JOURNAL OF PATHOLOGY, VOL.

179: 347-352 (1996)

DEVELOPMENT OF AN OPTIMAL PROTOCOL FOR

ANTIGEN RETRIEVAL: A 'TEST BATTERY' APPROACH
EXEMPLIFIED WITH REFERENCE TO THE STAINING
OF RETINOBLASTOMA PROTEIN (pRB) IN
FORMALIN-FIXED PARAFFIN SECTIONS
SHAN-RONG SHI*, RICHARD J. COTE*, CHRISTINA YANG*, CHEN CHEN*, HONG-JI xu?, WILLIAM F. BENEDICT? AND
CLIVE R. TAYLOR*

*Department of' Pathology, University of Southern California School of' Medicine, Los Angeles, California, U S.A.; +-Centrefor
Biotechnology, Baylor College of Medicine, Woodlands, Texas, U.S. A.

SUMMARY
The retinoblastoma (RB) gene, which encodes the nuclear RB protein (pRB), is believed to be involved in cell cycle control and cell
differentiation. Studies have demonstrated that loss of R B function may play a role in tumour formation and progression of a variety of
human turnours, such as bladder, lung, breast, and prostate cancers. The immunohistochemical detection of pRB expression in
formalin-paraffin sections of human cancer has potential advantages of convenience, economy, and compatibility with routine surgical
pathology practice. In practice, however, results using pRB antibodies on routinely processed, paraffin-embedded tissue have been
inconsistent. In this study, the antigen retrieval (AR) method has been applied to the immunohistochemical detection of pRB in
paraffin-embedded tissues and a 'test battery' approach has been developed to identify the principal variables that result in the optimal
AR protocol. This approach includes the use of buffered solutions at p H 1,6, and 10 with three different heating conditions (temperatures
120"C, 100"C,and 90°C). In the example described here with antibody RB-WL-1, the low pH solution with the microwave heating a t
100°C proved most effective. Both fresh and routinely processed formalin-paraffin tissues of normal and bladder carcinoma were used
for a comparison of the pRB immunostaining. The AR method was evaluated by comparing the immunohistochemical staining result on
routinely processed formalin-paraffin sections with frozen sections of the same tumour. A consistent intensity of immunohistochemical
staining for pRB was achieved using the identified optimal AR protocol on formalin-paraffin sections. All slides showed positive staining
of pRB in normal mesenchymal and epithelial tissues. The pattern of pRB localization and intensity of staining was similar to that
obtained in frozen sections, though the intensity obtained by AR treatment on paraffin sections was slightly to moderately stronger than
that obtained in frozen sections. Once the protocol was identified, it was tested using routinely processed paraffin tissue sections of 245
cases of bladder carcinoma, with consistent pRB immunostaining results. The protocol described is simple to perform and gives
reproducible results for evaluation of pRB expression by immunohistochemistry.
KEY worms-antigen retrieval; immunohistochemistry; R B microwave oven; paraffin section

INTRODUCTION fresh tissue samples, and by the particular difficulty of

The possible role of various tumour suppressor genes
in regulation of the cell cycle has been the subject of
many recent studies in both basic and clinical
The retinoblastoma (RB) gene encodes a
110 kD nuclear protein (pRB) that is expressed by a
variety of normal tissues. It is believed to be involved in
cell cycle regulation, through a phosphorylation step
that is controlled by several factors, including cyclin
D-CDK 4, 6 complexes, p53, p21, p16, and TGFp.3-10
Following the discovery of functionally significant muta-
tions of the RB gene in hereditary and sporadic retino-
blastoma,' 1 , 1 2 aberrant RB expression has been found in
many other tumour t y p e ~ . ~ -However,
~ , ~ J ~ the detection
of RB gene using the techniques of molecular biology
has been limited to a degree by a general requirement for
Addressee for correspondence: Clive R. Taylor, MD, PhD, Depart-
ment of Pathology, University of Southern California School of
Medicine, 2011 Zonal Avenue, HMR 204, Los Angeles, CA 90033,
U.S.A.
CCC 0022-34 17/96/070347-06
0 1996 by John Wiley & Sons, Ltd.
separating the abundant RB-positive normal tissue
components from tumour cells.
Immunohistochemical demonstration of the protein
product of the RB gene offers a more straight-
forward method for detecting RB expression in tissues,
principally because the pattern of expression of pRB
can be directly observed in normal and abnormal
cells.5 Unfortunately, in practice, application of the
immunohistochemical method to routinely processed
paraffin sections has proven difficult and results have
been inconsistent. Recently, the high-temperature heat-
ing antigen retrieval (AR) technique, developed by Shi
et al. in 1991,13 has been employed for the demonstra-
tion of epitopes that are otherwise difficult to detect
following fixation in formalin and paraffin-embedding.
The AR method has been successfully applied to the
detection of pRB in paraffin-embedded tissue by
immunohistochemistry. For example, Geradts et al. l 4
reported a protocol of pRB immunohistochemical
staining on archival tumour tissue sections using the
AR technique.
Received 10 July 1995
Accepted 13 December 1995
348 S.-R. SHT ET A L .
It is now apparent that several different factors affect
the efficiency of the AR process, including temperature,
time, pH, and the chemical nature of the buffer solu-
tion.I5 It is, therefore, of paramount importance to
select the optimal retrieval conditions for each antibody/
epitope system. With these factors in mind, we have
designed a 'test battery' that includes Tris-HC1 buffer
solution at three different values of pH, in combination
with three different high temperature heating conditions
of 120"C, 100°C, and 90°C. An optimal protocol for AR
immunostaining of pRB with antibody RB-WL-1 was
thereby determined, resulting in standardization and
enhanced reliability of the method. The 'test battery'
approach allows rapid assessment of the major variables
in the AR method and is applicable to any new antibody
under evaluation.
MATERIALS AND METHODS
Tissues
Routinely processed formalin-fixed, paraffin-
embedded tissue sections of human normal spleen, small
cell lung carcinoma, and bladder carcinoma were
obtained from files at the Norris Cancer Hospital and
Research Institute and the Baylor College of Medicine.
Paraffin sections were cut at 5 pm, mounted on poly-L-
lysine-coated slides or commercially available positively
charged slides (Fisher Scientific, Pittsburgh, PA,
U.S.A.), and dried at 60°C in an oven for 1 h before
use. Fresh (frozen) tissues from three cases of bladder
carcinoma were used for a comparative study with
routine paraffin sections from the same cases. The fresh
tissues were immediately embedded in OCT compound
(Miles Laboratories, Elkhart, IN, U.S.A.) and stored at
- 70°C. Frozen sections (5-7 pm) were cut with a cryo-
stat and mounted on poly-L-lysine-coated slides or com-
mercially provided charged slides. Before staining,
frozen sections were dried at room temperature for
10 min and then fixed in 10 per cent neutral buffered
formalin or cold acetone for 10 min, followed by a
PBS (phosphate and saline buffer) rinse and then the
immunostaining procedure.
In addition, RB-positive and RB-negative cell lines
obtained from the ATCC (Rockville, MD, U.S.A.) were
used as controls and were processed as paraffin-
embedded cell blocks of formalin-fixed cell pellets,
employing the formalin fixation procedures in routine
use at our institution.
Antibodies
The affinity-purified rabbit polyclonal antibody
RB-WL-1 (314 pl/ml), generated as previously docu-
mented,16 was used for this study. Monoclonal anti-
bodies to RB G3-245 were provided by PharMingen
(San Diego, CA, U.S.A.) and BioGenex (San Ramon,
CA, U.S.A.), and were used as a comparison with the
polyclonal antibody RB-WL-1. The ABC detection sys-
tem (Vector, Burlingame, CA, U.S.A.) was used
through-
out the study with an appropriate biotinylated 'link'
anti body.
Table I-'Test battery' used to develop optimal protocols for
PRB
-
~~
Tris buffer
pH1 pH6 pH10
Autoclave 120"C, 10 min NT* - +++
MW 10O"C, 10 rnin +++ -
--t
MW 9 0 T , 10 min + -
~~ ~
The procedure of microwave (MW) heating at 100°Cwas performed
as previously described13; the level of the AR solution should be
checked every 5 min.
*NT, low pH solution was not available for autoclave, as
tissue was completely detached from the slide.
?A prolonged heating time of 60 min for MW heating at
100°Cwith the use of Tris buffer at pH 10 could also yield high
staining intensity (+ + +) (see Table 11).
Results are shown for Tris-HC1: 0.01 M acetate buffer gave a
slightly superior intensity, especially at low pH.
The microwave heating AR technique was carried out
as previously d0c~mented.I~ In order to determine
the optimal protocol for pRB immunohistochemical
staining, a 'test battery' was applied.
'Test battery'
The basic principle of the 'test battery' approach is to
screen for the major variables involved in AR, i.e., the
heating conditions, temperature (7') and heating time ( t ) ,
the pH of the AR solution, and the buffer. Tris-HC1 was
the first choice buffer, based on its generally superior
performance over other buffers for a wide range of
antibodies as demonstrated by our previous studies.
Three pH values (pH 1, 6, and lo) of Tris-HC1 buffer
solution and three temperatures of super high 120°C
(autoclave), high 100°C (microwave, MW), and 90°C
(MW or water bath) were employed (Table I). An H2550
MW processor (Energy Beam Science, Agawam, MA,
U.S.A.) that is capable of monitoring both tempera-
ture and time was used. After the first screening test,
additional tests were performed using different buffer
solutions in order to improve the result. Once the
optimal heating conditions and pH were established,
other buffers were evaluated, including acetate, citrate,
and phosphate, to optimize the specific buffer type. We
found that the best results for pRB were achieved using
0.01 M acetate buffer solution at pH 1-2.
According to our observations of the relationship
between temperature and heating time,17 we postulated
that the high temperature (120°C) generated by auto-
clave might be equalled in effectiveness by MW heating
for a longer period of time. Therefore, when an optimal
heating condition was indicated by autoclave at 120°C
for 10 min, further tests of MW heating with extended
heating times were carried out as indicated in Table 11.
Optimal protocol of antigen retrieval
An optimal protocol for AR of pRB was developed
using the 'test battery' as discussed above. Briefly, the
 DEVELOPMENT OF AN OPTIMAL PROTOCOL FOR ANTIGEN RETRIEVAL
Table 11-Comparison of the intensity of pRB-AR-IHC stain-
ing using Tris-HC1 buffer at pH 10 with different MW heating
times
MW heating time
5 min x 2 5 min x 4 5 min x 12
Lung SCC - + +++
Bladder Ca - + +++
SCC=small cell carcinoma. The cancer cells were negative for pRB;
the score was graded for mesenchymal cells. Ca=carcinoma. -,
negative; +, weak positive; + +, moderate positive; and + + +, strong
positive results.
complete optimal protocol was as follows: (1) Tissue
sections were deparaffinized and rehydrated. (2) En-
dogenous peroxidase was blocked with H,O,-methanol
solution for 20 min. (3) Slides were washed with distilled
water for 5 min. (4) Slides were then placed in plastic
Coplin jars containing low pH buffer solution (0.01 M
acetate buffer at pH 1-2 showed the best result as
mentioned above) and were covered with loose-fitting
screw caps. ( 5 ) MW oven heating at 100°C for 10 rnin
was divided into two cycles as previously reported.I3 (6)
After heating, the Coplin jars were removed from the
oven and allowed to cool for 15 min. (7) Slides were then
rinsed twice in distilled water and in PBS for 5 min. (8)
The immunohistochemical staining procedure was
then performed. In total, sections from more than 300
paraffin blocks were tested with this AR immuno-
staining procedure.
Immunohistochemistry
Normal goat or horse serum diluted 1:20 was used for
a 20 min incubation in order to block the non-specific
binding of primary antibodies. Primary antibodies were
used at dilutions of 0-6,ug/ml and 5 pg/ml for polyclonal
and monoclonal antibodies to pRB, respectively.
Sections were incubated overnight at room tempera-
ture, followed by the ABC method as previously
documented. l8 AEC (3-amino-9-ethyl-carbazole) was
utilized as the chromogen. pRB-positive and pRB-
negative cell lines were used as tissue controls. Negative
control sections employing normal rabbit, mouse serum,
or PBS were tested with every batch after the AR
treatment.
Evaluation of immunohistochemical staining results
Only distinct nuclear immunostaining was interpreted
as a positive staining result for pRB. The intensity of
positive immunostaining was graded semi-quantitatively
as + + + , + +, +, or - , signifying decreasing intensity
of nuclear positivity (Fig. 1, R-U).
RESULTS
Both the polyclonal and the monoclonal antibodies to
pRB showed similar results. Data are presented for the
349
polyclonal antibody RB-WL-1. The strongest intensities
of pRB immunostaining were achieved under three
different conditions of AR: (I) the use of AR solution at
pH 1 with MW heating of 100°C for 10 rnin (Table I),
particularly the AR solution of acetate buffer (Fig. 1, A
vs. B and C); (2) the use of Tris-HCl buffer of pH 10 by
autoclave heating at 120°C for 10 min. (3) Further study
showed that the use of Tris-HC1 buffer of pH 10 by MW
heating at 100°C for 1 h gave equivalent results. The
intensity of pRB immunostaining after MW-AR at
100°C with Tris-HC1 buffer solution at pH 10 gradually
increased as the heating time was extended (Table I1 and
Fig. 1, V-X).
We selected low pH acetate buffer solution with 100°C
MW heating as the optimal protocol of choice and
tested a variety of normal tissues and bladder carci-
nomas. We observed that pRB-positive cells were widely
distributed in normal connective tissue, including
endothelial cells, smooth muscle cells, lymphocytes, and
also epithelial cells (Fig. 1, I-M). The patterns of
localization of pRB in the tissues correlated with those
obtained in frozen sections; for example, in stratified
squamous epithelium, pRB-positive nuclear staining
could be found only in the suprabasal layers, but not in
the basal layer (Fig. 1, I, J). In the case of the bladder
tumours, the paraffin-embedded tissue produced similar
results to the frozen tissue, though the staining intensity
obtained in paraffin sections after AR treatment was
usually stronger than that obtained in frozen sections
(Table 111 and Fig. 1, N-Q). In some cases, there was
complete loss of pRB immunoreactivity in tumour cells,
while immunoreactivity could be seen in surrounding
normal tissue (Fig. 1, N, 0, U).
The AR immunostaining results of pRB in 245 cases
of bladder carcinoma were reviewed; all displayed pRB
positivity in normal tissue components throughout the
entire section. In the case of the malignant cells, 31.4 per
cent (77 cases) of the tumours showed loss of pRB
positivity, while 68.6 per cent (168 cases) retained pRB
expression. In those cases retaining pRB expression,
there was significant inter-tumour heterogeneity, with
some cases showing very intense staining while other
cases showed weak to moderate staining (Fig. 1, R-T).
All negative control slides, including both tissue
negative controls and normal serum or PBS negative
control sections, gave no evidence of pRB immuno-
staining after the AR protocol (Fig. 1, D). The results
of AR-IHC staining by polyclonal and monoclonal
antibodies to pRB were similar (Fig. 1, E-H).
DISCUSSION
While high-temperature heating is the most important
factor for the AR technique,13~15~17~19~20 the pH value of
the AR solution is also a critical factor in influencing the
quality of the AR immunostaining result. 15,,* Based
on these two important factors, we designed a ‘test
battery’ for the development of an optimal AR
protocol when testing a new antibody for immunostain-
ing on archival paraffin sections. Using this approach,
we identified three sets of conditions that were
Fig. 1-Antigen retrieval immunohistochemical staining results on routeinely formalin-fixed, paraffin-embedded bladder carcinoma tissue
sections, showing nuclear staining, A-D show the results of a ‘Test battery’ used to select an optimal protocol for pRB, pH 1 showing the
strongest intensity (A), with p H 6 (B) and p H 10 (C) showing much weaker intensity. All were exposed t o MW heating for 10 min. D shows the
negative control after the AR treatment with p H 1 omitting the primary antibody. E-H show similar patterns of immunolocalization obtained
by monoclonal (E, G) and polyclonal antibodies to pRB (F, H). I-M show the immunolocalization of pRB in normal tissue components using
the low pH AR treatment prior to immunostaining; pRB was localized in the suprabasal layer of stratified squamous epithelium as indicated ( I ,
J), in smooth muscle (K), in the vessel wall (L), and in granulomatous tissue (M). N-Q compare pRB localization in frozen ( N , P) and parafin
sections (0,Q), showing similar patterns for the two kinds of sections. Cancer cells are negative in N and 0, and positive in P and Q (different
cases). In N and 0, infiltrating normal inflammatory cells are pRB-positive. R-U demonstrate the grading scale used for the intensity of
pRB-AR-IHC staining: + (R), + + (S), + + + (T). U shows a case in which cancer cells were negative and infiltrating lymphocytes were positive.
V-Y illustrate results obtained following autoclaving at p H 10 (cf. C with Y ) .Extending the MW heating time from 10 min (V, 5 x 2, negative)
to 20 min (W, 5 x 4. positive + ) o r 1 h (X, 5 x 12, strong positive + + +) with the use of Tris-HCl buffer at pH 10 also increases the intensity
of staining. Magnification: x 100 (A-F. I , J, U); x 200 (others). AEC was used as the chromogen and the counterstain was haematoxylin
 DEVELOPMENT OF AN OPTIMAL PROTOCOL FOR ANTIGEN RETRIEVAL
Table 111-Comparison of the immunostaining of pRB between frozen and paraffin sections with AR
treatment
Frozen
Case No. Epi M us Ves Cancer
1 ++ ++ + ++
2 + ++ + -
3 + + ++ ++
All three cases are transitional cell carcinoma of the bladder. Frozen sections were fixed with 10 per cent formalin and
acetone, respectively, both fixatives showing similar results. Epi=epitheliium; Mus=smooth muscle; Ves=vessel wall; all
are the normal components in the section.
approximately equivalent for the polyclonal antibody
pRB. We chose the low pH of acetate buffer as the
optimal protocol for this study, based primarily on
convenience.
Our recent study concerning AR immunohisto-
chemistry under the influence of pH demonstrated that
certain low pH (pH 1-2) AR solutions yield the best
result of AR immunostaining for several nuclear anti-
gens, such as ER, PR, and MIBl.I5 The strongest
intensity of pRB immunostaining was also obtained
using a low pH buffer solution as indicated above. In
addition, there was no evidence of false-positive nuclear
staining when performing pRB immunostaining on
archival paraffin sections using this protocol. This
optimal protocol with acetate buffer at p€I 1-2 yielded
reliable positive nuclear staining that was validated by
careful comparison of frozen and paraffin sections from
the same specimen, showing similar pRB immunolo-
calization in the same histological structures (Fig. 1,
N-Q). In addition, the patterns of pRB immunolo-
calization in normal tissues correlate well with the
results reported p r e v i o ~ s l y . For
~~,~ example,
~ Cordon-
Cardo and Richon22 reported an extensive immuno-
histochemical study of pRB on frozen sections of
normal tissues, noting that in normal squamous epi-
thelium, pRB could be identified in the suprabasal
layer, but not in the basal layer. Our results of pRB
immunolocalization with the use of AR treatment in low
pH buffer solution yielded an identical pattern (Fig. 1,
I, J). All negative control slides after AR treatment
showed absence of reactivity (Fig. 1, D).
The pattern of pRB immunostaining on 245 cases of
bladder carcinoma using AR treatment correlates well
with the published l i t e r a t ~ r e . ~From
. ’ ~ these data, the
optimal protocol identified for AR immunostaining of
pRB appears to provide a reliable method for achieving
consistent immunohistochemical staining.
This study exemplifies the use of the ‘test battery’
approach in identifying the optimal protocol for AR
immunohistochemistry, applicable to any new antibody
under evaluation. In the course of this work, it has
become clear that the use of a ‘test battery’ approach
may also contribute to the standardization of AR
immunostaining, by virtue of rapid identification of
that protocol which gives ‘maximal r e t r i e ~ a l ’ . ~ ~ . ~ ~
A number of recent studies regarding AR immuno-
histochemistry of ER, p53, PCNA, and MIBl have
351
Paraffin +AR treatment
Epi Mus Ves Cancer
++ +++ ++ +++
++ ++ + -
++ ++ ++ +++
examined the question of the ‘best’ AR protoc01.l~One
major factor underlying this question is the significantly
increased sensitivity generated by AR treatment, an
increase that can convert a low intensity stain to high
intensity, or can change a score of 57 per cent positive
cells to 93 per cent for p53 immunolocalization, or
can even convert a negative result (without AR) to a
positive result (post AR). 19,26 Some investigator^^^.^^
have found that the use of AR on paraffin tissues
actually produced a stronger immunostaining intensity
than the use of the same antibody on frozen tissue, just
as we found for pRB immunolocalization described
above (Table 111 and Fig. 1, N, P vs. 0, Q). One
explanation for the stronger intensity of some antibodies
on paraffin sections after AR treatment, as compared
with frozen sections, may be that some receptor mol-
ecules or epitopes are lost in the frozen sections by
decay, diffusion, or as a result of acetone/alcohol fixa-
tion, whereas in formalinized tissues all epitopes are
‘fixed’ in place and may be recovered to a great degree
by an optimal AR protocol.
The evidence that the degree of positivity is altered by
the differences in the AR method demands standardiz-
ation of the AR approach, if data are to be comparable
from case to case and from institution to institution. By
using an optimal protocol of AR, it can be shown
that archival tissue sections may be standardized at a
‘maximal retrieval’ level, regardless of the fixation
time or other institutional variables. The concept of
‘maximal retrieval’ may contribute to the standardiz-
ation of AR immunohistochemistry. The goal for
each institution would then be to identify that protocol
which gives ‘maximal retrieval’ for each antibody on
tissue processed at that institution. The ‘test battery’
approach facilitates this process; it also reveals that two
or three protocols will suffice for the great majority of
antibodies, but that no single protocol is optimal for all
antibodies.
The term ‘maximal retrieval’ is put forward to
describe the condition in which certain epitopes
previously ‘masked’ by formalin fixation have been
retrieved to the fullest extent possible by an optimal AR
protocol. The ‘test battery’ approach allows rapid evalu-
ation of variables in identifying the optimal protocol for
‘maximal retrieval’. This protocol then takes its place as
the laboratory standard for all antibodies shown to
benefit maximally under the protocol in question.
352 S.-R. SHI ET A L
A final issue raised by this study is the equally strong
intensity of staining achieved by three different protocols
of AR, as indicated in Table I. To some extent, this
result seems to be a function of the total amount of
energy delivered during the heating process, with equiva-
lency occurring at very high temperature for short times,
or lower temperatures for longer times, within the limits
described (90-120°C, 1 hr to 10 min). We have extended
this observation by demonstrating that under optimal
conditions there are no differences in AR-IHC staining
with the use of different heating methods such as MW,
conventional heating, or autoclave, provided that the
temperature and time are adjusted appropriately. This
observation of 'equivalence' in terms of the effectiveness
of different heating methods provides individual labora-
tories with great flexibility, provided that the optimal
protocol is always identified, using 'maximal retrieval' as
the end-point for standardization.
ACKNOWLEDGEMENTS
We thank Drs Nancy Warner and Peter Nichols for
help in taking microphotographs; Ms Lillian Young,
Moung Win, and Elizabeth Yee for technical assistance;
and Ms Linda Montellano for preparation of this
manuscript. This research was supported by the Univer-
sity Pathology Associates of the University of Southern
California and by the National Cancer Institute
(CA54672).
REFERENCES
1. Knudson AG. Antioncogenes and human cancer. Pro(,Nut/ Acud Sci USA
1993; 9 0 10914-10921.
2. Weiss LM. Retinoblastoma protein: another target for immunohisto-
chemists. Appl Immunohistochem 1994; 2 69-70.
3. Takahashi R, Hashimoto T, Xu H-J, er ul. The retinoblastoma gene
functions as a growth and tumour suppressor in human bladder carcinoma
cells. Proc Nurl Acad Sci USA 1991; 88: 5257-5261.
4. Kornblau SM, Xu H-J, Giglio A, er ul. Clinical implications of altered
retinoblastoma protein expression in acute myelogenous leukemia. C u n w
Re.$ 1992; 52: 45874590.
5. X u H-J. Hu S-X, Cagle PT, Moore GE, Benedict WF. Absence of
retinoblastoma protein expression in primary non-small cell lung carcino-
mas. Cunwr R1.s 1991; 51: 2735-2739.
6. Goodrich DW, Wang NP, Qian Y, Lee EY-HP, Lee W-H. The retino-
blastoma gene product regulates progression through the GI phase of the
cell cycle. CeN 1991; 67: 293-302.
7. Logothetis CJ, Xu H-J, Ro JY, ef ul. Altered expressoin of retinoblastoma
protein and known prognostic variables in locally advanced bladder cancer.
J , Nurl Crrncer lnsr 1992; 8 4 1256-1261.
8. Hunter T, Pines J. Cyclins and cancer 11: cyclin D and CDK inhibitors come
of age. Cell 1994;.79 573-582.
9. Trudel M, Mulligan L, Cavenee W, Margolese R, Cote J, Gariepy G.
Retinoblastoma and p53 gene product expression in breast carcinoma:
immunohistochemical analysis and clinicopathologic correlation. Hum
Pcrihol1992; 2 3 1388-1394.
10. Cordon-Cardo C, Wartinger D, Petrylak D, "i ul. Altered expression of thc
retinoblastoma gene product: prognostic indicator in bladder cancer. J Nut/
Cimcer Insr 1992; 8 4 1251-1256.
11. Fung Y-K, Murphee AL, T'Ang A, e t a/. Structui-al evidence for
the authenticity of the human retinoblastoma gene. .S<,itnr.e 1987; 236:
1657- 1661.
12. Cavenee WK, Hansen MF, Nordenskjold M, cr a/. Genetic origin of
mutations predisposing to rctinoblastoma.
13. Shi S-R, Key ME, Kalra KL. Antigen retrieval in formalin-fixed, parafin-
embedded tissues: an enhancement method for immunohistochemical stain-
ing based on microwave oven heating of tissue sections. J Hi.sro(,hem
Cytochem 1991; 3 9 141-748.
14. Geradts J, Hu S-X, Lincoln CE. Benedict WE, Xu H-J. Aberrant RB gene
expression in routinely processed, archival rumour tissue determined by
three ditferent anti-RB antibodies. h t J Cunwr 1994; 58: 161-167.
15. Sbi S-R, Imam SA, Young L, Cote RJ, Taylor CR. Antigen retrieval
immunohistochemistry under the influence of pH using monoclonal anti-
bodies. J Histochem Cyrochrm 1995; 4 3 193-201
16. Xu HJ, Hu SX, Hashimoto T. Takahashi R, Benedict WF. The retino-
blastoma susceptibility gene product: a charackristic pattern i n normal cells
and abnormal expression in malignant cells. Onc.ugcnc 1989; 4 807-X 12.
17. Shi S-R, Gu J, Kalra KL, Chen T, Cote RJ, Taylor CR. Antigen retrieval
technique: a novel approach to immunohistochemistry on routinely
processed tissue sections. Cell V%-'ision 1995; 2 6-22.
18. Hsu SM, Raine L, Fanger H. Use of avidin-biotin-peroxidase complex
(ABC) in immunoperoxidase techniques: a comparison between ABC and
unlabeled antibody (PAP) procedurcs. J Hi.sroc./ierfiC'1,rochem 1981; 2 9
577-580.
19. Lueassen PJ, Ravid R, Gonatas NK. Swaab DF Activation of the human
supraoptie and paraventricular nucleus neurons with aging and in
Alzheimer's disease as judged from increasing size of the Golgi apparatus.
Bruin Re.7 1993; 632: 105-113.
20. Igarashi H, Sugimura H, Maruyama K, er ul. Alteration of immuno-
reactivity by hydrated autoclaving, microwave trcalmcnt, and simple heat-
ing of paraffin-embedded tissue sections. A P M I S 1994; 102: 295-307.
21. Evers P, Uylings HBM. Microwave-stimulated antigen retrieval is pH and
temperature dependent. J Histoc.henf Cytochefn 1994; 4 2 1555 1563.
22. Cordon-Cardo C, Richon VM. Expression of the retinoblastoma protein is
regulated in normal human tissues. Am J Puthal 1994: 144 500-510.
23. Barbareschi M, Girlando S, Mauri FA, e t ul. Tumour suppressor gene
products, proliferation, and differentiation markers in lung neuroendocrinc
neoplasms. J Purhol 1992; 166: 343~~350.
24. Taylor CR. An exaltation of experts: conccrred efrorts in the standardi7-
ation of immunohistochemistry. Hunr Purhol 1994, 25: 2 11
25. Taylor CR, Shi S-R, Chaiwun B, Young L, Imam SA, Cote RJ.
Standardization and reproducibility in diagnostic immunohistochemislry
Hum Ptzriiof 1994; 2 5 I107 1109.
26. Tenaud C, Negoescu A, Labat-Moleur F. Lcgros Y, Soussi T. Brambilla E.
p53 immunolabeling in archival paratfin-enibcddcd tissues: optimal
protocol based on microwave heating foi- eight aiitibodio on lung carci-
nomas. Mod I'uthol 1994; 7: 853 859.
27. Battifora H. lmmunocytochemistry of hormone receptors in routinely
processed tissues. Appl Immunohisfochmn 1994; 2: 143.- 145.