till, J. E. and McCulloch, E. A. (1962) A direct measurement of the radiation sensitivity of normal mouse bone marrow cells. Rud. Res. 14,213-222. Anderson, S. M. and Scolnick,E. (1983) Construction and isolationof a transforming murine retrovirus containing the SYC gene of Rous sarcoma virus. J. Viral. 46,594605. Boettiger, D., Anderson, S., and Dexter, T. M. (1984) Effect of src infection on longterm marrow cultures: increased self-renewal of hemopoietic progenitor cells without leukemia. Cell 36,763-773. Wyke, J. A. and Quade, K. (1980) Infection of rat cells by avian sarcoma viruses: factors affecting transformation and subsequent reversion. Virology 106,217-233. Reimann, J., Burger, H. (19791 In vitro proliferation of hemopoietic cells in the presence of adherent cell layers. Experimental Hematology 7,45-51. Dexter, T. M., Testa, N. G., Allen, T. D., Rutherford, T., and Scolnick, E. (19811 Molecular and cell biologic aspects of erythropoiesis in long-term bone marrow cultures. Blood S&3,699-707.

2. 3. 4. 5. 6. 7. 8.

Chapter 23
High Proliferative Colony Forming Potential Cells

T, Ray Bradley, George S. Hodgson, and Ivan Bertoncello

1. Introduction
High proliferative potential colony forming cells (HIV-CFC) in mouse bone marrow (BM) were defined functionally by their ability to form large colonies, greater than 0.5 mm diameter, and containing an average of 5 x lo4 cells, in low-cell-density nutrient agar cultures after 10-14 d of incubation (I). HPP-CFC were initially detected in bone marrow cells taken from mice after 5-fluorouracil (FU) treatment by virtue of the fact that large colonies developed in culture in response to a combined growth factor stimulus of pregnant mouse uterus extract (PMUE) (2) or CSF-1 from L cell conditioned medium (3) (macrophage colony stimulating factors, M-CSF) plus a source of synergistic factor(s) (SF) (1,4,5), but not with either factor alone. Colonies that developed with M-CSF alone were low in numbers, less than 0.5 mm diameter, and contained from 50 to 5 x lo3 cells with the majority of colonies at the smaller end of the range (2). These are desig289

290

Bradley, Hodgson, and Bertoncello

nated as derived from low proliferative potential colony forming cells (LPP-CFC). At 2 d after FU treatment, the LPP-CFC are drastically depleted to less than 0.1% of the numbers present in normal bone marrow whereas the HPP-CFC are 4% of normal. HPFCFC have been shown to regenerate more rapidly than LPP-CFC, and their regeneration pattern correlates closely with the marrow repopulating ability (MRA) and the platelet repopulating ability (IRA) of post-FU BM, but not with CFU-S developing at 10 or 13 d posttransplantation, and became supranormal in numbers at 8-10 d post-FU before returning to normal at 14-16 d (4). The synergistic factor(s) sources used initially were human, rat, or mouse spleen conditioned media or human placenta (4,5). Conditioned medium from human bladder carcinoma cells, 5637, was also shown to be a source of SF in that together with CSF-1 it generated cells from post-FU BM that could be assayed by determination of the increased binding of YCSF-1 and the factor was designated hemopoietin-l(6). With M-CSF, 5637 CM also generates large colonies inagar cultures (7). Evidence has recently been produced suggesting that hemopoietin-1 is identical with interleukin-l (8). Reexamination of the production of large colonies with purified growth factors has shown that a major HP-CFC population of cells in FQ BM is optimally stimulated using a combination of the three purified growth factors, CSF-1 plus IL-l plus IL-3 (9). Cells taken later in post-FU regeneration, FU,,BM also respond to other combinations of growth factors, e.g., CSF-1 + ILa, CSF-1 + IL-1 to produce large colonies, but optimal large colony development still requires the three combined factors (9). HPP-CFC are detectable in normal bone marrow with the growth factor combination of CSF-1 plus IL-1 plus IL-3 at a low incidence: approximately 1 in 400 cells. Other combinations of growth factors can result in large colony formation, for instance, GM-CSF can substitute for M-CSF, and although qualitatively no differences are detectable, quantitatively the colony formation in post-FU bone marrow is less (9). HPP-CFC are among the most primitive hemopoietic cells in mouse bone marrow yet shown to proliferate and differentiate in vitro. The exact relationship of HPP-CFC to CFU-S,, is not rigidly defined at present. Using multiparameter cell separative strategies, it is possible to obtain fractions containing up to 30% HPP-CFC and at least 50% CFU-S,, (7). Cofractionation using these techniques indicates a close relationship, but does not necessarily indicate that they are identical.

Mouse Bone Marrow

HPP-CFC

291

2. Materials
1. Agar gels: Stocks of 1 and 0.66% Difco Bacto-agar (seeNote 1) are made for use over a month. Weigh out the solid agar into a sterile conical screwtopped flask containing a sterile magnetic stirrer bar. Add the requisite amount of sterile distilled water (DW). Stir and heat until the agar is melted and just commences to boil. Cool at room temperature. Repeat melting andboiling. Store at room termperature. It is important to make sure the caps are loose during melting and boiling to prevent shattering of the flasks. 2. Culture medium: Double strength medium is made from powdered media. We use alpha-modification of Eagles MEM (seeAppendix), but any basic medium that can be made to double strength may be used. Alpha medium has consistently given better results than most other media. For simplicity of daily use involving large amounts of medium, we prepare a more concentrated basic stock, with vitamin supplementation, for freezing and make up to the double-strength medium each day as follows: a. Concentrated Stock Use alpha powder without nucleosides for 10 L of medium (see Note 2). Stir in approximately 1500 mL sterile DW 4-6 h. Add 100 mL Eagles MEM Vitamins (x100) and 10 mL phenol red (1% aqueous). Determine osmolality and dilute to 1300 mosmol (seeNote 2), and gas with CO,. Sterile filter (0.22 pm filters) and store frozen at -2OOC. b. Double Strength Medium for Plating Concentrated stock 32 mL Sterile glutamine (200 mM) 2mL Sterile serum (seeNote 3) 40 mL Sterile sodium bicarbonate (5.6%) 8mL Gentamycin (seeNote 4) 8000 u Sterile DW 18 mL This is 100 mL of double strength medium ready for use (see Note 5). 3. Balanced Salt Solution: Hepes buffer (1.5M): 35.75 g Hepes (acid salt), 0.2 mL phenol red (l%, w/v>. Dissolve in approximately 60 mL distilled water and titrate to pH 7.2 with NaOH. Make up to 100 mL and filter through a 0.22 pm filter and store at 28C.

292

Bradley, Hodgson, and Bertoncello

4. Concentrated Stock Balanced Salt Solution: This is prepared as a x10 concentrated stock solution that is filtered through a 0.22 pm filter and stored at 2-BOC. 80 g/L NaC1,4 g/L KCl, 10 g/L glucose, 20 mL/L of a 1% (w/v) solution of phenol red, 2 x 105U/L gentamycin. 5. Isotonic Balanced Salt Solution (BSS): 10 mL Hepes buffer (1.5M and 5 mLNaHC0, (5.6%, w/v) is added to 100 mL BSS (x 10) and distilled water is added to bring the volume to 1 L. The solution is filtered through a 0.22 pm filter and stored in aliquots of 100 mL at 2-8C. 6. Growth Factors: These are prepared at suitable concentrations (see text) in single strength medium. 7. Incubation Boxes: Since the dishes should be incubated at 7% oxygen gas phase, we routinely use 5-L capacity polystyrene plastic boxes (see Note 6: Stewart Plastics, Purley Way, Croydon CR9 4HS,UK, Cat. No. 212 into which perforated stainless steel trays are placed). The tray is 2 cm above the bottom of the box. Two diagonally opposite holes (18 gage needle size) are drilled in the lids of the boxes. 8. Gas Mixture: Plastic boxes are gassed with a gas mixture of 5% oxygen, 10% carbon dioxide, and 85% nitrogen, which is ordered from regular suppliers (seeNote 12).

3. Methods
Plating of bone marrow cells in nutrient agar can be made using either a double or a single layer system. We routinely use a double-layer system since colony formation is better than in the single layer system.

3.1. Removal

of Marrow

1. Groups of at least five mice are used to provide femurs. The femoral shafts are stripped of all tissue and are flushed from one end and then the other with chilled BSS containing 2% newborn calf serum using a 1 mL syringe fitted with a 23-gage needle. 2. Marrow cells are collected at a concentration of one femur equivalent (2.5 x lo7 cells)/mL BSS and kept on ice until used.

3.2. Plating

Procedure

(Double

Layer)

1. Sterile growth factors to give the desired final concentrations are added directly to the non-tissue-culture-treated 35-mM Petri dishes (Note 7). Routinely the triple growth factor combination is used at the

Mouse Bone Marrow

HPP-CFC

2. 3.

4.

5.

6. 7.

8.

following doses: CSF-1 1 x 103 U, IL-la 1 x lo4 U (conversion assay units), IL-3 25 U per dish, but all growth factor preparations should be tested for the concentrations necessary to achieve optimal colony formation. Melt the 1% agar in a conical flask to just boiling and put in a 37OCbath to cool. Calculate the amount of double strength medium to be used for the number of dishes to be plated and allow extra to ensure that the medium-agar mixture will be adequate in volume. Place in a conical screw topped flask and warm in the 37OC bath. When the agar is cooled to 37OC (test it by agitating it in the flask and feeling the temperature), add an equal volume of it to the double strength medium mix thoroughly (either by pipet or with a sterile magnetic stirrer bar previously placed in the flask) and dispense 1 mL to each of the dishes, shaking the dishes from side to side to ensure complete coverage of the dish and mixing with growth factors (Note 8). This basal layer should gel within a few minutes at room temperature (2OOC). For the overlay (containing the bone marrow cells), melt and just boil the 0.66% agar and place in the bath to cool (Note 9). Add the required number of bone marrow cells to prewarmed medium (double strength) and immediately add an equal volume of agar. After mixing thoroughly, dispense 0.5 mL aliquots to the dishes; no shaking is required. After a few min the upper layer should be gelled and the dishes can be incubated (seeNote 10). For normal bone marrow we routinely use 2.5 x lo3 cells per dish: for FUzd, 2 x 104;for FQ, 1 x 103. For fractionated enriched bone marrow populations, the numbers are reduced appropriately (Note 11). 50 mL of sterile DW is placed under the stainless steel tray of the plastic box to provide humidity during the incubation. Stack the dishes in the box. Each box can take up to 150 dishes. The box lids are then taped on with three layers of tape which is left taut and wound around the lid-bottom box junction. No wrinkles should be evident and the taping should be finished by smoothing it down over the junction. The boxes are then gassed at the rate of 2.5 L/min for 30 min to give a final oxygen concentration of 7%, the gassing holes are sealed with three layers of tape, and the box is placed in the 37OCroom or incubator (Note 12).

294

Bradley, Hodgson, and Bertoncello

9. At 14 d of incubation, the colony numbers are counted (Notes 13-15)

using a dissecting microscope at 20x magnification with a calibrated grid in one eyepiece to measure colony diameters. Normally the colonies originating from HPP-CFC are clearly visible without magnification but are checked to observe that they are tightly packed with cells and to count any smaller colonies (Note 14).

4. Notes
1. Various agarose preparations may be used instead of agar, provided

2.

3. 4. 5. 6.

7.

that it is one gelling adequately at room temperature, but colony formation with myeloid cells is lower than in agar. Likewise autoclaving of agar reduces colony formation and should be avoided (10). Various other media have been used. It is important to measure the osmolality of media and sera to be used in order to ensure uniformity of culture conditions. On testing a range of media, we have found a final medium osmolatity of 280-300 mosmol to be optimal for mouse bone marrow colony formation. Batches of sera must be tested to choose a pool suitable for the next 6 mo to 1 yr work. We find some batches of newborn calf serum to be better than fetal calf serum. Gentamycin has been used routinely for several years to replace penicillin and streptomycin because of its spectrum and stability. The bicarbonate concentration used gives strong buffering with the 10% carbon dioxide in the gas phase. The plastic boxes are washed and dried before using and between incubations and have several advantages over normal incubators: a. Each experiment can remain undisturbed over the incubation period. b. Contamination during incubation is not a problem. c. They are cheap, and, if a 37C room, or nonhumidified incubator, is available the number of workers who can carry out clonogenie experiments is greatly increased for little change in cost. d. They can be used to test numerous gas concentrations for various cell and culture types. Growth factor preparations should not be added to the underlay at more than 0.3 mL/dish and preferably left to 0.15 mL maximum making 10% of the total double layer system in order to avoid the gels losing their firmness.

Mouse Bone Marrow

HPP-CFC

8. For plating large numbers of dishes, e.g., above 100 dishes, Cornwall continuous pipets are used for both underlays and overlays with aluminum foil covered or cotton plugged flasks replacing the screw topped flasks. 9, For the single layer system the 0.66% agar is melted and placed in the 37OC bath. The volume of double strength medium is placed in another flask and warmed, the requisite numbers of cells are added, and the equal volume of agar added immediately mixed and plated. 10. The most common failure of cultures arises from using agar that is too hot or over-heating pipets during routine flaming of them. On the other hand, care must be exercised to ensure that the agar-media mixture does not gel before dispensing into the dishes. Il. It has been observed that colony formation with fractionated enriched populations of bone marrow is usually better than with unfractionated marrow (7). 12. The gassing of the incubation boxes is calculated to pass 75 L of gas through the 5-L vol boxes. A brief passage of gas will not suffice to achieve the final 7% oxygen used nor an adequate CO, concentration. 13. One of the most important aspects of optimal colony formation in vitro is to ensure that low cell densities are used: a. to prevent possible secondary effects occurring such as stimulation of accessory cell production of growth factors, b. to prevent the smaller colony formation that can occur when too many colonies develop in the dish. When the incidence of cells present at low frequency in the population is being measured, the number of replicate dishes must be increased rather than the cell density per dish. In practice, 20-50 colonies/dish is an optimum to achieve, and especially with an enriched population, dishes should be set up at different cell densities. 14. Colonies from different bone marrow populations may develop at different rates, e.g., using FU,, BM cell cells clones are just starting to develop at 6 d of incubation and achieve their large colony formation rapidly over the next 4-8 d. FU,, BM develop more rapidly and are often large in diameter at 8-10 d of incubation. 15. HPP-CFC have been detected in all strains tested, e.g. BALB/C, C57B16,AKR, C,H/HeJ, CBA (C57Bl x BALB/C) -Fl, (C57Bl x DBA/2)-Fl, 16. Since colony formation takes place over a lengthy incubation period (10-14 d), the I-IPP-CFC may develop by sequential action of the growth factors initially placed in the cultures on cells generated with-

296

Bradley, Hodgson, and Bertoncello

in the colonies during colony development. Also, combinations of other growth factors than those discussed here may detect cells with high proliferative potential with their actions being either additive or synergistic. 17. At the present time the size of the colony, and more particularly the number of cells generated per single HPP-CFC, are the criteria for detection of these cell types. The development of large colonies in the primary cultures at 14 d does not exhaust their total proliferative potential since these colonies can be sampled, dispersed into single cells, and replated to give further colony formation, although the secondary colonies are smaller and no longer require all of the three growth factors. For example, large colonies developed from FU,,BM cells with CSF-1 plus IL-3 plus 5637 conditioned medium can be replated with CSF-1 alone to yield ultimately 27 x lo6 cells/initial HPP-CFC. This may demonstrate local exhaustion of the medium during the primary colony formation and/or local inhibitory effects within the tightly packed colonies.

References
1.
2.
3. 4. Bradley, T. R. and Hodgson, G. S. (1979) Detection of primitive macrophage progenitor cells in bone marrow. BZood 54,1446. Bradley, T. R., Stanley, E. R., and Sumner, M. A. (1971) Factors from mouse tissues stimulating colony growth of mouse marrow cells in vitro. Amt. J, Exp. Biol. Med. Sci. 49,595. Stanley, E. R. and Heard, P. M. (1977) Factors regulating macrophage production growth. Purification and some properties of the colony stimulating factor from medium conditioned by mouse L cells. I. Biol. Chem 252,4305. Bradley, R. R., Hodgson, G. S., and Bertoncello, I. (1980) Characteristics of macrophage progenitor cells with high proliferative potential, their relationships to cells with marrow repopulating ability in 5-FU treated mouse bone marrow. Experimental Hematiol. Today (1979) (Baum, S. J., Ledney, G. D., and van Bekkum, D. W., eds.), Karger, NY, p. 285. Kriegler, A. B., Bradley, T. R., Januscewicz, E., Hodgson, G. S., and Elms, E. F. (1982) Partial purification and characterization of a growth factor for macrophage progenitor cells with high proliferative potential in mouse bone marrow. Blood 60,503. Bartelmez, S. H., Sacca, R., and Stanley, E. R. (1985) Lineage specific receptors used to identify a growth factor for developmentally early hemopoietic cells: Assay for hemopoietin-2. I. Cell. Physiol. 122,362. Bertoncello, I., Bartelmez, S. H., Bradley, T. R., and Hodgson, G. S. (1987) Increased Qa-m7 antigen expression is characteristic of primitive hemopoietic progenitors in regenerating marrow. I. Immunol. 139,1096.

5.

6.

7.

Mouse Bone Marrow

HPP-CFC

297

8. Mochizuki, D. Y., Eisenman, J. R., Conlon, I. J., Larsen, A. D., and Tushinski, R. J. (1987) Interleukin-1 regulates hematopoietic activity, a role previously ascribed to hemopoietin-1. Proc.Nafl. Acad. Sci.84,2567. 9. Bartelmez, S. H., Bradley, T. R., Bertoncello, I., Mochizuki, D. Y., Tushinski, R. J., Stanley, E. R., Hapel, A. J,, Young, I. G., Kriegler, A. B., and Hodgson, G. S. (1989) Interleukin-1 plus interleukin-3 plus colony stimulating factor-l are essential for clonal proliferation of primitive myeloid bone marrow cells. Exp. Hmafol. in press. 10. Dixon, R. A., Linch, D., Baines, I., and Rosendaal, M. (1981) Autoclaved agar contains an inhibitor of granulocyte-macrophage colony growth in vitro. Exp. Ceil Res. 131,478.

Chapter 24 Murine Bone Marrow-Derived Macrophages

E. Richard Stanley

1. Introduction
The molecular phagocytic lineage comprises, in order of increasing maturity, the committed macrophage precursor cell, the monoblast, promonocyte, monocyte and the macrophage. Methods for the preparation and culture of bone marrow-derived macrophages, developed by Stanley and colleagues (l-3), provide large numbers of mononuclear phagocytes that are capable of extensive cell proliferation. Since their proliferation can be stimulated by colony stimulating factor-l (CSF-l), granulocyte macrophage colony stimulating factor (GM-CSF), or interleukin-3 (IL-3), they represent an important primary cell source for studies of the actions and interactions of these three growth factors. The principles underlying the method are: (1) to generate and expand primitive mononuclear phagocyte precursor cells by culturing bone marrow cells in a combination of partially purified CSF-I and IL-3 for a period of 3 d, (2) to remove contaminating red cells, fibroblasts and mature macrophages and disrupt aggregates of proliferating cells, by proteolytic digestion of the nonadherent cells at d 1 and 3 of culture and, (3) to obtain a population of mononuclear phagocytes that is relatively homogeneous with respect to their state of differentiation by recovering only those cells (i.e., monoblasts, promonocytes) that acquire the capacity to adhere to tissue culture plastic during d 4-5 of culture. About 95% of the bone marrow-derived macrophages so obtained possess the CSF-1 receptor, 93-98% proliferatein 299

300

Stanley

response to CSF-1 and 90% of cells die on removal of CSF-1 from the serum-containing medium (I, 4, 5). This latter observation reflects the absence in these cultures of contaminating, fibroblast-like, CSF-1 producing cells and enables this population to be used not only to study the effects of growth factors on macrophage proliferation, but also on their survival.

2. Materials
1. Dulbeccos Modified Eagles Medium (DMEM) 2. Pronase solution: 0.02% (w/v) Pronase (B grade, Calbiochem), 1.5 mM EDTA in phosphate buffered saline (PBS) 3. Complete medium: DMEM supplemented with 15% (v/v) fetal calf serum (FCS), 0.292 mg/mL glutamine, 0.02 mg/mL asparagine 0.5 I.&! 2-mercaptoethanol, 0.2 g/L penicillin, 0.2 g/L streptomycin 4. Stage 1 CSF-1: Stage 1 L cell CSF-1 prepared as in ref. (6). 5. 11-3: Purified recombinant murine IL-3 or medium conditioned by serum-free cultures of the myelomonocytic leukemia cell line WEHI3 (7) 6. Zwittergent stock solution: 1% (w/v) Zwittergent 3-14 (Calbiochem). Stored at 4C. 7. 0.005% Zwittergent: Zwittergent stock solution diluted 1 in 200 with PBS. Stored at 4C.

3. Method
The method is divided into (1) isolation of bone marrow cells, (2) generation of mononuclear phagocyte precursor cells, (3) differentiation of precursor cells to adherent mononuclear phagocytes, (4) culture of bone marrow-derived macrophages (2,2) and (5) determination of macrophage concentration (3). The last section is included since adherent cultured macrophages cannot be easily detached by trypsinization and alternative methods for quantitating cell number are required.

3.1. Isolation

of Bone Marrow

Chapters

Cells (See also 22 and 26)

the the out and

1. Remove the tibias and femurs from C3H/HeJ mice by cutting proximal end of the femur and the distal end of the tibia, leaving other ends intact. 2. Insert a 23-gageneedle into theintact ends and flush bone marrow through the cut ends with ice-cold DMEM. 3. Dispense marrow plug by three passes through a 22-gage needle centrifuge (12OOg,5 min, 4C).

Bone Marrow-Derived

Macrophages

301

4. Resuspend the pellet in ice-cold DMEM and determine the nucleated white cell concentration by counting an aliquot of the resuspended cells in 2% acetic acid.

3.2, Generation

of Mononuclear Precursor Cells

Phagocyte

1. Adjust the cell density to lo6 cells/mL in complete medium containing 500 U/mL (0.22 nM CSF-1) of stage 1 CSF-1, and 1 nMIL-3 (or 10% WEHI- conditioned medium). 2. Seed cells into tissue culture flasks at a density of 2.9 x 105 cells/cm* and incubate (37C, 10% CO, in air) for 24 h. 3. Collect nonadherent cells and discard adherent cells (fibroblast-like cells, mature mononuclear phagocytes, and other cells adhering to the flask). 4. Centrifuge the nonadherent cells (8OOg,10 min, 4OC), resuspend in 1 mL of Pronase solution/107 cells and incubate for 15 min at 37OC. 5. Stop Pronase digestion by the addition of horse serum (0.2 mL/ 10 mL Pronase solution) and layer the suspension onto 15 mL of ice-cold horse serum. 6. Incubate on ice for 15 min to allow clumped erythrocytes and debris to settle through the horse serum. 7. Remove the cell suspension from the horse serum by Pasteur pipet, overlay onto another 15 mL of ice-cold horse serum, centrifuge (1200 g, 10 min, 4C), disperse the pellet in DMEM as described above and seed into the original number of flasks. 8. Incubate the flasks at 37OC for 2 d, collect the nonadherent cells, and discard the adherent cells. Pronase treat the nonadherent cells as described above (steps 4-7) and resuspend the pellet in complete medium containing 0.22 nM CSF-1. 9. At this stage, if desired, adjust the cell suspension to 10% (v/v) with respect to dimethylsulfoxide and store at -196OC for later use.

3.3. Differentiation of Precursor Cells to Adherent Mononuclear Phagocytes

1. Adjust concentration of the pronase-treated nonadherent cells to 105 cells/mL. 2. Seed into tissue culture dishes at a density of 1.9 x 104 cells/cm* in complete medium containing 0.22 nM CSF-1 and incubate for 2 d at 37OC in 10% CO, in air. 3. Remove the nonadherent cells with two washes of sterile PBS.

Stanley

3.4. Culture of Bone Marrow-Derived Macrophages

1. Maintain log-phase growth by culturing in 1.3 nM CSF-1 in complete medium. At high cell densities it may be necessary to periodically change the medium [see ref. (I) for formula predicting the CSF-I consumption by bone marrow-derived macrophages]. 2. Log-phase cells may be rendered quiescent by washing once withPBS, replacing the growth medium with complete medium (without CSF1) and incubating for 16 h. 3. Remove adherent macrophages for subculture by scraping with a sterile cell scraper.

3.5. Determination
1. 2. 3. 4.

of Macrophage

Concentration

Wash the cell monolayer with ice-cold PBS. Pipet 0.005% Zwittergent onto the cells. Incubate for 5 min at 20C. Examine with a microscope to ensure cell detachment and remove by pipet using several up and down pipetings. 5. Dilute appropriately and count by electronic cell counter or hemocytometer.

References
I. 2. 3. 4. 5. Tushinski, R. J., Oliver, I. T., Guilbert, L. J., Tynan, P. W., Warner, J. R, and Stanley, E. R. (1982) Survival of mononuclear phagocytes depends on a lineage-specific growth factor that the differentiated cells selectively destroy. Cell 28,71-U. Guilbert, L. J. and Stanley, E. R. (1986) The Interaction of 1251-Colony-stimulating Factor-l with Bone Marrow-derived Macrophages. J. BioZ. Chem. 261,4024-4032. Tushinski, R. J. Zwittergent cell counting method, personal communication. Tushinski, R. J. and Stanley, E. R. (1983) The regulation of macrophage protein turnover by a colony stimulating factor (CSF-1). J. Cell. Physiol. 116,67-75. Tushinski, R. J. and Stanley, E. R. (1985) The regulation of mononuclear phagocyte entry into S phase by the colony stimulating factor, CSF-1. J. Cell. Physiol. 122,

221-228.
6. Stanley, Il. R. (1985) The macrophage colony stimulating factor, CSF-1, inMethods in EnzymoZogy-lmmunocheicaI Techniques, vol. 116 (Colowick, S. I?. and Kaplan, N. 0. eds.), Harcourt Brace Jovanovich, New York, pp 564-587. 7. Guilbert, L. J., Nelson, D. J., Hamilton, J. A., and Williams, N. (1983) The nature of 12-O-Tetradecanoylphorbol-13-Acetate(TPA)-stimulated hemopoiesis,colonystimulating factor (CSF) requirement for colony formation, and the effect of TFA on [12511 CSF-1 binding to macrophages. J. CeZZ.Physiol. 115,276-282.

Chapter 25 Long-Term B Lymphoid Cultures from Murine Bone Marrow

Establishment and Cloning by Using Stromal Cell Line AC 6.21
and Cheryl Christa A. WhitZock E. MuZZer-Sieburg

1. Introduction
Nearly all hematopoietic cells in mammals derive from precursors that undergo much or all of their development in the bone marrow. In vitro models for many lineages are available and represent modifications of the original bone marrow culture system designed by Dexter and Lajtha (I) and described elsewhere in this book. In this chapter, we describe a second bone marrow culture system, first reported in 1982 (2), that provides an in vitro environment selectively supporting long-term proliferation and differentiation of early B lymphocyte lineage cells. This method can be used to obtain heterogeneous populations of immature precursors of the B cell lineage greatly enriched from other hematopoietic cell types. Clonal populations can also be obtained by extension of this method to limiting dilution culture. 303

304

Whitlock and Muller-Sieburg

A recent major advance in the technology of B lineage cell culture has been the isolation of stromal cell lines that support B lymphopoiesis from the primitive precursors through the pre-B cell stage. These lines can substitute for the mixed stromal cell layer that provides the supportive microenvironment in the primary bone marrow cultures. We describe one such cell line, AC 6.21, and its use for expansion of established heterogeneous populations of B lineage cells and for limiting dilution culture to obtain clonal lines or determine the B cell precursor frequency of selected populations. The technique of long term bone marrow culture for the study of B lymphopoiesis is still in its infancy, but much progress has been made since its inception in 1982. We now know something about the characteristics of the stem cell precursors that establish the continuous B lineage cell lines and will hopefully someday learn more about how these precursors are committed to the B lineage by the culture environment. Many investigators are also isolating stromal cell lines that are able to support B lymphopoiesis and, through their efforts, the factors and cellular interactions that are important in supporting B cell ontogeny will be defined in the future.

2. Materials
1. Culture medium: RPM1 1640 medium,5% (v/v> selected batchof fetal calf serum (Note 1),5 x 10sM 2-mercaptoethanol, 2 mM glutamine, 1 mM sodium pyruvate, 100 U/mL penicillin, 50 pg/mL streptomycin. Antibioticsarerecommendedforinitiatingcultures,butmaybeomitted after 2 wk. Antifungal agents are toxic to the stromal layer and should not be used. B-mercaptoethanol is essential for viability and growth of the stromal cell line AC 6.21. 2. Mice: Balb/c mice provide the most consistent results, but continuous lymphoid cultures have been established using other mouse strains. Mice must be between 2-l /2 and 3-l /2 wkof age at the time of marrow harvesting. BALB congenic strains such as BALB.B, BALB.K, and BAB-14, are useful for preparing feeder layers (Note 3). 3. Enzyme: Dispase-Collagenase is purchased in 100 mg vials from Boehringer-Mannheim. One vial is dissolved in 20 mL serum-free medium to prepare a 10x stock, then filtered through a 0.45 pm filter before aliquoting and storing at -35 to -70C. Thawed and unused portions retain significant activity when refrozen or stored for a few days at 4OC.

Long-Term B Lymphoid

Cultures

305

Table 1 Antibodies and Their Specificities Antibody RA3-6B2 Ml /70 RA3-8C5 30H12 53-2.1 19XE5 31-11 GK1.5 5.3-6.72 Specificity B220 (pre-B and B cells) Mac-l (macrophages) Granulocytes (=Gran-1) Thy-l .2 Thy-l .2 Thy-l .l Thy-l (nonallelic) L3T4, on T, cells Lyt-2, on T, cells Cells in bone marrow % Reference 25-30 23-28 28-33 3-5 3-5 3-5 3-5 l-3 2-3

4. Reagents and equipment for immunofluorescent staining and cell sorting: Iscoves medium provides better viability of cells during these procedures. Propidium iodine is used in the final step of the staining procedure in order to detect and eliminate dead cells from the sorting. Stocks are prepared in PBS at 1 mg/mL and stored at -2OOC. Once stained, the cell suspension is filtered through a nylon screen with mesh3-60/45 (TetcoInc.,Elmsford,NY) inorder toremovecellclumps. Sorting requires a fluorescence-activated cell sorter (FACS) equipped with two lasers, one tuned to 488 nm to excite FITC, and the other to 590 nm to excite Texas red. 5. Antibodies: Antibodies used for enrichment of B lymphocyte progenitors are listed in Table 1 with their corresponding references. We have found that good quality antibody preparations are critical for successful purification of Thy-lloT-BG-M-cells (seeSection 3.2.4 and Note 4). We use antibodies derived from serum-free hybridoma culture supernatents that have been -loo-fold concentrated by ammonium sulfate precipitation (50% w/v). These antibodies can be labeled with biotin, FITC, or Texas red as described (10) and are always titrated prior to the FACS sorts to determine the optimal concentration for staining. All antibodies are diluted in Iscoves medium with 3 to 5% serum and sterile filtered through a 0.2 Frn disk filter directly before use. (Prewashing the filter with serum-containing medium will reduce loss of antibody by nonspecific binding to the filter.)

306

Whitlock and Muller-Sieburg

3. Methods
3.1. Lymphoid Long-Term Bone Marrow (Whitlock-Witte Cultures) Cultures

The following technique is remarkable for its simplicity and reproducibility. The most critical ingredient for success is the batch of fetal calf serum used. Selection of a suitable serum batch is described below (Note 1). A recommendation for those initiating cultures for the first time or screening serum batches is to use Balb/c mice because marrow from these mice has proven to be best for rapidly and reproducibly establishing B lineage cultures.

3.1.1. Harvesting

Marrow

and Initiating

Cultures

1. Sterilize the surgical area widely with ethanol. 2. Open the skin overlying the femur anteriorly with sterile scissors from abdomen to well below the knee. 3. Trim the thigh muscles away, then remove the bone by cutting just below the hip and knee joints. 4. Transfer the bones to a Petri dish containing medium or balanced salt solution with 5% serum and keep on ice until all the bones are removed. The knee joint is preserved until just before marrow harvesting because of the tendency of the marrow plug to be partially extruded when it is cut. 5. Scrape the interior of the marrow cavity with a 25-gage needle while flushing out the marrow plug with 3-5 mL of medium or balanced salt solution containing 5% serum. All clumps of cells and bony spicules should be saved and cultured. (A goal of marrow harvesting is to obtain as many stromal cells as possible.) 6. Pipet the cells and medium up and down to break up as many clumps as possible, then centrifuge the entire cell suspension at 1200 rpm for 10 min. 7. Suspend the cell pellet in culture medium at lo6 cells/ml. 8. Dispense the cell suspension at approximately 2.5 x 10s cells/cm2 of the culture vessel (an exception is in vessels less than 10 cm2, such as in 24- and 96-well plates). The volumes used for most conventional culture vessels are shown in Table 2. 9. Incubate cultures at 37OC,in 7-8% CO,, and in a well-humidified incubator. Cover multiwell plates in plastic wrap to decrease air circulation around them. Tissue culture plates are best handled in stacks on trays, and individual trays can be wrapped with foil. Wrapping cul-

Long-Term B Lymphoid

Cultures

Table 2 Culture Volumes for Standard Tissue Culture Vessels Volumes for initiation: Surface area, mL cm2 Vessel type 1.0 2 24-Well platesb 2.5 6-Well plates 10 5.0 20 60 mm dishes 57 14.0 100 mm dishes 25 7.0 T25 flasks 20.0 T75 flasks 75 The volume given is cell suspension needed to produce a similar density of cells per square centimeter in each vessel type. For initiation of high density cultures that will produce B lineage cells,a cell suspension of l@cells/mL is used. For preparation of low-celldensity mixed stroma, 3 x 105cells/ml is used. bFor24-well plates (and smaller vessels,if attempted) a higher density of cells per surface area must be used in order to seed enough cells to prevent well-to-well variability in the stromal layers. tures serves to decrease evaporation and concentration of medium, as well as aid in containing and preventing spread of airborne contaminants, such as mold spores.

3.1.2. Maintenance

of Cultures

Once initiated, long-term lymphoid bone marrow cultures need only to be fed with fresh medium until the time nonadherent lymphocytes are

harvested for analysis. The timetable for feeding long-term cultures is provided in Table 3. For 60- and loo-mm tissue culture plates, dehydration is a problem, and these cultures may require feeding twice a week. Flasks and multiwell plates have less of a dehydration problem and can be fed only once a week. 1. Initiate cultures as described above. 2. For biweekly feedings, add an amount of fresh medium equivalent to l/2 of the volume of culture medium used to initiate the culture without removal of any of the spent medium. (For once weekly feedings, this step is omitted.) 3. At the subsequent feeding, aspirate approximately 80% of the spent medium (with care not to remove the nonadherent cells), then add a volume of fresh medium equivalent to the amount used to initiate the
culture. (For once weekly feedings, cultures are aspirated at each feed-

ing and fed with a volume of fresh medium equivalent to the amount used to initiate the culture.)

308

Whitlock and Muller-Sieburg

Table 3 Schedule of Feeding for Culture Maintenance

Day 0 [3 or 41 LO or 111 14 Initiation of cultures [optional] Feed with l/2 volumeb Removal of 80% spent medium; feed with full volume Repeat of d 3 or 4 Repeat of d 7

Tissue culture dishes have a greater tendency for the medium to evaporate; therefore, they should be fed twice weekly to prevent concentration of the medium and poor culture growth. bDetailsof feeding stepsare given in Maintenanceof Cultures section.

3.1.3. Collection of Cultured Lymphocytes for Seeding Secondary Cultures or for Experiments
The whole process of establishment of B lymphopoiesis (Note 2) can take 3-6 wk, depending on the serum batch. Nonadherent cell numbers steadily increase, after a brief decline, and reach a maximum that is characteristic for each individual culture and ranges roughly between 105and 106/mL. When patches of small- to medium-sized lymphocytes predominate in the cultures (Fig. l), they can be harvested for assay without sacrificing the primary culture. 1. Gently pipet the surface of the stromal layer to dislodge the lymphoid cells that are loosely adherent to a subpopulation of stromal cells. 2. Once dislodged, remove 90% of the medium and suspended nonadherent cells. 3. Add fresh medium to the primary culture vessel, and the remaining lymphoid cells will continue to divide until they reach their characteristic maximum density in 4-7 d. 4. Transfer the harvested lymphoid cells to a new tissue culture vessel and incubate at 37OC for 2 h to allow any dislodged stromal cells to adhere. (It is best to leave the harvested lymphocytes in their condi tioned medium during this incubation. The adherence step is not required for routine expansion of lymphoid cultures.) 5. Harvest the lymphocytes a second time by gently pipeting the medium or swirling the culture vessel. 6. Repeat the adherence steps if a large number of stromal cells appear to be present. 7. Lymphocytes harvested from the primary cultures can now be used for experiments or subdivided onto established mixed stromal layers

Long-Term B Lymphoid

Cultures

309

Fig. 1. Established long-term lymphoid bone marrow culture initiated at lti cells/mL (high cell density). Patchesof small- to medium-sized lymphocytes predominate and are closely associated with underlying stromal cells. Scattered nonadherent cells not associated with stromal cells have poor viability. Often lymphocytes are seen beneath the stromal cells (center/bottom), a process termed pseudoimperipolesis.

310

Whitlock and Mulbr-Sieburg

or stromal cell lines (seebelaw) in order to expand the number of cells that can be obtained. The number of lymphocytes that is optimal to seed onto low-cell-density, mixed stroma has never been determined, but we know that primary cultures can be subdivided at least lo- to 20fold.

3.1.4. Preparation

of Mixed Stromal

Cell Feeder Layers

Mixed stromal cell layers devoid of lymphoid cells can be prepared by initiating bone marrow cultures at a lower cell density than used to initiate cultures that produce B lineage cells. Low-cell-density, mixed stroma proliferate to cover 50-75% of the culture vessel surface in about 3-4 wk. At this point, nonadherent lymphocytes harvested from high density cultures, as described above, can be seeded onto them. 1. Harvest bone marrow from young mice of the appropriate mouse strain or congenic type that will allow differentiation of the seeded lymphocytes from those that might derive from the feeder layer (see Note 3). 2. Suspend cells in culture medium at 3 x lo5 cells/ml. 3. Initiate cultures with the same volumes as indicated for high density cultures (Table 2). 4. Maintain cultures, as indicated, for high density cultures in Table 3 and Maintenance of Cultures section. 5. After 3-4 wk of culture, check cultures by phase microscopy for degree of confluence (should be at least 50% confluent), types of stromal cells present (should be a variety of stromal cell types), and the presence of foci of nonadherent lymphoid cells (Fig. 1; discard cultures with lymphoid colonies since these will contaminate the seeded cells). 6. Continue to regularly feed the feeder layers until they are ready to be used, and they should be functional for 2-3 mo.

3.1.5. Preparation

of Stromal Cell Line (AC 6.21) Feeder Layers

Stromal cell lines that can substitute for the mixed stroma are currently becoming available. One of several lines we have prepared, AC 6.21 (II), is especially good at supporting B lymphopoiesis from a very primitive precursor through the cytoplasmic u.-positive stage (seeNote 3). 1. Obtain confluent or nearly confluent stroma cell layers. 2. Aspirate medium. 3. Wash stroma 3x with serum-free RPMI-1640. 4. Add a vol Collagenase-Dispase (0.5 mg/mL in serum-free medium)

Long-Term B Lymphoid Cultures

311

5. 6. 7.

8. 9. 10.

11.

sufficient to cover the bottom of the vessel (approximately l/3-1 /2 of the vol used to initiate the culture, Table 2). Incubate at 37OC. Wait 15 min. Check by phase microscopy for cell detachment. Typically, AC 6.21 does not completely detach by this treatment until pipeted. If enzymatic digestion is complete, the membrane processes will be retracted to thin branching fibers that give a spider web appearance to the culture. If necessary, mix gently and wait 10-15 min more. Check by phase microscopy again for cell detachment. Once enzymatic digestion is complete, harvest the cells by vigorously pipeting the vessel surface with aPasteur pipet. (If most of the cells are still loosely adherent to the vessel before pipeting, then the Collagenase-Dispase solution can be removed by aspiration and the stromal cells harvested directly into a desired amount of fresh culture medium containing serum. This allows elimination of Step 12 below and simplifies routine passage of the cells.) Check by phase microscopy for efficient removal.
into the enzyme solution.

12. Pellet cells at 800 rpm for 7 min if harvested

13. Once in culture medium, count viable cells. 14. For preparation of feeder layers that are to be used within 1 or 2 d, dispense the harvested stromal cells into the new cultures at a density that is about l/4 to l/2 the density needed for preparing confluent stromal cell layers (Table 4). 15. For routine passage of the stromal cell line, it should be carried at subconfluency to prevent selection of variants that are not contact-inhibited. Our routine is to passage the cell line once per week and, at each passage, set up flasks that are 1:20 and 1:50 splits of a culture that is just nearing confluency. This is approximately 1 x lo4 and 4 x 1O3cells /mL, respectively.

3.2. Cloning of B Lineage CeZZs (Limiting Dilution Culture)

Precursors capableof long-term proliferation inlymphoid bonemarrow cultures can be cloned by limiting dilution culture on mixed stromal cell layers or, preferably, on an established stromal cell line, such as AC 6.21. The source of lymphocytes to be cloned can be either nonadherent cells from established Whitlock-Witte cultures or fresh bone marrow (Note 4).

312

Whitlock and Muller-Sieburg

Table 4 Approximate Numbers of AC 6.21 Needed for Confluency in Various Culture Vessels No. for confluency Maximum densitv
4x103
4x104

Vessel

96-Well 24-Well 6-Well 60 mm dish 100 mm dish T25 flask T75

2x105 4x105

4x105

8x105
2x106

1.1 x 106 5x105 1.5 x lo6

1 x106
3x106
that can cover a large area of the culture dish. Cul-

tures that appear confluent can increase in cell number by reduction of the surface area
occupied by each cell; therefore, the number of cells needed to achieve confluency and the maximumnumberof cells that canbe harvestedfromadensecultureare slightly different. This column is provided as an easy reference for determining how many flasks need to be prepared for large-scale experiments.

AC 6.21 has an extensive membrane

3.2.1. Preparation
1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11.

of Mixed Bone Marrow

Stroma

for Limiting Dilution Harvest marrow from the appropriate congenic mouse strain (Note 3) and prepare as for larger bone marrow cultures. Suspend cells in culture medium at 1.5 x lo6 cells/ml. Initiate g&well cultures with 3 x 105cells/well in a vol of 0.2 mL. Wrap stacks of plates in plastic wrap to decrease dehydration. Incubate at 37OC for 2-3 d. Pipet each well to suspend nonadherent cells. This is best accomplished by using a 12-space multiwell pipetor set at 0.15 mL. Remove all the medium and add 0.2 mL of fresh medium to each well. At weekly intervals, for a total of 3-4 wk, repeat steps 6 and 7 with the goal of depleting the stroma of any nonadherent lymphoid precursors prior to cloning. At l-2 d prior to cloning, remove the medium. Add 0.1 mL fresh medium containing 10% fetal calf serum to each well. During this 2-d period, screen the cloning plates for wells that contain patches of nonadherent cells that may be mistaken for clones later on and eliminate these wells from the limiting dilution experiment (that is, simply ignore these wells when selecting and counting clones).

Long-Term B Lymphoid Cultures 3.2.2. Preparation of Stromal Cell Line Feeders for Limiting Dilution

313

1. Harvest AC 6.21 stromal cells as described above. One T75 flask will be adequate for four or five 96-well plates. The approximate numbers of stromal cells that can be obtained from confluent vessels of different sizes are provided in Table 4. 2. Suspend the harvested stromal cells in culture medium with 10% fetal calf serum at 2 to 4 x lo4 cells/mL. 3. Aliquot 0.1 mL/well(75-100% confluency). 4. Use for limiting dilution experiments within l-2 d of initiation.

3.2.3. Cloning

of Established

Lymphoid

Cell Lines

1. Harvest and adhere lymphoid cells from established bone marrow cultures as described above. 2. Pellet and resuspend the lymphocytes in fresh medium with 10% serum at 3 x lo3 cells/mL. 3. Make five threefold dilutions in the same medium to prepare suspensions of 1000,333,111,33, and 11 cells/mL in quantities sufficient to plate 0.1 mL/well for the desired number of plates. (Often less than 10% of those colonies growing after 3 wk of cloning will continue to grow after expansion [Note 41; therefore, multiple plates at the lower cell numbers need to be set up in order to obtain enough colonies that have a reasonable chance of being clonal and continued growth.) 4. By using a 12-space multiwell pipetor, add 0.1 mL of each cell suspension to each of the appropriate wells. 5. Stack the plates, wrap in plastic wrap, and culture at 37C. 6. Weekly, carefully remove 0.1-0.15 mL of spent medium by using the multiwell pipetor so not to disturb the cells at the bottom of the wells. 7. Add an equivalent amount of fresh medium to each well. 8. After 3 wk, screen wells by phase microscopy to detect wells with lymphoid colonies. (The variety of colony types found on limiting dilution culture are described below [Note 51.) 9. Prepare a plot of the log frequency of negative wells vs the number of cells added to the wells to check for linearity at the lower cell densities, particularly in the range from which clones are to be selected. (Typically, there is negative interaction at high cell densities and deviation from linearity.) 10. To expand the clones, prepare stromal cell layers in 24-well plates and also in &well plates, 60 mm dishes, or T25 flasks for further expansion.

314

Whitlock and Muller-Sieburg

For mixed stroma, this process takes 3-4 wk, as described above, and should be started in all sized vessels around the time of initiating the limiting dilution cultures. Feeders of stromal cell line AC 6.21 can be prepared l-2 d prior to use. Check each mixed stromal feeder layer carefully for lymphoid cell growth before use. Change the medium to medium containing 10% serum l-2 d prior to using the mixed stromal layers for expansion of clones. Vigorously pipet the limiting dilution well with the clone to be harvested to disrupt the stromal layer and release any lymphoid cells intimately associated with it. Transfer the entire contents of the pipeted well to one well of a 24-well plate containing an established feeder layer. Feed the expanded clones weekly by aspirating 80% of the medium and replacing it with fresh medium. Care should be taken not to crosscontaminate wells, and the plates should be kept covered in plastic wrap to prevent dehydration. Any well showing patches of 50 or more nonadherent cells, indicative of continued proliferation of the transferred cells, can be expanded further.

11.
12. 13. 14. 15.

16.

3.2.4. Cloning of Selected B Cell Progenitors from Fresh Bone Marrow

A subpopulation of fresh bone marrow cells, designated Thy-l T-B-G-M- (Note 4) is enriched in progenitors capable of producing longterm B-lymphoid cultures. Our method for isolating these cells for limiting dilution is outlined below. 1. Harvest fresh bone marrow as described above for the initiation of long-term lymphoid cultures, except use Iscoves medium containing 5% fetal calf serum instead of RPMI-1640. (In Iscoves medium, bone marrow cells maintain good viability over a longer period of time.) 2. Dilute appropriate antibodies (Table 1) in Iscoves medium and sterile filter through a 0.2 pm disk directly before use. 3. Pellet cells in a conical tube at 1200 rpm for 10 min. 4. Suspend cells in rat anti-Thy-l antibody. (For this and all subsequent steps, 10 PL of antibody or avidin is used for each lo6 cells.) 5. Incubate on ice for 15 min. 6. Wash (i.e., dilute the antibody/cell suspension with 2 mL Iscoves medium containing serum, then underlay with 50-100% fetal calf serum. Pellet the cells through the serum, then carefully aspirate the

Long-Term B Lymphoid

Cultures

315

7. 8. 9. 10. 11. 12. 13. 14. 15. 16.

17. 18. 19.

20. 21. 22. 23.

liquid from the tube being careful not to disturb the cell pellet, but remove as much serum and antibody as possible.) Suspend the cell pellet in FITC-Goat anti-Rat Ig and incubate 15 min. Wash as above (Step 6). Suspend the cell pellet in normal rat serum (diluted 1:4 in Iscoves medium) and incubate 3-5 min. Add a cocktail of biotinylated antibodies: rat anti-B220, rat anti-GranI, rat anti-Mac-l, rat anti-Lyt-2, and rat anti-L2T4. Incubate 15 min on ice. Wash as above (Step 6). Suspend the cell pellet in Texas red-Avidin and incubate 15 min. Wash as above (Step 6). Suspend the cell pellet in Dulbeccos modified PBScontaining 20 pg/ mL propidium iodine. Filter the cell suspension through anylon screen to remove cell clumps before sorting. The cell population of interest is that which stains weakly with antiThy-l and not at all with the antibodies specific for B cells, macrophages, mature T cells, and granulocytes. Generally, we obtain 1 to 2 x lo4 Thy-l T-B-G-M- cells in a 2 h sort. Collect sorted cells into ice cold Iscoves medium with 5-10% serum. Count the sorted cells. Suspend the cells in culture medium with 10% fetal calf serum (now back to RPMI-1640 medium) at 300 cells/ml. Make threefold dilutions in quantities sufficient to plate 0.1 mL in each well of the number of desired limiting dilution plates. (If bulk cultured at this point, suspend at 103-lo4 cells/ml.) Aliquot 0.1 mL of each cell suspension on established feeder layers in 96-well plates (seeabove). Feed weekly as described above for limiting dilution cloning of established lymphoid cultures (Section 3.2.3). Screen wells after 3 wk of culture for lymphoid colonies, as described in Section 3.2.3. Select and expand clones as described in Section 3.2.3.

4. Notes
1. Selection of a batch of fetal calf serum that efficiently supports B-lymphocyte proliferation in this culture system is the key to the successof all the above procedures. Screenings we have done suggest that the highly defined serums do not work well in this system. Of those re-

Whitlock and Muller-Sieburg

maining, approximately 20% will work. Two basic criteria are useful for comparing serum lots. One is its ability to permit good establishment (50-75% confluency at 3 wk of culture) of low-cell density, mixed stromal cell layers using Balb/c bone marrow. The second criterion is outgrowth of lymphoid colonies in high cell density cultures. In a 60 mm dish, high cell density (IO6 cells/ml in 5 mL) culture at 3 wk, there should be a minimum of five discreet patches of lymphoid cells containing more than 500 cells each. Serum lots that have proven to work well for this culture system have been stable for up to 4 yr when stored at -35C. 2. Recognition of whether a culture is establishing adequately is important so that time is not wasted waiting for B lineage cultures to establish when the potential of successful establishment is low. The ability to assess how well a culture is doing comes only with experience, but the following gives a verbal description of how the cultures should look at different times after initiation, which may aid in assessment. The first phase is outgrowth of the stromal layer and death of the majority of the nonadherent cells. This progresses rapidly during the first 7-10 d of culture. Both lymphoid and myeloid precursors survive this initial phase to proliferate and form discrete foci of nonadherent cells. Lymphoid colonies tend to be intimately associated with the stromal cells that underlie them and are composed of cells that vary in size from small to medium-sized lymphocytes (Fig. 1). Myeloid colonies are more uniform in cell size and do not conform to the shape of the stromal cells beneath them. Myeloid cells also are larger than even the largest lymphoid cells. The second phase is a crisis phase in which nonadherent cell numbers decrease and, with some serum batches, essentially disappear. This occurs and lasts a variable amount of time after culture initiation, but a return of nonadherent cells should be seen after 3-4 wk of culture. Only lymphoid cell proliferation survives this second phase. With good batches of serum, the crisis phase is short, and often the first phase overlaps with the third phase, characterized by steady increases of nonadherent cell numbers to the maximum cell density characteristic of the individual culture (l-10 x KP/mL). The last or fourth phase is characteristic of Balb/c bone marrow cultures, but we have not determined whether it occurs in cultures initiated with other mouse strains. If the heterogeneity of immunoglobulin gene arrangements of the nonadherent cells is followed during phase three, it is found that the culture gradually become pauciclonal after 4-6 mo of culture. The beginning of the fourth phase is heralded

Long-Term B Lymphoid

Cultures

317

by a gradual increase in the maximum number of nonadherent cells in the culture and coincides with the cultures becoming more and more pauciclonal. The cultures progress to a stage when the nonadherent cells must be subdivided in order to maintain viability, and eventually a clonal, tumorigenic, stromal layer-independent cell line emerges. Because of the tendency of older cultures to be pauciclonal and less normal in their growth patterns, those interested in studying normal B cell development should limit their experiments to cultures that are between 3- and 12-wk-old. 3. Feeder Layers: Mixed stroma vs stromal cell lines and the use of congenic mouse strains. Proliferation of B lymphoid cells in this culture system is dependent on adherent stromal cells. In the primary bone marrow cultures, as described in the previous Note, a mixture of adherent cells establishes what we call a mixed stromal cell feeder layer. From one established culture, we isolated a number of stromal cell lines that can substitute for the mixed stroma. One such line, AC 6.21, is especially good at supporting proliferation. Use of the stromal cell line, AC 6.21, for limiting dilution culture has many advantages. First, a large number of 96-well plates can be prepared in a relatively short amount of time with a minimum of effort. Second, the microenvironment within each well is more consistent than with mixed stroma, and third, there is never any problem with contamination of the clones with stroma-derived lymphocytes. Irradiation of the stromal cells is not needed since this stromal cell line is contact inhibited and will remain viable for several weeks without passage if fresh medium is given weekly. This latter characteristic makes it highly useful for cloning and limiting dilution analysis of B lineage precursors. A disadvantage of this stromal cell line over mixed stroma is that dead cell debris accumulates on the AC 6.21. The failure of this debris to accumulate on the mixed stroma probably arises from the presence of macrophages that phagocytize the debris rather than a lower turnover of the lymphoid cells. We have used the AC 6.21 stromal cell line for most of our limiting dilution studies and know that it can support proliferation and differentiation of primitive progenitors (Thy-2*OT-B-G-M-) to the pm-B cell stage of differentiation with synthesis of cytoplasmic immunoglobulin (II). Differentiation to surface Ig-bearing cells does not proceed as well on the stromal cell line as it appears on mixed stromal cell feeders. AC 6.21 is grown in the same medium used for long-term lymphoid bone marrow cultures. P-mercaptoethanol is essential for via-

Whitlock and Muller-Sieburg bility of AC 6.21, and media should always contain 5-10 x WM fresh 2-mercaptoethanol. Stromal cell lines such as AC 6.21 have a tendency to select for variants that will overgrow the confluent cell layer if it is carried routinely in a confluent state. Therefore, cultures for passage should be no more than around 80% confluent at the time of subculture. It is also advisable to have a large number of vials of an early passage frozen for future use in the event the stromal cell line does start to overgrow. If mixed stromal cell layers are used as feeders for limiting dilution culture or routine passage or expansion of bulk cultures or clones, it should be kept in mind that the feeder layer may give rise to lymphoid cells that may contaminate the passaged cells. A simple method for being able to monitor feeder layer contamination is to use congenic mouse strains to prepare the feeders. Balb/c congenics at the histocompatibility locus (e.g., BALB.B and BALB.K) and the immunoglobulin locus (e.g., BAB-14) work equally well in this culture system. No problem has arisen in our experiments when we have seeded B lymphoid cultures onto feeders with a different H-2-type. Histocompatibility markers can be easily monitored by surface immunofluorescence. Balb/c and BAB-14 cells can be differentiated by restrictionlength fragment polymorphisms in the heavy chain constant region genes (12). 4. Limiting Dilution Culture Theories and Practice: Limiting dilution culture simply means to culture a cell population at progressively lower densi ties until the cells of interest are at a frequency of less than l/well. The statistical theory used at the basis for this technique is beyond the scope of this manuscript, but in practice two points should be kept in mind. One is that if proliferation of the cell of interest is dependent only on the microenvironment provided by the culture conditions (in this case the medium and stromal cells) and if there is no negative or positive interaction between the cell of interest and other cell types in the mixed population cultured at limiting dilution, then a plot of the log of the frequency of negative wells (those not containing the cell of interest) versus the number of cells per well cultured will be linear. The second point to remember is that at a plating frequency of one cell of interest per well, approximately two-thirds of the wells will contain that cell. Since one-third of the wells do not contain the cell of interest, then a significant number of wells will contain two or more. This latter point must be kept in mind when selecting the plates from which to choose clones.

Long-Term B Lymphoid Cultures

319

Approximately 1 in 1000 fresh bone marrow cells, when seeded at limiting dilution on an established bone marrow stromal layer, have the capacity to proliferate up to4 wk. As detailed above, this frequency has been derived from limiting dilution analysis using AC 6.21. Although the vast majority of nonadherent cells found in a well-established long-term lymphoid bone marrow culture are pre-B and B cells, bone marrow derived pre-B cells (purified by flow cytometry as cells that express B220) do not have the capacity to initiate long-term lymphoid cultures. We have found this capacity exclusively in a small population of bone marrow, characterized by expression of low levels of the cell surface antigen Thy-l. These cells lack B220, Mac-l, Gran1, L3T4, and Lyt-2 antigens, and are thus termed Thy-PTB-G-M-cells. One in 15 Thy-PTBG-M- cells, which comprise 0.1-0.2% of total bone marrow, give rise to a lymphoid culture (Table 5). Thy-llDT-B-G-Mcells are also highly enriched for pluripotent hematopoietic stem cells (13). This indicates that long-term lymphoid bone marrow cultures provide an in vitro system that allows the study of B lineage differentiation from stem cells through mature B cells. When established cultures are used for limiting dilution culture (3-8 wk post initiation), approximately 1 in 10 will proliferate to form a colony of 50 to several thousand cells after 3 wk. Only 10% or fewer of these will continue to proliferate after expansion. Therefore, when planning a limiting dilution experiment for obtaining clones for future experiments, a large number of plates at the dilutions that should have approximately 0.3 clonogenic cells/well should be initiated to obtain a large number of clones for expansion. Which clones will have long-term proliferative potential should be obvious after 2 wk of culture in the 24-well plate. 5. Selecting Lymphoid Clones from the Variety of Colonies Seen in Limiting Dilution Cultures: The most striking finding on limiting dilution culture of bone marrow cells on AC 6.21 is the complexity of the colony types observed. The colonies not only differ greatly in cell number, but also in cell morphology. The most prominent colony that will be observed is one that consists of large cells, approximately three times the size of a lymphocyte and are round and nonadherent in the 1st wk in culture (Fig. 2A and 2B). With time, the cells in these colonies become granular and vacuolated, and they adhere to the stromal cells beneath them (Fig. 2C). About 1 in 35-50 fresh bone marrow cells will form such a colony; therefore, these colonies present a problem in limiting dilution culture of lymphoid cells in whole bone marrow. Large numbers of them in a well will compete with growth of lym-

Whitlock and Muller-Sieburg

Table 5 Frequency of B Cell Precursors in Sorted Bone Marrow Populations Population Bone marrow, % 1/frequency Total bone marrow B220+ B-M-G Thy-l+b Thy-l T-B-G-M100 30 16 5 0.1-0.2 1200 >4000 150 70 15f5

B-M-G: cells depleted of B220, Gran-1, and Mac-l expressing cells. Thy-l: contains both T lymphocytes and Thy-lTB-G-M- cells. Frequencies of wells containing lymphoid colonies were determined at 2 wk.

phoid colonies. Since lymphoid cell precursor frequency in whole bone marrow is 1 in 300-1000, then every well containing a lymphocyte precursor also potentially contains six or more colonies of these large, granular cells. Fortunately, the large, granular cells do not have long-term growth capacity and, therefore, will be easily depleted with passage of the lymphoid clones. Lymphoid colonies can be recognized as those that consist of very small cells that vary in size and tend to adhere to the stromal cells (Fig. 2A). All colonies with this gross appearance can be shown to contain lymphocytes by Wrights-Geimsa staining and immunofluorescent staining of B220, a B lineage specific surface marker. A third group of colonies consists of cells that are more uniform in appearance and intermediate in size when compared to lymphocytes and large, granular cells (not pictured). Wrights-Geimsa staining of these cells most often shows them to contain predominantly primitive cell types with loose chromatin and basophilic cytoplasm. In some cases, these primitive cells are mixed with some lymphocytes or polymorphonuclear leukocytes. Attempts to expand such colonies have failed thus far. If nonadherent cells from primary lymphoid long-term bone marrow cultures are used as the source of cells for cloning, then the distribution of colonies obtained is more uniform. If young cultures are used (3-wk-old or less), then there is still a significant frequency of large, granular cell colonies that are obtained, but the frequency is about lo-fold less than for fresh bone marrow. In addition, the frequency of clonable B lineage cells is IO-fold higher; therefore, it is simpler to obtain a large number of pure B lineage clones by using established B lineage cultures. However, our experience has shown that

Fig. 2A

Fig. 2B

Fig. 2C

Fig. 2. Colony types in limiting dilution culture of fresh bone marrow cells on AC 6.21. (A) A colony of small lymphocytes is shown adjacent to a young colony of large-granular cells. The latter cells are at a stage in which most are still nonadherent, but many are beginning to increase their granularity. (B) A higher power view of the largegranular cell colony from (A). (C) Pictured is a high power view of a large-granular cell colony that is more mature than shown in (B). All of the cells are now highly granular and adherent. Many are swollen by a large intracytoplasmic vacuole that displaces the nucleus to one side (signet ring formation). Magnifications: (A) 25x, (B) 50x, (Cl 50x.

322

Whitlock and Muller-Sieburg

clones obtained from a single experiment can have the same immunoglobulin heavy chain gene rearrangements and, thus, are siblings. There are other cell types that undergo limited proliferation on AC 6.21 to give colonies of less than 50 cells. Once such cell type is oblong in shape and forms a colony where the cells characteristically migrate away from each other (not pictured). We have not been able to obtain enough of these cells for WrighYs-Geimsa stain or other methods of characterization.

References
2. Dexter,T.M.andLajtha,T.G. (1974) Proliferationofhaemopoieticstemcellsintitro. Br. I. Haematol. 28,525-530. 2. Whitlock,C.A.andWitte,O.N. (1982)LongtermcultureofBlymphocytesand their precursors from murine bone marrow. Proc. Natl. Acad. Sci. USA 79,3608-3612. 3. Coffman, R. L. and Weissman, I. L. (1983) Immunoglobulin gene rearrangement during pre-B cell differentiation. I. MoI. Cd. Zmmunol. 1,31-38. 4. Springer, T., Galfre, G., Secher,D. S.,and Milstein, C. (1979) Mac-l: a macrophage differentiationantigenidentifiedbymonoclonalantibody. Eur.J. Zmmunol. 9,301-306. 5. Holmes, K. L., Langdon, W. Y., Fredrickson, T. N., Coffman, R. L., Hoffman, P. M., Hartley, J.W., and Morse, H. C. (1986)Analysis of neoplasmen induced by CAS-BRM MuLV tumor extracts. 1. ZmmunoE. 137,679-688. 6. Ledbetter, J. A. and Herzenberg, L. A. (1979) Xenogeneic monoclonal antibodies to mouse lymphoid differentiation antigens. Zmmunol. Rev. 47,63-90. 7. Lostrom, M. W., Stone, M. R., Tam, M., Burnette, W. N., Pinter, A., and Nowinski, R. C. (1979). Monoclonal antibodies against leukemia viruses: identification of six antigenicdeterminantsonthep15(E)andgp7Oenvelopeproteins. Virology98,336-350. 8. McGrath, M. S., Pillmer, E., and Weissman, I. L. (1980) Murine leukemogenesis: Monoclonal antibodies to T cell determinants arrest T lymphoma cell proliferation.
Nafure 285,259-261. 9. Dialynos, D. P., Wilde, D. B., Marrack, P., Pierrces, A., Wall, K. A., Havran, W.,

10.
11. 12. 13.

Otten, G., Loken, M. R., Pierres, M. R., Kappler, J.,and Fitch, F. W. (1983) Characterizationof the murine antigenic determinant, designated L3T4a, recognized by monoclonal antibody GK1.5: Expression of L3T4 by functional Tcell clones appears to correlate primarily with classII MHC antigen restriction. Zmmunol. Rev. 74,29-55. Hardy, R. (1984)Purification and coupling of fluorescent proteins for use inflow cytometry, in HandbookofExperimentaZZmmunology, fourth edition (Weir, D. M., Blackwell, C., and Herzenberg, L. A., eds.), Blackwell Scientific, Oxford, pp. 146-155. Whitlock, C. A., Tidmarsh, G. F., Muller-Sieburg, C. E., and Weissman, I. L. (1987) Bone marrow stromal cell lines with lymphopoietic activity express high levels of a pre-B neoplasia-associated molecule. CeII 48,1009-1021. Nottenberg, C. and Weissman, I. L. (1981) Ccl gene rearrangements of mouse immunoglobulin genes in normal B cells occurs on both expressed and nonexpressed chromosomes. Proc. Natl. Acad. Sci USA 78,484-488. Muller-Sieburg, C. E., Whitlock, C. A., and Weissman, I. L. (1986) Isolation of two early B lymphocyte progenitors from mouse marrow: A committed pre-pre-B cell and a clonogenic Thy-l hematopoietic stem cell. Cell 44,653-662.

Chapter 26

CFU-C in Agar
Vincent Praloran and Anna Bartocci
1. Introduction
Agar culture systems for the clonal growth and differentiation of hemopoietic cells were first described 20 yr ago (1). The progenitor cells that developed into colonies in agar after several days of culture in the presence of a source of hemopoietic growth factor (2,3) were initially called Colony Forming Units in Culture (CFU-C). They are found in bone marrow, spleen, blood, fetal liver, and yolk sac. It was subsequently demonstrated that the CFU-C population was heterogeneous and contained progenitors giving rise to granulocyte/macrophage (CFU-GM), granulocyte (CFU-G), andmacrophage (CFU-M) colonies. Progenitor cells of other lineages (erythroid and megakaryocytic) have also been similarly demonstrated in hemopoietic organs. The progressive identification of different types of progenitors, the demonstration of their hierarchical distribution, the purification of the growth factors regulating their proliferation and differentiation, and the results of other in vitro and in vivo experiments led to a schematic threecompartment model for hemopoiesis involving: 323

324

Praloran and Bartocci

1. The stem cell compartment, with extensive self-renewal and commitment potential. 2. A committed progenitor cell compartment that is heterogeneous and hierarchically distributed from multipotential to unipotential progenitors. The differentiation of cells in this compartment is associated with a progressive loss of proliferative potential and is dependent upon the presence of specific growth factors. 3. A compartment of maturing and mature cells, composed of morphologically identifiable cells restricted to one lineage, endowed with very limited proliferative potential. These cells quickly acquire the phenotypic and functional properties of circulating mature cells (Fig. 1). The in vitro culture technique (1) described in this chapter has provided much information on the hemopoietic system over the last 20 yr. It will remain useful for: a. Determining the frequencies of different types of progenitors and their modification in vitro or in vivo by different agents. b. The screening of various substances for their effects on progenitor cell proliferation and/or differentiation. c. The identification and isolation of hemopoietic stem cells and the analysis of their growth factor requirements.

2. Materials
1. Media: a. Alpha (a) medium stock solution: dissolve the powder for 1 L of medium in 177 mL of deionized distilled water (DDW), filter through a 0.22 pm sterile filter and store at -2OOC (seeNote 1). b. x2 alpha medium: 32 mL Alpha stock solution 2mL Glutamine solution (200 mM) Fetal calf serum (seeNote 2) 40 mL Sodium bicarbonate 8mL (5.6% w/v solution) 0.5 mL Gentamycin solution (10 mg/mL) 17.5 mL DDW Mix thoroughly, filter through a 0.22 pm filter, and store at 4OC for less than 1 wk. c. xl Medium for cell collection and dilution: x2 alpha medium diluted with an equal volume of water and adjusted to pH7.35.

CFU-C in Agar

325

PI

Er

Eo

Ba

-.

Bly

Tly

-,

Fig. 1. Schematic representation of the hemopoietic system emphasizing granulocyte and macrophage production. IL1 actson very primitive hemopoietic cells, both directly onprimitiveprogenitorsandindirectly through the stimulationof theproductionof other growth factors by accessorycells. IL-1 alone does not stimulate hemopoietic cell proliferation. Furthermore, the proliferation of very primitive cells is not stimulated by either IL-l, IL-3, GM-CSF or CSF-1 alone. Combinations of IL-3, GM-CSF and CSF-1 with IL1 are required (seeChapter 23). IL-3 is a multipotential growth factor, active on all hemopoietic cell lineages. GM-CSF, CSF-1 (M-CSF), and G-CSF act on granulocyte or macrophage progenitor cells and maturing or mature granulocytes and macrophages. They not only stimulate proliferation and differentiation, but also survival and activation of functional properties of the mature cells of these lineages. (Pl): platelets; (Er) erythrocytes;(Eo) eosinophils: (Ba) basophils; (MB) monocytes-macrophages; (G) granulocytes; (Bly) B lymphocytes; (Tly) T lymphocytes. 2. 1% Agar Solution: In a 500 mL conical flask, mix 2 g of agar with 200

mL of DDW. Heat with stirring to boiling; allow to cool and reboil a second time. Allow to cool to 50C (seeNote 3). 3. 0.66% Agar Solution: Add 33 mL of prewarmed DDW to 66 mL of 1% agar at 50C and reboil once. 4. Ficoll Paque. 5. Gas Mix: 7% 02, 10% CO, 83% N,.

326

Praloran

and Bartocci

3. Methods
3.1. Bone Marrow CelZ Collection
3.1.1. Mouse Bone Marrow Sterilize surgical instruments by boiling or in 70% ethanol.
1. Kill the mice by cervical dislocation (seeNote 4). Sterilize the skin with

2. 3. 4.

1. 2.

ethanol and, using sterile scissors, remove the muscles from the femur and around the knee. Disarticulate at ball and knee joints. Cut the lower extremity of the femur at the epiphyse-cartilage junction. Using a 3 mL sterile plastic syringe with a 23-gage needle, flush the femoral cavity several times with medium, collecting cells and medium in a 5 mL sterile plastic tube. After flushing all femurs, disperse the cells to form a single cell suspension (seeNote 5). Count the white cells and adjust the cell concentration, as desired, with medium. 3.1.2. Human Bone Marrow Samples are aspirated from the posterior iliac crest or sternum in syringes previously rinsed with preservative-free heparin diluted 1:lO in heparinized (50-100 U/mL) xl medium. Carefully layer this aspirate on a Ficoll-Paque cushion and centrifuge at 400g for 30 min at 15OC (4). The mononuclear cells are aspirated from the interface, washed two times by centrifiguration in xl medium and diluted to the desired concentration prior to plating as described below.

3.2. PZating

Procedure

1. Number the Petri dishes (35 mm Falcon) and add growth factors (GF:0.15 mL/dish) (seeNote 6) according to the appropriate experimental protocol. Each experimental condition must be set up at least in triplicate, depending on the expected progenitor cell frequency. 2. Calculate the total volume of agar-medium needed for underlayers and overlayers according to the total number of dishes to be used. In order to have enough agar medium for all dishes, prepare the underlayer and overlayer agar-medium mix for 10 more dishes than required in the protocol, e.g., for 50 dishes: 2 x a medium, Agar, 37-40c 37-39C Cells GF Underlayer Overlaver ++ 30 mL 15 mL 30 mL (1%) 15 mL (0.6%) ++

CFU-C in Agar

327

Keeping the pool of agar medium at 37OC, distribute 1 mL/dish in all dishes using a pipet or an automatic pipeting device. Allow the underlayer to gel at room temperature for 5-15 min. 4. Overlayer: Prepare the agar medium (0.33% final) and quickly add the calculated volume of cell suspension, mix thoroughly (seeNote 7) and dispense 0.5 mL/dish keeping the temperature of the agar medium pool at 37OC (seeNote 8) and swirling the dishes as you proceed to ensure even distribution. 5. Allow the overlayer to gel at room temperature and incubate immediately.

3. Underlayer:

3.3. Incubation

Procedure

1. Transfer the dishes to grids over a thin DDW layer in plasticboxes (Fig. 2) (for a detailed description, seeChapter 23). 2. Seal the boxes with plastic tape and using holes in the cover (one as an inlet, the other as an outlet) gas with a 7% 02, 10% CO, 83% N, (5) gas mixture for 20 min, at a flow rate of 3 L/min. At the end of this time, the medium in the dishes should look slightly orange. 3. Immediately seal the two holes in the cover with plastic tape. 4. Incubate at 37OCin the dark for 7-10 d depending on the type of progenitor cell colony. 5. Use a dissecting microscope at x20 magnification to count colonies (>50 cell/aggregate) and clusters (4-50 cells/aggregate).

3.4. Morphological

Analysis

of Colony CeZZs

Procedures are time and give poor ._ _ for whole plate staining __ ._ consuming _. results with the agar system. For morphological studies, we suggest a collagen gel culture technique that allows good quality in situ staining of colonies (6): 1. Prepare a collagen solution from rat tail tendons according to the technique of Lanotte and Schor (6,7, seeChapter 19, this vol.). Briefly, the tail tendons are cut out, dropped in absolute ethanol for sterilization and transferred to 0.5M acetic acid for depolymerization. When all the collagen is solubilized, it is dialyzed against distilled water, filtered through a 0.22 pm sterile filter and sterilely stored at 4C. 2. The culture medium mix contains: x2 Medium 25% Collagen solution 25% 20% FCS 20 or 25% xl Medium 5% Cell suspension 5% (or added separately to the dishes) Growth factors

328 SUMMARY
1) PLATE UNDERLAYER. +015ml GF 2) PLATE OVERLAYER: + 0.5ml +lml

Praloran

and Bartocci

Cool Down

0 5% Agar at 40C

Cool Down

0.33% Agar + Cells at 40C

3) TRANSFER DISHESTO BOXES AND SEAL WITH PLASTICTAPE
-PLASTIC TAPE

4) GAS BOXESFOR 20 WITH GAS MIXTURE

5) COMPLETELYSEAL BOXESAND INCUBATE

6) RECORD:
Clusters

Fig. 2. Schematic

summary

of plating and incubation

procedure.

CFU-C in Agar All reagents, including the mixed medium, must be kept at 4C until distribution in dishes. Polymerization of collagen (gel formation) occurs when the temperature rises above 15OC. 3. Incubate as for agar cultures. 4. To prepare the gel for staining, it must be transferred from the plate to a glass slide, then dried and fixed. This part of the method is delicate and is described in detail in the figure in ref. 8. 5. The dried gels yield a very thin film of collagen containing the colonies and clusters. Standard staining andmost cytochemical reactions are as easily performed on the dried gel as on a smear and allow easy cellular identification (8).

4. Notes
1. The Alpha medium stock solution must be prepared without sodium bicarbonate in order to avoid amino acid precipitation upon thawing. 2. Several batches of fetal calf serum must be tested in this culture system in order to eliminate batches with inhibitory activity. 3. Agar must not be sterilized by autoclaving. Care must be taken to avoid burning during preparation. Discard the preparation if burning occurs. 4. For mouse bone marrow cells, even if a low number of cells is needed, it is necessary to use cells pooled from three different mice. 5. The cell suspension has to be carefully dispersed to single cells by repeated and patient pipeting. Before diluting for culture, tiny cell aggregates and debris may be eliminated by allowing them to settle under unit gravity. 6. If several growth factors are to be added in the same dish, be careful not to contaminate pipet tips with other growth factors. When dispensing the underlayer, avoid touching the dishes or the dispensed growth factor. 7. A homogeneous cell distribution in the agar mix culture medium is necessary for reliable results; because of its high viscosity, it requires careful mixing. 8. To prevent premature gelling, the temperature of the agar medium mix must be kept between 37 and 40C. A previously adjusted water bath is very useful for this purpose. Gelling of the agar underlayer must be complete before addition of the overlayer and the overlayer must gel before incubation at 37OC. Approximately 5 min at 20C is required for gel formation.

330

Praloran

and Bartocci

References
1. Bradley, T. R. and Metcalf, D. (1966)The growth of mouse bone marrow cells in vitro. Austalian Journal of Experimental Biology and Medical Science 44,287-299. 2. Metcalf, D. (1986) The molecular biology and functions of the granulocyte macro3. 4. 5. 6.

7. 8.

phage-stimulating factors. Blood 67,257-2&I. Clark, S.C. and Kamen, R. (1987) The human hematopoietic colony stimulating factors. Science 236,1229-1237. Boyum, A. (1968) Isolation of leukocytes from human blood. Scandinavian ]ournal of Clinical and Laboratory Investigation 21: Supp. 97,31-50. Bradley, R. R., Hodgson, G. S.,and Rosendaal, M. (1987) The effects of oxygen tension on haemopoietic and fibroblast cell proliferation in vitro. I. Cell Physiol. 94, 517-522. Lanotte, M., Schor, S.,and Dexter, T. M. (1981) Collagen gels as a matrix for haemopoiesis. Journal of Cellular Physiology 106,269-277. Schor, S. L. and Court, J. (1979) Different mechanisms in the attachment of cells to native and denatured collagen. Journal of Cell Science 38,267-281. Lanotte, M. (1984)Terminal differentiation of hemopoietic cell clones cultured in tridimensional collagen matrix in situ cell morphology and enzyme histochemistry analysis. Biology of the Cell 50,107-120.

Chapter 27 Human Long-Term Bone Marrow Culture

Armand Keating and Paul Toor

1. Introduction
Major advances in our understanding of human hematopoiesis have come from the development of semisolid in vitro culture techniques for the detection of progenitor cells capable of colony formation (seeChapter 28, this volume). However, colony assays select for hematopoietic precursors committed to a specific lineage(s) and, therefore, are of limited value in the assessment of very early progenitor cells. Also, they do not provide a means of assessing the influence of microenvironmental cells deemed essential for maintaining hematopoiesis in vivo (I). Furthermore, for the assays to be valid, cells must be plated at a sufficiently low concentration to ensure clonality (a condition in which each colony has arisen from a single progenitor cell). Consequently, cell-cell interactions that may be important in exerting regulatory effects on hematopoiesis are likely to be minimal.

331

332

Keating

and Toor

These limitations, in part, have been overcome by the adaptation of Dexters murine long-term marrow culture system (2) to human marrow (3). An essential feature of this liquid culture system is the development of an adherent layer of heterogeneous composition within 2-3 wk after culture initiation (3). The predominant cell type in the adherent layer shares many of the characteristics associated with smooth muscle cells (1). Loosely adherent and nonadherent hematopoietic cells residing above the adherent layer can be removed and assayed for colony forming progenitor cells (CFC). Under optimal conditions for the initiation and maintenance of human long-term marrow cultures (LTMC), granulocyte-macrophage progenitors (CFU-GM) can be regularly detected for 14 wk and erythroid (BFU-E) and multipotent progenitors (CFU-GEMM) for up to 8 and 4 wk, respectively (4). More primitive progenitors present within the adherent layer can only be assayed by disrupting the cultures (5). Thus, the human long-term marrow culture system enables the investigator to examine and manipulate interacting cell populations in a more physiologic manner and permits an assessment of the adherent cell layer, regarded as a suitable in vitro model of the hematopoietic microenvironment (1). Moreover, long-term marrow cultures may be initiated by precursor cells earlier in ontogeny than the progenitors detectable in semisolid clonal assays, since successful cultures can be generated from marrow depleted of CFU-GM (6). Although the system has limitations, such as a tendancy to favor granulopoiesis (4), it remains the best in vitro approximation currently available to in vivo hematopoiesis.

2. Materials
1. 2. 3. 4. 5. 6. 7. Fetal bovine serum. Horse serum. Percoll (Density 1.077 g/mL). McCoys 5A tissue culture medium. Sodium bicarbonate (7.5% w/v>. 100 mM sodium pyruvate. Media additions: Vitamins (100x concentrate), essential amino acids (50x concentrate) glutamine (200 mM) (as provided by the manufacturer, Gibco). 8. Antibiotic-antimycotic solution (each mL contains 10,000 U penicillin, 10,000 U streptomycin, 25 yg amphotericin B). 9. Hydrocortisone, 2.75 mM in Dimethylsulfoxide.

Long-Term Human Bone Marrow

Culture

333

3. Methods
The procedure for LTMC generation is outlined schematically in Fig. 1 (Note 1).

3.1. Long-Term
1. Place the following

Culture

Medium

ingredients in a 500 mL Erlenmeyer flask: mL Hydrocortisone (1 mg/mL Dimethylsulfoxide) 0.18 Sodium bicarbonate 5.0 Sodium pyruvate 5.0 Vi tamins 5.0 Essential amino acids 4.0 Nonessential amino acids 2.0 Glutamine 5.0 Antibiotic-antimycotic solution 5.0 Horse serum 62.5 Fetal bovine serum 62.5 McCoys 5A medium 343.82 500.00

2. Stir the contents of the flask gently for a few minutes to ensure thorough and uniform mixing. 3. Filter-sterilize the medium through a 0.2 pm filter, and store at 4OCfor use within 3 wk.

3.2. Isolation

of Cell Populations

(Note 2)

1. Aspirate 3-5 mL of bone marrow into a syringe containing 500 U preservative-free heparin. 2. Dilute the marrow 1:l with McCoys 5A medium. 3. Isolate the mononuclear cells by density centrifugation over Percoll (Note3). Collect the interface mononuclear cells after centrifigation at 4008 for 30 min, wash three times in medium, count, and resuspend in McCoys 5A medium.

3.3. Generation

of Long-Term

Culture

1. Place the cells into long-term culture medium in any of the following systems (Note 4) (Fig. 2):

334

Keating and Toor

Bone marrow cells (Collected In preservative-free hepann, 10 UnltSlmL, and diluted 1 1 with McCoys 5A medium) Density centrifugation (PercoIl* 1077 g/mL, 30 mmutes, 400g) cells (Suspended in long-term culture medrum)

I Mononuclear

25 cm2 tissue culture flask

20 x lo* cells in 10 mL

Flat-sided tissue culture tube (Ambitube) or

3-5 x 10 cells in 2.5 mL

6 well multi-well plate

10 x lo6 cells In 5 mL per well

or Multi-chamber/glass slide micro-culture system 2 x 10e cells in 1.0 mL per chamber

Incubate cultures at 37C In a 5% COB humidified atmosphere for 7 days, then transfer to Incubator at 33C Feed cultures weekly by replacing half of supernatant with fresh medium

Weekly plate cells collected from removed supernatant for presence of CFC.

Fig. 1. Scheme for generating human long-term marrow cultures.

a. 20 x lo6 cells in 10 mL medium in a 25 cm2 tissue culture flask. b. 3-5 x lo6 cells in 2.5 mL medium in a flat-sided tissue culture test-tube (Ambitube). c. 10 x lo6 cells in 5 mL medium in each well of a six well multiwell plate.

Long-Term Human Bone Marrow Culture

335

Fig. 2 Light micrograph of an adherent layer from a hematopoietically active human long-term marrow culture. (x100).

2. 3. 4. 5.

d. 1 x 106cells in 0.5 mL medium in each chamber of a multichamher/glass slide microculture system (Notes 5,6). Incubate the cultures at 37oC in a humidified incubator containing 5% co, Feed the cultures after 7 d by replacing half of the supernatant with fresh medium. Thereafter, transfer the cultures to an incubator set at 33OC. Continue feeding the cultures weekly in the same manner (Note 7). Assay nonadherent cells present in the supernatant removed during weekly culture feeding for CFC (CFU-GM, BFU-E, CFU-GEMM) (see Chapter 28 for method).

4. Notes
1. ThemethodforgeneratingL~Cdescribedinthischapterisbasedon our modification (6) of the Gartner and Kaplan technique (3). Assessment of hematopoiesis in LTMC with colony assays is generally confined to the investigation of the nonadherent cell population. How-

336

Keating and Toor

ever, hematopoietic progenitors can also be detected and enumerated from the adherent layer, but this involves sacrificing the culture (5). The cells to be assayed can be released by disrupting the adherent layer mechanically after incubation with 0.25% trypsin for 5 min. After washing in media containing 10% fetal bovine serum, the cells are assayed for CFC as previously described (seeChapter 28, this volume). 2. Several factors are particularly important for the generation of successful LTMC. Our practice has been to screen for the most suitable lots of both FBS and horse serum, although in our experience the quality of the FBS tends to be the more critical factor. Serum batches are screened by establishing LTMC with normal donor marrow using the protocol described. A culture generated with medium containing a suitable batch of serum gives rise to a confluent adherent cell layer as shown in Fig. 2, with numerous hematopoietic islands and round cells within 2-3 wk of culture initiation. 3. The method of isolating nucleated marrow cells also influences the quality of the culture. We have found that the use of the discontinuous Percoll gradient is more consistent in producing healthy cultures than other cell isolation methods, such as buffy coat generation or the use of Ficoll-hypaque gradients. 4. Cultures also vary according to the brand of the plastic culture flasks employed. We have found that Corning TMplastic flasks permit the best growth most consistently. 5. The water source for the preparation of culturemedium has long been known to be an important factor in successful tissue culture. Consequently, when the technique is first established in the laboratory, water from different sources should be screened in the same manner as described for serum. 6. A wide selection of tissue culture vessels offers considerable flexibility in the design of studies using the long-term marrow culture system. For example, small flat-sided tissue culture flasks (ambitubes) are particularly useful for generating LTMC from limited cell numbers. The culture of a large quantity of marrow as in the case of clinical in vitro purging (7) can be accomplished in 150 cm2 CorningTM flasks. Multichamber/glass slide microcultures permit the study of the adherent cell population with immunofluorescent or autoradiographic techniques. 7. A variation of the long-term marrow culture system (the two-stage system) involves the use of allogeneic or autologous adherent layers on which can be superimposed a marrow or peripheral blood-derived hematopoietic cell population of interest (8). The adherent layers are

Long-Term Human

Bone Marrow

Culture

337

derived from healthy 34wk-old LTMC generated from normal marrow and are irradiated (1100 cGy) to prevent proliferation of the hematopoietic cells within the layer. The overlying hematopoietic cells can then be assayed weekly for CFC in the usual manner for up to 6 wk. This two-stage LTMC system can be used as an indirect means of assessing multipotent stem cell toxicity. For example, the ability of marrow, depleted of CFC by a particular modality (cytotoxic monoclonal antibody, drug, antineoplastic agent), to regenerate CFC (CFUGM, BFU-E, CFU-GEMM) after incubation over an irradiated LTMC adherent layer may serve as a measure of the viability of early precursor cells not measurable in colony forming assays. Treated cells at a concentration of 2 x lOcells/mL (assuming no loss as a result of treatment with the particular modality) are placed over the irradiated layer, incubated at 33C, and assayed for CFC at weekly intervals for up to 4-6 wk. Alternatively, assays can be performed every 2 wk depending on the activity of the culture. The two-stage LTMC system may be particularly suitable for the assessment of hematopoiesis in marrow destined for autologous transplantation and treated ex vivo to eliminate neoplastic cells.

References
I. Singer, J. W., Keating, A., and Wright, T. H. (1985) The human hematopoietic microenvironment, in Advances in Haematofogy, vol. 4 (Hoffbrand, V., ed.) Churchill, London, pp. l-24. Dexter, T. M., Allen, T. D., and Lajtha, L. G. (1977) Conditions controlling the proliferation of hematopoietic cells in vitro. J. CeZZ.Physiol. 91,335-344. Gartner, S. and Kaplan, H. S. (1980) Long-term culture of human bone marrow cells. Proc. Natl. Acad. Sci. USA 77,4756-4759. Powell, J., Keating, A., Singer, J. W., and Adamson, J. W. (1983) Analysis of hematopoiesis in human long-term marrow cultures. International Society of Experimental Hematology, London, Exp. Hemutol. II, 6. Coulombel, L.,Eaves, A. C.,and Eaves, C. J. (1983) Enzymatic treatment of long-term marrow cultures reveals the preferential location of primitive hemopoietic progenitors in the adherent layer. Blood 62,291-297. Keating, A., Powell, J., Takahashi, M., and Singer, J. W. (1984) The generation of human long-term marrow cultures from marrow depleted of Ia (HLA-DR) positive cells. Blood 64,1159-l 162. Chang, J., Morgenstern, G., Deakin, D., Testa, N. G., Coutinho, L., Scarffe, J. H., Harrison, C., and Dexter, T. M. (1986) Reconstitution of haemopoietic system with autologousmarrow takenduringrelapseofacutemyeloblasticleukemiaandgrown in long-term culture. hncet 1, 294,295. Takahashi, M., Keating, A., and Singer, J. W. (1985) A functional defect in irradiated adherent layer from chronic myelogenous leukemia long-term marrow cultures. Exp. Hematol. 13,926931.

2. 3. 4. 5. 6 7.

8.

Chapter 28
In Vitro Clonal Culture of Human Hematopoietic Progenitor Cells
Armand Keating and Paul Toor

1. Introduction
Human hematopoietic tissue represents a complex developmental system that is closely regulated to ensure the maintenance of appropriate levels of circulating blood cells. At least nine distinct lineages of cells at various stages of maturation have been described (1). Substantial progress in the study of hematopoietic stem cell ontogeny has resulted from the development of semisolid culture methods for the detection of committed progenitor cells. The assays define specific individual precursor cells indirectly by the ability of these cells to form colonies consisting of clonally expanded progeny at varying stages of differentiation. In turn, these methods have led to the identification, characterization, and synthesis by recombinant DNA techniques of some of the growth factors capable of stimulating colony formation and infer to an in vivo regulatory role for such factors in hematopoiesis. Assays employing different semisolid supports have been established for the clonal culture of progenitor cells committed to granulocyte- macrophage (CFU-GM) (Z), erythroid (BFU-E, CFU-E) (3), megakaryocytic (4), and multipotential (CFLJ-GEMM) (5) lineages.
339

340

Keating and Toor

In this chapter, we describe a single method that conveniently allows the detection and enumeration of progenitor cells (CFC or colony forming cells) capable of colony formation. Thus, in a single dish, CFC can be detected with multlineage potential (CEU-GEMM), as well as progenitors committed to megakaryocytic (CFU-Meg), erythroid (BFU-E), and granulocyte-macrophage (CFU-GM) pathways. With this technique, the frequency of all colony types is similar to that obtained in theindividual clonal assays.

2. Materials
1. Erythropoietin (Step 1, lOOOU/mg: Terry FoxLaboratory, Vancouver, British Columbia). 2. Fetal bovine serum. 3. Ficoll-Paque (Density 1.077 g/mL). 4. Iscoves Modified Dulbeccos Medium (IMDM). 5. Lux Petri dishes. 4. Methylcellulose powder (density, 4000 cl?). 7. Percoll (density, 1.077 g/mL). 8. Phytohemagglutinin.

3. Methods
The procedure is outlined diagrammatically in Fig. 1 (Note 1).

3.1. Isolation

of the Cell Population

Best results are obtained if the marrow or peripheral blood specimen is used within 2-3 h after collection. If this is not possible, the specimen, anticoagulated with preservative-free heparin, may be left overnight at room temperature for use in the morning. 1. Draw 3-5 mL bone marrow or 5-10 mL peripheral blood into a syringe containing 500 U preservative-free heparin. 2. Dilute 1:l with Iscoves Modified Dulbeccos Medium (IMDM). 3. Isolate the mononuclear cells by density centrifugation over Percoll. Collect the interface mononuclear cells after centrifugation at 4008 for 30 min, wash three times in media, count, and resuspend in IMDM.

Human

Hematopoietic

Progenitor Cells

Bone marrow cells (Collected m preservative-free and diluted 1:l with IMDM)

hepann,

IO units/ml,

Density oentrlfugation I Mononuclear ceils (Final concentration

(Percoil: 1077 g/mL, 30 minutes, 4OOg) 2 x 105/mL mixture)

Mixture
0.9% methyl cellulose 30% human plasma 5 x 1g5M 2-mercaptoethanoi 5% leukocyte-conditioned 2 umts erythropoietin/mL cells in IMDM to volume

95 mm x 12 mm test-tube

medium

35 mm petri-dash

*I----.. ,Q
l. ( , --------: .-.--.- . -..* ..* #I

Vortex 15 seconds

and plate

i . ..---.. 00-a.--..;, /.c ,.- *:-. 7 , .--def - -;* -we .s_.- . - ---._Q

14 day mcubation at 37C In a 5% CO2 humidified atmosphere

Observe erythroid, granulocyte-macrophage, megakaryocyte and mixed ceil colonies under inverted microscppe x 37

Fig. 1. Scheme for CFC assay.

342

Keating and Toor

3.2. Preparation

of Leukocyte-Conditioned

Medium

Although recombinant growth factors are becoming commercially available, leukocyte-conditioned medium (LCM) containing hematopoietic stimulating factors is readily prepared (6,7) and may be used in the CFC assay. 1. Collect 50 mL blood samples in 500 U preservative-free heparin from each of three normal donors. 2. Dilute the blood 1:l with IMDM. 3. Isolate the lymphocytes by density centrifugation over Ficoll-Paque. Collect these cells from the interface after centrifugation at 4008 for 30 min, wash three times with media, count, and resuspend in IMDM. The lymphocytes from the three blood samples can then be pooled. 4. Resuspend the cells at a concentration of 1 x 106/mL in IMDM with 10% fetal bovine serum. Add reconstituted phytohemagglutinin stock solution to a concentration of 1% (v/v>. 5. Incubate 100 mL vol in tissue culture flasks for 7 d at 37OCin a humidified atmosphere containing 5% CO,. 6. Harvest the superna tant, filter-sterilize, and store at 4OCfor use within 4 wk, or at -7OOC for use within a year.

3.3. Preparation

of Methylcellulose

(2.1%)

1. Place 42 g of methylcellulose powder and a stir bar in a 2 L Erlenmeyer flask and autoclave, using the dry cycle. 2. Add 979 mL sterile distilled water to the flask and bring to boil. Let the water boil gently for 5-10 min, or until all the clumps of methylcellulose are broken up. 3. Place the flask on a magnetic stirrer for 3 h until cooled to room temperature. 4. Add 979 mL sterile double-strength IMDM, and stir for 1 h at room temperature. 5. Continue stirring overnight at 4C. 6. Aliquot the methylcellulose in 100 mL vol, and store at 4C for use within 4 wk, or at -70C, for use within 1 yr.

3.4. Collection

of Human

PZasma (Note 2)

1. Draw 50 mLor blood from a consenting adult donor into a syringe containing 500 U preservative-free heparin. 2. Centrifuge the plasma at 2008 for 10 min to remove the plasma. 3. Centrifuge the plasma at 4008 for 5 min to remove the platelets.

Human Hematopoietic

Progenitor Cells

343

4. Filter-sterilize the plasma, aliquot it, and freeze at -7OOCfor up to 6 mo. 5. Prior to use, thaw out the plasma, and filter-sterilize it through a 0.2 l..trnfilter to remove cryoprecipitates.

3.5. Assay for CFC

The culture conditions employed for the detection of CFU-GM, BFUE, CFU-Meg, and CFU-GEMM are based on a modification of the procedure originally described by Fauser and Messner (5). 1. Suspend the mononuclear cells (~10~ cells/mL final concentration: Note 3) in IMDM in a tube containing methylcellulose as viscous support (0.9% final concn,), 30% human plasma, 5% LCM, 5 x 10-W 2-mercaptoethanol, and 2U/mL humanurinary erythropoetin. Themethylcellulose can be handled easily with a syringe and 1Bgage needle. 2. Vortex the tube gently for 10 s to ensure thorough mixing. 3. Plate 1 mL amounts of the mixture into 35 mm Petri dishes. A minimum of three, but preferably six, dishes should be plated. 4. Incubate the dishes up to 16 d at 37OC in a humidified incubator containing 5% CO, (Note 4). 5. Examine and enumerate the colonies at d 14 using an inverted microscope (Figs. 2,3; Note 5).

4. Notes
1. The CFC assay described above offers several advantages over individual colony assays. The most obvious is convenience. A major advantage resides in the use of methylcellulose as the semisolid support. Methylcellulose is easy to handle and, unlike agar, will not harden if a delay occurs during plating. Moreover, colonies can be readily plucked from methylcellulose plates and the material removed without difficulty, thus permitting cytologic and cytogenetic assessment of the cells within colonies. Also, cell transfer and replating experiments can be more easily performed. Finally, multilineage colonies appear to grow better in methylcellulose. 2. Fetal bovine serum can be used instead of human plasma in the assay. However, inclusion of human plasma results in superior erythroid and, in particular, megakaryocyte colony growth. In our experience, few batches of fetal bovine serum facilitate the assay of CFU-Meg. The presence of human plasma in the assay tends to cause the assay mixture to gel. Consequently, colony morphology differs; the colonies

344

Keating and Tow

Fig. 2. Light micrograph (x125) of mixed cell colony (CFU-GEMM). tains erythroid cells, granulocytes, megakaryocytes, and macrophages.

The colony con-

Fig. 3. Light micrograph (x80) of erythroid colony (BFU-E, lower left) and a macrophage colony (CFU-M, note dispersed monocytes-macrophages, upper right).

Human Hematopoietic

Progenitor Cells

345

tend to be tighter and the cells within them less dispersed, particularly in the case of BFU-E. The major drawback in the use of human plasma is the need for extensive screening to determine specimens giving optimal colony growth. Indeed, some specimens inhibit colony formation, but this appears to be unrelated to the blood group of the plasma donor. The most suitable plasma is obtained from fasting donors and often a pooled batch from two or more donors gives the best results. 3. Marrow or peripheral blood cells can be depleted of macrophages in order to enrich the target population for CFC. Eliminating macrophages in the cells to be plated also reduces the background single cell population and facilitates enumeration of cluster (5-39 cells) as well as colonies (more than 39 cells). Adherent cells are removed by incubating the mononuclear cell fraction in IMDM with 10% fetal bovine serum for 2 h at 37OCin polystyrene culture dishes. The macrophagedepleted cells then can be obtained by gentle pipeting and rinsing of the dishes with medium. 4. A common technical pitfall in the CFC assay is to provide inadequate humidity during incubation, leading to drying of the methylcellulose and poor colony formation. This can be readily prevented by arranging, within a large covered transparent Petri dish, the assay dishes (35 mm Petri dishes) around a central open assay dish filled with sterile distilled water. 5. Colonies can be screened with an inverted microscope through the large transparent dish containing themethylcellulose dishes from d 10 onward without unduly affecting colony development. When colony formation is judged optimal, the dishes are removed and the colonies assessed and enumerated at magnification x 37. Figure 2 shows a mixed cell colony (CFU-GEMM), whereas Fig. 3 contains a hemoglobinized erythroid colony (BFU-E) and a loose macrophage colony (CFU-M).

References
2. Metcalf, D. (1985) The granulocyte-macrophage colony-stimulating factors. Science 229,x5-22. 2. Pike, B. L. and Robinson, W. A. (1984) Human bone marrow colony growth in agar gel. J. Cell. Physiol. 76,77-84. 3. Tepperman, A. D., Curtis, J. E., and McCulloch, E. A. (1974) Erythropoietic colonies in cultures of human marrow. Blood 44,659-669. 4. Vainchenker, V., Bouguet, J., Guichard, J., and Breton-Gorius, J. (1979) Megakaryocyte colony formation from human bone marrow precursors. Blood 54,940-945.

Keating and Toor

5. Fauser, A. A. and Messner, I-I. A. (1978) Granuloerythropoietic colonies in human marrow, peripheral blood and cord blood. Blood 52,1243-1248. 6. Aye, M. T., Niho, Y., Till, J. E., and McCulloch, E. A. (1974) Studies of leukemic cell populations in culture. Blood 44,205-219. 7. Parker, J. W. and Metcalf, D. (1974) Production of colony-stimulating factor in mitogen-stimulted lymphocyte cultures. J. ImmunoI. 112,502-510.

Chapter 29 Flow Sorting for Isolating CFU-E

Suzanne and John M. Wat-t M. Davis

1. Introduction
Erythroid progenitor cells have been classified into three groups of increasing maturity: the primitive burst forming unit (p-BFU-E), the mature burst forming unit (m-BFU-E), and the erythropoietin responsive colony forming unit (CFU-E). This classification is based on their time of maturation in vitro, their proliferative capacity, and their responsiveness to growth factors (I). The CFU-E can be distinguished from the more primitive erythroid progenitors by their ability to proliferate and mature in response to a single growth factor, erythropoietin. In clonal assays in vitro, the CFU-E form single or double clusters characteristically containing 8-64 mature or maturing erythroid cells 2 d after cultures have been initiated with mouse bone marrow or fetal liver (2,3). The availability of enriched populations of CFLJ-E is important to our understanding of the function and mode of action of growth factor receptors in normal cells, the influence and regulation of molecules, genes and viruses that are specific to the erythroid lineage, and as a baseline for understanding errors in gene regulation or function that govern the development of leukemic or preleukemic states. One approach to purifying CFU-E has relied on the use of fluorescently tagged monoclonal antibodies as probes to cell surface molecules together with flow cytometry (4,5). Al347

348

Watt and Davis

though probes that specifically identify CFU-E are not available, phenotypic analysis has revealed that CFU-E can be segregated from moreprimitive erythroid precursors, from morphologically recognizable erythroid cells, and from mature myelomonocytic cells and their progenitors with a series of rat monoclonal antibodies (4-6). Hemopoietic tissues vary in their content of different types of hemopoietic progenitor cells and of maturing cells of particular lineages (6). Since mouse fetal liver is a major site of erythropoiesis and contains high numbers of CFU-E, the strategy for isolating CFU-E described here relies on the fractionation of low density fetal liver cells on the basis of their forward light scatter characteristics and differential reactivity with two rat monoclonal antibodies, YBM 42.2.2 and YBM 10.14.9 using flow cytometry (4-6). The anti-T200 antibody YBM 42.2.2 does not react with the CFU-E or more mature erythroid cells (4) but reacts with all leukocytes, thus allowing segregation of CFU-E from both myelomonocytic and lymphoid cells and from all other hemopoietic precursors. The YBM 10.14.9 antibody is then used to separate CFU-E from more mature erythroid cells (5). This approach yields cell populations containing at least 60% CPU-E, whereas 80% of the cells have the morphology of early erythroid blast cells and do not stain with benzidine, which identifies hemoglobin containing cells (5-7; seeNote 1).

2. Materials 2.1. Reagents for Media Preparation

1. Powdered Iscoves Modified Dulbeccos Medium (IMDM) containing (per L) 3.024 g NaHCO, 60 mg penicillin, 100 mg streptomycin, 5 x 105M 2-mercaptoethanol. After preparation, do not adjust pH. Store at 4OC. Light sensitive. This should be prepared at lx and 2x strength. 2. IMDM without bicarbonate but with all other additions. Adjust to pH 7.3 with 5M NaOH or pH 5.1 with HCl. 3. Powdered modified Eagles minimal essential medium with Earles salts, but without phenol red containing 20 g/L BSA and sodium azide (0.02%). 4. Sodium bicarbonate (NaHCO,). 5. Sodium benzylpenicillin. 6. Streptomycin sulphate (745 U/mg). 7. 2-Mercaptoethanol. 8. IM HEPES buffer pH 7.3 (commercially available).

CFU-E Isolation

349

9. 10x PBS: 0.2M sodium phosphate buffer with 1.48M sodium chloride pH 7.3. 10. 4% (w/v> sodium azide.

2.2. In Vitro Culture

Reagents

1. Methylcellulose (65 HG, 4000 ml?a, Fluka, AG, Buchs, Switzerland) prescreened to support CFU-E growth as described below and in Note 2. 2. Bovine serum albumin (BSA). Store at 4OC. 3. Lipids a. Cholesterol (5-cholesterol-3-B-01). b. Oleic acid (cis-9-octadecanoic acid). c. L-a-phosphatidyl choline dipalmitoyl. 4. Human transferrin. 5. Erythropoietin (Epo). Epo step 1 (lOOOU/mg) or pure Epo (8O,OOOU/ mg) obtained from Terry Fox Laboratory Media Preparation Service, Vancouver, Canada or commercial recombinant Epo. Store in 200~PL aliquots at -70C in 0.1% BSA or prescreened FCS. Do not refreeze after thawing. 6. Commercial Ficoll-Hypaque (density 1.077 g/cc). Light sensitive. Store at room temperature. 7. Fetal Calf Serum (FCS). Prescreened for supporting CFU-E growth. Store in lo-mL aliquots at -2OOC.

2.3. AnimaZs
Day 12-13 pregnant CBAf/CaH mice 8-12 wk of age. Day 0 of pregnancy is taken as the day of appearance of the vaginal plug.

2.4. Hybridomas
1. YBM 42.2.2 (Rat IgG2a antibody). This does not bind protein A at neutral PH. 2. YBM 10.14.9 (Rat IgG2c antibody). This antibody binds protein A at neutral PH.

2.5. Antibody

Purification

1. Commercial rabbit anti-rat Ig. 2. Commerical Protein A-Sepharose CL-4B.

350
3. 4. 5. 6. 7. 8. 9. 10.

Watt and Davis

O.lM phosphate buffer pH 8. O.lM glycine-HCl buffer pH 3. 2M Trizma base in water. PBS: 0.02M phosphate buffer with 0.148M sodium chloride pH 7.3. Saturated ammonium sulfate solution pH 6.8. O.lM borate buffer pH 8.2. 0.2M triethanolamine pH 8.2. 20 mM dimethylpimelimidate dihydrochloride in 0.2M triethanolamine pH 8.2.

2.6. Antibody

Labeling

1. Commercial fluorescein isothiocyanate (FITC) coupled protein A. Store at 4OC. 2. Fluorescein isothiocyanate (Isomer 1). Store dessicated at 4OC.

2.7. Stains
1. Commercial May-Grunwald Stain prefiltered through a Whatman 1MM filter paper. 2. Commercial Giemsa R66 improved stain. 3. Benzidine stock solution: 0.2% (w/v) benzidine hydrochloride in 0.5M acetic acid. This can be stored in the dark at 4OCfor 34 wk. Caution, benzidine is a carcinogen. 4. 30% (w/w) 1$02 solution.

2.8. In Vitro Culture

Reagents

2.8.1. 2% Methylcellulose
1. Add 20 g methylcellulose to 500 g sterile glass freshly double-distilled deionized boiling water. 2. Boil for 2 min with great care to avoid excess foaming. Control the level of heating. 3. Add sterile double-distilled deionized water to 520 g to correct for water loss resulting from evaporation. 4. Cool to approximately 37C. 5. Add 500 mL of double-strength IMDM. Keep covered with foil. 6. Cool on ice with mixing for 2-3 h. 7. Stir on a magnetic stirrer at 4C overnight to allow the methylcellulose solution to clear. 8. Store in 50-mL aliquots protected from light at -2OOC for up to 4 wk.

CFU-E Isolation

351

1.
2. 3. 4. 5. 6. 7. 8. 9.

2.8.2. Deionized and Delipidated BSA Dissolve400 mg of Dextran T40 in 400 mL of glass double-distilled deionized water. Add 4 g of activated charcoal (Norit A or SX-1) and leave at room temperature for 30 min with occasional mixing. Add 20 g BSA to the surface of the dextran-coated charcoal solution and leave for 2-3 h at 4OCto dissolve without mixing. Titrate to pH 3.0 with concentrated hydrochloric acid to inhibit heatinduced polymerization of the albumin. Incubate for 30 min at 56OCin a shaking water bath. Centrifuge at 10,000 rpm for 20 min and Millipore filter the supernatant. Adjust the pH to 5.5 with 2M NaOH. Deionize the BSA solution overnight at 4OCover 40 cm3 of Amberlite MB-l mixed in exchange resin. Concentrate the solution to 150 mLon an Amicon UMlO membrane at at -2OOCor

4OC. 10. Adjust the pH to 7 with 2M hydrochloric acid. 11. Millipore filter and store in lo-mL aliquots indefinitely 4OC. 1.

2. 3.

4.

2.8.3. Lipids Dissolve 4 mg L-a-phosphatidyl choline dipalmitoyl, 3.9 mg cholesterol, and 2.8 mg oleic acid in a few drops of chloroform at room temperature or ethanol at 50C in a 25-mL glass beaker. Completely evaporate the solvent under a stream of nitrogen, leaving the film of lipid on the bottom of the beaker. Add 10 mL of bicarbonate-free IMDM pH 5.1 containing 1% of the deionized and delipidated BSA. Immerse the beaker containing the lipids in ice and sonicate under air for 10 min at maximumenergy just below the foaming point so that the lipids form small micelles. Millipore filter through 1.2 pm and then 0.45 pm filters. Store indefinitely at 4OC.

2.8.4. Other Reagents 1. 5 x 10-2M 2-mercaptoethanol in double-glass distilled deionized water. Millipore filter. Prepare fresh stocks before use. 2. Ferric chloride stock (FeCS). Since ferric chloride is hygroscopic,

352

Watt and Davis

weigh out a large piece of FeC1,*6%0 and immediately dissolve in 103M hydrochloric acid (HCl). Dilute to a 7.9 x lO3M stock solution in 1O3M HCl. Store at -2OOC. 3. Transferrin: Dissolve 360 mg of human transferrin in 4 mL of bicarbonate-free IMDM pH 7.4 and 1.15 mL of 7.9 x lOM FeCJ in 103M HCl. Millipore filter and store indefinitely at 4OC.

3. Methods
3.1. Antibody Preparation
1. Collect the supernatant from the hybridoma cell lines grown in IMDM with 1% FCS, 24 h after cells reach confluency. 2. Precipitate the immunoglobulin (Ig) by adding an equal volume of saturated ammonium sulfate solution (pH 6.8). Mix for 1 h at 4C. 3. Centrifuge at 10,OOOg for 10 min at 4OC. 4. Resuspend the pellet to one-tenth the original volume in PBS and dialyze against O.lM phosphate buffer pH 8.0 at 4OC. 5. Estimate the protein concentration by measuring the absorbance at 280 nm.

3.2. Antibody

Purification

1. Mix purified rabbit antirat Ig with Protein A-Sepharose to a final concentration of 11 mg Ig/mL of Sepharose beads in O.lM borate buffer pH 8.2 for 30 min at room temperature. 2. Wash the beads in excess borate buffer and then in 0.2M triethanolamine pH 8.2. 3. Resuspend the Sepharose in 20 vol of 20 mM dimethyl pimelimidate dihydrochloride freshly made in 0.2M triethanolamine pH 8.2. Mix for 45 min at room temperature. This will covalently crosslink the Ig to the protein-A and prevent it from leaching from the column. 4. Spin the beads at 5008 for 1 min and resuspend in an equal volume of 20 mM ethanolamine pH 8.2 for 5 min at room temperature. 5. Wash the beads three times in O.lM borate buffer pH 8.2. 6. At the same time, equilibrate Protein A-Sepharose with O.lM borate buffer pH 8.2. YBM 10.14.9 will bind to Protein A-Sepharose at neutral pH, whereas YBM 42.2.2 will not. 7. Apply the concentrated YBM 10.14.9 sample to a Protein A-Sepharose column and the YBM 42.2.2 to the rabbit anti-rat Ig-Protein A Sepharose column; lo-20 mg of Ig can be applied/mL of beads.

CFU-E Isolation

353

8. Elute the bound Ig with O.lM glycine-HCI buffer pH 3, and neutralize the eluted material immediately with Trizma base. 9. Dialyze the antibodies against PBS and store in small aliquots with 0.1% BSA. 10, If the purified antibodies are to be coupled to fluorescein isothiocyanate (FITC), dialyze against O.lM bicarbonate buffer pH 9.3 instead of the PBS, and couple with FITC immediately. Do not store antibodies for an extended period of time in the bicarbonate buffer.

3.3. Fluorescein

Labeling

of Antibodies

1. Dialyze the antibodies against O.lM bicarbonate buffer pH 9.3 for 2-5 h at 4OC. 2. Dissolve thefluoresceinisothiocyanateisomer 1 (FlTC) at 1 mg/mLin DMSO. Add 25 pg of FITC/mg of purified YBM 42.2.2 Ig for 2 h at room temperature with constant rotation to give a fluorescein to protein ratio of approximately 3:l. 3. Separate the FITC conjugated Ig from the free FITC by passing through a 5-mL Sephadex G-25 column equilibrated with PBS. Collect the Ig fraction and store in small aliquots containing 1% BSA and 0.02% NaN3 at -2OOC.

3.4. Cell Preparation

and Labeling

1. Dissect the livers from d 12-13 mouse fetuses using cataract knives and place in chilled bicarbonate-free IMDM containing 10% FCS. 2. Prepare a single-cell suspension by gently syringing the fetal livers through 19-, 21-, and 25-gage needles sequentially attached to a 2-mL syringe. 3. Place the cells in a IO-mL centrifuge tube and allow cell clumps to settle at 4OCfor 5 min. Pass the supernatant through sterile nylon gauze to remove smaller clumps. 4. Centrifuge 5 mL of cells (10 cells/ml) over 4 mL of Ficoll-Hypaque (density 1.077 g/mL) at 1600 rpm for 30 min at room temperature. 5. Collect the low density cells from the Ficoll-Hypaque interface and wash three times in Eagles-HEPES medium with BSA and sodium azide. Approximately lo5 cells are recovered/fetal liver processed. 6. Add normal mouse serum, heat inactivated at 56C for 30 min to the cell suspension at a final concentration of 10% to block Fc receptors. Incubate the cells for 20 min at room temperature, and wash the cells in Eagles-HEPES medium containing BSA and sodium azide.

354

Watt and Davis

7. Centrifuge the antibodies in a Beckman airfuge at 26 lb/i$for 10 min at room temperature to remove aggregrates and further minimize Fc binding. 8. Label cells (lO/mL) with saturating levels of FITC-tagged YBM 42.2.2 (approximately 400 pg/mL) at 4C for 30 min. Include propidium iodide (50 pg/mL) during the incubation. 9. Wash and resuspend the cells to 2 x lo6 cells/mL in Eagles-HEPES medium containing BSA and sodium azide at 4OC.

3.5. Cell Sorting

(See Note 1)

1. Separate the labeled cells on a fluorescence-activated cell sorter by the two parameters of forward light scatter and fluorescein fluorescence. 2. For a FACS-II cell sorter, cellular excitation is achieved with an argon ion laser at an output power of 0.3 W and an emission wavelength of 488 nm. Set the light scatter gain at approximately 4. For fluorescein fluorescence, set the photomultiplier voltage at 650 V with an amplifier gain of 8-16. These voltages and gains will vary with the instrument used. The fluorescein fluorescence is detected by placing a 530nm long pass interference filter and a 530-nm long pass filter in front of the appropriate photomultiplier tube. 3. Sterilize the tubing and nozzle by passing ethanol through the cell sorter for 30 min. Wash out the ethanol with sterile distilled water, and pass sterile distilled water through the tubing and nozzle for at least 1 h, following this with a 0.9% saline wash for 30 min prior to sorting. 4. Run the cells at 4C through the cell sorter, collecting cells with intermediate to high forward light scatter characteristics that are negative for YBM 42.2.2 labeling when compared to control cells labeled with an irrelevant FITC-tagged rat monoclonal antibody of the same subclass. SeeFig. 1A and 1B. 5, Collect the sorted cells in bicarbonate-free IMDM pH 7.3 containing 20-50% FCS at 4OC in earthed siliconized glass tubes. Note: CFU-E will die if the collection medium becomes alkaline. 6. Spin the sorted cells at 800 rpm for 15 min at 4OC. 7. Resuspend the cells to 2 x lo6 cells/ml in Eagles-HEPES medium with BSA and sodium azide. 8. Label with FITC-tagged YBM 10.14.9 or with unlabeled YBM 10.14.9 (using 400 pg/107 cells) for 30 min at 4OC. 9. Wash the cells twice with Eagles-HEPES medium with BSA and sodium azide. Resuspend to 2 x lo6 cells/ml in the same medium.

CFU-E Isolation
A

355

1sT SELECTION

FORWARD

LIGHT

b SCATTER

SORTED 0.

2ND
Lymphold, myelold & early progen,tor cells

SELECTION 2 2 NEGATIVE

FITC-YBM42

Erythrold cells

SORTED

CFU-E

3RD

SELECTION 14 9 POSITIVE

FITC-YBMlO

100 FLUORESCENCE b INTENSITY

256

Fig. 1. Typical histograms showing the regions selected for isolating CFU-E. The fetal liver cells are first selected on the basis of their intermediate to high forward light scatter characteristics (A) and the YBM 42.2.2 negative cells (B) are sorted. The isolated cells are labeled with FITC-tagged YBM 10.14.9, and the positive cells (0 are collected. Care is taken to exclude the highly positive cells that are stained with propidium iodide and represent the nonviable cell fraction.

10. When using the unlabeled YBM 10.14.9, a two-stage indirect labeling procedure is necessary. For this, incubate the cells with FITC-Protein A (20 pg/mL final concentration) for 30 min at 4OCprior to washing, and resuspension in Eagles-HEPES medium with BSA and sodium

356

Watt and Davis

azide. Include propidium iodide (50 pg/mL) in the final incubation step. Dead cells labeled with propidium iodide give a very high fluorescence signal in the fluorescein channel and can be gated out when only two parameters are available for sorting. 11. Sort the labeled cells selecting the YBM 10.14.9 positive cells (avoiding the very highly propidium iodide labeled nonviable cells) using conditions described for the first sort and shown in Fig. 1C. 12. Centrifuge the sorted cells at 800 rpm for 10 min at 4OC. Resuspend in bicarbonate-free IMDM containing 10% FCS at 105cells/mL. Keep at 4OCuntil cultured.

3.6. CFU-E Culture

(See Note 2)

1. Thaw the 2% methylcellulose stock at room temperature. 2. Set up triplicate cultures in 35-mm plastic Petri dishes in a final vol of 1 mL containing: 0.8% (w/v) methylcellulose, 10% (w/v) FCS, 1% (w/v) deionized and delipidated BSA, 0.3 mg transferrin saturated with FeCS, 5 xlOsM 2-mercaptoethanol, 0.05 U erythropoietin, 8 PgLa-phosphatidylcholine dipalmitoyl, 7.8 pg cholesterol, 5.6 pg oleic acid, and lo*-lo3 sorted fetal liver cells or 104unsorted fetal liver cells all in single-strength IMDM pH 7.3. 3. Incubate the cultures at 37OC in a humidified incubator gassed with 5% CO, in air for 2 d. 4. Score single- or double-cell clusters containing 8-64 mature or maturing erythroid cells using an inverted microscope with a loo-fold magnification.

3.7. Morphology
1. Cytocentrifuge sorted cells and air dry. 2. Fix the cells with methanol for 10 min at room temperature.

3. Incubate in May-Grunwald stain for 20 min at room temperature and then in 3% Giemsa in tap water for 20 min at room temperature.
4. Wash the slides in tap water. Air dry and mount the coverslips with DPX.

3.8. Ben&dine

Staining

1. Suspend the cells in tissue culture medium (without NaN3) to a concentration of 1.5 x lo6 cells/mL.

CFU-E Isolation

357

2. Pipet 150 PL of the cell suspension into the well of a 96-well-flat bottom microtiter plate. 3. Prepare the staining solution; add 10 PL of 30% I$OZ to 1 mL of benzidine stock solution. Mix and use within 30 min. 4. Add 50 PL of staining solution to cells, and mix quickly by pipeting up and down several times. 5. Wait 5 min. At this time, cells containing at least 10% hemoglobin will have stained a dark blue. Estimation of the number of hemoglobin containing cells is best achieved by photographing the cells at this stage, since both the color of the stain and the number of stained cells will change with time. 6. The early erythroid blast cells represent the most immature and largest erythropoietic precursors recognizable, containing a prominent nucleolus, basophilic cytoplasm, and loose chromatin pattern, and are negative for benzidine staining. More mature erythroid cells containing hemoglobin will stain with benzidine.

4. Notes
1. The method describes the isolation of CFU-E using sequential sorting with two monoclonal antibodies, YBM 42.2.2 and YBM 10.14.9. Relatively high recoveries of CFU-E (40%) can be achieved in this way. These studies could be done equally well using a single multiparameter sort with antiisotypereagents labeled with fluorescein and phycoerythrin or directly conjugated reagents. Substantial enrichment for CFU-E from both normal fetal liver and bone marrow can also be achieved with a set of monoclonal antibodies listed in reference 7. Indeed, more efficient purification may be obtained for CFU-E by combining three probes, such as YBM 10.14.9, YBM42.2.2, and YW 13.1.1, since all these antibodies exhibit different patterns of reactivity with normal hemopoietic cells. Studies using simultaneous two- and threecolor sorting with a variety of antibodies to human erythroid precursors show the potential benefit of such approaches to cell fractionation (8). Other procedures that allow substantial enrichment for CFU-E include multiparameter sorting using fluorescein-conjugated pokeweed mitogen and rhodamine-conjugated antineutrophil/monocyte antibodies (9). In addition, Nijhof and Wierenga (10) have obtained sufficient numbers of highly purified CFU-E for biochemical analysis

358

Watt and Davis

in a relatively short time by using density separation and elutriation of spleen cells from thiamphenol-treated mice. 2. The isolated cells are analyzed for CFU-E by their growth in methylcellulose (3). Details of methylcellulose preparation are also given in reference Il. Fetal calf serum can be omitted from the cultures, since the BSA, lipid, transferrin, and erythropoietin additives have been designed to allow CFU-E growth in serum-free conditions (3). The approximate concentrations of each additive are described, but it is essential to test and titrateeach additive in order to obtain the best conditions for CPU-E growth. 3. Details of antibody purification are given in reference 12. The benzidine staining technique described here was adapted from reference 13. Single cells from the purified CFU-E can also be sorted directly into 150 PL of the methylcellulose-supplemented culture medium in a microtiter tray. The colonies are allowed to develop for 2 d at 37OC, and the maturing erythroid cells can be stained after cellulase digestion of the methylcellulose (14). For this, 75 PL of FCS containing 1.08 mg/mL cellulase (1943 cellulase U/g> is added to each well, and the cultures are incubated overnight at 37OC. The following day, the contents of each well are transferred as drops to a glass microscope slide and allowed to air dry. The cells are fixed and stained with MayGrunwald/Giemsa stain. Alternatively, the cells may be stained in situ with benzidine with or without digestion of the methylcellulose with cellulase.

References
I. Gregory, C. J. (1976) Erythropoietin sensitivity as a differentiation marker in the hemopoietic system: studiesof three erythropoietic colony responses in culture. 1.
Cell Physiol. 89,289-301. 2. Eaves, A. C. and Eaves, C. J. (1984) Erythropoiesisin culture, in Clinics in Hematology (McCulloch, E. A., ed.), WB Saunders, London, pp. 371-392. 3. Iscove, N. N., Guilbert L. J., and Weyman, C. (1980) Complete replacement of serum in primary cultures of erythropoietin-dependent red cell precursors (CFU-E) by albumin, transferrin, iron, unsaturated fatty acids, lecithin and cholesterol. Exp. Cell Res. 126,121-126. 4. Watt, S. M., Gilmore, D. J., Metcalf, D., Cobbold, S. J., Hoang, T. K., and Waldmann, H. (1983a) Segregation of mouse hemopoietic progenitor cells using the monoclonal antibody YBM/42. J. Cell Physiol. 115,37-45. 5. Watt, S. M., Metcalf, D., Gilmore, D. J., Stenning, G. M., Clark, M. R., and Waldmann, H. (198313) Selective isolation of murine erythropoietin-responsive progenitor cells (CFU-E) with monoclonal antibodies. Mo2. Biol. Med. 1,95-115.

CFU-E Isolation

359

6. Watt, S.M., Gilmore, D. J.,Clark, M. R., Davis, J. M., Swirsky, D. M., and Waldmann, H. (1984) Hemopoietic progenitor cell heterogeneity revealed by a single monoclonal antibody YW13.1.1. Mol. Biol. Med. 5351-368. 7. Watt, S. M., Gilmore, D. J., Davis, J. M., Clark, M. R., and Waldmann, H. (1987) Cell surface markers on hemopoietic precursors. Reagents for the isolation and analysis of progenitor cell subpopulations. Mol. Cell Probes 1,297-326. 8. Loken, M. R., Shah, V. O., Dattilio, K. L., and Civin, C. I. (1987) Flow cytometry analysisof humanbone marrow: I.Normalerythroiddevelopment. Blood69,255-263. 9. Metcalf, D. and Nicola, N. A. (1984) The regulatory factors controlling murine erythropoiesis in vitro. Proceedings, NIH Conference on Aplastic Anemia, Airlie House (Young,N.S.,Levine,A.S.,andHumphries,R.K.,eds.),Liss,NewYork,pp.93-105. 20. Nijhof, W. and Wierenga, P. K. (1983) The isolation and characterization of the erythroid progenitor cell: CFU-E. J. Cell Biol. 96,386-392. II. Davis, J. M. (1986) A single step technique for selecting and cloning hybridomas for monoclonal antibody production. Mefhods in Enzymology 121,307-322. 12. Schneider, C.,Newman,R. A.,Sutherland, D.R., Asser,U., and Greaves,M. F. (1982) A one step purification of membrane proteins using a high efficiency immunomatrix. 1. BioZ. Chem. 257,10766-10769. 23. Orkin, S. H., Haroshi, S.I., and Leder, I. (1975) Differentiation in erythroleukemic cells and their somatic hybrids. PYOC.NatZ. Acad. Sci. USA 72,98-102. 24. Shillingstad, R. B. and Ragan, H. A. (1987) Cellulase slide preparation of methylcellulose cultures of hemopoietic cells. BZood CeZZs 12,657-660.

Chapter 30 Production of Human and Murine Eosinophils In Vitro and Assay for Eosinophil Differentiation Factors
Malcolm Strath, Elaine J. Clutterbuck, and Colin J. Sanderson
1. Introduction
Bone marrow cells from a number of animal species have been used extensively in liquid and semisolid cultures to study hemopoiesis and to produce functional mature cells and factor-dependent cell lines (for review seeref. I). Neutrophils and macrophages are produced without added growth factors from murine long-term bone marrow cultures (Z), while lymphoid cells (3,4) and megakaryocytes (reviewed in ref. 5) can be induced under certain conditions. When bone marrrow cells from mice or humans are established in tissue culture in the presence of eosinophil differentiation factor (EDF), mature functional eosinophils are produced and liberated into the non361

362

Strath, Clutterbuck,

and Sanderson

adherent cell population (6,7). This factor was originally proposed as interleukind (a), but is now accepted as interleukin-5. Other names for IL5 are eosinophil colony stimulating factor (CSF-Eo), B-cell growth factor type II (BCGFII), and T-cell replacing factor (TRF). Eosinophils are produced for only a relatively short time when marrow is cultured in the presence of IL5. This suggests that IL5 stimulates the differentiation of eosinophils from existing progenitor cells present in the marrow with no recruitment of eosinophil progenitor cells from stem cells. When the animal has a pronounced eosinophilia, subsequent culture of the marrow results in enhanced eosinophil production compared to cultures established from normal marrow. This is interpreted as an increase in eosinophil progenitors in the animal. We use the Helminth parasite Mesoc&&es corti Hoeppli 1925 to effect such an increase in eosinophil progenitor numbers in the mouse. This parasite has been used extensively to study host/parasite interactions and causes a well-documented eosinophilia (9), with peak numbers of eosinophil progenitors appearing in the bone marrow 10 d after infection (10). The assessment of eosinophil numbers from cultures can be done in one of two ways: (1) total cell counts (obtained from an electronic particle counter, e.g. Coulter counter, or by using a hemocytometer) and the percentage of eosinophils (obtained from a differential count on Giemsastained smear or cytocentrifuge preparation) are used to calculate the number of eosinophils, or (2) by assaying for eosinophil peroxidase, which can be related to eosinophil numbers (12). A microplate modification of the culture system (12) provides a method for assaying sources of IL5 since there is as yet no IL5-dependent cell line as there are for IL2 and IL3. This IL5 assay is based on assessment of eosinophil numbers by quantifying eosinophil peroxidase in microplate bone marrow cultures that have been incubated with IL5 samples. This assay also detects the eosinophil differentiation activity of GM-CSF and IL3. The B cell activity of mouse IL5 can be measured using the BCL, cell line (8) though this assay also detects interleukin-4 (B-cell-stimulatory factor [BSFl]).

2. Materials
1. Parasite: The second stage larvae (tetrathyridia) of the Cyclophyllidean Cestode M. corfi are maintained by intraperitoneal passage in mice, where it reproduces itself vegetatively. The larvae can be stored in Dulbeccos Phosphate Buffered Saline A (PBS) at 4OCfor several weeks, and for longer periods if fetal calf serum is added (13). To our knowledge there have been no recorded cases of human infections

Human

and Murine Eosinophils

363

with M. corti, although there have been several cases of infection by other species of Mesocestoides (24,25). Parasites for reinfecting mice are harvested from mice that have been infected for several weeks. The only noticeable visible effect of the infection is an increasing abdominal distension. It may be necessary to occasionally passage the parasite into another strain of mouse (e.g., CBA) as it seems to lose the ability to stimulate high eosinophilia after several months passage through the same strain. For availability of the parasite, refer to International Register of Living Helminth Species and Strains, published by WHO; or Register of Parasitic Protozoa, Helminths, and Arthropods of Medical and Veterinary Importance, produced by the British Society for Parasitology. 2. Mice: We routinely use Balb/c.nimr mice, 6-8 wk-old, which are maintained under SPF conditions and are allowed free accessto food and water. Any infections the mice contract before the marrow is harvested for culture may result in changes in the cell numbers and types seen in the marrow cultures. 3. Media: The basic medium used is RPM1 1640 purchased in powder form and reconstituted as recommended by the manufacturer. We use two basic forms of RPM1 as indicated below: (1) HEPES Medium HEPES buffer Sodium bicarbonate Glutamine Sodium pyruvate Penicillin Streptomycin Monothioglycerol (Sigma) 20 mM None 2mM None 100 U/mL 100 pg/mL None (2) Culture Medium 1OmM 24mM 2mM 1m.M 100 U/mL 100 pg/mL 75 w

The above media are stable at 4C for several months. Glutamine is unstable, so fresh glutamine is added from a frozen (-20C) 200 mM stock solution and the medium used within 2 wk. Glutamine is a general requirement in tissue culture medium and has been found to be necessary for hemopoietic cell differentiation (26). The media are further supplemented and used as follows: a. Bench medium: HEPES medium is supplemented with newborn calf serum to 5% v/v. This is used for cell and tissue collections and preparation, and is stored at 4C. It has the advantage over bicarbonate buffered medium that it does not change its pH while outside a CO, environment.

364

Strath, Clutterbuck,

and Sanderson

b. Bone marrow culture medium; Culture medium is supplemented with lO4M hydrocortisone and 15% fetal calf serum. The pH of this medium is maintained by culturing in an atmosphere of 5% CO, in air. (1) Media additives: (a) Hydrocortisone: A UYzM (48.45 mg/mL) stock solution of hydrocortisone sodium succinate in PBS is filter sterilized (0.22 pm pore) and stored in 50-100 PL aliquots at -20C. 10 PL stock solution is added to 100 mL bone marrow culture medium. Any remaining stock hydrocortisone solution is not refrozen, but discarded. (b) Fetal calf serum: This has to be selected for optimal eosinophil growth, since some batches result in negligible eosinophil production in cultures. Whether this is caused by inhibitors or lack of growth factors is not known. It is preferable to test the fetal calf serum in both the EDF assay and in long-term cultures. (c) Gentamicin: A stock solution of 5 mg/mL gentamicin sulfate in PBS is filter sterilized, stored at -20C, and diluted 1:lOO into medium for use. Gentamicin is occasionally required when assaying column fractions for IL5. 4. Sources of Interleukin-5: Native murine IL5 can be obtained from mitogen-or antigen-stimulated spleen cells (Zl), T cell clones (17), or the EL4 lymphoma cell line (18). However, these sources also contain other lymphokines. The T cell hybrid NIMP-THl produces IL5 in the apparent absence of other known lymphokines (29) and has provided the most useful source of IL5 until recombinant material from transfected monkey COS cells became available (20). There is no known source of native human IL5 and work on human eosinophils has been based on the cross reactivity of murine IL5 (21). However, recombinant human IL5 is now available from transfected COS cells (22). IL5 has also been detected in the serum of parasitized animals (B), and recent work has demonstrated the presence of a factor stimulating human eosinophil differentiation in serum from patients undergoing eosinophilia (23). Interleukin-3 and GM-CSF have some eosinophil differentiation activity (6) so limited number of eosinophils can be produced in cultures using WEHIconditioned medium or commercially available GM-CSF. When sources of IL5 such as crude spleen conditioned medium that contain other lymphokines as contaminants are used, large numbers of neutrophils and/or macrophages are produced so that eosinophils represent only a small percentage of the nonadherent cells.

Human

and Murine Eosinophils

Methods detailing production of the above sources of IL5 are given in ref. 24. 5. Eosinophil cultures: A Class II Microbiological safety cabinet is required for human cultures, and all waste materials should be autoclaved. Murine cultures are established under standard tissue culture conditions. A gassed 37C incubator is required, which has to be humidified for agar, cluster, and microplate cultures. a. Agar cultures: We use Difco Bacto-Agar that is preselected for colony growth. A 5% w/v stock solution is prepared by suspending the agar in distilled water and placing in a boiling waterbath for 5 min. The agar is stored at room temperature. We do not autoclave the agar since this seems to introduce some toxicity into the system. Leukocyte migration plates (Sterilin, Teddington) are used in place of Petri dishes as this allows the whole agar culture to be easily recovered and stained for assessing colony number and type. The small culture volume (400 PL) results in savings in materials and reagents. A square 1OOmMPetri dish is used as the container for the leukocyte migration plate. b. Long-term cultures: These cultures are established in either flasks or Cluster plates. We use 25-80 cm2 flasks from which large numbers of eosinophils can be produced for functional studies. Most experimental work is done using 24-well Cluster plates. A cytocentrifuge is useful to prepare slides for staining and differential counts, the morphology being clearer than on smears. c. Microplate cultures: The cultures are established in 96-well round (U) bottomed microtiter plates. The flat bottomed 96well microtiter plates are not satisfactory for this assay system. 6. Eosinophil Peroxidase Assay a. Peroxidase buffer: 0.05M Tris-HCl pH 8.0. Filtered (0.45 pm pore) and stored at room temperature. b. o-Phenylenediamine (OPD): a stock solution of 10 mg/mL in distilled water is stored in 1 mL aliquots at -70C, where it is stable for several months. c. 30% w/v hydrogen peroxide stored at 4C. d. Triton X-100: a 10% stock solution in water is stored at 4C. e. 4M sulfuric acid. f. Complete substrate solution: to 48.5 ml peroxidase buffer add 1 mL OPD stock solution, 0.5 mL Triton X-100 and 6 PL hydrogen peroxide. OPD is light sensitive and so this solution

366

Strath, Clutterbuck, and Sanderson

should not be prepared until it is required. An automatic microplate reader with a 490~nm filter is recommended for reading the plates. 7. Eosinophil Stains a. Giemsa: Buffer; Sorensens buffer concentrate pH 6.8, diluted to 3.3 mM in distilled water. Giemsa (10%) stock solution is diluted 1:5 with buffer for use. b. Congo Red: Dissolve 5 g Congo Red in 50 mL distilled water then add 50 mL ethanol. The solution is stable and can be reused. Different batches of Congo Red seem to vary in their ability to stain eosinophils. c. Toluidine Blue: Add 1 g Toluidine Blue to 100 mL methanol. Acidify by adding 5 mL2MHCl. The solution is stable and can be reused. d. Luxol-Fast-Blue: Add urea to 70% ethanoluntil saturated (approximately 250 g/L), then filter through a Whatman No. 1 filter paper. Dissolve 1 g Luxol-Fast-Blue in 100 mL of urea sat-

urated 70% ethanol. The solution is stable and can be reused. e. Harris Hematoxylin: Dissolve 1 g hematoxylin in 50 mL ethanol, and 100 g aluminum ammonium sulfate (or aluminium potassium sulfate) in 1 L distilled water (with gentle heating). Add the hematoxylin solution to the salt solution and bring to the boil rapidly. CAREFULLY add 2.5 g mercuric oxide (a violent reaction may occur if added too rapidly) and allow to cool. Filter. Add 4 mL glacial acetic acid to each 100 mL of stain and store at room temperature. The stain can be reused extensively.

3. Methods
3.1. Parasite Passage
With the parasite in a plastic Universal tube in PBS: 1. Aspirate 100-200 PL of PBS into a 1 mL syringe and expel the air bubble. 2. Allow the parasites to settle at unit gravity. 3. Insert the end of the syringe into the parasite pellet and fill the syringe. 4. Invert the syringe and allow the parasites to settle. Expel excess PBS and refill with more parasites if necessary. 5. Fit a 0.8 x 40 mm needle and expel air and excess PBS. 6. Inject 100 JJLparasite ip into each mouse.

Human and Murine Eosinophils

367

3.2. Parasite

Harvest

1. Kill the mouse by cervical dislocation. 2. Cut the abdominal skin and expose peritoneal wall. 3. Using a 5 or 10 mL syringe with a 0.8 x 40-mm needle inject 5-10 mL of PBS into the peritoneum. 4. Withdraw the parasite/PBS into the syringe. 5. Expel parasite/PBS into a sterile plastic Universal bottle. 6. Wash the parasites several times by filling the bottle with PBS, allow the parasites to settle, and pour off supernatant. 7. Use the parasites to infect more mice (seeSection 1) and store the remainder at 4OC.

3.3. Marrow
1. 2. 3. 4. 5. 6.

Collection

and Cell Preparation

7. 8. 9. 10.

3.3.1. Mouse Marrow (See Note 1) Fill a suitable container with 70% ethanol and immerse a pair of scissors and forceps. Use a tissue to dry the instruments before use and replace them into the alcohol between each procedure. Kill mice by cervical dislocation. Fill a 5-mL syringe with bench medium and fit a 0.4 x 12-mm needle. We use the syringe type, which is packed in an outer polypropylene case, the case being used as a sterile home during the procedures. Wet theanimalsfur with70% ethanol and place theanimalonits back. Pull up the abdominal skin with forceps and make a cut across the abdomen with the scissors held vertically. Pinch the skin on each side of the cut between thumb and forefinger, and enlarge the incision by pulling the skin toward the head and tail. The incision should enlarge around the animal and the skin finally break at the animals back. Continue to pull the skin back toward and over the tail and legs until the muscles of the upper and lower legs are exposed. Complete the skinning by grasping the tibia/fibula with forceps and the skin with the other hand and pulling the leg from the skin until the foot is clear and the whole leg completely skinned. Remove the legs from the animal by a combined movement that both dislocates the leg from the pelvis and cuts the muscles, ideally leaving the leg complete with femoral head. Holding the femur with forceps, cut across the femur with the scissors to remove the head. Flush out the marrow by holding the femur with forceps (foot held

368

Strath, Clutterbuck, and Sanderson

upwards) and inserting the needle as far as possible up into the femur, being careful not to put the needle through the knee joint. With the needle in the femur, bend the needle to an angle of about 45Oto prevent the cells running down the outside of the needle and onto the syringe. Position the leg above the collection tube and when the medium is injected into the femur thecells will be flushed out and drip into the tube. The cells often congregate as a lump at the end of the bone, so withdraw the needle and touch the cut end of the femur onto the tube and dislodge the cells using the syringe needle and flush them into the tube. Pool the marrow from several mice into one tube. Centrifuge the cells (3008 for 8 min) and remove the supernatant. Resuspend the cells in 5 mL of medium and break up any large cell clumps by repeated aspiration into a 10 mL pipet. If the cells are not to be established in culture immediately, then resuspend them in bench medium; otherwise use bone marrow culture medium. Aspirate the cells into a 10 mL syringe via a 9.8 x40-mm needle and expel them back into a tube through a 0.4 x 12-mm needle. Count the cells and adjust to the required density. A differential count can be done to assess the degree of marrow eosinophilia (seeNote 2).

11. 12.

13. 14. 15.

3.3.2. Human

Marrow
the iliac crest or the ribs removed during purposes requires the of the local Hospital

Marrow can be obtained from aspiration of sternum (during cardiothoracic surgery) or from thoracotomy. Use of marrow samples for research informed consent of the patient and the approval Ethics Committee. 1. 2. 3. 4. 5. 6.

Resuspend marrow in bench medium to 1-2 x 10 cells/mL. Layer 10 mL cells onto 10 mL Ficoll-Paque in Universal tubes. Centrifuge at 6008 for 35 min at room temperature. Collect theinterface cells (mononuclear cells) and add bench medium. Centrifuge (6008 for 10 min) and resuspend pellet in bench medium. Centrifuge (5008 for 10 min) and resuspend pellet in bone marrow culture medium at lo6 cells/mL.

Approximately 0.5-l .5 x 10 mononuclear cells are obtained from 1mL of normal marrow aspirate.

3.4. Eosinophil
3.4.1. Murine

Cultures
Cultures

Microplate

1. Adjust the marrow cell concentration to 1 x lo6 cells/mL in bone marrow medium.

Human

and Murine Eosinophils

369

2. Aliquot 10 j.tL of sample (or sample dilutions if titrating) into duplicate microplate wells. 3. Add 100 PL cells/well (including several wells without IL5 to use as control cultures) and incubatein a humidified atmosphere at 37OC and 5% CO, in air. 4. After 5 d assay the cultures for eosinophils by either total cell counts and differential counts or by peroxidase (seeSection 3.7). Human Microplate Cultures 1. Set up cultures as for mouse microplate cultures (steps l-3 above). 4. Every 7-10 d carefully aspirate 50 PL of medium and replace with fresh medium containing sample. 5. After 21-28 din culture, determine total cell numbers (Coulter counter or hemocytometer) and perform differential cell counts on a cytocentrifuge or smear preparation, or perform a peroxidase assay.
3.4.2.

1. 2. 3. 4. 5. 6. 7. 8.

3.4.3. Agar Cultures Melt the agar in a boiling waterbath and hold at 45OC. Adjust the bone marrow cells to 2 x lo5 cells/mL (or dilutions of cells if required) in bone marrow medium and hold in a 37OCwaterbath. Aliquot 40 PL of IL5 or other growth factors or their dilutions into each well of a leukocyte migration plate. Pipet 1 vol of agar into 15 vol of the cell suspension, mix well, and immediately aliquot 400 pL/well. Place the plate into the upturned lid of a lOO-mm-square Petri dish containing a moist piece of tissue or filter paper and cover with the square Petri dish base. Allow the agar to solidify, putting the plates for a short time at 4C if the laboratory temperature is high. Place in a humidified incubator at 37OCin 5% CO, in air. Fix and stain mouse cultures after 5-7 d, and human cultures after 14-21 d (Sections 3.53.6).
3.4.4.

Long-Term Cultures (See Note 6) 1. Adjust the bone marrow cells to 1.5 x lo6 cells/mL. If large numbers of eosinophils are required within 1-wk use marrow from parasitized animals, 9-20 d postinfection. 2. Aliquot the cells into flasks or Cluster plates, 10 mL of cells for a 25 cm2 flask, 30 mL for an80 cm* flask or 1 mL/well for 24-well cluster plates. 3. Add IL5 at a predetermined optimal dilution and place in a 37C incubator in 5% CO, in air. The Cluster plates will require a humidified incubator.

Strath, Clutterbuck,

and Sanderson

4. Every 6-7 d, gently tap the cultures to dislodge the nonadherent cells and remove all the supernatant and nonadherent cells. 5. Add fresh medium containing IL5 and place back into the incubator. 6. Determine the total cell number of nonadherent cells and prepare a smear or cytocentrifuge preparation.

3.5. Fixing

the Agar

Cultures

We find it necessary to stain the cultures to accurately count eosinophil colonies since assessing the cultures by morphology alone under the inverted microscope is misleading-we have found both tight and very loose colonies of eosinophils. Mouse cultures are stained using Congo Red and counterstained with either Harris Hematoxylin or Toluidine Blue. We have had little success using Luxol-Fast-Blue for staining mouse agar cultures. Human cultures can be stained with any of the above stains including Luxol-Fast-Blue. Colonies are defined as clusters containing more than 40 cells. 1. Set the leukocyte migration plate at an angle of about 30 in a retort stand. With a slide held horizontally and resting against the rim of a well, direct the agar disk onto the slide by flushing with a stream of PBS from a wash bottle. It is possible to mount three agar disks on one standard 26 x 76 mm microscope slide. 2. Cover the agar disk(s) with a piece of dry Whatman No. 1 filter paper and put onto a warm plate to dry. Do not have the warm plate too hot. 3. Remove the filter paper just before it is completely dry and allow the agar disk to dry completely. 4. Fix in fresh methanol for 15 min before staining (Section 3.6.).

3.6. Eosinophil
1. 2. 3. 4. 5.

Stains

3.6.1. Harris Hematoxylin and Congo Red Place the fixed slide in acidified Harris Hematoxylin for 5 min. Wash under gently running tapwater until blued. Place in Congo Red and stain for 15 min. Wash briefly under gently running tapwater and dry. Nuclei stain blue, eosinophil granules red. Blue

3.6.2. Congo Red and Toluidine 1. Stain the fixed slide for 10 min in Congo Red. 2. Wash in 50% ethanol for 5 min. 3. Stain in Toluidine Blue for 5 min.

Human

and Murine Eosinophils

371

4. Rinse in water and dry. 5. Eosinophil granules stain reddish-brown, mast cell granules dark blue. Neutrophils and macrophages should not have any cytoplasmic staining. I. 2. 3. 4. 3.6.3. Luxol-Fast-Blue Stain the fixed slide for 1.5 h in Luxol-Fast-Blue. Wash under gently running tapwater for 3 min. Dry. Stain with Harris Hematoxylin for 2 min (30 s for cytocentrifuge
preps).

5. Wash under gently running tapwater for 3 min. 6. Nuclei stain blue, eosinophils have green granules.

3.7. Eosinophil

Peroxidase

(EPO) Assay (See Note 4)

1. Aspirate most of the medium from the bone marrow cultures, taking care not to suck out the nonadherent cells of the pellet. The medium does not interfere with the assay, but the wells become overfilled if it is not removed. 2. Add 100 u.L of substrate solution to each well and leave at room temperature for 30 min. It is not necessary to incubate in the dark for this short period. 3. Stop the reaction by the addition of 50 ~.LL of 4M sulfuric acid. 4. Determine the absorbance at 490 nm. Once the acid has been added the color is stable for several hours.

3.8. Fixing

and Staining Preparations

Cytocentrifuge

1. Ensure the cytocentrifuge preparation (or smear) is dry then fix in methanol for 5 min. 2. Stain the fixed slide in Giemsa for 2 min. 3. Wash under tapwater. 4. Blot dry. 5. Examine with oil immersion, a green Kodak Wratten filter No. 11 may assist in identifying eosinophils. 6. Nuclei are reddish-purple, eosinophils have red to orange granules. Basophilic granules are blue. The eosinophil stains in Section 3.6. can be used instead of Giemsa for staining cytocentrifuge preparations and smears.

Strath, Clutterbuck,

and Sanderson

40

30

20

10

0~

1 0

I 2

I 7

10

14

I 17

21

24

28

31

Days Post-infection

of Marrow

Harvest

Fig. 1. Number of eosinophils in supernatant from murine bone marrow cultures after 7 d incubation. Marrow was harvested from mice on different days post-infection with M. corti. All cultures were established in the presence of IL5 from a T-cell clone conditioned medium. Eosinophil numbers represent the mean f 1 standard deviation of three replicate cultures.

4. Notes
1. When taking bone marrow we work in the open laboratory, using alcohol dried instruments. There are very few problems with contamination at this stage. The animals fur is wetted with alcohol to prevent possible contamination by the fur being flicked. 2. The importance of using parasitized mouse marrow for eosinophil production is illustrated in Fig. 1. It shows the production of eosinophils after 7 d incubationusing marrow harvested from mice on different days after infection. Uninfected mice usually produce relatively few eosinophils. We generally use marrow from mice infected for 9-20 d. 3. Use of donor horse serum (DHS) in long term cultures: If a good batch of fetal calf serum has been obtained, we do not find it necessary to

Human

and Murine Eosinophils

The Production and Table 1 of Eosinophils in the Presence Days in culture 7 35.90 58i9 27.10 41 f 10 46.50 63*5 52.50 64f7 14 41.94 43f6 37.50 33*4 54.00 38f12 53.19 32f 18 21 42.3 Sk4 38.61 7f4 54.84 llf5 54.26 Sk3 28 42.33 0.5 f 1 38.63 0.3 f 0.5 54.96 2913 54.30 If1

373

Absence of Hydrocortisone and Donor Horse Serum

35 42.34 0.2 f 0.4 38.64 0.2 f 0.1 55.02 If1 54.30 o*o

Medium DHS

additives Hc

+ + :z
+
+

(a> 09 (a) W

(+) indicates medium supplemented with 5% donor horse serum (DHS) or hydrocortisone (Hc), f-1 indicates absence from medium. b(a) = Cumulative total of eosinophils x lod/mL. (b) = eosinophils as percent of total nonadherent cells (mean + 1 standard deviation of six duplicate cultures). Bone marrow
from parasitized mice was incubated in the presence of T-cell clone conditioned medium.

include DHS in the bone marrow culture medium. Table 1 shows the cumulative numbers of eosinophils produced over 35 d in culture together with the percent of nonadherent cells that are eosinophils. It can be seen from Table 1 that the presence of DHS suppresses eosinophi1 numbers, though the percentage of nonadherent cells that are eosinophils is increased because of reductions in the other cell types. Hydrocortisone has a short-term suppressive effect in cultures without DHS, but is beneficial for the longer term cultures. 4. The short-term microplate culture system is used to assayIL5 samples, to find the optimal amount of IL5 to produce eosinophils in the longerterm cultures, and to test fetal calf serum batches. In assaying for IL5 in column fractions, it is advisable to include gentamicin in the bone marrow culture medium. Use marrow frommice infected for between 9-20 d for establishing the cultures. Cells of the neutrophil lineage constitute 30-60% of total normal human marrow cells and, in the absence of a suitable growth factor, such cells do not survive in themicroplate cultures for longer than 21 d. Since the eosinophil perioxdase (EPO) assay detects human myeloperoxidase (even in the presence of the enzymes inhibitors such as KCN), the assay is only suitable for use in the human microplate culture system after 21 d culture with sam-

Strath, Clutterbuck,

and Sanderson

80-

60-

40-

20-

I
0

I
7

I
14

I
21

1
28

I
35

I
42

49

56

Days in Culture
Fig. 2. Production of eosinophils from bone marrow cultures established fromM. c&i infected mice. Cultures were incubated continuously in the presence of NIMP-THl conditioned medium from day 0 of culture (01; day 7 of culture (A); and day 14of culture (W. Eosinophil numbers represent the mean + 1 standard deviation of three replicate cultures.

ples that are known not to contain neutrophil growth factors. Figure 2 shows the production of human eosinophils (and total cells) from microplate bone marrow cultures stimulated with NIMP-THl conditioned medium. Samples that are cytotoxic and kill the marrow inoculum may appear positive in the peroxidase assay, since eosinophil peroxidase has not been degraded during the culture period. For this reason it is prudent to examine the cultures before assaying for EPO, and wells that have no visible cell pellet noted. 5. Agar Cultures: Some workers use slides pre-coated with 0.3% agar to mount their agar disks on. This is to ensure that the agar sticks to the slide and is not pulled off when the filter paper is removed. 6. Figure 3 shows the time course of eosinophil production of cultures established in the presence of NIMP-THl conditioned medium. The

Human

and Murine Eosinophils

375

14

21

28

35

42

49

Days in Culture
Fig. 3. Production of total cells G-1 and eosinophils (- - -1 from microwell cultures of normal human bone marrow established in the presence (Cl) and absence (0) of murine EDF (NIMP-THl conditioned medium).

marrow was harvested from mice infected for 14 d with M. corti with amarrow eosinophilia of 45%. When IL5 was added from day 0 of culture, most of the eosinophils were produced in the first 2-3 weeks when over 80% of the nonadherent cells were eosinophils (85.7f 5.1% at 7 d and 83.7 f 1.2% at 14 d. Without IL5 the eosinophils present in the marrow samples disappear at the end of 7 d, and addition of IL5 for the first time at this point ensures that subsequent eosinophils are produced in vitro from the precursors and are not surviving eosinophils from the inoculum. In cultures that have the addition of IL5 de-

376

Strath, Clutterbuck,

and Sanderson

Table 2 Number of Eosinophils Produced from Mouse Bone Marrow After 7 d Incubation in the Presence of IL5 Source of IL5 EDF NIMP-THl rmIL5 dilution Number % NumbelP % 1:lOO 63.36 f 6.99 52.2 f 5.1 nd 1:200 nd 20.52 f 1.89 25.9 f 3.0 1:300 22.00 f 0.40 28.0 f 2.6 nd 1:400 nd 18.49f 0.54 21.4 f 4.8 1:lOOO 3.95 f 0.30 7.0 f 1.1 22.60 f 1.56 34.3 f 1.8
1:3000

none

0.76 f 0.17 0.11 f 0.15

1.9fO.l 0.2 f 0.3

10.14 f 0.65

14.3 f 0.1

Number given as eosinophils x lV/mL. standard deviation of 4 replicate cultures.

Both number and percent given as mean rt 1

layed for 7 d, it takes another 14 d before peak eosinophil numbers are produced, when the eosinophils may represent up to 90% of nonadherent cells. Fewer eosinophils are produced if the addition of IL5 to the cultures is delayed for 14 d, when at peak production time the eosinophils only represent 27% of supernatant cells. Recombinant mouse IL5 (rmIL5) is active in stimulating the production of eosinophils in murine marrow cultures. Table 2 shows the effect of different concentrations of rmIL5 and NIMP-THl conditioned medium on eosinophil production after 7 d culture. It must be noted that eosinophil production is very variable between experiments, as well as having large variations between replicate cultures. This variation between replicates tends to increase as the cultures get older, particularly in Cluster plate cultures. This may represent cloning of the very young committed eosinophil progenitors that are present in the marrow at a low frequency. A total lack of eosinophil production in the presence of a known IL5 sample may be due to an unsatisfactory batch of fetal calf serum, or a fault in the medium or its preparation. 7. Other sources of eosinophil progenitors: Spleen and fetal liver are focal points of hemopoiesis. We have no experience with fetal liver, but spleen cells have been used in agar cultures and in the IL5 assay. Uninfected mouse spleen cells produced no eosinophil colonies when incubated in agar in the presence of EL4 conditioned medium, but spleen cells from mice infected for 12 d do produce eosinophil colonies. Cells from the same spleen gave a positive assay result when

Human

and Murine Eosinophils

377

used in theIL5 assay, but had to be partially purified and the EPO level above background (i.e., unstimulated cultures), was not as great as marrow cells.

References
1. Greenberger, J. S.(1984) Long term hematopoietic cultures, inHuematupoiesis(Methodsin Huemafology, vol. 11) (Golde, D. W., ed.), Livingstone, New York, pp. 203-242. 2. Motomura, S.,Katsuno, M., Kaneko, S.,and Ibayashi, H. (1983) The effect of hydrocortisone on the production and differentiation of granulocyte/machrophage progenitor cells in long-term marrow cultures. Exp. Hematol. 11,56-62. 3. Dorshkind, K. and Phillips,R. A. (1982) Maturational stateof lymphoid cellsin longterm bone marrow cultures. 1. Immunol. 129,2444-2450. 4. Aspinall, R. and Owen, J. T. T. (1983) An investigation into the B lymphopoietic capacity of long-term bone marrow cultures. Immunology 48,9-15. 5. Levin, J. (1983) Murine megakaryocytopoiesis in vitro: An analysis of culture systems used for the study of megakaryocyte colony forming cells and of the characteristics of megakaryocyte colonies. Blood 61,617-623. 6. Sanderson, C. J., Warren, D. J., and Strath, M. (1985) Identification of a lymphokine that stimulates eosinophil differentiation in v&o. Its relationship to interleukin3, and functional properties of eosinophils produced in cultures. J, Exp.Med. 162,6074. 7. Clutterbuck, E.J. and Sanderson, C. J. (1988)Human eosinophil haemopoiesis studied in vitro by means of murine eosinophil differentiation factor (IL5): Production of functionally active eosinophils from normal human bone marrow. Blood71, M6651. 8. Sanderson, C. J.,OGarra, A., Warren, D. J.,and Klaus, G. G. B. (1986) Eosinophil differentiation factor also has B-cell growth factor activity: Proposed nameinterleukin4. Proc.Natl. Acad. Sci.83,437-440. 9. Johnson, G. R., Nicholas, W. L., Metcalf, D., McKenzie, I. F. C., and Mitchell, G. F. (1979) Peritoneal cell population of mice infected withMesocesfoidescorti as a source of eosinophils. hf. Arch. Allergy Appt. Immunol. 59,315-322. 20. Strath, M. and Sanderson, C. J. (1986) Detection of eosinophil differentiation factor and its relationship to eosinophilia in Mesocestoides corti infected mice. Exp. Hemato1.14, X-20. 22. Strath, M. and Sanderson, C. J.(1985) Production and functional properties of eosinophils produced from bone marrow cultures. J. Cell Sci.74,207-217. 12. Strath, M., Warren, D. J.,and Sanderson, C. J. (1985) Detection of eosinophils using an eosinophil peroxidase assay. Its use asan assayfor eosinophil differentiation factors. J. Immunol. Methods 83,209-215. tetrathyridia under ad23. Mueller, J. F. (1972) Survival and longevity of Mesocesfoides verse conditions. J. Parasitol.58,228. 14. Turner, J. A. (1975) Other Cestode Infections, in Diseases Transmitted from Animals to Man (Hubbert, W. T., McCulloch, W. F., and Schnurrenberger, P. R., eds.), Charles C. Thomas, Illinois, pp. 708-744. (Cestoda) in a child in Mis25. Hutchinson, W. F. and Martin, J. B. (19801Mesocestoides sissippi treated with Paromomycin sulphate (Humatin), Am. 1. Trap. Med. Hyg. 29, 478,479.

378

Strath, Clutterbuck, and Sanderson

16. Dass, P. D., Murdoch, F. E., and Wu, H. -C. (1984) Glutamine promotes colony formation in bone marrow and HL-60 cells; Accelerates myeloid differentiation in induced HL-60 cells. In Vitro 20,869-875. 17. Sanderson, C. J., Strath, M., Warren, D. J., OGarra, A., and Kirkwood, T. B. L. (1985) The production of lymphokines by primary alloreactive T cell clones; a coordinate analysis of 223 clones in seven lymphokine assays. Immunology 56,575- 584. 18. Dutton, R. W., Wetzel, G. D., and Swain, S. L. (1984) Partial purification and characterisation of a BCGFII from EL4 culture supematant. I. Immunol. 132,2451-2456. 19. Warren, D. J. and Sanderson, C. J. (1985) Production of a T-cell hybrid producing a lymphokine stimulating eosinophil differentiation. Immunology 54,615-623. 20. Campbell, H. D., Sanderson, C. J., Wang, Y., Hort, Y., Martinson, M. E., Tucker, W. Q. J., Stellwagon, A., Strath, M., and Young, I. G. (1988) Isolation structure and expression of cDNA and genomic clones of murine eosinophil differentiation factor. Comparison with other eosinophilopoietic lymphokines and identity with Interleukin-5. Eur. J. Biochem. 174,345-352. 21, Lopez, A. F., Begley, C. G., Warren, D. J., Vadas, M. A., and Sanderson, C. J. (1986) Murineeosinophil differentiation factor: An eosinophil-specific colony-stimulating factor with activity for human cells. J. Exp. Med. 163,1085-1099. 22. Campbell, H. D.,Tucker, W. Q. J., Hort,Y., Martinson, M. E., Mayo,G., Clutterbuck, E. J., Sanderson, C. J., and Young, I. G. (1987) Molecular cloning, nucleotide sequence, and expression of the gene encoding human eosinophil differentiation factor (interleukin-5). hoc. Nutl. Acud. Sci. 84,6629-6633. 23. Kern, P., Horstmann, R. D.,and Dietrich, M. (1987) Eosinophil production in human bone marrow cultures induced by 80-85 kDa serum component(s) of patients with eosinophilia. Br. 1. Huemafol. 66,165-172. 24. OGarra, A. and Sanderson, C. J. (1987) Eosinophil Differentiation Factor and its associated B cell growth factor activities, in Lymphokines and Interferons: A p~uctical upproud (Clemens, M. J., Morris, A. G., and Gearing, A. J. H., eds.), IRL, Oxford, pp. 323-343.

Chapter 31 Clonogenic Assays for Hematopoietic and Tumor Cells Using Agar-Containing Capillaries
H. Rainer Maurer and Christine Echarti
1. Introduction
The so-called hematopoietic stem cells from mouse and human bone marrow can be induced in vitro to form colonies, provided that the appropriate hormones (colony stimulating factors, CSFs) and culture conditions have been selected. Such hematopoietic colony cultures are of interest for in vitro cytotoxicity testing of a great number of drugs. In fact, it has been increasingly recognized that serious side effects of many drugs involve injury to hematopoietic organs. Moreover, several hormones and inhibitors (including chalones) that regulate hematopoiesis have attracted clinical interest and their identity can be determined by such in vitro cultures. Therefore hematopoietic stem cells seem to provide identified targets to assay for either stimulatory, inhibitory, or cytotoxic drug effects. Based on the pioneering studies by the groups of Sachs, Bradley, and Metcalf, (seeChapters 23 and 26 in this volume) methods have been developed during the last decades to clone, in agar or methylcellulose, the stem
379

380

Maurer and Echarti

cells of virtually all hematopoietic cells (erythrocytes, lymphocytes, granulocytes, macrophages, megakaryocytes) (2). Moreover, similar clonogenie cell assays in agar or methylcellulose using human tumor cells have been developed (human tumor clonogenic cell assay; HTCA) and are widely used to study tumor biology, to predict the sensitivities of patients tumors to anticancer drugs and to determine the potential activity of new cytostatic drugs (2,3). The in vitro colony assays of most research groups utilize Petri dishes into which the stem cells are seeded in an agar-medium on top of another agar-layer serving as a feeder layer (seeChapter 26 this volume). Since this double-layer technique is tedious for large-scale routine work, our group since 1976, has adopted and greatly refined a microprocedure (4). Instead of Petri dishes, glass capillaries are utilized, and these offer a number of advantages: 1. Less materials are consumed (-lo%), which includes cells, media, and test samples, with the result of a lower cost per assay and an increased number of samples that can be tested. 2. No agar double layer: Thus an easier and quicker test setup. 3. Lower agar concentration (0.18 vs 0.3% in Petri dishes) with the result of better cell staining after culture. 4. Linear alignment of colonies: Thus easier and quicker colony counting (with a dissecting microscope) and the potential for optical evaluation by simple light-scattering densitometry (5). 5. Reduced risk of contamination by bacteria and fungi. 6. Higher cloning (plating) efficiencies of up to fivefold (6; seebelozu). The steps of the procedure for cloning cells in agar-containing glass capillaries are shown in Fig. 1 for granulocyte-macrophage (7) and Tlymphocyte (a), in Fig. 2 for human tumor clonogenic assays (9). The dimensions of the capillaries, open at both ends, are selected such that homogeneous colony growth is maintained along the entire gel. The coefficient of variation of colony numbers does not exceed +_5% from capillary to capillary (20). Filling of the capillaries is easy: The capillaries are put onto the tip of an automatic pipet, by which the cell suspension in liquified agar is sucked in. Optical evaluation by light scattering densitometry is suited for special tasks (10,11). For this the capillaries are drawn through the light path of a spectrophotometer by means of a densitometer attachment. The light scattered by the colonies inside the capillaries is received via two mirrors (5). The agreement of visual colony counting by microscope and

Clonogenic Assays
T- Lymphocyfe - Colony -Assay

381

Humon bloof - Ficoll-centrifugofion 4 PHA-slimulalion in suspension culture for 24 h /37 OC 4 Add/lion of lest sample and liquefied agor 1 SeedIng inlo capWvres 1 Colony -counting ofler 7 days Isopoque

;., .:..*. ;
I,....,,._.,.1

1
1

Granulocyfe
:>; li ** ; .-.

- Co/any
Mouse

-Assay
bone ,marrow ceils

i-:

Addilion o/ CSA, fesf sump/e and liquified ogar Seeding 1 Colony 1 inlo capillaries 1 (and clusfer aller 7 duys ) counfing

)I))

*a .*... . *, ::. - ..*, .etc.

Fig. 1. Standardcolony assaysin glasscapillaries. light-scattering densitometry is excellent. The latter method allows the follow-up of daily growth of individual colonies of one capillary (IO). The cloning efficiencies obtained with the agar capillaries were in all cases tested higher than those found with the standard Petri dish (35 mm) or microtiter plate technique (96 wells) (6). Potential drawbacks of the agar capillary technique should not be ignored, however. These include: 1. It is impossible to substitute agar for methylcellulose, since highly viscous solutions of the latter cannot be drawn into the capillaries. 2. The closed incubation form of the capillary does not permit administration of samples after start of incubation.

382
e---+ Pathology

Maurer and Echarti

Solid

tumor

1
A;;.:-..:\*

mechanically enzymat/cally

or

Single

cell suspension

..:; : ;,y :.: 8 .,.

Controt

+Cytostatic

(div. doses)

I hr rncubafron I 2 x Washing and resuspending, addition of agar (0.2 % ) 1

0 (I :..*a . . ...y.* . . . . ., *., .-..

0
a

. .*.*. . ..*.
*,.a;. ..a..

7 I
J 1 I

Seeding in to glass capillaries

7 -21 Days incubofion

J Counting of colonies

1.4 rlr~ak1

1 Evaluafron:
Sensitivity Index 1, 2-L AlOO

(>40 cells )

Fig. 2. Tumor stem cell assay as oncobiogram (cytostatic drug sensitivity testing in vitro).

3. Since fewer colonies are usually counted per capillary (20-60) than in Petri dishes, it is essential to determine the required number of cells to be seeded in order to obtain statistically reliable data. Though one or the other obstacle may be relevant under particular conditions, it is certainly worthwhile to balance the pros and cons in each case. Nevertheless major advantages of the agar capillary assays have been acknowledged for large-scale clinical studies (12).

Clonogenic Assays

383

In this chapter we describe a microclonogenic assay developed for mouse hematopoietic cells. A similar technique for human bone marrow cells has been reported in detail (13). Further, we describea procedure that allows the cultivation of T-lymphocyte colonies in agar-capillaries (8). The method has proven suitable to screen for calf thymus inhibitors (chalones) (14) and for drug side effects (15). Similar techniques have been developed for T-lymphocytes from mouse thymus (16) and for B-lymphocytes from mouse spleen (17). The lymphocytes are isolated from human peripheral blood and prestimulated with phytohemagglutinin for 24 h. After adding a test sample, the lymphocytes are seeded into agar-containing glass capillaries, and the effect of the sample on colony formation is evaluated after 7 d. Finally the microtechnique for cloning tumor cells is described. It works with established tumor cells as well aswith cells from fresh tumor explants or biopsy material. Methods to establish in vivo-in vitro correlation rates and to determine in vitro sensitivity indices have been described and discussed (18). The reader is particularly referred to the pharmacokinetic problems involved with physiologically relevant drug exposure (19).

2. Materials 2.1. Materials Needed in All Procedures

1. Glass capillaries: inner diameter, 1.38 mm; length, 126 mm. Before use, the capillaries are put in a beaker containing double distilled water and cleaned in an ultrasonic bath at 90C for 30 min, rinsed twice with double distilled water, dried, and sterilized at 180C for 2 h. 2. Capillary holders (seeref. 5) 3. Dissecting microscope with special stage holder for capillaries and indirect lighting. 4. Bacto agar: 3 g of agar are dissolved in 100 mL of double distilled water using a boiling water bath. The solution is then distributed into 5 mL aliquots and autoclaved at 121OCfor 15 min.

2.2. Materials

for the Granulocyte-Macrophage Assay

1. Male C57B1/6 mice, ideally 6-8 wk old, but not older than 6 mo. 2. Dulbeccos Modified Eagles Medium (DMEM) with glucose (1 g/L), bicarbonate (3.7 g/L), glutamine (4 x lo3 M), penicillin (500 U/mL), and streptomycin (500 yg/mL); pH 7.2.
3. Eagles MinimumEssentialMediumIscovemodification (MEMIscove)

384

Maurer and Echarti

with bicarbonate (3.0 g/L), glutamine (4 x lOaM), penicillin, and streptomycin; pH 7.2. 1:l mixture of DMEM and Hams nutrient mixture F 12 (DMEM/F12) with bicarbonate (1.1 g/L) glutamine (2.5 x lOaM>, penicillin (500 U/ mL), and streptomycin (500 pg/mL). Saline: 9.1 g of NaCl/L of double distilled water. Lipopolysaccharide W (LPS) from Salmonella typhosu. Horse serum: inactivated at 56OCfor 30 min and pretested for optimal colony growth; stored at -2OOC. Glutamine solution: 200 mM in saline, stored at -2OOC. Bovine serum albumin (BSA) solution: 100 mg/mL from Boehringer Mannheim. Transferrin solution: 30 mg/mL, l/3 iron saturated from Boehringer Mannheim (seeChapter 21 for preparation). Soybean lipids: 2 mg/mL liposome suspension from Boehringer Mannheim. Bovine insulin: 500 pg/mL in 12 mM HCl, stored at -2OOC. Turks staining solution: 10 mg of gentian violet are dissolved in a few mL of double distilled water. After adding 1 mL of glacial acetic acid, the vol is made up to 100 mL with double distilled water. Stored at room temperature. All media and solutions are stored at 4OCunless otherwise stated.

4. 5. 6. 7. 8. 9. 10. 11. 12. 13.

2.3. Materials

for the X-Lymphocyte

Assay

1. Dulbeccos Modified Eagles Medium (DMEM; see Materials for granulocyte-macrophage assay, item 2). 2. Hanks balanced salt solution (HBSS; see Appendix). 3. Saline: 9.1 g NaCl/L %O. 4. Ficoll-Paque (density 1.077 g/mL). 5. Phytohemagglutinin M @HA-M): dissolve contents of one vial in 5 mL of sterile double distilled water; 300 PL aliquots are stored at -2OOC. 6. Heparin solution: 5000 U/mL. All media and solutions are stored at 4OC unless otherwise stated.

2.4. Materials

for the Tumor

Cell Assay
(2.1 x

1. Female Balb / c mice, 6 wk old. 2. Rl?MI 1640 medium with bicarbonate (0.98 g/L), glutamine

Clonogenic Assays
104M), penicillin, and streptomycin; pH 7.2.

385

3. CMRL 1066 medium with bicarbonate (2.17 g/L) and glutamine (7 x 10 -4M); pH 7.2. 4. Enriched CMRL 1066 medium: Supplements for 100 mL of CMRL 1066 liquid medium: 7% NaHCO, solution, 3.1 mL. 2.2% CaC1, solution, 4.0 mL. 0.6% asparagine solution, 1.5 mL. 4.4% sodium pyruvate solution, 0.5 mL. 2.1% serine solution, 0.2 mL. 3% Tryptic Soy Broth (TSB), 10.0 mL. 2.9% glutamine solution, 10.0 mL. (All % = w/v.> Insulin (e.g. Depot-Insulin Hoechst; 10 mL ampule/400 IU solution for injection), 5.0 mL. Horse serum, 15.0 mL. Penicillin-streptomycin solution 50,000 IU/mL and 0.05 g/mL, respectively, 1.0 mL. pH 7.2. 5. Mineral oil (e.g. Esso Extra). 6. Metal filters 100 and 50 mesh. 7. Collagenase-DNase solution (in CMRL 1066 medium; ref. 21): collagenase type I, 0.15%, and DNase type III, Sigma, 0.015% (final concentrations); prepared freshly. 8. Dispase: Grade I, from Boehringer Mannheim, 2.4 U/mL; prepared freshly. 9. Horse (HS) and fetal calf (FCS) serum: inactivated at 56OCfor 30 min and stored at -2OOCin 1 mL aliquots. Sera should be pretested for optimal colony growth. 10. Saline: 9.1 g NaCl/L H20. 11. Trypan blue solution: 0.2% in saline. All media and solutions are stored at 4C unless otherwise stated.

3. Methods
3.1. Granulocyte-Macrophage Colony Assay
3.1.1. Preparation of Colony Stimulating Factor The type and source of colony stimulating factor (CSF) will determine which species of colony cells will develop in the cultures (see2 for more

386

Maurer and Echarti

details). Highly purified recombinant CSFs are now available from various commercial sources and may be used. In this section we describe a relatively simple procedure to prepare CSF from mouse-lung-conditioned medium (MLCM) according to a modified method described by Sheridan and Metcalf (20). Injection of endotoxin in vivo induces formation of CSF in mouse lung cells incubated in vitro. The CSF produced may then be purified from the culture supernatant (conditioned medium). Low molecular weight inhibitors, which are also formed, are removed by dialysis. 1. 60 to 70 strain C57B1/6 mice are injected intraperitoneally with 100 PL each of an endotoxin solution containing 50 pg/mL of LPS in sterile saline. 2. The mice are killed 4 h later by cervical dislocation. The lungs are removed and washed in saline. 3. Each lung is placed in a Petri dish containing 7 mL of DMEM and minced to small pieces with sterile scissors and forceps. The cultures are incubated at 37C and 7.5% of CO, in a fully humidified atmosphere for 48 h. 4. After incubation the conditioned medium (MLCM) is separated from the lung tissue pieces by filtration. The filtrate is inactivated at 56OC for 30 min. Precipitates formed are removed by centrifugation at 30,OOOg at 4C for 30 min. The supernatant is then dialyzed against saline for 3 d. The saline should be changed frequently. 5. It is recommended to concentrate the MLCM by l/3 of its volume by ultrafiltration through a YMlO membrane in an Amicon chamber at 4OC. The filtrate is discarded. The concentrated MLCM is sterile filtered, distributed into 1 mL aliquots and stored at -2OOC. 6. The optimal dose of CSFis determined by performing a dose-response curve. About 5-30 PL of MLCM/300 VL assay mixture usually are sufficient for maximal stimulation of granulocyte-macrophage colony growth. 3.1.2. Assay Granulocyte-macrophage colonies from mouse bone marrow may be grown in serum-containing and serum-free culture conditions. Both systems are described here. 1. Before use, add 100 PL of glutamine solution/l0 mL of MEM Iscove. 2. A mouse is killed by cervical dislocation, the femur removed, and adhering muscle tissue is scraped off with a scalpel. The bone marrow

Clonogenic Assays

387

3. 4. 5.

6.

7.

8.

cells are flushed out aseptically from the marrow cavity with 2 mL of culture medium (MEM Iscove for serum-containing cultures, DMEM/ F12 for serum-free ones) using a disposable syringe. The cells are suspended by forcing them through a al-gage needle several times. 90 PL of Turks staining solution are mixed with a 10 PL aliquot of cell suspension. The cells are counted in ahemocytometer (e.g., Neubauer chamber) and the remainder adjusted to 1.2 x lo6 cells/ml. For a serum-containing assay the following components are pipeted into sterile Eppendorf reaction tubes: a. Saline, 100 PL. b. Horse serum, 60 FL. c. CSF, 30 pL. d. MEM Iscove, 10 pL. e. Cell suspension, 25 pL. f. Agar-medium mixture, 75 pL. Final vol, 300 pL. The saline may be replaced by a solution of test agent (seeNote 1). For a serum-free assay (seeNote 2) the composition of assay mixtures is as follows: a. DMEM/F12,96 pL. b. BSA solution, 30 pL. c. Transferrin solution (diluted 1:lOO with D/F12), 35 pL. d. Soybean lipids, 3 pL. e. Insulin solution, 3 PL. f. L-Glutamine, 3 pL. g. CSF, 30 pL. h. Cell suspension, 25 PL. i. Agar-medium mixture, 75 pL. Total vol, 300 pL. The medium may be replaced by test agent solution. For preparation of the agar-medium mixture 3% agar is melted in a boiling water bath. 1.2 mL are mixed with 3.8 mL of the medium to be used prewarmed at 37OC. The mixture is then added to the components listed above, which have also been prewarmed to 37OC. After mixing carefully, three capillaries are filled with 75 PL of assay mixture from each Eppendorf tube. For this each capillary is put on a plasticpipet tip fitted onto an automatic microliter pipet (200 pL), and the assay mixture is sucked up. The capillary is then removed horizontally and one end lowered, so that the gel moves to the middle.

388

Maurer and Echarti

Then the capillary is left on a cooling plate for a few minutes, 9. For incubation the ends of the capillaries are wiped with 70% ethanol and the capillaries mounted in special capillary holders disinfected with 70% ethanol. 10. Incubation is for 7 d at 37OC, 7.5% of CO,, and relative humidity of < 100%. 11. Using a dissecting microscope granulocyte-macrophage colonies are counted at 15 x magnification (objective 1.5 x, eyepiece 10x). A colony is defined as an aggregate of more than 40 cells; smaller aggregates (840 cells) are termed clusters. There are both diffuse and compact colonies. In control capillaries without test agent 18-25 colonies are usually found. 12. From the colony counts of three capillaries the median and mean absolute deviation are calculated. The median of the untreated control (without test agent) is set equal to 100%. For dose-response curves the percentages of colony numbers compared to the control are calculated from the medians and mean absolute deviations of the assay mixtures. 13. Gels containing colonies may be transferred to microscope slides and, after drying, may be fixed with methanol and stained using the MayGriinwald-Giemsa method. Peroxidase staining with benzidine and H20z is also possible.

3.2. T-Lymphocyte
3.2.1. Prestimula

Assay
tion

1. 20 mL of peripheral blood are obtained by venipuncture from a healthy donor and mixed with 0.1 mL of heparin solution. 2. A 1 mL sample is removed to obtain erythrocytes (seebelow, step 5). The remaining blood (about 19 mL) is diluted with 13 mL of HBSS. 3. In each of 4 centrifuge tubes, 8 mL of the blood-HBSS mixture is carefully layered on top of 6 mL of Ficoll-Paque. The tubes are closed with parafilm and centrifuged at 400g at room temperature for 40 min. After this, four layers appear in the centrifuge tubes. The autologous plasma is removed with a sterile Pasteur pipet and kept at room temperature. 4. The lymphocyte layer is removed, washed twice with DMEM, and resuspended in 2 mL of DMEM, and the total number of lymphocytes

Clonogenic Assays

389

determined. From this the final volume of the prestimulation mixture is calculated. The final cell density is adjusted to 1.25 x lo6 cells/mL. 5. For prestimulating the lymphocytes, erythrocytes are added at a ratio of 1:l. For this purpose the erythrocyte concentration of the 1 mL blood sample has to bemeasured. A small aliquot is diluted 1:lOO with saline and counted. After determining the concentration the needed volume of blood (= erythrocyte suspension) is calculated. 6. The following components are also needed for prestimulation: autologous plasma: 375 pL/mL of final volume. PHA-M solution: 12.5pL/mL of final volume. To prepare the prestimulation mixture the calculated amounts of plasma, erythrocytes, PHA-M, and the 2 mL of lymphocyte suspension are mixed and made up to final volume calculated with DMEM. 7. The mixture is incubated in plastic tubes at 37OC,5% CO,, and I 100% humidity for 24 h. 1. 2. 3.2.2. Assay (See Note 3) The cells are washed twice with DMEM, resuspended in 1 mL of DMEM and readjusted to 1.2 x lo6 cells/mL. For the assay the following components are pipeted into sterile Eppendorf reaction tubes: a. Saline, 100 PL. b. Autologous plasma, 63 pL. c. PHA-M solution, 2 pL. d. DMEM, 23 /JL. e. Cell suspension, 37 PL. f. Agar-medium mixture, 75 pL. Final vol, 300 pL. The saline may be replaced by a solution of test agent. Preparation of the agar-medium mixture and filling the capillaries are performed as in steps 7-9 of the granulocyte-macrophage assay except that 30 PL of assay mixture are sucked into the capillaries. Incubation and evaluation are done as described in steps 10-13 of the granulocyte-macrophage assay. Colonies are counted at a magnification of 37x (objective3.7x, eyepiece 10 x). Normally40-60 colonies are counted per capillary, usually located at the bottom of the agar gel. A colony is defined as an aggregate of more than 50 cells. T-lymphocyte colonies are mostly compact (seeNote 4).

3. 4.

390

Maurer and Echarti

3.3. CZonogoenic Assay Using Tumor Cells for Cytostatic Drug Screening
3.3.1. Preparation of lkmor Growth Factor (TGF)
TGF is used for activating the tumor stem cells in cloning assays with fresh human tumor stem cells. 1. About 20 Balb/c mice are injected intraperitoneally with 0.1 mL/ mouse of mineral oil. After 5 wk the mice are killed by cervical dislocation. The spleens are removed aseptically, minced with scissors and forceps, and sieved through a metal filter (100 pm mesh). The resulting cell suspension is adjusted to 2 x lo6 cells/mL in RPM1 1640 medium. 2. 5 mL aliquots of this suspension are pipeted into 60 mL Petri dishes and incubated at 37OC and 5% of CO, for 2 h. The supernatant is then discarded. 3. Adherent cells are carefully washed twice with fresh RPM1 1640 medium by gently moving the Petri dishes and sucking off the medium with a Pasteur pipet. Fresh RPM1 1640 medium supplemented with 15% of FCS is added to the dishes. 4. The cultures are now incubated for 3 d under the same conditions mentioned in step 2. After this the supernatant is collected and filtersterilized. 5. The activity of this conditioned medium is measured by performing a dose-response curve on a standard tumor cell line (e.g., mouse lymphocytic leukemia L 1210). From the curve, its optimal concentration for cloning fresh tumor cells is determined (e.g., 50 u.L/30Oj~L of cloning mixture).

3.3.2. Established Tumor Lines: Propagation of Suspension Cultures

Stock suspension cultures are kept in 25 cm2 tissue culture flasks. 1. A 1-mL sample is removed from the culture to be passaged, the cells suspended evenly, and counted with Trypan blue for assessment of viability. 2. The cell density is then adjusted in such a way that the final concentration in the new culture will be a value dependent on the cell line concerned, e.g., 1 x lo* viable cells/ml for L 1210. 1 mL of HS or FCS is added, and the volume made up to 10 mL with medium (e.g., RPM1 1640).

Clonogenic Assays 3. After incubation repeated. 3.3.3.

391 at 37OC and 5% CO, for 3-4 d the procedure is Tumor Lines: Cloning Assay

Established

1. The following components are mixed in sterile Eppendorf reaction tubes: a. RPM1 1640 medium, 141 pL. b. HS or FCS, 45 pL. c. Cell suspension, 60 PL. d. Agar-medium mixture, 54 PL. Final vol, 300 pL. The medium may be replaced by a solution of test agent. 2. The cell density is adjusted depending on the cell line used. For instance, L 1210 cells are seeded at a final concentration of 2.4 x 103 cells/ml; thus the suspension used for the assay is adjusted to 1.2x 104 viable cells/mL to allow for dilution. For this purpose, 1 mL is removed from a liquid stock culture, a small sample of this is mixed with Trypan blue solution, and viable cells counted. The viability of the established cell line and the primary human tumors should not be below 20%. 3. The agar-medium mixture consists of 1 mL of 3% agar and 2 mL of RPM1 1640 medium and is prepared as described in step 7 of the granulocyte-macrophage assay. 4. Capillaries are filled with 30 PL of assay mixture as described in step 8 of the granulocyte-macrophage assay. 5. Incubation and evaluation are done as described in steps 10-13 of the granulocyte-macrophage assay. Normally colony yields should be 30-60 colonies/capillary. Colonies are aggregates of more than 50 cells. 3.3.4. Cloning of Primary Human Tumor Cells 1. Solid tumors should be placed into sterile screwcap tubes containing RPM1 1640 medium at room temperature, and if possible, sent to the laboratory immediately after removal. Before processing the tumor is thoroughly washed with RPM1 1640 medium. Macroscopically visible necrotic tissue, remainders of tumor capsules, and connecting tissue are removed with sterile scissors and forceps. 2. The solid tumor is then placed in 5-10 ml of RPM1 1640 medium in a 60 mm Petri dish and cut with sterile scalpels and scissors, until a seemingly homogenous cell suspension arises. Do not stop cutting the

392

Maurer and Echarti pieces too early! The cells are sucked up several times through first 21gage and then 23-gage needles, and finally filtered through a metal filter (50 pm pore size) or through several layers of gauze. The mechanically pretreated tumor sample is now transferred to a plastic tissue culture flask (25 cm2) with sterile forceps and Pasteur pipets. An 8-mL portion of sterile collagenase-DNase solution or 5-10 mL of sterile dispase solution is added, and the mixture is stirred at room temperature for 90 min. Then the suspension is homogenized with 21gage and 23-gage needles and filtered as described above. The cells are washed twice with RPM1 1640 medium and resuspended in 1 mL of enriched CMRL 1066 medium. Cell concentration and viability are determined as described for established tumor cell lines. For the cloning mixture the following components are mixed in sterile Eppendorf reaction tubes: a. Enriched CMRL 1066 medium, 136 PL. b. TGF, 50 pL. c. Cell suspension (106-107viable cells/mL), 60 PL. d. Agar-medium mixture, 54 pL. Final vol, 300 pL. The medium may be replaced by a solution of test agent. Further steps are as described for established tumor cell lines (steps 3-5), except that the capillaries are filled with 50 PL of cloning mixture and the incubation is for up to 3 wk.

3. 4.

5. 6.

7.

4. Notes
1. In all instances, it is recommended to compare unknown test agents with a substance of known effect on the same test cells (internal standard, positive control). 2. In serum-free granulocyte-macrophage assays the capillaries should be left on the cooling plate for lo-15 min before mounting them in the holders. They should then be incubated in an exactly horizontal position. Otherwise it is possible that the agar gel will slip out of one end of the capillary during the incubation. 3. In the T-lymphocyte assay, the prestimulated cells are very delicate to handle. Although they are sometimes difficult to resuspend they must not be vortexed. Pasteur pipets should be used for this purpose. Vibration of the incubator, especiallv during the first few days, will

Clonogenic Assays

393

lead to irregular colony growth. Therefore the capillaries must not be moved during this time. Concerning the reproducibility of the assay and the use of cryoconserved lymphocytes, seeref. 22. 4. The median and mean absolute deviation are determined as follows, First, the three colony counts from each assay point are put into sequential order. The value in the middle is the median. If there are an even number of colony counts (e.g., six values from a control run in duplicate), the median is defined as the mean from the two middle values. For the mean absolute deviation the absolute differences between the colony counts and the median are determined and the mean calculated.

References
1. Metcalf, D. (1984) TheHemopoietic Colony Stimulating Factors (Elsevier, Amsterdam). 2. Salmon, S. E. and Trent, J. M. (eds.) (1984) Human Tumor Cloning (Grime &Stratton, Orlando, Florida). 3. Courtenay,V. D. and Mills, J. (1978) An in vitro colony assay for human tumors grown in immune-suppressed mice and treated in viva with cytotoxic agents. Br. 1, Cancer 37,261-268. 4. Maurer, H. R. (1979) In vitro colony growth of granulocytes, macrophages, T- and B-lymphocytes in agar capillaries. Acfu Haemafol. 62,322-325. 5. Maurer, H. R. and Henry, R. (1976) Automated scanning of bone marrow cell colonies growing in agar-containing glass capillaries. Exp. Cell Res. 103,271-277. 6. Echarti, C., and Maurer, H. R. (1989) Manuscript in preparation. 7. Maurer, H. R. and Henry, R. (1976) Colony growth of mouse bone marrow cells in agar contained in glass capillaries. Bluf 33,11-22. 8. Maurer, H. R., Maschler, R., Dietrich, R., and Goebel,B. (1977)In vitro culture of lymphocyte colonies in agar capillary tubes after PHA-stimulation. J. Immunol. 18, 353-364. 9. Maurer, H. R. and Ali-Osman, F. (1981)Tumor stem cell cloning in agar-containing capillaries. Naturwissenschaften 68,381-383. 10. Maurer, H. R. and Henry, R. (1978) Drug evaluation on haemopoietic cells in vitro. Arzneimittel-Forsch. (Drug Red 28,601-605. 11. Maurer H. R., Henry, R., and Maschler, R. (1978) Chalone inhibition of granulocyte colony growth in agar: Kinetic quantitation by capillary tube scanning. CeZ2 Tiss. Kinet. 11,129-138. 12. Von Hoff, D. D., Forseth, B. J., Huong, M., Buchok, J. B., and Lathan, 8. (1986) Improved plating efficiencies for human tumors cloned in capillary tubes versus Petri dishes. CancerRes.46,40124017. 13. Neumeier, R., Maurer, H. R.,Maschler,R., and Bombik, B. M. (1981)A micromethod to culture granulocyte colonies from human bone marrow in agar-containing glass capillaries. Exp. Hemufol. 9,44-51. 14. Maschler, R. and Maurer, H. R. (1978) Screening for specific calf thymus inhibitors (chalones) by different assays.Hope-Seylers Z. Physiol. Chem. 359,825-834.

394

Maurer and Echarti

15. Tortarolo, M., Braun, R., Hiibner, G. E., and Maurer, H. R. (1982) In vitro effects of epoxide bearing valepotriates on mouse early hematopoietic progenitor cells and human T-lymphocytes. Arch. Toxicol. 51,3742. 26. Matthiessen, I?.,Echarti, C., Gerber, J., and Maurer, H. R. (1978) T-lymphocyte colony formation of murine thymocytes in agar contained in glass capillaries. Nut 55, 517-522. 17. Ulmer, A. and Maurer, H. R. (1978) The formation of B-lymphocyte colonies in agarcontaining glass capillaries. Immuno&y 34,919-925. 18. Ali-Osman, F., Bier, J., and Maurer, H. R. (1983) In tifvo cytostatic drug sensitivity 1 testing in the human tumor stemcell assay:amodified method for the determination of the-sensitivity index. Tumor Diagn. Ther. 4,1-6. 19. Ali-Osman, F., Giblin, J., Dougherty, D., and Rosenblum, M. L. (1987) Application of in vivo and in vitro pharmacokinetics for physiologically relevant drug exposure in a human tumor clonogenic cell assay.CancerRes.47,3718-3724. 20. Sheridan, J.W. and Metcalf, D. (1973) A low molecular weight factor in lung-conditioned medium stimulating granulocyte and monocyte formation in vitro. J. CelI. Physiol.Bl, ll-24. 21. Leibovits, A., Liu, R., Hayes, C., and Salmon, S.E. (1983) A hype-osmotic medium to disaggregate tumor cell clumps into viable and clonogenic single cells for the human tumor stem cell clonogenic assay.Infer&. J. Cell Cloning 1,478-485. 22. Maschler, R. and Maurer, H. R. (1981) Lymphocyte chalone from calf thymus: problems of large scale extraction and assay.Z. Nafurforsch. 36 c, 568-571.

Chapter 32 Cytogenetic for Human

G. John

Techniques Leukemias
Swansbury

1. Introduction
It is appropriate that a chapter on cytogenetic techniques for studying leukemic tissue appears in a series on molecular methods since these subjects are closely related, both in history and in application. It is through molecular methods that the significance of specific recurring chromosome abnormalities begins to be understood, but it is these specific recurring chromosome abnormalities that suggest the areas of interest for molecular studies. For example: when the breakpoint cluster region (bcr) for the Philadelphia chromosome translocation, t(9; 22) (q34; qll), was identified and DNA probes were developed (I), it was supposed that the labor-intensive cytogenetic screening for the presence of the Philadelphia translocation (and similarly for other translocations) might become redundant. Events have shown, however, that the correspondence between abnormal bcr and detected Philadelphia translocation is not 100% (2), so for the time being, at least, cytogenetic and molecular studies must continue to support, confirm and stimulate each other. It is said that if there are many methods towards one end, then not only is none much better than the others, but none is very good at all.
395

396

Swansbury

This currently happens to apply to cytogenetic studies of leukemic tissues: there are probably as many methods as cytogeneticists and none gives the consistently good quality that can be obtained with, for example, PHA*stimulated lymphocytes. The observation that consistently applied technique can give vastly different results in consecutive cases and that populations of cells with different morphology can occur on the same slide indicates that there is an intrinsic, or disease-, or treatment-derived effect on marrow kinetics and chromosome morphology. Most effects obtained by varying techniques are slight and are swamped by these variations. A survey of laboratory methods undertaken by the Association of Clinical Cytogeneticists (3) revealed that almost every aspect of the processing of marrows for cytogenetic studies was subject to great variation-and it can be assumed that all the variations worked for at least some cases. One of the most significant factors in improving results is to set up multiple cultures using different media, culture conditions, and harvesting procedures-a dozen cultures per sample is not unusual in some laboratories (Hagemeijer, A., personal communication). This, however, requires that an adequate sample is provided and that there is time available to cope with this large amount of work-beyond the scope of most cytogeneticists in routine or service laboratories. It used to be considered that direct preparation of marrow samples gave the most reliable picture of the in vivo state of the marrow. However, it has been shown that using various culture times is highly desirable since different cell lineages predominate at different times (4,5). Erythroid cells tend to predominate in the direct marrow preparation and granulocyte cells after culture. Except in erythroleukemia, where the converse is true, increased proportions of chromosomally abnormal cells have been directly associated with loss of erythroblasts. Berger et al. (4) also observed that normal, erythroblast mitoses had good chromosome definition whereas myeloblast/monoblast mitoses had fuzzy chromosomes. Insome ANLLs, * Abbreviations frequently used are: ALL Acute lymphoblastic leukemia; ANLL Acute nonlymphocytic leukemia; CLL Chronic lymphocytic leukemia; CML Chronic myeloid leukemia; FluorodeoxyUridine; FdUr PHA Phytohemagglutinin; TPA 12-O-tetradecanoylphorbol-13-acetate.

Cytogenetic Techniques for Human Leukemias

397

and especially those with a prior myelodysplastic phase, all lineages are affected; erythroblasts have been shown to be involved by using radioactive Fe uptake and heterozygosity for the enzyme GGPD (6). The importance of considering the cytogenetic effect of culturing was clearly demonstrated in the study of the t(15; 17) translocation in acute promyelocytic leukemia (M3). It was thought to have a geographic variation in incidence (7) but this is now thought to arise from varying culture techniques. It is possible that it occurs in 100% of M3 if enough material is studied (4,8). The translocation is found much more often in overnight cultures than in direct preparations. In addition it can easily be missed because mitoses with it often have poorer chromosome morphology than accompanying normal cells and the translocation produces only small changes that can easily go undetected in the morphology of chromosomes 15 and 17. Divisions from ALL often have particularly poor morphology, especially when from a clone with high hyperdiploidy. A very detailed methodology has been published by Williams et al. (9) who have a relatively high successrate. The techniques for cytogenetic studies in leukemia are simple and robust; the analysis of the chromosomes obtained comes with a little practice; the interpretation of the results, however, requires experience and so this aspect is given a substantial proportion of the notes section.

2. Materials
Most of the solutions should be kept in the dark at 4OC. Unless stated otherwise, add 0.1 mL to a 10 mL culture. 1. Containers: Sterile, capped, plastic, centrifuge tubes, universals, or sterilized glass bottles that have been siliconized and have silicone cap liners. 2. Pipets: Plastic or siliconized [with dimethyldichlorosilanet, e.g., Repelcote] to prevent adhesion of fixed cells. 3. Medium: RPM1 1640, McCoys 5A, Hams FlO or F12, TC199, Iscoves, and so forth, are all usable. This laboratory now routinely uses Hepesbuffered RPM1 1640, which was developed specifically for leukemic cells, except where described below. To each 100 mL bottle, add 1 mL +Indicates known or potential carcinogens/poisons: handle with care!

Swansbury of preservative-free heparin and antibiotics (e.g., 1 mL penicillin + streptomycin). Medium used for longer cultures should have Lglutamineadded (final concentration 0.15 mg/mL); this is an essential amino acid that is unstable and has a short life at room temperature. Iscoves low folate medium can be used without serum. Serum: Fetal calf serum is preferred, but human serum of blood group AB can be used. The proportion routinely added is 1 part serum to 4 parts medium but 5 or 10% may be adequate. For longer-term cultures, each batch of serum should be tested to ensure that it supports cell growth. Mitogens: Phytohemagglutinin (PHA) stimulates T-lymphocytes to divide, acting via monocytes that produce interleukin II. For B-cell disorders, other mitogens are used, usually pokeweed mitogen, which indirectly stimulates both T and B-lymphocytes, and TPA, which is specific for B cells. PI-IA and pokeweed mitogen for cytogenetic use are obtained freeze-dried or lyophilized, ready for reconstitution to the appropriate concentration. 12-0-tetradecanoylphorbol-13-acetate (TPA)? (also Phorbol 12myristate 13-acetate): 10 yg/mL in 0.5% ethanol, stored in aliquots at -2OOC. Prepare by dissolving in 10% ethanol and then further diluting 1:19 with water. FluorodeoxyUridine (FdUr)+: Stock solution: 1 part FdUr (25 pg/mL) to 3 partsuridine (1 mg/mL), giving final concentrations of 0.1 and4.0 mM. Thymidine: 0.05 g in 100 mL distilled water; filter sterilize (0.22 I.IM millipore filter); store in aliquots at -20C (do not refreeze); thawed solution keeps at 4OCfor at least 1 mo. Arresting agent: Colcemid or colchicine: Colcemid+: stock solution 1 pg/mL. Also called demecolchicine, from deacetylmethylcolchicine. Colcemid at greater than 0.1 pg/mL can be toxic (20) but colcemid is said to be less toxic than colchicine, so it may be more suitable if a long exposure (e.g., overnight) is used. Colchicine+: Stock solution: 20 pg/mL in saline (0.9% NaCl). Ethidium bromide+: 10 pg/mL in saline (0.9% NaCl). Hypotonic solution: 0.075M KC1 (5.59 g/L). For delicate membranes, and if Iscoves medium has been used, a mixture of three parts KC1 to 1 part 1% aqueous citric acid may be preferable. Note that the effectiveness doesnt derive from just the osmolarity: the K+ ions have a physiological action, so no advantage is obtained by diluting further.

4.

5.

6. 7. 8.

9. 10.

Cytogenetic Techniques for Human Leukemias

399

11. Fixative: Absolute methanol and glacial acetic acid (3:l). This should be freshly prepared just before use, although it may be kept for a little while if chilled. 12. 2 x SSC: NaCl(17.53g) + sodium citrate (8.82g) made up to 1 L of aqueous solution. 13. Trypsin: (2.5%) stored frozen in 1 mL aliquots. Diluted 1:50 in PBS (Ca*+and MgZ+ free) when required. 14. Phosphate buffered saline (PBS; seeAppendix): pH 6.8 for stains, pH 7.3 for trypsin, and pH 5.2 and 6.5 for fluorescence. If necessary, use N%HI?O, (0.07M) and K2HI?04 (0.07M) in the ratio 1:49 for pH 5.2 and 8:17 for pH 6.5. 15. Slides: Frosted-end variety are preferable for convenience of labeling. The slides must be free of dust and grease. Wash in detergent, rinse well in water, then in dilute HCl and alcohol. Store dry, protected from dust, or in acidified alcohol. 16. Coverslips: 1.5 grade thickness or less for G-banding, grade 0 for fluorescence work; 22 x 50 mm size. 17. Stains: Giemsa, Wrights, and Leishmans stains (all commercially available) are all suitable. Leishmans stain is said to be better for the fuzzy chromosomes that occur in leukemias. Fluorescence stains: Atebrin: (quinacrine dihydrochloride)+ 50 pg/ mL in 2% KCl. A&dine Orange? 0.01% in phosphate buffer at pH 6.5. 18. McIlvanes buffer (pH 5.6): Atebrin mounting solution: 0.466 g citric acid, 0.825 g N%HPO,, 50 mL glycerol; make up to 100 mL of solution with distilled water. 19. Mounting medium: Gurrs neutral mounting medium is routinely used in this laboratory; other suitable mountants are XAM, DPX, Histamount, and so on.

3. Methods
As already stated, every cytogeneticist has his/her own variations on basically similar processes. This is a summary of the general technique; variations may need to be used to suit particular conditions. In summary, chromosomes are prepared from dividing cells (mitoses) since at metaphase, just before division, they shorten and become recognizable, discrete units. The cells may be already dividing in the tissue supplied or, in certain circumstances, may be stimulated into division. They are arrested and accumulated in metaphase or prometaphase by

Swansbury destroying the spindle, e.g., with colcemid. The cells are treated with a hypotonic solution to encourage spreading of the chromosomes. They are then fixed, after which they can be stored indefinitely. Fixed cells are spread on slides and air-dried. They can be stained immediately, but are usually first treated to induce banding patterns on the chromosomes to assist in their identification.

3.1. Collection

of Samples

Bone marrow is the preferred tissue for investigation of most hemopoietic disorders. Valuable results can sometimes be obtained from blood, spleen, and so forth, and these are better than marrow for CLL and myelofibrosis. Details relevant to culturing blood are given in Note 1; for the methods described here it will be assumed that a sample of bone marrow is being studied. It is very important that the marrow aspirate for cytogenetic study is taken before any cytotoxic therapy is given: Except for CML, clones in leukemias almost always disappear during remission, often reappearing at relapse. Cytogenetic studies of bone marrow are expensive because they are so labor-intensive, and a lot of time can be wasted on inadequate samples. Ideally, a generous portion of the first spongy part of the biopsy should be sent: later samples tend to be heavily contaminated with blood. Resiting the needle, through the same puncture if necessary, gives better results than trying to get more material from the same site. If there is plenty of material, consider storing some in liquid nitrogen (details in Note 5). Heparinized bone marrow samples can be transported with or without medium. However, use of medium may reduce the likelihood of the sample clotting or drying out. The samples should be sent to the laboratory as quickly as possible without exposure to extremes of temperature. However, a result can usually be obtained even from samples a few days old (except for ALL, which seems to need more immediate attention), blood being more robust than marrow. The samples should not be refrigerated: the cells may take several hours to start dividing again after exposure to cold conditions.

3.2. Handling

Samples

The risk from aerosol formation from marrow/blood is low but the risk following accidental introduction of infected serum (e.g., through a cut on the skin or by penetration with a needle or glass pipet) is high. All samples should be handled carefully, as if they might have hepatitis B or

Cytogenetic Techniques for Human Leukemias

401

HTLV/HIV (AIDS). Gloves should be worn and quills should be used (rather than pipets or needles) while processing unfixed tissue. A laminar flow cabinet should be used if available, otherwise a clear, draft-free bench not used for other purposes. Careful sterile technique should be developed. Since most cultures are short-term, contamination is not usually a problem.

3.3. Setting

Up the Cultures

I. If the marrow aspirate arrives in medium known to be fresh and the patient is not on chemotherapy, then serum and more medium can be added directly. Otherwise, spin down the sample and resuspend it in warmed medium and serum. 2. At least two different cultures should be set up, the two most easy being a direct and a 24-h culture, incorporating ethidium bromide as described below. (For ALLs, direct and 2-4-d cultures may be better.) The cultures most likely to succeed, however, are exposure to colcemid overnight and FdUr synchronization. Where possible, it is well worthwhile using all four types of culture. In addition, using various culture times increases the likelihood of detecting abnormal clones, as well as of getting any result at all, since divisions from different lineages tend to occur at different times. The longer the cells are exposed to colcemid, the more divisions will be collected. However, the chromosomes continue to contract and the metaphases may eventually become unusable. 3. Use 5 or 10 mL cultures with 5-20% serum and a cell density of up to 106/mL. Stand cultures at an angle, rather than upright, since this increases the surface area of the deposit and reduces local exhaustion of the medium. 4. For the direct preparation, incubate at 37OC for an hour or so, if possible, before adding colcemid. 5. Cultures can be gassed with 4% carbon dioxide in air to help maintain the pH of the culture medium if it is bicarbonate buffered. An increased partial pressure (pp) of CO, is not as important for cell growth as a decreased pp of oxygen -2.5% may be optimal for longer cultures.

3.4. Harvesting
Use capped 10 mL plastic centrifuge tubes and do all centrifugation steps at 1000 rpm for 10 min. Avoid fierce treatment of living cells with hypodermic needles or fine-bore pipets as far as possible: use a syringe fitted with a plastic filling tube (e.g., a Kwill).

402

Swansbury

1. Add 0.1 mL of colcemid (the syringe need not be sterile for this) and incubate at 37OC for 10 min to 2 h (or overnight). A very useful variation is to add 0.1 mL ethidium bromide (11) with the colcemid and leave the culture for 2-4 h. This retards chromosome condensation while allowing divisions to accumulate. 2. Spin down the cells and add hypotonic solution (at not less than room temperature or above 37OC),mixing thoroughly by tapping the baseof the tube after the first few drops have been added. Make up to 10 mL. 3. After lo-15 minutes, spin again, remove the supernatant, and tap the base of the tube to loosen the pellet. Add a drop or two of hypotonic solution if necessary to ensure dispersal of the cells, since they are useless if they are fixed in a lump. Add a few drops of fresh fixative while tapping the tube. Make up to 10 mL and invert the tube to ensure thorough mixing. If the volume of red blood cells was large and inadequately dispersed in the hypotonic solution, they may fix into a gelatinous mass, trapping the white cells. If this happens, decant off (and use) the supernatant and then try adding 1:l hypotonic solution: fix to the clump, which may then be dispersed. 4. Leave in this first fix for a little while before spinning again and adding fresh fixative. After this, the sample can be stored at 4OC with further changes of fix when convenient. The fix should be changed at least three times (until the solution is completely colorless) before spreading. Dehydration of the cells improves if left in fix for at least overnight and up to a week if time permits. For longer periods, store at -2OOC. If there was a lot of fat in the specimen, it should be removed by washing once with Carnoys fixative (ethanol 60%, chloroform 30%, acetic acid 10%).

3.5. Spreading
1. A supply of clean slides should be kept either dry in a dustproof container or in a jar of industrial alcohol acidified with a little HCl. Dry slides are usually left in a freezer for a while prior to use, as the film of frost helps spreading. If glass pipets are used, they must be siliconized since the cells are particularly sticky at this stage. Change the fixative shortly before spreading, then spin again and add any of the following to obtain a slightly cloudy suspension: a. 6:l methanol:glacial acetic acid. The slide is then held at an angle and briefly passed through a gentle Bunsen or spirit

Cytogenetic Techniques for Human Leukemias

403

lamp flame. It is usually possible to see the fixative vapor burn. This method usually gives good spreading, but possibly at the expense of some banding quality and definition of outline. b. Fix: The suspension is dropped onto dry or cold-frosted slides. Spreading may be improved by huffing on each slide immediately or waving in the hot moist air above a flame or by adding a drop of fresh fix on top of the spread. c. 60% aqueous acetic acid. This method usually gives fair spreading of chromosomes, but also has the disadvantage of sometimes affecting the quality of banding. Prolongation of the time in 60% acetic acid before spreading may improve spreading, but possibly at the expense of further deterioration in banding quality. The slides can be dried in the hot air above a flame. 2. A couple of drops onto each slide is adequate. Any loss of band definition caused by heating is of more consequence when fluorescence stains are to be used. Spreading may be helped if the suspension is dropped from a foot or more above the slide, but this is risky and should not be done if the sample is small and precious. Place the slides on a hotplate or in an oven (not greater than 60C) for a few hours to dry* 3. If possible, check the slides immediately under phase-contrast microscopy for cell density and chromosome spreading. If these are not optimal, try varying the spreading procedure. An added factor that affects spreading is the weather-particularly the humidity. 4. Some of the slides may be stained immediately, but banding is not very effective until the slides have aged for a few days. Some aging effect can be obtained more quickly by incubating the slides at 60C overnight in an oven.

3.6. Banding
There are many methods of producing banding on chromosomes. Band patterns are broadly grouped into G (Giemsa) bands (seeFig. 1) and R (reverse) bands which are largely complementary (12). Some methods specific for identifying certain chromosome regions (seeFig. 2) are not described here (see,for example, 13). Sequential banding is possible in the order fluorescence -+ GTG + C-bands, but not in the reverse.

404

Swansbury

Fig. 1. GTG-banded karyotype, showing duplication (arrow) of part of the long arms of a No. 1 chromosome.

Cell synchronization techniques are described in Notes 2 and 3. This laboratory routinely uses GTG-banding produced by incubation in 2 x SSC followed by treatment with trypsin (modified from ref. 24). By this method the trypsin serves to enhance the bands produced by the 2 x SSC, and may be omitted. Trypsin may be used alone, but without the incubation the banding tends to be unpredictable and it is difficult to control the amount of digestion needed. Chromosomes banded with SSC may have a slightly hairy appearance; trypsin alone tends to produce rounded-off bands and thus an apparently clearer outline.

3.6.1. SSC Banding

1. Heat up a Coplin jar containing 2 x SSC to 60C in a water bath. Incubate the slides for 60-90 min for bone marrow, an hour for blood. 2. Cool the slides in SSC and rinse in water. (They may be dehydrated through alcohols and left to dry-overnight if wished-before continuing with trypsin treatment.)

Cytogenetic Techniques for Human Leukemias

405

Fig. 2. C-banded metaphase of the same case.

3. Add 1 mL of trypsin to 50 mL of PBS (Ca*+- and MgWree): Hanks buffered salt solution may also be used. The temperature should not be over 10C. 4. Transfer the slides to a jar containing buffer. Each slide is then immersed in the trypsin solution for a few seconds, washed in alcohol or in a copious volume of buffer to arrest the enzyme action, and placed in a rack to dry. The time needed in trypsin is quite variable and needs to be determined for each laboratory. It is affected by variations in spreading technique, age of the slides, degree of contraction of the chromosomes, general chromosome morphology, and so forth. If the trypsin appears to act very quickly (within a few seconds), some of the conditions can be altered to slow it down-e.g., use a lower temperature or a suboptimal buffer (pH 6.8).

3.6.2. Dypsin Banding

Slides should be incubated for at least an hour in an oven at 60C immediately before banding. The duration of exposure to trypsin needs

406

Swansbury

careful attention, being about 5-10 s for BM. A variation is to dip the slide for 30 s into 25 vol hydrogen peroxide (1:3 in tap water) then rinse in tap water before proceeding to use trypsin for 8 s for both blood and marrow. It may be helpful to watch a metaphase under x40 phase contrast to see when the chromatids start to swell.

3.7. Staining and Mounting of GTG Banded Preparations

1. Dip the slides in PBS first, to reduce the risk of picking up oxidized stain from the surface. Use freshly made stain, diluted in pH 6.8 buffer: Giemsa: 1 part stain to 25 parts buffer for 10-15 min. or Leishman: 1 part stain to 4 parts buffer for 3-4 min. or Wright: 1 part stain to 4 parts buffer for 3-5 min. 2. Rinse the slides in buffer, blot gently, and stand to dry. 3. Mount in a little Gurrs neutral mounting medium (in our experience, this does not lead to leaching of stain, but XAM, DPX, Histamount, and so forth, are probably also suitable) using grade 1.5 (or less) thickness coverslips. If possible, leave to dry overnight before examining the slide with an oil immersion objective, since oil and soft mountant tend to mix. Do not leave the slides exposed to strong sunlight. Do not thin the mounting medium with xylene unless necessary, since it may result in loss of stain.

3.8. Fluorescence

Banding

Fluorescence is not recommended for routine cytogenetic work with leukemic marrow. The interpretation requires considerably more experience and, because the abnormalities found are often complex, the rapid fading during exposure to ultraviolet light is a major drawback to the prolonged analysis of complex abnormalities down the microscope, and photography gives a poor record because of the low contrast. However, there are instances when fluorescence is helpful inidentifying particular parts of chromosomes involved in translocations. For example, Atebrin causes heterochromatin, satellites, and Y-bodies to fluoresce particularly brightly, whereas DA-DAPI (20) (seeFig. 3) shows only the heterochromatic regions of chromosomes 1,9, and 16 and the short arms of chromosome 15. Fluorescent R-banding is particularly useful for checking on abnormalities near the ends of the chromosome arms.

Cytogenetic Techniques for Human I;eukemias

407

Fig. 3. DAPI-banded metaphase of a PI-IA-stimulated blood cell.

3.8.1. Fluorescent

G-Banding

The simplest method of fluorescence G-banding, adapted from ref. (22), is: 1. Stain the slide in Atebrin for 5-20 min. (The stain may be kept and reused.) 2. Rinse in 2% KCl. 3. Add a drop of mounting solution (M&wes buffer), place a grade 0 coverslip, press, blot, and seal the edges with rubber solution or nail varnish. 4. The slides can be examined immediately, using ultraviolet light, or first refrigerated overnight at -2OOC to improve the result. 5. Restaining is simple: peel off the sealant, float off the coverslip, and restain as before.

408

Swansbury 3.8.2. Fluorescent R-Banding

For R bands with good contrast, use the following procedure, adapted from ref. (22): 1. Incubate in phosphate buffer pH 5.2 at 87C for 20 min (longer for fresh slides, less for old), 2. Stain in 0.01% A&dine Orange in phosphate buffer, pH 6.5, for 5 min. 3. Rinse in pH 6.5 buffer for 2 min. 4. Mount in buffer and seal the coverslip with rubber solution or nail varnish. 5. An alternative method of producing R bands is to substitute bromodeoxyuridine for thymidine when rescuing after treatment with Methotrexate or FdUr.

3.8.3.

Chromosome Analysis

Use a 10x objective for screening and an oil-immersion 100x objective for studying the metaphases. Analysis directly with the microscope is possible with experience, but at least one cell from each case should be photographed as a record. If the workload permits, at least one cell should be karyotyped, the photographed chromosomes being cut out and arranged in order (12), especially in apparently normal cases. When screening, remember to bias toward cells with poor morphology where there is a mixed population. Selection of only good mitoses can lead to failure to detect the presence of an abnormal clone. If the quality of metaphases is poor, then full analysis may not always be possible. However, even if the chromosomes can only be counted and/or grouped, useful information can sometimes be obtained. The analysis of polyploid mitoses may appear daunting, but in some cases their morphology is better and so abnormalities may be more apparent. In some cases, especially after exposure to radiation, a high frequency of chromosome abnormalities is found. Note that a clone is assumed to be present if at least two cells have the same abnormal or increased chromosome or at least three cells with the same gained chromosome, assuming that random loss is not at a high frequency (seeNote 4).

4. Notes
1. Blood cultures are used for the following purposes: a. T-Lymphocytes which are cultured with PHA (0.1 mL/lO mL culture) in the presence of monocytes for at least 48 h are

Cytogenetic Techniques for Human Leukemias

409

transformed and start dividing. These cells can usually be used to establish the patients constitutional karyotype, against which any acquired abnormalities found can be compared. Other tissues, e.g., skin, may be used instead, but blood lymphocytes are the most convenient. If the patient is on chemotherapy or the T-cells are affected by the disease, then PHAstimulated cultures may not be successful. A Philadelphiachromosome-positive result may sometimes be obtained from PHA-stimulated T-lymphocytes, useful if other cultures fail. b. In the leukemias, cultures of leukocytes unstimulated by any mitogens may be useful if marrow is not available or has failed to produce a result. Mitoses may be found in unstimulated cultures in approximately 10% of cases that have primitive cells in the blood from the bone marrow or extramedullary sites of hemopoiesis. If a chromosomally abnormal clone is found, these cells can be scored as bone marrow cells. It must be borne in mind, however, that some of these cells may have left the marrow weeks or months previously, so that the population of mitoses found may not represent an up-to-date picture of the state of the marrow. Mitoses also occur as a result of in vivo transformation of normal lymphocytes, e.g., as a result of infection, so the origin of cells with normal chromosomes cannot generally be determined. These must be interpreted with care. Other malignancies, e.g., multiple myeloma, may also release dividing cells into the blood. c. Cytogenetic studies of CLL and B cell disorders can be made on 4-5-d blood cultures, in preference to marrow, stimulated with pokeweed mitogen, which transforms both B and T lymphocytes, or TPA, which is specific for B cells. Use of TPA also reduces the toxic effect of colchicine on CLL cells (15). Difficult cases may respond to cocktails of two or three mitogens. 2. The methotrexate (MTX) method of obtaining better-quality chromosomes was introduced for blood lymphocytes and adapted for other tissues (16). Although promising results were initially claimed for hematologic disorders, the method is falling into disuse since it does not appear to be helpful in most cases. MTX was thought to arrest cell division at a stage when thymidine is required, thus bringing dividing cells into synchrony, the block being later released by adding thymidine, The results are now thought to derive from a slowing-

410

Swansbury

down of early S-phase, retarding the contraction of the chromosomes, or through an effect on the chromosome morphology and the cell membrane. The time after rescuing with thymidine is different for different tissues. A release of 5 h was defined for normal lymphocytes. Gallo et al. (27) have shown that the time should be 9.5-11.5 h for myeloid and leukemic cells; Morris and Fitzgerald (18) have shown that the time varies between patients and that the cell cycle time is generally shorter in CML than ANLL. 3. The fluorodeoxyuridine (FdUr) (29) method is gaining popularity over the MTX procedure, particularly for ALL. FdUr is less toxic than MTX, so it is not necessary to wash the cells after exposure and cell loss is reduced. a. Culture the sample for 6-24 h in a low-thymidine medium (i.e., not Hams FlO or F12; seeAppendix, this vol). b. Add FdUr + Uridine stock solution: 0.1 mL/lO mL culture. c. Incubate overnight. d. Add thymidine or bromodeoxyuridine (BrdU: stock solution 200 mg/mL) to release the block or spin down and resuspend in Hams FlO or F12. e. Harvest 7-8 h later, with colcemid for the last half hour. f. Further processing is as already described. 4. Interpretation of results: Most centers find that about 50% of ANLLs have a detectable chromosome abnormality, rising to about 90% if enough cells can be analyzed. It has been suggested that all cases may eventually be shown to have an abnormality, but in practice there will always be some cases in which an abnormality cannot be demonstrated. Detection of a karyotypically abnormal clone is evidence for the presence of a malignancy. Note that it doesnt automatically mean that the patient has a poor prognosis: some abnormalities are actually associated with a better prognosis than a normal karyotype. The finding of only karyotypically normal cells does not, however, mean that there is no malignant clone present. If the patient has acute leukemia and has had any cytotoxic therapy, for example, then it is highly probable that any clone will become undetectable until the disease relapses and so a normal result will be obtained. In contrast, in CML, the Philadelphia chromosome persists after treatment but subsequently acquired abnormalities disappear until the acute phase supervenes.

Cytogenetic Techniques for Human Leukemias

411

Bear in mind the possibility that the clone was present, but not detected: it has been estimated that in an adult about 40 thousand million (billion) new marrow cells are produced every hour, so only a very small proportion is being examined. The more cells examined, the more confident you can be that there isnt a clone present. At least 25 metaphases should be analyzed, whenever possible, unless the presence of a clone can be established with fewer metaphases. It has been calculated that if 29 cells are normal then you can be 95% confident that a clone involving 10% or more of the population is not present (23). Even if a cloneis detected after analyzing just a few cells, more should be analyzed in caseof clonal variation and even multiple clones. In some cases karyotypically normal cells of good morphology are found on the same slide as karyotypically abnormal cells of poor morphology. When two morphologically distinct populations occur without any detectable chromosome abnormality, one may assume that the poorer cells derive from the leukemic clone. However, the converse, that the good-morphology, karyotypically normal cells are not leukemic, does not follow. In fact, in some instances, e.g., ANLL subsequent to a myelodysplasia, it is quite likely that all the divisions, abnormal or apparently normal, derive from the neoplasia. It can be discouraging to work in leukemia cytogenetics when even experienced workers can find that up to one third of samples may fail to give a useful result because of (a) inadequate material, (b) absence of mitoses, or (c) mitoses being completely unanalyzable. The work remains interesting, however, because of the high frequency and variety of chromosome abnormalities and because there is still so much to be learned about their biological and clinical significance. 5. Use of frozen material: The custom, in some hospitals, of freezing samples of blood and marrow in liquid nitrogen has been of great help to those working with DNA. It is also possible to perform cytogenetic studies on frozen material if sufficient care is taken in the freezing and thawing processes. A minimum procedure is given here: refinements are according to the facilities available. a. Freezing: Suspend the cells in: 10% Dimethylsulfoxide (DMSO), 20% serum and 70% RPM1 or other medium. Start freezing straight away: 1 degree/min for 30 min then 3 degrees/min to the end.

412

Swansbury

b. Thawing: Warm up quickly (shake the vial in warm water) and wash with medium + serum to remove the toxic DMSO. Cultures may need an extra day before harvesting. Although some samples fail to recover, this Department has occasionally obtained chromosomes from frozen material that were of better morphology than those from a fresh portion of the same biopsy.

Acknowledgments
Methodology is a constant topic of conversation among cytogeneticists and I am indebted to the many, particularly fellow-workers in this Department, who have contributed to this chapter by sharing experiences, theories and advice. I am also grateful to Professor Sylvia Lawler, former head of the Department of Cytogenetics, and the Royal Marsden Hospital, London, UK, under whose auspices this chapter was written.

References
2. Groffen, J., Stephensen, J. R., Heisterkamp, N., de Klein, A., Bartram, C. R., and Grosveld, G. (1984) Philadelphia chromosome breakpoints are clustered within a limited region-bcr--on chromosome 22. Cell 36,93-99. 2. Bartram, C. R. and Carbonell, F. (1986) bcr rearrangement in Ph-negative CML. Cancer Genet. Cyfogenet. 21,183,184. 3. Harrison, C. J., Fitchett, M., Potter, A. M., and Swansbury, G. J. (1987) A guide to cytogenetic studies in haematological disorders. Eugenics Society OccusiomZ Papers, New Series No. 1. 4. Berger, R., Bernheim, A., Daniel, M. T., Valensi, F., and Flandrin, G. (1983) Cytological types of mitoses and chromosome abnormalities in acute leukemia. I..e~kx~& Res. 7,221-235. 5. Keinanen, M., Knuutila, S., Bloomfield, C. D., Elonen, E., and de la Chapelle, A. (1986) The proportion of mitoses in different cell lineages changes during short-term culture of normal human bone marrow. Blood 67,1240-1243. 6. Fialkov, P. J., Singer, J. W., Adamson, J. W., Vaidya, K., Dow, L. W., Ochs, J., and Moohr, J. W. (1981) Acute non-lymphocytic leukemia: heterogeneity of stem-cell origin. Blood 57,1068-1073. 7. Mitelman, F. and Levin, G. (1978) Clustering of aberrations to specific chromosomes in human neoplasms. III. Incidence and geographic distribution of chromosome aberrations in 856 cases. Hereditus 89,207-232. 8. Swansbury, G. J., Feary, S. W., Clink, H. M., and Lawler, S. D. (1985) Cytogenetics of acute promyelocytic leukemia: incidence of t(15; 17) at the Royal Marsden Hospital, London. Leukemia Res. 9,271-278. 9. Williams, D. L., Harris, A., Williams, K. J., Brosius, M. J., and Lemonds, W. (1984) A

Cytogenetic Techniques for Human Leukemias

413

direct bone marrow chromosome technique for acute lymphoblastic leukemia. Cancer Genet. Cytogenet.13,239-257. 10. Wiley, J. E., Sargent, L. M., Inhirn, S. L., and Meisner, L. F. (1984) Comparison of prometaphase chromosome techniques with emphasis on the role of colcemid. In
Vitro 20,937-941. 11. Misawa, S., Horiike, S., Taniwaki, M., Abe, T., and Takino, T. (1986) Prefixation

treatment with ethidium bromide for high resolution banding analysis of chromosomes from cultured human bone marrow cells. Cancer Genet. Cytogenet. 22,
319-329. 22. ISCN (1978) An international system for human cytogenetic nomenclature. Birth Defects: original article series, XIV, No. 8, New York, The National Foundation. Rooney, D. E. and Czepulkowski, B. H. (1986) Human Cytogenetics: a practical 23.

approach. IRL Press,Oxford, UK and Washington, DC.

14. Sumner, A. T., Evans, H. J., and Buckland, K. A. (1971) A new technique for distinguishing between human chromosomes. Nature (Nezu Biology) 232,31,32.

15. Connor, T. W. E. (1985)Phorbol ester-induced lossof colchicine sensitivity in chronic lymphocytic leukemia lymphocytes. Leuk. Res. 9,885-895. 16. Yunis, J.J.(1981)Newchromosome techniquesinthestudyof humanneoplasia. Human Pathol. 12,540-549. 17. Gallo, J.H., Ordonez, J. V., Grown, G. E., and Testa, J. R. (1984) Synchronization of human leukemic cells: relevance for high-resolution banding. Hum. Genet. 66, 220-224. 18 Morris, C. M. and Fitzgerald, l?.H. (1985) An evaluation of high resolution chromosome banding of hematologic cells by methotrexate synchronization and thymidine release. Cancer Genet. Cytogenet. 14,275-284. 19. Webber, L. M. and Garson, 0. M. (19831Fluorodeoxyuridine synchronization of bone marrow cultures. Cancer Genet. Cytogenet. f&123-132. 20. Schweizer, D. (1980) Simultaneous fluorescent staining of R bands and specific heterochromatic regions (DA-DAPI bands) in human chromosomes. Cytogenet. Cell Genet. 27,190-193. 21. Caspersson, T., Gahrton, G., Lindsten, J., and Zech, L. (1970) Identification of the Philadelphia chromosome as a number 22 by quinacrine mustard fluorescence analysis. Exp. Cell Res. 63,238-240. 22. Verma, R. S.and Lubs, H. A. (1975) A simple R banding technic. Am. J. Hum. Genet. 27,11&117. 23. Hook, E. B.(1977) Exclusion of chromosomal mosaicism: tables of 90%,95% and 99% confidence limits and comments on use. Am. 1. Hum. Genet. 29,94-97.

Chapter 33 Time-Lapse Cinemicroscopy

Peter N. RiddZe

1. Introduction
Cinematography commenced as a scientific technique used as a system for slowing down observed movement. Marey in 1888 (1) constructed, following a number of other ideas, a Chambre Chronophotographique, which had practically all the elements of the modern tine camera, With this he made serial photographs (not transparencies) of various biological phenomena (2). The mounting of the tine camera to the microscope was achieved shortly after the turn of the century, and apparatus producing film that would be difficult to surpass today was in use in the early 1920s. No opportunity to see the films made by such workers as Canti at that time should be missed. Such equipment in fact differs from present-day cinemicroscopes only in general sophistication and the (possibly excessive) use of chips where motor-driven switches are quite adequate. (A motordriven control unit lives at the back of my laboratory shelf and is still brought out when the electronic units fail. It is more clumsy to use, but it does work on and on and on with no attention.) Also, more sophisticated optical systems such as phase contrast and DIC have become available 415

416

Riddle

since earlier cinemicroscopes came into use. It is less obvious that some older but equally valuable methods have been largely forgotten in the flurry of modern developments. An example is Rheinberg illumination (3), which is k nown to relatively few users of microscopes (as opposed to microscopists). Similarly, it is rarely remembered that a thin biological subject can often be seen in considerable detail using simple but properly adjusted bright field optics. Most workers turn to the phase contrast (despite its inherent artifacts) without considering possible alternatives. Thus, much of the work done now would have been possible with equipment available half a century or more ago. It is hoped that this chapter may show that the system is far from demanding, and that some dusty but perfectly adequate equipment will be brought back into use from laboratory cupboards. The maxim that seeing is believing is frequently difficult to apply to scientific observation. So often, results are inferred from effects that are secondary to the data actually sought. One example of this is the estimation of cell growth by measuring the incorporation of Tritiated Thymidine in the (not unreasonable) hope that the number of cells entering the S phase is an accurate predictor of ensuing divisions. Such a procedure not only loses the culture, but also adds extra variables to the system that may cause changes to the very parameter being measured (4). In contrast, cinemicrographic recording of cellular activity yields a record that can be seen directly while not destroying the specimen. Data can be reassessed later for different parameters, and last but not least, the visual record often allows unsuspected activity to be seen and investigated if relevant.

1 .I. Basic Considerations

The theory of conventional cinematography is that a series of individual photographs (transparencies) are taken with a brief time spacing (usually l/25 s) apart. When projected at the same time spacing and in register, the image appears to move in the same way and at the same rate as the original. The fact that the record consists of disconnected images is hidden by the persistence of vision of the observer. If the speed of projection is greater than the speed used when filming, the apparent rate of movement of the image may be changed either to appear slightly more rapid as in the Keystone Cops or increased much more, as for cinemicroscopy of cells in tissue culture. Alternatively, if the filming interval is shorter than the projection interval, the movement is slowed and high-speed films can even record the passage of a bullet. When observing cells in tissue culture, it is

Time-Lapse Cinemicroscopy

417

not unusual to have a time-lapse interval of between 1 and 300 s which causes an apparent increase in rate of between 25x and 7500x. Films taken at these intervals show movement and serial divisions. The development of cell clones can be recorded with the longer intervals so that a 30-m length of film can record over 2 wk of activity. Thus, the time-lapse cinemicroscope offers a method to compress time and subsequently extract a vastly increased amount of data. The majority of movements within cells and cultures are so slow that it is difficult to perceive them at all by eye, although the associated biochemical phenomenamay occur much more rapidly. With gross translocational movement, even the fastest moving of mammalian cells, leukocytes, do so at speeds measurable in pm/min (5), whereas fibroblasts seldom exceed 60 l..tm/h (6); many remain practically stationary. These rates are considerably affected by environmental factors such as medium, substrates, temperature, toxins, and so on (7). The division of the majority of cells from start to finish of mitosis takes about 1 h, whereas in contrast, the whole period between cell divisions lasts from 8 h upward, for example (8). Unfortunately, it is rarely possible to obtain worthwhile data about one slow and one fast parameter on the same film. It is essential to recall, when looking for rapid changes in a film that has been taken at a long time interval, that the individual frames are brief records of conditions at widely separated times and that, in consequence, there are many intermediate situations that will not have been recorded. Optical considerations also affect changing appearance and cell behavior as seen at the projection screen. The most important of these is the relation of total magnification of the image to the time interval. For a particular interval, the speed of a point on the screen travels proportionately to the linear magnification. The image of a cell organelle that may travel 10 cm on the screen when photographed with a 10x objective will move 1 m in the same time if a 100x lens is used. Although these considerations appear obvious, they can be overlooked when time intervals and magnifications are being chosen at the start of an experiment. What data are required from the film dictates the parameters when filming, but the basic principles are similar for all time-lapse photography.

2. Methods
Cultures for cinemicroscopy are assumed here to be already growing in a suitable container (see below) and ready for examination.

418

Riddle

2.1. Filming
2.1.1. Parameters
Before any further bench work is attempted, the experimental parameters to be studied should be listed. Without these details, it is not possible to choose suitable microscope and camera settings. Many biologists want to have a film of some phenomenon without previous thought of what they want from the result. Many even seem surprised that films can be of greater analytical than pictorial value. Such unplanned material is seldom of little worth other than of idle enjoyment to those who have not seen such films before. Sound technique demands an accurate appreciation of what must be filmed, whether it is the length of time between cell divisions, the rate of cell movement, or another feature.

2.1.2. Equipment
The microscope and photographic equipment for cinemicroscopy can be purchased, in most cases, from the manufacturer as a unit. Alternatively, it can be assembled piecemeal from components of ones own preference. Experience with both approaches suggests that the former is likely to give more consistent results and the latter more flexibility (and satisfaction). It is essential that the system should be dedicated to this particular use since films can take a long time to produce and cannot be casually interrupted for irrelevant purposes. The entire system, however, requires the following basic elements listed outward from the specimen: 1. 2. 3. 4. 5. 6. specimen chamber environmental control for specimen microscope with appropriate lens systems and illumination comfortable operator seating tine camera film

2.1.2.1. Specimen Chambers. More than ten years ago in the early days of science, articles on time-lapse cinemicroscopy usually contained a description of at least one typeof cell growth chamber designed for specific observations. Generally, these consisted of two coverslips or glass plates separated by a washer or gasket to contain a minute amount of medium with the cells adherent to one surface. Microscopes generally used were of conventional rather than theinverted form and required the specimen cells

Time-Lapse Cinemicroscopy

419

(and the medium) to be on the upper surface and not where gravity normally dictated. This involved growing the cells on the surface to be observed and then inverting it. Medium was contained by surface tension or total filling of the space. Some special chambers were, however, essentially different in that they could provide conditions such as raised pressure, which was not possible in other designs (e.g., 9,10). The more general use of the inverted microscopes that are now offered by the majority of microscope manufacturers has removed the need for growing the cells on the ceiling. Conventional Petri dishes or plastic cultureflasks canbeputonto themicroscopestage. Lossof detail by distortion from less than perfect plastic is seldom a problem, and optical distortion resulting from increased thickness can be compensated for by the use of lenses corrected for thicker substrates. If plastic surfaces are found to be inadequate, there are optically flat glass Petri dishes in which the base is designed to suit the long focus of such special lenses (seeNote I). Use of the same type of container as that in which other cells are grown in parallel experiments removes the risk of differences, resulting from substrates, both in cell behavior and in more subtle effects. In the authors opinion, there is seldom any need now to venture into the realms of special chambers as long as inverted microscopes are used (for those who wish to use special chambers (see10). It should be noted also that the use of conventional plastic dishes is not possible for high-magnification photography with oil immersion objectives. Even so, high magnification is now possible if dishes with bases made of thin stretched plastic membranes are used. These are gas permeable but oil resistant (Petriperm, made by Heraeus), and allow the immersion objectives to come as close to the cells as do conventional coverslips. The image quality is occasionally somewhat marred by faults in the plastic, which is also easily damaged, but such faults, even if an unspoiled area cannot be found, are usually outside the depth of field of the system when the specimen is in focus.

2.1.2.2. Environmental Control 2.1.2.2.1. TEMPERATURE.Temperature requirements for growth are

well known for most cell lines, although in the authors experience, most laboratory incubators seem to be set at the normal human temperature of 37OC, rather than that appropriate to any other species being grown. Usually, provided the temperature is stable the cells do not appear to suffer. In addition, the ability to change the temperature of the culture environment rapidly may be required experimentally with, for example,

420

Riddle

temperature-sensitive mutants. Several methods are available for heating cultures. The once popular heated stage has drawbacks for long-term work, and the crucial part of the culture, that under observation, gets no heating at all. The most useful systems are those that enclose the microscope stage or the stage and as much of the body as is physically necessary. A box of Perspex with doors to allow access to the controls and insertion of the culture is convenient and easy to construct. Some manufacturers supply these incubators with the microscopes, but it is as well to ensure that the design was not created around small hands, making the access to the stage very difficult. With large access ports opened, it has been found that temperature fall is seldom significant, provided that adequate heating backup exists. It is as well to be sure that the available stage equipment will not need to be replaced with something larger, requiring complete rebuilding. Temperature within the box can be maintained by any available heat source, provided that it does not cause local overheating or vibration. Heaters inside the box can be the source of considerable convection and poor distribution unless a fan is included too. Although fans that are nearly vibration free are available for this role, the risk of slight vibration evident at high magnification remains, especially if the subject is suspended in a gel medium and not attached. External blown heat has been found to be an excellent way to warm the incubator box. The domestic hair dryers mounted in a box with a long flexible tube connected to a plastic bag for the head have proven to be both incredibly long lasting and easy and cheap to convert to this use. The tube is led to the incubator box through a hole and an air baffle placed within to disperse the hot-air stream. The fan motor is run continuously, often for years on end (its life being mainly limited by environmental dirt). Motor maintenance, of occasional lubrication, should prevent the motor from seizing up, overheating, and in serious cases, melting through the bottom of the box. The heating element is controlled by a solid-state thermostat (9) and a simple thermistor device has given temperatures stable at the stage to within O.lOC as indicated by a conventional thermometer. This is adequate for the most demanding tissue culture. A hair dryer heating unit when used in this way causes no vibration to the microscope unless it shares the same bench. In practice, mounting the heaters on a separate, wall-mounted shelf beneath the microscope bench, with the pipes coming up through holes, is both convenient and adequate (seeNote 2). 2.1.2.2.2. GAS. The medium in use dictates the gas requirement. Some media containing buffers such as HEPES or TRIS, which are used in

Time-Lapse Cinemicroscopy

421

the absence of any carbon dioxide, might appear at first sight to be the ideal solution to the problem of pH control during cinematography. However, there is little point in introducing what can only be described as a further variable into conditions that are of themselves different to those in which the cells are normally grown. With sealed culture chambers, there may s till be a need for a minute flow of gas past the drop of medium to compensate for any loss of carbon dioxide. The Roberts/Trevan chamber (II) is one that is designed to have the medium in the center of the space between the cover slips with the gas circulating around it. Such gas flow must be very small (of the order of 1 cc/5 min) and has proven very difficult to regulate and measure with any accuracy. Gas supplies cannot be fed directly to the incubator box surrounding the microscope, since problems of leakage, corrosion, and condensation demand that the humid gas be restricted to a smaller, sealed space. It has been found that Petri dishes can be gassed adequately in boxes small enough to fit on the stage. These have optically flat glass or plastic windows in the lid and open viewing ports in the base. The holes are covered by the dish. Two such designs are shown (Fig. la and lb), and variations could be made according to particular circumstances. The square design was thought to be unlikely to be sufficiently gas tight to maintain pH, but in practice it has been found to be simple, easy to clean, and effective. The alternative stainless-steel circular boxes (which are far more elegant) are little if at all more effective. Neither chamber can be purchased commercially but can be made from readily available materials (seeNote 3). 2.1.2.2.3. HUMIDITY. Whether small or large, the gassed volume must be maintained as near 100% relative humidity as possible to avoid concentration of the medium with all the untoward chemical and biological changes. This humidity is almost impossible to achieve, but something very close can be reached by: (1) placing a reservoir of water close to or around the culture, and (2) bubbling the incoming gas supply through a Wolff bottle containing water. It is desirable to heat the water slightly above the temperature of the culture to slightly super-saturate the gas. Extra complications can be avoided by mounting the humidifier bottle within the incubator, but even then a small amount of additional heat is needed to compensate for evaporative cooling. A torch bulb in a finger worked into the side or base of the bottle (1 or 2 W) is usually enough. A wire around the bottle would serve equally well or a painted on heating element as used in ref. 12. Any connecting pipes between the humidifier bottle and the environmental chamber must be kept short, of small internal diameter, and warmed to avoid moisture resulting from condensation in

422

Riddle

4a

4ia

Fig. la. Circular metal environmental chamber. Above, lid seen vertically. Middle, chamber in section. Below, base from above. (1) Optically flat glass insert (cemented in), (2) stainless-steel top, (3) inlet hole for humidified gas, (4a) hole for locating pin, (4b) pin, (5) metal base, (6) water reservoir in moat form, (7) viewing port.

the tube. The tube itself should be of PVC or polythene, rather than silicone rubber since the latter is somewhat permeable to carbon dioxide. A warmed chamber of water, such as the humidifier reservoir, does tend to acquire unwanted flora and fauna in a remarkably short time. Copper sulfate in the reservoir (1% would be adequate) is nonvolatile and

Time-Lapse Cinemicroscopy

423

lb

06 ,--I
2

I 4 I ,-- -1

-\ / 8 / \\ I \0 / -A/ 0 ;::j
7 4 3

-t&l-

Fig. lb. Dismountable square environmental chamber. Above, dismantled. Middle, assembled, in section. Bottom, vertical appearance. (1) Glasslid, (2) perspex body, (3) aluminum alloy base, (4) water reservoir (bottle lid), (5) recessed corner for stage clips, (6) inlet hole for gas supply, (7) viewing port, (8) spring clip (ex. microscope stage) on extended spigot.

424

Riddle

suppresses such growth, but it reacts with the incoming carbon dioxide to form a very unsightly blue/green precipitate. Potassium dichromate has been found to be equally effective, and the same reservoir solute is simply topped up to replace evaporation losses without any need to clean out the bottle. More sophisticated antibacterial agents, e.g., Dichlorophen have proved effective but quite unnecessary. 2.1.2.2.4. LIGHT. Some types of cells are sensitive to light, and laboratory lights should be dim except when setting up the apparatus. An alternative light source from a low-power lamp not only keeps sensitive cultures in the best condition, but also helps viewing when only low levels of microscope light are available for viewing (seeNote4). Environmental light from the laboratory is of low intensity at the culture surface compared to the amount that assaults the cells each time the microscope lamp is turned on. The effects of this concentrated beam of light can sometimes be seen as physical damage (inactivity or even death), but such extremes are more usually the result of local heating in the light path close to the microscope condenser. It is salutory to place ones hand in the same position as a culture to understand this. There are two ways to counteract the heating resulting from the conventional tungsten light beam: (I) insert a heat filter (heat absorbing or dichroic heat reflecting glass) in the light path, and/or (2) have the light source playing on the culture for minimum periods of film exposure and observation only (13). Most modern camera control units have means to switch the illuminating beam on or operate a shutter in the light path. If the latter method is used, the lamp should be allowed to heat up to full brightness before the exposure is made. If this is not done, exposures made when the lamp is on for observation may be significantly more heavily exposed than those when the switching is in operation. Film made then has sharp variations in intensity between frames, which is very distracting during projection. 2.1.2.2.5. VIBRATION. Vibration is an ever-present artifact that is very difficult to eliminate. It affects both cultures and microscopes. Some (most) cultures when settled on the culture dish are little affected by background vibration, but those that settle slowly take much longer to settle and may even be made to collect in the center of the dish if the movement is severe. The effect of vibration more usually considered is that upon the equipment itself. Microscope images are spoiled by vibration especially at high magnification. Some types of culture are relatively resistant to such effects, but the image of cells grown in agar gel can become

Time-Lapse Cinemicroscopy

425

a blur if the natural period of resonance of the system approaches the vibration frequency. Elimination of vibration depends upon its source and the way in which the equipment is mounted. The first precaution when setting up such apparatus is to ensure that the bench is adequate. Observation of a sensitive specimen (e.g., at high magnification) soon settles the question of how much of the image is being spoiled. It should be realized that some forms of vibration are relatively short-lived and may last only a few seconds. Others are persistent, or only occur when the microscope is in use. The persistent vibration that affects a building (traffic shake, elevators, central heating fans, and so on) is often only present in certain places, and a search reveals that more suitable siting of sensitive equipment is the best way to overcome the problem (seeNote 5). When a suitable position is found (a basement often being ideal), the use of a firm bench is advisable. It should be free-standing if possible and heavy, although exceptions have been seen where cinemicroscopes have been mounted on relatively flimsy tables and yet showed no evidence of vibration. Extremely heavy and relatively inexpensive benches can be made from concrete garden furniture, and a pub table weighing some 100 kg, with the seat removed has proved very stable indeed. Vibration caused by the electrical equipment in use (intervalometer cooling fans, heaters, and so on) is best eliminated by placing the offending source on a separate shelf. Finally, there is a vibration caused by the camera shutter mechanism. This is brief, and occurs at the beginning and end of an exposure. Where the camera and microscope are integal, this cannot be eliminated by changing the bench, although where the camera is mounted on a separate pillar (not necessary in the authors opinion), the two units can be separated on two distinct surfaces. (The relative movement of camera and microscope form a separate problem here.) Cures for camera-induced shake are: (1) using a relatively long exposure. Short periods of vibration occur while the camera and microscope are still moving because of the operation of the camera shutter. A longer exposure (greater than l/2 s) allows the majority of the exposure to be taken after movement has died away, (2) using flash exposures that freeze the relative movement. Although this measure ensures that the image on each frame is not blurred, it does not compensate for differences in register caused by vibration between each frame. None of these measures are needed for the majority of low-power cinemicrographs where surprising stability is achieved on the simple bench.

426

Riddle

Fig. 2. Inverted microscope with tine camera mounted on triple eyepiece. Camera drive is situated on the camera body and the projection from the microscope is used to insert titles. Camera control is governed by the box on the right.

2.1.2.3. Microscopes. It has already been mentioned that the conventional microscopes, with objectives above the stage, are of less use than the inverted types to the biological cinemicroscopist. Earlier examples of inverted cinemicroscopes were conversions of inexpensive conventional types (14), and the modern instruments have only been common laboratory equipment for a relatively few years. However, unless special cell chambers are to be used, they are essential. Many of the well-known manufacturers now offer inverted microscopes, some with tine attachments and control units (Fig. 2). Choice is as much a matter of preference as available budget. Units from the larger makers do not differ very greatly in terms of their potential optical quality-which is always high.

Time-Lapse Cinemicroscopy

427

It should be remembered that, if a culture is grown in a plastic dish of less than perfect optical quality and where the thickness is not especially matched to the objective lens, then no amount of microscope superiority will improve on the image obtained. Also, the image is usually quite adequate for analytical purposes, even if less than perfect photographically. Choice should depend upon such features as accessibility of controls, robustness, ease of cleaning, and available accessories such as a separate still (PolaroidTM) camera. Optical components must be adequate. Of the accessories to check, a very long working distance condenser for phase contrast (clearance at least 2.5 cm) and very low magnification objectives with phase contrast are probably the most important. This microscope must be dedicated to its time-lapse role, preferably used by only one or two people. Any wider use nearly always leads to deterioration of the microscope and consequently to inferior or even ruined film. It cannot be overemphasized that, if second, third, or more cinemicroscopy units are available, the experimental time can be cut disproportionately. The single unit can never produce a true control film to compare with one of a test situation because it is made at a different time. For further reading upon microscopy in general and biological film production, the reader is referred to James (3) and Michaelis (2) (seeNote 6). 2.1.2.4. Seating. Films show progress of the subject afterward. However, prolonged periods of observation by eye are inevitable, especially when the subject changes frequently. Suchobservationis an invaluable aid for personal education about cells. There are few subjects more exciting to observe than the mitosis of a cell seen at high magnification, yet only a very few biologists can claim to have watched this elemental phenomenon. To do so, however, requires some hours on the stool by the microscope. Similarly, for observation and during the setting up and adjustment for filming, it is essential to have comfortable seating. The most important factor is to have seat and eyepieces at a relative height that allows the user to sit comfortably without the back being stretched or arched. Only experience with mismatched combinations of laboratory stool and microscope can impress this fact on the mind and elsewhere. 2.1.2.5. Cameras. A tine camera can be mounted permanentlyon the microscope using the microscope body to support or at least hold the camera body. There must be integral viewing to see and focus the actual image for the film. The image seen through the normal microscope lenses covers a larger field and is seldom in the same plane of focus. The majority of bioscientific filming is currently done on 16-mm film since the format

428

Riddle

and size are both convenient for handling and storage and are about the minimum that will give results of acceptable standards. The camera functions, namely time between exposures and length of exposure, together with the film advance are controlled by an Intervalometer, recently produced units being electronic rather than electromechanical. These may be small and mounted upon the camera itself or almost ludicrously large and smothered with flashing lights. It is usual to incorporate here a means to run the film on without delay, and in the more modern instruments, automatic exposure control and even an indication of color temperature for the users of color film. The cameras do not need to be capable of functioning as normal cinecameras, because the only part of the mechanism (usually) used in timelapse work is the film advance and shutter. The remaining components (for focusing and motor speed control) are not used. Where the camera control unit operates a motor drive that is fitted to a conventional camera back, the motor and control components are not used. As a result, it is sometimes possible to obtain the camera back, i.e., all but the motor drive and without any lens systems, for a price that is far less than that of the entire camera. Choice of camera is often dictated by the microscope manufacturer, but several types of varying degrees of sophistication are available. For laboratory work, it is unlikely that anything more than the most basic and simple camera back is required (seeNote 7). 2.1.2.6. Film. Film is available in both color and black and white types. For the majority of routine cell culture uses, black and white is adequate. Most culture specimens have no natural color and the image is monochromatic if obtained by phase contrast. Admittedly, the color effects produced by DIC (Differential Interference Contrast) are sometimes very elegant, but that often indicates that the system is not being used properly! As the majority of tine cameras are 16-mm format, the film does need to be of fine grain to avoid loss of detail. Also, the contrast of the optical system is such that a low-contrast film is not adequate for negative viewing (see Note 8) (although the additional contrast achieved during printing makes more conventional types suitable, the grain is more pronounced than that of microfilm). For this reason, it has been found that the film emulsion used in microfilms is adequately fast, practically free of grain, and nowadays produced on a very strong base that avoids damage in the projector. One such film (if not the only one now available in perforated form) is KodakTM Infocapture AHU microfilm 1454. Kodak has ceased to

Time-Lapse Cinemicroscopy

429

market this microfilm (in UK if nowhere else) in perforated form. Should a large enough order be required, the company has been willing to perforate a special batch. This has given most satisfactory results. Closed circuit videorecording of microscope images is also amenable to the time-lapse mode. The method currently used involves stopping and starting the movement of the normally continuously moving tape, while the rotating tape head over which the tape passes during recording remains in motion. This causes relatively rapid wear of that most expensive component of the equipment. Also, the tape suffers. In addition, reverse viewing of the resulting record is not as convenient as with film, except where very expensive videorecorders are used. There is the compensating advantage in that the record is available immediately, but given in-lab processing facilities for film, the differences are not great. These factors may not be of great significance where very short time intervals are required, but the longer intervals more often associated with the filming of tissuecultures can produce problems. Nevertheless, the method is used. The greatest disadvantage in videorecording lies in the rate at which changes are made to standard techniques. X-mm tine film is much the same now as it was 50 years ago and will still fit the projectors of that era, Video techniques are changing, and it is quite possible for two adjacent laboratories to be using incompatible apparatus. At present, the VHS system seems to be the most popular, but what the industry is going to offer us 10 years hence is open to debate. 2.1.2.6.1. DEVELOPING FACILITIES. Where only a small number of films are produced from time to time, development of the film should be entrusted to a specialist film laboratory. It should be noted that some of these now deal exclusively with color film, and obtaining first-class prints or copies of black and white films is much more difficult than some few years ago. Within the laboratory, short lengths (up to 10 m) can be developed in some special spiral developing tanks. This length is too short for most subjects, but there is one type that is capable of accepting a full 30-m length. Even so, the film has to be cut halfway along its length, which does not improve its analytical potential. More usefully, bench top developing units are available, and if space (not necessarily in a dark room) can be found, then the time from camera to projector can be reduced to 20 min. It is also cheaper provided enough developing is required. 2.1.2.6.2. SPECIMEN MARKER. Filming of cultures can be extended over many days, and the culture may need to be removed from the microscope for change of medium or treatment. Proper analysis demands that the

Riddle

,.-s-4 6
I 7 ,

;z' \

I c
3a

2 3

3
4

4
a

3b

Fig. 3. Objective markers. (3a) Eccentric diamond type. Mounted on the nosepiece, the diamond canbe rotated manually to mark the dish. (1) Diamond scribe, (2) adjustment for needle position, (3) ring for turning needle, (4) nosepiece thread. (b) ink marker type, (1) Lid(mustbereplaced whennotinuse),(2) triple fiber-tippenintriangularform, (3)plastic replaceable pen (Zeiss, Jena), (4) body to hold pen, (5) nosepiece thread.

same area be filmed before and after treatment. It is essential to mark the area under study to relocate it after such procedures. Simple location by stage coordinates is rarely adequate. Several microscope manufacturers offer objective markers at a great variety of prices. These units are of two types: (I) a rotatable diamond (or tungsten carbide) pointer mounted eccentrically on a block (Fig. 3a). The block is screwed into the microscope nosepiece like an objective and is concentric with the optical path. To mark a selected area, the marker is swung into position and a circle is scribed around the area under study, (2) a similarly mounted nosepiece block that bears three felt-tip pens in a triangle (Fig. 3b). These mark the area by simple pressure, but are not resistant to water or even wiping. It is possible to create a simple version of such a device by mounting the 0 of a hand printing set (e.g., John Bull printer) on an old objective and then inking it before bringing into contact with the surface. With both methods, when using an inverted microscope, the dish is lifted from the stage before a mark is made. One way to avoid this is to rack the microscope condenser down firmly onto the top of the dish (before inserting in an environmental chamber) before making the mark. Alterna-

Time-Lapse Cinemicroscopy

431

tively, a sheet of plate glass (warmed to stop condensation) on the top of an exposed dish will hold it in place. 2.1.2.6.3. TIME SWITCH. Where the intervalometer lacks any means to stop the film at a given time, a simple household time-switch of the sort used to turn lamps or heating units on and off is suitable. If a process under study is going to extend for some inconvenient length of time into the evening, let alone a weekend, a lot of film can be saved by switching off the intervalometer at an appropriate time. 2.1.2.6.4. PERFUSION SYSTEMS. Perfusion of the cell culture with fresh medium would appear at first sight to be the ideal method to obtain either constant conditions (16) or rapid changes in conditions. When fitted to a cinemicroscope and perfusing a Petri dish, the amount of plumbing required can be forbidding. It is necessary to arrange a pump to take medium from a reservoir to the dish, warming the medium en route, and a separate (or integral) pump to suck off medium from the dish at a predetermined height to maintain level. Failure of either can lead to a very messy disaster. Gassing of the reservoir and the piping to the dish is required where flow is slow enough for carbon dioxide escape through the tube walls. This involves a gassed jacket around the pipe, inside the pump, and around the culture. The medium must remain sterile and it is advisable to have a means to stir the medium while in the dish. 2.1.2.6.5. GRATICULE. In many films, the actual screen dimensions are of vital importance, as in the measurement of cell movement (6). It is possible to calculate the dimensions of the projected image using knowledge of the various lenses used, the microscope conformation, projection lens details and its throw, but these parameters are complex to calculate and relatively inaccurate in the answer obtained. In addition, factors such as film shrinkage (slight) at development all introduce further errors. It is always useful to have a list of various field sizes available with the microscope lenses available, but these are of use only when selecting lenses during setting up. It is better and absolutely accurate to insert at the beginning of any film (preferably just after the title and at the same magnification to be used) a short section of 3-5 frames of the image of a stage micrometer. These are scales photographed or etched on glass and mounted in a holder like a conventional microscope slide. For most cell biological purposes, a scale of 1 mm in length, subdivided into 1Oths and IOOths, gives a suitable image. 2.1.2.6.6. DEMISTING SYSTEM. Where culture chambers have an air space above the medium, as in those using Petri dishes, any maneuver that cools the lid or its container will cause misting. This can be so severe that

Riddle
the specimen is impossible to see because scattered light degrades the performance of the condenser. Before the dish is placed on the microscope, the mist can be dispelled rapidly, if somewhat hazardously, by gentle application of a Bunsen burner flame. For the less adventurous microscopist, a warm (not hot) plate of glass or metal, kept nearby can be placed on the lid to dispel1 the mist in a few moments. Such plates can be kept in an incubator or heated by an electric element (22). Once the dish is on the microscope, the mist is usually kept at bay by the close proximity of the warm condenser. This is usually enough to clear already misted dishes if not noticed before a film was started, but demisting by this method may take a long time.

2.2. Cinematography
2.2.1. Preliminary Set Up
Itis anadvantage to put a titleonto thefilmbeforesettingup themicroscope, especially if several cameras are in use. If there is no optical port available in the microscope system to this, the camera must be removed from the microscope and a film made of a title written on a board using a conventional lens. An alternative system is to make a short length of film with an England Finder on the stage. These finders are slides marked with circles containing rows of miniature letters and figures. Films can then be numbered serially from Al, A2, and so on. If the required magnification is known in advance, a short section showing a graticule is invaluable. Decisions made after preliminary examination either require the culture to be removed, the graticule inserted at the selected magnification, and then the culture to be replaced. An alternative is to put the graticule onto the film at the end of the film with the risk that there may not be enough left (e.g., after overnight or weekend filming). Knowing the basic requirements, the culture is then placed on the microscope stage and environmental requirements are adjusted. Following approximate microscope adjustment, a preliminary observation shows whether the culture is uniform with evenly distributed cells or whatever particular feature is to be demonstrated. The choice of a field almost inevitably introduces considerations of random fields. In some situations, it is difficult not to select a random field, which is a contradiction in its own terms. There is always the greener grass in the next field, and the only way to avoid such hazards is to ensure that there are simply enough cells present, that none are atypical, and

Time-Lapse Cinemicroscopy

433

that the photogenic aspect of the film is, at this stage, considered to be more important than the scientific.

2.2.2. Location
The field to be filmed must be marked, either by hand using a felt-tip pen inserted under the stage or with one of the devices described under equipment. This allows the same field to be located at a later date. A l?olaroidTM picture of the field is useful for accurate location before any operation involving the removal of the culture. Preferably this should be made through the viewing eyepiece, so that the eyepiece graticule is superimposed upon the picture of the cells. After a procedure involving the removal of the dish, the area can be relocated and orientated so accurately that it is hard to detect any movement on the subsequent film. A brief blank of unexposed frames to mark the time of change is valuable. This procedure usually involves a specially adapted PolaroidTM camera. The extra time involved saves a lot later. 2.2.3. Magnification The parameters studied dictate the magnification to be used for filming. If estimates of statistical value are needed, the sheer number of cells in the field is vitally important. Conversely, detailed study of single areas may need maximum magnification. Knowing the type of image to be seen from the film, the relevant selection of lenses is made and changes aremade at either the objective or projection eyepiece (projective) or using either of the others with a magnification control built into the microscope, which usually allows linear magnification changes from 1-1.5x. 2.2.4. Eme Interval Since various cell processes take a great variety of times, the interval must be chosen accordingly. No rules can be given, but remember that 30m of film carries 4000 frames plus a lead and tail each of 200 frames. A frame interval of 90 s uses 40 frames/h, and one of 60 s uses 60 frames/h. Whatever procedure needs to be filmed should be placed on a manageable length of film for analysis. It is time-consuming, boring, and wasteful to have to run repeatedly over more film than is needed. It has been found that a length of 10 ft, i.e., 400 frames, gives a manageable rate for analysis of a single parameter (for example, a complete cycle from one division to the next). The procedure then takes 16 s to project. Where the same process is repeated (e.g., repeated cell divisions), the time is added onto and not compressed into the original time. Full analysis of a film 4000 frames in

434

Riddle

length is possible but incredibly wearisome, largely because of the time taken in rewinding the film. (There are systems to give continuous circular, end to end projection of a film, but these are not appropriate for long films.) It is not possible to study two parameters of incompatible length, e.g., the total cell cycle and thelength of mitosis, on single film. Any attempt to do so results in one proceeding too fast for analysis and the other too slow for comfort. 2.2.5. Equilibration

Aculture inevitably suffers a relatively violent upheaval on its way to the microscope and it is desirable to allow it a period to equilibrate with its new surroundings before starting the film. If nothing else, aPetri dish that has been in a moist incubator will tend to dry on its lower surface if used in the system described above (seeEquipment); this differential moisture situation does lead to slight distortion and loss of focus. Any equilibration minimizes the effect. 2.2.6. Microscope Adjustment

It is only possible to suggest a system for microscope adjustment in the most general terms as different equipment varies. However, a strict routine for the setting up of such systems is invaluable. It is helpful to record the various settings on a note sheet, which is a checklist and future reference of both culture and microscopy (Fig. 4). A useful routine for phase contrast microscopy of a tine specimen would be: 1. 2. 3. 4. 5. 6. Locate the field. Focus on specimen. Mark the field (seeabove). Close field stop to within field of view. Move condenser to bring edge of field stop into sharp focus. Open field stop to be outside area used either in filming or in photometry (if automatic exposure is used). 7. Insert phase telescope and superimpose images of annulus and ring. 8. Observe image for defects not visible before adjustment.

Strict adherence to such a routine avoids errors of setting up with loss of image quality. It must be repeated whenever the specimen is moved or the objective changed.

Time-Lapse Cinemicroscopy
Number DM.

435

Date

Collaborators Disposal

Purpose

Film
Unit ,.....*....... ,.....*...
5Pe

I
..
. . ....................................... ..........................................

I
.................. ....................................

I
. . . . . . . . ..a......
.................

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ..I.........

Cells ..,...........,......................................,,.......,.........I................. ..a....... .. Msdlum .*....,.,.,... .*....,.,. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ..a.... Treatment

. . ....................................... .................. .................

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ..I................. Gas ..., . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .*...........**.. Temp.

.......................................... ........ ......... ................. .......................................... ........ ......... .................

Illumination . . . . . . . . . . . . ..I...................................... Filter . . . . . . . . . . . . ..I...................................... E$psure . . . . . . . . . ..)..f................................... or ASA . . . . . . . .. .. . . . .. . . .. . .. . . . . . .. .. .. . . . . . . . .. .. . . . . . . . . Intensity ,....,.. *...*,.,,*......,......,..***..............# Condanrar .,....*......* I...,................................,, Objective ,. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Prqective . . . . . . . . . . . ..I...................................... Intsrv.31 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ..I Prism . . . . . . . . . . . . ..I.....................................~ Fieldwidth . . . . . . . . . . ...* I...........,......................,.., Pilm stat . . . . . . frame . . . -7 . . . . . ..,......*........ I I ..*....*a c *.*....*.

...................................

........

.........

.................

........

.........

. .

..............

........

.........

......

..........

..................

.................

..................................

.........

.........

.................

........

.........

.................

.........

.........

.................

.........

..........................

...................................

.............. .................. ......... ......... ........ ......... ................. ......... / ................. / g;;;%s: .... / .................. 1 1......... / ......... 1........ 1
Fig. 4. Sample data sheet used to record experimental details that are otherwise so easily forgotten. It also actsas a checklist during setting up.length is possible but incredibly wearisome, largely because of the time taken in rewinding the film. (There are systems to give continuous circular, end-to-end projection of a film, but these are not appropriate for long films.)

c . . . . . . . . I . . . . . . . . . I . . . . . . . . . I...**...

2.2.7. Photometry Some systems available for control of exposure have been mentioned under equipment. With an automatic system, it is only necessary to ensure that the light level is adequate for the exposure chosen and that the correct

436

Riddle

film speed rating is set on the intervalometer. Manual adjustment must follow the instructions for the particular photometer in use. The speed rating marked on the film packet has been found to be an unreliable guide in cinemicrographic use (of one manufacture). In microscope use, it has been found better to make a test strip using a similar specimen to that for future study. Exposures are then based on the most satisfactory of the images obtained. To do this, select a test field, and either expose at the same light level and vary the film speed settings (on an automatic unit) while exposing a few frames at each or, on a manually controlled intervalometer, vary the length of exposure with the light constant. Put a blank frame between each change for location purposes. Changes of one full stop (i.e., a doubling of exposure at each change) is best on the first test, and if the result leaves any uncertainty between two levels, a more finely graded test at l/3 stop intervals should resolve all doubts. 2.2.8. Start of Film

It has been found useful to record on the checklist (Fig. 4) both the time of start of an experiment and also the frame number indicated on the intervalometer(s) at the start of an experiment. These indicate whether film transport is occurring at the expected rate and also whether enough film is left for any period when the unit is to be left running unattended. 2.2.9. Surveillance

The amount of attention required by the cinemicroscope when running depends (apart from individual idosyncrasies of the equipment) upon the specimen and to a large extent on the magnification. Once a culture has equilibrated with its surroundings, very little change may be visible in focus at low magnifications (less than 10x objective). At 4-10x that magnification (possibly using oil immersion), the relatively short depth of field may cause an unacceptable drift of focus within as short a period as a few minutes. Unless automatic focusing is introduced (a possibility that is scarcely worth considering because of the expense, even if reliable), the only answer for high-power work has been the adjustment at frequent intervals. Overfrequent readjustment, unless done very gently, can ruin the appearance of a film almost as much as a slight focus error. Fortunately, high-magnification filming is usually relatively brief. In contrast, the studies at lower magnifications can last for days to weeks and fortunately can often be left over a weekend without any checking.

Time-Lapse Cinemicroscopy

437

2.2.10. Removal of Film 16-mm tine cameras take daylight loading film. The spools have solid sides, and the outer five or so turns of film are sacrificed as a leader that keeps light from the exposed portion. The camera can be loaded in the light and-after the end of a film-a further 200 frames should be run on as a tail so that the film can be taken from the camera in the light. If no such lead is left and the film is developed at a processing laboratory, the last section will be lost. Be sure therefore to replace film before the final 400 frames have been used, or be prepared to remove film and start development in total darkness. 2.2.11. Development Assuming that development is being done in ones own dark room, it is essential to have used several test lengths of film to establish such variables as temperature of the baths of developer. Any variation in the quality of routine output calls for a repeat test. X-ray type developer and rapid fixer have been found ideal for the high-contrast film used in phasecontrast filming. The developer and fixer both have a limited life in terms of the amount of film they can process before exhaustion and also in absolute age once diluted. As a rough guide, assume for 1.5-2 L of working solution that the passage of 2 d is equivalent to the development of 30 m of film. The capacity of such developers varies. Although a developer that is approaching the end of its lifetime is capable of producing apparently satisfactory negatives, a follow-on film developed in fresh materials may show an unacceptable change in density when the two sections are spliced together. Fixers (seeNote 9) used as recommended usually last the equivalent time to the developer. 2.2.12. Storage Film for scientific purposes is seldom required in lengths of greater than 30 m, The cans in which the film is supplied are ideal for storage and can be kept in the original boxes, which have space for a note of the content on the outside. Although there are optimum conditions of humidity for film storage, it has been found that film kept in normal laboratory conditions does not suffer unduly. For demonstration, longer lengths of film are sometimes needed, and it is wise to have a small stock of 60 and 120 m spools available. 2.2.13. Splicing To join lengths of film, for whatever purpose, a splicer is needed. The most usual types, which work with conventionally backed film, use film

438

Riddle

cement to make the join. The ends to be joined are cut to overlap with the perforations in correct register for both sections. The overlapping portions are then scraped or ground to remove the emulsion from the one side and a little surface material from the other side. Film cement is applied to one or both abraded surfaces before bringing the two into firm contact. Some types use a heated anvil under the mating surfaces to speed up the setting time. Pressure must otherwise be maintained for a minute or so until the cement hardens. Film made upon a base that is not soluble in conventional film adhesives requires a different technique (which can also be used for conventional film). The joins, which are every bit as satisfactory as the cemented type, are made with an adhesive tape. The splicer for this method is used to trim the ends of the film, so that they butt accurately (usually across a pair of perforations). The film is then clamped with the ends in contact and a clear adhesive tape is placed over the joint. Downward pressure with a trimmer block presses the tape on, cuts the tape from the side of the film and perforates the tape in the same place as the film. Although it is not considered necessary to tape both sides, it is the authors practice to do so, since film for laboratory use has much more severe treatment than that for simple viewing. When purchasing a tape splicer, a double- or single-perforation type should be specified. Use of a single-perforation model involves an extra stage to perforate the second piece of tape, but a twin-perforated type is not suitable for some situations. However, these are unlikely to be encountered in time-lapse work.

2.3. Observation

and Analysis

2.3.1. Projectors
It is advisable to have a projector that is capable of both normal and analytical viewing. Unfortunately, all the projectors available use the conventional intermittent feed systems and rotating prism viewers, since those used in the editing of professional films are both very expensive and not as capable of producing a large, detailed image. These machines advance the film continuously (with consequently less wear and often tear), and it is surprising that a large screen unit of this type has not emerged for analytical cinematography. Several types of analytical projectors are available. Some of these have been developed from conventional projectors and others developed purely for analysis. Since the behavior of the various makes can be very different,

Time-Lapse Cinemicroscopy

439

it is advisable to consult other users before purchase. Some are superb; others tear, burn, or scratch film, and/or fail to operate for long periods. The conventional projector is, after all, only expected to run for half an hour at a stretch, whereas film analysis can continue all day and every day. An ideal analytical projector should be practically silent (few are, but some are worse than others, making work very tiring). They should be able to project the film in forward or reverse mode both at the normal projection speeds of 18 and 25 frames/s (fps), and a range of longer intervals, e.g., 12 fps down to 1 fps. They should be able to project a single frame (still frame) without any physical change such as warping to the film, let alone damage. More rapid film movement in either direction would be an advantage, but is seldom found. Control of the projector functions should be on a remote hand-held unit to allow the viewer to sit close by the screen. A frame counter mounted on the hand-control unit (as opposed to the projector itself) is invaluable to avoid Wimbledon neck. This is said in all seriousness and with experience. Automatic feed projectors are of dubious advantage in the laboratory situation, especially if it is required to remove lengths of film without rewinding onto either spool. Some automatic projectors are also very difficult to unload if a film has become damaged or displaced during projection. The projectors and projector analyzers in the laboratory are best used with a back projection system in such a manner that the screen is on the bench in front of the observer. Units to hold projector, mirror, and screen are available commercially. Alternatively, the projector can be mounted at a convenient height above the bench and a mirror fitted to project the image down onto a paper screen on the bench. A photographic tripod head with ball joint is a convenient way to mount the mirror to adjust it as it is often displaced by the head of the observer! With this arrangement, the image falls with the long axis across the bench and not along it as is customary. This does not matter when film of cells in culture is the subject, but can be confusing when trying to decipher a written title (seeNote 10). 2.3.2. Analytical Procedures

It is difficult to make any recommendations about the way to approach the observation of a film. The method depends on the parameters being investigated. However, it may help to indicate some general principles that have been found valuable. If the film is for demonstration purposes, a copy negative or at the very least a print should be obtained before putting the negative through the projector. The copy should be used for

440

Riddle

analysis if the negative is of great value. When ready to project, a preliminary scan of the film at normal projection speeds (25 fps) should bemade several times in forward and reverse modes to obtain a mental picture of what is happening. It is also possible at this stage to notice additional points of interest for the future. For example, a film made to study intermitotic times under changing conditions may show concommitant changes of movement and shape that were not suspected when the experiment was designed. Later investigationof these points may be fruitful. After the original survey has revealed whether the cells have shown the required behavior, a systematic analysis, cell by cell, is nearly always needed. Just what parameter is followed should have been decided before the culture ever reached the microscope. The amount of potential data on a film of crowded cells can appear daunting to the beginner and the experienced observer alike, but a systematic approach nearly always proves matters to be far more easily handled than feared. Common to most analyses, it is necessary to locate the cells at the first frame position, and the projector is used in the still frame mode while this is done. A projection screen upon which the observer can write becomes essential. For back-projection sys terns, remove the usual ground glass, and replace that with clear glass holding tracing paper held by clips or sticky tape. Number the sheet, and then before starting analysis, choose a frame. With the projector in still-frame mode, mark two recognizable and wellseparated fixed points. This enables the screen to be aligned for the same film with the minimum of effort should the systembe disturbed or another film have to be put temporarily in its place. If a simple growth curve is required (a useful starting point in studies of proliferation), a grid drawn on a clear plastic sheet can now be superimposed. The cells are ticked off with a wipe-off pen making counts at times chosen to be suitably separated for the rate of growth. This method has been found far easier than attempts to count the cells, even against a grid, without marking off, and a crowded field containing over 1000 cells becomes manageable. For other analyses, e.g., of intermitotic times, the paper screen becomes a cell map and the cells are numbered at their starting positions. Each cell is followed individually, and the place at which, e.g., mitosis occurs is noted and given its cell number. The beginner will find difficulty in following more than one cell at a time, but with experience and given a not too active culture, up to four cells (daughters and granddaughters of a division) can be followed at one time. Data can be written onto a family tree in this instance (26).

Time-Lapse Cinemicroscopy

441

Another more sophisticated analysis is that of speed and direction of movement of whole cells or cell components. A simple but time-consuming method is to draw the tracks of the cells with pencil upon the cell map and then to return to the beginning. Then advancing the film to frames that correspond to appropriate times, the point reached is marked on the original trace. Subsequently, each trace is measured with amap measure. Final conversion to pm/h is made with reference to the image of the graticule filmed at the start of the film. A far quicker method can be applied when a digitizer/computer combination is available by the screen. Given suitable software, such apparatus can measure the tracks of the hand-held cursor and also will convert the data into statistically analyzed summaries. The time taken to gathering the data onto disk, using one such system in the authors laboratory is approximately l/2 to 3/4 h/50 cells. Print out takes a short time. One form of analysis that is difficult if not impossible in any true sense is that of cellular morphology. The various individual factors that combine to make up this rather vague concept are all amenable to analysis, e.g., area, number of pseudopodia, number and shape of nuclei, and a host of others but it is seldom possible to reduce this to a numerical description. One exception to this rule occurs where a distinctive feature can be used to denote a particular morphology, and the change from motile, to non motile cells with a web of proliferations as described by Small et al. (17) is an example.

3. Notes
1. If long focus lenses are not available, it is possible to introduce a thinner area into the base of a plastic Petri dish. A hole is made with a heated cork borer or better with a hole saw. The hot tube method can leave a burr around the hole and often causes cracking as the plastic cools (remove the burr, carefully, with a scalpel). The hole saw method often leads to shattering of the dish. The hole is covered with a conventional coverslip that is sealed on using either a silicone rubber or two-part epoxy resin adhesive. Subsequent sterilization after the adhesive has cured is achieved by thorough washing, concluding with double-distilled water and finally with 70% ethanol for some 30 min. The vessel is then allowed to dry in clean air. A microwave oven can also be used to sterilize this dish. 2. Hair dryers of this type are now less popular, since high wattage mini-

442

Riddle

ature designs are the current fashion. If a suitable unit cannot be purchased, a box-mounted blower with a heater in the air path can be made into a similar device. 3. Problems of changing focus while cultures are being filmed (and which usually arise on Friday evening) have, in some cases, been found to be the result of the container. Considerable effort revealed that the focus varied because of distortion of the base on which the culture chambers were placed. This was caused less by temperature change than by humidity variation and differential humidity effects either side of the perspex bases that were first used. Substitution of firm, stainless-steel, or aluminium alloy bases has cured this problem almost completely. If the surface upon which the dish is seated cannot bend and the microscope controls do not slip, unexpected alteration in focus must arise in the culture dish base. It has been observed that dishes that are freshly seeded and previously dry do distort over a period of several hours (presumably) as a result of absorption of water from the medium into the wet side. Further slight changes can sometimes be seen with changes in atmospheric humidity and are avoided by full laboratory air-conditioning. High-magnification filming shows up these distortions far more than low magnifications, and there comes a time when constant correction at intervals of a few minutes (or automatic focus) must be used. Otherwise, variation must be accepted. The dishes with the stretched membrane bases (seeabove)are less prone to distortion causing change of focus. 4. A thought: Total darkness, even if possible in the technique being described, may be almost as unnatural as bright light. Skin cells can be exposed to very high light levels at times, and the light undoubtedly penetrates several cm below the surface. It is doubtful if the very center of a laboratory mouse is ever in total darkness during daytime! 5. Tests for vibration do not require the microscope to be moved from place to place. Simple feeling (as lightly as possible) with the back of the hand will detect slight movement. It is said that the tip of the nose or the forehead is more sensitive to vibration than other parts, but somewhat embarrassing to use if seeking a suitable site around a building! An easier test is to place a shallow dish of water on the surface being examined and to observe the water surface by reflected light. When there is vibration present, concentric rings form on the surface; that test is unfortunately sensitive to the periodicity of the

Time-Lapse Cinemicroscopy

443

movement of the water in the dish, and several different size dishes should be tried. 6. Filming of two microscope fields with one camera using a comparator head has been advocated but is complex. The only other possible means to obtain images of several areas on the same film is to use a computer-controlled stage that tracks repeatedly from area to area. If the wish is to make a final film comparing two different situations side by side, it is easier to prepare separate negatives and then have split screen prints made, showing half of one film and half of the other on the same print (15). 7. The differences in final film obtained with cameras of the various types can only be summarized under the headings of imperfections. Faults that are usually, but not always, brought about by the camera itself are: a. Scratching-check the various places where the film comes into camera contact to make sure that they are clean, not damaged, and free to rotate if in the form of rollers. Do not confuse this with scratches caused by processing systems (unlikely if made at a professional film laboratory), but easily caused in a bench unit if not perfectly clean. b. Inadequate or irregular film advance-this may appear as a wedge- shaped clear or dark area on some or all frames. Check that the shutter movement is correct, as seen when the camera is removed from the microscope and any lenses. Do this with film in place since the drag of the film can slow shutter movement. If faulty, first check the loading of the film and that the camera is clean. If still faulty, the relevant servicing should be obtained. Some camera-control systems can also cause similar effects and also require servicing. c. Unequal exposure produces single or multiple frames appearing darker or lighter than the rest (seeunder Environment). The only way to detect the cause is to observe the camera action, and again servicing may be indicated. This fault is also seen occasionally in film made with units that have automatic exposure. The cause there can be the result of extra light reaching the photometry system through secondary light paths (titling systems, other eyepieces, and so on). It can also stem from systems that have photometric devices that switch between sensitivity levels. Where the exposure being used is on the borderline at

444

Riddle

the changeover point, the exposure can alternate. The answer here is to set the light level so that the critical threshold is not approached. A somewhat different variation of film density can be the result of unequal rates of passage of the film through a laboratory developing system. The indication here is that darker sections have not been drawn throughout the developer at a steady rate. These sections (on the negative) tend to start at any part of the frame and are not related to fogging of the border. Overall fogging caused by light leakage onto the film appears somewhat similar, but in this case, the darkened areas cover both frames and the film between and around the perforations. d. Unequal spacing between frames. This can be because of dirt on the film guides, excessive film tension from any cause, or even badly perforated film. The latter does occur, but is very rare indeed. In all cases where faults are found, the value of more than one identical cinemicroscopy unit is incalculable. To be able to exchange components is the best way to locate the one that is faulty. 8. For much of biological microscopy, where phase contrast is used, the light densities of the images are not true black to white density differences, but contrasts induced by optical means. In this case, there is no theoretical objection to looking at the images in negative or positive form. Whereas such crimes are totally unacceptable in the realms of entertainment or illustrative photography, the scientist who wishes to extract data can do so even if the film becomes somewhat scratched, and can save a lot of time, space, and expense by viewing and analyzing the negatives. This has the added advantage that the image in many cases is more acceptable visually, because the grain of the film is less accentuated (prints are not made upon the fine grain stock of the microfilm) and the phase contrast halo of brightness around a darker area becomes an apparent shadow that is more acceptable to our perception of an image. 9. Some fixers are normally used with a hardener that is stored as a separate concentrate. Making up solutions without preliminary dilution of the fixer before adding the hardener causes a precipitation of sulfur (from the thiosulphate), which damages the film. Since fixer tanks in

Time-Lapse Cinemicroscopy

445

the small bench-top developing units do collect an encrustation of sulfur in the most awkward of places, a routine rinse out between changes of fixer may prove insufficient to remove anything but the more floccular debris. It is sometimes necessary to chip rock-solid sulfur away. A note of how much film can be developed before any undesirable deterioration occurs is invaluable. A record of the amount of film that has been processed after each change of solutions allows changing the solutions in adequate time. 10. Suitable mirrors can be made from bathroom mirror tiles, available in a variety of sizes at hardware shops. The inevitable ghost image caused by reflection from the unsilvered front surface is not significant. However, superior mirrors can be made of plate glass sheets surface coated with aluminium in a standard EM evaporation unit. The reflectance is superb and undistorted, but the surface is easily damaged.

References
1. Marey, E. J. (1894) Le Mouvement (Masson, G., ed.), Paris. 2. Michaelis, A. R. (1955) Research Films in Biology, Anthropology, Psychology, and Medicine (Academic Nijhoff Press, New York).

3. See J. James (1976) Light Microscupy Techniques in BioZogy and Medicine, Martinus 4. Richmond, 5. 6. 7. 8. 9. 10. 11. 12. 13.
Medical Division, Netherlands, pp. 151,152. K. M. V., Riddle, P. N., and Brooks, R. F. (1984) Apparent desensitization of Swiss 3T3 cells to the mitogens FGF and vasopressin. J. CeZI Physiol. 121,547-557. Wilkinson, C. and Lackie, J. M. (1983) The influence of contact guidance on chemotaxis of human neutrophil leukocytes. Exp. Cell Res. 145,255-264. ONeill, C. H., Riddle, P. N., and Rozengurt, E. (1985) Stimulating the proliferation of quiescent 3T3 fibroblasts by peptide growth factors or by agents that elevate cellular cyclic AMP level has opposite effects on motility. Exp. Cell Res. 156,65-78. T.ackie, J. M. (1986) Cell Movement and Cell Behavior (Allen and Unwin, London, Boston, and Sydney). Brooks, R. F., Riddle, P. N., Richmond, F. N., and Marsden, J. (1983) The Gl distribution of Gl-less V79 Chinese hamster cells. Exp. Cell Res. 148,127-142. Salmon, E. D. and Ellis, G. W. (1975) A new miniature hydrostatic pressure chamber for microscopy. J CeZZ BioZ. 65,587-602. Riddle, P. N. (1979) Time-Lapse Cinemicroscopy (Traherne, J. E. and Rubery, P. H., eds.), Academic Press, London and New York. Roberts, D. C. and Trevan, D. J. (1961) A versatile microscope chamber for the study of environmental changes on living cells. 1. R. Microsc. Sot. 79,361366. Riddle, P. N. (1983) A device for demisting Petri dishes. Lab. Practice 32,80. Riddle, P. N. (1977) Beam deflection as an alternative to light shuttering in cinemicroscopy. Lab. Practice 26,865.

446
Microsc. Sot. 79,367,368.

Riddle

24. Trevan, D. J. (1961) A simple inverted microscope for use with a tine camera. 1. R. 25. Riddle, P. N. (1974) Comparison cinemicrographs prepared by optical split screen printing, Report of2nd International Colloquium of Znterkamers 73 (Serb, V. and Sibalova, J., eds.), In Vitro vs CSSRHradci Kralove. 26. Brooks, R. F. and Riddle, P. N. (1988) Differences in growth factor sensitivity between individual 3T3 cells arise at high frequency: possible relevance to cell senescence. EXQ. Cell Res. 17~4378-388. 17. Small, R. K., Riddle, P.N., andNoble, M. (1987) Evidence for the migration of oligodendrocyte-type-2 astrocyte progenitor cells into the developing rat optic nerve.
Nature 328,155-157.

Chapter

34

Transmission and Scanning Electron Microscope Preparation of Whole Cultured Cells

Josef Neumiiller
1. Introduction
Despite the enthusiasm of the first investigations of cell ultrastructure, morphological studies have since lost some of their importance for biomedical research. The development of quantitative biochemical methods has been the cause of this reduced interest in morphology. Biochemical reactions, however, take place in compartments of the cell and the extracellular matrix. This compartmentation, in an ultrastructural dimension, is the prerequisite for a systemic discharge of metabolic processes in temporal continuity. This compartmentation is provided by phospholipid biomembranes serving as support for sets of enzymes or receptors and as vessels for internalized or synthesized substances. Investigations of the biochemical processes in compartments are only possible if these compartments can be separated by ultracentrifugation or by ultrahistochemical methods. Routine preparation for transmission electron microscopy (TEM) preserves only the morphology of cells and limits the possibilities 447

448

Neumilller

for staining to electron-dense reaction products. Furthermore, the electron beam causes considerable contamination and damage to the object. In the past decade several attempts have been made to overcome these problems in order to obtain realistic morphological representations of cells and to perform ultrahistochemistry. First of all the development of cryomethods has made it possible to retain and preserve substances at their original site. However, the use of only soft aldehyde fixativesand the renunciation of crosslinking reagents such as 0~0, causes a loss of contrast of the ultrastructural components. The newer techniques of: (a) cryotransfer from the ultramicrotome to the EM in a vitrified state, (b) cryosubstitution that involves a gradual substitution of ice with organic solvents and finally with resins that polymerize at low temperatures, (c) controlled freeze drying over a molecular sieve, and (d) the replica techniques taken from freeze-fractures can partially resolve the problems of conventional preparation methods. The problem of contrast can also be solved in some cases by electronic contrast enhancement and image processing. The lack of information about absent ultrastructural details in sectioned cells can to some extent be compensated for by investigations using SEM (Scanning Electron Microscopy) or HVEM (High Voltage Electron Microscopy). The following chapter deals with preparation schedules for several anchorage dependent and suspension cell cultures as well as for cell cultures grown in gels. In the early days of EM, whole cultured cells on formvar coated grids were investigated but gave poor resolution under the normal 50 kV electron beam. After the development of ultramicrotomy these methods receded into the background. Metallurgists, however, were faced with the need to develop a HVEM allowing penetration of thin metal foils. Appropriate electron microscopes using a voltage of I MeV were constructed. Biologists, too, became interested in the features of these microscopes. Above all, the microtrabecular lattice was clarified by Porter and his colleagues. Only few biological investigations have been carried out because of the small number of HVEMs available in the world. New generations of high resolution electron microscopes provided a compromise by working at a high voltage (HV) up to 400 kV. With this HV setting it is also possible to penetrate through cells that have been spread on grids (1-5). This method is very powerful, particularly for investigation of cell motility, the cytoskeleton, and its interactions with the extra-cellular matrix (ECM) using immunogold particles and contrasting with silver stain.

Transmission

and Scanning Electron Microscope

449

2. Materials
1. Fixative: 2% glutaraldehyde (GA) in O.lMNa-cacodylate-HCl-buffer + O.lM sucrose (pH 7.2), total osmolarity: 510 mOsM, vehicle osmolarity: 300 mOsM. GA is commercially available as a 25% aqueous solution (prepare fresh before each use); 1% 0~0, in Na-cacodylate-HCl buffer (can be stored at 4OCfor some months). 2. Na-cacodylate-HCl buffer as in ref. (I) without GA and sucrose as a washing solution (can be stored at 4OCfor some months). 3. A gradual series of ethanol: 30,50,70,80,90, and 95% ethanol in double distilled water; absolute ethanol that has been previously purified of any water and particles, using a molecular sieve that absorbs molecules with an effective diameter ~0.3 nm (store at 4C). 4. Intermediate fluids: Freon TF; gas flasks with Freon 13 or 116, or pure co,. 5. Reagents for embedding: Propylene oxide (1,2-epoxypropane); Nbutylglycidyl ether; ultralow viscosity resin (ULVR): 100 g ERL 4206 resin (vinylcyclohexene dioxide), 200 g HXSA (hexenylsuccinic anhydride), 25 g Araldite RD2 (DY026) resin, and 25 g DMAE (dimethylaminoethanol) (store at 4OC). 6. Formvar resin, 0.2%, in water-free chloroform in order to prepare supporting film on EM grids (prepare fresh). 7. Materials spheroids, microcarriers, or coverslips for sticking cells, onto the specimen mount for SEM: 10 mg poly+lysine in 100 mL double distilled water; a double-sided adhesive tape; a silver-loaded epoxy adhesive. 8. Reagents for removal of resin from embedded cells in gels: 0.5 g crystalline NaOH in 50 mL absolute ethanol; a gradual series of 30,60,70, and 90% amyl acetate in absolute ethanol and pure amyl acetate (prepare fresh). 9, Chemicals for the freeze-fracture: A 5 mM aqueous solution of polyL-lysine (MW 2000-4000); 0.5% boiled and filtrated starch solution in double distilled water (both to be prepared fresh); Freon 22; 10% hydrofluoric acid (store at 4OC). 10. Materials for the preparation of microchambers: Beem capsules (size 00); nylon gaze (mesh width 100 pm); dialysis tubes (10 mm diemeter, separation at MW 50,000). 11. Small Beem capsules (size 3) for direct embedding of selected areas of monolayers.

450

Neumiiller

12. Other materials: Round coverslips (10 mm diameter); glass fiber grids (mesh width 0.5 mm) for handling of floating monolayers; gold grids G 200; copper grids with a hexagonal pattern (repeat distance 460 lines/in.); appropriate flat embedding forms made of silicone rubber.

3. Methods 3.1. Method 1: Simultaneous Preparation of Cell Cultures for SEM and TEM (6)
3.1.1. Fixation 1. Fix in 2% glutaraldehyde (GA) in O.lM Na-cacodylate-HCl buffer plus O.lM sucrose (pH 7.2), total osmolarity: 510 mOsM, vehicle osmolarity: 300 mOsM (seeNote 1). Add prewarmed (37OC)fixative in equal volume to the culture medium for 5 min without moving the culture vessel (seeNote 2). 2. Replace the fixative/mediummixture with pure fixative at room temperature for 1 h and gently agitate the culture vessel. 3. Change the fixative and fix for a further 24 h at 4OC. 4. Rinse 3 x with Na-cacodylate-HCl buffer at room temperature. 5. Carry out postfixation in 1% 0~0, in O.lM Na-cacodylate-HCl buffer (pH 7.2) at room temperature. 3.1.2. Dehydration

Use gradual steps of ethanol as follows: 2 x 5 min in 30% and in 50%; 10 min in 70%, 80%, 90%, 95%; 2 x 10 min in 100% (seeNote 3). 3.1.3. Drying at the Critical PointPreparation for SEM

1. Incubate with Freon TF as intermediate fluid (can be omitted if polystyrene vessels are used). 2. Transfer the sample to the specimen boat of the critical point drying apparatus (CPDA). In order to avoid artifacts the specimen has to be kept immersed in Freon TF or absolute ethanol (Fig. 1). 3. Put the specimen boat into the CPDA cooled to IO-15C. Close CPDA and fill with CO, or Freon 13 or 116. 4. Allow the fluid in the specimen boat to escape and change theintermediate fluid 3x at 5 min intervals.

Transmission and Scanning Electron Microscope

451

Fig. 1. Human skin fibroblasts grown on glass coverslips. Artifact by drying (bar=10 m). 5. Increase the temperature until the critical point is reached (it differs in respect to the intermediate medium used): Carbon dioxide (CO,>: Tc = 31.3OC PC = 75.5 kg/cm2 Freon 13 (CClFJ: Tc = 28.9OC PC = 39.5 kg/cm2 Freon 116 (CF,-CF,): Tc = 19.7OC PC = 30.4 kg/cm2 (Tc = critical temperature; PC = critical point.) 6. When the critical point is reached, the liquid interface becomes opalescent and disappears. Allow the gas to escape slowly in order to avoid recondensation.

3.1.4. Sputter Coating

1. Place the specimen onto a specimen mount using a conductive adhesive. 2. Put the specimen mount into a sputter coater and cover it with a goldpalladium @O/20) layer of 1.5-2 nm (seeNote 4).

452 3.1.5. Alternative

Neumiiller

(continue

Embedding-Preparation for TEM after dehydration step described above)

1. Put the specimen into propylene oxide for 10 min (this step may be omitted if polystyrene vessels are used). 2. Change the ULVR (ultralow viscosity resin) 3x at 4 h intervals or transfer the specimen directly from 100% ethanol to the resin in a flat embedding form. The resin is miscible with absolute ethanol. 3. Polymerize in an appropriate oven at 70 C for 12 h.

3.2. Method 2: Preparation of Cell Pellets, Spheroids, Cell Suspensions and Cells Grown on Microcarriers for SEM
3.2.1. Cell Harvesting
1. Filter the cell pellets, spheroids, and microcarriers through a nylon gaze (mesh width: 100 l.rrn) placed in a small glass funnel. 2. Rinse with 20 mL culture medium (without FCS), prewarmed to 37OC. 3. Incubate a dialysis tube (separation at MW 50,000,10 mm diameter) in PBS for 30 min before use. Knot one end of the tube, cut out the tip of the filter with the cell aggregates or microcarriers, and place it into the tube. Fill the tube with medium without FCS and knot the second end of the tube. 4. Alternatively, centrifuge cell suspensions at 800x g for 10 min, resuspend in a small amount of culture medium without FCS, and fill directly into the dialysis tube.

3.2.2. Fixation,

Dehydration,

and Critical

Point Drying

Perform according to Method 1. The cells, aggregates, or microcarriers remain in the dialysis tubes, which are placed into 20 mL tumbler vials, completely filled with the respective liquid and moved by means of a rotation tumbler. Avoid flushing too rapidly and changing the pressure in the CPDA too fast during outstreaming of gas in order to prevent bursting of the tubes.

3.2.3.

Coating

with Gold-Palladium

1. Cover the SEM specimen mounts with a double-coated adhesive tape (the tape should be somewhat smaller than the top surface area of the specimen mount). Open the dialysis tube and disperse the contents of the filter over the sticky surface of the tape.

Transmission

and Scanning Electron Microscope

453

2. Invert the specimen mounts and remove the excess of particles by generating of a soft air stream from an air puffer. 3. Connect the edge of the tape to the specimen mount with a droplet of conductive epoxy adhesive. 4. Put the samples in a desiccator which contains silica beads, evacuate and allow the glue of the tape to dry for at least 24 h. 5. Perform sputter-coating according to Method 1.

3.3. Method 3: Preparation of Cell Pellets, Spheroids, Cell Suspensions (Fig. 2a-c) and Cells Grown on Microcarriers (Fig. 3a-f) for TEM
1. Cell pellets from monolayers or centrifuged cell suspensions, spheroids and sectionable microcarriers can alternatively be processed for TEM. Instead of the dialysis tubes, Beem capsules covered with a lOOpm-mesh-width nylon gauze are used. Cut the gauze to an appropriate size and clamp to the capsule with the bored capsule cap (Fig. 4). 2. Fix and dehydrate according to Method 1 (the chambers made from the Beem capsules must be submersed completely in the respective fluids of the tumbler vessels). If polydextran beads are used, the incubation time should be extended 2-3x to permit sufficient infiltration. 3. After dehydration in absolute ethanol submerse the capsule chambers in changes of ULVR for 4 x 4 h. Each change of resin should take place in a separate tumbler vial moved by means of a rotation tumbler. 4. Perform the infiltration with resin at 4OC. After the last incubation bring the vials into a vertical position and clean the outsides with a filter paper wetted with propylene oxide. 5. Allow the particles in the capsule to settle and perform polymerization at 70C for 12 h. 6. Trim the blocks. This is quite easy because the particles are concentrated in the tip of the capsule. 7. For sectioning, the use of a diamond knife is recommended, above all if microcarriers are embedded.

3.4. Method 4: Preparation on Coverslips or Plastic

of Monolayers Grown Petri Dishes for SEM

1. Use plastic Petri dishes with four ring-divisions fitted with 1 cm diameter coverslips previously marked with an asymmetric letter and sterilized (Fig. 5) or 6-cm diameter Petri dishes to grow cells.

454

Neumiiller

Fig. 2a. Attachment of human erythrocytes to an adhesive tape. Preparation by using dialysis tubes @ar = 5 pm).

Fig. 2b. The same preparation as Fig. 2a, but of human granulocytes (bar = 5 ~1.

Fig. 2c. The same preparation as Fig. 2a, but of human buffy coat cells (bar = 4 ~1.

Transmission and Scanning Electron Microscope

455

Fig. 3a. HeLa cells grown to a high density on glass microcam ers (bar = 10 pm).

Fig. 3b. Cell contactsand surface projections of HeLacells (detail of Fig. 3a;bar = 5~1.

456

Neumiiller

Fig. 3c. Human

skin fibroblasts grown on micmcarriers

(Cytodex 3) (bar = 100 pm).

2. Fix and dehydrate cells after spreading and multiplication according to Method 1. 3. (a) Displace the coverslips carefully from the dish with a preparation needle and a tweezer with angled parallel edges and put into a special coverslip tray (Fig. 6) that fits the specimen boat of the CPDA. Drying can be easily performed (Method 1) in this tray that can be made by anyone with do-it-yourself experience. (b) Ifcellswereplatedonplastics,stripsofthebottomofthedishescan be cut out with a hot scalpel blade. This has to be carried out very quickly to avoid desiccation of cells. The plastic strips can be marked as previously mentioned and processed in tumbler vials. They can be dried without the use of Freon TF, e.g., directly from absolute ethanol to co*.

Fig. 3d. HeLa cells grown on microcarriers

(Biosilon)

(bar = 30 pm).

Fig. 3e. TEM at low magnification of human tendon fibroblasts grown to a high density on Biosilon microcarriers which have not been stirred (* = microcarrier; bar = 1.5 pm).

I I I Fig. 3f. As Fig. 3e, but at higher magnification (bar = 5 prr$

457

458

Neumiiller

Fig. 4. Beem capsule with a filter (F) clamped to the perforated cap (C) containing microcarriers (ML

4. Mount the coverslips or plastic pieces on to the specimen mounts with a conductive epoxy adhesive, place into a desiccator under vacuum for 24 h, and sputter coat as described under Methods 1 and 2.

3.5. Method

5: Simultaneous TEM Preparation of Cultured Cells

Cells Grown on Round Coverslips

Process as in Method 4 until dehydration in 100% ethanol.

3.5.1.

1, Fill the rings containing the coverslips with a large drop of ULVR, which is changed 3x at 10 min intervals at 4OC. 2. Fill a Beem capsule completely with ULVR, inverse rapidly and place over the coverslip in a vertical position. The diameter of the Beem capsule should be cl cm. 3. Polymerize in this position at 70C for 12 h.

Transmission and Scanning Electron Microscope

459

Fig. 5. Petri dish with ring-divisions

containing

marked coverslips. (F)

4. Break away the Beem capsule from the Petri dish. Split the coverslip from the resin block with a scalpel blade or by rapid cooling with CO, snow. 5. Carry out appropriate trimming after orientation under an inverted light microscope.

3.5.2. Cells Grown on Plastic Dishes (7)

1. Mark the edge of a selected area of cells inside the Petri dish with a thin preparation needle. 2. Replace the 100% ethanol in the dish by 2 mL ULVR. 3. Change it at 3x 10 min intervals at 4OC. 4. Cover the marked area with an inverted and resin filled Beem capsule as mentioned in step 2 above (3.5.1) and then process further as described there. 5. When the Beem capsule containing the resin block is split from the dish, the marks are visible as replicas on the flat surface at the aperture of the Beem capsule and serve as a guideline for trimming. 6. Orientate the block in the ultramicrotome at an exact 90 angle to the cutting axis. After one semi-thin section, perform ultrathin sections parallel to the monolayer plane.

460

Neumftller

Fig. 6. Specimen boat of a CPDA with metal grid trays CT).

3.6. Method of Monolayers

6: !Floating Sheet* Preparation for TEM (Fig. 7a, b) (8) (See Note

5)

1. Grow cells in 6-cm diameter plastic dishes. 2. Fix and dehydrate according to Method 1. 3. Add 5 mL propylene oxide or N-butylglycidyl ether to the dishes. Put a white porcelain plate under the dishes. 4. Move the dishes gently when the plastic appears rippled (after about 5 min). The monolayer floats up to the surface of the liquid. 5. Harvest the floating monolayer with a glass fiber grid (mesh width 0.5 mm), invert the grid and place it carefully in a flat embedding form filled with ULVR. Press the grid slightly against the resin surface using a glass rod in order to separate the monolayer from the grid. 6. Polymerize at 70C for 12 h.

3.7.

Method 7: Whole Mount Preparation of Cultured Cells on Grids (9) 3.7.1. Preparation of Grids

1. Wash slides in distilled water and dry with Kleenexm. Remove remnant dust particles by using a gas jet duster. Do not clean with soap,

Transmission

and Scanning Electron Microscope

461

Fig. 7a. TEM preparation of floating sheets of human ligamentous fibroblasts. Note the electron-dense glycoprotein layer (AA,) and the overlapping of these cells of a confluent monolayer (bar = 2 PI.

Fig. 7b. TEM preparation of floatingsheets of human ligamentous fibroblasts. the production of collagen fibrils (?I in deep recesses of the cell (bar = 0.5 ~1.

Note

462

Neumiiller solvent, or acids because in this case the separation of films is very difficult. Prepare a 0.2% solution of formvar resin in water-free chloroform. The resin must be dissolved completely in the solvent. Use a magnetic stirrer. Run a thin preparation needle along the edge of the slide. Seize the coated surface and dip the slide slowly at an angle of 45Ointo a glass trough filled with water. Under appropriate illumination the floating of the formvar film separating from the slide can be distinctly seen. Transfer the films to gold grids using a grid filming device. The grids are held by a TeflonTM plate at the bottom of the trough. The surface of the liquid is slowly depressed and the films come in contact with the grids without any folds forming. Remove excess water and dry the grid plate. Stabilize the filmed grids with a carbon layer using a vacuum coater. Make the carbon surface hydrophilic by irradiation with a UV lamp for 30 min.

2.

3.

4.

5. 6. 7.

3.7.2. Attachment
1.

of Cells to Grids

a. Cells are grown on the slides by incubation in small plastic dishes. Avoid sliding together of grids. Petri dishes with divisions (Fig. 4) may be used. b. Nonadherent cells can be attached to the carbon coated surface by dipping the grids into 10 mg/mL 30 poly+lysine and washing 3x in distilled water. Place the grids into the plastic dishes and allow the cells to settle and adhere to the coated surface. 2. Transfer the grids for further processing into small Beem capsules that have been perforated with some holes at the conical part and on the sides in order to permit a good fluid exchange. As an example, HeLa cells grown on filmed grids are shown in SEM (Fig. 8a) and in TEM mode as whole cell mounts (Fig. 8b).

3.7.3. Fixation, Dehydration, Critical and Carbon Coating

Point Drying,

All techniques are performed according to Method 3. The grids with the attached cells are coated a second time with carbon.

Transmission

and Scanning Electron Microscope

463

Fig. 8k HeLa cells in TEM mode prepared by the whole cell mounting technique (bar =lprn).

Fig. 8b. HeLa cells prepared in SEM mode by the whole cell mounting technique (bar =5op.m).

464

Neumiiller

3.8. Method 8: Preparation of Cells Cultured on, or in, GeZs for TEM and SEM (See Note 6)
1. Perform fixation in situ according to Method 1. 2. Extend the rinsing with Na-cacodylate-HCl buffer to twice the usual time. 3. a. Cut the gel into small cubes that then are transferred to tumbler vessels and dehydrate according to Method I. b. If the cells are grown on the surface of the gels, dry the cubes in small gaskets in the CPDA. 4. a. Embed cells grown inside the gel in ULVR after dehydration in tumbler vials: 100% Ethanol 3x 1 h Propylene oxide 1 h 1 part ULVR + 3 parts propylene oxide 1 h 1 part ULVR + 1 part propylene oxide 1 h 3 parts ULVR + 1 part propylene oxide 1 h ULVR alone 2x 3 h b. Prepare cells grown on the surface of the gels according to Method 1. 5. Place the cubes in silicone embedding molds containing ULVR and polymerize at 70 C for 12 h. 6. Trim the blocks and make sections for TEM. 7. After the TEM observation, shorten the resin block to a flat disk on which the trimmed pyramid remains (compare with [lo]). 8. Incubate the disk in 0.5 g of crystalline NaOH in 50 mL of absolute ethanol for 30-60 min while the solution is agitated with a magnetic stirrer. Avoid evaporation of the solvent. 9. Perform critical point drying when the removal of the resin is completed, using intermediate fluids as indicated in Method 1. 10. Incubate 3x in absolute ethanol for 10 min, and change stepwise to amyl acetate (7:3,4:6,3:7,1:9, and 100% amyl acetate) for 10 min at every step. 11. Dry in the CPDA with CO, for 30 min. 12. Sputter with gold-palladium (seeMethod 1).

Freeze Fracture

3.9. Method 9: of Monolayer (ll) (See Note 7)

1. Attach the cells to a glass coverslip using an aqueous solution of 5 mM poly-L-lysine (MW 2,000-4,000).

Transmission

and Scanning Electron Microscope

465

2. Coat the slide with a thin film of a 0.5% boiled and filtered starch solution. 3. Bring the coverslip side with the attached cells into contact with the filmed side of the slide without applying any pressure. 4. Place the slide with the attached coverslip in a freezing container filled with Freon 22 that has been precooled with liquid nitrogen to -160C. 5. Separate the coverslip from the slide using a razor blade mounted on a polyethylene sheet. 6. Place the broken parts in a freezing stage that is put in a vacuum unit on a liquid N2 cooled shield. N2 is released and the system evacuated at 10 torr. 7. Increase the temperature to -8OOCusing an internal heater inside the vacuum sys tern. 8. Remove the cooling shield and shadow the specimen with platinum at a 20 angle and with carbon at a 90 angle. 9. Float off the replica in 10% hydrofluoric acid, rinse with distilled water and mount on grids according to Method 7 (3.7.1[4]).

4. Notes
1. There is one basic principle that needs to be followed in order to get good preparation results. As mentioned already above, cultured cells are very sensitive to changes in pH, osmolarity, and variations in temperature. These parameters have to be maintained during the aldehyde fixation steps at a level that is similar to that found in living cells. Aldehyde fixatives do not completely break down the semipermeability of the plasma membrane. Therefore the osmolarity should be at 510 mOsM (vehicle osmolarity 300 mOsM) (6,12) and it is recommended to check the osmolarity with an osmometer. Even in cryomethods, in which cells are frozen with the addition of a cryoprotectant (13), the osmolarity is of great importance. 2. During fixation, changes in temperature or vigorous shaking can completely alter the shape of the cells. Therefore one should add GA at 37C to the complete medium for only a few minutes. This precaution stabilizes the plasma membrane and its projections. If the mixture of GA and medium remains in the culture vessels for more than a few min, protein aggregates from fetal calf serum cover the cells and cannot be removed. In a second step GA is added to an adequate fixation buffer (always at 37C) that is allowed to cool down to room temperature after 1 hr, the fixative is replaced again and the fixed cells are kept at 4C for 12 h.

466

Neumiiller

If one is interested in a good preservation of microvilli and other cell membrane projections, postfixation with 0~0, is indispensable in order to avoid bubble-like swelling. 0~0, on the other hand, increases the rigidity and fragility of the cell membrane. This leads to clefts and fragmentation. Therefore, the appropriate fixation and the right concentration of 0~0, have to be found (Fig. 9a-d). 3. For dehydration in ethanol or acetone the cells do not need longer than a few min, but care has to be taken to make sure that the substratum is also well infiltrated and dehydrated, above all if microcarriers made of DEAE cellulose are used. 4. After critical point drying the specimen is placed onto a specimen mount, using a double-sided adhesive tape or a drop of silver-loaded epoxy adhesive. In order to avoid charge problems it is recommended that conductivity is provided between the specimen and the specimen mount by a droplet of a conductive adhesive. Such adhesives are available and they dry in a few min. The specimen mount with the specimen should be stored in a vacuum desiccator overnight before coating. For coating a sputter coater is usually used. In this apparatus the specimen is put into a vacuum jar in a holder that is at the same time the anode of alarge diode. The cathode is situated above the specimen and holds a gold target. The system is evacuated until co.05 torr are reached. Pure argon is let in several times in order to replace the remaining air, When the vacuum of 0.05-0.07 torr is reached again, a high voltage of about 1 kV is set. Because of the electron-push the argon atoms are ionized before they reach the gold target and cause an emission of gold from the cathode to the specimen surface at the anode. This process causes a violet light in the jar. The particles collide and are thereby slowed down. Thus the vacuum decreases and is kept at 0.1-0.2 torr by a minimal influx of argon through a needle valve. The argon influx triggers the electric current that should be maintained at about 10 mA during sputtering. By these three parameters (voltage, electric current, and sputter time) the thickness of the gold layer can be gaged. It should be about 0.15 nm in order to avoid burying fine surface detail under the metal coating. The size of the gold grains can be diminished by using gold-palladium targets (22). Mostly the observation in SEM is performed using secondary electrons. In immunohistochemical methods with gold particles bound to receptors or antigens, the detector for back-scattered electrons is used,

Transmission and Scanning Electron Microscope

467

Fig. 9a. Spread human skin fibroblast. preserved microvilli (bar = 4 WI.

Cell fixed with GA and OsO,. Note the well

Fig. 9b. Human fibroblast grown on a microcarrier coated with gelatin (Cytodex 3) fixed only with GA. Note the bubble-like contracture of microvilli (bar = 4 ~1.

468

Neumuller

Fig. 9c. Human synovial fluid cells attached to glass coverslips. Fixation with cooled GA in cacodylate buffer (4C). Note the retraction of the cytoplasm at thelobopodia. Only radial remnants of the cytoskeleton are visible (bar = 10 cun).

Fig. 9d. Good preservation of the cell shape of human skin fibroblasts (bar = 20 ~1.

Transmission

and Scanning Electron Microscope

469

Fig. 10. Macrophage attached to a glass coverslip labeled with a monoclonal antibody against HLA DR (BMA 020) and in a second step with a goat-anti-mouse immunogold probe (VI (bar = 2 ~1.

which gives good contrast between the gold particles and the rest of the cell surface (14-22) (Fig. 10). 5. For the simultaneous processing of cell cultures for TEM, there are numerous possibilities. Cells can be grown in a second Petri dish of the same size but without divisions, and prepared as a floating sheet (8). By this method cells are placed in several layers in which they maintain their contacts and are sitting on a thin glycoprotein film, which in cross-sections is visible as a distinct, electron-dense line (Fig. 7a). With this method it is possible to show the formation of collagen fibrils in deep recesses of the cell surface of fibroblasts in confluent cell cultures (Fig. 7b). 6. If cells are grown in semisolid gels no cubes can be cut. Therefore, embedding is performed in situ in the culture vessel. Appropriate cubes are cut after polymerization. Further processing is the same as the preparation of cells in gels. 7. The main problem in cryomethods is the formation of ice crystals. To overcome this, sucrose, DMSO, or glycerol is added to the culture medium before freezing (13,23). These substances, however, particu-

470

Neumiiller

larly alter the lipid and protein constitution of the cells. Therefore freezing methods without cryoprotectants have been developed. Immediate immersion into liquid nitrogen is not useful because the contact of the warm surface bearing the monolayer causes the formation of bubbles. For this reason precooled polished copper plates are commonly used to which the monolayers are rapidly attached (24,25). The procedure described in Method 9 has the advantage that the two complementary parts of the broken cells can be investigated separately. This method is very powerful in combination with immuno-histochemical techniques using antibody or protein A labeled gold particles.

References
1. Wolosewick, J. J. and Porter, K. R. (1979) Microtrabecular lattice of the cytoplasmic ground substance. Artifact or reality. J. Cell Bid. 82,114-139. 2. Pawley, J. and Ris, H. (1987) Structure of the cytoplasmic filament system in freezedried whole mounts viewed by HVEM. J. Microsc. 145 (Pt 3), 319-332. 3. Schliwa, M. (1986) Whole-mount preparations for the study of the cytoskeleton in Electron Microscopy 2986 Vol. 3 (Imura, T., Marusche, S., and Susuki, J., eds.), J. Electron Microsc. 35, Suppl. pp. 1905-1908. Japanese Sot. Electr. Microsc. Tokyo, Japan. 4. Porter, K. (1986) Section VII: High Voltage Electron Microscopy. J. Electron Microsc. Techn. 4,142-145. 5. Buckley, I. K. (1975) Three dimensional fine structure of cultured cells: possible implications for subcellular motility. Tissue 6 Cell 7,51-72. 6. Collins, V. I., Fredriksson, B. A., and Brunk, U. T. (1981) Changes associated with the growth stimulation of in vitro cultivated spheroids of human glioma cells. Scan. Electr. Microsc. 1981/Q 187-196. 7, Miller, G. J. and Jones, A. S. (1987) A simple method for the preparation of selected tissue culture cells for transmission electron microscopy. 1. Electron Microsc. Techn. 5,385,386. 8. Arnold, J. R. and Boor, I. J. (1986) Improved transmission electron microscopy (TEM) of cultured cells through a floating sheet method. J. Uftrusfruct. Molec.
Sfruct. Res. 94,30-36.

Hyatt, A. D., Eaton, B. T., and Lunt, R. (1987) The grid-cell-culture technique: the direct examination of virus-infected cells and progeny viruses. J Microsc. 145 (I? l), 97-106. 10. Cajander, S. B. (1986) A rapid and simple technique for correlating light microscopy, transmission and scanning electron microscopy of fixed tissues in Epon blocks. J. Microsc. 143 (Pt 3),265-274. 22. Edwards, H. H., Mueller, T. J., and Morrison, M. (1986) Monolayer freeze-fracturea modified procedure. J. Electron Microsc. Techn. 3,439-451. 12. Arro, E., Collins, V. I., and Brunk, U. T. (1981) High resolution SEM of cultured cells: preparation procedures. Scan. Electr. Microsc. 198l/II, 159-168.
9.

Transmission

and Scanning Electron Microscope

471

23. Gelderblom, H. R., Kocks, C., LAge-Stehr, J.,and Reupke, H. (1985) Comparative immunoelectron microscopy with monoclonal antibodies on yellow fever virusinfected cells: pre-embedding labeling versus immunocryoultramicrotomy. J. Viral. Meth. 10,225-239. 24. Hodges, G. M., Southgate, J., and Toulson, E. C. (1987) Colloidal gold-a powerful tool in scanning electron microscope immunocytochemistry: an overview of bioapplications. Scanning Microscopy 1,301318. 15. Goode, D. andMauge1, T. K. (1987) Backscatteredelectron imaging of immunogoldlabeled and silver-enhanced microtubules in cultured mammalian cells. 1.Elecfron
Microsc. Techn. 5,263-273.

16. ,Handley, D. A. (19851Ultrastructural studies of endothelial and platelet receptor binding of thrombin-colloidal gold probes. Europ. J. Cell Bid. 39,391398. 17. Handley, D. A. (1987)Receptor-mediated binding, endocytosisand cellular processing of macromolecules conjugated with colloidal gold. Scanning Microscopy 1,359367. 18. Bohn, W., Mannweiler, K., Hohenberg, H., and Rutter, G. (1987) Replica-immunogold technique applied to studies on measles virus morphogenesis. ScanningMicroscopy1,319-330. 29. Paatero, G. I. L., Miettinen, H., Klingstedt, G., and Isomaa, B. (1987) Scanning electron microscopic detection of colloidal gold labelled surface immunoglobulin on mouse splenic lymphocytes following treatment with the amphiphilic agent CTAB.
Cell. Molec. Biol. 33,13-20.

20. Handley, A. D., Arbeeny, C. M., and Witte,L. D. (1985)Intralysosomal accumulation of colloidal gold-low density lipoprotein conjugates in chloroquine-treated fibroblasts, in Proceedings of fhe 43rd Annual Meeting of the Electron Microscopy Society of Amerzku (Bailey, G. W., ed.), San Francisco Press, San Francisco, pp. 546,547. 21. de Harven, E.,Soligo, D., and Christensen, H. (1987) Should we be counting immunogold marker particles on cell surfaces with the SEM?J.Microsc. 146 (Pt 2),183-189. 22. Silver, M. M. and Hearn, S.A. (1987) Postembedding immunoelectron microscopy using protein A-gold. Ulfrastrucf. P&d. 11,693-703. 23. Linner, J.G., Livesey,S. A., Harrison, D. S.,andsteiner, A. L. (1986) A new technique for removal of amorphous phase tissue water without ice crystal damage: a preparative method for ultrastructural analysis and immunoelectron microscopy. J. Hisfothem. Cytochem 34,1123-1135. 24. Bearer, E. L. and Orci, L. (1986) A simple method for quick-freezing. 1. Electron
Microsc. Techn. 3,233-241.

25. Lawson, D. (1986) Myosin distribution and actin organization in different areas of antibody-labeled quick-frozen fibroblasts. J. Cell Sci. Suppl. 5,45-54.

Chapter 35
Double IndirectImmunofluorescent Labeling of Cultured
Christine A. Boocock

Cells

1. Introduction
Immunofluorescence is a powerful technique for identifying and localizing intra- and extra-cellular components both in histological sections and in cultured cells of plant or animal origin. Briefly, an antibody, raised against a specific component, is used as a label to map the distribution of the component in the specimen, and then visualized under the light microscope using a fluorescent dye (a fluorochrome) such as rhodamine or fluorescein. These dyes are excited to fluoresce by microscope illumination of the appropriate wavelength. By using fluorochromes that differ both in the wavelength required for excitation and in the color of light emitted, several components can be mapped within the same specimen. In this way it is possible, for example, to distinguish cell types within tissues, to identify components involved in cell motility, adhesion and cell-cell recognition or, using monoclonal antibodies, to detect small variations in antigen structure. Specimens must first be fixed to preserve structure and to immobilize components that would otherwise be crosslinked and aggregated by the antibodies used to label them. Cell mem473

474

Boocock

branes are then permeabilized by detergent extraction. This allows antibodies to reach and label components within the cell. These procedures can be carried out with minimal damage to cellular structure, so that the distribution of an antigen labeled by immunofluorescence can be related to information given by other optical techniques, such as phase contrast and interference reflection microscopy (2,2). Furthermore, by fixing a series of replicate cultures at intervals during an experiment, active cellular processes can be observed, the components involved being recognized by their molecular structure (1,3). Some of the different methods available for immunofluorescent labeling are briefly compared below. In direct immunofluorescence, a fluorochrome is linked to the primary antibody. This fluorescently conjugated antibody binds directly to antigenic sites in the fixed and permeabilized specimen. Direct immunofluorescence is the quickest, simplest method of immuno-labeling, but gives rather weak specific labeling and a relatively high background from nonspecific binding. In indirect immunofluorescence, the fluorochrome is linked instead to a secondary antibody directed against the primary antibody. Several secondary antibody molecules bind to each primary antibody molecule, thus enhancing the brightness of specific labeling and reducing the relative contribution of background labeling. To distinguish several components in the same specimen, primary antibodies must be raised in separate species, and secondary antibodies must be strictly species-specific and conjugated to different fluorochromes whose excitation and emission spectra overlap to a minimal extent (seeTable 1). Protein A (4,5) is a cell wall protein (mol wt 42,000) of Stu@zyZococc~s uureus and binds with high affinity the Fc regions (seeNote 8) of immunoglobulin (Ig) molecules, especially of the class IgG. Fluorescently conjugated protein A can be used instead of a secondary antibody to cut down background labeling, and is especially useful where background labeling results from the binding of secondary antibodies to cell surface Fc receptors (found on granulocytes, B-lymphocytes, and macrophages) in the specimen. However, only about twofold amplification is obtained in this way, compared with the 7- to 8-fold amplification obtained using a secondary antibody. Moreover, the usefulness of protein A is limited by its lower affinity for immunoglobulins of some species, particularly sheep, goat, and rat, and for some subclasses of IgG in other species (4,6). Biotin-avidin and biotin-streptavidin systems (5,7&biotin, a small vitamin (mol wt 224) that can be linked to antibodies with minimal effect

Immunofluorescent

Labeling

475

Table 1 Excitation and Emission Maxima of Common Fluorochromes Emiss. Color of Abs. max. max. observed Fluorochrome fluorescence (nm) (nm) Texas red 596 615 Deep red Lissamine-rhodamine-B 570 590 Red Rhodamine isothiocyanate (RITC) 554 573 Red Fluorescein isothiocyanate (FITC) 492 515 Green Aminomethyl coumarin acetic acid 350 450 Blue on their biological activity. Egg-white avidin and streptavidin (from cultures of Sfrepfomyces avidinii) are tetrameric proteins, both able to bind four molecules of biotin. Streptavidin (mol wt 60,000) binds biotin with an affinity 6-10 orders of magnitude greater than that of an antigen-antibody interaction, and with less nonspecific binding than avidin. Using a biotinylated primary (or secondary) antibody and a fluorescently labeled streptavidin, three (or four) amplification steps are achieved: each antigen molecule binds several antibodies; each antibody has several biotins attached, each of which can bind one streptavidin molecule; and each streptavidin can have several fluorochromes attached. Alternatively, a bridging streptavidin (7) can be used to achieve a further threefold amplification: when bound by a biotinylated primary (or secondary) antibody, the bridging steptavidin retains three free sites able to bind fluorochromeconjugated biotin molecules. Whichever of these methods is used, the quality of immunofluorescent labeling depends ultimately on that of the antibodies. Polyclonal antisera generally give a high density of labeling because they contain different antibodies able to bind several sites (determinants) on each antigen molecule. Affinity purification removes most of those that bind to molecules other than the intended target. Monoclonal antibodies have the advantage of specificity for a single antigenic determinant and therefore generally cross-react less with other molecules. Another critical factor is the affinity of the antibody for its antigen. This varies greatly according to the species and method of preparationof specimen material, so that quantitative comparisons can be misleading. Optimal conditions and antibody dilutions for immunofluorescence must be found, with a certain amount of trial and error, for each new specimen and antibody, and those given here are intended as a guide, not as a set of hard-and-fast rules.

476

Boocock

2. Materials
1. Cells cultured on coverslips. Quantities used in this procedure are suitable for 15-mm diameter coverslips (circular coverslips are easiest to handle, and most microscope objectives are optically corrected for coverslip thickness no. 1.5). 2. A coverslip rack is useful to ensure equivalent processing of numerous replicate cultures. 3. PBS (phosphate-buffered saline): 0.14M NaCl, 2.7 mM KCl, 1.5 mM KH,PO,, 8.1 mM Na,HPO,. 4. PBSB: PBS containing 0.05% bovine serum albumin (BSA). Protein is added to all antibody solutions to compete with nonspecifically binding antibodies. Fetal calf serum (10% w/v) may be used instead of BSA. 5. Fixation buffer: 0.25% (v/v> glutaraldehyde in PBS. 6. Extraction buffer: 0.1% (v/v) Triton X-100 in PBS. 7. Sodium azide: 0.01% in PBS. (EXTREMELY TOXIC. Handle with great care.) 8. Sodium borohydride: 0.05% in distilled water. (TOXIC and liberates hydrogen on contact with water.) Store sodium borohydride dessicated, open as briefly and infrequently as possible, and only dilute immediately before use. Add a small spatula-full to 15 mL of ice-cold water and shake in a sealed container. 9. Clean microscope slides. 10. A water-miscible mountant, such as Uvinert- (Gurr) or Citifluor (Goodwin & Davidson, Department of Chemistry, The City University, London). These contain anti-fadeingredients that retard the fading of fluorochromes in light of their excitation wavelength. 11. Primary antibodies (either antisera or monoclonal antibodies purified from ascites fluid) raised against each antigen to be labeled. Try to obtain antibodies raised against material of the same species as your specimen. Antibodies raised against an avian protein may, for example, have very weak affinity for the mammalian equivalent. Use of affinity-purified antisera can eliminate labeling of cross-reacting antigenic determinants. However, even monoclonal antibodies can cross-react with similar determinants on completely different molecules. Crossreactivity of antibodies can be checked on immunoblots of specimen material. 12. Preimmune sera obtained from the same host species (preferably from the same animal) as each orimarv antibodv. On control specimens, to

Immunofluorescent

Labeling

477

test for nonspecifically binding antibodies, these are substituted for the corresponding primary antibodies. 13. Fluorescently conjugated secondary antibodies raised against immunoglobulins of the same species as each primary antibody. If more than one component is to be labeled, secondary antibodies must be sufficiently species-specific to distinguish totally between primary antibodies. To test for cross-reactivity between secondary antibodies, each should be used, on separate control specimens, in combination with each primary antibody. The use of secondary antibodies preadsorbed against sera of other species can obviate the need for this control. Secondary antibodies must be conjugated with fluorochromes separated as far as possible inexcitation and emission spectra (seeTable 1). In combination with fluorescein, Texas red is better than rhodamine in this respect. The blue dye coumarin, recently applied in double-label immunofluorescence, additionally eliminates the excitation of one fluorochrome by fluorescence emitted by the other (8).

3. Methods
3.1. Double Indirect-ImmunofZuorescent of Cultured Labeling

Cells (See Note 1)

3.1.1. Dilution of Antibodies 1. Antisera vary greatly in antibody titer and the ideal dilution to use varies also with antibody affinity, specimen thickness, and so on. In a preliminary trial, test on replicate specimens a few dilutions between, 10 and 100 pg/mL. Store all antibodies frozen or lyophilized, and dilute only just before use. Solutions for double-labeling of tubulin and vimentin are described in Table 2. These are given merely as examples and are not intended for use as general recipes. 2. Centrifuge diluted antibodies 30 min at 10,000 x g, to pellet any precipitates. Keep antibody solutions on ice, and take care not to disturb the pellet. 3.1.2. Fixation and Permeabilization of Cells Volumes of solutions are not critical. Use enough to immerse coverslips and avoid their drying out. About 1 mL/coverslip should suffice (see Note 2). 1. Rinse coverslips briefly in PBS, then pre-fix 5 min at 37*C in fixation buffer.

478 Table 2

Boocock

Suggested Dilutions of Antisera for Double Indirect Immunofluorescent

Labeling of Tubulin and Vimentin in Cultured Cells Primarv antibodies Rabbit antitubulin antiserum Goat antivimentin antiserum PBSB Preimmune sera Rabbit preimmune serum Goat preimmune serum PBSB Secondary antibodies FITC-mouse IgG antirabbit IgG RITC-rat IgG antigoat IgG
PBSB

Volume
10 pl 5 PI 235 ~1 10 kl 5 PI 235 J.L~
6

Final dilution l/25 l/50

l/25 l/50

11

4 PI 230 ~1

l/40 l/60

2. Rinse briefly in PBS, and then permeabilize for 15 min in extraction buffer. Steps 2-4 may be performed at room temperature. 3. Rinse briefly in PBS, and then fix 5 min in fixation buffer.

4. Dip briefly in PBS,and then rinse in three changes of PBSover 30 min. 5. Reduce any remaining aldehyde as follows: Immerse coverslips in FRESHLY MADE UP, ice-cold sodium borohydride solution. Incubate 5 min on ice, shaking dishes continuously. Bubbles should form if the solution is fresh. Drain dishes, add fresh borohydride solution, and repeat. 6. Rinse as in Step 4. Continue rinsing until autofluorescence (resulting from residual aldehyde) is completely quenched (check with a fluorescence microscope). 3.1.3. Antibody Labeling 1. Drain coverslips thoroughly and place, close together but not touching, in lidded Petri dishes lined with moist filter paper (filter paper may be omitted if a humidified incubator is to be used for incubations). 2. Add to each coverslip 60 PL of primary antibody solution (preimmune sera to controls) and incubate 30 min at 37OC(or 1 h at room temperature). 3. Drain excess antibody solution from coverslips (touch edge to filter paper), dip briefly in PBS, and then rinse 45 min to 1 h, using at least three changes of PBSand agitating occasionally. This rinsing must be

Immunofluorescent

Labeling

479

4. 5.

6. 7.

very thorough. Because IgG antibodies are divalent, secondary antibodies can cross-link any primary antibodies remaining in solution, forming immunoprecipitates, which will appear as spurious fluorescent structures. Repeat Steps 1-3, using secondary antibodies. Continue rinsing in distilled water, checking occasionally under the microscope for removal of background fluorescence. Use 0.01% sodium azide for the final rinses, to prevent bacterial growth after mounting. Invert each coverslip onto a small drop of water-miscible mountant on a clean glass slide. Observe, using a microscope equipped with epifluorescence illumination and fluor type objectives (seeNotes 3 and4). Onlyonefluorochrome may be visualized at a time, although photomicrographs of structures labeled with different fluorochromes may be superimposed on color film (seeNote 5). The corresponding beam-splitter/ interference-filter combination is inserted (and must be correctly oriented) into the microscope body, just behind the objective. This transmits light of the excitation wavelength to the specimen and light of the emitted wavelength to the microscope eyepiece or camera.

4. Notes
1. The procedure given is for cultured cells, but very similar methods are applicable to sections (l-10 pm thick) of frozen or wax-embedded tissue (8,9). The use of frozen-sectioned material avoids some of the loss of antigenicity incurred by fixation. 2. Rigorous cleanliness is important to avoid contamination with crossreacting antigens. Use acid-washed glass culture substrata (plastics are easily scratched, trapping reagents, and may themselves be fluorescent), separate, disposable dishes for all incubations and rinses, and separate, clean instruments to avoid cross-contamination. Wear gloves at all times. The amplification inherent inmost methods makes early contamination more serious, but dust or air-bubbles included at any stage, and in the optical system, can cause reflections with ruinous effect on image contrast. 3. If labeling is very faint, first check that the fault is not in the optical system. Second, try repeating the method using a higher concentration of the primary antibody. Third, raise the concentration of the secondary antibody. If all else fails, try a third incubation with a tertiary

480

Boocock

antibody directed against the secondary antibody and conjugated with the same fluorochrome. The resulting chain of three antibodies should be stabilized, before mounting, by post-fixation: Swirl for 15 s in 5% acetic acid/80% ethanol, and then briefly in (large volumes of) 70% ethanol, 50% ethanol and water. 4. High background labeling may be the result of insufficient rinsing, specimen contamination, cross-reactivity or nonspecific binding of antibodies, or antibody binding to cell surface Fc receptors. Figure 1 shows the effect of these factors on the fluorescence image. Unwanted labeling of Fc receptors can be diminished by using a biotin-avidin system, or by replacing the secondary antibody with fluorescently labeled protein A. Alternatively, all antibodies can be replaced with corresponding Fab or F(ab), fragments (seeNote 8). However, some of the antigenicity of the primary antibody resides in the Fc region, and fluorescently conjugated secondary antibodies (or their Fab fragments) raised against Fab or F(ab), fragments are rarely commmercially available. 5. Fluorochromes, especially fluorescein, gradually fade when exposed to light of the excitation wavelength, or even to daylight or bright artifical light. Keep specimens in the dark, e.g., foil-wrapped, and avoid illuminating for longer than is necessary in order to focus before taking micrographs. Successive exposures of the same area will need increasing exposure times. If using several fluorochromes, dont panic: illumination of the excitation wavelength for one fluorochrome should not cause fading of another. Fading can be retarded by antifade mountants although these tend to have high refractive indices, marring the image obtained with phase-contrast optics. Some watermiscible mountants also allow bacterial growth, which may in time destroy the specimen. Avoid this tragedy by using 0.01% sodium azide for the final rinses before mounting. 6. Poor fixation can impair both the antigenicity of cellular components and the structural integrity of the specimen, and several buffering systems have been developed specifically to preserve cytoskeletal structure during fixation (10). Glutaraldehyde is a good fixative for preserving cellular structure. Alternatives are paraformaldehyde (2,3), ice-cold absolute methanol and/or acetone (3), and ethylene glycolyl bis (succinamidyl succinate) (EGS). EGS preserves antigenicity better than either glutaraldehyde or paraformaldehyde, because it forms more widely spaced cross-links in the specimen material. For this

Immunofluorescent

Labeling

481

reason, fixed specimens remain rather fragile. Dissolve EGS initially in dimethyl sulfoxide and dilute to 5-10 mM in warm PBS immediately before use: it tends to precipitate once diluted if allowed to cool below 37OC. Fixation with methanol or acetone obviate the need for a separate permeabilizing step. Another alternative is to include detergent in the fixation buffer, although this accelerates fixation, making timing and temperature more critical. Try 0.05% Triton X-100, 0.5% glutaraldehyde in the buffer described by Small (10). Fix for 4 min at 37OC. 7. Permeabilization extracts a significant amount of cell-surface protein, indirectly disrupting cellular structure. Cell-surface determinants, such as growth factor, hormone, or neurotransmitter receptors and cell adhesion molecules are best labeled on cells that have not been permeabilized. Using permeabilized and unpermeabilized replicates, they can be distinguished from receptors that the cell has internalized. 8. Digestion of the immunoglobulin molecule with papain gives rise to fragments of two types: Fc (fragment crystallizable), which contains domains able to bind staphylococcal protein A and cell-surface Fc receptors, but has no antigen-binding specificity; and Fab (fragment antigen-binding), which is a monovalent antigen-binding fragment, consisting of just one of the two antigen-binding arms of the immunoglobulin molecule. F(ab), is a divalent antigen-binding fragment prepared by digestion of immunoglobulin with pepsin.

4.1. Variations

of the Technique.

Immunofluorescence has many applications beyond the mapping of a few antigenic determinants within fixed cells and tissues. Modified methods can be applied to living cells (12,12) or permeabilized but unfixed cell models (23), giving a more dynamic picture of cellular processes. Antibodies can also be used as tools to manipulate cellular structures, e.g., to cross-link cell-surface receptors (14) or to immunoprecipitate intracellular elements (25), and thus investigate the connections and functional relationships between cellular components. Unless (monovalent) Fab fragments are used, the (divalent) primary antibody cross-links cell surface components to form a patch. The movement of patched determinants in the cell surface membrane can be followed by subsequent fixation and labeling with a secondary antibody (14) or by image-intensified video microscopy (H), and is of interest in understanding the role of the cell membrane in cell motility.

482

Boocock

Fig. 1. (a) Chick heart fibroblast labeled with antibodies to tubulin. Cells were fixed 4 min in 0.05% Triton X-100,0.5% glutaraldehyde in the buffer described by Small (1981) (10). (b) Chick heart fibroblast badly fixed with cold methanol and labeled with antibodies to tubulin. Poor fixation is indicated by disintegration of microtubules, and distortion of cellular structure was apparent by phase-contrast microscopy. (cl Chick heart fibroblast labeled with antibodies to tubulin. Incomplete removal of unbound primary antibody has allowed the formation of fluorescent immunoprecipitates, obscuring detail of the cell. Note: Incomplete removal of unbound secondary antibody would instead have caused overall brightness of cells and background. (d) Chick heart fibroblast labeled with antibodies to vinculin, a component of cell-substratum adhesions. Not only the marginal adhesions of this cell are labeled but also its nucleus, indicating cross-reactivity with components other than vinculin. Indeed, this antiserum cross-reacted with several nuclear

Immunofluorescent

Labeling

483

The links between cytoskeletal components and the ventral membrane of the cell are particularly important in the study of cell adhesion and motility. In immunofluorescently labeled preparations, detail is often obscured by out-of-focus images from other planes within the cell. This problem has been overcome first by studies of isolated ventral membranes of adhering cells (16), and second by an ingenious optical technique giving a three-dimensional view of intact, fluorescently labeled cells (17). Further variations of the technique have been used to throw light on the mechanism of cell adhesion in motile and phagocytic processes. Cells are plated onto surfaces derivatized with a carpet of a primary antibody (E?), or of a protein that can later be immunofluorescently labeled (19). The resulting pattern of dark patches reveals the regions where tight adhesion of the cell restricts access of the secondary antibody, or where the cell has removed the protein carpet. The spatial relationships between cytoskeletal and extracellular components have been investigated in ever-increasing detail by the more recently developed technique of immuno-electron microscopy (20,21 and Chapter 38, this vol). This differs from immunofluorescence in that electron-dense markers such as colloidal gold or ferritin replace fluorochromes, allowing cellular components to be mapped at the much higher resolution of the electron microscope (21). Because of the different artifacts produced by preparation for different types of microscopy, immunofluorescence and immuno-electron microscopy are especially valuable in combination (20). Other techniques closely related to immunofluorescence include the labeling of cellular components with fluorescent cytochemical markers, such as the phallotoxins (3,22) and heavy meromyosin (14), or their substiproteins on immunoblots of chick heart material. (e) Mouse macrophagelabeled with antibodies to the membraneprotein, talin, associatedwith cell-substratum adhesions. The antiserum was raised againstchick gizzard talin and showedlow affinity for mammalian talin. Cell-substratumadhesionsarenot notice-ablylabeled. Instead, the whole
cell is labeled by antibody binding to cell surface Fc receptors. Note especially bright labeling of membranous folds-ruffles-at both ends of the cell. (0 Mouse macrophage labeled with antibodies to the colony-stimulating factor, CSF-1. Although Fab fragments of the primary antibody were used to avoid labeling cell-surface Fc receptors, specific labeling of cell-associated CSF-1is obscured by a general cytoplasmic fluorescence because of incomplete quenching of residual glutaraldehyde used in fixation. Exclusion of this fluorescence from intra-cellular vesicles indicates that, unlike that in (e), it is not confined to the cell surface. All scale bars 10 pm.

484

Boocock

tution with fluorescent analogues, which may be incorporated into living cells directly (12) or by microinjection (23,24). These techniques again are most useful in combination with immunofluorescence. The high contrast of the fluorescence image makes it suitable for video image analysis, and the clear distinction between labeled and unlabeled cells is useful in systems for cell-counting (flow cytometry) and fluorescence-activated cellsorting (FACS) (seeChapters 40 and 41, this vol.).

Acknowledgment
This work was supported by the Cancer Research Campaign.

References
1. Schliwa, M., Nakamura,
2.

3.

4. 5. 6.

7.

8.

9. 1oa

T., Porter, K. R., and Euteneuer, U. (1984) A tumor-promoter induces rapid and coordinated reorganization of actin and vinculin in cultured cells. 1. Cell Biol. 99,1045-1059. Wehland, J., Osborn, M., and Weber, K. (1979) Cell to substratum contacts in living cells: A direct correlation between interference reflexion and indirect immunofluorescence microscopy using antibodies against actin and alpha actinin. I. Cell Sci. 37, 257-273. Schlessinger,J.andGeiger,B. (1981)Epidermal growthfactor inducesredistribution of actin and alpha-actinin human epidermal carcinoma cells. Exp. Cell Res. 134,273279. Goding, J. W. (1978) Use of staphylococcal protein A as an immunological reagent. J. lmmunol. Mefh. 20,241-253. Hsu, S. M. and Raine, L. (1981) Protein A, avidin and biotin in immunohistochemistry. J Hisfochem. Cyfochem.29,1349-1353. Lindmark, R., Thoren-Tolling, K., and Sjoquist, J. (1983) Binding of immunoglobulins to protein A and immunoglobulin levels in mammalian sera. J. immunol. Mefh. 62,1-13. Bonnard, C., Papermaster, D. S., and Kraehenbuhl, J. -I. (1984) The streptavidin-biotin bridge technique: Application in light and electron microscopic immunocytochemistry, idmmunolabellingfov Electron Microscopy (Polak, J. M. and Varndell, I. M., eds.), Elsevier, Amsterdam, pp. 95-111. Khalfan, H., Abuknesha, R., Rand-Weaver, M., Price, R. and Robinson, D. (1986) Aminomethyl coumarin acetic acid: A new fluorescent labellingagent for proteins. Hisfochem. J l&497-499. Hayman, E. G., Pierschbacher, M. D., Ohgren, Y., and Ruoslahti, E. (1983) Serum spreading factor (Vitronectin) is present at the cell surface and in tissues. Proc. Nut. Acad. Sci. USA SO, 4003-4007. Small, J, V. (1981) Organisation of actin in the leading edge of cultured cells: Influence of osmium tetroxide and dehydration on the ultrastructure of actin meshworks. 1, Cell Biol. 91,695-705.

Immunofluorescent
12. 12.

Labeling

485

13.

14.

15.

26.

17.

18. 19.

20. 21.

Heath, J. P. (1983) Direct evidence for microfilament-mediated capping of surface receptors on crawling fibroblasts. Nature 302, 532-534. Spiegel, S., Schlessinger, J., and Fishman, P. H. (1984) Incorporation of fluorescent gangliosides into human fibroblasts: Mobility, fate, and interaction with fibronectin. J. CeZI Biol. 99,699-704. Weber, K., Rathke, I. C., Osbom, M., and Franke, W. W. (1976) Distribution of actin and tubulin in cells and in glycerinated cell models after treatment with cytochalasin B. Exp. Cell Res. 102,285-297. Ash, J. F., Louvard, D., and Singer, S. J. (1977) Antibody-induced linkages of plasma membrane proteins to intracellular actomyosin-containing filaments in cultured fibroblasts. Proc. Natl. Acad. Sci. USA 74,5584-5588. Mangeat, P. H. and Burridge, K. (1984) Immunoprecipitation of nonerythrocyte spectrin within live cells following microinjection of specific antibodies: Relation to cytoskeletal structures. J. Cell Bid. 98,1363-1377. Avnur, Z. and Geiger, B. (1981) Substrate-attached membranes of cultured cells: Isolation and characterization of ventral membranes and the associated cytoskeleton. J Mol. Biol. 133,361-379. Osborn, M., Born, T., Koitzsch, H. J., and Weber, K. (1978) Stereo immunofluorescence microscopy: I. Three dimensional arrangement of microfilaments, microtubules, and tonofilaments. CelI 14,477-488. Wright, S. D. and Silverstein, S. C. (1984) Phagocytosing macrophages exclude proteins from the zones of contact with opsonised targets. Nature 309,359-361. Chen, W.-T., Olden, K., Bernard, B. A., and Chu, F. F. (19841 Expression of transformation-sensitive proteases that degrade fibronectin at cell contact sites. J. Cell Biol. 98,1546-1555. Avnur, Z. and Geiger, B. (1985) Spatial interrelationships between proteoglycans and extracellular matrix proteins in cell cultures. Exp. Cell Res. 158,321-332. Chen, W.-T. and Singer, S. J. (1982) Immuno-electron microscopic studies of the sites of cell-substratum and cell-cell contacts in cultured fibroblasts. J. Ceil BioZ. 95,205-

222.
22. Wulf, E., Deboben, A., Bautz, F. A., Faulstich, H., and Wieland, T. (1979) Fluorescent phallotoxin, a tool for the visualisation of cellular actin. Proc. Nutl. Ad. Sci. USA 76,

4498-4502.
23. 24. Burridge, K. and Feramisco, J. R. (19801 Microinjection and localization of a 130K protein in living fibroblasts: A relationship to actin and fibronectin. Cell 19,587-595. Scherson, T., Kreis, T. E., Schlessinger, J., Littauer, U. Z., Borisy, C. G., and Geiger, B. (1984) Dynamic interactions of fluorescently labelled microtubule-associated proteins in living cells. J. Cell Biol. 99,425-434.

Chapter 36 Gene Mapping to Chromosomes by Hybridization In Situ

John R. Gosden

1. Introduction
Since the technique of hybridizing labeled nucleic acid sequences to cytological preparations (hybridization in situ) was first described in 1969 by Pardue and Gall (I), it has undergone considerable refinement. In those early days, the technique was only capable of detecting and locating highly repeated target sequences such as satellite DNAs (2) or the genes for ribosomal DNA (3). The introduction of methods that combined positive identification of every chromosome in each cell with a quantitative analysis of the distribution of autoradiographic grains (4) opened the way to more sensitive uses of the technique and ultimately to the present situation, in which hybridization in situ is routinely used to map the precise chromosomal location of DNA sequences as little as 1000 base pairs (one kilobase, 1 kb) in length. The technique described here can be broken down into six stages: 1. Chromosome preparation. Chromosomes are prepared from cell cultures, which may be of permanent cell lines (fibroblasts, lymphoblastoid cells, and so on), or primary cultures. The cell type most frequently 487

488

Gosden used for mapping human DNA sequences is the phytohemagglutinin (PHA) stimulated peripheral blood lymphocyte. The chromosome preparations are spread onto clean glass slides. Since it is essential to be able to identify each chromosome, some form of banding procedure is needed. The method I use, which produces bands on the chromosomes after hybridization in situ, requires specific treatment during culture and preparation. The method used must produce a high mitotic index with well-spread metaphase plates and chromosomes of adequate length for good band resolution. It is obviously important that the method minimizes any loss of DNA from the chromosomes, while ensuring that the DNA is readily accessible to the probe. Labeling the probe. The label of choice is tritium, because this provides the best combination of sensitivity and resolution. Alternative labels are discussed in the notes. The maximum specific radioactivity that can be obtained without compromising the integrity of the probe is the ideal. The method of labeling may be either nick translation or random oligonucleotide primed synthesis. Hybridization. This comprises two steps: denaturing the chromosomal and probe DNA, and annealing them together. The denaturing steps must minimize loss or degradation of the DNA while maximizing the separation of the DNA duplex, yet still retain chromosome morphology. The annealing conditions are selected to reduce the chance of producing a poorly matched hybrid duplex, yet obtain the greatest possible amount of specific hybridization. In practice, the hybridization conditions are standard, and the posthybridization wash conditions are varied to increase or reduce the stringency of the reaction as required. Autoradiography. This involves coating the slides with sensitive emulsion, exposing them for a suitable length of time, and developing the autoradiograph. Staining-banding chromosomes after hybridization. Analysis-of grain distribution.

2.

3.

4. 5. 6.

2. Materials Chromosome Preparation from Human Peripheral Blood

2.1.
1. Medium: RPM1 1640 (store at 4OC). 2. Phytohemagglutinin (PHA): dissolved in 5 mL sterile H20 (store at 4C).

Gene Mapping

to Chromosomes

489

3. Fetal calf serum (store at -2OOC). 4. 5-Bromo-2-deoxyuridine (BUdR): 10 mg/mL in sterile water, filtered through a sterile 0.2~pm millipore filter (store at 4OC). 5. Tris-EDTA (TE): 1OmM Tris-HCI pH 7.5; 1mM EDTA, (autoclaved). 6. Thymidine: 1O3M in TE, filtered through sterile millipore filter (store at 4OC). 7. Colcemid: 10 pg/mL (store at 4OC). 8. Methanol, absolute. 9. Glacial acetic acid. 10. 0.075M KCl: 0.56 g/100 mL 30.

2.2. Labeling Probe DNA by Random Oligonucleotide Primed Synthesis

1. Deoxy(1,2,2,8-3H) adenosine 5-triphosphate (50-100 mCi/mmol) (dATl?) (store at -2OOC). 2. Deoxy(1,2,5-3H) cytidine 5-triphosphate (50-85 mCi/mmol) (dCTP) (store at -20C). 3. (MethyZ,l,2-3H) thymidine 5-triphosphate (90-130 mCi/mmol) (dTTE) (store at -2OOC). 4. 2-deoxyguanosine 5-triphosphate (dGTP): 0.1 mM in TE (store at -2OOC). 5. Mixed sequence hexadeoxynucleotides: 1 mg/mL in TE (store at -20C). 6. DNA Polymerase 1: large fragment, Klenow enzyme (store at -2OOC). 7. 10 x concentration salt solution: 0.5M Tris-HCl pH 7.8; 50 mM MgCl,; 10 mM B-mercaptoethanol; 500 pg/mL bovine serum albumin (BSA) (store at -2OOC). 8. TNE column buffer: 02MNaCl; O.OlM Tris-HCl pH8.0; 1 .OmM EDTA (store at room temperature (RT). 9. Water-saturated, distilled phenol (store at RT). 10. Chloroform: octan-2-ol(24:l) (store at RT). 11. Sonicated salmon sperm DNA: 10 mg/mL in TE (store at 4OC). 12. 3M ammonium acetate (store at 4OC). 13. Scintillation fluid made as follows: 2(4- t-Butylphenyl)-5-(4-biphenyl)1,3,4-oxadiazole (Butyl-PBD), 8 g, dissolved in 1000 mL of toluene, to which is added 600 mL 2-ethoxyethanol (Scintillation grade). 14. Nick column (Pharmacia).

2.3. Hybridization
1. 20 x SSC: 3M NaCl; 0.3M tri-sodium citrate.

490

Gosden

2. Ribonuclease A (10 mg/mL in 2 x SSC, boiled 5 min to inactivate DNAse) (store at 4OC). 3. Formamide (Analar). 4. Dextran sulfate. 5. 10 x SSCB: 1.5M NaCl; 0.15M tri-sodium citrate; 0.25M Tris-HCl pH 7.4; 0.5mM EDTA (store at 4C). 6. E. coli tRNA: 4 mg/mL in TE (store at 4OC). 7. Rubber solution (e.g., Pang Super Solution, Fritz Hesselbein Chemischefabrik, Nurderstedt).

2.4. Autoradiography
1. Dark room with safelight (Kodak S902filter), water bath at 45C, lighttight cabinet. 2. Ilford L4 Nuclear Emulsion (KodakNTB2 emulsion can besubstituted) (store in dark at 4C). 3. Dipping vessel (flat Pyrex tubing, 30 x 7 mm inside measurements, 80 mm long, sealed and rounded at one end). 4. Light-tight slide containers. 5. Kodak D19 developer. 6. Kodafix.

2.5. Staining-Post-hybridization

Banding

and Analysis

1. Bisbenzimide H33258 fluorochrome (H33258): 500 pg/mL in HO (store in dark at 4OC). 2. Giemsa stain: 2% in Gurrs buffer pH 6.8, diluted fresh on each occasion. 3. 2 x SSC: 0.3M NaCl; 0.03M tri-sodium citrate (store at RT). 4. D.P.X. mountant. 5. Printed chromosome idograms for grain distribution analysis (seeFig. 1).

3. Methods
3.1.

Chromosome Peripheral

Preparation from Human Blood Igmphocytes

All solutions used prior to harvest must be sterile. 1. Medium: To a sterile medical flat or other closed glass vessel containing 200 mL sterile RPMI 1640 medium, add 2mL PHA (dissolved in

Gene Mapping

to Chromosomes

491

2.

3. 4. 5. 6. 7. 8. 9.
10. 11.

12. 13.

sterile distilled water as directed on container) and 30 mL sterile fetal calf serum. For culturing, this is then dispensed with a sterile pipet in 9-mL aliquots into sterile glass vessels of 20-25 mL capacity. Culture: Blood is obtained by venipuncture with a sterile syringe and needle and placed in a vial containing lithium heparin as an anticoagulant. Of this fresh whole blood 1 mL is added to 9 mL of medium, gently mixed, and placed in a water bath or incubator at 37C for 72 h. It will be found convenient to do this at about 3 PM on Friday. After 72 h (3 PM Monday) 200 ~1 of BUdR solution is added, gently mixed, and incubation is continued for a further 18 h (seeNote 1). At 8 AM the following day (Tuesday), transfer the cultures to sterile capped centrifuge tubes and spin them in a bench centrifuge at 1800 rpm for 5 min. Gently decant the supernatant, without disturbing the pellet, and gently resuspend the cultures in RPM1 1640 (without PHA or serum). Repeat the centrifugation and resuspension, and then centrifuge for a third time. On this occasion, the cultures are resuspended in complete medium and transferred to fresh, sterile glass culture vessels. To each, add 100 PL of sterile 1PM thymidine. Incubate the cultures for a further 5 h 15 min at 37C, at which time add 100 PL colcemid solution to each culture and continue incubation for a final 40 min. Transfer the cultures to fresh, sterile capped centrifuge tubes, and spin at 1800 rpm for 5 min. Gently remove the supernatant by pipet, and add 5 mL freshly made hypotonic (0.075M) KC1 solution at RT drop by drop while mixing on a vortex mixer. Leave the cultures at RT for 10 min, and then centrifuge again. Carefully remove the supernatant and add (drop by drop on a vortex mixer as before) 3 mL of a freshly made mix of methanol: glacial acetic acid (3:l). Top the tube up to the 10 mL level with fix, and mix well. Leave in this first fix for up to 30 min, then spin again, decant the supernatant and add a further 10 mL of 3:l fix, mixing well. Centrifuge again, decant the supernatant, and add 10 mL of fix. At this stage, a test slide can be made by dropping one drop of the resuspended chromosome preparation onto a clean glass slide and examining it under phase contrast. If phase optics are not available, an approximation can be made by removing the condenser from a standard microscope. This will let the observer see how many cells are in metaphase, how well the chromosomes are spread, and how much cy-

492

Gosden toplasm is still attached to the chromosome spreads. It is worth staining one slide (2 min in 2% giemsa) to check that nuclear envelopes are removed to permit access of probes to chromosomes. This cannot be readily determined under phase contrast. A further two or three changes of fix will be necessary in most cases. After the final change, resuspend the pellet in 1 mL of fix. Double-frosted glass slides should be used. The slides are cleaned prior to use by soaking in absolute ethanol to which 1% HCl has been added, and polishing with fine muslin. (The same treatment is used for coverslips. These will be termed clean slides and coverslips.) One drop of cell suspension is dropped into the center of a clean slide. The height from which it is dropped, and whether the slide should be warm or cold, wet, dry, or damp (from being breathed on) will vary from day to day according to the local climatic conditions. Each worker will find the best treatment for the local conditions by trial. The slides should be dried at room temperature and then stored in light-tight containers in desiccators under vacuum. Under these conditions, a batch of slides should remain useful for at least 2 mo, and may remain so for much longer.

14. 15.

16.

17.

3.2. Labeling

Probe DNA by Random (See Note 2)

Priming

1. A minimum of 0.3 pg of linear DNA is needed, and the maximum to be labeled at any one time is 1 pg. 2. In a sterile plastic microcentrifuge tube, place 0.5 nmol of each of 3Hlabeled dATP, dCTl?, and dTTP. Dry down by blowing a gentle stream of air into the tube (just sufficient to ruffle the surface of the fluid) from a glass Pasteur pipet connected to an aquarium aeration pump. This procedure takes about 3 h, and may conveniently be done overnight. 3. Mix the probe DNA, 2 PLof hexanucleotide solution and sterile, distilled Hz0 to a total volume of 16 PL in a second sterile 1.5 mL plastic microcentrifuge tube. Spin briefly to collect, pierce a fine hole in the lid with a heated syringe needle, and place in a boiling water bath for 3 min. Remove from the bath, cool to room temperature, and spin again to collect. 4. Transfer the contents of the second tube to that in which the labeled nucleotides have been dried. Add 2 PL of 10 x salt solution and 1 PL of unlabeled dGTl?. Mix well (ensuring that all the dried nucleotides

Gene Mapping

to Chromosomes

493

5.

6.

7.

8.

9.

10.

are redissolved), and add 1 PL of Klenow enzyme. Mix gently and incubate in a water bath at 37OC for 90 min to 2 h. Meanwhile, prepare a Nick column by washing with three reservoirs full of TNE to replace storage buffer with column buffer, and ensure column bed is properly equilibrated. Add 80 PL of TNE and 50 PL of water-saturated phenol to the incubation mix. Mix well and incubate for a further 10 min at 37C. Add 50 ~.LL of chloroform: octan-2-01, mix again, and spin for 2 min to separate the aqueous and phenol phases. With a glass Pasteur pipet, carefully transfer the upper (aqueous) phase to the Nick column being scrupulously careful to avoid picking up any of the lower (phenol) phase. Having allowed the aqueous phase to run into the column bed, add 400 JJL TNE to the column and let this run through. Add a further 400 uL of TNE and collect the drops from the column (total volume should be approximately 400 pL) in a clean, sterile plastic microcentrifuge tube. Take a 4-VL aliquot from this, place in a scintillation vial containing 3 mL scintillation fluid, mix well, and count in a liquid scintillation counter. There should be no less than 50,000 cpm (about 100,000 dpm) in this sample. If there are fewer counts than this, incorporation of label may well be insufficient to detect in the hybridization. If the counts are adequate, add 10 uL of sonicated salmon sperm DNA (to act as coprecipitant) and 50 yL of 3M ammonium acetate to the microcentrifuge tube, mix well, and add 1 mL of absoluteethanol. Mix again and place the tube at -2OOC for at least 1 h. Centrifuge for 5 min, decant the supernatant (into a new tube) and place the sample in a lyophilizer (Freeze-drier) to evaporate the residual alcohol.

3.3. Hybridization

(See Note 3)

Since this takes a long time to dissolve and seems stable at room temperature for several months, it is worth making a large volume such as 10 mL. 2. Treat the slides with ribonuclease A to remove any endogenous RNA that might otherwise compete for the probe. Make a solution of 100 pg/mL RNAse A in 2 x SSC by adding 1% of the stock solution to the appropriate volume of 2 x SSC. For a normal experiment, using six slides, 50 mL final volume in a Coplin jar is sufficient. Place the jar in a water bath at 37C and incubate the slides in it for 1 h. Remove the

1. Make a stock solution of 20% dextran sulfate in formamide.

494

Gosden slides and pass them through 70%, 90%, and 100% ethanol (5 min each). Dry the slides under vacuum. Denature the chromosomal DNA by placing the slides in a solution of 0.6 x SSC; 70% formamide at 70C for 2 min, transferring them instantly to 70% ethanol after this time, and passing through the alcohol series as above, before drying under vacuum. At this stage, provided the slides are kept dry, they are stable for at least several hours, if not days. Meanwhile, make up the hybridization mix as follows: For each slide, allow 20 PL of mix, plus an extra 40 pL, i.e., for six slides (6 x 20) + 40 = 160 pL. Of this total, 20% is the labeled probe dissolved in sterile distilled water, 20% is 10 x SSCB, 10% is E. co2i tRNA solution, and 50% is 20% dextran sulfate in formamide. For the six slides, dissolve the dry, labeled probe in 33 u.L sterile distilled water, leave 5 min to allow complete solution, and take 1 PL and count in a scintillation counter (leaving the 32 PL needed for the hybridization mix). Add 32 PL 10 x SSCB, 16 PL tRNA and 80 PL dextran/formamide. This last is veryviscous: it is best to cut off the end of a micropipet tip to dispense this material, and use the same tip to mix it with the remainder of the hybridization mix by pipeting up and down. Mix very thoroughly and place in a water bath at 70C for 5 min. At the end of this time, transfer to an ice bath. Prepare six 20 x 40 mm clean coverslips. Place 20 PL of hybridization mix on each coverslip, and invert a slide over each. Allow the mix to spread, ensuring that there are no air bubbles; when the area under the coverslip is covered, seal with rubber solution to prevent evaporation of hybridization mix. When the rubber solution is dry, transfer the slides to a thin metal vessel floating in a 37OCwater bath with a lid, or a closed container in a 37OC incubator. Leave overnight. Meanwhile, prepare the washes. In 1000 mL reagent bottles or flasks, make (a) 50% formamide; 1 x SSC and (b) 1 x SSC. Place in a water bath at 42OC. The following morning, carefully peel the rubber solution away from the coverslips, and remove them from the slides. (They should come away easily if the rubber solution provided a good seal.) Place the slides in a glass stain rack, and place this in a 250-mL capacity stain dish in the 42OC water bath. Add 50% formamide; 1 x SSC, agitate gently, and leave for 5 min.

3.

4.

5.

6. 7. 8. 9.

Gene Mapping

to Chromosomes

495

10. Repeat three times, and then give four washes of 5 min each in 1 x SSC. 11. Transfer to 70% ethanol and pass through the series to absolute alcohol as before. Air dry or dry under vacuum.

3.4. Autoradiography
1. In a darkroom under a safelight, heat the water bath to 45OC. In the water bath place: (a) a 25-mL measuring cylinder containing 5 mL distilled water and (b) a lOO-mL beaker full of water, in which the dipping vessel is supported. Place the slides to be dipped face up on a hot plate that is just warm to the touch. (This helps the emulsion to spread thinly and uniformly.) 2. With a pair of sterile plastic forceps, transfer emulsion from the container to the measuring cylinder until the volume in the latter is 10 mL. Allow to stand 5 min to reach melting temperature. Mix very gently, avoiding air bubbles, and pour the diluted emulsion into the dipping vessel. Take two slides, back to back, and dip into the emulsion, so that the frosted area is just clear of the emulsion surface. 3. Remove gently (without scraping the sides of the vessel), separate the slides, and place in a rack to dry. (A standard test-tube rack, with the slides diagonally across the apertures will do.) 4. When all the slides have been dipped, place the rack in a locked, lighttight cabinet, and leave for 2.5-3 h to dry. 5. When dry (still under safelight), transfer the slides to a light-tight box containing a small amount of silica gel separated from the slides by a piece of tissue paper. Seal the box and place at 4C in a refrigerator free from all sources of radiation. 6. After 6 d, bring the slides to room temperature (this takes about an hour) and, in the dark room, under safelight, remove one slide from the box. 7. Develop for 5 min in Kodak D19, agitating once a minute, stop in water, and fix for 5 min in Kodafix, diluted 1:3 with water, again agitating once a minute. 8. Wash under gently running cold water for 6 min. 9. Allow to dry, and stain in 2% Giemsa in Gurrs buffer for 5 min. 10. Air dry, and examine under a microscope. If there are more than 2 or 3 autoradiographic grains/metaphase, develop the remainder of the slides. If not, leave a further 6 or 7 d, by which time, if the experiment has worked, there should be enough grains.

496

Gbsden

3.5.

Staining-Banding

1. After the slides have been developed and washed, allow them to dry thoroughly-at least overnight-to ensure that the emulsion is hard again. 2. Stain slides for 15 min inH33528 diluted to 50 pg/mL in 2 x SSC. Rinse in2xSSC. 3. Mount in 2 x SSC under a 20 x 40 mm coverslip and, in a darkroom, place under UV lamp, set to long wave, at a distance of lo-12 cm, for 30-60 min. 4. Place slides in a Coplin jar containing 2 x SSC for 35-45 min at room temperature. 5. Dry slides, and stain for 15 min in freshly diluted 2% Giemsa. It is probably wise to treat only three or four slides at a time. The chromosomes should show clearly defined G-bands. 3.6.

Analysis

(see Note

4)

1. When completely dry, dip the slides in xylene for 5 min before covering with a 20 x 40 mm coverslip mounted in D.P.X. When they are completely dry, examine under microscope. 2. A suitable form of idiogram is shown in Fig. 1. It is important that the resolution implied by the bands on the idiogram is equivalent to that seen down the microscope! 3. Identify those chromosomes on which grains can be seen, and mark the grains in the correct location on the idiogram. Only grains in contact with a chromosome should be scored. In this way, the total distribution of grains is accumulated, and it is possible to distinguish specific signals from the background. 4. The number of cells needed to obtain a significant result varies from probe to probe. In general, the highest proportion of grains seen on the specific chromosome is about 20%, i.e., only one grain in five is actually found on the correct chromosome, and of these between 60 and 90% may be located at the correct band. It is therefore rarely possible to get a significant result without scoring at least 20 cells, and it may be necessary to score many more-perhaps as many as 80 or 100. It should be possible to resolve the location of a single copy sequence to within a major band, and in most cases to within one or two minor bands.

4. Notes
1. Chromosome preparation: It is important to remember that the BUdR treatment makes the chromosomal DNA sensitive to light, and there-

Gene Mapping

to Chromosomes

497

Fig. 1. Hybridization in situ of ?&ll to human metaphase chromosomes. Mcll is a cloned genomic probe containing 5.3 kb of human DNA. It was labeled with tritium by random priming and hybridized as described in the text. The autoradiographic exposure was for 7d. Forty-nine cells were analyzed, with a total of 100grains on the chromosomes. Of this total, 17 (17%) were on chromosome 6, and of these,7 grains (41%) were on band 6~12.

fore to minimize exposure to light, the slides should be stored in lighttight containers. Storage of the slides in dessicators under vacuum seems to slow down the aging process which eventually reduces their hybridization efficiency. Other methods of banding can be used that do not require the BUdR treatment. These include methods of banding the chromosomes before hybridization, either by Lipsol-Giemsa (5) or Trypsin-Giemsa (6). These have several disadvantages: First, selected cells must be photographed to record the bands, since these are lost during the hybridization process. The photographed cells are relocated after autoradiography, and the labeled chromosomes identified by reference to the photograph. Only a proportion of the selected cells prove suitable for analysis, which means some labor and time has been wasted. Secondly, the staining treatment results in some loss of DNA from the chromosomes. Should the lost sequences include that of interest, the efficiency of hybridization will be reduced. A method of banding after hybridization using Wrights stain on cells synchronized with methotrexate has been described (7), but in my hands, this does not give the

498

Gosden

consistency or quality of bands produced by the method I have described. Some methods for producing elongated chromosomes have depended on the incorporation of ethidium bromide into the DNA. This seems to reduce the efficiency of hybridization, probably because the dye is intercalated into the DNA, and the use of such methods is not encouraged. 2. Labeling the probe: In order to obtain the highest specific activity with H, three labeled nucleotides are used. Since the incorporation of dGTP is less efficient, this base is not normally labeled. Of the two possible methods of labeling DNA, nick translation (8) and random oligonucleotide primed synthesis (9), the latter usually gives higher specific activity, but in some circumstances (seebelow), nick translation may be preferable. Random priming requires the substrate probe to be linearized, by cutting with a restriction endonuclease with a single recognition site. This method gives longer stretches of intact labeled DNA than nick translation. Most experiments with irz situ hybridization are attempts to map the location of a sequence that exists in a single copy at one site in the haploid genome. However, if the probe is derived from a cloned genomic DNA sequence it may, in addition to the single copy sequence of interest, include sequences that are repeated and dispersed throughout the genome. Even if the probe is derived from a cDNA, it may still include sequences that are repeated elsewherein the genome. In either of these cases, it may be difficult or impossible to distinguish the site of interest from the other labeled sites and background. In this case, hybridization of the repeated elements can be reduced or even eliminated by preannealing the probe with an excess of unlabeled sheared total DNA of the same species (20). This process (known as stripping) requires that the fragments of labeled probe containing the repeats can form duplexes with the unlabeled total DNA (and thereby be prevented from taking part in the hybridization reaction) without compromising the unique sequence of interest. For this purpose, the lower mol wt product of nick translation may be preferable. Alternative radioisotopes are %I and %. These have the advantage over 3H that higher specific activities can be obtained using only one labeled nucleotide. However, because of the greater energy of emission from these isotopes, the resolution is poorer and background higher. Nevertheless, some groups have success, particularly with *?I (11). Nonradioactive systems, most of which are based on the use of biotin incorporation into the probe, and its detection by immunologi-

Gene Mapping

to Chromosomes

499

cal methods, hold hope for the future (12). Despite encouraging reports (23), however, they are not yet sufficiently sensitive for routine use. 3. Hybridization: All glass pipets and containers, micropipet tips, microcentrifuge tubes, and solutions must be sterilized by autoclaving or filtration to reduce the risk of contamination with DNase. The number of slides suggested for an average experiment (six) is not an arbitrary figure. It provides one or two slides for flat staining (unbanded) to check the progress of the autoradiographic exposure, plus four slides for banding and full analysis. This should give an adequate number of analyzable metaphase chromosomes to obtain a result (see
above).

4. Analysis: It is possible to carry out analysis of grain distribution by photographing suitably banded and labeled cells, identifying the labeled chromosomes on the photograph, and transferring the accumulated data to an idiogram. This has the advantage of providing a permanent photographic record of the analyzed cells, but the disadvantage that the autoradiographic grains are frequently in a different focal plane from that of the chromosomes. In half-tone photographs, these are not always readily distinguished from stain debris. The method that I use, however, stores the permanent data in the form of the slides themselves, with microscope vernier references to identify the analyzed cells.

Acknowledgments
I would like to thank Professor H. J. Evans for his continued support during the course of the development of the techniques described here, Derek Rout for his excellent technical assistance, and Alison Brown and Derek Rout for their critical reading of the manuscript.

References
1. 2. 3. 4. Pardue, M. L. and Gall, J. G. (1969) Molecular hybridization of radioactive DNA to the DNA of cytological preparations. Pm. Nafl. Acud. Sci. USA 64,6oo-604. Pardue, M. L. and Gall J. G. (1970) Chromosomal location of mouse satellite DNA. Science 168,1356-1358. Henderson, A. S., Warburton, D., and Atwood, K. C. (1972) Location of ribosomal DNA in the human chromosome complement. Pm. N&l. Ad. Sci. USA 69,33943398. Gosden, J. R., Mitchell, A. R., Buckland, R. A., Clayton, R. I., and Evans, H. J. (1975) The location of four human satellite DNAs on human chromosomes. Eqd. Cell Res. 95148-158.

500
5. 6. 7. 8. 9. 10.

Gosden
Gosden, J*R., Middleton, P. G., Rout, D.,andDe Angelis, C. (1986) Chromosomal localization of the human oncogene ERBA2. Cytogenet. Ceil Genet. 43,150-153. Garson, J. A., van den Burghe, L. J. A., and Kernshead, J. T. (1987) High-resolution hybridization technique using biotinylated NMYC oncogene probe reveals periodic structure of HSRs in human neuroblastoma. Cytogenet. Cell Genet. 45,X&15. Harper, M. E. and Saunders, G. M. (1981) Localization of single-copy DNA sequences on G-banded human chromosomes by in situ hybridization. Chromosoma 83, 431-439. Rigby, I. W. J., Diekman, M., Rhodes, C., and Berg, I. (1977) Labelingdeoxyribonucleic acid to high specific activity in vitro by nick translation with DNA polymerase I. J. Mol. Biol. 113,237-251. Feinberg, A. I. andvogelstein, B. (1983) A technique for radiolabeling DNA restriction endonuclease fragments to high specific activity. And. Biochem. 136,613. Porteous, D. J., Bickmore, W., Christie, S., Boyd, I. A., Cranston, G., Fletcher, J. M., Gosden, J. R., Rout, D., Seawright, A., Simola, K. 0. J., Hastie, N. D., and van Heyningen,V. (1987) HEASI-selected chromosome transfer generates markers that colocalize aniridia- and genitourinary dysplasia-associated translocation breakpoints and the Wilms tumor gene within band 11~13. Proc. Nufl. Ad. Sci. USA 84, 5355-5359. Buckle, V. J. and Craig, I. W. (1986) In situ hybridization, in Human Genetic Disease: II Practical Approach (Davies, K. E., ed.), IRL Press, Oxford, pp. 85-101. Gosden, J. R. andPorteous,D. J. (1987) HRASI-selected,chromosomemediated gene transfer; in situ hybridization with combined biotin and tritium label localizes the oncogene and reveals duplications of the human transgenome. Cyfogenet. CeEZ Genet. 45,44-51. Landegent, J. E., Jansen in de Wal, N., van Ommen, G. -J. B., Baas, F., de Vijlder, J. J. M., vanDujin, P., and van der Ploeg, M. (1985) Chromosomallocalizationof a unique gene by nonautoradiographic in situ hybridization. Nature 317,175-177.

22. 12.

23.

Chapter 37 In Situ Hybridization with Radiolabeled cRNA Probes, Using Tissue Sections and Smears
Carolin Adams, Qutayla A. Hamid, and Julia M. Polak
1. Introduction
Immunocytochemistry., together with biochemical procedures, has ensured rapid and accurate identification of tissue components. This has led to a better appreciation of cellular events, particularly storage and secretion of products. However, such techniques have certain drawbacks, in particular the impossibility of monitoring the intracellular processes concerned with protein synthesis. Thus, there was a need for a method that would provide more detailed information about the functional morphology and gene expression in tissues at cellular level. Advances in molecular biology allowed the development of in situ hybridization, a procedure that localizes specific nucleotide sequences (DNA or RNA) in tissue preparations using labeled complementary probes (DNA or RNA). 501

502

Adams, Hamid,

and Polak

Initial in situ studies (1,2,3) involved the visualization of ribosomal DNA, using radiolabeled cRNA probes. Since then, the technique has expanded to include the hybridization of both DNA and cytoplasmic RNA with a variety of nucleic acid probes (4,5). For localization of RNA, there are four types of nucleic acid probes: double-stranded DNAprobes (5), single-stranded cDNA probes (6), cRNA probes (7), and finally, synthetic oligonucleotide probes (5). Doublestranded probes require melting and may reanneal during hybridizations, thus reducing the amount of probe available for reaction with the mRNA. The advantages of cRNA probes include the high thermal stability and affinity of RNA-RNA hybrids, a constant probe size, no vector sequences, and the availability of RNase to remove unhybridized probes (8). In addition, a probe with an identical sequence to the mRNA can be prepared and used as a sense probe. The probes can be labeled with radioactive or nonradioactive tags. The most successful of the nonradioactivemethods uses biotin-substituted nucleotides. However, at present, autoradiography appears to be the most sensitive detection method available for hybridization histochemistry. In the following pages, we will outline the method we follow using radioactive labeled riboprobes and give some examples of in situ hybridization for the investigation of the diffuse neuroendocrine system (9,10,11).

2. Materials
2.1. Tkanscription
1. 5 x Transcription Buffer: 0.2M Tris-HCl, pH 7.5,30 mM Mg Cl, 10 mM Spermidine. 2. 100 mM dithiothreitol (DTT). 3. RNasin (Human Placental Ribonuclease Inhibitor): 25 U/pL. 4. Nucleotide Mixture: 2.5 mM each of ATP, GTP, and UTP. 5. 100 cln/l cytidine triphosphate (CTP). 6. 1 mg/mL linearized plasmid template DNA in water or Tris-EDTA Buffer. 7. Cytidine (~-~~l? triphosphate) 10 mCi/mL. 8. SP6 RNA Polymerase, T7 RNA Polymerase, or T3 RNA Polymerase: 10 U/ pL. These enzymes are very labile and should be out of the -20C deep freeze for a minimal time. 9. DNase (RNase free): 1 pg/pL. 10. t RNA: 10 pg/j,tL.

In Situ Hybridization

503

11. 4MNaCl. 12. Phenol: Melt the solid at 68C, add 8-hydroxy-quinoline (antioxidant) to a final concentration of 0.1%. Extract several times with an equal volume of buffer (1 .OM Tris HCl, pH 8.0, followed by O.lM Tris HCl, pH 8.0 with 0.2% p mercaptoethanol) until the pH of the aqueous phase is 7.6. 13. Chloroform:isoamyl alcohol (24:l). 14. 7M Ammonium Acetate. 15. Absolute Ethanol. 16. 10% TCA (Trichloroacetic acid). 17. Bovine Serum Albumin: 10 pg/pL. Store solutions l-10,15,17at -2OC, and 12 at 4C. All other solutions may be stored at room temperature.

2.2. Fixation
1. O.lM phosphate-buffered saline (PBS): Dissolve 87.9 g NaCl, 2.72 g KH,l?O, and 11.35 g N%Hl?O, (or 23.9 g Na, HPO,* 1230) in 10 L of distilled water. 2. 4% Paraformaldehyde: Dissolve 4 g of paraformaldehyde in 100 mL of hot PBS (50-6OOC) with stirring. Continue stirring until the solution is clear. If necessary, add 1ONNaOH dropwise until the solution clears. Cool, check pH (7.2), and use immediately. 3. 15% Sucrose in phosphate-buffered saline. 4. Polyz-lysine-coated glass slides: Soak slides in detergent overnight. Wash in running tap water for 4-6 h. Rinse in several changes of double-distilled %O. Bake at 250C for 4 h. Coat slides with 0.01% poly+lysine (Sigma, mol wt 300,000) (stored at -20C). Air dry.

2.3. In Situ Hybridization

1. 2. 3. 4.

with cRNA Probes

Phosphate-buffered saline, O.lM Glycine in phosphate-buffered saline. 0.3% Triton X 100 in phosphate-buffered saline. Proteinase K solution: 1 ug/mL, O.lM Tris HCl pH 8.0,500 mM EDTA pH 8.0 (Stock proteinase K: 500 pg/mL, store at -20C). 5. 4% Paraformaldehyde (freshly prepared). 6. Acetylation solution: 0.25% (v/v) Acetic Anhydride, O.lM Triethanolamine pH 8.0. Use immediately. 7. Formamide.

504

Adams, Hamid,

and Polak

8. lOxstandardsodiumcitrate(lOxSSC)stock: 1.5MNaCl,O.l5MTrisodium citrate. Dilute as required. 9. Hybridization solution: 50% deionized formamide, 12.5 x Denhardts, 10% dextran sulfate, 250 mMTris HCl pH 7.5,0.5% sodium pyrophosphate, 0.5% sodium dodecyl sulfate (SDS) and 250 pg/mL denatured salmon sperm DNA. This is prepared freshly from the following stock solutions, which are stored at -2OOC: a. 100% deionized formamide. b. 100x Denhardts: 2% bovine serum albumin, 2% polyvinylpyrrolidone (WI?-360), 2% Ficoll400. c. 5M Tris-HCl, pH 7.5. d. 30x standard sodium citrate (SSC). e. 20% sodium dodecyl sulfate (SDS). f. 50% dextran sulphate. g. Salmon sperm DNA: 20 mg/mL. 10. RNAase A solution: 20 pg/mL in 0.5MNaC1,lO mM Tris-HCl pH 8.0, 1 mM EDTA pH 8.0 (Stock RNase A: 10 mg/mL, store at -2OOC). 11. Dimethyl-dichlorosilane-coated coverslips: Dip coverslips in 5% dimethyl-dichlorosilane in chloroform. Rinse several times with ddT0. Dry. 12. 70,90, and 100% ethanol containing 0.3M ammonium acetate. 13. Autoradiography emulsion: Kodak NTB-2 or Ilford K-5 diluted 1:l with double-distilled HZO. 14. Kodak D19 developer.

3. Methods
3.1. Synthesis
of High

Activity Single-Stranded Probes (lkanscription)

cRNA

The following protocol is a modification of that given by Promega Biotee for synthesis of RNA probes (seeNotes 1-4; Figs. 1 and 2). 1. To a sterile microfuge tube, at room temperature, add in the following order: 4.0 PL 5 x Transcription Buffer, 2.0 PL 100 mM dithiothreitol (DTT), 0.8 PL RNasin, 4.0 PL Nucleotide Mixture, 2.4 PL 100 pm cytidine triphosphate (CTP), 1.0 PL linearized plasmid template DNA (1 pg), 5.0 PL w~~I?-CTP (50 PCi), 0.5-0.8 PL SP6 RNA polymerase, T7 RNA Polymerase or T3 RNA Polymerase, and DEPC-treated water to 20 PL final vol.

In Situ Hybridization Hind =,Pst I (568)* Hint 11(259)

-rr \

Y
t-

Eco RI Pvu II

*NUMBERING REFERS TO RAT ANP cDNA SEQUENCE

Fig. 1. Schematic diagram of the SP6 plasmid (ANP-cDNA) used for in vitro transcription.

2. Incubate for l-1-1/2 h at 37-4OC. 3. To terminate transcription, add 1 PL of RNase-free DNase and 1 FL of RNasin. Incubate for 10 min at 37OC. 4. Add: 1 PL of tRNA, 175 PL of DEPC-treated water, and 5 JJL of 4M NaCl. Extract with an equal vol(200 pL> of phenol/chloroform (1:l v/ v). Mixby vortexing. Separate the phases by centrifugationin amicrofuge (5 min). 5. Remove the upper aqueous phase (200 I.~.L> and extract this with an equal vol(200 pL) of chloroform. Mix and spin as above. 6. To the upper aqueous layer add 100 PL of 7M ammonium acetate (2.5M final concentration), 750 l.tL of absolute ethanol (2.5 vol, -2OOC). Mix and leave at -2OOC overnight. 7. Spin in a microfuge for 30 min. Discard the supernatant. Dry the RNA pellet under vacuum. When dry, dissolve the pellet in 20 PL of DEPC-

506

Adams, Humid,

and Polak

TCGATCG

C-DNA Vector
SP6 T 7 T a polymerase

Transcription Biotin

Fig. 2. Diagrammatic representation of cRNA synthesis.

treated water. Remove 1 PL for assessment of incorporation of radioactivity. Store SE? probes at -7OC, 32Pprobes at -2OOC. The maximum storage time will depend on the radioisotope used. However, background increases with storage time. 8. Incorporation of radioactivity is estimated by determination of trichloroacetic acid (TCA)-precipitable counts. Mix 1 PL labeled RNA probe, 50 PL Bovine Serum Albumin, and 100 PL 10% trichloroacetic acid. Vacuum filter on GF/C (Glass microfiber paper, Whatman). Wash the filter twice with 10% TCA and twice with absolute ethanol. Dry the filters. Count the radioactivity on the filters and from this determine the percent of incorporation of radioactivity.

3.2. Fisation

of Material

(See Note

5)

3.2.1. Tissue We have evaluated the use of various fixatives, including 10% formalin, Bouins solution, 2.5% glutaradehyde/paraformaldehyde mixtures,

In Situ Hybridization

507

acetic acid/alcohol, and 4% paraformaldehyde. Formalin and Bouins gave the best morphology but poor RNA retention, whereas glutaraldehyde fixation retained more RNA, but the resultant morphology was poor. Thus, we compromised with 4% paraformaldehyde, which gave the best results when considering both RNA retention and cellular morphology. 1. Tissues for hybridization must be collected as fresh as possible. Fix small pieces of tissue (1 x 1 x 0.5 cm) by immersion in freshly prepared 4% paraformaldehyde for 4 h at 4OC(prolonged fixation reduces the hybridization signal). To avoid RNase contamination, wear gloves, use sterile equipment, and use DEPC treated water. 2. Transfer the tissue to phosphate-buffered saline containing 15% sucrose and store at 4C (maximum 2 mo). 3. After washing the tissue, prepare cryostat blocks, cut sections (15 pm), thaw-mount on poly-L-lysine (PII,)-coated glass slides, and allow to dry at 37OCovernight before processing for hybridization. Use these slides as soon as possible; otherwise store in a container with desiccant at -7OOC. 1. 3.2.2. Culture Cultured cells can be grown on coverslips, slides, or in suspension. Rinse coverslips or slides, with attached cultures, in cold phosphatebuffered saline, and fix by immersion in 4% paraformaldehyde for 60 min at 4C. Rinse several times in phosphate-buffered saline and finally in doubledistilled H20. Dry the cultures for 6 h at 37C and store in a dessicated box at -70C. Cell lines, grown in suspension, are cytospun onto poly-L-lysinecoated slides, which are then air-dried for 5-10 min before being fixed and stored as above.

2. 3. 4.

3.3. In Situ Hybridization

with cRNA Probes

Precautions should be taken to avoid RNase contamination until hybridization is complete. Coverslips with cultured cells grown on them should be attached to slides with paper clips (12). 3.3.1. Tissue Preparation 1. Rehydrate in phosphate-buffered saline (PBS) (pH 7.2) for 5 min. 2. Immerse in O.lM glycine/PBS (5 min).

508

Adams, Hamid,

and Polak

3. Permeabilize by immersion in 0.3% Triton X 100 in PBS for 15 min. 4. Wash with PBS (2x 3 min). 5. Deproteinize by incubation with 1 pg/mL Proteinase Ksolution for 20 min at 37OC. 6. Stop deproteinization by immersion in 4% paraformaldehyde/PBS (5 min). 7. Immerse in freshly prepared acetylation solution for 10 min to reduce nonspecific binding. 8. Prehybridize in 50% formamide, 2x SSC (37OC at least 15 min) to enhance signal to noise ratio. 1. 2. 3.3.2. Hybridization Drain the slides briefly (do not dry). Apply 20 PL of hybridization mixture preheated to 37OC, containing 2-3 ng radiolabeled cRNA probe (5 x lo5 cpm/section) diluted in hybridization buffer. Cover the sections with suitably sized dimethyl dichlorosilane-coated coverslips. Incubate at 37-43OC for 16 h. 3.3.3. Posthybridization Washing Remove the coverslips by immersion in 4 x standard sodium citrate (SSC). Wash the slides with 4 x SSC (37C, 3 x 20 min) with gentle shaking. Remove unhybridized single-stranded cRNA probe by treating preparations with RNase A solution for 30 min at 37C. Wash the slides with 2 x SSC (37OC, 30 min), with gentle shaking. Wash the slides with 0.1 x SSC (37OC,30 min), with gentle shaking. Dehydrate in 70,90, and 2 x 100% ethanol containing 0.3M ammonium acetate (10 min each at room temperature). Air dry (60-90 min). 3.3.4. Autoradiography Dip the slides in emulsion. Air dry for l-2 h. Store in light box for 2-5 d, depending on the radioisotopes used (2-3 d for 321?, 5 d for =S>. Develop in Kodak D19 developer prepared and used according to manufacturers instructions. Fix as appropriate. Wash well with water.

3. 4. 1. 2. 3. 4. 5. 6. 7. 1, 2. 3q 4. 5.

In Situ Hybridization

509 3.3.5. Counterstaining

1. Preparations are usually lightly counterstained with hematoxylin or hematoxylin/eosin. However, eosin may cover fine grains. Other counterstains may be used where appropriate (e.g., Coomassie blue for myocytes, Toluidine blue for brain preparations). 2. Dehydrate, clear, and mount with DPX. Examples of in situ hybridization results are shown in Figs. 3-6.

4. Notes
1. Control experiments are very important to assess the specificity of the hybridization and should include the following: a. Sense probes: Probes identical to the coding strand of the mRNA under investigation are transcribed and hybridized as above. b. Ribonuclease treatment: Sections or cultures are treated with RNase A (20 pg/mL, 37OC, 30 min) before the prehybridization step. A remnant of the ribonuclease could result in probe degradation and thus invalidates the results. c. Inappropriate probe for the tissue in question. d. Inappropriate tissue for the probe in question. e. Northern Blot Analysis: The presence of the particular mRNA in the tissue may be confirmed by Northern Blot hybridization. f. Several probes, coding for different regions of the same gene. g. Immunocytochemistry: The correlation of immunocytochemistry results with thoseobtained by in situ hybridization is one of the most useful indications of the specificity of the signal. 2, Signal/ Background Ratio: Although probes labeled with % give better subcellular resolution than those labeled with 3?l?, there is an increase in background. The background may be reduced by the following manipulations: a. Decreased autoradiography time. b. Minimize the amount of probe used for hybridization. c. The inclusion of dithiothreitol(50 mM) in one or all the following: prehybridization solution, hybridization mixture or posthybridization washings. This is particularly true with the use of ?S-labeled probe. d. The addition of cold cytidine triphosphate to the prehybridization buffer.

510

Adams, Hamid,

and Polak

Fig. 3. In situ hybridization of ANP-mRNA in cultured ventricle using ?P-labeled ANP-cRNA probes.

myocytes of rat atrium

and

Fig. 4. Autoradiographic preparations of cultures, fixed in4% paraformaldehyde, counterstained with hematoxylin x 2!?10.

and

In Situ Hybridization

511

Fig. 5. In situ hybridization of CGRP-mRNA in rat colon (fixed by perfusion with 4% paraformaldehyde) using a T-labeled cFNA probe. Positive hybridization signal in submucous plexus (SMP) x 300. MM = muscularis mucosa, IG = intestinal glands.

Fig. 6. The localization of bombesin-mRNA in culture of small cell carcinoma lung, cytospun onto slides, using 9 bombesin-cRNA probe.

of the

512

Adams, Hamid, and Polak

3. In situ hybridization: a. Combined immunocytochemistry and in situ hybridization. Immunocytochemistry may be done subsequent to in situ hybridization (a), the difference being the ommission of dextran sulfate from the hybridization buffer. After the final posthybridization wash, slides are processed as normal for immunocytochemistry. On completion of these protocols, they are dehydrated and dipped for autoradiography as described above. b. Quantitation of autoradiographic signal (12). Before attempting to quantify the autoradiographic preparations, one should take into consideration many factors that could affect hybridization signal such as the thickness of tissue or emulsion, loss of mRNA in tissue processing, and the efficiency of in situ procedure. Densitometry or computer imaging analysis give semi-quantitative estimates. 4. Transcription: A major problem when working with mRNA preparations is RNase contamination. Thus, gloves should be worn throughout the transcription and hybridization protocols. Glassware should be baked at 250C for 4 h; batches of plasticware should be set aside exclusively for RNA work and autoclaved where appropriate before use. All solutions should be prepared with DEPC-treated water. DEPC (O.l%, final concentration) is added to distilled water and left at room temperature for 12 h. Residual DEPC is destroyed by autoclaving this water for 15 min. Solutions prepared for transcription with this water are then aliquoted into sterile tubes and stored at -20C. Other labeled rNTPs can be used (10 mCi/mL, 400 Ci/mmol). This reaction can be run in the absence of unlabeled cytidine triphosphate. For a 20 PL reaction, 100 PCi of 400 Ci/mmol CX-~~P-CTPis 12 pm. However, the yield of full-length transcripts drops as the concentration of limiting nucleotide cytidine triphosphate falls below 12 l.t.m. The size of the probes may be reduced by alkaline hydrolysis (7). 5. Fixation: In animal experiments, perfusion with 4% paraformaldehyde gives the best results: Anesthetize rats with ether and perfuse intracardially with PBS followed by 4% paraformaldehyde. Remove the appropriate tissue and continue fixation in 4% paraformaldehyde for 1 h as above. Human tissue could also be perfused, for example, perfusion of the bowel through mesenteric vessels and the brain through the Circle of Willis.

In Situ Hybridization

513

Acknowledgments
The authors are grateful to J. Dixon, Purdue University, West Lafayette, USA; E. Spindel, Harvard Medical School, USA; and S. Amara, Yale University, New Haven, USA, for supplying ANP, bombesin, and CGRl?, cDNA, respectively. This work was supported by Amsersham International and the Upjohn Company.

References
1. Gall, J. and Pardue, M. (1969) Formation and detection of RNA-DNA hybrid mole-

culesin cytological preparations. Proc.Nutl. Acad. Sci. USA 63,378-383.

2. John, H. A., Birnstiel, M. L., and Jones, K. W. (1969) RNA-DNA hybrids at the cytological level. Nature 223,582-587. 3. Buongiorno-Nardelli, S. and Amaldi, F. (1970) Autoradiographic detection of molecular hybrids between RNA and DNA in tissue sections. Nature 225,946-948. 4. Coghlan, J. I., Aldred, I., Haralambidis, J.,Niall, H. D., Penschow, J. D., and Tregear, G. W. (1985) Hybridization histochemistry. Anulyf. Biochem. 149,1-28. J., Darling, P. E.,Darby, I. A., Wintour,E. M.,Tregear, 5. Penschow, J. D.,Haralambidis, G. W., and Coghlan, J. P. (1987) Hybridization histochemistry. Experienfiu 43, 741-750. 6. Vamdell, I. M., Polak, J. M., Sikri, K. L., Minth, C. D., Bloom, S. R., and Dixon, J. E. (1984)VisualisationofmessengerRNAdirectingpeptidesynthesisbyinsifuhybridization using a novel single-stranded cDNA probe. Hisfochemisfry 81,597-601. 7. Cox, K. H., De Leon, D. V., Angerer, L. M., and Angerer, R. C. (1984) Detection of mRNAs in sea urchin embryos by in situ hybridization using asymmetric RNA probes. Deal. Biol. 101,485-502. M. R., Lechan, R. M., Goodman, R. H., and 8. Hoefler, H., Childers, H., Montminy, Wolfe, H. J. (1986) In situ hybridization methods for the detection of somatostatin mRNA in tissue sections using antisera RNA probes. Hisfocbrn. J. l&597-604. 9, Terenghi, G., Polak, J. M., Hamid, Q. A., OBrien, E., Denny, I., Legon, S., Dixon, J., Minth, C., Palay, S. L., Yasargil, G., and Chan-Palay, V. (1987) Localisation of neuropeptide Y-mRNA in neurones of the human brain cortex using in situ hybridization and cRNA probes. Pm. Nufl. Acud. Sci. USA 84,7315-7318. 10. Hamid, Q. A., Wharton, J., Terenghi, G., Hassalle, C. J., Aimi, H., Taylor, K. M., Nakazato, H., Dixon, J. E., Bumstock, G., and Polak, J. M. (1987) Localization of atria1 natriuretic peptide mRNA and immunoreactivity in rat heart and human atria1 appendage. Proc. Nufl. Acud. Sci. USA 84,6760-6764. 11. Hamid, Q. A., Bishop, A. E., Springall, D. R., Adams, C.,Giaid, A., Denny,P., Ghatei, M., Legon, S., Cuttitta, F., Rode, J., Spindel, E., Bloom, S. R., and Polak, J. M. (1989) Detection of human probombesin mRNA neuroendocrine (in small) cell carcinoma of the lung: in situ hybridization with cRNA probe. Cancer 63,266271. 12. McCabe, J, T., Morrell, J. I., and Pfaff, D. W. (1986) In situ hybridization as a quantitative autoradiographic method: Vasopressin and oxytocin gene transcription in the Brattleboro rat, in In situ Hybridization in Bruin (Uhl, G. R., ed.), Plenum, New York, Ch. 5, pp. 73-97.

Chapter 38 Immunogold Electron

John

Labeling Microscopy
Pacy

for

1. Introduction
Colloidal gold immunocytochemistry was first introduced by Faulk and Taylor (I), and has rapidly become a major technique in electron microscopy, covering many aspects of biological research. The rapid expansion of this technique in electron microscopy is first the result of its simplicity compared with other labeling techniques, and secondly, the result of the properties of the gold probes themselves. These probes are usually 3-15 nm in diameter and coated with immunologically active proteins. They have the advantages of being very electron dense, giving a characteristic appearance that cannot be confused with other biological structures; they are highly sensitive, producing very specific labeling of both monoclonal and polyclonal antibodies; they are also permanent, nonhazardous, and can be easily quantified. Several techniques now exist for obtaining different types of information from the sample. The successof these techniques depends on using very specific antibodies that have a high affinity or binding capacity to their antigen. Any unwanted antibodies present in the preparation should be of a lower affinity than the specific antibody. The choice of antibody, whether monoclonal or polyclonal, is still mainly determined by availability. Fixation reduces antigenicity, and polyclonal antibodies are less
515

516

Pacy

affected than monoclonals by the fixation procedures. However, greater specificity can be obtained with monclonal antibodies. The choice of using secondary antibody/gold or protein A/gold is still a complex one, especially as more secondary antibody/gold probes are becoming commercially available. In general, secondary antibody/ gold produces less background than protein A/gold when used in higher dilutions. Some signal amplification is found using the secondary antibody/gold probes because of the binding of more than one probe to one primary antibody. This results in clusters of gold particles that can be helpful when the marker density is low. The following techniques are now well established and will be dealt with in this chapter:

1. The preembedding technique for the localization of external antigens

on whole cells. 2. The postembedding technique for the localization of both internal and external antigens. 3. Cryo-techniques for sensitive antigens that would be destroyed by conventional processing. 4. The immunonegative staining technique for localization of surface antigens on small specimens, such as viruses and bacteria, which lend themselves to negative staining. 5. The double-labeling technique for localizing more than one species of antigen by using different sized probes. Thesetechniques listed above are the most common, but other techniques exist such as those involving freeze-fracture and immunoreplica.

There is now a wide range of commercially available gold probes,

and consequently, methods for producing these probes will be dealt with only briefly after the immunolabeling techniques.

2. Materials
1. Fixatives: a. 1% glutaraldehyde in O.lM sodium cacodylate buffer at pH 7.2 for polyclonal antibodies. b. 4% paraformaldehyde plus 0.05% glutaraldehyde in O.lM sodium cacodylate buffer at pH 7.2 for monoclonal antibodies. c. 2.5% glutaraldehyde in O.lM sodium cacodylate buffer at pH 7.2. (For use in the preembedding and cryo-techniques.) 2. 2% osmium tetroxide in distilled water. 3. Ethanol solutions: 30,50,70,90,100% by volume.

Immunogold

Labeling for Microscopy

517

4. Embedding media: a. L. R. White embedding resin. b. Spurr embedding resin: NSA (nonyl succinic anhydride) 26 mL, ERL 4206 (vinyl cyclohexane dioxide) 10 mL, DER (Diglycidyl ether of polypropylene glycol) 6 mL, Sl (dimethylaminoethanol) 0.4 mL. c. 10% gelatin in O.lM sodium cacodylate buffer pH 7.2 (dissolve the gelatin powder by warming and store in microcentrifuge tubes at 4OC). 5. Stains: a. 2% aqueous uranyl acetate. b. 2% neutral uranyl acetate (2). Prepare by mixing 4% aqueous UA (O.lM) and 0.3M oxalic acid in equal proportion. Add a small amount of 10% ammonium hydroxide to adjust the pH to 7.2-7.4. c. 2% aqueous ammonium molybdate. d. Reynolds lead citrate (3). To prepare: Add 1.33 g lead nitrate and 1.76 g sodium citrate to 30 mL boiled distilled water in a 50 mL volumetric flask. Shake the suspension vigorously for 1 min, and then allow to stand for 30 min with intermittant shaking. Add 8 mL of 1 NaOH (freshly prepared) and dilute to 50 mL with boiled distilled water. Mix by inversion. 6. Phosphate immunobuffer: The phosphate immunobuffer used for incubations is PBS plus 1% BSA. Phosphate buffered saline (PBS): 0.14M NaCl, 3 mM KCl, 10 mM Na,HPO,, 3 mM KH2POI, 0.1% NaN, (sodium azide) can be added as a preservative. 7. Tris immunobuffer: Stock buffer: 6 g Tris (tris-hydroxymethyl-aminomethane), 9g NaCl, 1 g gelatin (dissolved in 10 mL of distilled water by warming), 990 mL distilled water, pH to 7.4 with cont. HCl. 1 g of NaN, can be added as a preservative. Final buffer: Add 1% (w/v) ovalbumin and 0.01% (v/v) Tween 20 to the stock buffer and pH to 8.2. (1% Bovine serum albumin (BSA) can be used instead of ovalbumin.) TheTris immunobuffer used for incubations is therefore: 0.05M Tris, 0.01% Tween 20,0.1% gelatin, 1.0% ovalbumin, made up in 0.9% NaCl solution. 8. 1% (w/v) gelatin in PBS. 9. 0.02M glycine in PBS. 10. 2.3M sucrose. 11. 0.5M NH,Cl.

518

Pacy

12. Primary antibody prepared in serial dilutions of immunobuffer. Start with 1 in 10. 13. Gold conjugate prepared in serial dilutions with immunobuffer. Start with 1 in 4 (e.g., for rabbit primary antibody use Goat Anti Rabbit or Protein A; for mouse primary antibody, use Goat Anti Mouse or Protein A). The gold conjugate can be obtained commercially or made in the laboratory. 14. 2% aqueous uranyl acetate plus 1.5% methyl cellulose (1:l mixture). Store at 4OC. 15. 0.01% chloroauric acid (gold chloride) in distilled water. 16. 1% (w/v) trisodium citrate in distilled water. 17. 0.2M potassium carbonate. 18. Ether solution of white phosphorus. Small pieces of white phosphorus sticks are cut under water and quickly transferred to 20 mL diethyl ether. After stirring for 2 h, the supernatant is centrifuged for 20 min at 4800 g. One part of the supernatant is added to 4 parts &ethyl ether. 19. 1% (w/v) polyethylene glycol.

3. Methods
The labeling methods dealt with in this chapter are shown in the form of a schematic diagram (seeFig. 1). 3.1. The Preembedding Technique This method is for the localization of external antigens with the advantage that conventional electron microscopy (E. M.) can be used to give good preservation and contrast to the sample. 1. Prepare antigen for immunolabeling. 2. Wash sample in PBSpH 7.2 plus 1% Bovine Serum Albumin (BSA) for 5 min. 3. Incubate with a primary antibody of suitable dilution in PBSplus 1% BSA for 1 h. (SeeChapter 35 for details of dilutions.) 4. Wash in PBS plus 1% BSA, five changes of 1 min each. 5. Incubate with a gold probe of suitable dilution with PBSplus 1% BSA for 1 h. (Start with a 10 nM sized probe.) 6. Rinse in PBS, three changes of 1 min each. 7. Fix the sample in 2.5% glutaraldehyde in PBS for 5 min. 8. Wash well in distilled water, five changes of 1 min each. 9. Transfer to 2.5% glutaraldehyde in O.lM sodium cacodylate buffer pH 7.2 for 20 min.

Immunogold

Labeling for Microscopy

519

F zig.1. Schematic diagram of gold labeling methods.

10. Wash in O.lM cacodylate buffer, three changes of 5 min each.

11. Postfix in 1% osmium. tetroxide in O.lM sodium cacodylate buffer

12. 13. 14. 15. 16.

for 1 h. Wash in 30% ethanol for 5 min. Dehydrate through a series of ethanols, 50-70-90-lOO-lOO-100% for 15 min each. Clear in propylene oxide, two changes of 5 min each. Infiltrate with Spurr resin by increasing the resin to propylene dioxide ratio slowly over a few hours. Infiltrate with 100% Spurr resin for 2 h. (This is an arbitrary time dependent on the sample.)

520

Pacy

17. Embed in Beem capsules or equivalent. The sample can be spun down through the resin to form a pellet once inside the capsule. Polymerize in a 60C oven for 2 d. 18. Section and stain as for conventional E. M. (See Chapter 34, this vol. for details.) Apart from the polymerization of resin, all the other procedures are carried out at room temperature.

3.2.

The Postembedding

Technique

This method is used for the localization of both external and internal antigens that can survive the fixation and embedding procedures. Osmium tetroxide is usually omitted resulting in the problem of obtaining good contrast in the specimen. (SeeFigures 2a and 2b.)

3.2.1. Fixation

and Embedding

1. Prepare antigen for immunolabeling. 2. Fixation: For monoclonal antibodies, fix the sample in 4% paraformaldehyde plus 0.05% glutaraldehyde in O.lM sodium cacodylate buffer at pH 7.2 for as long as required. For polyclonal antibodies, fix the sample in 1% glutaraldehyde in O+lM sodium cacodylate buffer at pH 7.2 for half an hour. 3. Wash in cacodylate buffer, three changes of 5 min each. 4. Postfix in 1% osmium tetroxide for 1 h only if labeling is not affected. This stage is usually omitted. 5. Dehydrate in an ethanol series, 10 min each step: 50-70-90-100100-100%. 6. Transfer to L. R. White resin (4). Four changes of half an hour each. NOTE: pour out enough resin from stock to cover the infiltration and embedding steps. Be prepared to throw away any excess that is left. 7. Embed in airtight molds, preferably oven-dried gelatin capsules. Leave in a 60C oven for 22 h. 8. Section onto gold or nickel grids. Copper grids react with Tris buffer. Use carbon/formvar coated grids if the sections are delicate.

3.2.2. Antibody

Incubation

During this procedure the grids are floated, sections downwards on droplets. Allow two drops/grid at each stage. The time in the first drop need only be brief enough to wash off the previous solution. With Tris Immunobuffer:

Immunogold

Labeling for Microscopy

521

Fig. 2. (a) Localization of myosin antigens on B granules of Islet of Langerhans using the postembedding technique (70,OtXIx). fb) Localization of 4 protein antigens in PseudoG specific monasputida using the postembedding technique (30,000~). (c) Localizationof antigens on the surface of G organisms of B-hemolytic streptococci using the cryo-labeling technique (50,000x) (Courtesy of Dr. J. E. Beesley, Wellcome Research Laboratories, Beckenham, Kent.). (d) Simultaneous localization of Ca and HMFG 2 antigens in breast tissue using the double-labeling technique. Detection of Ca is with a 20 nM probe and HMFG 2 is with a 5 nM probe (30,000x) (Courtesy of Dr. J. E. Bee&y, Wellcome Research Labs.) Published in Proc. Roy. Mic. Sac. 20 (41,187-196 (permission granted).

1. Float grids on 0.5M NH,Cl for 10 min to neutralize free aldehyde groups. 2. Wash sections with Tris immunobuffer (No. 7 in materials section), two changes of 2 min each. 3. Incubate sections on drops of primary antibody diluted with immunobuffer, for 1 h at room temperature.

522

Pacy

4. Wash with immunobuffer, five changes of 1 min each. This is to remove unbound primary antibody. 5. Transfer grids to drops of gold conjugate (either secondary antibody gold, e.g., GAR. or protein A) diluted with immunobuffer, for 1 h at room temperature. 6. Wash on drops of distilled water, five changes of I min each. This is to remove unbound gold conjugate. 7. Obtain contrast with conventional E. M. staining procedures. With PBS Immunobuffer: 1. Float grids on PBS plus 0.02M glycine for 5 min to neutralize any free aldehyde sites. 2. Transfer to PBS plus 1% gelatin for 10 min to neutralize sticky sites on the sample, and thus reducing nonspecific binding. 3. Wash in PBS plus 1% BSA, two changes of 3 min each. BSA attaches to nonspecific sites, thus competing with the primary antibody and reducing nonspecific background labeling. 4. Incubate grids in a suitable dilution of antibody in PBS plus 1% BSA for 1 h. 5. Wash in PBS plus 1% BSA, five changes of 1 min each. 6. Incubate in a suitable dilution of gold conjugate in PBS plus 1% BSA for 1 h. 7. Wash in distilled water, five changes of 1 min each. 8. Obtain contrast as for conventional E. M. sections.

3.3. The Cryo-Technique

This method is for antigens that would not survive conventional processing and presents us with the inherent difficulties of cryo-sectioning, It does, however, give the best preservation of antigens (seeFig. 2~). 1. 2. 3. 4. Prepare antigen for immunolabeling. Fix sample, depending on the antibody. See previous method. Wash in cacodylate buffer, three changes of 5 min each. Embed in 10% gelatin in cacodylate buffer. Heat the gelatin (stored at 4C) to 37OC before use. Warm pipets and tubes before use. Resuspend the sample in the warm gelatin and spin down for 2 min into a pellet. Place in ice to set the gelatin. 5. Cut out the pellet and refix small pieces of it in the original fixative for 10 min to cross-link the gelatin.

Inamunogold

Labeling for Microscopy

6. Cryo-protect in 2.3M sucrose for 1 h. 7. Freeze in nitrogen slush. Prepare slush by placing a wide-mouthed polystyrene container with about l-in depth of liquid nitrogen into a vacuum chamber. Evacuate for a few minutes until the liquid N, turns to slush. Return the chamber quickly to atmospheric pressure, and use the slush immediately. 8. Section at about -80C, with the knife slightly colder at about -9OOC. 9. Collect sections on a droplet of saturated sucrose solution, and mount onto a carbon/formvar coated grid (3). 10. Transfer the grid to a droplet of PBS plus 1% BSA to remove sucrose (3). NOTE: It is advisable to check that there is a sample in the sections. Take preliminary sections, mount on a grid, wash with three droplets of distilled water and negative stain with 2% ammonium molybdate. l-6. 7. 8. 9. 10. 11. 12. 3.3.1. Incubation Follow steps 14 of the postembedding technique with phosphate immunobuffer. Rinse in PBS, two changes of 2 min each. Fix in 2.5% glutaraldehyde in PBS for 2-10 min. Rinse in distilled water, five changes of 1 min each to remove phosphate. Stain in uranyl acetate oxalate for 5 min (3). Rinse in distilled water, three changes of 20 s each. Stabilize and enhance the contrast of the sections by floating them on a 1:1 mixture of 2% aqueous uranyl acetate and 1.5% methyl cellulose for 2-3 min. This has to be done on ice to keep the methyl cellulose from setting.

3.4. The Immunonegative

Staining

Technique

This method is for the localization of surface antigens on small specimens such as bacteria and viruses. Fixation and embedding procedures are not involved and therefore the maximum labeling potential of the untreated antigens is retained. This means that low levels of antigen can be labeled. Negative staining is a fast and simple technique to perform, giving high contrast and high resolution. 1. Prepare antigen for immunolabeling. 2. Dry the sample down on to a carbon/formvar coated 400 mesh grid.

524

Pacy

3. Rinse in PBSplus 1% BSA, 2x 1 min. 4. Incubate grids with primary antibody diluted with PBS plus 1% BSA for 15 min. 5. Wash with PBS plus 1% BSA, 4x 1 min. 6. Incubate grids with gold probe diluted with PBS plus 1% BSA for 15 min. 7. Wash with PBS, 4x 1 min. 8. Negative stain with 2% ammonium molybdate.

3.5. The Double-Labeling

Technique

This method is for localizing more than one antigen by using different sized gold probes (seeFig. 2d). Double-labeling can be carried out using any of the previous techniques. The incubations can be simultaneous, providing there is no cross-reaction of reagents, or sequential, with a blocking step if there is cross-reaction. In general, the smaller probe is incubated first. Sections on uncoated grids are incubated on one side and then carbon coated. The second incubation can then be carried out on the reverse side without risk of contamination.

3.6. Preparation

of Gold Probes

The production of gold probes is in two distinct steps. The first is the production of colloidal gold. This involves the reduction of gold salts into metallic gold spheres that can vary in size depending on the method used for their preparation. The second step is the coating of the spheres with the chosen protein/antibody. This procedure stabilizes the colloidal gold and is more complicated than the reduction process. The preparation of protein A-gold complex will be used as an example. 3.6.1. Preparation
of Colloidal

Gold water

All glassware must be thoroughly cleaned, and double-distilled should be used in all solutions. For 5 nM spheres (5):

1. Add 1.5 mL of 1% HAuCl, to 120 mL distilled water. To this add 2.7 mL 0.2M potassium carbonate. 2. Add 1 mL of ether-saturated white phosphorus slowly and shake for 15 min at room temperature. The mixture becomes brownish-red. 3. Gently heat over an open flame until the color changes to a wine-red color.

Immunogold

Labeling for Microscopy

525

For 10 nM spheres (6): 1. Add 15 mL of 1% trisodium citrate to 247.5 mL boiling water. Continue boiling for 5 min. 2. Quickly add 2.5 mL of 1% HAuCl, (chloroauric acid) while stirring. The solution changes from blue to red; continue boiling for a further 10 min to complete the reduction. 3. Allow to cool and pH to 6.9 with 0.2M potassium carbonate. For 15,30, and 40 nM spheres (6): 1. Heat 100 mL 0.01% HAuCl, to boiling. 2. Add 3.0 mL, 1.5 mL, and 1.2 mL of 1% sodium citrate, respectively, for 15,30, and 40 nM spheres. 3. Continue boiling for 10 min, cool, and pH to 6.9 with 0.2M potassium carbonate.

3.62.

Preparation

of Protein A-Gold

Complex (7)

1. Dissolve 0.5 mg purified protein A in 0.1 mL distilled water and mix with 10 mL colloidal gold. 2. After 2-3 min, 1 mL of 1% aqueous polyethylene glycol is added, and centrifuged for 1 h at 20,OOOg. 3. The dark red, loose pellet is resuspended in 6 mL of 0.2 mg/mL of polyethylene glycol in phosphate buffered saline (pH 7.4) and stored at 4C.

4. Notes
1. It is imperative to include appropriate controls in all immunolabeling techniques. Negative controls are straightforward to carry out. They could include replacing the primary antibody with an alternative antibody from the same species raised against an antigen known to be absent in the sample, replacing the primary antibody with nonimmune serum, or replacing the primary antibody with buffer only. Positive controls, where possible, should also be included. 2. Longer incubation with higher dilutions of primary antibody reduces nonspecific binding and, thus, produces more specific binding. Higher dilutions of gold conjugate can reduce nonspecific binding to the highly charged components of the sample (e.g., necrotic cells, nucleic acids, and so on). 3. Clustering of gold particles can occur. This may result from the natural amplification of the gold conjugate when using IgG gold

526

Pacy

probes, and does not occur with protein A gold. Clustering may also result from clumping of the primary antibody because of age or storage conditions. 4. Trouble Shooting a. No label (1) The antigen is absent or destroyed by the preparative procedures. Change procedure. (2) The antb 1 o d y is not working. Could be the result of age, poor storage, wrong dilution, wrong antibody, excessive freezing and thawing. Change the antibody and run a positive control. (3) The antigen is present, but in very low amounts. Extend the incubation times and use more concentrated primary antibody. (4) The gold probe is not working. Could be the result of its poor condition or even a wrong probe. Repeat the incubation to check. (5) The pH of the solutions is wrong. Check this. b. Excessive background labeling (1) Concentration of primary antibody and/or gold conjugate is too high. Dilute at least 10 times. (2) Insufficient washing between incubations. Increase the washing times substantially. (3) Nonspecific binding is occurring. Add ovalbumin or BSA or increase their concentrations if already present. (4) Nonspecific charge attraction of the antibody is occuring. Use a detergent in all solutions (e.g., 1% Tween 20). Osmium tetroxide fixation can introduce excess charge into the sample. Omit the osmium. (5) Binding to free aldehyde groups in fixed tissue might have occurred. Neutralize these with glycine or NH&l before incubation. (6) Bad fixation can sometimes produce background labeling in dead or damaged cells. Improve fixation. 5. Low temperature embedding can be achieved using Lowicryl K4M resin (8). The fixation and embedding procedures are carried out at 4C with the polymerization of resin completed with UV radiation at the same temperature. Low temperatures enhance preservation of antigens (9), and proteins are less likely toleachoutof the tissue. There

Immunogold

Labeling for Microscopy

527

are some doubts, however, that because of a rise in temperature during polymerization (10) true low temperature embedding is achieved. 6. A number of useful general references (1Z-1 8) are included at the end of this chapter.

Acknowledgments
I wish to thank Trish Dopping-Hepenstal for her assistance in preparing this chapter. I would also like to thank Dr. Julian Beesley for his micrographs and for his help in the past.

References
1. Faulk, W. P. and Taylor, G. M. (1971) An immunocolloid method for the electron microscope. immunochemistry 8,1081-1083. 2. Tokuyasu, K. T. (1978)A study of positive staining of ultrathin frozen sections. J. of Ultrastructure Res. 63,287-307. 3. Reynolds, E. S. (1963) The use of lead citrate at high pH as an electron opaque stain in electron microscopy. J. Cell. Biol. 17,208. 4. Timms, B. G. (1986) Postembedding immunogold labeling for electron microscopy using L. R. White resin. Amer. J, of Anat. 175,267-275. 5. Zsigmondy, R. (1905) Zur Erkenntnisse der Kolloide. Z. Physik. Chem. 56,65. 6. Frens, G. (1973) Controlled nucleation for the regulation of particle size in monodisperse gold suspensions. Nature Phys. Sci. 241,20-22. 7. Roth, J. and Binder, M. (1978) Colloidal gold, ferritin and peroxidase as markers for electron microscopic double labeling lectin techniques. I, Histochem. Cyfochem. 26, 163-169. 8. Valentino, K. L., Crumrine, D. A., and Reichardt, L. F. (1985) Lowicryl K4M embedding of brain tissue for immunogold electron microscopy. J.Histochem. Cyfochem. 33,
969-973. 9. Armbruster, B. L., Garavito, R. M., and Kellenberger, E. (1983) Dehydration and

embedding temperatures affect the antigenic specificity of tubulin and immunolabelling by the protein A-colloidal gold technique. 1, Hisfochem. Cytochem. 31, 1380-1384. 10. Ashford, A. E., Allaway, W. G., Gubler, F., Lennon, A., and Sleegers, J. (1986) Temperature control in Lowicryl K4M and glycol methacrylate during polymerisation: is there a low temperature embedding method? 1.Microscopy 144,107-126. 11. Beesley,J. E. (1987) Colloidal gold electron immunocytochemistry: its potential in medical microbiology. Seriodiug.and lrnmuno therapy1,239-252. 12. Roth, J. (1982) The protein A-gold technique. Tech. in Immunocytochem. 1,108133. 13. Polak, J. M. and Van Noorden, S.(1984) An introduction to immunocytochemistry; current techniquesand problems, inM. S.Handbook22(Royal Microscopical Society, Oxford, England).

528

Pacy

14. Bullock, G. and Petrusz, P. (1983,1984,1985) Techniques in immunocyfochmisty, ~01s.1,2,3 (Academic, London: New York). 25. Tokuyasu, K. T. (1986) Immunocryoultramicrotomy. J. Microsc. 143,139-149. 26. Polak, J.M. and Varndell, I. M, eds. (1984) Immunolabeling for electronmicroscopy. Elsevier, Amsterdam, BV., pp. 129-142. 17. Beesley, J. E. (1986) The use of gold markers in immunocytochemical studies of microbiological organisms: a review. 1. Microsc. 143,177-186. 18. Polak, J. and Van Noorden, S. (1986) hnmunocytochemisfry: Modern Methods and Applications. J. Wright and Sons.

Chapter 39
Human Chromosome Analysis and Sorting
Judith A. Fantes and Da@ K. Green
I. Introduction
Flow cytometry has provided the cytogeneticist with a fast and accurate method of measuring the quantity of DNA in each human chromosome (I). Almost all the chromosomes in the human complement can now be resolved and abnormal chromosomes and aneuploidies (13,21, and X) recognized. A flow karyotype shows apattern of peaks and troughs that is unique for each individual or cell line because of the variation in heterochromatic regions of the chromosomes (2). When combined with family studies, flow cytometry has been able to resolve homologues differing in DNA content by as little as l/2000 of the human genome (3,4), less than a metaphase band. In addition, the sorting capabilities of most flow machines have provided a method for the purification of small but useful quantities of individual chromosomes, for example, 2x lo6 average sized human chromosomes are equivalent to 500 ng of DNA. Using recombinant DNA techniques, this material can be used to generate a large number of DNA probes to produce a chromosome-specific library, which can be
529

530

Fantes and Green

used for the molecular analysis of genetic disease (5,6). More recently, molecular biologists have experimented with gene mapping by sorting small quantities of individual chromosomes onto filters for spot-blot hybridization with DNA probes (7). The sample preparation and flow machine techniques relating to all the above biological objectives will be discussed here. The art of producing an enriched sample of a particular group of human chromosomes by flow cytometry lies in bringing a clean, well-separated chromosome suspension to a clean, sterile, and well-adjusted flow machine. Debris, unbalanced stain/chromosome concentration or clumps of aggregated chromosomes in the suspension or noise in the flow machine in the form of obstruction to the flow, optical misalignment, or electronic noise all contribute to a reduction in purity of the sorted sample. The methods described here are aimed at producing the best possible sample obtainable from a starting material of peripheral blood lymphocytes or lymphoblastoid cells to the end point of verifying the identity and purity of the chosen sorted chromosome group.

2. Materials
1. Complete culture medium: RPM1 1640,10% fetal calf serum, 12.5 mM MOPS, lOOU/mL penicillin, 100 pg/mL streptomycin, 2. Phyto-hemagglutinin (R-IA): reagent grade from Wellcome. 3. Lymphoprep from Nyegaard. 4. Polyamine buffer (Bl): 15 mM Tris, 0.2 mM spermine, 0.5 mM spermidine, 2 mM EDTA, 0.5 mM EGTA, 80 mM KCI, 20 mM NaCl, 14 mM (0.1% v/v) B-mercaptoethanol. Adjust to pH 7.2 with 1N HCl before adding B-mercaptoethanol. Prepare fresh every week; store at 4OC. 5. Polyamine buffer plus digitonin (B2): 0.1% solution of digitonin in Bl. Prepare a saturated solution by heating to 37OC and filtering through a 0.2 pm filter to remove any undissolved digitonin. The best source of digitonin is Fluka, since some batches of digitonin from other sources are difficult to dissolve. 6. Hoechst 33258: 100 w in distilled water. Store in dark at 4OC. 7. Ethidium bromide: 1 mg/mL in distilled water. Store in dark at 4OC. 8. Chromomycin A3: 1 mg/mL in distilled water. Leave overnight in cold to dissolve. Store in dark at 4OC. 9. Colcemid: 0.01 mg/mL in distilled water. Filter sterilize and store at 4OC.

Chromosomes, Analysis and Sorting

531

10. DAPI (4,6-Diamidino-2-phenylindole l 2HCl): 50 pg /mL in distilled water. Store in dark at 4C. 11. Spermidine Es-acridine (8): Dissolve 5 mg in 2 mL methanol. Make up to 100 mL with 10 mM disodium orthophosphate, adjust to pH 6.5 with 0.2M HCl. Store in dark at 4*C. 12. Decon or 7X detergent diluted 1:25with sterile filtered distilled water. 13. Activated glutaraldehyde (Cidex). 14. Distilled water sterilized by filtration through 0.2 pm filter. 15. Phosphate buffered saline (Dulbecco A); filter sterilized.

3. Methods
The method of isolating chromosomes described here uses polyamines to stabilize the chromosomes and detergent treatment to lyse the cells (9). Since the lysis of interphase nuclei is minimal, this method can be used successfully with all types of cell lines, including suspension cultures with a significant proportion of interphase cells. Although the isolated chromosomes are highly condensed, they maintain most of their in vivo structure, and it is possible to prepare high mol wt DNA from them.

3.1. Cell Culture

Human chromosomes are usually prepared from lymphoblastoid cell lines (EBV transformed B lymphocytes, seeChapter 5, this volume) or from PHA-stimulated peripheral blood lymphocytes. The best chromosome suspensions and hence the flow karyotypes with the greatest resolution are prepared from cultures with a reasonable proportion of mitotic cells (>20%, seeNote 1). 1. Lymphoblastoid cells: set up cells from a stationary phase culture at 3 x 105viable cells/mL. Thirty hours later, add colcemid to a final concentration of 0.1 pg/mL to block the cells at metaphase. Sixteen to eighteen hours later harvest cultures and disperse any cell clumps by gentle pipeting. 2. Lymphocytes: Defibrinate a 40-mL sample of peripheral blood with orange sticks. Spin down the cells at 4008 for 10 min, and remove se rum (=20 mL). Replace with an equal volume of medium without FCS. Layer 10 mL of suspension onto 7 mL lymphoprep in a sterile plastic centrifuge tube. Spin for 15 min at SOOg.The lymphocytes collect at the interface and can be removed with a pastette. They are washed

532

Fantes and Green twice by centrifugation with excess complete medium and counted before setting up at 0.5 x W cells/ml in complete medium containing 15% fetal calf serum and 1% PHA. Colcemid is added to 0.1 pg/mL after 46 or 66 h of incubation, and the cultures are harvested 17 h later.

3.2. Chromosome

Preparation

1. Take an aliquot of cell suspension for a cell count. Centrifuge cells at 180s for 10 min. Pour off supernatant and resuspend cells in fresh icecold complete medium. Centrifuge cells at 18Ogfor 10 min; this washing step removes some dead cells and debris. 2. Pour off the supematant and resuspend cells in hypotonic 0.075M KC1 solution to swell the cells. Use 10 mL of hypotonic for every lo7 cells. 3. Incubate the lymphoblastoid cells for 20 min at 37OC;peripheral blood lymphocytes require 10 min incubation at room temperature at this stage. 4, Remove 0.25 mL for mitotic index determination: add 5 mL of 3:l methanol:acetic acid and allow to stand for 10 min. Centrifuge at 3008 for 5 min, pour off the supernatant, and resuspend the pellet in a small volume of fixative (see Note 2). Drop onto a clean slide, and dry quickly in air. Stain in 2% Giemsa in pH 6.8 buffer and air dry. Count the number of divisions in 500 cells. 5. After incubation in hypotonic solution, centrifuge the cell suspension at 1808 for 5 min. All further steps should be carried out at 4OC. 6. Pour off the supernatant and resuspend the pellet in cold polyamine buffer (Bl), 1 mL/ 10 cells. Centrifuge at 1808 for 5 min. 7. Pour off the supernatant. Resuspend the pellet in cold polyamine buffer plus digitonin (B2), 1 mL/ lo7 cells (seeNote 3). 8. Vortex vigorously for 30-60 s to break the cell walls. Monitor cell lysis by phase-contrast microscopy or by fluorescence microscopy. In this case, drop the chromosome suspension onto a slide previously spread with a drop of fluorochrome such as Hoechst 33258 or ethidium bromide. Place a coverslip in position, seal with rubber solution, and examine. Most of the chromosomes should be free and in suspension after 60 s vortexing; further vortexing will only cause an increase in chromosome degradation and stickiness. 9. Nuclei should be removed from the chromosome suspension before flow analysis and sorting, since they can contaminate sorted fractions. Spin down the nuclei at 1808 for 10 min, and transfer the

Chromosomes, Analysis and Sorting

533

supernatant carefully to another tube. Add I mL of B2 buffer to the pellet and resuspend by a 5s vortex. Centrifuge at 180s for 5 min, remove supernatant, and add to the first supernatant. Check for the presence of nuclei as described above in stage 8 (seeNote 4). 10. This chromosome suspension can be stored for 24 wk at 4OCwith little loss of resolution when analyzed by flow cytometry.

3.3. Staining
1. If the chromosome suspension was prepared from cells with a high mitotic index (> 30%), dilute I:1 with fresh B2 buffer before staining (seeNote 5). 2. For single fluorochrome analysis, add Hoechst 33258 to 0.5 pg/mL or ethidium bromide to 50 pg/mL. 3. For dual fluorochrome analysis (seeNote 6), add chromomycin A3 (see Note 7) to 20 pg/mL, magnesium chloride to 1mMand Hoechst 33258 to 0.5 pg/mL from stock solutions. Leave for at least 1 h for the fluorochromes to equilibrate before analyzing or sorting the chromosomes.

3.4. Chromosome

Identification

Although chromosomes prepared in polyamine buffer (seeNotes 8 and 9) after exposure to 16 h colcemid are condensed, it is possible to band and identify them if they are swollen and elongated slightly by prior exposure to phosphate buffered saline. Not all the chromosomes on a slide will be sufficiently decondensed to give adequate banding for identification, but over 50% should have sufficient bands and examples are shown in Fig. 1. 1. Sort 60,000 chromosomes into a cold Eppendorf tube containing 0.25 mL of buffer B2. On our machine, this quantity of chromosomes will be sorted in 0.25 mL of PBS sheath fluid so the sorted chromosomes will finally be exposed to 1:l buffer B2:PBS. 2. Fix chromosomes by adding 40% formaldehyde to give a final concentration of 4%. Leave for 10 min on ice. 3. Spin chromosomes onto alcohol-cleaned slides using a Shandon cytocentrifuge at 150s for 7min. 60,000 chromosomes can be split between two slides. 4. Allow slides to air dry. Wash briefly in deionized water and air dry. 5. Fix in 3:l methanohacetic acid for 5 min and air dry.

534

Fantes and Green

Fig. 1. Selected photomicrographs of sorted chromosomes 1,2,11,13 and Y. The chromosomes were sorted by flow cytometry, and deposited on microscope slides according to the method described in the text. Dapi staining was used for chromosomes 1,2,11, and 13 and spermidine his-acridine for chromosome Y. The distinctive banding of the chromosomes associated with these dyes can be seen.

6. Either: (a) Stain in 0.5 PgJmLDAPIin distilled water for 10 min, wash, and air dry. Mount in citifluor/glycerol; or (b) Stain in 0.005% spermidine his-acridine for 10 min, wash, and air dry. Mount in deionized water.

3.5.

Flow

Cytometry

The following methods do not relate to any particular commercial flow cytometer, but given that a machine consists of a light source (usually a laser beam), an optical train, a liquid flow arrangement to deliver the chromosome suspension, and a signal detection system, they are generally applicable to all machines. A general text dealing with the basic principles and a wide variety of applications of flow cytometry is recommended for a newcomer to the subject (10). A typical bivariate flow karyotype of normal human chromosomes is shown in Fig. 2. The analysis of human chromosomes requires the flow cytometer to be performing as well as or better than specification. It is not good enough to have the instrument roughly tuned to performance, which often suf-

Chromosomes, Analysis and Sorting

535

Human Chromosomes

CHROMOMYCIN

A3

Fig. 2. Human chromosome fluorescence distribution for two DNA-specific fluorescent dyes Hoechst 33258 and Chromomycin A3. Differential amounts of AT- and GC-rich DNA on each chromosome, which determines the intensity of Hoechst and Chromomytin fluorescence, respectively, produce further separation of the chromosome peaks than is seen with a single DNA-specific fluorochrome.

fices for cell analysis, and attention to detail will be rewarded with a reduced coefficient of variation (CV) of the chromosome peaks, provided of course that the suspension of chromosomes is well prepared. It will be assumed that the reader has some knowledge of a particular flow cytometry machine, and is able to adjust the signal detection and display system, clean all lens and signal detection surfaces, and optimize the output of the lasers or other light sources. Those parts of the system that are most likely to give rise to performance loss when they are not optimized are dealt with in detail below.

536

Fantes and Green

3.5.1. Sample Delivery The arrangement of nozzles and pressurized sheath and sample stream is designed to deliver chromosomes in single file through a welldefined position in the cytometer, where a beam of light can be focused. All tubes, nozzles, and reservoirs therefore must be clean, aligned, and purged before injecting a series of samples. This will ensure an unobstructed laminar flow, which leads to precise positioning of chromosomes and a stable droplet formation. Any liquids introduced into the cytometer must be filtered through a 0.2 pm filter. 1. Cleaning (see Note 10): Flush the following solutions through the machine for 30 min each. This should include backflushing through the sample stream. a. Warm Dilute 7-X (or equivalent) solution to remove chromosome material from previous experiments. b. Cidex to sterilize. c. Use distilled water to rinse out unwanted detergent or Cidex. d. Sheath buffer chosen for the experiment. 2. Alignment: In many cases, the sample and sheath nozzles have a fixed geometry, but in cases where adjustment is provided, make sure that the sample injection nozzle is placed centrally inside the sheath nozzle and terminates where the inside diameter of the sheath nozzle is widest. This can be checked by injecting a concentrated fluorescent dye (preferably the dye to be used in subsequent experiments) through the sample stream and focusing the laser beam somewhere near the signal detection point such that fluorescent light piping occurs right up to the sample nozzle tip. Figure 3 shows an example of sheath and sample nozzles both in and out of correct adjustment. Take care with plastic sheath nozzles not to melt the plastic when looking for a light piping effect. 3. Purging: Every flow cytometer is equipped with a liquid exhaust port well above the sample injection nozzle, which can be opened to release trapped air bubbles. These bubbles must be exhausted before a stable flow and droplet formation can be established. 4. Sample pressure: Too high a sample pressure can cause the sample stream to balloon, as shown in Fig. 3, and leads to a broad final sample stream and consequently a high chromosome peak CV. Most commercial flow cytometers are equipped with a simple pressurized sample delivery system, which is adequate but sometimes oversensitive to adjustment when low sample to sheath pressure differentials

Chromosomes, Analysis and Sorting SHEATH LIQUID I

537

Widest . Sheath Diameter

.

Fig. 3. The example shown is for a glass nozzle system for in air flow where the laser beam is focused on the edge of the sheath nozzle to produce light piping in the concentrated dye sample. On the left, the sample nozzle is shown misaligned, and a high sample pressure is also shown to be causing ballooning of the emerging sample stream. On the right, the dye sample emerges at an acceptable relative pressure and passes centrally down to the in air detection point. Alignment and sample pressure are equally important for cuvette detection systems.

are needed. A more satisfactory way of controlling sample flow is to use a motorized syringe driver capable of delivering less than 1 mL from a l-mL syringe. A cooling jacket around the syringe will maintain the chromosomes in good condition throughout an experiment. 5. Sample stream coating: A stable flow karyotype will be achieved more rapidly when, prior to the introduction of a chromosome sus-

Fantes and Green

pension, a dummy sample of buffer and dye (at twice the final concentration) is injected through the sample stream for approximately 5 min. Using a dummy sample between different chromosome suspensions also helps to flush out remaining chromosomes from previous samples. Generally, flow cytometers have a sheath liquid pressure of about one atmosphere. Under these conditions, a chromosome suspension containing lOchromosomes/mL flowing at 0.4 mL/h, which leads to a flow rate of about 1000 chromosomes/s, should result in a wellresolved flow karyotype. A slower sample flow rate may produce even better resolution, and depending on the quality of cytometer adjustment and the intensity of the laser beam(s), a faster flow rate may not necessarily spoil the resolution. As high a flow rate as possible should be achieved for chromosome sorting experiments.

3.5.2. Beam Alignment

1. Single beam experiments: It is important that the laser beam passes precisely along the optical axis of the focusing lens system. Check this by first removing the forward light scatter detector and then adjust the laser beam to pass through an aperture at the geometric center of the final focusing lens, and at the same time check that the beam is incident on the center of a screen some distance away from the stream axis. Swing the liquid stream to one side for this adjustment. A mark, known to be on the optical axis, on an adjacent wall often provides a useful target for beam alignment checking. Swing back the liquid stream and adjust the beam focus at the stream to be at its narrowest by observation through the stream viewing optics. 2. Dual beam experiments: Here, the lower wavelength beam, which in our example is the ultraviolet beam, is aligned in a similar way to the single beam experiments. The higher wavelength beam is aligned sufficiently off axis to allow the spherical and chromatic aberrations to compensate and bring the two beams into focus along the liquid stream axis. Figure 4 shows how the longer focal length of the higher wavelength beam and the shorter focal length of off axis rays produce the desired alignment of focal points.

3.5.3.

Fluorescence Detection

To optimize the detection of each fluorescent color, it is important to select suitable long and short pass filters. The following combinations

Chromosomes, Analysis

and Sorting
Liquid Stream

539

Fig. 4. The focusing arrangement of an ultraviolet (U.V.) and a blue 458 nM wavelength beam is shown. A converging lens that could be cylindrical or convex focuses the U.V. beam on axis and the 458 nM beam at a slight angle and off axis. The scale of distances and beam widths are somewhat exaggerated in the figure, as are the chromosomes in the liquid stream.

were used for the specific fluorescent dyes described here: (a) Hoechst 33258475 nM long pass + 550 nM short pass, (b) Chromomycin A3-515 nM long pass.
3.5.4. Final Adjustment

The final adjustment of laser beam, stream position, flow rate, and photomultiplier detector positioning needs an actual sample of flowing objects. Standard practice is to use a suspension of fluorescent microbeads. This is not recommended prior to chromosome analyses and sorting, since there is a risk of chromosome aggregation around stray microbeads left behind after the tuning process. Use instead a portion of the chromosome suspension. The distribution of signal pulses appearing on the oscilloscope will soon become familiar, particularly the prominent cluster of signals, all with the same peak height, arising from the human chromosome groups 9 to 12. Optimize the peak height and minimize the peak width of the signal pulses by adjusting the optical components and stream position in an ordered fashion. Readjustment of the stream position between adjustment of each optical component will usually ensure a steadily improving signal height and width.

540

Fantes and Green 3.5.5. Sorting

A working day will, as a rule, produce lo6 sorted chromosomes. Reaching this target depends on a clean chromosome suspension, a steady flow rate of 1000 to 2000 chromosomes/s, and a degree of stability of the live parts of the flow machine. The purity of each sorted fraction will depend on the resolution of chromosome peaks seen in the flow karyotype and on prior identification of the chromosome or chromosomes appearing in particular peaks of interest. Identification requires a sorted fraction of about 60000 chromosomes into 0.25 mL of sheath buffer followed by the procedure described in the chromosome identification section. Contamination of a sorted fraction with chromosomes from adjacent peaks can be minimized to a limited extent by drawing narrow sort windows, but the researcher must be sure that the extra purity gained is justified by a possible further days sort to accumulate the required quantity of chromosomes. It is usually possible to be occupied on another task while each sample of chromosome suspension is sorted, but it must be possible to check frequently on the relative positioning of sorting windows, chromosome peaks, and the droplet stability. No amount of care and attention to the machine will improve sorted fraction purity if there are large numbers of undividing nuclei or large quantities of debris in the original suspension.

4. Notes
1. The starting point for a good chromosome preparation must be a rapidly dividing cell culture with very few dead cells and free from bacterial or mycoplasma contamination. Mycoplasma contamination will cause the chromosomes in the final suspension to stick together and form large clumps. 2. It is important to treat the cells gently during the preparation; centrifuge at low speeds, resuspend the cell pellet by tapping the tube not by vortexing, and ensure that all the cells are uniformly exposed to hypotonic and buffer solutions by maintaining a single-cell suspension. important; 3. The concentrationof cells to buffer 82inmethods,stage7is if the amount of B2 is decreased, cell breakage will be incomplete. 4. The differential centrifugation step described in stage 9 will remove most of the contaminating nuclei. Centrifugation of the chromosomes at higher speeds to concentrate them or remove debris will increase the number of clumps and degraded chromosomes giving

Chromosomes, Analysis and Sorting

541

5. 6. 7.

8.

9. 10.

poor resolution. It is better to start again from a culture with a higher mitotic index. Do not attempt to analyze very concentrated chromosome suspensions. Staining irregularities will occur as well as excessive signal coincidences and nozzle blockages. Dilute with buffer B2. Addition of 10 mM sodium citrate and 10 mM sodium sulfite (11) to stained chromosomes has been reported to increase the resolution of dual beam flow karyotypes. Chromomycin A3 intercalates into the DNA, and this has sometimes reduced the efficiency of subsequent DNA manipulations. This problem has been overcome by dialysis of the sorted chromosomes against two changes of BI to remove the stain. Chromosomes stabilized by polyamines are not suitable for chromosome-mediated gene transfer into other cells. An alternative buffer containing 15 mM Tris, 3 mM calcium chloride, and digitonin should be used (12). Chromosomes stabilized with polyamines are also not suitable for use with antibodies against chromosomal proteins; other buffers are recommended (23 ). We have found that the best routine for cleaning the tubes of the flow machine involves washing first with warm detergent solution to remove residual chromosomes without fixing them to the tube walls, and then sterilizing with activated glutaraldehyde (Cidex). Extensive washing with sterile distilled water is then necessary to remove all traces of Cidex. Using 70% ethanol, which is commonly used by flow cytometer operators for flushing and sterilizing, will fix residual chromosomes to the tube walls. The method recommended here avoids this problem.

References
karyotyping of human chromosomes by dual beam flow cytometry. Proc. N&Z. Ad. SC!. USA 79,7876-7880. 2. Green, D. K., Fantes, J. A., Buckton, K. E., Elder, J. K., Malloy, P., Carothers, A., and Evans, H. J. (1984) Karyotypingand identification of human chromosome polymorphisms by single fluorochrome flow cytometry. Hum. Genef. 66,143-146. 3. Harris, P., Boyd, E., Young, B. D., and Ferguson-Smith, M. A. (1986) Determination of the DNA content of human chromosomes by flow cytometry. Cyfogenet. Cell Genet. 41,14-21.

1. Langlois, R. G., Yu, L. C., Gray, J. W., and Carrano, A. V. (1982) Quantitative

542
4.

Fantes and Green

Harris, I., Cooke, A., Boyd, II., Young, B. D., and Ferguson-Smith, M. A. (1987) The potential of family flow karyotyping for the detection of chromosome abnormalities. Hum. Genet. 76,129-133. 5. Krumlauf, R.,Jeanpierre, M., and Young, B.D. (1982)Construction and characterisation of genomic libraries from specific human chromosomes. Proc. NutZ. Ad. Sci.
USA 79,2971-2975.

6. Cooke, H. J.,Fantes,J.A., and Green, D. K. (19831Structure and evolution of human Y chromosome DNA. Differentiation 23, S48-S55. 7. Lebo, R. V., Gorin, F., Fletterick, R. J.,Kao, F., Cheung, M., Bruce, B. D., and Kan, Y. W. (1984) High resolution chromosome sorting and DNA spot-blot analysis assign McArdles syndrome to chromosome 11. Science 225,57-59. 8. Van de Sande, V. H., Lin, C. C., and Dengau, K. V. (1979) Clearly differentiated and stable chromosome bands produced by a spermine bis-a&dine; a bifunctional analogue of quinacrine. Exp. Cell Res. 120,439-444. 9. Sillars, R. and Young, B. D. (1981) A new method for the preparation of metaphase chromosomes for flow analysis. J Histochem. Cytuchem. 29,74-78. 10. Melamed, M. R., Mullaney, P. F., and Mendelsohn, M. L. teds.) (1979) Flow Cytometry and Sorting (Wiley, New York). II. Van den Engh, G., Trask, B.,and Gray, J.W. (1987) Improved resolution of bivariate flow karyotypes by manipulation of staining conditions. Cytometry Suppl. 1,3. 12. Porteous, D. J., Morten, J. E. N., Cranston, G., Fletcher, J. M., Mitchell, A., van Heyningen, V., Fantes, J. A., Boyd, P. A., and Hastie, N. D. (1986) Molecular and physical arrangements of human DNA in HRASl-selected, chromosome mediated transfectants. Mol. Cellular Biol. 6,2223-2232. 13. Trask, B., Van den Engh, G., Gray, J. W., Vanderlaan, M., and Turner, B. (1984) Immunofluorescent detection of histone 2B on metaphase chromosomes using flow cytometry. Chromosoma 90,295-302.

Chapter Flow

40

Cytometry

Michael G. Ormerod and Patrick R. Imrie

1. Introduction
There are several commercial flow cytometers on the market. They all operate on the same basic principle, but there are important differences in their design, and the methods for alignment, and so on, depend on the type of instrument. In this chapter, therefore, we describe procedures for preparing samples and limit description of the flow cytometry to comments about important features of the analysis. One of the exciting applications of the techniques is the ability to sort cells. Good separation of cells demands good preparation of the sample and adequate analysis. Although procedures for sorting cells are not given for reasons given above (see Chapter 41, this vol.), the methods described in this chapter and the comments about analysis are applicable. A description of the basic principles involved in the design and operation of a flow cytometer can be found in references 1 and2. An associated microcomputer is important, since multiparametric data cannot be analyzed adequately without this accessory. The comments on analysis assume that there is an adequate facility for computing. Flow cytometry measures properties of cells in a flow system. It is therefore a prerequisite of the method that the particles (cells or nuclei) are in suspension and free of clumps. The quality of the information obtained will largely depend on the quality of the preparation of the sample. This comment applies to all the methods described. 543

544

Ormerod and Imrie

2. Materials
1. Phospate buffered saline (PBS). 8.5 g NaCl, 1.07 g Na,HPO, (or 2.7 g Na,Hl?O,~lZ~O), 0.39 g NaH,PO, in 1 L of water. 2. Ribonuclease A (RNase). Type III-A. Stock solution: 1 mg/mL in PBS. Store at -2OOC. 3. Collagenase. Type 1A. Stocksolutionl mg/mLinPBS. Storeat-20C. 4. Pepsin. Lyophilized powder from stomach mucosa. Prepare solutions as needed by dissolving the enzyme in 0.9% NaCl and adjusting the pH to 1.5 with HCl. 5. Propidium iodide (PI). Stock solution: 100 pg/mL in PBS. Store at 4C. 6. Flourescein diacetate (FDA). Stock solution: 1 mg/mL in acetone. Store at -20C. 7. Stain-detergent solution for preparing nuclei from fresh tissue: 1 g trisodium citrate, 564 mg NaCl, 300 PL Nonidet P-40,10 mg ethidium bromide in 1 L distilled water. Just before use, dissolve 1 mg of RNase in 100 mL solution. 8. Solution for diluting antibodies in the Bromodeoxyuridine method: 1% BSA, 5% normal goat serum, 0.5% Tween 80, and 20 mM EDTA in PBS, pH 7.5. All the figures shown in this chapter were recorded using an Ortho Cytofluorograf 50H equipped with a 50 mw Lexel argon-ion laser and a 5W Coherent laser driving a dye laser. The instrument was interfaced to an Ortho 2150 computer system, and the figures were taken directly from the computers graphics screen.

3. Methods
3.1. Labeling Cells with Antibodies
3.1.1. Analysis of a Single Antigen (See Notes l-3) The basic method for cell labeling is straightforward-a typical procedure is included in the method for studying DNA labeled with bromodeoxyuridine. The method of analysis given below is suitable for cells labeled with a fluorescein conjugated antibody. It is assumed that propidium iodide (PI) has been added in order to distinguish dead cells. In cell mixtures, light scatter can be used to distinguish between different types of cell (seeFig. 1.).

Flow Cytometry

545

Fig. 1. Human keratinocytes labeled with a mouse monoclonal antibody (23.10) that reacts with basal cells (3) followed by fluoresceinated goat anti-mouse immunoglobulin; PI has been added. A shows a histogram of red fluorescence; a region 1 has been set to include cells that have taken up PI and are therefore fluorescing red. These are excluded from further analysis. B is a cytogram of light scattered orthogonally against that scattered over a narrow forward angle. Each dot represented a cell recorded by the instrument. Region 1 has been set to include the smaller cells and region 2 to include the larger. The green (antibody) fluorescence of the small cells is shown in histogram C and that from the large in histogram D. Cells supplied by Dr. J. Burman.

1. With a flow cytometer employing an argon-ion laser, tune the laser to 488 nm and set up the instrument to measure four parameters-light scattered orthogonally, and in a forward direction, green (fluorescein) and red (PI) fluorescent light. The green fluorescence should be selected using a band pass filter centered on 520 nm and the red with a long pass filter with a cutoff close to 600 nm. 2. Display red fluorescence and set a gate to exclude labeled cells from further analysis. 3. Display a cytogram of orthogonal vs forward light scatter, and set a region to include single cells of the desired type and to exclude, as far as possible, clumps and debris. 4. Display a histogram of green fluorescence of the single cells of interest.

546

Ormerod and Imrie

RED

FLUOR.

RED

FLUOR.

Fig. 2. Single cells prepared from the human mammary gland and labeled with rabbit antiepithelial membrane antigen (6) and mouse anti-common acute lymphoblastic leukemia antigen (7) followed by fluoresceinated goat antirabbit and phycoerythrin conjugated sheep antimouse Ig. The single cells have been selected using light scatter. A is a cytogram showing green vs orange fluorescence; inB the computer has been used to correct for overlap between the green and orange channels. Cells prepared by Dr. M. OHare.

3.1.2. Simultaneous Analysis of Two Antigens (See Notes 4 and 5) If two antigens are to be visualized, it is customary to label one antibody with fluorescein and the other with either phycoerythrin (PE) (4) or Texas red (a derivative of rhodamine) (5). PE can be excited by the same laser line as fluorescein. There is some overlap between the emission spectra of these two dyes (that is, fluorescein will be detected in the PE channel and vice versa), and a correction must be applied either electronically or by the computer. Figure 2 shows an example of this. The procedure for the simultaneous analysis of fluorescein and phycoerythrin is given below. 1. Tune the argon-ion laser to 488 nm, and set up the instrument to measure light scatter (orthogonal and forward), green fluorescence at 520 nm and orange fluorescence at 580 nm. 2. Display a cytogram of the light scatter, and set a suitable region to include the cells of interest. 3, Display a cytogram of green vs red fluorescence of the selected cells. 4. Display a cytogram of corrected green vs orange fluorescence.

Flow Cytometry

3.2. Analysis

of DNA

A variety of methods have been published for measuring the DNA content of cells (8). Those used in this laboratory are given below. The important features of the analysis are to ensure that the profiles are as sharp as possible (as measured by the coefficient of variation acrosss the peak), and that two or more cells stuck together are excluded from the analysis. The width of the peaks arising from cells in Gl or G2 of the cell cycle will depend on the flow rate, and this should be kept aslow as possible. For this reason, the concentration of cells should be between 5.105and 106/mL. Two cells or nuclei stuck together will have the same DNA content as a single cell in G2 of the cell cycle, and the two must be distinguished if the DNA histogram is to reflect accurately the state of the cell cycle. In instruments that focus the laser beam down to an ellipse with a cross-section of 5 pm or less, the difference can be resolved by analyzing the shape of the signal generated as the particle crosses the laser beam. The width of the signal generated is given by the sum of the width of the laser beam and nuclear diameter. As a cell progresses through its cycle, while the content of DNA doubles, the nuclear diameter will not increase by more than 25%. Consequently, there is only a small increase in the signal width. Because of the nature of the flow system, two particles stuck together will align one behind the other; their combined diameter will be at least two nuclear diameters. This is illustrated in Fig. 3. Using the Ortho Cytofluorograph, the width of the signal is best measured by comparing the signal height to its area. This can be displayed on a cytogram on which the clumps can be distinguished from single cells. A region can be selected to include only the latter (Fig. 4). On instruments in which the laser beam is focused to a circular spot of diameter 60 PM or more (for example the FACS series made by Becton-Dickinson), this type of analysis is not possible. Whenmeasuring ploidy, the concentration of cells from one sample to another should be kept the same so that the dye/DNA ratio does not vary too much. A standard sample is also needed. Some workers use either chicken or trout erythrocytes (9). These are not entirely satisfactory, since their DNA content is considerably lower than that from mammalian cells. We use ethanol-fixed sheep lymphocytes stained in parallel, and run a sample before and after the unknown. If there is any doubt or a small change in ploidy is suspected, the standard and the unknown are mixed and rerun. It is often possible to distinguish the sheep lymphoctyes from other cells on the basis of light scatter so that the two populations can be

548

Ormerod and Imrie

Iarer J

Ln In In
limo

Gl

G2

2xQl

Fig. 3. The signal shape generated at the photomultiplier by the fluorescence froma nucleus in Gl, a nucleus in G2, and two nuclei inG1 stuck together. The area under the curve generated by a nucleus in G2 and that from two nuclei in Gl are the same and equal to twice that of the curve from a single nucleus in Gl. However, the peak height of the G2 signal is twice that from the clumped nuclei.

selected from the mixture, the histograms of their DNA content displayed separately, and their peak channels compared. When samples of cancerous tissue are studied, there are usually some normal diploid cells present, and these can act as an internal standard. Having recorded a DNA histogram, a computer program is often used to deconvolute the curve into the separate contributions from the different phases of the cell cycle (GO/Gl, S, and GZ). A variety of programs have been written; we have recently published an aligorithm that is particularly fast and handles perturbed histograms (20). Measurement of the DNA Content of Cultured Cells and Cells in Suspension (See Notes 6-8) 1. Prepare a suspension of single cells in 200 PL of PBS. 2. Add vigorously 2 mL of ice-cold 70% ethanol, 30% PBS. Leave for at least 30 min in the cold. 3.2.1.

Flow Cytometry

549

Fig. 4. Sheep lymphocytes fixed in ethanol, treated with RNAase and labeled with PI. The cytogram in A shows the signal peak displayed against the signal area. The region was drawn to include single cells and exclude clumps. Note the appreciable contribution madebytwocellsstuck together(arrowedincytogram1. ThehistogramB shows theDNA histogram (signal area) of the single cells only, and C shows the histogram that resulted when all the cells were included. 3. Harvest the cells by centrifugation.

4. Add 100 PL of RNase (1 mg/mL)

cubate at 37OC for 30 min.

Resuspend in 800 clr, of PBS. and 100 PL of PI (400 pg/mL). In-

5. Analyze. of the DNA Content of Fresh Tissue Several methods for extracting nuclei from fresh tissue have been published. We use one adapted from that of Petersen (II). 1. Take 200 mg of wet weight or more of tissue. Mince with scalpels in tissue culture medium.
2. Filter through a stainless-steel tea-strainer.

3.2.2. Measurement

3. Centrifuge at 8OOgfor 5 min. 4. Aspirate and discard the supernatant, and resuspend the pellet in 10 mL of stain-detergent solution.

550

Ormerod and Imrie

5. Mix on a rotator at 4OCfor 1-4 h. The exact time will depend on the tissue used, and must be found by trial and error. 6. Filter through 60qM pore nylon mesh. 7. Analyze.

3.2.3. Measurement of the DNA Content of Fixed Tissue Embedded in Parafin Wax (See Notes 9 and 10)
Tissue removed during an operation or an autopsy is often fixed in formalin and embedded in paraffin wax prior to cutting tissue sections for staining and histopathological examination. These blocks are stored for many years, and the development of a method for extracting nuclei from them has given flow cytometrists access to much archival material (12). 1. Cut three or four 40-pm sections from each block. Store in xylene in glass containers. (Check the lids. Xylene will dissolve the rubber washers often found in screw-top lids.) 2. Harvest tissue by centrifugation in glass tubes. Reextract twice with xylene to complete removal of paraffin wax. 3. Wash the tissue twice in ethanol, once in 50% (v/v) ethanol and twice in PBS. 4. If the tissue is particularly fibrous, incubate for 1 h at 37OCin 1 mg/mL collagenase in PBS. 5. Centrifuge and resuspend in 0.9% NaCl, 0.5% pepsin adjusted to pH 1.5 with HCl. Incubate at 37OC for 1 h. 6. Pass through a 23-gage needle to break up clumps. Wash once in IM phosphate buffer, pH 7.0 and once in PBS. 7. Finally, resuspend in PBS, containing 100 pg/mL RNase, 10 pg/mL PI, and incubate for 30 min at 37C. 8. Analyze, measuring orthogonal and forward light scatter and peak and area of the red (PI) fluorescent signal.

3.2.4.

Combined Analysis

of a Surface Antigen

and DNA

This can be used to observe any cell cycle dependence of expression of an antigen. 1. Prepare a suspension of 1-2 x lo6 single cells. 2. Incubate in an appropriate dilution of primary antibody (for example, a murine monoclonal antibody) in tissue culture medium (TCM) plus 10% fetal calf serum (FCS) at 0C for 1 h. 3. Wash the cells by centrifugation twice in TCM plus FCS.

Flow Cytometry

551

4. Incubate the cells in an appropriate dilution of fluorescein conjugated second antibody (for example, goat antimouse Ig) in TCM plus 10% FCS for 1 h at OOC. 5. Wash the cells once in TCM plus 10% FCS; once in PBS. 6. Resuspend the cells in 200 FL PBS, and take up and down through a pipet tip to ensure a good suspension of single cells. 7. Add vigorously 2 mL of ice-cold 70% ethanol, 30% PBS. Leave at least 30 min in the cold. 8. Harvest the cells by centrifugation. Resuspend in 800 FL of PBS. 9. Add 100 FL of RNase (1 mg/mL), 100 PL of PI (400 pg/mL). Incubate at 37C for 30 min. 10. If necessary, pass the cells through a 25-gage needle. 11. Analyze on the flow cytometer using an argon-ion laser tuned to 488 nm and recording light scattered in a forward angle, green (antibody) fluorescence (520 nm) and red (PI) (>630 nm) fluorescence on peak and area (seeabove). On an instrument capable of recording 5 parameters, also measure orthogonal light scatter. 12. On analysis, gate out any debris and large clumps using light scatter, select single cells using peak vs area for the red fluorescence, and then display green vs red fluorescence. If the antibody staining is weak, it may be necessary to correct for red fluorescence in the green channel. An example of this analysis is shown in Fig. 5. Combined Analysis of Cytoplasmic or Nuclear Antigen and DNA If the antigen is inside the cell, the cell has to be fixed before incubating with the first antibody The fixative used should be chosen to avoid damage to the antigen being studied. The DNA in a fixed cell is susceptible to attack by nucleases, and this can cause problems. It can be avoided by a clean technique, adding EDTA to the buffers, and incubating the cells on ice. A procedure for observing progesterone receptors is given (worked out in collaboration with Dr. J. Mansi). 3.2.5. 1. Prepare a suspension of l-2 lo6 single cells in 910 PL of PBS. 2. Add 90 PL of 40% formaldehyde solution. Leave at room temperature for 15 min. 3. Harvest the cells by centrifugation, and wash once in PBS. 4. To the cell pellet, add 2 mL of absolute methanol at -2OOC. Leave for 4 min.

552

Ornaerod and Imrie

Fig. 5. Cultured cells from the human mammary gland labeled with a murine monoclonal antibody against CALLA (71 followedby fluoresceinated sheep antimouse Ig, fixed in 70% ethanol, and incubated with RNase and PI. Cytogram A shows the signal peak against area for red (DNA-PI) fluorescence. The single cells have been included in Region 1 and all other particles havebeenexcluded from further analysis. CytogramB shows the orthogonal vs forward light scatter. The larger cells in the elliptical fluorescence displayed against red fluorescence in cytogram C. Cells prepared by Mrs. S. Davies.

5. Centrifuge, remove the methanol, and add 2 mL of acetone at -2OOC. Leave for 2 min. 6. Centrifuge, remove the acetone, and wash the cells twice in PBS, finally resuspending in 100 PL of PBS, 10 mM EDTA. Take the cells up and down through a pipet tip to ensure a suspension of single cells. 7. Add 200 PL of 5% normal goat serum in PBS/EDTA. Incubate on ice for 15 min. 8. Add 200 PL of monoclonal rat antibody (antiprogesterone receptor, Abbott Diagnostic). Add an irrevelant rat antibody to the control. Incubate for 1 h at 0C. 9. Wash the cells twice in PBS/EDTA/O.l% bovine serum albumin (BSA).

Flow Cytometry

553

10. Resuspend final pellet in a fluoresceinated goat antirat Ig diluted appropriately in PBS/EDTA/BSA. Incubate for 1 h at OOC. 11. Wash the cells twice in PBS/EDTA, finally resuspending in 800 of yL PBS. 12. Add 100 PL of PI solution (100 ug/mL), and 100 PL of RNase solution (1 mg/mL). 13. Incubate for 15 min at room temperature before analysis by the same procedure described in the preceding section.

3.2.6. Cells with DNA Labeled with Bromodeoxyuridine (Brd U)

This is another example of the method given above. It is also described because of its importance (seeNotes 11 and 12). If BrdU is injected into an animal or added to a cell culture, it will be incorporated into the DNA of cells synthesizing DNA. The BrdU can be visualized subsequently by reaction with an antibody directed against bromouracil. This is a powerful technique for studying the cell cycle and perturbations therein caused by radiation and cytotoxic drugs. It can also be used to estimate the growth fraction of a culture. Most antibodies available will not react with BrdU in native DNA. It is necessary to disrupt the structure of the DNA; one of several published methods can be used (seereference 13). 1. Prepare a suspension of about 2 x lo6 cells. 2. Resuspend in 200 PL of PBS pipeting up and down to ensure a good suspension of single cells. 3. Vigorously add 2 mL of ice-cold 70% ethanol. 4. Stand on ice for at least 15 min. (The cells can be stored at this stage for several days at 4OC.) 5. Harvest the cells by centrifugation and wash once in PBS. 6. Resuspend the cell pellet in 2 mL of 2M HCl. Incubate at 37OC for 30 min. 7. Harvest the cells; wash once in 5 mL of 1Mphosphate buffer, pH 7, and once in PBS. 8. Resuspend cells in 200 PL of incubation buffer (seeMaterials section). Add 10 PL of rat anti-BU (ICR2, Seralab, Ltd., Crawley). Incubate on ice for 1 h. 9. Wash the cells twice in PBS, 0.5% Tween 80. 10. Resuspend in 200 PL of fluoresceinated, goat antirat Ig diluted appropriately in incubation buffer. Incubate on ice for 1 h. 11. Wash once in PBS/Tween and once in PBS.

554

Ormerod and Imrie

12. Resuspend in 450 PL of PBS. Add 50 PL of PI solution (100 pg/mL). 13. Incubate at 0C for 15 min. 14. Analyze according to protocol described above for combined analysis of surface antigen and DNA. A typical result is shown in Fig. 6.

3.3. Monitoring

the Ebctropermeabilization

of Cells

Electroporation is the permeabilization of cells by a high voltage pulse. On subsequent incubation, if the correct conditions are used, the cells will reseal their plasma membranes. The method is an efficient way of introducing foreign DNA into cells. A variety of parameters require optimization. For effective transfection, it is necessary to know whether holes have been punched in the plasma membrane, for how long they stay open, how efficiently they reseal, and the effects of this treatment on cell viability. Different types of cell require different conditions for electroporation in order to optimize these conditions (24,X). The method is based on the observation that propidium iodide is excluded by an intact plasma membrane. If it is added immediately after electropermeabilization, cells whose membrane has been ruptured will fluoresce red. After the cells have been incubated in warm medium, fluorescein diacetate is added. This is taken up by live cells and hydrolyzed to fluorescein, which fluoresces green and will be retained by cells with an intact membrane. The details of the flow cytometric method are given. For information about the method of electroporation, seeref. 15. 1. 2. 3. 4. 5. Electroporate a precooled suspension of single cells at 0C. Add PI to a concentration of 10 pg/mL. Incubate at 0C for 10 min. Dilute the cells in culture medium at 37OC. Incubate for 10 min. Wash the cells in PBS and resuspend in 1 mL of PBS. Add FDA to a final concentration of 10 ng/mL. Incubate at room temperature for 10 min. 6. Analyze, recording orthogonal and forward light scatter, red and green fluorescence. Gate on the light scatter from single cells and display green vs red fluorescence. 7. In the final cytogram, three classes of cells are distinguished: fluorescein positive, PI negative (unpermeabilized); fluorescein positive, PI positive (permeabilized and resealed membranes); fluorescein negative, PI positive (permeabilized and failed to reseal) (Fig. 7 and Note 13).

Flow Cytometry

555

Fig. 6. Chinese hamster V79 cells labeled in culture for 30 min with 10 @I BrdU. The cells were labeledwith a rat anti-BU antibody and PI and analyzed as in Fig. 5. An isometric display of the final histogram of green vs red fluorescence is shown. The cells in S phase of the cell cycle have incorporated BrdU and are labeled with the antibody. Cells prepared by Mrs. P. Loverock.

4. Notes
1. Dead cells (including fixed cells) tend to take up antibodies nonspecifically; this can be reduced by adding an extraneous protein to the solutions of antibody and taking particular care in washing the cells. 2. When antigens on the surface of the cell are studied, it is usual to label live cells. In such cases, the cells should be kept cold; all incubations should be at 0C and all buffers kept on ice. At higher temperatures, the antigens in the plasma membrane may become crosslinked by antibody (called capping) and then be pinocytosed. 3. Since clumps of cells and dead cells may label nonspecifically, they should be excluded as far as possible from the analysis by measuring scattered light. Dead cells can also be distinguished from their live counterparts by adding propidium iodide (PI) at a typical concentrationof 10 pg/mL. This dyeis excluded by theintact plasma membrane in live cells and fluoresces red when bound to nucleic acid. For some types of cells, such as lymphocytes, the addition of PI is unnecessary

556

Ormerod and Imrie

t FLUORESCEIN 1 l PROPIDIUH IODIDE I

Fig. 7. A fibroblastic rat cell line electoporated at 5 kV/cmin low conductivity medium (15. The cytogram shows green (fluorescein) vs red (PI) fluorescence. Three groups of cells can be identified: fluorescein +ve, PI -ve; fluorescein +ve, PI +ve; and fluorescein -ve, PI +ve. Cells supplied by Drs. T. McMillan and M. J. OHare.

since dead cells can be distinguished by their reduced light scatter. 4. Texas red is excited by orange light and fluoresces red (peak emission at 620 nm). A second laser must be used, either a krypton laser tuned to 568 nm or a dye laser using rhodamine 6G and tuned to 580 nm. This has the disadvantage of using additional expensive hardware. The advantage is that the two laser beams are separated spatially (usually set between 30 and 130 pm apart), so that the signals from them are resolved both optically and in time; cross-talk between the red and green channels is usually undetectable. With a dual laser cytometer, with the combination of either fluorescein, phycoerythrin and either Texas red or allophycocyanin, three antibodies can be detected simultaneously (26,Z7). 5. When measuring two antibodies, PI may be added to exclude dead cells (28). An anticipated difficulty in correcting for overlap between PI fluorescence and that from PE or Texas red should not be a problem since only the excluded (dead) cells are stained with PI. 6. If there are an excessive number of clumps, these may be reduced by passing the sample through a 25-gage hypodermic needle. 7. If PI or ethidium bromide is used to label DNA, the profiles may be broadened because of the presence of residual RNA, particularly in

Flow Cytometry

557

8. 9.

10. 11. 12.

13.

larger cells. This can usually be resolved by replacing the stock solution of RNase. In particularly difficult cases, it may be necessary to add mithramycin (final concentration: 50 pg/mL, 7.5 mM MgCl,) and excite at 457 nm (mithramycin binds to DNA but not RNA) (19). Another problem is the degradation of DNA by contaminating nucleases. Attention should be paid to the cleanliness of all equipment used. Storage of prepared samples is not generally advisable. Occasionally it is found that the DNA in the extracted nuclei is badly degraded. This appears to be related to the treatment of the tissue prior to fixation, so little can be done about it. Sometimes light scatter can be used to exclude the more badly degraded material. The amount of PI bound is affected by the conditions of fixation of the original tissue (20) and account should be taken of this when determining ploidy. Failure to stain the BrdU-labeled DNA usually relates to a failure to disrupt the DNA sufficiently. Sometimes there are problems with background stain observed by staining cells that have not been labeled with BrdU. Attention should be paid to procedures for washing the cells-increasing the number of washes and/or the protein concentration in the buffer used. In Fig. 7, the unpermeabilized cells were present because a small number of cells were retained in the tip of the pipet and were not subjected to the electric field. They served as a useful built-in control.

Acknowledgments
This was supported by a joint program _ work _ - grant from the Cancer Research Campaign and the Medical Research Council.

References
2. Shapiro, H. M. (1985) PracticalFlow Cytometry (Liss, NY, New York). 2. Dilla, A. V. D., Dean, P. N., Laerum, 0. D., and Melamed, M. R. feds.1 (1985) Flow Cytometry: hsfrumentution and Dafa Analysis (Academic, London). 3. Gusterson, B. A., McIlhinnney, R. A. J., Patel, S., Knight, J., Monaghan, P., and Ormerod, M. G. (1985) The biochemical characterisation and immunocytochemical characterisation of an antigen on the membrane of basal cells of the epidermis. Difierenfidion 30,102-110. 4. Oi, V. T., Glazer, A. N., and Stryer, L. (1982) Fluorescent phycobiloprotein conjugates for analyses of cells and molecules. J. Cell Biol. 93,981-986. 5. Titus, J. A., Haugland, R., Sharrow, S.O., and Segal, D. M. (1982) Texas red, a hydro-

558

Ormerod and Imrie

philic red-emitting fluorophore for use with fluorescein in dual parameter flow microfluorometric and fluorescence microscopic studies. J. ImmunoI. Methods 50, 193-204. Sloane, J.P. and Ormerod, M. G. (1981) Distribution of epithelial membrane antigen in normal and neoplastic tissuesand its value in diagnostic tumor pathology. Cancer 47,1786-l 795. Gusterson, B. A., Monaghan, P., Mahendron, R., Ellis, J., and OHare, M. J. (1986) Identification of myoepithelial cells in human and rat breasts by anti-commonacute lymphoblastic leukemia antigen antibody A12. J. NUB CancerInst. 77,343349. Taylor, I. W. and Milthorpe, B.K. (1980)An evaluation of DNA fluorochrome, staining techniques and analysis for flow cytometry. J. Hisfochem. Cytochem. Z&12241232. Vindelov, L. L., Christensen, I. J., and Nissen, N. I. (1983) Standardisation of highresolution flow cytometric DNA analysis by the simultaneous use of chicken and trout red blood cells as internal reference standards. Cytometry 3,321-328. Ormerod, M. G., Payne, A. W. R., and Watson, J.V. (1987) Improved program for the analysis of DNA histograms. Cyfomefy 8,637-641. Petersen,S. E. (1985) Flow cytometry of human colorectal tumors: nuclear isolation by detergent technique. Cytomefy 6,452-460. Hedley, D. W., Friedlander, M. L., Taylor, I. W., Rugg, C. A., and Musgrove, C. A. (1983)Method for analysisof cellular DNA content of paraffin-embedded pathological material using flow cytometry. J. Histochem.Cytochem.31,1333-1335. Gray, J. W. and Mayall, B. H. (eds.) (1985) Monoclonul Antibodies Against Bromode Oxyuridine (Liss, New York). OHare, M. J., Asche W., Williams, L., and Ormerod, M. G. (1989) Optimisation of transfection of breast epithelial cells by electroporation. DNA, in press. OHare, M. J.,Ormerod, M. G., Imrie,P. I., Peacock,J. H., and Asche,W. (1989)Electropermeabilisation and electrosensitivity of different types of mammalian cells, in Electroporufion and Elecfrojksion in Cell Biology (Neumann, E., Sowers, A. E., and Jordan, C., eds.), Plenum, New York, in press. Parks, D. R., Hardy, R. R., and Herzenberg, L. A. (1984) Three-color immunofluorescence analysis of mouse B-lymphocyte poplulations. Cytometry 5,159-W. Loken, N. R. and Lanier, L. L. (1984) Three-color immunofluorescence analysis of Leu antigens on human periphereal blood using two lasers on a fluorescence activated cell sorter. Cytometry 5,151-158. Sasaki,D. T., Dumas, S.E., and Engelman, E. G. (1987) Discrimination of viable and non-viable cells using propidium iodide in two color immunofluorescence. Cytomety 8,413-420. Barlogie, B.,Spitzer, G., Hart, J.S.,Johnston, D. A., Buchner, T., Schumann, J.H., and Drewinko, B. (1976) DNA histogram analysis of human hemopoietic cells. Blood 48, 245-258. Schutte, B., Reynders, M. M. J., Bosman, F. T., and Blijham, G. H. (1985) Flow cytometric determination of DNA ploidy level in nuclei isolated from paraffin-embedded tissue. Cyfomehy 6,26-30.

6. 7. 8. 9. 20. 11, 12. 23. 14. 15.

16. 17. 18. 19. 20.

Chapter 41 Fluorescence-Activated Sorting According to Receptor Density

Application for Isolating Transfected Cell Lines
Stella Clark

Cell

1. Introduction
Isolation of cells expressing transfected cDNAcan be extremely difficult and tedious if the efficiency of expression is low (less than l-5%). If the receptor being selected for is a cell surface or integral membrane protein at least partly exposed to the exterior, then it is possible to use the extremely powerful technique of fluorescence-activated cell sorting (FACS) to select these cells from a receptor-negative background. The technique depends on being able to tag the receptor-expressing cells with a fluorescent label. A detector identifies fluorescent cells in a stream of single cells passing through a laser beam. An electric charge is applied to this stream and positive cells can be deflected into a collecting tube. The entire procedure is carried out under sterile conditions, so that the positive cells can be expanded for further sorting and ultimately examination of the expressed receptor. Theoretically, this method could select out a single fluorescent cell from a population of nonexpressing cells; in practice one is looking at a level of detection of l:l@-104. 559

560

Clark

There are two types of fluorescent probes that are used to label receptors on cells. First, one can use a fluorescently labeled ligand, and indeed, several groups have used the FACS to measure receptor numbers and to identify subsets within a cell population based on cell size and receptor expression using this approach (Z-5). However, ligand receptor binding is usually reversible and thus the continued presence of the ligand in solution during a sort would be necessary. As yet, using the ligand to detect transfected receptor expression by FACS has not been reported. Secondly, antireceptor antibodies can be used in conjunction with a fluorescent second antibody. In general, antibodies bind to receptors with higher affinity than ligands, so dissociation of the label during the sort is unlikely. By using a double antibody system, a much higher sensitivity can be achieved. For example, from one transfection, we observed no positive cells above background at the first sort, but by the fifth sort, a large percentage of the cells were expressing the desired antigen (Fig. 1). This method has been used both to isolate cells expressing transfected receptors and to select high expressers for the transferrin receptor (6) and the insulin receptor (7). We have also used this approach to select for EGF-receptor expression (8). This chapter describes the preparation of cells for FACS using a fluorescent double-antibody method.

2. Materials
1. Complete cell medium, e.g., alpha minimal essential media plus 10% v/v fetal calf serum. 2. Phosphate Buffered Saline (PBS): 0.14M NaCl, 2.7 mM KCl, 1.5 mM KH2P04, 8.1 mM Na,Hl?O,. Sterilize. 3. Trypsin solution: 0.05% Trypsin in PBS/O.5 mM EDTA. 4. Earles Balanced Salt Solution (10 x EBSS): Make up as a 10x stock, sterilize, and store at room temperature. 1.2M NaCl, 54 mM KCl, 12 mM NaH$O,, 8 mM MgSO,, 18 mM CaCl,. If using a commercial source, specify without bicarbonate and phenol red. 5. FACS Buffer A: EBSS (1:lO dilution of stock) containing 10 mM Hepes pH 7.4, filter sterilize, and add 2% (v/v) heat inactivated (30 min/ 50C) fetal calf serum. 6. FACS Buffer B: FACS Buffer A without serum. NOTE: For sorting one cell population, about 100 mL of buffers A and B are needed. Make up on day of sort and cool to 4C.

Fluorescence-Activated

Sorting
SORT NUMBER

561

-SCATTER

Fig. 1. FACS Sorting of Chinese Hamster Ovary (CHO) Cells Transfected with EGFReceptor cDNA. G418 resistant CHO cells from one transfection were pooled for the first sort. The cells were prepared for the sort as described in the Method section. Approximately lo7 cells were used in each sort (+Rl) and 10s cells for the negative control (-Rl). A431 cells labeled with Rl were used to align the machine. The data is presented as a series of dot plots with forward scatter on the horizontal axis (allows one to gate out dead cells and cell aggregates) and fluorescence intensity on the vertical axis (measures level of receptor expression). Sort 1 showed no difference between k first antibody. The top 5% of cells were collected and expanded for resorting. From sort 2 onwards, an increasing fraction of the cells showed fluorescence levels above the control, i.e., sort 38% positive; sort 5,56% positive.

7. Antireceptor Antibody (first antibody): In general, use a high titer antireceptor antibody directed against the extracellular domain of the receptor. Use at a concentration of 100'M. For EGF-receptor, the monoclonal Rl (9) is used at a 1:50 dilution. 8. Second Antibody Conjugated to a Fluorescent Probe. Choose an appropriate second antibody. When using Rl as the first antibody, we use goat antimouse IgG conjugated to fluorescein isothiocyanate (FITC). The dilution of second antibody required for best results needs to be empirically determined. We routinely use a 1:20 dilution.

562

Clark

3. Method
The Method describes sorting cells for EGF-receptor expression. The entire procedure is carried out under sterile conditions. 1. Following marker selection of transfected cells (seeChapter 42, this vol.), the resultant colonies are removed from the plates and replated on 150-mm dishes. 2. When the cells become confluent, themedia is removed and the adherent cells are washed once, on the dish, with 10 mL PBS. 3. The cells are removed with trypsin/EDTA (seeNote 1) and washed 3x in 10 mL Buffer A by centrifuging at 600g for 5 min at 4OC. This and all subsequent procedures are carried out on ice. 4. On resuspending the cells (seeNote 2) for the third wash, an aliquot of the suspension is removed for a cell count. The cells from the final wash are resuspended at lO/mL in Buffer A. At least lo7 cells are required for a satisfactory sort. 5. From the sample to be sorted, remove lo6 cells (100 pL) for a negative control (no first antibody), and to the rest of the sample add EGF-Rl (1:50) for 30 min at 4OC. 6. Wash cells twice in 10 mL Buffer A. 7. Make up a small volume of second antibody (1:20 FITC-goat antimouse IgG) in Buffer A, and filter through a 0.22~pm membrane. Resuspend the cells (sample and negative control) at 107/mL in the second antibody containing solution. Incubate 30 min at 4OC. 8. Wash cells 3x in 10 mL Buffer B, and resuspend at 107/mL in this buffer. 9. Once the FACS analysis has been performed on the cells (seeNote 3) and the difference (if any) between f first antibody observed, a decision has to be made about what fraction to collect in the sort (seeNote 4). If the cells are being sorted for the first time, both positively staining and negative cells should be collected. As a rough guide, collect separately the brightest 5% of cells and the bottom 10%. The cells are collected into a tube(s) of relevant culture medium. 10. Following sorting, the cells should immediately be plated onto a dish of suitable size for the number of cells recovered. The culture medium will have been substantially diluted by the buffer the cells were sorted in, so after 1 h (or when the cells have reattached), the medium should be changed.

Fluorescence-Activated

Sorting

4. Notes
1. The EGF-receptor in intact cells is not susceptible to digestion by trypsin; however, other receptor proteins may be. If this is the case, it is possible to trypsinize the cells the day before the sort and replate them on bacteriological dishes. Cells do not attach well to this surface and can be removed simply by washing. It may also be possible to remove cells by using EDTA alone. 2. It is very important to make sure the cells are in a single-cell suspension. Cell aggregates cannot be sorted properly and may in fact clog up the machine. This can be a serious problem with some cells and may require passing the cell suspenion gently through a needle. 3. A positive control is required by the FACS operators to align the laser and photomultiplier tubes to get optimum sensitivity from the machine prior to sorting. In general, use cells expressing moderate to high levels of the receptor under study. To sort cells expressing EGFreceptors, we use A431 cells (106cells in 100 pL), processed with Rl exactly as described. 4. The number of rounds of FACS sorting required to obtain a population of cells expressing the transfected receptor varies depending upon the efficiency of the initial transfection. We keep sorting until over 50% of the cells are positive. In the case illustrated in Fig. 1, this may require five sorts, but more usually two or three. To obtain pure cell lines, the cells are then cloned by limiting dilution. Do not be disheartened if there is no difference between + 1st antibody on the initial sort (seeFig. 1 bottom set); collect the top 5% of cells and expand for resorting. However, if no difference is seen on the second sort, then it is not worth proceeding further. 5. Although the scope of this article (and author) does not cover a detailed description of the FACS machine and the actual sort, these being carried out by a separate cell-sorting laboratory, a brief description of the equipment used to sort the authors samples is presented (seealso Chapter 40 of this volume). All analysis and cell sorts were performed with a modified FACS I (Becton Dickinson, Mt. View, CA), which operates with an argon laser used at a wavelength of 488 nM. For the fluorescence signal a 530 nm filter was used, and for the scatter signal, a neutral density filter (1%). Ten thousand cells were analyzed at 500-2000 cells/s to determine the window settings for the sort. The nozzle tip was 100 pm in diameter.

564

Clark The windows were set from the dot plots of the analysis, with all dead cells and cell aggregates excluded. Approximately 10 cells were sorted over a 90-min period.

Acknowledgments
The author holds a C. J. Martin traveling fellowship from the National Health and Medical Research Council of Australia. With many thanks to Dorothy Cheng for maintaining all the cell lines and assisting with the sorts and to David Capillaro and the FACS team at the Imperial Cancer Research Fund for carrying out the sorts.

References
1. Sklar, L. A. and Finney, D. A. (1982)Analysis of l&and-receptor interactions with the fluorescence activated cell sorter. Cytometry 3,161-165. 2. Macinnes, D. G., Green, D. K., Harmer, A., Nairn, E. G., and Fink, G. (1983) Neuroendocrine receptor -l&and binding using quantitative video-intensification microscopy and fluorescence-activatedcell sorting. Quart. J. Exp. Physiol. 68,463-474. 3. Due, C., Linnet, K., Langeland Johansen,N., and Olsson, L. (1985) Analysis of insulin receptorson heterogeneous eukaryotic cell populations with fluorochrome-conjugated insulin and fluorescence-activated cell sorter. Diubetologia 28,749-755. 4. Chatelier, R. C., Ashcroft, R. C., Lloyd, C. J.,Nice, E. C., Whitehead, R. H., Sawyer, W. H., and Burgess,A. W. (1986)Binding of fluoresceinated epidermal growth factor to A431 cell sub-populations studied using a model-independent analysis of flow cytometric fluorescence data. EMBO J. 5,1181-1186. 5. Tubiana, N., Mishal, Z., Le Caer, F.,Seigneurin, J. M., Berthiox, Y., Martin, P. M., and Carcassonne,Y. (1986) Quantification of oestradiol binding at the surface of human lymphocytes by flow cytometry. Br. J. Cancer54,501~504. 6. Newman, R., Domingo, D., Trotter, J., and Trowbridge, I. (1986) Selection and properties of a mouse L-cell transformant expressing human transferrin receptor.
Nu ture 304,643-&5. 7.

Ellis, L., Clauser, E., Morgan, D. O., Edery, M., Roth, R. A., and Rutter, W. J. (1986) Replacement of insulin receptor tyrosine residues (1162 and 1163) by site-directed mutagenesis compromises insulin-stimulated kinase activity and uptake of 2-Deoxyglucose. Cell 45,721-732. 8. Clark, S, Cheng, D. J.,Hsuan, J.J.,Haley, J.D., and Waterfield, M. D. (1988) Loss of three major auto phosphorylation sites in the EGF-receptor does not block the mitogenic action of EGF. J. Cell Physiol. 133,421X25. Waterfield, M. D., Mayes, E. L. V., Stroobant, l?., Bennett, I. L. l?., Young, S., Goodfellow, P. N., Banting, G. S.,and Ozanne, B. (1982) A monoclonal antibody to the human epidermal growth factor receptor. 1. Cell Biochem. 20,149-161.

Chapter 42

Transfection and Transformation of Human Thyroid Epithelial Cells

Nicholas and David Robert Lemoine Wynford-Thomas

1. Introduction
High efficiency gene transfer into mammalian fibroblasts is a routine procedure that can be performed by a variety of methods (I), but transfection of epithelial cells has been more difficult to achieve. Each of the standard techniques has associated problems; special equipment is required for electroporation (2) and also for direct microinjection (3). Protoplast fusion (4) is suitable only for cloned genes. Retroviral transduction of genes (5) is a method of high efficiency for animal cells, but again is applicableonly to cloned genes of restricted size, and safety considerations are likely to restrict greatly the use of retroviral vectors suitable for human cells. Calcium phosphate coprecipitation has been used with success in several epithelial lines (6,7,8) and, more recently, strontium phosphate coprecipitation (9). These techniques are very simple to perform, require no special equipment, and can be applied to genomic DNA transfection. 565

566

Lemoine and Wynford-Thomas

The human thyroid follicular cell appears to be a particularly suitable epithelial cell for gene transfection by coprecipitation, in which insoluble complexes of DNA with calcium or strontium phosphate are taken up by recipient cells, because phagocytosis can be specifically increased by thryoid-stimulating hormone (TSH). We have recently shown that this maneuver enhances the efficiency of gene transfer, with a fivefold increase in the frequency of cells transiently expressing SV40 large T antigen as assessed by immunocytochemical assay 48 h after transfection (10). In addition, these cells appear remarkably resistant to the toxic effects of calcium phosphate, with plating efficiencies of 295% even after 16 h exposure to coprecipitate.

2. Materials
1. Plasmid DNA: closed circular plasmid DNA (prepared by cleared lysate method with banding on a CsCl, gradient) (II) is dissolved in 1.0 mM Tris-HCl, 0.1 mM EDTA pH 8.0 at a concentration of 0.1 pg/ NJ* 2. Carrier DNA: high mol wt (seeNote 1) carrier DNA (prepared by a method of caesium chloride banding) (12) is dissolved in 1.OmM TrisHCl, 0.1 mM EDTA pH 8.0 at a concentration of 0.5 pg/pL. 3. 25 rnM Hepes: dilute 2.5 mL of sterile 1M Hepes to 100 mL with sterile water. Store at 4*C for up to 3 mo. 4. 2x Hepes-buffered saline (2 x HBS): dissolve 1.64 g NaCl plus 0.023 g Na,HPO,*2H,O in 90 mL of water. Add 5 mL of 1M Hepes, then: (a) for CaCl, coprecipitation, adjust pH to 7.1 with 0.5M NaOH. (b) for SrCl,coprecipitation, adjust pH to 7.8 with 0.5MNaOH. Adjust vol to 100 mL with water (seeNote 2). Filter sterilize (seeNote 3), and store at 4*C for up to 3 mo. 5. 2.5M CaCl,: 10.8 g CaCl,~G~O in 15 mL of water. Adjust vol to 20 mL with water. Filter sterilize and store in aliquots at -2OOC. 6. 2M SrCl,: 10.665 g Sr Cl, in 15 mL of water. Adjust vol to 20 mL with water. Filter sterilize and store in aliquots at -2OOC. 7. 1.OmM Tris-HCl, 0.1 mMEDTA, pH 8.0: 100 PL l.OM Tris-HCI pH 8.0 plus 20 I.~L0.5M EDTA, pH 8.0 in 100 mL water. Filter sterilize and store at 4*C. 8. SF-12 medium: This must contain the appropriate concentration of sodium bicarbonate for the conditions of incubation: for a pH of 7.4 at 37C in 5% CO, add 2.9 mL of 7.5% NaHCO, solution/100 mL of medium (seeNote 4).

Human Thyroid Epithelial

Cells

9. RPM1 medium: prepared with sodium bicarbonate concentrations as above. 10. 11.8% (v/v) Glycerol: 11.8 mL Glycerol, diluted to 100 mL in Hanks balanced salt solution, (HBSS). Filter sterilize and store at 4C. 11. TSH: thytropar (Armour Pharmaceuticals, Tarrytown, NY) dissolved in HBSS at 1 U/mL as stock solution, filter sterilized and stored in 100 PL aliquots at -2OOC. 12. 2MNH,Cl: 1.06 g of NH&l in 8 mL of water. Adjust pH to 7.1 and adjust vol to 10 mL with water. Filter sterilize and store at 4OCfor up to 3 mo.

3. Methods
1. Thaw human thyroid epithelial cells (seeChapter 13, this vol.) from frozen stock, wash once with RPM1 medium and plate out as a monolayer at 5 x lo5 cells /60-mm dish in RPM1 medium supplemented with 10% fetal calf serum (FCS). 2. Transfection is performed 4 d after plating (seeNote 5). Six hours before transfection, replace the medium with 5 mL of warm SF-12 supplemented with 10% FCS (seeNote 6).

3.1. Cakium

Phosphate

Coprecipitation

1. Prepare DNA for transfection by mixing x PL of plasmid DNA (containing 0.5-10 pg/DNA) solution with (200 -x> PL of 25 mM Hepes in a sterile plastic bijou. If carrier DNA is used, then this is included to bring the final concentration of DNA to 10 pg/O.5 mL of coprecipitation suspension. Add 50 PL of CaCl, and swirl to mix. 2. Add this solution dropwise through a syringe and needle to 250 PL of 2 x HBS in another sterile plastic bijou with continuous mixing by a stream of air bubbles via a plugged sterile pipet with plastic tip (see Note 7). 3. Leave the mixture to stand for 30 min, after which time the fully formed coprecipitate will have settled to the bottom of the container. Coprecipitates containing high MW carrier DNA have a coarser consistency than those containing plasmid DNA only. 4. Five minutes before application of coprecipitate, add 50 PL of 1 U/mL TSH solution to the medium and swirl gently to mix (seeNote 8). 5. Gently take up the suspension in a 1-mL Gilson pipet a few times, so that the coprecipitate is uniformly suspended through the solution. 6. Add 100 PL of 2M NH&l solution to the medium and swirl gently to mix (seeNote 8).

Lemoine and Wynford-Thomas 7. Sprinkle the suspension over the medium of the epithelial monolayer. 8. Incubate at 37OC for 16 h to allow uptake of the coprecipitate. 9. Remove the medium containing coprecipitate and wash the cells twice with warm HBSS. Refeed with warm RPM1 + 10% FCS.

3.2. Strontium

Phosphate

Coprecipitation

1. Prepare DNAfor transfection by consecutively mixingx PLof plasmid DNA (containing 0.5-10 pg DNA), 30 PL of 2M SrCl,, and (220-x) PL of sterile water in a sterile plastic bijou. Seeabove for details of carrier DNA. 2. Add this solution dropwise through a syringe and 23-gage needle to 250 PL of 2 x HBS in another sterile plastic bijou with continuous mixing by a stream of air bubbles (via a plugged sterile pipet with plastic tip). 3. Leave the mixture to stand undisturbed at room temperature and observe the development of the coprecipitate. After 5-10 min, the precipitate is seen as a fine dust and is then ready for application to the cells. 4. Gently take up the suspension in a Gilson pipet a few times, so that the coprecipitate is uniformly suspended throughout the solution. 5. Sprinkle the suspension over the medium of the epithelial monolayer in a 60-mm dish. 6. Incubate at 37OC for 90 min to allow adsorption of the coprecipitate. 7. Remove the medium containing coprecipitate, and wash the cells twice with warmed HBSS. Apply 1.5 mL of 15% glycerol in HBSS and incubate at room temperature for 30 s, then wash three times with warm HBSS, and refeed with warm RPM1 + 10% FCS.

3.3. Transient

Expression

Assay

The transient expression of genes that have been successfully transfected is maximal at 48-72 h after transfection in these cells. A convenient assay system that allows calculation of transfection frequency is immunocytochemical demonstration of SV40 large T expression (Fig. 1, see Chapter 44, this vol. for alternative transient assay systems) following the transfection of a plasmid such as pSVorP(13).

3.4. Stable Expression

Assay

The transfection of a dominant selective marker such as the neo gene that codes for resistance to geneticin in eukaryotic cells allows assay of stable transfection frequency. Primary human thyroid epithelial cells have

Human

Thyroid Epithelial

Cells

569

Fig. 1. Pair of recently divided thyroid epithelial cells showing stong nuclear immunoreactivity with antiSV40 large T antibody pAB 419,48 h after transfection with a plasmid containing the whole SV40 sequence. (1mmunocytochemistry was performed on acetone-fixed monolayers using pAB 419 followed by rabbit antimouse immunoglobulin conjugated to horseradishperoxidase. Positivityisshownbydepositionof thebrownperoxidasecatalyzed polymer of diaminobenzidine).

poor cloning efficiency and require the use of a feeder layer (of geneticinresistant fibroblasts) in order to develop viable clones from the low density required for geneticin selection. Transfection of plasmids containing SV40 leads to the outgrowth of clones with extended life span that continue to proliferate after the untransformed cells undergo senescence (after an average of 3-6 doublings). These clones can be shown to express SV40 large T by the immunocytochemical assay mentioned above.

4. Notes
1. DNA prepared by methods that involve phenol/chloroform extraction of proteinase K digests for example will be of lower mol wt, and this will reduce the efficiency of transfection.

570

hmoine

and Wynford-Thomas

Fig. 2. Relationship between time of addition of TSH (10 mU/mL) to monolayer culture and efficiency of transfection (as assessed by proportion of cells showing transient expression of SV40 T antigen 48 h after transfection). q lno TSH (addition of NH&l had no effect on this result). 4 TSH (10 mU/mL) added at indicated times. A TSH (10 mU/mL) added at indicated times plus NH&l (20 m&f) added at time 0. All points show mean +/ - SE derived from 6 separate experiments.

2. The effect of pH is critical, and it is recommended that pH of stock solutions should be checked at intervals after preparation, and if incorrect, new reagents should be prepared. 3. 0.22 pm filters should be flushed with sterile water before use. This has been observed to reduce subsequent aggregation of the DNA-cation coprecipitates. 4. This medium has been found to be ideal for the culture of cells during transfection by coprecipitation. Some other media (including RPMI) are unsuitable because the product of divalent cation -phosphate concentrations is too high, causing excessive coprecipitation. 5. The growth rate of these cells (as determined by thymidine labeling) shows a peak on d 4 after stimulation with serum. The highest efficiency of transfection is at this time (10). 6. Media and wash solutions should be prewarmed to enhance growth and transfectability of cells. 7. Even mixing of the solutions is best achieved by a gentle stream of bubbles, and the pipet should have a plastic tip since the precipitate adheres to glassware.

Human

Thyroid Epithelial

Cells

571

8. The addition of TSH to stimulate phagocytosis significantly

improves the efficiency of transfection by calcium phosphate coprecipitation when given within a critical time-window (seeFig. 2): no effect is seen if it is given more than 20 min before (or at any time after) addition of the coprecipitate. Lysosomal function inhibitors, such as ammonium chloride, have no influence when used alone, but have a synergistic effect when used in concert with TSH in this system. Interestingly, no effect of TSH has been demonstrable on the efficiency of strontium phosphate coprecipitation.

Acknowledgments
We are grateful to the Cancer Research Campaign and to the Welsh Scheme for the development of Health and Social Research for grant support.

References
1. Pollard, J. W., Luqmani, Y., Bateson,A., and Chotai, K. (1984) DNA transformation of mammalian cells, in Methods in Molecular Biology, vol. 2 (Walker, J. M., ed.),

Humana Fress, NJ, pp. 321-332. 2. Tur-I&spa, R.,Teicher, L., Levine, B.J.,Skoultchi, A. I., and Shafritz, D. A. (1986) Use of electroporation to introduce biologically active genes into primary rat hepatocytes. Mol. Cell Biol. 6,716-718. 3. Garcia, I., Sordat, B., Rauccio-Farinon, E., Dunand, M., Kraehenbuhl, J. -P., and Diggelmann, H. (1986) Establishment of two rabbit mammary epithelial cell lines with distinct oncogenic potential and differentiated phenotype after microinjection of transforming genes. Mol. Cell Biol. 6,1974-1982. 4. Yoakum, G. H., Lechner, J. F., Gabrielson, E. W., Korba, B. E., Malan-Shibley, L., Willey, J. C.,Valerio, M. G.,Shamsuddin, A. M., Trump, B.F., and Harris, C. C. (1985) TransformationofhumanbronchialepithialcellstransfectedbyHarveyrasoncogene. Science 22, 1174-1179. 5. Wolff, J. A.,Yee J. -K., Skelly, H. F., Moores, J. C., Respess,J. G., Friedmann, T., and Leffert, H. (1987)Expression of retrovirally transduced genes in primary cultures of adult rat hepatocytes. Proc. N&l. Acad. Sci. USA 84,3344-3348. 6. Hynes,N. E.,Jaggi,R., Kozma, S.C., Ball, R., Muellener, D., Wetherall, N. T., bavis, B. W., and Groner, B. (1985) New acceptor cell for transfected genomic DNA: oncogene transfer into a mouse mammary epithelial cell line. Mol. Cell Biol. 5, 268-272. 7. Storer, R. D., Stein, R. B., Sina, J. F., DeLuca, J. G., Allen, H. L., and Bradley, M. 0. (1986) Malignant transformation of a preneoplastic hamster epidermal cell line by the EJ c-Ha-ras oncogene. Cancer Res. 46,1458-1464. 8. Summerhayes, I. C., Malone, P., and Visvanathan, K. (1986) Altered growth properties and cell surface changes in ras transformed mouse bladder epithelium. Int. J. Cancer37,233-240. 9. Brash, D. E., Reddel, R. R., Quanrud, M., Yang, K., Farrell, M. P., and Harris, C. C.

572

Lemoine and Wynford-Thomas

(1987) Strontium phosphate transfection of human cells in primary culture: stable expression of the simian virus 40 large-T-antigen gene in primary human bronchial epithelial cells. Mol. Cell Bid. 7,2031-2034. Lemoine, N. R. and Wynford-Thomas, D. (1987) The thyroid follicular cell as a recepient for DNA transfection. Br. J. Cancer 55,342. Boffey, S. A. (1984) Plasmid DNA isolation by the cleared lysate method, inMethods in Molecular Biology, vol. 2 (Walker, J. M., ed.), Humana Press, pp. 177-183. Kaiser, K. and Murray, N. E. (1985) The use of phage lambda replacement vectors in the construction of representative genomic DNA libraries, in DNA Cloning, vol. 1 (Glover, D. M., ed.), IRL, pp. l-74. Gluzman, Y., Sambrook, J. F., and Frisque, R. J. (1980) Expression of early genes of origin-defective mutants of SV40. Pm. Nufl. Ad. Sci. USA 77,3898-3902.

10. 11. 12. 13.

Chapter 43
Expression of Foreign Genes in Cultured Insect Cells Using a Recombinant Baculovirus Vector
Martin J. Page and Vivienne F. Murphy
1. Introduction
An important consideration for the expression of cloned genes in recombinant expression systems is the ability of the foreign host to produce the protein faithfully in a form that is similar or identical to that found in the cell type from which the gene was cloned. For eukaryotic proteins, this frequently involves many posttranslational modifications of the protein, such as glycosylation, phosphorylation, processing, and secretory events. Additionally, very precise interactions are essential for the correct folding of the polypeptide to achieve the final tertiary structure. If the folding is incorrect, then the molecule will often be biologically inactive. 573

Page and Murphy These considerations have led to the increasing use of recombinant eukaryotic expression systems to express cloned genes accurately. In particular, recombinant mammalian expression systems have been used extensively to achieve this aim, with considerable success. More recently, an alternative higher eukaryotic expression system involving the use of a recombinant baculovirus and insect tissue culture cells has demonstrated considerable advantages. These include the speed of obtaining expression of the recombinant product, the potential high yield, and the option of expressing proteins that would otherwise be toxic at high levels in mammalian cells (e.g., c-myc, seeref. 1). Examples of recombinant proteins produced using the insect expression system have so far shown them to be appropriately modified, processed, secreted, and correctly folded to give high yields of biologically active proteins, such as human p interferon (2), P-galactosidase (3), c-myc (4), interleukin-2 (5), and influenza hemagglutinin (6,7). The insect baculovirus used for these experiments is called Autogvapha Ca2ifornica Nuclear Polyhedrosis Virus (AcNPV), and it has a 128 kdaltons double-stranded circular DNA genome, which replicates within the nuclei of lepidopteran species with a biphasic life cycle. The first phase (W-24 h postinfection) involves the formation of mature nucleocapsids that bud through the cellular membrane to produce extracellular virus (EV). The second phase (18-24 h postinfection) involves high-level expression of the AcNPV late gene called polyhedrin. This protein can account for up to 50% of the protein mass of an infected insect cell and is responsible for embedding mature virus particles within the nuclei, generating extremely large viral occlusion bodies (seeFig. 1). These occluded viruses (OV) are released only after cell death and are essential for lateral transmission of the virus. Several features of this biphasic life cycle make the insect baculovirus system amenable for geneticmanipulation. First, the polyhedrin gene product is totally unnecessary for the EV form of the life cycle, which is the only infectious form in insect cell cultures. Therefore, the polyhedrin gene can be replaced with a foreign gene, so that it is suitably positioned under control of the powerful polyhedrin promoter. Second, whereas wild-type AcNPV will make polyhedrin within infected cells and give rise to viral plaques easily identified visually as occlusion positive, the recombinant AcNPV lacks the polyhedrin gene and gives rise to occlusion negative viral plaques. This forms the basis for rapidly screening and identifying rare recombinant viruses that contain and likely express a cloned foreign gene.

Foreign Gene Expression

in Insect Cells

575

Fig. 1. The contrast between uninfected Sf insect cells (top) and cells 48 h postinfection with wild-type AcNPV (bottom). The polyhedrin occlusion bodies are easily visible within the nuclei of the infected cells and give rise to occlusion-positive plaques that have a characteristic silver-gray sheen when examined against a light source.

This chapter is aimed at introducing this technology to researchers who may have little or no previous experience of the insect baculovirus expression system. Throughout, it is assumed that some knowledge of basic recombinant DNA and tissue culture techniques are known, although important aspects of each will be emphasized.

576

Page and Murphy

2. Materials
1. The insect cell line Spo&rWu frugiperda Clone 9 (S. f. 9) may be obtained from the American Type Culture Collection (12301 Parklawn Drive, Rockville, MD 20852, USA) or alternatively from any research group working in this area. 2. Insect cell stocks are maintained at room temperature in modified BML TC-10 medium (8) supplemented with 10% fetal calf serum and antibiotics as log-phase monolayer cultures in glass tissue culture flasks of approximately 150 cm* surface area. The modifications to the culture medium are as follows. The concentrations of L-arginine, Lhistidine, and calcium pantothenate are 550 mg/L, 3.4 g/L, and 0.11 mg/L, respectively. Choline chloride is replaced with cyanocobalbumen at 0.01 mg/L. 3. lx TE buffer: Prepare a solution containing 10 mM Tris HCI, pH 7.4, 1 mM EDTA, autoclave, and store at room temperature. 4. Transfection buffer: Prepare a solution containing 25 mM Hepes, pH 7.1,150 mMNaC1,125 mM CaCl, autoclave, and store at 4C. The pH of this solution is critical and should be 7.1+ 0.05. 5. Neutral red stock solution is obtained from BDH. 6. Low melting temperature ultrapure Seaplaque agarose.

3. Methods
To obtain successful expression of cloned genes using the insect expression system, it is critical that the principles underlying each experimental step are clearly understood. For this reason, each of the steps necessary to engineer the baculovirus DNA are shown schematically in Fig. 2, and described in detail below.

3.1. Cloning

of Foreign mansfer

Genes into an AcNPV Vector

To obtain expression of recombinant proteins using this system, it is necessary to position the foreign gene under control of the strong polyhedrin promoter. Owing to the very large size of the AcNPV genome (about 128 kdaltons), and lack of a suitable restriction site at the desired position, this cannot be achieved by a direct cloning route. Instead, the foreign gene is first subcloned into an AcNPV transfer vector. This consists

Foreign Gene Expression

in Insect Cells

577

TRANSFER VECTOR

FOREIGN GENE

BACULOVIRUS

RECOMBINANT TRANSFER VECTOR

POLYHEDRIN GENE

IRAL NA

HOMOLOGOUS RECOMBINATION IN WV0

-INSECT

CELL

\ t t + t

PROGENY RELEASED INFECTED

VIRUS FROM CELL 0 l-l% CONTAIN FOREIGN GENE -LIQUID OVERLAY AGAROSE OVERLAY CELL MONOLAYER (INFECTED WITH PROGENY VIRUS)

Co PETRI DISH c - - - - - - - - - - - - - - -i, 27% NEUTRAL 3 days

RED STAIN

Fig. 2. Generation of a recombinant baculovirus for expression of a foreign gene. (1) Construction of the recombinant transfer vector. (2) Extraction of wild-type baculovirus DNA. (3) Cotransfectionof recombinant transfer vector and baculovirus DNA. (4) Plaque assay for selection of recombinant baculoviruses. Steps l-4 are detailed in Methods sections 3.1,3.4., respectively.

578
0 / Hind III

Page and Murphy

3.96 EcoR+ 4.0 BamH I -8 / +177 I Hind III 4.1

\ Kpn I 4.43

Fig. 3. The baculovirus transfer vector pAc373. The map units shown are in kilobase pairs relative to the Hind III site at map unit 0. The insertion site for cloning foreign genes is at the BamHI site, which is map unit 4.0, and this represents fusion of coordinates-8 and +177 relative to the original polyhedrin ATG translation codon via a BamHI linker. The remainder of the polyhedrin gene is shown as a hatched box, whereas flanking baculovirus DNA is shown as stippled boxes.

of a cloned region of the AcNPV genome encompassing the polyhedrin gene, in which a unique restriction site has been engineered immediately upstream of the polyhedrin coding body. A suitable and widely used AcNPV transfer is pAc373 (51,which has a unique BamHl restriction site engineered at position -8 relative to the polyhedrin ATG translation initiation codon (seeFig. 3). The following steps should be performed using standard cloning techniques (9, and vol. 2 and 4 of this series) to insert the foreign gene into pAc373: 1. Digest pAc373 with the restriction enzyme BamHl. 2. Phosphatase the protruding 5 overhang with calf intestinal phosphatase. This step is necessary to reduce the high frequency of vector self-ligation that would otherwise occur. 3. The foreign gene to be cloned must be prepared or manipulated so that it has homologous cloning ends for the phosphatased pAc373 BamHl

Foreign Gene Expression

in Insect Cells

579

4. 5. 6. 7.

digested transfer vector. Ideally, the gene should have BamHl, BglII, or BclI ends to facilitate sticky end cloning. If these sites are not present, then there are two options. First, the ends could be altered by blunting and cloning suitable linkers onto the ends (making sure that there are no internal sites within the gene), or secondly, the ends could be left as blunted ends and cloned into a pAc373/BamHl digested transfer vector that has also had the BamHl ends blunted. Ligate the foreign gene into the phosphatased pAc373 transfer vector using T4 DNA ligase. Transform competent E. cdi to ampicillin resistance using an aliquot of the ligation mix. Prepare mini-prep plasmid DNAs from picked colonies and identify by restriction enzyme analysis recombinants that have the foreign gene inserted in the correct orientation. Prepare full-scale DNA preparations from an identified positive clone and band through a cesium chloride density gradient. Perform a series of restriction enzyme digests to confirm that the recombinant transfer vector is correct. The foreign gene is now correctly inserted in the proper orientation immediately downstream of the polyhedrin promoter within a recombinant AcNPV transfer vector.

3.2. Preparation of High Molecular Weight Infectious DNA fkom Wild-5Qpe AcNPV
The foreign gene contained within the AcNPV recombinant transfer vector now has to be introduced precisely into the AcNPV genome by homologous recombination in vivo. This procedure requires the use of goodquality, intact AcNPV DNA, which should be capable of initiating aninfection when transfected into insect cells. 1. Seed a 100-mL spinner flask with insect cells at 1 x 106/mL and incubate overnight at 27OC. 2. Centrifuge at 2,500 rpm for 10 min to pellet the cells and debris using a bench centrifuge. Discard the cell pellet. 3. Pellet the extracellular virus (EV) from the supernatant at 100,000 x g in a swing-out rotor (e.g., 27,500 rpm in a Beckman SW40 rotor) for 30 min at room temperature. 4. Resuspend the viral pellet in 10 mL of 0.1 x TE, and pellet the EV as before. 5. Resuspend the viral pellet in 4.5 mL of 1 x TE containing 0.2M NaCl,

580

Page and Murphy

and then add 200 pg of proteinase K. Incubate at 50C for 1-2 h. 6. Add 0.5 mL of 10% sarkosyl, and incubate for a further 2 h or overnight. 7. Add an equal vol of 1:l phenohchloroform, and mix the phases very gently, to prevent shearing of the DNA, for five min. 8. Repeat step 7, and then reextract in the same way with chloroform alone. Transfer upper aqueous phase to a clean tube. Centrifuge at 100 rpm for 5 min; then transfer upper aqueous phase to a clean tube using a wide bore pipet. 9. Add 10 mL of cold (-20C) ethanol and mix gently. A cotton-wool like precipitate should form at this stage. Centrifuge at 1000 rpm for five min to pellet the precipitate, discard the supernatant, and wash the pellet in 5 mL 80% ethanol. 10. Vacuum dry the DNA pellet, and dissolve it in 0.5 mL of 0.1 x TE. The DNA is now ready for use. The yield should be around 200-500 pg. Store at 4OC. The quality of the purified AcNPV DNA can be monitored by its ability to initiate an infection after transfection into insect cells (seebe2ow).

3.3. Cotransfection of Wild-Type AcNPV DNA and Recombinant AcNPV lkansfer Vector into Insect Cells
This step is necessary to introduce into cultured insect cells both DNA molecules consisting of the wild-type AcNPV DNA and the smaller recombinant transfer vector containing the foreign gene to be expressed. Once inside the cell, a small proportion of the molecules (approximately 1 in 100-1000) will undergo double-reciprocal homologous recombination in which the polyhedrin gene of the wild-type AcNl?V DNA is replaced by the foreign gene. 1. Seed 2 x lo6 insect cells into a 25 cm2 tissue culture flask, and leave at 27OC for 1 h to attach. 2. Into a sterile 1.5 mL Sarstedt tube, add 1 pg of high mol wt wild-type AcNPV DNA and 2 pg of recombinant transfer vector. Add 0.75 mL of transfection buffer, and mix by pipeting up and down. 3. Remove all of the medium from the flask of insect cells, and replace with exactly 0.75 mL fresh medium. Be sure to cover the entire monolayer with a film of medium. Slowly add the 0.75 mL of DNA solution directly onto the cell monolayer dropwise using a I-mL

Foreign Gene Expression in Insect Cells

pipet. The DNA will form a calcium phosphate DNA precipitate in situ because of the phosphates in the medium. 4. Leave at 27OCfor 4 h, then remove the transfection solution and carefully wash the cell monolayer twice with 5 mL of culture medium. Finally, add 5 mL fresh medium and incubate (with cap tightly closed) in a 27C incubator for 3 d. 5. At the end of this period, a significant proportion (usually about 10%) of the cells should be showing positive signs of viral infection as evidenced by large granular occlusions of polyhedrin in the nuclei of infected cells. By this stage, the culture medium will contain virus particles at a concentration of about 2 x 10 pfu/mL. Only 0.1-l% of these will consist of recombinant viruses, and these have to be identified by visual inspection following a plaque assay.

3.4. Plaque

Assay

Recombinant viruses present in the medium harvested following cotransfection can be identified and isolated by plaque assay. This method can also be used to determine the titer of an unknown virus stock solution. The principle of the method is to obtain well-isolated viral plaques in a confluent cell monolayer, which can then be screened for recombinants or simply counted for determination of virus titer. 1. Prepare a series of lo-fold serial dilutions in culture medium of the virus stock. An appropriate range for culture supernatant harvested from cotransfections is 1 in 100-l in 10,000, and for determination of virus titer, 1 in 100-l in l,OOO,OOO. 2. Seed 35-mm plastic Petri dishes with 2 x lo6 insect cells, and incubate at 27OCfor 1 h to allow the cells to attach. Seed enough for 5-10 plates/ dilution for screening recombinant plaques and two plates/dilution for determination of virus titer. 3. Remove most of the culture medium from the dishes, leaving a small amount to prevent drying out of the monolayer. Gently drip 100 PL of the diluted virus onto the cells in the center of the dish, removing any bubbles that form. Incubate for 1 h at room temperature to allow adsorption of the virus. 4. Prepare a 1% agarose/culture medium overlay solution as follows: Autoclave an appropriate volume of 3% Seaplaque agarose solutionin distilled water (0.5 ml/plate) and cool to 37OC. Warm twice this volume of culturemedium to 37C. Mix the agarose and the medium, and maintain at 37OC.

582

Page and Murphy

5. Remove the inoculum from the Petri dishes using a sterile Pasteur pipet. Add 1.5 mL of the overlay mix to the center of each plate, allowing it to spread evenly over the monolayer. Allow it to set for 20 min at room temperature, and then overlay with 1 mL of culture medium. Incubate undisturbed in a humid environment at 27C for 3-4 d. Movement of the plates during incubation can cause smearing of the plaques. 6. Add 1 mL of a neutral red stock solution, diluted 1 in 10 in PBS, to the liquid overlay. Incubate at 27OC for 1-2 h. Tip off the liquid overlay containing stain, invert the plates, and leave for several hours or overnight at 4C for the plaques to clear. Plaques appear as circular regions of weak staining about 1-3 mm in diameter in a darker stained monolayer, and may show retarded cell growth. Cells in the center of wild-type plaques will contain nuclear occlusion bodies (seeFig. 1) clearly visible under low-power magnification. The rare recombinants generated following cotransfection will lack occlusionbodies and can be selected visually by this phenotype. Each plaque represents one virus or pfu in the inoculum; therefore, the titer of the original stock can be determined allowing for the dilutions made. 7. To screen for recombinants, select plates with well-isolated plaques at a density of 20-200 plaques/plate. With practice, it is possible to pick out occlusion-negative plaques by eye, and then confirm these by microscopic analysis. Wild-type plaques containing occlusion bodies have a characteristic refractile appearance, giving a slight silvery sheen when viewed at an angle against a light source. Plaques lacking this sheen can be marked for closer inspection under 400x magnification with an inverted microscope. Check every cell in the plaque carefully for occlusion bodies, and if none are seen, mark the plaque for plaque purification.

3.5.

Plaque

Purification

Occlusion-negative plaques may still contain some wild-type virus contamination because of diffusion from neighboring plaques. Thus, to isolate a pure viral stock, recombinants are purified by successive rounds of plaque purification. 1. Using a sterile Pasteur pipet, pick a plug of agarose from directly over the potential recombinant occlusion-negative plaque into 0.5 mL of culture medium. Vortex briefly, and leave at room temperature for 30

Foreign Gene Expression in Insect Cells

583

min or longer to allow the virus to diffuse from the plug. Keep an aliquot at 4C for long-term storage. 2. Carry out a plaque assay as in section 3.4., using the plaque suspension and lo-fold dilutions of it down to 1 in 1000. 3. Stain and screen as described in the previous section. 4. Repeat until the plaques generated are 100% occlusion negative. Generally, three rounds of plaque purification are carried out to ensure elimination of wild-type virus, but the presence of the foreign gene can be confirmed by Southern Blotting or product analysis at an earlier stage if required.

3.6. Derivation of High-Titer Medium and Determination of pj%lmL

Once an occlusion-negative plaque has been successfully identified and plaque purified, it is necessary to obtain culture medium containing high titers of the recombinant virus for further infections. This is achieved by a series of scale-up infections and the resulting high-titer culture medium is then titerd by plaque dilution assays to determine the pfu/mL of infectious recombinant virus. 1, The isolated pure recombinant AcNPV plaque is picked and transferred into a sterile Bijoux bottle (5 mL size) containing 1 mL of culture medium, and left overnight at room temperature to allow diffusion of the virus particles into the medium. 2. The next day, 0.5 x lo6 insect cells are seeded onto a 35-mm tissue culture dish and left for 1 h at 27OCto attach. The medium is removed and replaced with 0.8 mL of the culture medium containing the virus plaque. The cells are left for a further hour at 27OCto allow infection to occur. Then the medium is removed and replaced with 2 mL fresh medium. The cells are left at 27OCfor 4 d. The remaining 0.2 mL of culture medium containing the original plaque is stored at 4OCas a reserve in case of contamination at this stage. 3. After 4 d, the medium is removed and spun at 1500 rpm in abench-top centrifuge for 5 min and the culture medium transferred to a labeled sterile tube. Of this, 1.8 mL is used to infect a T-25 flask seeded with 5 x lo6 insect cells, as described above, then replaced with 5 mL of culture medium, and left for a further 3 d at 27C. 4. The medium is then removed and spun as described above. Of this, 4.8 mL is used to infect a T150 flask seeded with 30 x lo6 insect cells,

Page and Murphy as described above, then replaced with 50 mL culture medium and left for a further 3 d at 27C. 5. The medium is removed and spun as described above. At this stage, one has 50 mL of culture medium containing virus at a titer in the range of 1-4 x lo8 pfu/mL. Of this, 1 mL is frozen at -8OOC for longterm storage, the remainder is kept at 4OCand should retain its high infectivity for at least 1 yr. 6. It is usual to determine the titer of the virus in the culture medium at this stage, and this is done by a dilution plaque assay. Assuming that the titer is about lo8 pfu/mL, carry out the necessary dilutions to give 10-100 plaques/35-mm plate using a standard plaque assay as detailed in section 3.4.

3.7. Isolation of Recombinant AcNPV DNA for Characterization by Southern Blotting

It is good practice to verify that the recombinant virus, identified by virtue of producing occlusion-negative plaques, does indeed contain the inserted foreign gene, in the correct orientation, and without rearrangement or deletion. Fortunately, it is not necessary to isolate pure viral DNA as described in section 3.2. for this purpose. Instead, since about 25% of the total nucleic acid from cells late in infection consists of viral DNA, this can be used to isolate a total DNA preparation, which is very suitable for restriction enzyme and Southern Blot analysis. 1. Infect 30 x lo6 cells in a T-150 flask at a multiplicity of infection of 5 pfu/ cell, and leave for 3 d. 2. Smack the flask hard a few times to dislodge the insect cells, spin at 1000 rpm for 5 min, and then wash the cell pellet twice with 25 mL of PBS. Do not vortex or pipet the cells vigorously; otherwise, considerable cell lysis will occur. Finally, allow the cell pellet to drain for a few minutes. 3. Resuspend the cell pellet in 5 mL of TE buffer. Then add 0.1 mL of proteinase K solution (1 mg/mL in TE), mix quickly, rapidly add 0.25 mL of 10% SDS, and mix rapidly by inverting the tube several times. Leave the mixture at 37OC for 4 h. 4. Add 5 mL of phenol:chloroform (l:l), mix by inverting and rolling the tube (do not vortex), spin in a 30-mL glass sterile Corex tube at 10,000 rpm in a Sorvall centrifuge for 10 min, then remove the lower organic layer with a pipet, and discard. Repeat this procedure once more with phenol:chloroform and then once with chloroform alone.

Foreign Gene Expression

in Insect Cells

5. Remove the top aqueous layer this time, into a new labeled sterile Corex tube. Add 0.15 mL of 5M NaCl and 12.5 mL of ethanol, and swill or invert to mix. A fluffy white precipitate of DNA should form at this stage. This is spun, washed in 70% ethanol, and then air dried. 6. The pellet will contain RNA and DNA at this stage. Restriction enzyme digests can be performed on this material, but must include RNase A (preheated at 70C for 10 min to inactivate any DNases present) at a final concentration of 10 pg/digest. Alternatively, the total nucleic acid preparation can be dissolved in TE buffer, incubated with RNase A (50 pg/mL) for 3 h at 37OC,then deproteinized again by adding NaCl to O.l5M, SDS to 0.1% and proteinase K to 20 pg/mL, and leaving for a further 2 h at 37C. The solution is then extracted once with phenol:chloroform and once with chloroform before precipitating with ethanol. The DNAobtained after this step will be purer and more likely to digest completely with restriction enzymes. The final DNA pellet is dissolved in 1 mL sterile distilled water and quantitated by UV absorbance at 260 nM. 7. 2 t.tg aliquots of DNA are digested with restriction enzymes, electrophoresed, and blotted using standard procedures. The blot is hybridized with a radioactive probe consisting of the cloned foreign gene. The signals obtained are usually very strong because of the high proportion of pure viral DNA within the total DNA preparation. Control digests using the cloned foreign DNA in the original recombinant transfer vector should be included to identify correct co-migrating bands.

4. Notes
1. Subculturing of the insect cell monolayer stocks is required at approximately 1-wk intervals, while cells are still in log phase growth (i.e., subconfluent). Cells are detached from the flask by sharply smacking the flask several times or by scraping the cells off with a sterile rubber policeman. Trypsin/versene should not be used. Viability counts should be done at this stage using Trypan blue stain. The resuspended cells are normally subcultured 1/lO for weekly maintenance. 2. Spinner cultures can be used for virus preparations or for large scale production of recombinant protein. These shouldbe grown at 27-28C, which is the permissive temperature for virus replication. Standard glass spinner flasks are seeded at 5 x 104-1 x 106 cells/mL, stirred at 50-100 rpm. Aeration is not required for volumes up to 500 mL, but

586

Page and Murphy for larger scale cultures, refer to reference 20 and Chapter 6 of this vol. for methodology. We have noticed considerable variation in the growth and plaquing properties of several Spodopteva frugiperda cell lines from different sources. If problems are encountered, then a different cell isolate from another research group should be tried. Some protocols recommend the use of TClOO or Graces insect culture medium. However, in our experience, commercial sources of these support only poor growth of the Spodopteru fiugiperda cell line. The culture medium recipe given in the Materials section (and in detail in the Appendix) has, in our experience, proven to be the most reliable for good cell growth. Although the insect cells grow optimally at 27OC,we prefer to restrain the growth of the cells by maintaining the stock cultures at room temperature. These slower growing cells seem to perform consistently better in plaque assays (carried out at the optimal temperature of 27C). The insect cells are considerably more fragile than most mammalian cells. For this reason, the cells should not be subjected to rapid pipeting or vortexing. Otherwise, substantial cell death could occur. It is recommended that viability counts be performed regularly on the cells during passaging. If difficulties are consistently found in attempting to identify the rare occlusion-negative recombinant plaques among the wild-type occlusion-positive plaques, then it is advisable to obtain a pure recombinant virus stock from a research group and plaque this just to familiarize oneself with plaques totally lacking polyhedrin. This could be further reinforced by mixing together wild-type and recombinant virus in a 1O:l ratio, carrying out a plaque assay, and then practice identifying the 1 occlusion-negative plaques in every 10 occlusion positive plaques. The condition and number of cells used in a plaque assay is critical for obtaining well-formed, easily identifiable occlusion-positive or negative plaques. The cells must be subconfluent on the first day of the assay. If a confluent monolayer is formed on the first day (Le., by seeding too many cells), then the plaques will be small and poorly formed. If the cells fail to reach confluence by the end of the assay, identification of plaques is often very difficult. This is usually the result of unhealthy stock cells, or obtaining high numbers of dead cells upon harvesting with which to set up the assay.

3.

4.

5.

6.

7.

8.

Foreign Gene Expression in Insect Cells

587

9. Occasionally, occlusion-negative plaques can be generated by events other than inactivation of the polyhedrin gene by insertion with the foreign gene. It is strongly recommended that Southern Blot analysis be performed on final plaque purified recombinants to ensure identification of true recombinants. 10. The factors affecting the levels of foreign gene expression with the baculovirus/insect cell system are still poorly understood. Although polyhedrin can be expressed from the wild-type virus at levels approaching 1 mg/mL, the levels of foreign gene expression are usually in the range of l-50 pg/mL using the transfer vector pAc373. 11. pAc373 can only be used to express mature foreign proteins (i.e., from genes that contain their own ATG translation start codon followed by the uninterrupted coding open reading frame). However, there are vectors available, such as pAc360 (II), that use the polyhedrin gene ATG translational start followed by a unique cloning site shortly downstream of theN terminus. The foreign genes are cloned in frame into this site and will now be expressed as polyhedrin/foreign gene fusion proteins. The levels of these proteins can often be considerably greater than that obtained with mature proteins generated from pAc373 and can easily provide large quantities of material suitable for antigenicity evaluation. The pAc373 vector is most suitable for expression of therapeutic proteins where biological activity is more important.

References
1. Wurm, E M., Gwinn, K. A., and Kingston, R. E. (1986) Inducible overproduction of
the mouse c-myc protein in mammalian cells. Proc.Nutl. Ad. Sci. 83,5414-5418. 2. Smith, G. E., Summers, M. D., and Fraser, M. J. (1983) Production of human J3 interferon in insect cells infected with a baculovirus expression vector. Mol. Cd. Bid. 3, 2156-2165.
3. Pennock, G. D., Shoemaker, C., and Miller, L. K. (1984) Strong and regulated expression of Escherichia coli P-galactosidasein insect cells with a baculovirus vector. Mol. Cell. Biol. 4,388X%. 4. Miyamoto, C., Smith,G. E.,Farrell-Towt, J.,Chizzonite,R.,Summers, M. D.,and Ju,

G. (1985) Production of human c-myc protein in insect cells infected with a baculovirus expression vector. Mol. Cell. Biol. 5,2860-2865. 5. Smith, G. E., Ju, G., Ericson, B. L., Moschera, J., Lahm, H. W., Chizzonite, R., and Summers, M. D. (1985) Modification and secretion of human interleukin-2 produced in insect cells by a baculovirus expression vector. Proc. N&l. Ad. Sci. 82,
8404-8408.

588

Page and Murphy

6. Possee, R. D. (1986) Cell surface expression of influenza virus hemagglutinin in insect cells using a baculovirus vector. Virus Research 5,43-59. 7. Kuroda, K., Hauser, C., Rott, R., Klenk, H. D., and Doerfler, W. (1986) Expression of the influenza virus hemagglutinin in insect cells by a baculovirus vector. EMBO. J. 5,1359-1365. 8. Gardiner, G. R. and Stockdale, H. (1975) Two tissue culture media for production of lepidopteran cells and nuclear polyhedrosis viruses. 1. Znvertebr. Path. 25,363-370. 9. Maniatis, T., Fritsch, E., and Sambrook, J. feds.) (1982) Molecdur Cloning-A Labovatovy Manual (Cold Spring Harbor Laboratory, New York). 10. Weiss, S. A. and Vaughn, J. L. (1986) Cell culture methods for large-scale propagationofbaculoviruses,inTheBiologyofBaculoviruses(Granados,R.andShapiro,M., eds.), CRC Press, Florida. 22. Summers, M. D. and Smith, G. E. (1987) A manual of methods for baculovirus vectors and insect culture procedures. Texas Agriculture Experiment Station, bulletin No.: 1555.

Chapter 44
Chloramphenicol Acetyltransferase as a Reporter
in Mammalian
Christopher and Bruce

Gene

Transfer

D, Corsica H. Howard

1. Introduction
Reporter genes can be used to advantage in a variety of different kinds of gene transfer experiments. One of the most common applications is the optimization of transfection methods. Owing to the large number of variables that influence the uptake and expression of exogenously introduced genes, it is extremely helpful to have available rapid, sensitive methods for determining transfection efficiency. A second frequent application of reporter genes is to study promoters and transcriptional enhancers. To rigorously map functional domains in transcriptional control elements it is often necessary to generate large numbers of deletions, insertions, and point mutations, and then carry out functional assays on each construct. As in optimization studies, an appropriate reporter gene can greatly facilitate the rate with which information is accumulated. Another gene regulation application involves measuring activity of transacting transcriptional regulatory proteins. In such experiments, one plasmid codes for the transacting factor, while the second cotransfected plasmid carries a target 589

590

Corsica and Howard

&-acting element coupled to a reporter gene. Examples of more novel reporter gene uses include introducing translational stop codons into a reporter coding sequence to measure suppressor function in mammalian cells (1,2), and measuring transcriptional competence of UV-irradiated reporter plasmid DNA (3). Finally, reporter genes can be used to normalize expression where other transfected genes are measured at the RNA level by Sl nuclease, primer extension, or Northern analysis (4). Given this multiplicity of applications, there has been a steady increase in the number of studies in which reporter genes are used. Of the many different reporter genes available for gene transfer studies in mammalian cells, the chloramphenicol acetyltransferase (CAT) gene (5) has probably been the most widely used. Another reporter gene that is frequently employed is the E. coli B-galactosidase gene (6). The E. coli xanthine-guanine phosphoribosyltransferase (7,8) and galactokinase (9) genes, and recently, the firefly luciferase gene (20) represent alternative reporter systems. There are several variables in the choice of an enzymatic activity as a reporter system. Perhaps the first consideration is whether there is a corresponding endogenous enzyme activity in recipient cells. The absence of endogenous activity eliminates the possibility that the investigator will encounter interference in the detection of the expression of exogenously introduced genes. An additional concern is whether the assay is appropriate for the experiments contemplated, e.g., whether cost or sensitivity is of primary importance. Finally, an ancillary but often crucial point is whether a given reporter system provides functions in addition to an assay for soluble enzyme activity. With respect to these considerations, each of the above-mentioned reporter systems has attendant advantages and disadvantages. For example, the E. coli P-galactosidase must be distinguished from endogenous P-galactosidase activity and is generally less sensitive than CAT as a reporter (Padmanabhan, R. and Howard, B., unpublished results); on the other hand, an excellent in situ detection system for P-galactosidase based on the X-gal substrate is available (12). E. coli xanthine-guanine phosphoribosyltransferase is also somewhat less sensitive than CAT, but can be used simultaneously as a dominant selectable marker (12). Firefly luciferase is a novel activity in mammalian cells and offers a more sensitive assay than CAT, but at present requires relatively expensive equipment to obtain maximum sensitivity. This chapter will present details concerning the use of the CAT reporter system in mammalian gene transfer applications (seeNote 1). The

Chloramphenicol

Acetyltransferase

591

focus will be on the commonly used CAT enzyme assay developed by Shaw (13) and subsequently adapted by Cohen and coworkers (14). This assay measures acetylation of chloramphenicol at the l- and 3-positions of the carbon side chain. CAT initially catalyzes acetylation at the 3-position. The 3-acetoxy chloramphenicol can then undergo a nonenzymatic rearrangement that results in the formation of l-acetoxy chloramphenicol. In a second CAT catalyzed reaction, the l-acetoxy chloramphenicol can be reacetylated at the 3-position yielding 1,3-diacetoxy chloramphenicol(Z5,Z 6).
Chloramphenicol + AcCoA CAT 3-acetoxy chloramphenicol 1-acetoxy chloramphenicol 1,ddiacetoxy chloramphenicol + CoA + CoA

3-acetoxy chloramphenicol lacetoxy chloramphenicol

t-------3 CAT

+ AcCoA -

2. Materials
Acetyl &enzyme A, Lithium. Store at -7OOC. 14C-chloramphenico140-60 mCi/mmol. Store at -70C. Ethyl acetate. Baker flex silica gel 1B TLC plates. Chloroform. Methanol. Phosphate Buffered Saline (PBS) (seeAppendix, this vol.)-without Mg2+ or Ca2+. 8. TEN buffer: 0.04.M Tris HCl, pH 7.4/l mM EDTA/O.lSM NaCl. 9. 0.25M Tris HCl, pH 7.5. 10. 4 mM Acetyl Coenzyme A-Prepare in 0.25M Tris HCl, pH 7.5. It is preferable to prepare a fresh solution each time the assay is done; however, storage of the solution at -7OOCfor up to 2 wk is possible. 1. 2. 3. 4. 5. 6. 7.

3. Methods
1. Wash transfected cells in a T-25 tissue culture flask three times with 5 mL of PBS. 2. Add 1 mL of TEN, and let the flask sit at room temperature for 5 min. 3. Gently scrape the cells off the bottom of the flask with a tissue culture scraper, and transfer them to an appropriately labeled Eppendorf tube (seeNote 2). Immediately centrifuge the cells at 4OC, aspirate the supernatant, and resuspend the cell pellet in 0.10 mL of 0.25M Tris

592

Corsica and Howard HCl, pH 7.5. At this point, the suspension can be stored at -2OOC for later use. For cells grown in suspension, the washing can be done in centrifuge tubes. After the final wash, resuspend the cells in 1 mL of 0.25M Tris HCl, pH 7.5, and transfer to an Eppendorf tube. Lyse the cells by freeze-thawing them three times. To ensure complete lysis, cells can be sonicated briefly. A cuphorn apparatus is preferable to a sonication tip; if a tip is to be used, it is imperative that it be washed thoroughly in between each sample to prevent crosscontamination. Pellet the cellular debris by centrifugation at 12000 x g at 4OC for 10 min. Transfer the supernatant, being careful not to disrupt the pellet, to a new Eppendorf tube. At this point, the supernatant can be stored at -20C. The supernatants can now be assayed radiochemically. Place a new Eppendorf tube on ice and add the following: l-50 I.LL Extract 78-28 /.tL 0.25M Tris HCl, pH 7.5 4C-chloramphenicol W-J 20 /.tL 4 mM Acetyl Coenzyme A The final reaction vol is 100 pL; acetyl CoA is the last reagent that should be added to the reaction mixture. Generally, it is only necessary to assay 1 PL of an E. coli (pRSVcat) extract from a positive control sample (seeNote 5); 1 PL of extract can usually convert 80-90% of the 4C-chloramphenicol to its acetylated product in less than 1 h. Centrifuge the reaction mixture at 12000 x g for 15 s. Incubate the reaction mixture at 37C for 30 min. Incubation times range from as short as 10 min to as long as 18 h. To enhance the accuracy and the sensitivity of the assay, time points can be taken during the course of the incubation. By plotting the formation of acetylated 14C-chloramphenicol as a function of time, one can minimize random variation and ensure that all the assays are within the linear range. When taking time points, stop the reaction by adding 1 mL of ethyl acetate to the reaction mixture. Process the samples as outlined below. The reaction is terminated by adding 1 mL of ethyl acetate to the reaction mixture (seeNote 4).

4. 5.

6. 7. 8.

9.

10. 11.

12.

Chlorarnphenicol

Acetyltransferase

593

13. Vortex the reaction mixture vigorously for 30 s, and separate the phases by centrifugation for 30 s. 14. Remove 950 PL of the top organic layer, and transfer it to a fresh Eppendorf tube (seeNote 2). 15. Evaporate the ethyl acetate overnight in a fume hood. This step can be expedited by evaporating the samples in a Speed Vat concentrator (Savant). 16. Once the sample is thoroughly dried, resuspend it in 30 PL of ethyl acetate. Vortex vigorously for 30 s, and spot onto a TLC plate using a capillary pipet. Although not absolutely necessary, a preequilibrated TLC plate can be used. To prepare it, place a fresh TLC plate into a chromatography tank containing 95:5 chloroform:methanol. Once the solvent has run to the top of the plate, remove it from the tank and let it dry in a fume hood. Spot as indicated above. 17. The TLC plate is placed in a chromatography tank that has been preequilibrated with 95:s chloroform:methanol (see Note 3). The ascending chromatography should be completed in less than 3 h. The time will vary depending on the commercial source of plates used. 18. The plate is dried in a fume hood and then exposed to film overnight. 19. The conversion of 4C-chloramphenicol to its acetylated products can be quantitated by scintillating counting. Using the film as a guide, mark the acetylated andnonacetylated spots on the TLC plate. Cut the spots out from the plate, place into separate scintillation counting vials, and count. The percent conversion can be calculated as follows:
% acetylated Wchloramphenicol = 100 x acetylated counts total counts (acetylated + nonacetylated counts)

4. Notes
1. The focus of this chapter on the standard CAT assay notwithstanding, it should be emphasized that overall success in gene transfer experiments (seeChapter 50; Vol. 2, Chapter 25; and Vol. 4, Chapter 42 of this series) depends not only on the reporter system, but also on adherence to proper technique with respect to DNA plasmid preparation, tissue culture, and transfection protocols. In particular, the efficiency of transfection can vary over a wide range depending on the competence of recipient cells, which in turn is strongly dependent on tissue culture parameters. Readers interested in details concerning

594

Corsica and Howard gene transfer efficiency are referred to other discussions on that topic (l7,18). Streaking of spots on the TLC plates has several causes: a. Failure to avoid the interface when transferring the organic layer (Method, step 14); b. Inadequate drying of sample (Method, step 15); and c. Use of tubes from certain manufacturers (e.g., Sarstadt) during the ethyl acetateextraction (Method, steps 12 and 13)-we speculate that the polypropylene in some tubes has been treated to make it more hydrophilic, leading to more protein contamination after ethyl acetate extraction and subsequent gumming of the TLC. Abnormally slow migration of acetylated chloramphenicol derivatives (manifested by compression of those spots) is usually the result of inadequate solvent saturation of the atmosphere in the TLC tank. To prevent this problem, the inside of the TLC tank should be lined with chromatography paper. Place the paper so that the lower edge sits in the solvent. Ethyl acetate rapidly attacks polystyrene. This reagent should be transferred with glass pipets only. It is often useful to prepare a bacterial extract that contains CAT activity as a positive control for the assay. Bacteria carrying a plasmid that codes for chloramphenicol resistance (e.g., pBR325 or pRSVcat) provide a convenient source. Prepare a 5-mL overnight culture, collect in a 1.5-mL Eppendorf tube, lyse by several freeze-thaw cycles and brief sonication, and clear by centrifugation in a microfuge at 4OCfor 10 min. Aliquot and store at -2OOC. Low CAT activity in cell extracts can result from premature cell lysis with consequent loss of CAT during scraping of cells from the plate and centrifugation. One solution to this problem is to remove cells from the plate by trypsinization. Addition of serum and pelleting of cells eliminates most trypsin activity; moreover, trypsin concentrations as high as 0.025% have no observable effect on CAT activity under standard assay conditions (Holter, W., Howard, T., and Howard, B. H., unpublished results). CAT has been used as a reporter in extracts from a wide variety of mammalian tissue culture cells and in extracts from multiple tissues prepared from transgenic mice (19). In some extracts, however, inhibitors of CAT activity have been detected. These inhibitors have

2.

3.

4. 5.

6.

7.

Chloramphenicol

Acetyltransferase

595

not been characterized, but may degrade acetyl CoA or acetylated chloramphenicol derivatives. In most cases, such inhibitors can be inactivated by incubation of extracts at 65OC for 5 min (19,20). 8. Other methods for determining CAT activity that avoid requirements for Y-chloramphenicol and thin layer chromatography have been described (21,22).

References
1. Burke, J. F. and Mogg, A. E. (1985) Construction
2. of a vector, pRSVcatamb38, for the rapid and sensitive assay of amber suppression in human and other mammalian cells. Nucleic Acids Res. 13,1317-1326. Capone, J. P., Sedivy, J. M., Sharp, P. A., and RajBhandary, U. L. (1986) Introduction of UAG, UAA, and UGA nonsense mutations at a specific site in the Escherichia coli chloramphenicol acetyltransferase gene: use in measurement of amber, ochre, and opal suppression in mammalian cells. Mol. Cell. Bid. 6,3059-3067. Protic-Sabljic, M. and Kraemer, K. H. (1986) Host cell reactivation by human cells of DNA expression vectors damaged by ultraviolet radiation or by acid-heat treatment. Carcinogenesis 10,1765-1770. Gaynor, R. B., Feldman, L. T., and Berk, A. J. (1985) Transcription of class III genes activated by viral immediate early proteins. Science 230,447-450. Gorman,C.M.,Moffat,L.F.,andHoward,B.H. (1982)Recombinantgenomes which express chloramphenicol acetyltransferase in mammalian cells. Mol. Cell. BioZ. 2, 1044-1051. Hall, C. V., Jacob, P. E., Ringold, G. M., and Lee, F. (1983) Expression and regulation of Escherichia coli 1acZ gene fusions in mammalian cells. J. Molec. Appl. Genef. 2, 101-109. Mulligan, R. and Berg, P. (1980) Expression of a bacterial gene in mammalian cells. Science 209,1422-1427. Chu, G. and Berg, P. (1985) Rapid assay for detection of Escherichia coli xanthineguanine phosphoribosyltransferase activity in transduced cells. Nucleic Acids Res. 13,2921-2930. Schumperli, D., Howard, B. H., and Rosenberg, M. (1982) Efficient expression of Escherichia coli galactokinase gene in mammalian cells. PYOC. NafI. Acad. Sci. USA 79, 257-261. de Wet, J. R., Wood, K. V., DeLuca, M., Helinski, D. R., and Subramani, S. (1987) Firefly luciferase gene: structure and expression in mammalian cells. Mol. CeII. Biol. 7,725-737. Price, J,, Turner, D., and Cepko, C. (1987) Lineage analysis in the vertebrate nervous system by retrovirus-mediated gene transfer. PYOC.Nafl. Acad, Sci. USA 84,156-160. Mulligan, R. and Berg P. (1981) Selection for animal cells that express the Escherichia coli gene coding for xanthine-guanine phosphoribosyltransferase. Proc. Nafl. Acad. Sci. USA 78,2072-2076. Shaw, W. (1967) The enzymatic acetylation of chloramphenicol by extracts of R factor-resistant Escherichia coli. J. Biol. Chem. 242,687-693.

3. 4. 5.

6.

7. 8.

9.

lO#

11. 12.

13.

596

Corsica and Howard

14. Cohen, J. D., Eccleshall, T. R., Needleman, R. B., Federoff, H., Buchferer, B. A., and Marmur, J. (1980) Functional expression in yeast of the Escherichia coli plasmid gene coding for chloramphenicol acetyltransferase. Proc. Natl. Acad. Sci. USA 77,1078-1082. 15. Shaw, W. V. (1983) Chloramphenicol acetyltransferase: enzymology and molecular biology. CRC Crit. Rev. B&hem. 14,146. 16. Kleanthous, C. and Shaw, W. V. (1984) Analysis of the mechanism of chloramphenicol acetyltransferase by steady-state kinetics. Evidence for a ternary-complex mechanism. Biochem. J. 223,211-220. 17. Gorman, C., Padmanabhan, R., and Howard, B. H. (1983) High efficiency DNAmediated transformation of primate cells. Science 221,551-553. 18. Fordis, C. M. and Howard, B. H. (1987) Use of the CAT reporter gene for optimization of gene transfer into eucaryotic cells. Methods Enzymol, 151,382-397. 19. Overbeek, P. A., Chepelinsky, A. B., Khillan, J. S., Piatigorsky, J., and Westphal, H. (1985) Lens-specific expression and developmental regulation of the bacterial chloramphenicol acetyltransferase gene driven by the murine alpha A-crystallin promoter in transgenic mice. Proc. Natl. Acad. Sci. USA 82,7815-7819. 20. Mercola, M., Goverman, J., Mire& C., and Calame, K. (1985) Immunoglobulin heavy-chain enhancer requires one or more tissue-specific factors. Science 227,

266-270.
22. 22. Sleigh, M. F. (1986) A nonchromatographic assay for expression of the chloramphenicol acetyltransferase gene in eucaryotic cells. Anal. Biochem. 156,251-256. Young, S. L., Jackson, A. E., Puett, D., and Melner, M. H. (1985) Detection of chloramphenicol acetyltransferase in transfected cells: a rapid and sensitive HPLCbased method. DNA 4,469-475.

Chapter Immunizing for Hybridoma

Bjiirn

45 Schedules Production

Gustafsson

1. Introduction
Immunization protocols for generating activated B-lymphocytes suitable for hybridoma production vary widely. The optimal amount of antigen given, the way of antigen presentation, and the timing between injections must be determined for each antigen. Often, the antigen of interest is not available in a pure form, and the influence of contaminating antigens in the preparation should be considered when the immunization schedule is designed. Generally, a lower immunization dose will give rise to antibodies with higher specificity as compared to the total antibody response obtained after immunization with higher doses. Furthermore, antibodies obtained after one or two injections often show lower affinity than those obtained when a prolonged immunization schedule is used. The number of antibody-producing hybridomas obtained after fusing Blymphocytes with myeloma cells depends on how well the mouse was immunized, and it is therefore worthwhile to put some effort into optimizing the immunization schedule. 597

598

Gustafsson

2. Materials
1. 2. 3. 4. Antigen, dissolved in O.OlM phosphate-buffered Freunds complete adjuvant (FCA). Freunds incomplete adjuvant (FlA). Female BALB/c mice, 6-8 wk of age. saline, pH 7.2.

3. Methods
3.1. Soluble
1.

Antigens

Mix lo-100 pg of antigen in FCA (1:l; vol:vol). The antigen solution is sucked into a glass syringe and mixed with FCA in a test tube by pushing the mixture back and forth. Continue until a thick creamy solution is obtained. If the solution is properly mixed, it should remain as one phase after a 1 h incubation at 4OC. If two phases are obtained, the solution was not properly homogenized. 2. Inject 50-100 PL intraperitoneally (ip) to each mouse. Two to four weeks later the same dose of antigen is given ip, but this time it is mixed with FIA. Two to four weeks later, the mice are boosted with 100-200 pg ip without adjuvant. Continue to immunize until a satisfactory immune response is obtained. Alternatively, the animals could be immunized subcutaneously with 50 yL at each of four lymph node sites. Immunization in the footpads causes great pain to the animal and should be avoided. For highly immunogenic antigens, it is possible to omit adjuvants which is also more lenient to the animal. Fuse spleen cells with myeloma cells 34 d after the last booster dose.

3.2. Cellular

Antigens

1. Inject approximately 2 x lo7 cells ip. Boost the animals at 34 wk intervals with the same or a lo-fold higher dose. In general, the animals should be rested for 4 wk before the final booster dose. 2. When optimizing the immunization schedule, the development of a specific immune response should be measured by a method suitable for rapid screening of multiple samples. Collect serum samples by retrobulbar punction (eye-bleeding) or tail-bleeding 1 wk after each injection, and measure the increase of specific antibodies in enzyme-linked immunosorbent assay, radio-immunoassay, or some other method suitable for rapid screening of multiple samples (see Chapter 48).

Immunization

for Hybridoma

Production

599

4. Notes
1. Individual animals may sometimes respond differently to the same immunization schedule. When optimizing the immunization program, each schedule should be tested in groups containing at least five mice/group. 2. Adjuvants are used to enhance the immune response to molecules that are poorly immunogenic. FCA is a mixture of mineral oils containing killed and dried Mycobactevium tubercu2c~is bacteria. FIA is devoid of the bacteria. Both FCA and FIA cause local inflammatory response at the site of injection. Precautions should be made to avoid accidental injections into the hand or splashing of droplets into the eye, since this may cause irreparable damage. 3. Collecting serum by eye-bleeding should be performed only on sufficiently anesthesized mice under the supervision of a person who is thoroughly skilled in this technique.

Chapter 46
Fusion Protocol for the Production of Mouse Hybridomas
Bj6rn Gustafsson

1. Introduction
Hybridoma technology makes it possible to produce almost unlimited amounts of monospecific antibodies. Each hybridoma represents only one of several B-lymphocytes responding to a particular antigen, and for this reason, monoclonal antibodies can also be produced against impure antigen preparations. Monoclonal antibodies are produced by hybridomas derived from the successful fusion of B-lymphocytes and mouse myeloma cells (1). Most myeloma cell lines used in hybridoma work are mutants lacking the enzyme hypoxanthine guanine phosphoribosyl transferase (HGPRT), which is utilized in the hypoxanthine-aminopterin-thymidine (HAT) selection system (2). Aminopterin blocks the main biosynthetic pathways for DNA and RNA synthesis. By providing thymidine and hypoxanthine in the growth medium, cells can usually continue to multiply by using the salvage pathway. However, a HGPRT- myeloma cell line will not grow, since the cell cannot utilize hypoxanthine supplied with the medium. Only hybrids formed between a myeloma cell (HGPRT-), and a B-lymphocyte (HGPRT) will survive and multiply. 601

602

Gustafsson

The production of hybridomas begins by immunizing a mouse with a selected antigen (seeFig. I). B-lymphocytes from the spleen are prepared and mixed with myeloma cells in the presence of polyethylene glycol to promote fusion. The cells are suspended in a growth medium containing HAT. Unfused myeloma cells and myeloma-myeloma hybrids, which lack the enzyme HGPRT, cannot survive in the HAT-medium. Unfused spleen cells and spleen-spleen hybrids die naturally after a few replications. The only surviving cells are hybrids of myeloma cells and B-lymphocytes. These hybridomas are screened for specific antibody production, and hybrids of interest are cloned. Following recloning, antibody-producing hybridomas are expanded, and the cells are frozen in liquid nitrogen for future use. A hybridoma cultivated in tissue culture bottles produce l-50 pg/mL of monoclonal antibodies. For large-scale production, the hybridomas are injected into the peritoneal cavity of mice, in which they produce ascites tumors that generate monoclonal antibodies in concentrations of l-10 mg/mL. The entire process of establishing antibody producing hybridomas takes 34 mo. A number of HAT-sensitive mouse myeloma cell lines are available for hybridoma production. However, many myelomas produce their own immunoglobulin, and this makes themless suitable as fusion partners. The optimal myeloma should be a nonimmunoglobulin-producing cell line with a high capacity of hybrid formation. The cell line Sl?2/0-Ag 14 (3) is such a myeloma and is commonly used for the generation of hybridomas. Hybridoma production involves a large number of steps requiring patience and strict adherence to sterile tissue culture technique. This and following chapters (46-51) describe the basic methods required for the successful production and characterization of hybridomas.

2. Materials
1. RPM1 1640 tissue culture medium. Store at 4OC. 2. Fetal calf serum. Store at -20C. 3. Stock solutions of 100 x L-glutamine: 2.92 g L-glutamine dissolved in 100 mL distilled water. 4. Stock solutions of 100 x sodium pyruvate: 1.l g sodium pyruvate dissolved in 100 mL distilled water. 5. Stock solution of 100 x PEST: 10,000 U pencillin and 10 mg streptomycin/ml of distilled water.

Hybridoma

Fusion Protocol

603
MYELOMA CELL CULTURE

IMMIJNI~ATI~N

SPLEEN

CELLS

HGPRT+

FUSION

IN PEG

SELECTION OF HYBRID IN HAT-MEDIUM

CELLS

ASSAY A

FOR ANTIBODY

4 13 CLONE @jj'b z& HYBRIDS ANTIBODY-PRODUCING

FREEZE HYBRIDOMA FOR FUTURE USE MONOCLONAL ANTIBODY

Fig. 1. The generatioxuof

hybridomas

and production

of monoclonal

antibodies.

6. Stock solution of 50 x I-IT solution: Dissolve 68 mg of hypoxanthine and 19.5 mg of thymidine in 100 mL of distilled water under gentle heating. If the hypoxanthine does not dissolve readily, add a few drops of LOM NaOH. Adjust to pH 7.2 with 1M HCl when the hypoxanthine is solubilized.

Gustafsson 7. Stock solution of 50 x HAT solution: Dissolve 0.9 mg of aminopterin in 100 mL of 50 x HT solution. All stock solutions should be sterilized by membrane filtration and stored in aliquots at -20C. 8. Standard medium: 500 mL of RPM1 1640 tissue culture medium is supplemented with 50 mL of fetal calf serum, 5 mL of L-glutamine, 5 mL of sodium pyruvate, and 5 mL of PEST. 9. HAT medium: Add 2 mL of HAT (50x) stock solution to 100 mL of standard medium. 10. 50% PEG solution: 1 g polyethylene glycol4000 (PEG) in a test tube is autoclaved for 20 min. While the PEG is still liquid, add 1 mL of RPM1 1640 medium.

3. Methods
1. Give the immunized mouse a final booster dose without adjuvant, 3-4 d before fusion. 2. Dilute the myeloma cells the day before fusion to a concentration of 2-3 x lo5 cells/mL. 3. Kill the mouse by cervical dislocation or CO, asphyxiation. 4. Swab the mouse with 70% ethanol and remove the spleen, aseptically. Place the spleen in a Petri dish containing 5-mL of RPM1 1640 medium. 5. Put a sterile sieve on top of a Petri dish containing 5-mL medium. Prepare a single-cell suspension by placing the spleen on the sieve and gently tease the spleen through the sieve using forceps and spatula. All handling of cells should be performed in a laminar flow hood. 6. Transfer the spleen cells to a conical tube. Exclude large clumps of tissue. Allow the tube to stand for l-2 min, while clumps settle out. Transfer the cell suspension to a new tube and centrifuge at 300 xg for 5 min. Wash the cells once in RPM1 1640 medium and count them using a cell counting chamber. Approximately lo* cells are obtained from a mouse spleen. 7. Harvest the myeloma cells by centrifugation at 300 x g for 5 min, wash once in RPM1 1640 medium, and count. 8. Mix the myeloma cells with the spleen cells in a 50-mL centrifuge tube. The ratio of myeloma cells: spleen cells may range from l:l-1:lO. Centrifuge the mixed cell suspension at 300 x g for 5 min and discard the supernatant.

Hybridoma

Fusion Protocol

9. Add 1 mL of the PEG solution (37OC) to the pellet over 1 min, using a I-mL pipet, continuously stirring the pellet with the tip. Continue with gentle stirring for 1 min. 10. Dilute the fusion mixture by adding 1 mL of RPM1 1640 medium (37C) dropwise, over 1 min. Stir gently to ensure an even dilution. 11. Repeat step 10. 12. Add 10 mL of medium over 5 min to the mixture. Centrifuge the suspension at 300 x g for 5 min and discard the supernatant. Resuspend the pellet gently in 30-40 mL of HAT-medium and distribute the cells into four 96-well tissue culture trays, at a concentration of approximately 2.5 x lo5 cells/well. 13. Incubate the trays at 37OC,5% CO, 80% humidity in a CO,-incubator. 14. Feed the cells twice a week by removing half of the medium in each well using sterile Pasteur pipets connected to a filter pump, taking care not to disturb the cells. Add fresh HAT-medium to the wells by gently dropping medium from a 10-mL pipet. A color change of the medium from red to yellow is indicative of cell growth. Visible clones will appear within 2-3 wk.

4. Notes
1. Stock solutions of medium supplements are commercially available. 2. The myeloma cells must be in exponential growth prior to fusion in order to obtain high yields of hybrid clones. Keep the cells at a density of 2-6 x l@ cells/mL. The cells should never be allowed to exceed 8 x 105cells/mL. The cells can be cultured in the same bottles for several months under the condition that they are split and fed with fresh standard medium regularly. The size of a B-lymphocyte is about onetenth of a myeloma cell, and the two cell types are easily distinguished in the microscope. Hybridomas and myeloma cells are of the same size and appear in the microscope as round, heavily light-scattering cells (seeFigs. 2 and 3). 3. Most myeloma cells grow as suspension cultures. However, a certain percent of the cells attach to the plastic of the bottle. By vigorously shaking the bottle, the adherent cells are released from the plastic surface. The cells are split by removing all the spent medium and replacing it with fresh standard medium. There will still be enough cells left in the bottle to ensure the culture.

606

Gustafsson

Fig. 2. Hybridoma cells in tissue culture (100 x magnification).

Fig. 3. Hybridoma

cells in tissue culture (400 x rnagnificationh

Hy bridoma Fusion Protocol

607

References
2. K&ler, G. and Milstein, C. (1975) Continuous cultures of fused cells secreting antibody of predefined specificity. Nature 256,495-497. 2. Littlefield, J,W. (1964) Selection of hybrids from matings of fibroblasts in vitro and their presumed recombinants. Science 145,709. 3. Schulman, M., Wilde, C., D., and Kiihler, G. (1978) A better cell line for making hybridomas secreting specific antibodies. Nature h-m.lon) 276,269,270.

Chapter 47
Cloning
Bjiirn

of Hybridomas
Gustafsson

1. Introduction
Hybrid clones will appear within 2-3 wk after fusion (seeChapter 46). It is of utmost importance that a newly established hybridoma is cloned thoroughly to ensure that the cells growing in the tissue culture are of monoclonal origin and not a mixture of two or more hybridomas. A mixture of cells may result in a gradual decline of specific antibody production because of overgrowth of contaminating hybridomas, until finally the hybridoma of interest is lost. Culture supernatants from the hybridomas should be tested for specific antibody production by appropriate methods. Hybridomas are cloned by limiting dilution using thymocytes as feeder cells as soon as the cell density in the well reaches one-half to two-third confluence. The hybridomas will have previously been cultivated in HAT medium. Before the cells can be switched to standard medium, they must be cultivated in HT medium for 2 wk. If the HAT constituents are excluded from the medium, there will be a risk that the hybrids will die out because of remaining amounts of aminopterin in the cells. By maintaining the cells on HT medium for a couple of weeks, they can still use the salvage pathway for nucleic acid synthesis while aminopterin is metabolized. In order to avoid accidental feeding of cells with HT medium instead of HAT medium

610

Gustafsson

and vice versa, we routinely add HAT medium to all fusion trays, I-IT medium to all cloning trays, and standard medium to all tissue culture bottles.

2. Materials
1. Standard medium, HAT medium and sterile tissue culture equipment (seeChapter 46). 2. HT medium: Add 2 mL of HT (50x) stock solution to 100 mL of standard medium.

3. Methods
1. A mouse (BALB/c), 3-4 wk of age, is killed by CO, asphyxiation. 2. Swab the mouse with 70% ethanol and remove the thymus, aseptically. Prepare a single-cell suspension of thyrnocytes as described for splenocytes (seeChapter 46, points 4-6). Approximately lo8 thymocytes are obtained from a young mouse. Use one thymus for each hybridoma to be cloned. 3. Suspend the thymocytes in 10 mL of HT medium. Distribute 0.1 mL of the cell suspension to each well of a 96-well tissue culture tray. 4. Remove the hybridoma cells from the microtiter well with a sterile Pasteur pipet by gently flushing the medium up and down a couple of times, and transfer the cells to a sterile test tube. Add fresh HAT medium to the well, and put the tray back into the CO, incubator. 5. Count and dilute the cells in I-IT medium to a concentration of 50 and 5 cells/mL, respectively. Add 0.1 mL of the suspension containing 50 cells/mL to eight wells in the thymocyte-conditioned tray and 0.1 mL of the suspension containing 5 cells/mL to the rest of the wells. Incubate the tray in the CO, incubator. 6. Feed the cells twice a week with HT medium. Visible clones appear within 1-2 wk. A successful cloning should not show growth in more than approximately 50% of the wells. 7. The clones are screened for antibody production, and hybridomas of interest are expanded gradually from the 96-well tissue culture well, over 24-well trays, to 25- and 80-cm2 tissue culture flasks. The hybridoma cells can be cultivated in the same tissue culture flasks for several months if they are handled properly. Monoclonal antibodies can be harvested twice a week from such cultures by simply replacing spent medium, containing antibodies, with fresh medium. Newly estab-

Cloning of Hybridomas

611

lished hybridomas should be frozen and stored in liquid nitrogen as soon as possible.

4. Notes
1. Hybridoma cells grow poorly at low cell densities (< lo4 cells/mL), and growth is supported by the use of feeder cells. Thymocytes are effective as feeder cells, although peritoneal macrophages or splenocytes may be used equally well. Feeder cells are short-lived in culture, but provide necessary growth supporting factors during the early stages of growth of the hybridoma cells. 2. Cloning of hybridomas should routinely be performed twice, and monoclonality should be verified by different methods, e.g., isoelectric focusing of monoclonal antibodies, immunoglobulin isotype determination, and/or titration of antibodies produced during prolonged cultivation of hybridomas. The frequency of chromosome loss decreases with time in culture, but may still occur during long-time cultivation. Hybridomas that are cultivated over longer periods should be tested for antibody production and recloned if the titer drops significantly. 3. Scaling up of hybridomas may occasionally be troublesome, since some hybridomas do not always transfer well to growth bottles. The easiest way to get the cells started is to incubate the flask in an upright position for 24 h and then continue cultivation with the flask lying down. Feeder cells may help. Keep the cells in log-growth phase (3-6 x lo5 cells/mL). They should never exceed lo6 cells/ml. Overgrowth may result in a dead culture.

Chapter 48 Enzyme-Linked Immunosorbent Assay for Screening of Antibodies in Hybridoma Supernatants

Bjiirn Gustafsson

1. Introduction
Enzyme-linked immunosorbent assay (ELISA) is a rapid and convenient method for screening of antibody producing hybridomas (1,2). The method is highly sensitive and can be applied to detect antibodies directed against soluble antigens as well as cell-bound antigens. Over a thousand culture supernatants can be tested in one day if the assay is performed in microtiter trays. In principle, the soluble antigen (protein, glycoprotein, lipopolysaccharide) is attached to the plastic by physical adsorption. Whole-cell antigens may have to be bound by poly-L-lysine to the plastic followed by fixation with glutaraldehyde. Remaining binding sites are blocked by bovine serum albumin (BSA). Culture supernatants are added to the wells followed by incubation with enzyme-conjugated rabbit antimouse immunoglobulin antibodies. Unbound conjugate is washed away, and an enzyme substrate is added to each well. The amount of colored product 613

614
Antigen-ELISA,

Gustafsson

C-------

Substrate

Enzyme-ab conjugate

Monoclonal antibody

-e
Fig. 1. ELBA principle.

BSA

formed is proportional to the amount of antibodies produced by the hybridoma cells (seeFig. 1).

2. Materials
1. Coating buffer: 0.05M carbonate buffer, pH 9.6; 1.56 g Na.&03, 2.93 g NaHCO, 0.2 g NaN, in 1000 mL of distilled water. 2. PBS (stock): Phosphate-buffered saline (PBS), pH 7.2; 36.04 g Na,HPO,*Z&O, 9.18 g KHJ?O,, 145.8 g NaCl, 0.6 g NaN, in 3 L of distilled water. 3. PBS (15 mM; working dilution): 500 mL PBS (stock) + 2.5 L of distilled water. 4. Washing buffer: 0.05% Tween-20 in PBS.

Enzyme-Linked

Immunosorbent

Assay

615

5. Blocking buffer: 1% bovine serum albumin (BSA), 0.05% Tween-20 in PBS. 6. Substrate solution: Dissolve 55 mg of 1,2-phenylenediaminedihydrochloride in 100 mL of 40 mM Tris-HCl, pH 7.6. Add 40 yL of 30% H202 immediately before use. 7. 96-well flat bottom microtiter trays (ELISA grade). 8. Conjugate: Peroxidase-conjugated rabbit anti-mouse immunoglobulin antiserum. 9. Stop solution: 1M H$04. 10. Microplate spectrophotometer. 11. Cell bound antigens: a. 10 pg/mL poly+lysine in PBS. b. 0.5% glutaraldehyde in cold PBS. Prepare immediately before
use. C.

0,lM glycine, 0.1% bovine serum albumin in PBS.

3. Methods 3.1. Soluble

Antigens

1. Add 100 JJL of antigen (l-10 pg/mL in coating buffer) to the wells. Incubate overnight at 22OC. The trays can be stored at 4OC. 2. Wash three times with washing buffer. Add 150 PL of BSA-solution and incubate for 15 min at 22C. 3. Wash. Add 100 PL of hybridoma culture supernatant, and incubate for 1 h at 37OC. 4. Wash. Add 100 JJL of conjugate diluted according to manufacturers instructions. Incubate for 1 h at 37C. 5. Wash. Add 100 FL of substrate solution. Incubate for 10 min at 22C. 6. Stop the reaction by adding 50 PL of 1 M H,SO, and measure the absorbance at 492 nm.

3.2. Cell Bound

Antigens

1. Add 100 PL of poly-L-lysine solution to each well. Incubate for 30 min at 22C. 2. Aspirate. 3. Prewash the cells in PBS. Add lo6 cells/well in 100 PL of PBS. 4. Centrifuge the trays at 200 x g for 5 min.

616

Gustafsson

5. Add 100 l.tL of cold glutaraldehyde solution to each well. Incubate for 15 min at 22C. 6. Wash twice with PBS. Add 125 PL of PBS-glycine-BSA solution, incubate for 30 min at 22C, and then follow the procedure described for soluble antigens, steps 3-7. 7. If the trays are to be stored: wash twice with PBS. Add 100 u.L of 0.1% BSA in PBS, then store at -20C.

4. Notes
1. Carbonate buffer is the standard buffer used for coating of antigen to microtiter trays. In our experience however, PBS often works equally well. Microtiter trays made of polystyrene are commonly used. If the antigen sticks poorly to the plastic, it may be worthwhile to try flexible polyvinyl trays. Sterilized (irridiated) microtiter trays often show higher affinity to the antigen, although sometimes accompanied by an increased background level. 2. The enzyme reaction can be performed using different enzymeantibody conjugates. Horseradish peroxidase (HRP) and alkaline phosphatase (ALP) are the most commonly used enzymes. HRP is popular because of its rapid formation of colored product. However, the substrate o-phenylenediamine has been suggested to be a carcinogen, and gloves should be used when handling this chemical. 3,3,5,5-tetramethylbenzidine is not a known carcinogen and can replace o-phenylenediamine as substrate in the HRP-reaction. When using this substrate, 10 mg of 3,3,5,5-tetramethylbenzidine is dissolved in 1 mL dimethylsulfoxide, and 100 PL of this solution is added to 10 mL O.lM sodium acetate, pH 6.0. I!-$02 is added to a final concentration of 0.01% immediately before use. The ALP reaction is equally sensitive, but somewhat slower. None of the enzymes should be used if endogenous activity of these enzymes in the antigen is expected. Other enzymes that can be used are glucose oxidase, galactosidase (3), and urease (4). 3. All ELISAs must be optimized for each antigen, and the instructions given above should only be regarded as guidelines. The use of polyL-lysine and glutaraldehyde fixation in whole cell ELISAs may sometimes be omitted, and the number of cells/well may have to be altered in order to obtain optimal sensitivity.

Enzyme-Linked

Immunosorbent

Assay

617

References
1. Engvall, E. and Perlman, l? (1972) Enzyme-linked immunosorbent assay, ELISA. III. Quantitation of specific antibodies by enzyme-labeled anti-immunoglobulin in antigen coated tubes. 1. Irnmunol. 109,129-135. 2. Gustafsson, B. (1984) Monoclonal antibody-based enzyme-linked immunosorbent assays for identification and serotyping of Vibrio cholerae 01. J. Clin. Microbial. 20, 1180-1185. 3. Chandler, H. M., Cox, J. C., Healey, K., McGregor, A., Premier, R. R., and Hurrell, J. G. R. (1982) An investigation of the use of urease-antibody conjugates in enzyme immunoassays. J, Immunol. Methods 53,187-194. 4. Kincade, P. W., Lee, G., Sun, L., and Watanabe, T. (1981) Monoclonal rat antibodies to murine IgM determinants. J. Zmmunol. Methods 4217-26.

Chapter 49
Cryopreservation of Hybridomas
Bjiirn Gustafsson

1. Introduction
Hybridomas are exposed to many threats, such as contamination with bacteria and fungi, loss of chromosomes coding for antibody production, overgrowth by nonsecreting mutants, and cell death resulting from overgrowth. Therefore, newly established hybridomas should be frozen and stored in liquid nitrogen at -196OC as soon as possible. Alternatively, they can be stored in the vapor phase above the liquid nitrogen where the temperature will be between -120 and -195OC. Generally, eucaryotic cells are very sensitive to the freezing procedure, and programmable freezing devices are available that lower the temperature from 0 to -3OOC and from -30 to-100C at different time rates. However, in our experience, hybridoma cells will survive by simply incubating the vials at -2OOC for 2 h followed by overnight incubation at -70C. The vials are then transferred to liquid nitrogen for long-time storage, Cells frozen in this way have survived for more than 5 yr. However, sometimes hundreds of hybridomas are obtained after a successful fusion experiment. Since it is not possible to clone all hybridomas within the time available before the cultures become overgrown, one can freeze hybridomas directly in the culture trays (1). This may also be useful when time-consuming screening protocols are necessary. Culture supernatants can be analyzed, while the cells are preserved frozen.
619

Gustafsson

2. Materials
1. 1.0 mL cryopreservation vials. 2. Liquid nitrogen container . 3. Freezing medium: 10% dimethylsulfoxide

(DMSO) in fetal calf serum.

3. Methods
3.1. Freezing

Cells in Cryotubes

1. Day 1. Remove half of the cell suspension, and add fresh standard medium (seeChapter 46). 2. Day 2. Centrifuge the cells at 300 xg for 10 min. Resuspend the cells in ice-cold freezing medium at a concentration of 2-5 x IO6 cells/mL. 3. Put 1 mL of cell suspension into the cryopreservation vials. Incubate immediately at -2OOCfor 2-3 h, and then transfer the vials to -7OOCfor overnight incubation. 4. Day 3. Put the vials in the liquid nitrogen container. To reconstitute frozen cells, carry out the following procedure: 1. Day 1. Thaw rapidly by placing the vial in a beaker of water at 37OC. As soon as the cells are liquified, wash the cells in 10 mL of cold standard medium (seeChapter 46). Resuspend the cells in 5 mL of standard medium, and transfer the cells to a 50-mL tissue culture flask. 2. Day 2. Check the cells for viability. Split if necessary or transfer to a larger tissue culture flask.

3.2. Freezing of Hybridomas in 96-Well Microtiter Tkays

1. Day 1. Remove half of the medium from the wells, and add fresh medium. 2. Day 2. Remove all medium from the wells. Add 0.1 mL of ice-cold freezing medium to each well, and wrap the tissue culture tray in plastic. Incubate at -2OOC for 2 h, and then at -7OOC. Cells frozen in trays have remained viable for at least 2 mo. To reconstitute the cells, carry out the following steps:

Cryopreservation

of Hybridomas

621

1. Add 0.1 mL of RPM1 1640 medium, prewarmed to 37OC. 2. Aspirate the freezing medium as soon as it is liquified. Add HAT-medium to the wells (seeChapter 46). Centrifuge the trays at 300 xg for 5 min. to the wells. 3. Aspirate all medium from the wells. Add HAT-medium Incubate the trays in the CO, incubator.

4. Notes
1. Freeze only cells that are in mid-log phase (2-5 x lo5 cells/mL). Rapidly growing cultures can be identified because they contain cells of varying size. 2. Freeze several vials each time. Reconstitute one vial after a couple of days, and check for viability (by cultivating the cells for a couple of days) that the freezing was successful before discharging the cell cultures. 3. Recently thawed cells may sometimes have difficulties in getting started. Such cultures may improve by increasing the fetal calf serum content to 20%. Once they are started, they can be switched back to 10%. 4. An alternative to the freezing procedure described above is to put the vials in a polystyrene box with l-cm thick walls, and put the box overnight in -70C before transferring the vials to liquid nitrogen.

Reference
I. Wells, D. E. (1983) Simple rapid method for freezing hybridomas in 96-well microculture plates. 1. Immunol. Melhods 59,49-52.

Chapter

50

Isotype Determination of Monoclonal Antibodies by Immunodiffusion

Bjiirn Gustafsson

1. Introduction
Immunodiffusion is an important qualitative method for the detection of antigens and antibodies. The technique may be used to determine the immunoglobulin heavy and light chain class and subclass of hybridoma antibodies. The monoclonal antibodies are allowed to diffuse toward anti-mouse isotype specific antisera in an agarose gel. In a positive reaction, a precipitin line will develop in the agarose between two wells. A monoclonal antibody should show only one heavy chain and one light chain. Knowledge of the isotype is also useful for purification purposes.

2. Materials
1. Cell culture supernatants, lo-fold concentrated by ammonium sulfate precipitation. 2. Specific antisera against mouse IgM, IgA, IgGl, IgG2a, IgGILb, IgG3, kappa and lambda light chain. 623

624 3. 4. 5. 6. 7. 8.

Gustafsson

1% agarose in 15 mM phosphate-buffered saline (PBS), pH 7.2. Plastic film (Gel BondTM). Gel pouch (4 mm). PBS. Washing solution: 0.9% NaCl in distilled water. Staining solution: 1.5 g Coomassie Brilliant Blue R250, in 300 mL destaining solution. Stir for 1 h, 22OC. Filter before use. The solution can be used several times. 9. Destaining solution: 350 mL 95% ethanol, 100 mL glacial acetic acid. Make up to 1000 mL with distilled water. The destaining solution may be reused if the dye is first removed by passing through a funnel of activated charcoal.

3. Methods
3.1.

Immunodiffusion

1. Add 1 g of agarose to 100 mL PBS. Heat to boiling while stirring. When the agarose is completely dissolved, transfer to a 60C waterbath. 2. Cut the plastic film to appropriate size. 3. Place the plastic film on a horizontal table with the hydrophilic side facing upward (check with a drop of water). 4. Pour the melted agarose solution onto the film (0.15 mL/cm2). Allow the agarose to solidify (lo-15 min). 5. Stamp and suck out wells, 4-5 mm apart in a circle and with one well in the center. 6. Fill the center well (approximately 15 FL) with mouse isotypespecific antisera and the outer wells with concentrated hybridoma supernatants. Incubate overnight at 22C, in a moist chamber. 7. Check for precipitin lines on a light box, or stain in Coomassie Blue as described below.

3.2.

Staining

of Agarose

Gels

Precipitates may sometimes be difficult to discover. However, the precipitates will be visualized by staining the gels with a suitable reagent. 1. Flush the gel, and fill the wells with distilled water. 2. Cover the gel with a wet filter paper (Whatman No. 1). Make sure that no air bubbles are trapped between gel and filter.

Monoclonal

Antibodies

625

3. Cover the filter paper with a 2-cm layer of cellulose wadding, and place a thick glass plate on top of the cellulose wadding. Apply a lo-20g/cm2 pressure on the plate for 15 min. 4. Remove the cellulose. Flush the filter paper with distilled water, and remove it carefully from the gel. 5. Wash the gel twice in 0.9% NaCl for 20 min, and then once in distilled water for 10 min. 6. Repeat steps 24. 7. Dry the gels with a heating fan (hair dryer). 8. Put the gel in the staining solution for 5 min at 22C. 9. Transfer the gel to the destaining solution. Stir gently. Change the destaining solution once or twice. Protein will retain the dye, whereas the agarose will be destained. 10. Dry the gel and check for precipitin lines.

4. Notes
1. Monoclonal antibodies can be concentrated by adding an equal vol of saturation (NHJ,SO, to the culture supernatant. Allow the antibodies to precipitate for 1 h, followed by centrifugation at 10,000 x g for 10 min. Dissolve the pellet in distilled water to one-tenth of the original vol, and apply the sample directly to the agarose gel without prior dialysis. 2. Thin gels, e.g., isoelectric focusing gels, can be washed in 0.9% saline, overnight, followed by washing in 95% ethanol for 10 min. When the gel is dry, it can be stained according to steps B-10 described above.

Chapter 51
Isoelectric Focusing and Immunofixation of Monoclonal Antibodies
Bjiirn Gustafsson

1. Introduction
Isoelectric focusing of monoclonal antibodies in agarose gel is one of several ways to characterize a hybridoma. Isoelectric focusing is performed on each monoclonal antibody, and the mouse immunoglobulin bands are immobilized by soaking the gel with rabbit antimouse immunoglobulin before staining with Coomassie Blue. A single band obtained after isoelectric focusing and immunoprecipitation is a strong indication of monoclonality. However, results from isoelectric focusing are sometimes difficult to interpret, since amonoclonal antibody often show multiple bands. Nevertheless, these bands often tend to show a characteristic grouping that also remains constant after extensive recloning (I).

2. Materials
1. Electrofocusing apparatus. 2. Thermostatic circulating cooling bath (flow capacity > 6 L/min). 3. Agarose, IEF-grade. 627

Gustafsson
4. 5. 6. 7. 8. 9. 10. 11. 12.

13. 14. 15. 16.

Ampholytes, pH 3.5-9.5. pH surface electrode (optional). Glass plates 250 x 125 x 3 mm and 270 x 125 x 3 mm. Spacers, 250 x 6 x 0.5 mm. Plastic film (Gel-BondTM) for agarose gels. Electrode filter strips. Sample application strips; 5 x 10 mm pieces of Whatman No. 1 filter paper. Rubber roller. Hemoglobin marker: Add 100 PL of mouse whole blood to 1 mL of O.lM citrate in PBS, pH 7.2. Centrifuge at 2008 for 10 min. Wash the erythrocytes twice, in 5 mL of PBS, and resuspend the cells in 1 mL of distilled water. Store at -2OOC. Electrode solutions: 0.5M acetic acid (anode) and 0.5M NaOH (cathode). 0.9% NaCl solution. 99.5% ethanol. Coomassie Brilliant Blue staining and destaining solutions (seeChapter 50). Rabbit antimouse immunoglobulin antiserum.

3. Methods
3.1. Gel Preparation
1. Add 0.18 g of agarose to 16.6 mL of double-distilled

2. 3.

4.

5.

water. Heat to boiling while stirring. When the agarose is dissolved, transfer to a 75OC water bath. Add 1.4 mL of ampholytes pH 3.5-9.5 when the agarose solution is at 75C. Add the ampholytes immediately before casting the gel (steps 6-10). Prior to preparing the mold, wash and dry the glass plates thoroughly. Pour a few mL of water onto the larger glass plate. Place the plastic film with the hydrophilic side facing upwards on the glass plate, and squeeze out excess water and air bubbles with the rubber roller. Make sure that the surface is dry and free of dust. Place the spacers along both sides, and put the shorter glass plate on top of the spacers, leaving 1 cm free on both sides. Fix the plates together by placing clamps along the long sides. Place the mold in a heating cabinet at 75C for 20 min before casting the gel.

Isoelectric Focusing of Antibodies

629

6. To cast the gel take the mold from the heating cabinet and place it at an angle of 30C. 7. Fill a preheated pipet with the agarose solution. Put the tip of the pipet to the slit between the glass plates, and let the gel solution flow into the mold. Move the pipet along the slit to ensure an even distribution of the agarose solution and so that no air bubbles are formed. 8. Lower the mold to a horizontal position, and leave i t for 30 min at 22OC. 9. Remove excess agarose with a spatula. Grasp protruding plastic film and tear the glass plates gently apart. 10. Place the gel in a moist chamber.

3.2. Isoelectric

Focusing

and Immunoffxation

1. Connect the isoelectric focusing apparatus to the circulating water bath (+lOC). 2. Pour a few mL of water on the cooling plate and place the gel on top. Avoid trapping air bubbles. 3. Soak the electrode strips with electrode solutions. Apply the electrode strips to the gel. Cut off parts of the strips that protrude beyond the end of the gel. 4. Apply sample application strips 2 cm from the cathode. Sample volumes of < 5 PL can be applied directly on the gel surface. 5. Apply samples (2-20 PL vol). Hemoglobin can be used as marker. Add 2 PL of hemoglobin in two separate samples, one at the anode side and one at the cathode side. 6. Run the gel at 15 W for 15 min, then turn off the power, and remove the application pieces with forceps. Continue the isoelectric focusing for approximately 30 min at the same wattage. 7. Turn off the power supply, and remove the electrode strips. Measure the pH gradient at l-cm intervals across the gel with the surface electrode. 8. Place the gel in a box, and soak it with rabbit antimouse immunoglobulin antiserum. Incubate for 30 min at 22OC. 9. Wash the gel overnight in 0.9% NaCl solution. 10. Wash the gel in 95% ethanol for 10 min at 22OC, and then dry the gel in front of a heating fan. 11. Stain by soaking the gel in Coomassie Blue solution for 5 min and destain in destaining solution for 5-15 min (seeChapter 50, steps S-10).

630

Gustafsson

4. Notes
1. Gel casting kits are commercially available. 2. The quality of the agarose and distilled water is essential for successful isoelectric focusing. 3. It is not important to measure the pH gradient unless an exact p1 of the antibody is of interest, since it is the banding pattern of the antibody that is important.

Reference
1. Gustafsson, B. and Holme, T. (1983) Monoclonal antibodies against group- and typespecific lipopolysaccharide antigens of Vibrio cholerue 0:l. J. Clin. Microbiol. l&480485.

Chapter 52
Large-Scale Production and Storage of Monoclonal Antibodies and Hybridomas
Stuart A. Clark, J. Bryan Grifiths, and Christopher B. Morris
1. Introduction
The end product of all the selection and cloning procedures described in Chapters 4548 will be a monoclonal hybridoma culture growing in a 0.2-mL culture well. From this single well, sufficient cells must be grown for storage in liquid nitrogen. It is advisable that at least six ampules, divided between two liquid nitrogen freezers, be stored. After the security of the cell line has been assured, then the clone can be further expanded for bulk antibody production. Large quantities of antibody may be produced either in ascitic fluid or in vitro in various culture vessels. The following sections describe the various areas covered in the chapter.

1.1. Freeze Preservation

The maintenance of a cell line in continuous culture is not only impractical in terms of resources, but generally undesirable because of the risk of microbial infection and of genetic drift. The ideal situation is to have a cell stock that has been fully characterized and frozen at the lowest passage possible. In view of the strict regulations concerning the manu631

632

Clark, Griffiths,

and Morris

facture of diagnostic and therapeutic products from animal cell lines, e.g., monoclonal antibodies and enzymes, it becomes increasingly important to ensure that cell lines used in production processes have been fully screened for microbial contaminants.

1.2.

Clonal

Expansion

The expansion of the clone from a small well to larger culture vessels can be difficult. The rate of cell growth, which can vary significantly between clones, and the cell density are critical factors. Fortunately, there are a range of plastic vessels that permit a gentle stepwise scale-up of the cultures. The normal route is: 0.2-mL well > l-mL well > 25-cm2 flask (up to 15 mL) > 752 flask (up to 50 mL). At all expansion steps, it is important that the cells have reached a density of between 5 x lo5 and 1 x lo6 cells/ml before transfer to the next vessel. During the initial stages of this progression, especially at the lower end of this cell density, murine thymocyte feeder cells may be required to ensure that cells survive the transfer. Although the main aim is to expand the culture, it is very important to ensure that the cell density does not increase above 1 x lo6 cell/mL. Above this density, the cells are under stress caused by the limited availability of nutrients in the medium and may as a result alter in their secretory or replicative capacity.

1.3. Hybridoma

Production

in Ascites

The hybridomas are derived from plasmacytoma and spleen cells of an inbred strain of mouse, and are therefore immunologically compatible withother mice of the same strain. Moreover, the parental plasmacytomas were originally induced in the peritoneal cavity of mice treated with Staphylococcal adjuvants (I). By injecting the hybridoma cells into the peritoneal cavity of mice pretreated with Pristane or incomplete Freunds adjuvant (FIA), rapid cell proliferation occurs generating an ascitic fluid containing large quantities of cells and high concentrations of antibody (l-3 mg/mL).

1.4. Production

of Monoclonal

Antibodies

In Vitro

Production in ascitic fluid, although convenient for producing small quantities of high concentration monoclonal antibody (50-500 mg), does have the disadvantage of being contaminated with other immunoglobulins (and also possibly infectious agents), lacks reproducibility, may be unsuitable for therapeutic use, and is unsuitable for human MCAbs. The method is inconvenient for extensive scale-up because it is costly in manpower and facilities with no scale-up economics, in addition to the ethical

Large-Scale Hybridoma

Production

633

considerations of using animals. Thus, in vitro culture is the scale-up route for quantities of over 1 g, and certainly for the commercial operations now producing l-100 kg quantities. This scale of production is becoming necessary now that the use of antibodies has extended from diagnostic to affinity Durification and therapeutic use.

2. Materials
1. Freezing medium, e.g., growth medium with 25% serum and 7-10% cryoprotectant; alternatively, whole serum, i.e., newborn calf with 7-10% cryoprotectant (seeNote 1). 2. One Balb/c or Fl (Balb/c x CBA) mouse 4-8 wk old. 3. One sterile stainless-steel mesh (Expanded Metal Co., Hartlepool), Type 940,200-mm wide or a tea strainer. 4. Pristane (2,6,10,14-Tetramethypentadecane). 5. 20 mM EDTA containing 1% (w/v) sodium azide solution. 6. Mature Balb/c or Fl (Balb/c x CBA) mice. 7. Diethyl ether.

3. Methods
3.1. Freeze Preservation (See Notes l-5)
1. Harvest the cells in the same manner as for routine subculture (see Note 5). 2. After counting the viable cells, pellet the required number in centrifuge tubes. Trypsinized cells must be washed in medium containing serum to neutralize the enzyme. The number of ampules required is determined by allowing a final concentration of 4-10 x lo6 cells/ ampule. 3. Prepare the freeze medium while the cells are being centrifuged, allowing I mL/ampule. 4. Clearly mark your ampules (artline for plastic and ceramic ink for glass) with an identification code and preferably some designation that can be identified with that particular freeze run. Always maintain full records of freeze runs, with dates and locations. 5. The medium is decanted from cell pellets and freeze medium added. The pellets can be ampuled individually or pooled. If the number of ampules is small, i.e., c 20, dispensing can be by plastic pipets. 6. Transfer the ampules to the programmable freezing apparatus or twostage freezer (Linde BF-S), and cool at a steady rate of between 1 and 3C/min. With a programmable freezer, it is possible to reduce the

634

Clark, Griffiths,

and Morris

7.

8. 9.

10. 11.

problem of a temperature rise during phase transition, i.e., liquid-solid state, by temporarily increasing the cooling rate to 15 -2OC/min, and then return to the original slow cooling rate. On completion of freezing, the ampules are stored either in the gas or liquid phase of nitrogen depending on whether the ampules have been hermetically sealed. For temporary handling use a polystyrene container with liquid nitrogen to keep the ampules cold, e.g., while transferring from the freezer to the nitrogen refrigerator. Wear protective clothing at all times. One ampule should be examined for viability, i.e., Trypan blue exclusion test, and growth capacity. Thaw by placing immediately in a 37OC water bath, preferably in a rack to prevent total submersion. Resuspend the cells by slowly adding 1 mL of growth medium (dropwise), and then transfer to a flask or tube with 5-10 mL of medium. Take an aliquot, i.e., l-2% of the total volume, and measure cell viability. If it is necessary to remove the cryoprotectant, pellet the cells at low speed to prevent damaging them, i.e., 50-70g. Start a new culture by resuspending the viable cells at 40-50% of their usual saturation density. If the viability is below 70%, new stocks should be prepared. Attached cells can be medium changed after 24-48 h to remove cell debris.

3.2. Clonal Expansion

1. A mouse (4-8 wk old) is sacrificed by cervical dislocation, pinned ventral surface uppermost to a dissection board, and swabbed in 70% alcohol. 2. The skin is dissected away from the abdomen and thorax, the skin flaps pinned back, and the underlying tissue swabbed with 70% alcohol. 3. The peritoneum is opened and scissors used to cut through the ribs, which are removed to expose the heart and lungs. Care must be taken not to puncture the heart. The thymus is a cream colored bilobed organ, found lying above the heart (seeFig. 1). 4. The thymus is removed and immediately placed in 10 mL of medium containing 20% FCS (seeNote 6). The thymus, with the medium, is dissociated by pushing the tissue through a sterile tea strainer or steel mesh placed in a Petri dish, using the barrel of a 5-mL plastic syringe. 5. The cells are diluted to 85 mL in medium with 20% FCS. of eight g&well plates and in6. 100~pL aliquots are dispensed/well cubated at 37C in 5% CO,. (The outer wells can be filled with distilled

Large-Scale Hybridoma

Production

635

Cerwcal

kmph

node

lngvlnrl

lymph

node

Fig. 1. The main lymphoid organs of the mouse.

water containing 2x concentration of antibiotics to reduce dehydration in nonhumidified CO, atmospheres.) 7. Thymocyte feeder cells should be used to expand the positive hybridoma clones within 24 h of production.

3.3. Hybridoma Production (See Notes 7-14)

in Ascites

1. The mouse is injected ip with 0.5 mL of Pristane using a 23 g x l-in needle. 2. These pretreated mice can be used between 4 and 60 d after injection. 3. For each mouse, approximately 20 mL of cell suspension is harvested and centrifuged at 15Ogfor 10 min. The cells (approximately 1 x 10 cells) are resuspended in 0.5 mL of PBS and injected ip into the mouse using a 21 g x l-in needle. 4. Within 7-10 d after injection, the peritoneum of the mouse will become distended and the ascitic fluid can be drained. 5. The mouse is anesthesized with diethyl ether vapor, and a 19 g x l-in needle attached to a 5-mL syringe is inserted into the peritoneal cavity through the left flank of the mouse. The mouse is gently tilted so

636

Clark, Griffiths,

and Morris

that the liquid drains to the site of the needle and the excess fluid is gently removed. 6. Add 10% (v/v) of EDTA/sodium azide solution to the ascitic fluid to prevent clot formation. 7. The ascitic fluid is centrifuged for 10 min at 2008 and the supernatant

recovered (seeChapter 53 for the purification of the monoclonal antibody from this supernatant) and Notes 15 and 16.

3.4. Production

of MonocZonaZ

Antibodies

In Vitro

3.4.1. Method 1. Small-Scale Culture System Where less than 100 mg of monoclonal antibody is required, a few liters of end-culture supernatant may suffice. End-culture supernatant is that obtained after the cells are grown to the maximum cell density, and includes antibody secreted from viable and lysed cells. Antibody concentrations in such supernatants tend to reach between lo-20 &mL (see Note 17). 1. A lOOO-mL Spinner culture flask (TechneLtd., Note 18) is seeded with an initial culture volume of 200 mL containing 1 x 105cells/ml, stirred at 100 rpm and incubated at 37OC for 5 d. 3.4.2. Method 2. Bioreactors for Scale-Up (Table 1 and Fig. 2) The basic cell culture techniques described in this chapter (and Chapter 6, this volume) apply to the production of MCAb in all the culture systems reviewed. In addition, manufacturers instructions are now very comprehensive. The aim of this section is mainly to make one aware of the range of culture systems and concepts that are available, and to introduce this range in order that a rational choice can be made (2) (seeNotes 20-22). The following reactor types are available: a. Techne BR06 Bioreactor: This retains the spinner culture simplicity of use, but with the option of controlling pH and oxygen

and of using it as a continuous-flow culture. Oxygen supply is

good because of a very low aspect-ratio (i.e., high surface area/

unit vol) and the use of a surface aerator. The surface stirrer impeller is suitable for most fragile cells and is convenient for
building up the culture volume within the vessel as the cells grow (from 500 mL-3 L). It is currently only available as a 3-L system. b. Stirred Fermentation: There are a wide range of commercial reactors available in the range l-30 L and upwards (Applicon,

Table 1 Scale-Up Culture Systems for the Production

Culture system Bioreactor (Techne) Fermentors, stirred Airlift fermentor Opticell Ultrafiltration fibers Encapsulation Membroferm# B B B C C B/SC C Scale (L) 1-3
l-1000

of Monoclonal

Antibodies Process Vol (L) for Ig MCAb

lo-100 lo-100 10-100

Process intensity cells/ml x lo6 2-3 2-3 2-3 50

>lOO >lOO

Contamination Y Y Y Y N N/Y N

Ifs

5-1000
1

0.5-5
0.1-l 0.1-l 1

2.5
1-4-O

l-6

50-100

a See also Fig. 3. b B = batch, C = continuous, SC = semi-continuous processes, Y = yes, N = no. c Concentration in cell compartment, which is M-30% of total medium volume.

638

Clark, Griffiths,

and Morris

Fig. 2. High process intensity culture systems for producing monoclonal antibodies (further data inTable 1). (A) Opticell ceramic cartridge with coarse surface configuration for trapping cells against the medium flow, (B) hollow fiber cartridge with recirculating medium via a reservoir (R). Harvest product through (H), (Cl membroferm unit showing three-compartment configuration for medium, cells, and product, (D) encapsulation method in a stirred fcrmentor with the option of containing the product (a) or allowing the product to diffuse out of the capsule (b).

Large-Scale Hybridoma

Production

639

LH Fermentation, Life Science Laboratories, Contact Flow, New Brunswick, SGI, Chemap). Models specially adapted for animal cells have a modified marine, or large flat blade, impeller, curved (or domed) bottom configuration, slow stirring rates (10-200 rpm), and water jacket or heat pad temperature control. Above 10 L, special laboratory facilities are needed for in situ steam sterilization and so on. These units are not always suitable for fragile cells. c. Airlift Fermenter: Models are available from 500-mL (Fischer Scientific) through 2-, 5-, lo-, 30- to 90-L volumes (LH Fermentation, Chemap). They are very efficient for oxygenation, provide very gentle mixing, and are relatively easy and troublefree to operate. A good air flow (and mixing) control unit is needed. d. Opticell (CRBS): These systems are based on a ceramic cartridge that is a honeycomb of l-mm square longitudinal channels with an uneven surface. This provides a large surface area for the entrapment of cells as the media flows continuously through the unit (Fig. 2). The system comes in a highly sophisticated computer-controlled package, and the cost is such that it can only really be considered for commercial operation. e. Ultrafiltration Fibers: Complete turn-key units (Amicon, CellPharm CD Medical Inc.), specialized high performance units (Endotronics), or a wide range of biochemical and kidney dialysis cartridges are commercially available. High cell densities (> 108/mL) can be achieved, and maintained for long periods (months) allowing a daily harvesting of antibody. Advantages of all high density systems are the reduced serum requirement and a high antibody titer in the process fluid. These cartridges cannot be steam sterilized and need aseptic washing procedures before use. f. Encapsulation: The entrapment of cells within gel beads is a commercially used procedure for producing MCAbs (e.g., Damon Biotechnology, Karyon Technology), but is a skilled and complicated operation. There is a choice of suitable gels depending upon whether the product is required to diffuse out of the capsule containing the concentrated cell suspension (alginate gels), or retained with the cells (poly-L-Lysine). Other materials that have been used are agarose, collagen, and fibrin. The gel spheres can be suspended in growth medium in both stirred fermentors and fluidized beds. A fluidized bed biore-

Clark, Griffiths,

and Morris

actor is available with in situ gel entrapment (Bellco Glass Inc.). Procedures for entrapping and culturing cells in gels are published (e.g., 3). g. Membroferm Fermenter (MBR Bioreactor Ltd.): This fermenter contains a stack of parallel semi-permeable membranes that form a series of closed compartments (each of 20-25 mL). Membranes with two molecular cut off values are available so there is a choice of keeping the product compartmentalized separately from the cells (Fig. 2), (3 chamber reactor), or only having the cells compartmentalized (2 chamber reactor). Fresh medium continuously diffuses across the membranes form a recirculating reservoir into the cell compartment, thus maintaining high cell densities (over 50 million/ml). Units are available with up to 300 compartments.
3.4.2.1.

Procedures

1. Seed cultures at between 1-2 x 105 cells/mL with the expectation of achieving a yield of 1-3 x lo6 cells/mL after 5 d (mean generation time will be within the range 11-36 h, with a mean of 17-20 h). 2. pH range tolerated is 6.5-7.2. Initiate the culture at pH 7.0-7.2. A hepes buffered medium will be beneficial, especially in the absence of pH control. 3. The specific rate of antibody synthesis increases during the death phase (Fig. 3) for most hybridomas, thus keep the culture during the death phase of the cells. 4. Expected antibody yields will be within the range of 5-100 pg/mL, with a mean of about lo-20 pg/mL. With process control of pH and oxygen, the yield should increase to 2040 pg/mL. 5. Scale-up increases the need for more sophisticated downstream processing, especially for initial clarification. Filtration (especially tangential flow ultra-filtration) of the supernatant should be used. With large volumes, flocculation of the cells (e.g., with polygalacturonic acid or ammonium sulfate) increases filtration efficiency.

4. Notes
4.1. Freeze Preservation
1. Several cryoprotectants are available, the most commonly used are dimethyl sufoxide (DMSO), glycerol, and polyvinyl pyrrolidone (PVP). Only PTFE membranes can be used to filter sterilize DMSO; it must never be autoclaved because it will explode! Use only the highest

Large-Scale Hybridoma

Production

641

Time Fig. 3. Typical hybridoma

---+ antibody production kinetics.

cell growth and monoclonal

2. 3. 4. 5.

grades available, e.g., ultra pure and ensure that the newest stocks available are supplied. DMSO can be stored at -2OOCafter aliquotting. The condition of the cells at the time of freezing is essential to their survival. Therfore, only cells that are healthy and in log-phase growth should be frozen. For suspension cultures, the cell density should be 5-8 x l@/mL and viability > 90%. Attached cell lines should be no more than 90% confluent. Cells should be in medium free of antibiotics, which might otherwise mask low levels of infection. Remember that it is advisable to perform all culture work in class 2 cabinets. This reduces the risk of cross-contamination between cell lines, and also meets the requirements necessary to protect the user from extraneous pathogens. For further reading, the following texts are recommended (43).

4.2. CZonuZ Expansion

6. A thymus from a recently weaned mouse provides approximately lo* thymocytes, which should be used at 5 x 105/mL. Thymocyte feeder cells have the advantage that they die within 7-10 d in culture.

Clark, Griffiths,

and Morris

4.3. Production

in As&tic

Fluid

7. Mice treated in this way rapidly deteriorate in health. The ascitic fluid must be drained regularly (every l-3 d). Frequently, solid tumors are produced in addition to, or instead of, ascites. Such solid tumor formation tends to be a property of the clone. Mice should be sacrificed as soon as signs of distress are observed. The volume of ascitic fluid obtainable from each mouse will depend largely on the biological properties of the clone, but between 15-30 mL can be expected. 8. Some hybridomas may not produce a great deal of ascitic fluid in the first instance and may tend toward solid tumor production. Any liquid harvested from such mice when transferred directly into secondary pris tane primed mice should enhance the level of liquid accumulating in the peritoneal cavity. 9. When a liquid tumor has developed, cells may be recovered and reestablished in in vitro culture or frozen in aliquots. The cells frozen directly from the mice can be reintroduced into pristane primed mice directly from the frozen state. 10. It is also possible to clean up clones when they have reached over lo6 total cell number, should they become infected, by passaging them through pris tane primed mice. 11. Alternatively, Freunds incomplete adjuvant (FIA) can also be used to pretreat the mice for ascitic fluid production. In this case, the cells are injected 24.48 h following the injection of FIA. Although this method successfully produces ascitic fluid with fewer cells and in less time than the pristane method, in our experience some monoclonal antibodies lose significant activity in the process. 12, It is essential to monitor the antibody concentration throughout the process of purification. Monitoring antibody activity at regular intervals, using a suitable screening procedure, is necessary when the purification techniques may denature the monoclonal protein or when the preparation is contaminated with high concentrations of other proteins, especially immunoglobulins. Such is the case in supernatants concentrated from bulk cultures, unless the cells have been adapted from growth in medium supplemented with little, or no, FCS. However, the screening procedures are relatively time-consuming and give no indication of the degree of contamination with other proteins. 13. At high concentrations (above 1 mg/mL), the monoclonal antibody can be seen as a discrete band, distinct from other contaminating proteins, after gel electrophoresis. Several electrophoretic systems are commercially available based on thin agarose gels supported with

Large-Scale Hybridoma

Production

643

plastic films. Results are obtainable within 90 min with such systems and allow rapid monitoring of downstream processes. 14. Figure 4 illustrates the results of the purification of several monoclonal antibodies from ascitic fluids, run on a Paragon Electrophoresis System (Beckman Ltd). 15. It is important to remember that storage conditions under which the activity of polyclonal antibodies will survive can easily destroy the activity of monoclonal antibodies. This is because polyclonal antibodies have a range of physical properties; therefore, only a proportion of the antibodies will be denatured by mistreatment, whereas all of the molecules of a monoclonal antibody have the same properties. 16. Repeated freezing and thawing causes immunoglobulin aggregation with concommitant loss of antibody activity. It is, therefore, essential to freeze supernatants, ascites fluid, or purified antibody in small aliquots. Such aliquots are best stored at -5OOCor below. This inhibits enzymic proteolytic activity and helps to stabilize IgM isotypes. Alternatively, the antibody can be diluted 1:1 in glycerol and stored at -20C; at this temperature, the mixture remains liquid. Despite these precautions, some monoclonal antibodies will still not be recoverable after freezing. Such antibodies should be diluted in O.lM Tris-HCl buffer, pH 7.2, and stored at 4C; Tris has considerable antibacterial activity (6).

4.4. Production

of MCAbs In Vitro

17. By adapting the hybridomas to grow in a serum substitute, such as CPSR-1 (Sigma Chemicals, Ltd.), cells can be maintained in a serum protein concentration of 1% or less. End-culture supernatants from these cultures may then be concentrated (up to 30-fold) using ultrafiltration prior to further purification. 18. The Techne MCS spinner vessels are very useful and efficient first-step culture units, which provide a gentle but efficient agitation and are a preferred alternative to the use of stationary flasks or roller bottles. However, the next step in scale-up must allow the introduction of environmental control equipment (02, pH) into the process. In choosing a suitable system, consideration has to be given to the relative fragilityof the cell (i.e., canit withstand conventionalmarine-typeimpellers or does it need gentle agitation methods), oxygen, and nutrient requirements. 19. The scale-up of animal cells has been reviewed in Chapter 6, and the contents that refer to suspension culture are applicable to hybridoma cells and should be read in conjuction with this chapter.

644

Clark, Gri@ths, and Morris

Fig. 4. Monitoring antibody purification. Track 1. Normal mouse serum (l/5 dil.) Track 2. IgM monoclonal as&tic fluid. Track 3.lgM monoclonal purified using AcA 22 Column (0.75 mg/mL). Track 4. IgG,, monoclonal ascitic fluid. Track 5. I& monoclonal purified using protein-A-Sepharose (1 mg/mL). Track 6. IgG, monoclonal ascitic fluid. Track 7. IgG, monoclonal purified using DEAE Trisacryl column (0.8 mg/ mL). Track 8. IgGzr,monoclonal ascitic fluid. Track 9. IgG, monoclonal ascitic fluid. Track 10. Normal mouse serum (a 115 dil.). The isotypes were determined by ELISA (seeSection 6.1.2.). The position of monoclonal bands cannot be used to determine the isotype, seetracks 6 and 9.

20. If only milligram quantities of MCAb are required, then the mouse ascitic fluid method or simple replicate culture vessels (flasks, roller bottles) are adequate. 21. Once the requirement exceeds 100 mg, then scale-up methods should be considered. Otherwise, one has to use over 50 mice or 30 roller bottles. 22. Until the requirement exceeds 0.5 g, the simple versions of fermenters (stirred or airlift) can be conveniently used, but once 0.5-l g has to be produced (equivalent to 50-100 L of low process intensity culture), then the more sophisticated and expensive culture systems listed in Table 1 will be needed.

Acknowledgments
The authors wish to thank Miss A. Williams for the artwork on Fig. 1 and Mrs. Heather Hogton for typing the manuscript.

Large-Scale Hybridoma

Production

645

References
1. Potter, M. and Robertson, C. L. (1960))x~naZ of the National Cancer Institute 25,847. 2. Griffiths, J. B. (1988) Overview of cell culture systems and their scale-up, in Animal Cell:Biotechnology, vol. 3 (Spier, R. E. and Griffiths, J.B.,eds.), Academic, London, pp. 179-220. 3. Nilsson, K. (1987) Methods for immobilizing animal cells. Trends in Biofechnol. 5,
73-78.

4. Ashwood-Smith, M. J. and Farrant, J. (eds.) (1980) low Temperature Presentation in Medicine and SioZogy (Pitman Medical). 5. Doyle, A., Morris, C. B., and Armitage, W. J. (1988) Cryupreservation of Animal Cells in Advunces in Biotechnological Processes, vol. 7 (A. R. Liss, New York), pp. l-17. 6. Newell, D., McBride, B. W., and Clark, S. A. (1988) Making Monoclonals (PHLS, London).

Chapter 53
Purification Monoclonal of Murine Antibodies

Michael G. Baines, Andrew J. H. Gearing, and Robin Thorpe

1. Introduction
Hybridoma technology has made possible the production of highly specific, homogeneous antibodies withpredefined binding characteristics, which can be produced in large amounts, from immortal cell lines. They probably represent the immunochemists ideal as reagents, and monoclonal antibodies (MoAbs) are invaluable for the investigation and assay of the entire antigen spectrum. However, it must be remembered that, although MoAbs can be exquisitely specific, they are far from pure unless specific procedures are used for their isolation. If produced as ascitic fluid, they contain proteins derived from the host animal, whereas in vitro production as cell culture supernatant produces MoAbs contaminated with tissue culture additives, nonimmunoglobulin secretion products, and material derived from dead disrupted cells. Preliminary purification procedures (1,2) generally utilize a precipitation technique allowing the easy production of a moderately pure, stable and concentrated MoAb preparation. Such procedures are the ideal startingpoint for further purification, for storage, and since the precipitates are particularly stable at room temperature, they are suitable for distribution or exchange between laboratories. Although precipitation techniques are 647

648

Baines, Gearing, and Thorpe

excellent for producing purified, concentrated MoAb preparations they rarely produce the highly pure antibodies necessary for many immunochemical procedures, e.g., radiolabeling or enzyme conjugation for use in immunoassays. For this, it is usually necessary to carry out some type of chromatographic technique. Both ion-exchange and gel filtration have been used for this, and there are many adaptations of these for use in conventional chromatography and HPLC systems (3,4,5,6). Affinity chromatography techniques (7) are also extremely valuable for the purification of MoAbs. At present, mouse and rat MoAbs are by far the easiest to prepare because of the availability of inbred strains, good fusion partners, and ease of immunization. However, mouse and rat immunoglobulins are generally less stable than those of higher mammals, and can be more difficult to purify and successfully fragment.

2. Materials
1. PBS: 0.14M NaCl, 2.7 mM KCl, 1.5 mM KH,PO, 8.1 mM Na,HPO,. Store at 4C. 2. PBSA: PBS containing 0.015M NaN,. 3. Saturated Ammonium Sulfate Solution: Add excess (NH,),SO, to distilled water and stir overnight at 4OC. This solution (in contact with solid salt) is stored at 4C. 4. PEG Solution: 20% w/vPEG 6000inPBS;forIgMprecipitationuse6% PEG 6000 in PBS. 5. Silicon Dioxide. 6. Glycerol. 7. Phosphate Buffer: 2 mM Na,HPO,, 2 mM NaH,PO,*2%0, pH 6.0. Add the dihydrogen salt to the disodium salt until the required pH is obtained. Phosphate buffer is also used at O.lM, pH 8.0. Store at 4OC. 8. Sephacryl S-300. 9. DEAE - Sephacel. 10. HCl: lOaM, 0.5M, 1M. 11. NaOH: 0.5M, lM, 4M. 12. Tris-HCl Buffer: O.O5M, lM, 2M Tris, pH 8.0. Adjust pH with 1M HCI or concentrated HCl as appropriate. For ionic gradient separations use Tris-HCl buffer containing 0.3M NaCl. Store at 4OC. 13. Triethanolamine Buffer: 0.02M triethanolamine, pH 7.7. For ionic gradient separations by FPLC, use triethanolamine buffer containing 1M NaCl. 14. Sepharose 4B.

Murine Monoclonal

Antibodies

649

15. Sodium Carbonate Buffer: 0.5M NaHCO, pH 10.5. 16. Cyanogen Bromide. 17. Sodium Citrate Buffer: O.lM Na&H,O,*2HO, pH 2.5,3.5,6.5. Adjust
pH with citric acid. 18. Ethanolamine Buffer: 2M ethanolamine. 19. Glycine-HCl Buffer: O.lM glycine, pH 2.5. Adjust pH with 1M HCl. 20. Protein A Sepharose. 21. Mini-Gel Solution: 1M Tris, 1M Bicine (2.0 mL), 50% w/v acrylamide (2.5% w/v his-acrylamide) (4.0 mL), 1.5% w/v (NHJ2 S,O, (0.4 mL), 10% w/v SDS (0.2 mL). Make up to 20 mL with distilled water. 22. Mini-Gel Running Buffer: 1M Tris, IM Bicine (2.8 mL), 10% w/v SDS (1.4 mL). Make up to 140 mL with distilled water. Store at 4OC. 23. Mini-Gel Sample Buffer: Sucrose (1 g), 1M Tris, 1M Bicine (0.2 mL), 10% w/v SDS (1 mL), Zmercaptoethanol(O.25 mL). Made up to 3 mL with distilled water and add 0.001% w/v bromophenol blue. Store at -2OOC. 24. Coomassie Blue R Stain: Add Coomassie brilliant blue R (0.025 g) to methanol (50 mL) and stir for 10 min. Add distilled water (45 mL) and glacial acetic acid (5 mL). Use within 1 mo. 25. Destain Solution: Glacial acetic acid (7.5 mL), methanol (5 mL). Make up to 100 mL with distilled water. Store at room temperature. 26. TEMED. 27. High and Low Molecular Weight Markers. All reagents should be of Analar grade.

3. Methods
3.1. Preliminary

Purification Techniques

(Precipitation)

Generally, ammoniumsulfateprecipitationis themethodof choicefor most IgG and some IgM MoAbs; however, euglobulin precipitation is particularly successful with some, but not all IgM MoAbs. The use of PEG or less often sodium sulfate can be useful for some MoAbs, and techniques for these procedures are also described below. Other techniques such as the use of rivanolor ethanol can be of use in a few cases; however, since they are not generally applicable to most MoAbs, they will not be considered further here. 3.1.1. Ammonium Sulphate Precipitation 1. Prepare the ascitic fluid or culture supernatant for fractionation by centrifugation for 20 min at lOOOOg,,(seeNote 1).

650

Baines, Gearing, and Thorpe

2. Cool the antibody containing fluid on ice (or to 4C), and stir slowly using a magnetic stirrer. 3. Add saturated ammonium sulfate solution (equilibrated to 4C) dropwise to give a 35-45% final saturation, i.e., for 45% saturation, add 8.2

mL saturated ammonium sulfate for each 10 mL of antibody containing fluid. Alternatively, add solid ammonium sulfate to give desired saturation. (45% saturation = 2.7 g ammonium sulfate/l0 mL fluid.)

Stir in the cold for 2-4 h.

4. Centrifuge for 15-20 min at 20004000gaV at 4OC and discard supernatant. Drain pellet until dry. 5. Dissolve the precipitate in lo-20% of the original volume of PBS (or other buffer) by agitation with a spatula or drawing repeatedly into a wide gage Pasteur pipet. Make up to 25-50% of original volume and dialyze against 100 vol of PBS (or other buffer) with repeated buffer changes overnight at 4OC. Alternatively, the precipitate may be stored at 4 or -20C if not immediately required (seeNotes 2-4). 1. 2. 3. 3.1.2. Precipitation with PEG (See Note 5) Prepare a 20% w/v solution of PEG 6000 in PBS and cool it to 4OC. Prepare the ascitic fluid or culture supernatant for fractionation. Cool the antibody solution on ice, and add an equal vol of 20% PEG solution with slow stirring. Stir on ice for 15-30 min. This works well for most IgG MoAbs; if the preparation is very impure, reduce the amount of PEG added. For IgM MoAbs, less PEG is needed and a final concentration of 4-6% PEG works in most cases. Centrifuge for 30 min at 2OOOg,,at 4C and discard the supernatant. Drain the pellet until dry and resuspend. Dialyze against PBS (or other buffer) as described for the ammonium sulfate fractionation.

4.

3.1.3. Euglobulin Precipitation of IgM (See Note 6) 1. Prepare the ascitic fluid or culture supernatant for fractionation. 2. Dialyze MoAb solution against lo-50 vol of 2 mA4 sodium phosphate pH 6.0 at 4OC, changing the buffer frequently until a precipitate develops. 3. Centrifuge for 10 min at 2000gaVat 4C, discard the supernatant, and drain the pellet. Resuspend carefully in PBS or other suitable buffer.

3.2.

Chromatography Techniques or Charge Separation

Based

on Size

3.2.1. Purification of IgM MoAbs by Gel Filtration Gel filtration is not particularly effective for the purification of IgG MoAbs, since they tend to elutein a rather broad peak that is normally quite

Murine Monoclonal Antibodies

651

heavily contaminated with albumin (derived from dimeric albumin, M, 135,000). For this reason, gel filtration for isolation of IgM, which can be a very useful purification procedure for IgM MoAbs, will be the only gel filtration procedure described in this section (seeNote 7). 1. Carefully pack a column of Sephacryl S-300 (seemanufacturers instructions and 2). Pump through with column buffer (PBSA) at 20-30 mL/h. 2. Carry out a preliminary purification of approximately 10 mL ascitic fluid by euglobulin precipitation, PEG precipitation, or ammonium sulfate fractionation as described above. 3. Pump partially purified MoAb preparation onto the column at 20-30 mL/h and collect approximately 6 mL fractions. Monitor absorbance at 280 nM. The void peak will contain the IgM (seeFig. la). 4. Check purity of fractions and pool fractions accordingly. 3.2.2. Ion-Exchange Chromatography (See Note 8) Ion-exchange chromatography has been widely used as a method for the purification of immunoglobulins from both polyclonal antisera and preparations containing high concentrations of MoAb. Anion exchange groups, e.g., diethylaminoethyl (DEAE) and cation exchange groups (5), e.g., carboxymethyl (CM), have both been used coupled to a cellulose support for the fractionation of serum proteins, although DEAE cellulose is the most commonly used matrix for this purpose. DEAE has also been covalently linked to a beaded crosslinked agarose support together with the dye cibacron blue F3GA (DEAE Affi-gel blue). In this case,the blue dye displays a differential affinity for a number of serum proteins and has been used as a method for the purification of immunoglobulins (4). 1. 3.2.3. Preparation and Equilibration of Ion-Exchanger Gently stir the ion-exchanger into approximately five times its swollen volume of 0.5M HCl (anion exchanger) or 0.5M NaOH (cation exchanger). Leave for 30 min at room temperature with occasional swirling. Filter out the resin by suction through a Buchner funnel with a WhatmanNo. 54 filter paper. Wash the resin cake with distilled water until the pH of the filtrate is greater than 4 (after acid) or less than 8 (after base). Gently stir the ion-exchanger into 5x its swollen volume of 0.5M NaOH (anion exchanger) or 0.5M HCl (cation exchanger). Leave for 30 min at room temperature with occasional swirling. Repeat step 2. Add ion-exchanger to an equal volume of 10x concentrated starting

2.

3. 4. 5.

652
20

Baines, Gearing, and Thorpe

41

34

40

46

52

58

64

70

FRACTION NUMBER
Fig. la. Purification of IgM MoAb by gel filtration using Sephacryl S-300. Ascitic fluid containing an IgM MoAb was partially purified by ammonium sulfate precipitation. The crude IgM obtained from 9 mL of ascitic fluid was applied to a Sephacryl S300 column (85 x 2.8 cm). The IgM elutes near the void volume (fractions 3545).

buffer, e.g., Tris-HCl, pH 8.0,0.05M (or other buffer). Mix thoroughly and leave at room temperature for 30 min. 6. The ion-exchanger will absorb some buffer ions in exchange for protons or hydroxyl ions and hence alter the PH. The pH must be restored to the original value by gently stirring the slurry, and adding the acid or basic forms of the buffer (HCI for an anion exchanger and IM Tris for a cation exchanger). 7. Leave the slurry at room temperature for another 30 min and recheck its PH. Adjust it if necessary. The ion-exchanger is now at the required pH with the counterion bound but the ionic strength will be too high. 8. Wash the resin with 5x its volume of starting buffer through a Whatman No. 54 filter paper.

Murine Monoclonal

Antibodies

653

9. Degas the slurry using a Buchner side-arm flask under vacuum for 1 h with periodic swirling. 10. Suspend the resin in about 5x its volume of starting buffer and leave to stand until the majority of beads have settled. Remove the fines by aspirating the supernatant down to about twice the settled slurry volume. 11. Carefully pack a clean column by first filling it with 10 mL starting buffer and then by pouring the resuspended slurry down a glass rod onto the inside wall until the column is filled. Allow to settle under gravity. 12. Wash the column with starting buffer at the operating temperature until the pH and conductivity of the eluant is exactly the same as the starting buffer. 1. 2. 3.2.4. Sample Application and Elution Apply the precipitate ascitic fluid sample to the column (seeNote 9), and wash the ion-exchanger with two column volumes of starting buffer. Collect fractions and monitor eluant for absorbance at 280 nM. Using a commercial or homemade gradient maker, elute the proteins with a gradient of increasing ionic strength. For example, 200 mL of Tris-HCl (pH 8.0; 0.05M) in chamber A of the gradient maker and 200 mL of Tris-HCl (pH 8.0; O.OSM>containing 0.3M NaCl in chamber B of the gradient maker. Collect fractions of about 5-mL volume and monitor for absorbance at 280 nM. The first major peak of the elution profile will contain IgG antibody if DEAE-resin is used. If cationic exchangers are used, the position of the MoAb peak must be determined empirically. Pool the protein containing fractions determined by absorbance at 280 nM, and dialyze against PBS (or other buffer).

3.

4.

3.3. HPLC and ReZated Techniques

Fractionation of proteins by HJ?LC (3) utilizes the same principles as chromatography in standard columns by gel filtration, ion-exchange or affinity chromatography. Therefore, in HPLC, separation occurs in a column of small cross-sectional area containing the chromatographic material in the form of very fine particles (the stationary phase). The solvent (buffer) or mobile phase is pumped through the column using medium to high pressure pumps. This enables the sample molecules in the mobile phase to interact reversibly with the stationary phase in a continuous fashion.

654

Baines, Gearing, and Thorpe

The advantages of HPLC over conventional chromatographic techniques are speed because of the small high capacity columns, improved reproducibility because of the sophisticated pumps and accurate timers, and in many cases increased resolution because of the fine resins and control systems. The technology of the resins in the stationary phase has advanced considerably in recent years, and the entire fractionation process can be performed very quickly indeed. Fast Protein Liquid Chromatography (FPLC) is a variant of HPLC, which has proved useful in the purification of mouse MoAbs (6 and Note 10).
3.3.1. FPLC Purification of Mouse IgG MoAbs

1. Prepare ascitic fluid by ammonium sulfate precipitation. In the final step, use triethanolamine buffer (pH 7.7; 0.02M) to redissolve and dialyze exhaustively overnight at 4OCusing 100 vol of this buffer. Filter sample before use (0.2 cln/r). 2. Assemble FPLC system according to manufacturers (Pharmacia) instructions for use with a Mono-Q anion exchange column. 3. Equilibrate the column using the Mono-Q ion-exchange equilibration method, which is programmed into the liquid chromatography controller. This takes about 13 min. 4. Wash through the sample injection loop with buffer A (triethanolamine, pH 7.7; 0.02M). 5. Load the sample injection loop with the filtered sample. (Usually up to 500 PL can be loaded using the appropriate loop.) 6. Set the sensitivity on the UV monitor control unit to 0.2 and zero the chart recorder, (The setting will vary depending on concentration of sample; try 0.2 to start with and adjust as required.) 7. Run the Mono-Q ion-exchange MoAb purification method, which is programmed into the liquid chromatography controller. Collect fractions (automatic, seeFig. lb). This takes about 1 h. Elution buffers are buffer A (triethanolamine, pH 7.7; 0.02M) and buffer B (buffer A containing 1M NaCl). 8. Wash the Mono-Q ion-exchange column with NaOH and NaCl (methods programmed). 9. Store column in 0.02% NaN, (method programmed).

3.4. Affinity

Chromatography

Purification of MoAbs by affinity chromatography is a very powerful method that requires little specialized equipment. Essentially all procedures require a solid matrix to which ligands can be covalently coupled.

Murine Monoclonal Antibodies

655

-41

,/

20

30 TIME(mins3

40

50

Fig. lb. FPLC purified MoAb elution profile. A typical elution profile for an FPLC purified MoAb. This profile represents the purification of an antiprolactin MoAb by the authors. The first major peak is IgG, which was eluted after about 30 min.

Ligands are chosen that display specific and reversible binding to some portion of the antibody molecule (seeNotes 11-13). of Sepharose with Cyanogen Bromide 1. Wash 10 mL of Sepharose 4B by vacuum filtration on a sintered funnel with 1 L of water, and then resuspend it immediately in 18 mL of water in a 50 mL beaker. 2. Add 2 mL of 0.5M sodium carbonate buffer pH 10.5 and a magnetic flea. Place in a fume cupboard on a magnetic stirrer with a pH electrode in the solution. 3. CAUTION: Dispense 1.5 g cyanogen bromide (CNBr) into an airtight
container. Weigh in fume cupboard, and wear gloves. (CNBr is ex-

3.4.1. Activation

Baines, Gearing, and Thorpe tremely toxic by inhalation or by skin contact.) Contaminated equipment should be immersed in 1MNaOH overnight in the fume hood to neutralize the CNBr. Add the CNBr to the gently stirred Sepharose. Maintain the pH between 10.5-11.0 by dropwise addition of 4M NaOH until all the CNBr is dissolved and the pH stabilizes (lo-15 min). If the pH rises above 11.5, the Sepharose must be discarded. Wash themixtureon a sintered glass or Buchner funnel with 2L of cold O.IM sodium citrate pH 6.5. CAUTION: The filtrate contains CNBr, so discard it carefully. Do not allow it to dry out. Ifusing commercial CNBr Sepharose, swell 3 g of powder and wash in 600 mL lO3M HCI for 15 min on a sintered glass filter. Quickly add the cake of activated Sepharose to the protein solution (2-10 mg/mL) in O.lM citrate buffer pH 2.5. Mix it gently on a rotator overnight at 4OC. NOTE: This is a relatively inefficient pH to couple at, but results in much less deformation of the proteins and hence better affinity ligands. Block remaining active sites with 2M ethanolamine over 1 h. Calculate the percentage of protein unbound by measuring the absorbanceat280 nMof asampleof thesupernatant. Normally less than5% will be unbound. 3.4.2. Purification of MoAbs on Afinity Columns

4.

5.

6.

7. 8.

1. Preelute column with dissociating buffer, e.g., O.lM glycine-HCl pH 2.5, followed by PBS. 2. Slowly apply sample containing MoAb, and allow to equilibrate for 15-30 min. 3. Elute unbound proteins with PBS until no protein is detected by absorbance at 280 nM. 4. Disrupt ligand-MoAb complex with dissociating buffer and collect peak of protein (MoAb) into IM Tris (120 pL/l mL fraction). Dialyze against PBS to remove glycine-HCl. 5. Wash column in PBSA. Store at 4OC. 3.4.3. Protein A Column Purifications of IgG

1. Preelute a column containing 1 mL of protein A Sepharose with 5 mL of O.IM phosphate buffer pH 8.0,2 mL citric acid, and 5 mL of O.lM phosphate buffer pH 8.0. A 1 mL column should contain about 2 mg protein A, which can retain 20 mg IgG (2 mL ascites or 500 mL culture supernatant).

Murine Monoclonal Antibodies

657

2. If ascites fluid, adjust pH to 8.0 with 2M Tris base, add equal volume of O.lM phosphate buffer pH 8.0, and apply to the column at 5 mL/h. Collect 0.5 mL fractions and monitor the absorbance at 280 nM. 3. Wash with 5 mL of O.lM phosphate buffer pH 8.0 IgM. IgG, or IgA antibodies are eluted at this stage. 4. Elute with 5 mL O.lM citrate buffer pH 3.5 to obtain all other IgG subclasses. Fractions should be collected into 60 PL 1M Tris. The protein peak should be pooled and dialyzed against PBS to give the final antibody solution.

3.5. Assessment Polyacrylamide

of Purity

by Sodium Gel Electrophoresis

Dodecyl Sulfate (SDS-PAGE)

After the purification procedure, it is then necessary to obtain some index of sample purity. The simplest method for assessing the purity of a MoAb is by SDS-PAGE (seeNote 14). Although full size slab gels (see Figs. 2 and 3) can be used with discontinuous buffer systems and stacking gels, the use of a mini-gel procedure is much easier, quicker and is perfectly adequate for assessing purity, monitoring column fractions, and so on (seeFig. 4). 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13, 14. Prepare the sample buffer. Adjust the MoAb preparation to I mg/mL in O.lM Tris, O.lM Bicine. Mix the sample in the ratio 2:l with the sample buffer. Heat at 100C for 24 min. Prepare gel solution and running buffer as described in Materials. Assemble the gel plates and mini-gel apparatus. Add 50 PL TEMED to 10 mL of gel solution, and use this to set the trough in the mini-gel apparatus. Leave for 10 min. Add 30 PL TEMED to the remaining 10 mL of gel solution, pour the gel into the gel plates, and set the comb. Leave for 10 min. Remove comb and locate gel plates into the electrophoresis tank. Add running buffer. Load the sample prepared as described in 14 (30-50 l.tL per track) and run standard mol wt markers (10 PIJtrack of manufacturers recommended stock solution) in parallel. Run at 150 V for 1.5 h. Remove gel, and stain in excess Coomassie blue R stain for 2 h (gently rocking) or overnight (stationary). Pour off the stain and rinse briefly in tap water. Add excess destain and 2-3 pieces of sponge to the gel. Leave until destaining is complete (usually overnight gently rocking).

658

Baines, Gearing, and Thorpe

ABC

Fig. 2. Polyacrylamide gel electrophoresis of Protein A purified IgG, MoAb. The separating gel contained a gradient of 5-15% total acrylamide and was run under reducing conditions. Track A shows mol wt markers (from top to bottom: myosin heavy chain M, 200 kdaltons; b-galactosidase M, 116 kdaltons; phosphorylase b M, 97.4 kdaltons; bovine albumin Mr 66 kdaltons; egg albumin Mr 45 kdaltons and carbonic anhydrase M, 29 kdaltons). Tracks B and C show purified MoAb and tracks D and E show hybridoma culture supematant containing MoAb.

4. Notes
1. Crude preparations of MoAbs, whether ascitic fluid or cell culture supernatant, normally contain insoluble material that must be removed before MoAb purification is commenced. This should be removed by centrifugation (20-30 min at 1OOOOgJ. Ascitic fluid often contains a considerable amount of lipid that sometimes forms a pellicule after centrifugation and can be removed with a spatula. Alterna-

Murine Monoclonal

Antibodies

659

Fig. 3. Affinity purification of two IgM class MoAb by antigen columns. The figure is a Coomassie blue stained SDS polyacrylamide gel run under reducing conditions. Tracks A and C show proteins present in ascitic fluid containing two different IgM MoAbs directed against human IgM. Tracks Band D show the low pH eluate after applying the ascitic fluid to a polyclonal human IgM affinity column.

660

Baines, Gearing, and Thorpe

Fig. 4. SDS-PAGE of FPLC purified MoAbs. SDS mini-gel electrophoresis analysis of an FPLC purified mouse MoAb stained with Coomassie blue R. The left-hand track shows standard mol wt markers (see Fig. 2 for details). The remaining tracks show different loadings of the same MoAb. Characteristic heavy chains (50,000) and light chains (22,000) are clearly seen free of contamination by other proteins.

tively, and if lipid contamination is a par titular problem, the addition of silicon dioxide powder (15 mg/mL) followed by ten-trifugation (20 min; 2OOOgJ readily eliminates most of the fatty material present in such ascitic fluids (2). 2. Crude preparations of MoAbs (a&tic fluid or culture supernatant) should be stored at -20C and repeated freeze-thawing avoided. Purified IgG MoAbs should be aliquoted and stored at -20 or -7OOC. Pure IgM antibodies can be stored frozen, but some are better stored at 4 or -2OOC in 50% glycerol, i.e., not frozen. Some MoAbs can be freeze

Murine Monoclonal

Antibodies

661

dried with no apparent loss in antigen binding properties. If highly purified it is often necessary to add carrier protein, e.g., 0.5% bovine or human serum albumin. It should be remembered that all MoAbs are different, and so the most efficient storage conditions must be found empirically-some unpurified MoAb containing culture supernatants can be stored at 4OCfor 1-2 yr with no apparent adverse effects. If MoAbs are stored at 4C, then 0.02-0.1% sodium azide or 0.01% thimerosal should be added to prohibit bacterial growth. 3. Precipitation of antibodies with ammonium sulfate is one of the most useful purification procedures. The technique is particularly applicable to murine MoAbs since it is fast, can be carried out in the cold, and is efficient. It produces a fairly pure product, particularly if repeat precipitation with 40% saturation is used and leads to minimal denaturation of antibody in most cases. Dependent upon whether purity or yield are most important, precipitation is carried out using 35-45% saturation with ammonium sulfate. The former will produce a relatively pure immunoglobulin preparation but usually leads to some loss of MoAb, whereas the latter enables almost total recovery of most antibodies but this will be contaminated with other proteins, mainly transferrin and some albumin. Since the latter proteins are easily removed by further purification procedures, precipitation with 45% saturated ammonium sulfateis theideal starting point for further purification. The desired saturation with ammonium sulfate can be obtained either by addition of solid salt, or by addition of a saturated solution. For most MoAbs, the latter is preferable because it generally causes less denaturation. The saturated solution should be prepared by adding excess ammonium sulfate to distilled water and stirring overnight at 4C. The solution, which should be in contact with solid salt, must be stored at 4OC.In some cases,adjusting the pH to 7.0 using sodium hydroxide has been advised, but the authors have not found this to be advantageous. 4. Precipitation of human and rabbit serum with sodium sulfate produces a fairly pure IgG preparation in good yield with little denaturation. However, fractionation of mouse or rat ascitic fluid with this salt is generally less successful, and both yield and purity are almost always lower than for human or rabbit serum. Since the procedure is carried out at 25-37OC, some denaturation of the less stable mouse or rat IgG may also occur. The fractionation is performed as for ammonium sulfate precipitation except that it is carried out at all stages (including centrifugation) at 25C and solid sodium sulfate is added to give 15-18% saturation (18% sat = 1.8 g salt/l0 mL fluid). A pre-

Baines, Gearing, and Thorpe cipitation time of 30-60 min is normally used. It is emphasized that a pilot scale fractionation should be carried out to determine the effectiveness of this procedure for purification of MoAb, and it is essential to check purity, yield, and whether extensive denaturation has occurred. The procedure is not suitable for fractionation of cell culture supernatants in most cases. 5. Precipitation of antibody with PEG (2) is a procedure well known to those working with liquid-phase radioimmunoassay. It can be a very effective purification procedure for MoAbs and is particularly mild resulting in very little denaturation. Although it usually works well, a pilot scale experiment should always be carried out to check for unexpected problems. The procedure is suitable for both IgG and IgM MoAbs, although the percent of PEG required is different. 6. Although IgM MoAbs can be fractionated by the above techniques, the preparation obtained is usually impure (ascitic fluid fractionated in this way will contain much IgG), and is often denatured and/or difficult to redissolve. Many (but not all) IgM antibodies can be precipitated by dialysis against very low salt concentrations, which effectively removes most salts from the solution resulting in precipitation of the macromolecular IgM. This is the basis of euglobulin precipitation, and again pilot scale studies are necessary to establish the efficiency and efficacy of this procedure for particular MoAbs. Some IgM MoAbs are not precipitated by this process, and others are precipitated but will not redissolve or become denatured. If the procedure does not produce precipitation, try precipitation with ammonium sulfate (40% saturated) or 6% PEG 6000. 7. The macromolecular size of IgM (Mr about 9 x 105-106)causes MoAbs of this class to elute in the void volume of many molecular seive gels, e.g., Sephadex G-200, Bio gel P300, Sephacryl S-300. Since the purity of the IgM MoAbs will be compromised if the full fractionation range of most gels is not obtained, it is usually necessary to use well-constructed commercially available columns equipped with flow adaptors (Pharmacia, LKB, and so on). Although Sephadex G-200 is often used for the preparation of IgM, this gel is rather fragile and the authors prefer to use the more robust Sephacryl S-300. Alternatively, Bio-Gel P300, Ultragel AcA 22, or Sepharose 6B can be used. It is important not to overload the column or to use too fast a flow rate. Sepharose 4B or 6B and gels of the Ultragel A series have fractionation ranges such that IgM is within the exclusion volume and will elute after the void volume. This is usually not advantageous for purifica-

Murine Monoclonal

Antibodies

663

tion of IgM MoAbs, but can be of use in some cases. For a detailed discussion of gel filtration, see manufacturers literature and (I). It is usual to carry out preliminary purification using a precipitation technique; euglobulin, PEG, or ammonium sulfate precipitation as described are the most appropriate. The precipitate is redissolved in column buffer at l-3 mg/mL, and this is pumped directly onto the column. The buffer for fractionation is relatively unimportant, and the authors have used PBSA successfully for several different IgM MoAbs. Buffers should contain at least 100 mA4 sodium chloride or a similar salt to avoid binding of the proteins to the gel matrix. Fractions containing the IgM often appear opalescent, and it is best to avoid further manipulation of the purified MoAb, e.g., by concentration or further precipitation. A procedure using Sephacryl S-300 is described and the profile obtained in the authors laboratory is shown in Fig. la. 8. There are a number of criteria to consider when setting up an ionexchange chromatography sys tern, e.g., the choice of ion-exchanger, matrix support, buffer, and column. The most useful ion-exchange matrix for the purification of IgG antibody (only MoAbs of this class can usually be successfully purified by ion-exchange chromatography) is DEAE cellulose, of which there are several variants. The DEAE reactive group was originally coupled to fibrous cellulose (Whatman, Biorad), which resulted in poor flow rates and clogging of the matrix. It appears that coupling the DEAE to beaded cellulose (DEAE-Sephacel, Pharmacia) results in a matrix with a higher capacity and greater resolving potential than the fibrous forms. In general, an ion-exchange matrix will need to be regenerated and swollen before use according to the manufacturers instructions. Cellulose exchangers should be regenerated before and after use, which will ensure that they are washed free of residual protein, and protons or hydroxyl ions left bound weakly to the charged groups of cation or anion exchangers, respectively. DEAE-Sephacel, however, is supplied preswollen and ready to use. When choosing a buffer, the pH should be within the operational range of the ion-exchange matrix (pH 2-12 for DEAE-Sephacel) and suitable for use with the protein in question, in this case IgG. For a protein to bind to the ion-exchange matrix, the pH should be approximately one unit above its isoelectric point for anion exchangers (one unit below for cation exchangers). The ionic strength of the application buffer should be between O.Ol-0.05A4, and for reproducible

Baines, Gearing, and Thorpe fractionations, the water usedin the buffers should be as pure as possible. Furthermore the addition of charged bacterial inhibitors, such as sodium azide or thimerosal, should be avoided. When choosing a column for ion-exchange chromatography, there are a few rules to follow. The column should be of controlled diameter glass with as low a dead space as possible. Column length should be 20-30 cm and internal diameter 1.5-l .6 cm. Homemade columns are quite acceptable for most purposes, although commercial columns are available from several manufacturers. The column must be packed with the regenerated or ready to use ion-exchanger and equilibrated fully with the chosen starting buffer before use. Once the column is ready for use, the sample can be applied after preliminary purification, e.g., by ammonium sulfate precipitation and exhaustive dialysis against the starting buffer. The sample can then be applied to the column (volume is not an important consideration) and the eluant monitored for its absorbance at 280 nM, either manually or by passing the eluant through the flow cell of a UV absorbance monitor. If there is a high concentration of proteins in the eluant, then the ion-exchanger or sample were not fully equilibrated or the absorbing capacity of the matrix has been exceeded. After the sample has been applied, the ion-exchanger is washed with two column volumes of starting buffer to ensure complete elution of any unbound protein. The bound proteins are then eluted by changing the pHof the buffer or increasing its ionic strength. In general, variation in ionic strength is easier to control than variation in pH, and hence, this method is preferable; a final ionic strength of 1.0 is usually enough to elute most proteins. There are two ways of altering the ionic strength, either by increasing the concentration of the starting buffer or by increasing the concentration of other ions, e.g., NaCl. In the latter case, the buffering capacity and pH is kept constant throughout the separation procedure. The entire process can be automated by using a commercial gradient maker and fraction collector, although a laboratory made gradient maker is perfectly adequate. In general, the total volume of the gradient should be between five and ten times the column volume and the size of collected fractions should be about 1/5-l /lO of the column volume. The elution profile can be plotted by recording the optical density at 280 nM of the eluted fractions and the first major peak contains the IgG antibody. IgG of different isotypes will elute at different times in the gradient, e.g., IgG, elutes later than IgG,,. 9. Although ion-exchange chromatography works very well for rabbit IgG isolated from serum, it does not seem to be as effective in purify-

Murine Monoclonal

Antibodies

665

ing mouse IgG from either serum or samples containing MoAbs. One problem is that some immunoglobulins are unstable at low ionic strength, e.g., mouse IgG,, and hence, precipitation occurs during the sometimes lengthy ion-exchange procedure. Similarly, low ionic strengths have been reported to be deleterious to antibody activity. A number of studies have shown that DEAE cellulose chromatography after ammonium sulphate precipitation does not remove contaminating protease and nuclease activity. Furthermore, transferrin is not well separated by ion-exchange methods, and a further purification step involving gel filtration often has to be employed (6). The use of hydroxyapatite (hydroxylapatite) has been described for purification of MoAbs, but the authors have not found this to be very effective in most cases. Coupled with the need for long equilibration and recycling procedures inherent in ion-exchange techniques and the fact that only certain MoAbs will have a suitable charge for use in this method (all MoAbs are by definition different), conventional ion-exchange chromatography cannot be considered a rapid and efficient general method for the purification of mouse MoAbs. Sometimes (but not always) the use of an alternative support gel, e.g., DEAE Sepharose improves the efficiency of the procedure. 10. The FPLC system may be used in a fully automated mode, a manual mode, or a combination of both. The details of the FPLC hardware and its use are comprehensively described in the manufacturers manuals. Briefly, however, the system consists of two high-precision pumps connected to mixer for gradient formation, an automatic motor valve for injection of samples (samples are delivered manually to the valve via a loop with a syringe attachment), a single-path UV monitor (and control unit) for monitoring samples, and a programmable liquid chromatography controller for the full or partial automation of the FPLC system. A fraction collector and chart recorder complete the hardware requirements together with the column of choice, e.g., ion-exchanger, gel filtration, and so on. Simple and rapid purification of mouse IgG MoAbs may be achieved using the above apparatus and the Mono-Q HR5/5 anion exchange column. Several criteria require further consideration. It is essential that all buffers and the sample to be loaded are filtered prior to use (0.2 JJM millipore). All methods for column equilibration, elution of sample, and so on, are programmed into the liquid chromatography controller. The programs have to be put in by the user and the parameters of each program can be varied, e.g., flow rate, gradient times, and so on. In practice, for purification methods, we have found that a flow

Baines, Gearing, and Thorpe rate of 0.5 mL/min with a gradient of 35% buffer B for 2 min followed by a gradient of 100% buffer B for 8 min is adequate for rapid and single-step purifications of IgG MoAb from ascitic fluid. Although the method itself takes about 1 h to run, the MoAb is eluted after about 25 min, thus equilibration of column to the elution of the MoAb takes only about 40 min. A typical elution profile of mouse IgG MoAb is shown in Fig. lb. It is often useful to run a dilute sample of a&tic fluid preparation by the method outlined above in order to ascertain the sensitivity required in the system. Assuming ascitic fluid preparations contain about 5-10 mg/mL of MoAb, a 1 /lO dilution is appropriate. 11. In affinity chromatography techniques, the solid matrix used must have a loose porous mesh to allow good flow rates, should be chemically and physically stable, and should have chemically derivatized groups for interaction with ligands. For most protein ligands, agarose beads such as Sepharose 4B meet the requirements for MoAb purifications. Sepharose can be activated by cyanogen bromide (CNBr) at high pH and subsequently linked to the amino groups of proteins at neutral pH to yield a stable covalently linked complex. For some small ligands, it is sometimes better to couple to the gel matrix via a spacer arm that allows more efficient interaction with the MoAb. A number of spacer arm derivatized gels are available commercially. Polyacrylamide beads may also be used for coupling, but are not as versatile and will not be discussed further. Sepharose can be activated in the laboratory, provided a fume hood is available. The activated Sepharose formed is unstable and should be used immediately. A more convenient, but more expensive alternative is to use commercially available CNBr-activated Sepharose, which comes as a fairly stable freeze-dried powder. Once coupled, the matrix should be packed into simple columns either made from syringe barrels fitted with a sintered polythene sheet to retain the gel or in commercially available columns. These procedures can readily be scaled up for large-scale purification. 12. Three basic groups of affinity ligands can be used to purify MoAbs: class-specific reagents such as antiimmunoglobulin antisera, bacterial proteins, such as protein A, C or G, which bind to the Fc portion of some immunoglobulin classes, and antigens that specifically bind to individual MoAbs. In addition, certain gel matrices that form hydrophobic interactions with proteins can be used to purify MoAbs. The choice of ligand depends on the source of the MoAb, its subclass, and the degree of purity required. For culture supernatants, if fetal calf serum or serum-free conditions are used, the MoAb is usually the

Murine Monoclonal

Antibodies

667

predominant antibody species present. Less specific methods, such as protein A or antiimmunoglobulin columns, can then give a very pure product. For ascitic fluid where other host-derived immunoglobulins are present, a substantially pure MoAb will only be obtained by antigen-based affinity columns. For many immunochemical applications, however, it is sufficient to use MoAbs from ascites that have beenpurified on the basis of class or subclass either by protein A or hydrophobic chromatography. The immunoglobulin contaminating such preparations will consist of a diverse range of antigen specificities. The MoAb would therefore be in great excess over any other antibody species, and the preparation can be used in practice as a specific reagent. 13. All affinity methods involve a number of basic steps, all of which should be carried out at 4C. The column should first be preeluted with buffer containing the dissociating agent, which will be used to disrupt the ligand-MoAb complex. For MoAbs, O.lM glycine HCl pH 2.5 is most common, although 3M potassium thiocyanate, high salt, O.lM sodium borate pH 10, or 0.05M diethylamine pH 11.5 are also used as dissociating agents. The preelution step purges the column of any loosely bound ligand that would otherwise contaminate the purified MoAb. After preelution, the column is washed with running buffer such as PBS. The sample is then slowly applied and allowed to incubate for 15-30 min. Unbound proteins are eluted by washing with several column volumes of buffer until no protein is detected by absorbance at 280 nM. The MoAb is finally eluted from the column in the dissociating buffer, and the peak of protein is collected, neutralized,

and dialyzed to remove the dissociating agent. The column should be washed in buffer containing O.lM sodium azide prior to storage. The use of a protein A column for the purification of MoAbs of various subclasses is described. The effect of such a column in purifying an
IgG MoAb from ascitic fluid is also illustrated in Fig. 2, which shows an SDS polyacrylamide gel separation of proteins in ascitic fluid, and in the pH 2.5 eluate from the column. Figure 3 shows a similar gel of an IgM class MoAb directed against human IgM purified from ascitic fluid on a human polyclonal IgM column. Both of these figures clearly illustrate the power of affinity chromatography in the purification of

MoAbs.
14. The mini-gel is easily and quickly prepared consisting only of a resolving gel, and takes approximately 1.5 h to run once set up. The sample (MoAb) is prepared for electrophoresis under reduced conditions and is run in parallel with standard mol wt markers. The sample is loaded onto the mini-gel and run at 150 V for approximately

668

Bakes, Gearing, and Thorpe 1.5 h. The gel is then stained using Coomassie blue R to visualize protein bands, followed by destaining, after which the gel can be photographed and/or dried onto filter paper. A typical gel showing three different loadings of an FPLC purified mouse IgG MoAb is shown in Figs 3 and 4. Run under the reduced conditions described, the heavy chains have characteristic mol wt of approximately 50,000 and the light chains mol wt of approximately 22,000. The gel shows that the purified MoAbs are free from any detectable, contaminating proteins. The method described clearly indicates the purity of the MoAb. When this has been established, the biological activity of the MoAb must then be considered. Clearly, assays utilizing antibodies are extremely varied, and researchers will have to devise their own methods for assessing biological activity of purified MoAbs.

Acknowledgments
This investigation received financial support from the Special Program of Research, Development and Research Training in Human Reproduction, World Health Organization. We would like to thank Chris Bird for excellent technical assistance and Miss L. Hudson for typing the manuscript.

References
1. Johnstone, A. and Thorpe, R. (1987) Immunochemistry in Practice. 2nd Ed. (Blackwell Scientific Publication, Oxford). 2. Neoh, S. H., Gordon, S., Potter, A., and Zola, H. (1986) The purification of mouse monoclonal antibodies from ascitic fluid. J. Immunol. Mefhods 91,231-235. 3. Burchiel, S. W., Billman, J. R., and Alber, T. R. (1984) Rapid and efficient purification of mouse monoclonal antibodies from ascites fluid using high performance liquid chromatography. J. Immunol. Methods 69,33-42. 4. Bruck, C., Portetelle, D., Glineur, C., and Bollen, A. (1982) One-step purification of mouse monoclonal antibodies from ascitic fluid by DEAE affi-gel blue chromatography. J. Immunol. Methods 53,313-319. 5. Carlsson, M., Hedin, A., Inganas, M., Harfast, B., and Blomberg, F. (1985) Purification of in vitro produced mouse monoclonal antibodies. A two-step procedure utilizing cation exchange chromatography and gel filtration. J. Immunol. Methods 79,8998. 6. Clezardin, P., McGregor, J. L., Manach, M., Boukerche, H., and Dechavanne, M. (1985) One-step procedure for the rapid isolation of mouse monoclonal antibodies and their antigen binding fragments by fast protein liquid chromatography on a mono-Q anion-exchange column. J. Chromatography 319,67-77. 7. Dean, P. D. G., Johnson, W. S., and Middle, F. A. teds.1 (1985) Affinity Chromatogruphy u Prucficd Approach (IRL Press Limited, Oxford, Washington, DC).

Chapter 54
Rat x Rat Hybridomas
Christopher J. Dean

1. Introduction
There is an increasing interest in the preparation of rat x rat hybridomas, because they have been found to be more stable in culture than mouse hybridomas and they secrete consistently high levels (10 &mL and above) of monoclonal antibody. In addition, certain subclasses of rat IgG have been found to interact efficiently with human Fc receptors and the human complement system. More important perhaps in practical terms is that, when grown as an ascites, up to 30 mL of fluid containing 1-5 mg/mL specific antibody can be obtained from each rat. Furthermore, when nude rats are used for ascites production, the levels of endogenous immunoglobulin can be less than 1 mg/mL. Two rat myelomas are in general use today, both derived from ileocecal plasmacytomas of the LOU strain of rats, namely R21O.Y3 Ag 1.2.3, abbreviated Y3 (I), which secretes Kappa light chains, andIR983F (Z), which is a nonsecretor. Although the IR 983F grows normally in suspension, the Y3 myeloma is strongly adherent to glass and tissue culture plastics. To obtain cells that fuse well, it is essential to grow cells in spinner cultures. For a general introduction to hybridoma production, the reader is referred to the articles by Galfre and Milstein (3) and Bazin (4), as well as the chapters in this volume. Essentially, the procedure is to fuse myeloma cells that lack hypoxanthine phosphoribosyl transferase (HGPRT) and so
669

670

Dean

cannot grow in media containing hypoxanthine, aminopterin, and thymidine (HAT), with specific antibody producing B-cells that contain the HGPRT gene. Only stable hybrids containing the HGPRT gene will grow in HAT selection medium. Perhaps the two most important requirements for successful hybridoma production are first the ability to produce specific antibody-producing B-cells in the spleen or lymph node of the immune donor, and second the need for a rapid, sensitive, and specific assay for the antibody required. The appearance of high titers of antibody in serum will show that immunization has been successful, but it is not always possible to elicit at the same time the specific B-cells required for fusion. In this case, it is better to allow the response to decline (weeks to months) before reimmunizing just before fusion. Our experience with proteins and intact cells shows that two immunizations via the Peyers patches 14 d apart followed by cell fusion 3 d later is a useful and quick way of generating the specific B-cells. We immunize routinely via the Peyers patches because the stimulated mesentericnodes yield ahigher proportion of IgG-producing hybridomas as well as a proportion of IgA secretors (5). Rat x rat hybridomas grow rapidly with generation times often similar to that of the parental myeloma (lo-12 h), and it is essential to clone cells from positive wells as soon as possible. We plate fusion cultures into 4 x 24 well plates and pick individual hybridoma colonies from antibody positive wells by aspiration with a Pasteur pipet. After rescreening the picked colonies, those producing specific antibodies are cloned when about lo6 cells have been obtained (equivalent to about two confluent wells of a 24-well plate). One final point to note is the need to freeze down stocks of cells as soon as possible. Even the best run laboratory suffers with contamination at times, and there is nothing more frustrating than loss of a promising hybridoma.

2. Materials
1. Rats: Any strain may be used, preferably of lo-12 wk of age at commencement of immunization. 2. Immunization: Mammalian cells and membrane vesicles can elicit good antibody responses without further treatment. Soluble antigens, however, usually require injection as an emulsion in an oil base (Freunds adjuvant) to achieve good responses. a. Live mammalian cells (2-5 x 106/immunization/rat) or membrane vesicles as a suspension in phosphate buffered saline (PBS) or Dulbeccos Modified Eagles Medium (DMEM).

Rat x Rat Hybridomas

671

b. Macromolecules (e.g., protein, carbohydrates, DNA-methylated BSA precipitates at XI-100 pg/immunization/rat) in normal saline or PBSand emulsified just before use, by highspeed vortexing, blending, or ultrasonication, with an equal volume of Freunds adjuvant to give a stable emulsion. Emulsion stability can be demonstrated by allowing a drop from a Pasteur pipet to fall onto water. With a stable emulsion, the drop contracts and remains floating on the water surface, whereas an unstable emulsion disperses. c. Peptides or other low mol wt haptens conjugated to carrier proteins, such as ovalbumen, keyhole limpet hemocyanin, or bovine thyroglobulin using glutaraldehyde or carbodiimide. Solutions or suspensions in PBS should be emulsified with Freunds adjuvant before use, 3. Cell fusion: a. Growth medium: DMEM is prepared from dry powder and supplemented with 50 IU penicillin, 50 pg streptomycin, 100 u.g neomycin/mL, and then filter sterilized. Medium is stored at 5-6OC and used within a 2-wk period. b. PEG solution: Weigh 50 g of polyethyleneglycol(1500 MW) into a capped 200-mL bottle, add 1 mL of water, and then autoclave for 30 min at 12OOC. Cool to about 70C, add 50 mL of DMEM mix, and after further cooling to ambient temperature, adjust the pH to about 7.2 with NaOH (the mixture should be colored orange). Store as I-mL aliquots at -20%. c. HAT selection medium: Prepare 100 x HT by dissolving 136 mg of hypoxanthine and 38.75 mg of thymidine in 100 mL of 0.02N NaOH prewarmed to 60C. Cool, filter sterilize, and store at -2OOCin l- and 2-mL aliquots. Prepare 100x A by dissolving 1.9 mg aminopterin in O.OlNNaOH, then filter sterilize, and store as 2-mL aliquots at -2OOC.HAT medium is prepared by adding 2 mL of HT and 2 mL of A to 200 mL DMEM containing 20% fetal calf serum. d. HT medium: Add 1 mL of HT/lOO mL DMEM containing 20% fetal calf serum. e, Fetal calf serum: Use only batches that have been shown to support the growth of hybridomas. f. Feeder cells for fusion cultures (seeNote 1): 3 h to 1 d before fusion, 2-4 x lo6 rat fibroblasts in 10 mL DMEM and contained in a 30-mL plastic universal are irradiated with about

672

Dean 30 grays (3000 rad) of x or y-rays. Dilute with 90 mL of HT (3 h before fusion) or DMEM-10% FCS (I d before fusion) and plate l-mL aliquots into four 24-well plates. g. Y3 Ag 1.2.3 myeloma: About 5 x lo6 cells of Y3, stored frozen in liquid nitrogen as 1-mL aliquots in 5% dimethyl sulfoxide/95% FCS are thawed quickly at 37OC, centrifuged down (5008 x 2 min), then resuspended in 100 mL DMEM10% FCS placed in a 200-mL spinner flask. Stand for 2 d at 37OC to allow cells to attach to base of vessel, and then place spinner flask on a Bellco magnetic stirrer running at about 160 rpm. Cells can be maintained in exponential growth for up to 1 mo by fourfold dilution daily with DMEM-10% FCS. Myeloma cells are ready for use&5 d after commencement of stirring (seeNote 1). h. IR 983F myeloma: Cells growing exponentially in flasks, roller bottles, or spinner culture are maintained by daily fourfold dilution with DMEM-10% FCS. Unlike theY3 myeloma, these cells do not readily attach to glass or plastic. i. Freezer medium: Freshly prepared 5% dimethyl sulfoxide/ 95% fetal calf serum.

3. Methods
3.1. Immunization
for Spleen

Cell Fusions

1. Rats are anesthetized and a 0.5-mL sample of blood (preimmune) is taken from the jugular vein into a 1.5 mL capped microcentrifuge tube. Subsequently, after clotting at room temperature, the serum is removed, clarified by centrifugation and stored at -20C. 2. Antigens, either in suspension (cells) or emulsified with complete Freunds adjuvant (soluble materials), are injected at five sites (4x subcutaneous and 1 x intraperitoneal) using 0.1 ml/site. 3. Fourteen days later, the rats are again anesthetized, bled (0.5 mL), and then rechallenged at five sites with antigen using incomplete (no bacteria) Freunds adjuvant where necessary. 4. Repeat immunization at 14 d or longer intervals until test bleeds show high titers of specific serum antibodies using the preimmune serum as control. 5. When satisfactory titers of antibody have been achieved, give the rats an intravenous boost of antigen (no adjuvant), and 3 d later, remove the spleens for cell fusion.

Rat x Rat Hybridomas

673

3.2. Immunization

for Mesenteric -(See Note 3)

Node Cell Fusion

1. Rats are anesthetized, bled from the jugular, and then the lower abdo2. 3. men opened along the central line. Carefully extend the small intestine and locate the 8-16 Peyers patches (strain dependent) that lie along the peritoneal wall of the small gut. Take up antigen-containing samples (in PBS or emulsified with complete Freundls adjuvant) into a l-mL syringe using a 27-gage needle, inject between 10 and 15 PL into every other Peyers patch to give a total dose of between 0.05 and 0.1 mL/animal (seeNote 4). To prevent the formation of adhesions, place 2 mL of sterile 0.9% NaCl (containing antibiotics if necessary) in the peritoneal cavity and close the abdomen. Two weeks later, test bleed the rats and rechallenge intraperitoneally with antigen in 0.5 mL PBS or incomplete Freunds adjuvant. Four weeks after the initial challenge, anesthetize the rats, test bleed, open the abdomen parallel to the initial incision, and reimmunize using theunchallenged Peyers patches as recipient for antigen. Remove the mesenteric nodes 3 d later, and use for cell fusion.

4. 5. 6.

3.3. Hybridoma
3.3.1. Preparation

Production
of Cells for Fusion

1. Centrifuge exponentially growing cells of Y3 (see Note 1) or IR 983F in 50-mL aliquots for 3 min at 4008, wash twice by resuspension in se-

rum-free DMEM, count in a hemocytometer, and resuspend in this

medium to l-2 x 10 cells/ml. 2. Kill immune rat by cervical dislocation or CO, inhalation, test bleed,

and open abdominal cavity. Remove mesenteric nodes and spleen by blunt dissection.
3. Disaggregate nodes or spleen by passing through a fine stainless-steel sieve (e.g., a fine tea strainer) into 10 mL of serum-free DMEM using a spoon-head spatula. (Dip it into ethanol and flame to sterilize.) 4. Harvest cells by centrifuging for 5 min at 4008, wash twice in serumfree DMEM, and resuspend in 10 mL of the same medium. 5. Count viablelymphoid cells in a hemocytometer. Immune rat spleens generally yield between 3-5 x lo* cells total, whereas the nodes can yield up to 2 x lo8 lymphocytes.

674

Dean 3.3.2. Fusion Protocol Mix lo8 viable lymphocytes with 5 x 10 myeloma cells in a lo-mL sterile capped tube, and centrifuge for 5 min at 4008. Pour off the supernatant, and while draining remove any surplus medium carefully with a Pasteur pipet. Release cell pellet by gently tapping the tube on bench. Over a period of 1 min, stir 1 mL of PEG solution, prewarmed to 37OC, into the pellet. Continue mixing for a further minute by gently rocking the tube. Dilute the fusion mixture with DMEM (2 mL over a period of 2 min, then 5 mL over 1 min). Centrifuge for 5 min at 4OOg,then resuspend the cells in 200 mL HAT selection medium, and plate 2-mL aliquots into four 24-well plates seeded with irradiated fibroblasts. Screen cultures for specific antibody&14 dafter commencement of incubation at 37C in 5% CO,-95% air (seeNote 5). With a Pasteur pipet, pick Individual colonies into 1 mL HT medium contained in 24-well plates. Split and feed with HT when good growth commences. Freeze samples in liquid nitrogen. Rescreen the picked colonies and expand positive cultures. Freeze samples of these in liquid nitrogen and reclone twice. 3.3.3. Cloning of Rat Hybridomas Prepare a suspension of irradiated rat fibroblasts in DMEM-10% FCS, and seed at 5 x lo3 cells/well into 96-well plates. The next day, centrifuge down cells from at least two wells of a 24-well plate that contains a confluent layer of hybridoma cells. Carefully count the number of cells, and then dilute to give about 50 cells in 20 mL of HT or DMEM-10% FCS. Carefully flick off the supernatant medium from the feeder cells, and plate 0.2-mL aliquots of hybridoma cells into each of the 96 wells. Examine plates 5-10 d later, and screen those wells that contain only single colonies. Pick cells from positive wells into 24-well plates, expand, and freeze in liquid nitrogen. Repeat by recloning the best antibody-producing colonies.

1.

2. 3. 4. 5. 6. 7, 8.

1. 2.

3. 4. 5. 6.

3.3.4. Bulk Cultures Rat hybridomas grow well, but more slowly, in DMEM containing3% fetal calf serum, and they can be quickly adapted by reduction in concentrations of serum over a period of 1 wk. To bulk up cultures in vitro, seed

Rat x Rat Hybridomas

675

about lo6 cells into a 24-cm2 flask containing 10 mL DMEM-10% FCS. Use cells from this flask to seed an 80-cm2 flask containing 3 mL DMEM and cells for two 80-cm2flasks to seed an 800-cm2roller bottle or 200-mL spinner flask containing 100 mL DMEM-10% FCS. Hybridomas produced with either rat myeloma grow well in both roller and spinner cultures, and antibody yields of up to50 pg/mL can be obtained. Cultures are fed daily with DMEM-10% FCS and on alternate days when DMEM-3% FCS is used.
3.3.5. Ascites

In general, rat x rat hybridomas grow well in the peritoneal cavity of nude rats, which is an advantage because the levels of endogenous immunoglobulin are usually low (<lmg/mL). 1. Inject 5 x 106-lo7 hybridoma cells, taken either from roller/spinner culture or ob tained from a previously grown ascites and stored frozen, into the peritoneal cavity of each nude rat. 2. When tumor growth is visible by the swelling of the abdomen, drain off the ascitic fluid into a syringe containing heparin to prevent clotting. If tumor does not grow as an ascites, use rats that have been injected ip with 1 mL pris tane 7-10 d previously. 3. Centrifuge down cells at 6008 for 10 min, aspirate off the supernatant (seeNote6), add theproteaseinhibitorphenylmethylsulfonylfluoride to a concentration of lOaM, and store at -2OOC. Collect the cells into 5% DMSO-95% FCS and freeze in liquid nitrogen.

4. Notes
1. Feeders for fusion culture. A useful source of rat cells that can bemaintained in culture can be prepared from the xiphoid cartilage that terminates the xiphisternum. Cartilages from 6-8 adult rats are chopped into 2-3 mm pieces with a scalpel and stirred for 45 min at room temperature in 15 mL DMEM containing 0.5% trypsin (bovine pancreas type III) and 1% collagenase (type II). Add FCS to 10% and filter through sterile gauze to remove debris. Wash cells in DMEM-10% FCS and plate into same medium. Passage in DMEM-10% FCS after removing cells by incubation for 2-3 min in PBS-0.05% N%EDTA containing 0.25% trypsin. Store cells in liquid nitrogen as aliquots of lo6 cells in 5% DMSO-95% FCS. 2. The Y3 myeloma attaches firmly to glass and plastic surfaces in static culture, and cells produced in this way do not fuse well. Our experience suggests that good fusions are obtained only using cells taken from exponentially growing spinner culture.

676

Dean

3. Surgical and immunization procedures: The mesenteric lymphatic system can be visualized by injecting a small amount of lymphography dye (Guerbets patent blue V, May and Baker, Dagenham, England) into a Peyers patch. 4. Administration of the antigen to the Peyers patch is facilitated if the region of the gut containing the Peyers patch is gently squeezed either with forceps or between thumb and forefinger, so that thehypodermic needle can be inserted just beneath the capsular membrane of the Peyers patch. 5. The test system selected for screening of hybridoma supernatants, e.g., RIA or ELISA using cells or antigen-coated multiwell plates or immunostaining of tissue sections and blots of polyacrylamide gels, will depend on the antigen used for immunization. We recommend the use of sheep or rabbit antibodies to polyclonal rat F(ab>, as the second stage reagent for the primary screen, because neither protein A (which binds with high affinity only to rat IgG,) nor the mouse antirat kappa monoclonals (MARK-l and MARK-3) bind to all rat antibodies. If a particular antibody isotype is required, specific antiheavy chain antisera should be used either at the primary screen or after the hybridomas have been picked from the fusion wells. 6. Purification of rat monoclonals is most easily achieved by affinity chromatography on antigen linked to Sepharose 4B (seeChapter 53 this volume). Where this is not possible, IgG, and IgGgb antibodies can beisolated from culturesupernatants and ascitic fluid by precipitation with (NH,),SO, added to give 40% of saturation and purified by ionexchange chromatography on Whatman DE52, Pharmacia Mono-Q,, or similar product. Alternatively, affinity chromatography can be carried out using Protein A-Sepharose for IgGZCantibodies or Sepharoselinked antiheavy chain antibodies for other isotypes. Immunoaffinity columns of MARK-l are a useful general reagent for the isolation of rat monoclonal antibodies. However, when hybridomas have been made using the Y3 myeloma, it is important to check the protein prepared in this way both by SDS-PAGE and for antigen binding because MARK1 has a high affinity for the Y3 kappa chain (secreted by most Y3 hybridomas), whereas not all of the rat monoclonal antibodies bind to MARK-l.

References
2. Galfr6, G., Milstein, C., and Wright, B. (1979) Rat x rat hybrid myelomas clonal anti-Fd portion of mouse IgG. Nature 277,131-133. and a mono-

Rat x Rat Hybridomas

677

2. Bazin,H. (1982)Production of monoclonal antibodies with the LOU rat nonsecreting IR 983Fmyeloma cell line, in Profidesoffhe BiologicalFluids, vol. 29 (Peeters,H., ed.), Pergamon, Oxford, pp. 615-618. 3. Galfre, G. and Milstein, C. (1981) Preparation of monoclonal antibodies: strategies and procedures, in Methodsin Enzymology,, vol. 73 (Langone, J.J. and van Vunakis, H., eds.), Academic Press,New York and London, pp. 346. 4. Bazin, H. (1987) Rat x rat hybridoma formation and rat monoclonal antibodies, in MethodsofHybridoma Formation (Bartal, A. H. and Hirshaut,Y., eds.), Humana Press, Clifton, New Jersey, pp. 337-378. 5. Dean, C. J.,Styles,J.M., Gyure, L. A., Peppard, J.,Hobbs, S.M., Jackson,E.,and Hall, J. G. (1984) The production of hybridomas from the gut associated lymphoid tissue of tumor bearing rats. I. Mesenteric Nodes as a Source of IgG Producing Cells. Clin. Exp. Immunol. 57358-364.

Chapter 55 Monoclonal Antibodies Against Glycosphingolipids (GSLs)-Gangliosides

Bruce Fletcher
1. Introduction
Free GSL molecules are poorly immunogenic using conventional immunization procedures (1,2). Although immunogenicity has been increased by immunizing with purified glycolipids coated onto the surface of SaZmone2Za minnesota organisms (3,4), this has not always been successful (5). Theoretically, one can take advantage of GSLs absorbed on silica as occurs in thin-layer chromatography (t. 1.c.) separation (6), and then to use the silica-absorbed GSL spots scraped from the plate as immunogen. This proposition is based on the principle that aggregation and/or presentation on solid surfaces often leads to increased immune responses (7), and is supported by the demonstration that silica enhanced the response to poorly immunogenic proteins in an in vitro system (8). Most attempts at inducing high-titer antibody responses and monoclonal antibodies (mAbs) of the IgG type involve immunization schedules consisting of several injections given over a fairly long period. An adjuvant is often employed, at least for the primary immunization. Spitz et al. (9) found that intrasplenic immunization had the advantages that a very small amount of immunogen without adjuvant is required, and that it induced 679

Clark

680

Clark

a rapid response, a 34-d period only, being necessary between immunization and fusion. Such a short schedule, which is equivalent to the primary immunization of conventional procedures, is likely to produce antibodies of the IgM rather than the IgG class. This represents a disadvantage, since IgG antibodies are often preferred. However, in seeking to improve the efficiency and speed of producing antibodies against GSLs, this is not considered a major disadvantage, especially in view of the possibility of class switching by selection of somatic cell variants that arise spontaneously or can be induced by mutagens (ZO,Zl). Also, it is possible to administer a second intrasplenic injection at a later time, thereby increasing the probability of obtaining antibodies of the IgG class. A modified version of the technique described by Galfre and Milstein (12), the principles of which have been outlined earlier in this series (13), was used to fuse immune mouse spleen cells with the azaguanineresistant, nonantibody secreting, mouse myeloma cell line, NS-1. Hypoxanthine-aminopterin-thymidine (HAT) medium was used to select monoclonal antibody-producing hybrid cell lines. These principles have been successfully applied to the generation of mABs against gangliosides from fetal brain, monosialoganglioside (GMI) from a commercial source, and total glycolipids from colonic mucosa.

2. Materials
2.1. Stock Solutions
1. CulturemediumRl?MI1640 withantibiotics (penicillinlOOU/mLand streptomycin 100 pg/mL). Store at 4OC. 2. Fetal Calf Serum (FCS) (Note 1). Inactivated by heating at 56C for 30 min and dispensed into aliquots of 20 mL (store at -2OOC). 3. Supplements to Media (100x strength): a. Oxaloacetic Acid (15 mg/mL). b. Sodium Pyruvate (5 mg/mL). c. Insulin (100 iu/mL). d.Tris Base (2&I). 4. For Selection media (100x strength): a. Hypoxanthine plus Thymidine: Dissolve 0.1361 g and 0.0388 g, respectively, in double-distilled water at 70-BOOC. b. Aminopterin: Dissolve 17.6 mg aminopterin in 80 mL water, adding O.lM NaOH if solution is not readily achieved. Adjust the pH to 7.2 with HCl and make up to 100 mL.

Gangliosides
c. Sterilize solutions of supplements and selection agents by filtration through single-use 0.22~pm membrane filters and dispense into 2 mL aliquots. Store at -2OOC.

681

2.2. Extraction and Purification of Glycolipids and Immunization

1. Chloroform: methanol (2:l v/v and 1:2 v/v with 5% water). 2. Aluminium-backed silica-coated high-performance thin-layer chromatography (HPTLC) sheets (Art5553; E. Merck, Darmstadt, West Germany). 3. Authentic GSLs (for use as references) (Supelchem UK, Radley and Co. Ltd., London, England). 4. Orcinol-ferric chloride reagent. 5. D&isopropyl ether/l-butanol/50 mM aqueous sodium chloride (6:4:5 v/v/v>. 6. Silica gel chromatogram carrying GSLs. 7. 10 mM sodium phosphate buffer, pH 7.2,0.17M NaCl.

2.3. Fusion of NM Cells with Immune Spleen and Cultivation of Hybridomas

Cell

1. P3/NSl/l-Ag4-1 cells (22), and immune spleen cells. 2. Waterbath (made by placing a beaker of water inside another containing water and kept at 37OC, until needed). 3. 50% solution of polyethylene glycol (PEG, MW, 1540), prepared as follows: weigh 1 g of PEG in a glass universal bottle and me1t in a microwave oven with the bottle top loose. As the PEG cools, but before it solidifies, add 1 mL RPM1 1640 and mix. Store at 37OCbefore use.

2.4. Cloning,

Growing,

and Storing

Hybrid

Cells

1. Glass capillary pipet (made by stretching a 100~PL accupette [DADE Diagnostics Inc., Aguada, Puerto Rico] in a Bunsen burner flame to a diameter, estimated visually, about l/4 that of the low power field of an inverted microscope). 2. 10% analar DMSO (dimethylsulfoxide)/90% FCS. Prepare by adding DMSO dropwise to FCS with mixing to prevent precipitating proteins. Make up fresh and cool on ice before use. 3. Sterile transfer pipets (Pastettes, LW4010, Apha Laboratories, Eastleigh, Hants).

682

Clark

3. Methods
3.1. Extraction and Separation
1.

of Glycolipids of Gangliosides

2. 3. 4. 5. 6. 7.

This method is essentially as described by Ladisch and Zi-Liang (14). Homogenize with a Polytron homogenizer 1 g tissue in 1 mL water. Add 40 mL of chloroform/methanol (2:l v/v>; leave for 2 h, and then vacuum fiter through sintered glass. Reextract the remaining homogenate with 10 mL chloroform/methanol (1:2 v/v plus 5% water). Pool extracts. Dry extract under N, at 7OOC. Partition the dry extract (30 vol/g tissue) in di-isopropyl-ether/l-butanoll50 mM aqueous sodium chloride (6:4:5 by vol). Remove the lower phase containing gangliosides, repartition, separate, lyophilize, and dissolve in distilled water (1 mL). Remove low mol wt components by Sephadex G-50 filtration (300 x 15 mM column) (15). Lyophilize the void volume, and redissolve the gangliosides in 1.0 mL chloroform/methanol (I:1 v/v). Centrifuge for 5 min at 10,000 x g at 4C to remove trace residual protein.

3.2. Thin Layer Chromatography of Total GSLS and Gangliosides

In advance, equilibrate a thin-layer chromatography tank lined with chromatography paper, overnight at 4OCwith about 100 mL chloroform/ methanol containing 0.2% KC1 (60:35:8, by vol) or, for greater resolution of gangliosides, with chloroform/methanol/0.2% aqueous calcium chloride (60:40:9 by vol) at room temperature. 1. Apply chloroform/methanol extracts containing glycolipids (gangliosides) equivalent to lo-20 mg wet wt of tissue, or authentic individual GSLs (l-10 pg), to 0.5-l-cm wide lanes on a chromatogram sheet. This is best achieved using a short piece of solvent-resistant plastic tubing attached to a 50-PL Hamilton syringe. (Designate parallel lanes as reference tracks to be visualized chemically.) 2. Develop chromatogram in appropriate solvent system over a distance of 8.0 cm. This takes about 45 min. 3. Dry the chromatogram in a fume hood, cut reference lanes from the sheet, and visualize GSLs with orcinol-ferric chloride reagent as

Gangliosides

683

described on the reagent bottle. (Use an acid-resistant fume extraction hood, and wear protective gloves and glasses.)

3.3. Immunogen and Immunization Procedure (Essentially as Described by Spitz et al. 191; See Notes 2 and 3)
1. Gently scrape silica and associated GSLs from an unstained lane of the chromatogram with a scalpel blade (Shape 13 x 23), using a parallel stained lane to locate the region(s) of interest. 2. Collect the silica into a glass Bijou bottle with the aid of a glass funnel (Note 4). 3. Suspend the silica plus associated glycolipids in sodium phosphate buffer (10 mg wet wt equivalent of tissue/mL) by using sonication until homogeneous by visual inspection. Pick up in a 1-mL glass syringe with a l/2 in x 27-gage needle. 4. Anesthetize adult Balb/c mice by ip injection of 6-7 PL of nembutal/g body wt. 5. Swab the abdomen with 70% ethanol, and place the animal on its right side. 6. Make a skin incision, and blunt dissect skin from the body wall to expose the abdominal wall. 7. Incise the body wall to expose the spleen. 8. Lift the lower pole of the spleen and insert the l/2 in x 27-gage needle deeply into the spleen, and inoculate 50 PL from a I-mL glass syringe of the suspension of silica and associated GSLs as the l/2 in x 27-gage needle is pulled out (Note 5). 9, Push back the spleen, suture the body wall with thread, approximate and close the skin edges with stitches. lO# Repeat the immunization after 1 wk.

3.4. Production

of Hybridomas

In advance, grow NS-1 cells (Note 7) using the following procedure: 1, About 1 wk before fusion, remove and thaw about 4 x 105NSI cells from liquid nitrogen storage (seebelow). 2, Wash the cells twice by centrifugation in RPM1 1640 for 5 min at 1200 X8* 3. Transfer the cells, in a small volume of medium, to a 750-mL culture flask with 20 mL RPM1 1640plus 20% FCS and supplements (put 2 mL

684

Clark

of each stock solution of oxaloacetic acid and sodium pyruvate; 40 iu Insulin and 200 PL 2M Tris Base to 200 mL of medium). 4. Put the flask on its side, with the cap loose, in a humidified CO, incubator (7% CO, in air gas mixture). 5. Feed every 2 d by addition of medium to expand the culture, keeping the cell density between 104/mL and 2 x 105/mL. (Cultures with densities less than 104/mL grow slowly and lose viability above 2 x 105/ mL.) This should produce about 2 x lo5 cells for fusion, in log-phase growth at about lo5 cells/mL. Prepare a backup culture in case of miscalculation or accidental loss. Also in advance, prepare and plate out feeder cells (peritoneal macrophages) by the following procedure: 1. Twenty-four h before the day of the fusion, sacrifice four, large, mature, Balb/c mice (enough for eight culture plates), sterilize by immersion in 70% ethanol, blot dry, and put in cell culture hood. 2. Make a mid-line cut along the skin of the abdomen, and then makefurther cuts to allow the skin over the entire abdominal area to be dissected away exposing the body wall. 3. Inject ip (small 23-gage needle) 5.0 mL of RPM1 1640 plus 20% FCS and supplements (ice-cold to avoid cells sticking to walls of containers), and gently massage the abdomen for about 1 min. 4. Aspirate the fluid from the abdominal cavity using the same syringe with the larger IPgage needle pushed through the same hole in the body wall. This avoids leakage. Take care not to puncture internal organs (visible through the body wall) while manipulating the fluid into the syringe. Pool the aspirates and make up a volume with cold medium (about 10 5. mL/plate), so that drops of about 100 PL from a lo-mL plastic disposable pipet can be delivered to each well of 8 x 96-well culture plates. One mouse provides enough cells for two plates (2-5 x lo4 cells/well). 6. Put the plates in an incubator overnight.

3.5.

Fusion

of Immune Spleen with NSl Cells

Cells

Three days (Note 8) after the second injection of immunogen animals are sacrificed and immune splenocytes immortalized by fusion with NSl cells:

Gangliosides

685

1. Warm supplemented RPM1 1640 with 20% FCS and serum free medium to 37C. 2. Harvest NSl cells by rubbing the bottom surface of the culture flask with a bent Pasteur pipet and then repeatedly rinsing gently with the medium using a lo-mL pipet. Do this slowly to minimize damage to the cells. 3. Wash twice with serum-free medium by centrifugation in a bench centrifuge at 1200 x g for 5 min, and resuspend in medium. 4. Count viable NSl cells using an improved Neubauer hemocytometer. View the cells directly using a phase contrast microscope (viable cells appear surrounded by a bright halo). Alternatively, dilute the cell suspension l/2 with 0.1% Trypan blue in PBS, and examine by light microscopy (viable cells appear free of stain). The formula for calculating the number of cells/mL of medium is 2 (if the medium is diluted with Trypan blue) x number of cells within the outermost triple boundary lines x 104. 5. Remove the spleen(s) aseptically. 6. Release cells from l-2 spleens by teasing the tissue with forceps in RPM1 1640 in a bacterial (Note 9) culture plate. 7. Gently transfer the medium to a centrifuge tube, and allow fragments of tissue to sediment. 8. Transfer supernatant medium containing splenocytes to another tube, and wash cells 3 x by centrifugation in serum-free medium at about 1200 x g for 5 min. 9. Mix together 2-4 x lo7 NSl myeloma cells and the spleen cells in a 50 mL centrifuge tube, and centrifuge at 400 x g for 10 min. 10. Remove supernatant, and put tube in water bath at 37OC for subsequent manipulations. Il. Add 1 mL of warm 50% PEG over 1 min, gently stirring the cell pellet with the pipet as the PEG is added. Continue to stir gently for another minute. 12. Using the same pipet, slowly stir in 1 mL serum-free medium at 37C and repeat. 13. Stir in 7 mL of 37OC serum-free medium over 2-3 min. 14. Centrifuge at 400 x g for 10 min at room temperature, and remove the supernatant. 15. Aim a lo-mL pipet of 37OCsupplemented medium with 20% FCS at the cell pellet. Release the medium directly at the pellet, and then by stirring produce a suspension of fine clumps of cells.

686

Clark

16. Make up a volume with more medium such that, using a 10-mL pipet, drops of about 100 PL can be distributed into each well of 8 x g&well culture plates containing feeder cells. 17. Put the plates in the incubator at 37OC.

3.6. SeZection of Hybrid

Cells

1. On the day after fusion, make a 3 x solution of selection agents (HAT) in supplemented medium with 20% FCS (6 mL of each stock solution/ 200 mL medium). 2. Using a lo-mL disposable plastic pipet, distribute one drop of selection medium into each culture well (already containing two drops), thereby diluting the HAT to the required strength. 3. Every 2-3 d later, aspirate the medium and replace with fresh HAT medium.

3.7. Primary Screening to Identify Cultures Producing ReZevant Antibodies

1. Commencing 10 d after fusion, examine wells visually with an inverted microscope every 3-4 d for up to 30 d. 2. Remove aliquots of supernatant medium, for antibody screening, from those wells that contain sufficient hybridomas cells to cover at least half the bottom of the well. The cells can easily be distinguished from the feeder cells, since they should be round and bright. 3. The method chosen for screening will depend on the purpose of the investigation. A recent rapid, micro-enzyme-linked-immunoassay (16) has the advantage of high sensitivity, allowing the assay of as little as 5 yL culture medium. This has been found to work satisfactorily in detecting antibodies against GSL target antigens absorbed to the plastic wellplates.

3.8. CZonaZ Isolation by Manual Single

of Hybridomas

Cell Isolation

1. Remove one drop of hybrid cell suspension from a g&well plate with a pastette (plastic transfer pipet; seeitem 3 in Materials 2.4.) and disperse into 10 mL of the same medium in a lo-cm bacterial culture plate (hybridoma cells do not stick to bacterial plates).

Ganghides 2. Examine with an inverted microscope, and dilute into further plates until only 1-3 cells are seen per low-power microscope field. 3. Using a fine pipet drawn from an acupette (capillary pipet; seeitem 1 under Materials 2.4.) transfer single cells to wells of a 96-well plate containing feeder cells. 4. Expand the remaining cells in order to have enough to freeze, and store in liquid nitrogen as a backup stock in case of later failures. 5. Retest the supernatants of the clones for antibody after culture for an appropriate time, feeding every 2-3 d (Note 10). 6. Transfer at least two of the positive clones into 5-mL flasks with 1 mL of medium and put in the incubator, standing the flasks on end to concentrate the cells. When the culture becomes dense (1-2x l@/mL) add4 mL medium and lay the flask on its side. When the culture again becomes dense, transfer to a larger flask and continue feeding until enough cells are available for freezing in aliquots of about l-5 x 10 cells.

3.9. Freezing
1.

and Thawing

of Cells

2. 3. 4. 5. 6. 7. 8.

Count cells in growth medium- harvest sufficient cells to provide aliquots of l-5 x 10 cells. Spin cells in a bench centrifuge (1500 rpm for 5 min). Discard supernatant, and resuspend the cell pellet in the small amount of medium left in tube. Immediately add (dropwise from a pastette with mixing) sufficient 90% FCS/lO% DMSO (at 4OC)to resuspend cells such that 1 mL contains l-5 x 10 cells. Place diluted cells on ice while aliquoting out. Working quickly, aliquot about 1 mL to liquid nitrogen storage tubes on ice. Cap tightly. Immediately transfer capped tubes to dry ice (-70C) for 5-8 min (bury tubes in dry ice). Transfer tubes to liquid nitrogen storage (in drawers and racks for ease of access). Store at -196OC. To retrieve frozen cells from liquid nitrogen, remove storage tube and thaw rapidly by moving the tube(s) from side to side in a water bath at 37OC(wear protective gloves and glasses). Wash immediately with medium and resuspend at l-2 x lo6 cells/mL for optimum culture.

688

Clark

4. Notes
1. FCS should be screened and known to support growth of NSl cells. A good serum will support 70-80% cloning efficiencies of NSl. 2. These procedures were first developed using mixtures of gangliosides separated by t.1.c. Various checks discounted the unlikely possibility that the relatively abundant yield of antibodies (9-14% of wells containing hybrid cells were positive) was because of protein(s) in the immunogen. Thus, positive supernatants obtained after immunization with a single ganglioside (monosialoganglioside; GMI), obtained from a commercial source, showed similar activity when retested on the GM1 preparation after removing nonlipid material by reversephase chromatography. Also, amino acid analysis indicated negligible protein in the immunogen. Similar yields of positive wells were also obtained using chromatograms of total glycolipid extracts. It is important that the procedures using animals comply with legisla3. tive regulations. 4. Successful antibody production has been achieved with a complete spectrum of gangliosides from fetal brain associated with about 30 mg silica. Similarly, selected GSLs from colonic mucosa or 25 pg GM1 alone (associated with 15 mg silica) produced a successful yield of posi tive hybridomas. 5. Two intrasplenic injections have usually been given successfully 2 wk apart. 6. The use of an intermittant Bunsen burner, with instant on/off control, inside a sterile laminar flow hood is helpful and interferes minimally with the air flow. However, a solenoid gas safety inlet must be fitted. Keeping an open container of formaldehyde solution (38% w/v) in the hood overnight is also felt to be useful, as is the use of plastic sleeve covers in addition to surgical gloves. Any and every reasonable precaution should be taken, including not sharing the use of the hood with other workers. 7. Not all NSl cells are equivalent, and care must be taken to ensure that a clone that is able to fuse well with immune spleen cells is used. NSl cells are &azaguanineresistant and hencelack hypoxanthine-guanine phosphoribosyl-transferase. Long-term culture gives rise to genetic drift, which may affect continued resistance. Growth of cells for a few days in medium containing azaguanine (20 yg/mL) prior to growing for a fusion may, therefore, be necessary.

Gangliosides

8. The optimum time for obtaining maximum numbers of fused immune spleen cells does not coincide with that for getting high serum titers of antibody (25). 9. Spleen cells do not stick to the surface of bacterial culture plates. 10. It may be appropriate at this time to start weaning the hybrid cells from HAT to media containing, first, lx HT then, l/2 HT.

References
1.
2. 3. Weigandt, H. (1985) Gangliosides, in Glycolipids (Weigandt, H., ed.), Elsevier, Amsterdam, pp. 199-245. Alving, C. R. (1977) Immune reactions of lipids and lipid model membranes, in The Antigens, 4 (Sela, M., ed.), Academic, London, pp. 3-63. Young, W. W., MacDonald, E. M. S., Novinsky, R. C., and Hakomori, S. I. (1979) Production of monoclonal antibodies specific for two distinct steric portions of the glycolipid ganglio-N-tryosylceramide (asialo-GM21 I. Exp. Med. 150,1008-1019. Young, W. W., Portoukalien, T., and Hakomori, S. I. (1981) 2 monoclonal anticarbohydrate antibodies directed to glycosphingolipids with a lacto-N-glycosyl type-11 chain. J, Biol. Chem. 256,10967-10972. Brodin, T., Thurin, J., Stromberg, N., Karlsson, K. A., and Sjogren, H. 0. (1986) Production of oligosaccharide-binding monoclonal antibodies of diverse specificities by immunization with purified tumor-associated glycolipids inserted into liposomes with lipid-A. Eur. J. Immunol. 16,951-956. Magnani, J. L., Brockhaus, M., Smith, D. F., Ginsberg, V., Blaszczyk, M., Mitchell,K. F., Steplewski, Z., and Koprowski, H. (1981) A monosialoganglioside is a monoclonal antibody-defined antigen of colon-carcinoma. Science 212,55,56. Crumpton, M. J. (1974) Protein Antigens: The molecular basis of antigenicity and immunogenicity, in The Anfigens (Sela, M., ed.> Academic, London, pp. l-78. Van Ness, J., Laemmli, U. K., and Pettijohn, D. E. (1984) Immunization in vitro and production of monoclonal antibodies specific to insoluble and weakly immunogenic proteins. Proc. NaB Acad. Sci. USA S&7897-7901. Spitz, M., Spitz, L., Thorpe, R., and Eugui, E. (1984) Intrasplenic primary immunization for the production of monoclonal antibodies. J. Immunol. Methods 70,3943. Morrison, S. L. and Scharff, M. D. (1981) Mutational events in mouse myeloma cells, a review. C. R. C. Cmf. Rev. lmmunol. 3,1-22. Yelton, D. E. and Scharff, M. D. (1982) Mutant monoclonal antibodies with altered biological functions. 1. Exp. Med. 156,1131-1148. Galfre, G. and Milstein, C. (1981) Preparation of Monoclonal Antibodies: Strategies and Procedures, in Methods in Enzymology 73 part B. Immunochemical Techniques (Lagone, J. H. and Van Vunakis, H., eds.), Academic, London, pp. 346. Wood, J. N. (1984) Immunization and fusion protocols for hybridoma production, inMefhods inMoleculav Biology 1: Proteins (Walker, J. M., ed.), Humana Press,Clifton, New Jersey, pp, 261-270. Ladisch, S. and Zi-Liang, W. (1985) Detection of tumor-associated ganglioside in plasma of patients with neuroblastoma. Luncef i, 136-138.

4. 5.

6.

7. 8.

9. 10. 11. 12.

13.

14.

690

Clark

15. Ueno,K.,Ando,S.,andYu,R.K. (1978)Gangliosidesof human,cat,andrabbit spinal cords and cord myelin. 1. Lipid Res.19,863-871. 16. Durbin, H. and Bodmer, W. F. (1987) A sensitive micro-immunoassay using p-galactosidase/anti P-galactosidase complexes. J. ImmunoZ. Methods 97,19-27.

Appendix
Compiled by

Jeffrey

W. Pollard

Table 1 Balanced Salt Solutions Earles, g/Lb KC1 =w4 MgCl.p6H,O MgSO,*qO CaC1, NaCl NaHCO, NaH,PO,*H,O Na,HPO, l 7H,O Glucose Phenol Red Reference
For detailed description @Sterilize by filtration. Sterilize by autoclaving. d12H,0 0.4
-

Hanks, g/Lb
0.4 0.06

Geys, g/Lb
0.37 0.03 0.21 0.07 0.17 7 2.27 0.226

Pucks, g/Lb
0.4 0.15
-

Dulbeccos PBS, g/L

0.2 0.2

0.2 0.02 6.68 2.2 0.14

0.1 0.1
0.14 8 0.35

0.154 0.012 8

0.1 0.1 8

2.31
-

1 0.01 1
of derivations,

0.09 1 0.01
2 see ref. 6.

1
3

Table 2 Eagles Minimum Essential Medium and Derivatives Eagles MEM, Component Amino acids L-Alanine L-ArginineeHCl L-Asparagine L-Aspartic acid L-Cysteine HCl L-Cysteine d&odium salt L-Glutamic acid L-Glutamine Glycine L-Hi&dine HClmqO L-Isoleucine L-Leucine L-Lysine*HCl L-Methionine L-Phenylalanine L-Proline LSerine L-Threonine L-Tryptophan L-Tyrosine L-Tyrosine: disodium salt L-Valine
l l l

Dulbeccos modification DMEM,

mg/L

CiMEMb
mg/L

Iscoves modified DMEM,

mg/L

Jokliks MEM,
mg/L

mg/L

tp zi s E! R

126.4

84

25 126.4 50 30

25 84 30

89.74
28.42 56.78
-

28.42 75

292.3
42 52.5 52.5 73.06

584 30 42 104.8 104.8 146.2 30 66

292
50 42 52.5 52.5 73.06

56.78 75 584 30 42 104.8 104.8 146.2 30 66

29.60 294 42 52 52 72.5 15 32

-

14.9
33.02
-

14.9
33.02

42 47.64 10.2 36.22

40 25
47.64 10.2 36.22

40
42

95.2
16 72

95.2
16
-

48

10
37.8

104.2

46.9

93.6

46.9

93.6

46

Eagles Minimum Eagles MEM, mg/L

Table 2 Essential Medium and Derivatives

Component Vitamins and lipids L-Ascorbic acid Biotin D-Ca pantothenate Choline chloride Folk acid i-inositol Nicotinamide lyridoxal HCI Riboflavin Thiamine HCl Vitamin B12 Cholesterol Caqe%0 CaCt, Fe(Nb,),*9&0 KCl MgSO, (anhyd) NaCl NaHCO,

Dulbeccos modification DMEM, q/L

CiMEMb mg/L 50 0.1 1

1

(cuntm~ed) Iscoves modified DMEM, mg/L

-

Jokliks MEM, mg/L

1 1 1 2 1 1 0.1

1
-

4 4 4 7.2 4 4 0.4 4

1
2 1

1
0.1 1 1.36 264.9 200 400 200 6800 2000 158.3

264.9 400
200 6800 2000 158.3

264.9 0.1 400

200 6400 3700 141.3

0.013 4 4 4 7.2 4 4 0.4 4 0.013 0.02 165 330 4-00 242.2 6500 2000 1500

MgScw-40

97.67
4505 3024 125 0.076 0.0173

Others Adenosine
-

10

10 Cytidihe 10 Reoxyadenosine 10 Deoxycytidine 10 Deoxyguanosine Dihydrostreptomycin sulfate Bovine serum 0.4 albumin 1000 4500 4500 Glucose 1000 10 Guanosine 5958 HEPRS 0.2 Lipoic acid . Penicillin G Sodium Phenol 15 10 15 17 Red Sodium 110 110 110 Pyruvate 0.1 Soybean lipid 10 Thymidine 0.001 Transferrin 10 Uridine 9 7 8 10 Reference For full lists of tissue cuhure medium and references, see ref. 27. It should be noted that medium formulations from company to company baMEM is often supplied without ribo and deoxyribonucleosides. cAntibiotics are often supplied with the medium or they can be added during preparation (see ako Table 4).

50

2000

75000 ru 10 110 21 vary somewhat

Table 3 Other Useful Media RrMl 1640, mg/L HAMS F12, mg/L 8.9 200 50 20 211 15.01 13.3
-

Component L-alanine L-arginine free base L-arginine HCl L-asparagine L-asparaginee w L-aspartic acid L-cysteine (free base) L-cysteine* HCI*qO L-cysteine HCl L-cystine L-cysteine, disodium salt L-glutamic acid L-glutamine Glycine L-Hi&dine free base L-Hi&line HClGJO L-Hydroxyproline L-Isoleucine
l

CMRLb 1066, mg/L 25 70

McCoys 5A,
mg/L

Medium 199,
mg/L

Waymouths MB752/1,
mg/L

13.9 42.1 45 19.97 31.5

25
-

70 30

75 60 61

30

50

35.12 -

260 20

0.099 -

15
-

20 300 10 15

14.7 146 7.5

-

75 100 50

22.1 219.2 7.5

23.66 66.82 100 50

150 350 50 128

20.96 20 50 3.94

20 10 20

20.96
-

21.88
10

39.36

20

19.67 25

L-Leucine L-Lysine HCl L-Methionine L-Phenylalanine L-Proline L-Serine L-Threonine L-Tryptophan L-Tyrosine L-Valine Vitamins L-Ascorbic Biotin D-Ca Pantothenate Choline chloride Folk acid i-inositol Menopthone sodium bisulpite trihydra te Nicotinic acid Nicotinamide p-Aminobenzoic acid Pyridoxal l HCl PyridoxineeHCI Riboflavin ThiamineeHCl
l

50 40 15 15 20 30 20 5 20 20

13.1 36.5 4.48 4.96 34.5 10.5 11.9 2.04 5.4 11.7
-

60 70 15 25 40 25 30 10 40 25 50 0.01 0.01 0.5 0.01 0.05

39.36 36.5 14.9 16.5 17.3 26.3 17.9 3.1 18.1 17.6 0.5 0.2 0.2 5 10 36

60 70 15 25 40 25 30 10 40 25 0.05 0.01 0.01 0.5 0.01 0.05

50 240 50 50 50 75 40 40 65 17.50 0.02 1 250 0.4 1

0.2 0.25 3 1 35

0.0073 0.48 13.96 1.3 18

0.04 1 1 0.2 1
-

0.025 0.05 0.025 0.025 0.01 0.01

0.5 0.5
1

0.019 0.025 0.025 0.05 0.025 0.025 0.01 0.01

0.062 0.062 0.038 0.34

0.5 0.5 0.2 0.2

1 1 10

Table 3 (Continued) RPMI 1640, Component DL-a Tocopherol phosphate, disodium salt Vitamin A acetate Vitamin B12 Cholesterol Inorganic salts CaCl,eanhyd. CaC1,*2H,O CaNO, l 4H,O CuSO, .5H,O Fe(NOJ,*9H,O FeSO,m7H,O KC1 KIvo4 MgCJe6H,O MgSO,*7H,O NaCl NaHCO, NaH,PO,eH,O Na,HPO,e7H,O ZnS04*7H,O
mg/L

HAMS F12,
mg/L

CMRLb 1066,
mg/L

McCoys 5A
mg/L

Medium 199,
mg/L

Waymouths MB752/1,
mg/L

0.005 100 400 100 6000 2200 1512 -

0.01 2 0.115 0.2 264.9

-

1.36 200 -

0.2
-

44 0.0025
-

0.834 223.6 122 7599 1176 268 0.863

400 -

0.1 400 60 200 6400 2200 530 200 6800 2200 158.3

200 6799 2200 140

-

150 80 240 200 6000 2240 566

Other Components Adenine sulpha te

10

5-AMP ATE, disodium salt Bactopeptone 2-deoxyribose Glucose Glutathione GuanineaHCl Hypoxanthine Linoleic acid Lipoic acid Phenol Red Putrescine
l wcl

0.2 10 0.5 1000 0.3 0.3 25

-

1802

20 0.161 -

17
-

Ribose Sodium acetate Sodium glucoronate Sodium pyruvate Triphosphopyridine nucleotide Thymidine Thymine Tween 80 Uracil Uridine triphosphate Xanthine CoCarboxylase
4%*

83 4.2
-

0.5 36.71
-

110

1
-

Component CoEnzyme A Deoxyadenosine Deoxycytid.ine*HCl Deoxyguanosine Diphyosphopyridine nucleotidee FP Ethanol (for lipid component) Flavin adenine n nucleotide 5-methyl-deoxycytidine Reference

RrMl 1640, mg/L

HAMS F12, mg/L

-

CMRLb 1066, mg/L 2.5

10 10

McCoys 5% mg/L

Medium 199, mg/L

Waymouths MES752/1, mg/L -

10

7 16
1

0.1

12

13

11

14

15

16

For a full list of the medium, their modifications and references, seerefs. 11 and 23. bCan be made with Hanks salts rather than Earles salts.

Appendix
Table 4 Useful Antibiotics for Tissue Culture Recommended concentration pg/mL
1 100

701

Antibiotic Ampho tericin B Ampicillin Chloramphenicol Erythromycin Gentamycin

Spectrum of action

Approximate stability, d

Kanamycin Nystatin Penicillin G Rifampicin Streptomycin Tetracycline

Fungi and yeast Gram-positive and negative bacteria Gram-negative bacteria Gram-positive bacteria and mycoplasma Gram-positive and gram-negative bacteria and mycoplasma Gram-positive and negative bacteria Fungi and yeast Gram-positive bacteria Gram-positive and gram-negative bacteria Gram-positive and gram-negative bacteria Gram-positive, gram-negative bacteria and mycoplasma

5 100

50 100 50 100 50
100

3 3 4

10

The concentration given is sufficient to control a mild infection for the length of time stated at 37OCwithout undue toxicity to cells (see ref. 27), for greater detail). Most media contain streptomycin and penicillin G. The use of other antibiotics, particularly clinically relevant ones such as erythromycin, should not be encouraged unless absolutely necessary. This is because media invariably goes into the drainage system and so increase the range of drug-resistant wild bacteria.

702
Table 5 Insect Cell Medium Graces insect tissue culture medium (ref. 18) Ingredient L-Isoleucine L-Phenylalanine L-Tryptophan L-Leucine L-Histidine*HCl*H,O L-Methionine L-Valine L-Arginine HCl L-Lysine l HCl L-Threonine L-Asparagine l H,O L-Proline L-Glutamine DL-Serine Glycine L-Alanine I)-Alanine L-Cystine disodium salt L-Tyrosine disodium salt L-Glutamic acid L-Aspartic acid D-Sucrose D-Fructose D-Glucose L-malic acid a-Ketoglutaric acid D-Succinic acid Fumaric acid p-Aminobenzoic acid Folic acid Riboflavin Biotin Thiamin HCl D-Calcium pantothenate Pyridoxine HCl Nice tinic acid I-Inositol Choline chloride NaH,PO, 2H,O CaCl, 2H,O
l l l l l

Pollard

mg/L 50 150 100 75 3378 50 100 700 625 175 397.7 350 600 1100 650 225 200 22.69 62.15 600 350 26680 400 700 670 370 60 55 0.02 0.02 0.02 0.01 0.02 0.02 0.02 0.02 0.02 0.2 1140 1325

Appendix
Table 5 (Continued) Graces insect tissue culture medium Ingredient MgCl, .2H,O MgSO, l 7H,O KC1 NaHCO,
mg/L

703

2280 2780 2240 350

BML-TC/lO for insect cell culture and production of nuclear polyhedrosis virus (ref. 19) Ingredient Fetal calf serum Tryptose broth KC1 NaH,PO, CaCl,*2H,O MgCQ6H20 MgSO, 7H,O NaHCO, Glucose
l

Values/L 100 mL
2.6 g 2.87 g 1.14 g 1.32 g 2.28 g 2.78 g 0.35 g lg

Tlus Graces vitamins

and amino acids without j3-alanine and o-serine as above.

704

Pollard

Appendix
I. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 22. 13. 14. 15. 16. 17. 18. 19.

References

Earle, W. R. (1943) Production of malignancy in vitro: IV The mouse fibroblast cultures and changes in the living cell. J. N&l. Cancer Inst. 4,X5-212. Hanks, J. H. and Wallace,R. E. (1949) Relation of oxygenand temperature in the preservation of tissues by refrigeration. Proc. Sot. Exp. Biol. Med. 71,196-200. Gey, G. 0. and Gey, M. K. (1936) The maintenance of humannormal cells and tumor cells in continuous culture. 1. Preliminary report: cultivation of mesoblastic tumors and normal tissues and rates or method of cultivation. Am. 1. Cancer 27,45-76. Puck, T. T., Cieciura, S. J., and Robinson, A. (1958) Genetics of somatic mammalian cells: III Long-term cultivation of euploid cells from human and animal subjects. J. Exp. Med. 108,945-955. Dulbecco, R. and Vogt, M. (1954) Plaque formation and isolation of pure lines with poliomyelitis viruses. J, Exp. Med. 99,167-182. Bashor, M. M. (1979) Dispersion and disruption of tissues. Methods Enzymol. 58, 119-131. Eagle, H. (1955) Amino acid metabolism in mammalian cell cultures. Science 130, 432437. Dulbecco, R. and Freeman, G. (1959) Plaque production by polyoma virus. Virology 8,396-397. Stanners, C. P., Eliceiri, G. L., and Green, H. (1971) Two types of ribosomes in mousehamster hybrid cells. Nufure Nezu Biol. 230,52-54. Iscove, N. N. and Melchers, F. (1978) Complete replacement of serum by albumin, transferrin, and soybean lipid in cultures of lipopolysaccharide-reactive b-lymphocytes. 1. Exp. Med. 147,923-933. Jakoby, W. B. and Pastan, I. H., eds. (1979) Cell culture. Methods Enzymol. 58. Moore, G. E., Gerner, R. E., and Franklin, H. A. (1961) Culture of normal human leukocytes. JAMA 199,519-524. Ham, R. G. and McKechan, W. L. (1979) Media and growth requirements. Methods Enzymol. 58,44-93. McCoy, T. A., Maxwell, M., and Kruse, P. F. (1959) Amino acid requirements of the Novikoff hepatoma in vitro. Proc. Sot. Exp. Biol. Med. 100,115-118. Morgan, J. F., Campbell, M. E., and Morton, H. J. (1955) The nutrition of animal tissues cultivated invifro. I. A survey of natural materials as supplements to synthetic medium 199. J. Nafl. Cancer Inst. 16,557-567. Waymouth, C. (1959) Rapd proliferation of sublines of NCTC clone 929 (Strain L) mouse cells in a simple chemically defined medium (MB 752/l). 1, Nafl. Cancer Inst. 22,1003-1015. Perlman, D. (1979) Use of antibiotics in cell culture media. Methods Enzymol. 58, 110-119. Grace, T. D. C. (1962) Establishment of four strains ot cells from insect tissue grown in vitro. Nature 195,788,789. Gardiner, G. R. and Stockdale, H. (1975) Two tissue culture media for the production of Lepidoptcran cells and nuclear polyhedrosis viruses. 1. Invert. Path. 25, 363-370.