Histochemical Journal 18, 565-575 (1986)

Differential expression of keratin 19 in normal human

epithelial tissues revealed by monospecific monoclonal
antibodies
J. BARTEK 1, J. BARTKOV,~ 1, J. T A Y L O R - P A P A D I M I T R I O U 2, A. REJTHAR 1,
J KOVAI~IK ~, Z. LUKA~ 3 and B. VOJTI~EK ~
1Research Institute of Clinical and Experimental Oncology, ~'.lut~ Kopec 7, 602 O0 Brno, Czechoslovakia
2Imperial Cancer Research Fund, PO Box 123, Lincoln's Inn Fields, London W C 2 A 3PX, UK
32nd Department of Pathology, Medical Faculty of the Purkyne University Children's Hospital, 662 63 Brno, Czechoslovakia

Received 9 April 1986 and in revised form 5 July 1986

Summary
Three monospecific monoclonal antibodies (BA16, BA17 and A53-B/A2) recognizing different epitopes of the human keratin 19
were used to determine tissue distribution of this 40 kDa keratin polypeptide. Immunohistochemical methods revealed four
different staining patterns among normal human epithelial tissues: firstly, complete negativity of the epidermis, sebaceous glands,
hepatocytes and other tissues; secondly, homogeneous positivity as seen for example in the gall bladder and urinary bladder
epithelium, endometrium and many other epithelia; thirdly, a mosaic of positive and negative cells among mammary gland
luminal cells, prostate epithelia and some other epithelia and fourthly, a more complex heterogeneous pattern found in non-
keratinizing squamous epithelia and hair follicles with generally the basal layer being the most strongly or sometimes exclusively
stained. The pattem seen in non-keratinizing squamous epithelia varied considerably according to the fixation method and the
antibody used as well as among different donors and in different areas of the same organ. The other three staining patterns were
on the other hand nearly identical with all three antibodies on both frozen sections and sections of methacarn-fixed paraffin-
embedded tissues. Our results provide evidence for differential expression of the human keratin 19 at the single cell level, an
observation which could be exploited in the study of epithelial differentiation and pathology.

Introduction
Intermediate filament distribution has been shown to
follow histogenetic principles, and thus determination of
the type of intermediate filament component(s) present in a
cell provides information relevant to the origin of that cell
and is therefore of importance in both cell biology and
routine pathology (Lazarides, 1980; Osbom & Weber,
1982). Among the five known classes of intermediate
filaments, the cytokeratins represent the most complicated
family of proteins (19 in the human). These have been
catalogued by Moll et al. (1982a) according to molecular
weight and isoelectric point and their tissue distribution
determined by biochemical analysis (reviewed by Quinlan
et al., 1985). Different epithelial cell types differ in their
keratin spectra and tumours generally tend to preserve the
keratins expressed by the cell of origin which offers a new
powerful tool for subdivision of epithelia and carcinomas
(Moll et al., 1982a; Cooper et al., 1985; Quinlan et al.,
1985). Although extremely important, the biochemical
techniques of keratin estimation have some limitations
especially concerning tissue topography and the ex-
pression of the particular keratin polypeptide at the single
cell level. An alternative approach to the single-keratin
identification is provided by immunohistochemical tech-
niques using hybridoma-derived monoclonal antibodies
specific for a particular keratin polypeptide. Although
limited, a number of monoclonal antibodies specific for
particular keratin polypeptides have been described. Many
of these are directed to the human 45 kDa keratin
(keratin 18) (Lane, 1982; Debus et al., 1982; Ramaekers et
al., 1983), but antibodies are also available which
specifically react with keratins 4 and 13 (Van Huijen et al.,
1986), keratin 8 (Lane et al., 1985) and with keratin 19
(Bartek et al., 1985a; Karsten et al., 1985; Lane et al., 1985).
The list of monospecific antibodies is expanding (see
Cooper et al., 1985 for review).
We have reported the preparation of mouse hybridomas
producing two monoclonal antibodies BA16 and BA17
specific for the human 40 kDa keratin (Bartek et al., I985a)
which has been classified as keratin 19 (Moll et al., 1982a).
Recently Karsten et al. (1985) prepared a mouse monoclon-
al antibody of apparently similar specificity, although
direct proof (e.g. immunoprecipitation or immunoblotting

0018-2214/86 $03.00+.12 © 1986 Chapman and Hall Ltd.

566
data) was lacking. Comparison of the three monoclonal
antibodies in various assays including immunoblotting
confirmed their specificity for the 40 kDa keratin and
suggested that each of them recognizes a different epitope
of the polypeptide (Karsten & Bartek, unpublished results,
1986). Using antibodies BA16 and BA17 we have
demonstrated the usefulness of the immunohistochemical
visualization of keratin 19 at the single cell level in the
identification of so far unrecognized subpopulations
among human breast epithelial cells (Bartek et al., 1985a, b)
as well as in the subdivision of human tumours according
to the pattern of reactivity with our antibodies (Bartek et
al., 1985b, 1986).
The observation of three different staining patterns in
human tumours in situ as well as in cultured cells with
antibodies specific for keratin 19 (Bartek et al., 1986) led us
to examine the reactivity of a wide range of normal tissues
with antibodies BA16, BA17 and A53B/A2. In the present
study we report the immunohistochemical localization of
the human 40 kDa cytokeratin (keratin 19) in normal
human tissues as detected with these three monospecific
monoclonal antibodies.
Materials and methods
Tissues
Samples of morphologically normal tissues distant from tumours
were obtained during operations carried out at the Departments
of Gynaecology and Surgery of the Research Institute of Clinical
and Experimental Oncology in Bmo. Some normal tissues were
obtained from autopsy material taken from peopl e killed in road
accidents (supplied by the Department of Forensic Medicine, J. E.
Purkyne University Medical School, Bmo) and from children
who died of congenital heart diseases and were autopsied at the
2nd Department of Pathology, J. E. Purkyne University
Childrens Hospital, Bmo. Small pieces of tissue were either
soaked in Tissue-Tek II OCT Compound (Miles Laboratories,
Wien, Austria) for 30 min and frozen on solid carbon dioxide or
fixed in methacam (a mixture of methanol, chloroform and acetic
acid, 6: 3:1 by vol) and embedded in paraffin. Samples of several
frozen tissues came from the tissue bank of the Department of
Epithelial Cell Growth Regulation, Imperial Cancer Research
Fund, London. Frozen tissue blocks were kept either in liquid
nitrogen or at -- 70° C before sectioning.
Monoclonal antibodies
The preparation and basic characterization of two mouse
monoclonal antibodies BA16 and BA17 (both IgG 1 with a kappa
light chain) reacting with the human 40 kDa cytokeratin
(keratin 19) was described previously (Bartek et al., 1985a).
Another monoclonal antibody A53-B/A2 (IgG2a) specific for
human keratin 19 was kindly provided by U. Karsten (Central
Institute of Molecular Biology, Berlin, GDR). Antibody TF-1
(IgG1) against pig transferrin (Bartek et aL, 1982) was used as a
negative control. In all immunohistochemical methods undiluted
hybridoma tissue culture supematants were used.
Immunofluorescence microscopy
Indirect immunofluorescence assay was performed on either
B A R T E K et al.
cryostat sections fixed in a precooled mixture of methanol and
acetone (1:1 v/v) for 10 rain or deparaffinized sections of tissues
that were fixed in methacam and subsequently embedded in
paraffin. The staining procedure used was essentially as
described previously (Bartek et al., I985c) using FITC-conju-
gated rabbit anti-mouse immunoglobulin antiserum (Miles
Laboratories) diluted 1:30 as the second antibody.
Indirect immunoperoxidase staining
The indirect immunoperoxidase technique used on both frozen
and methacam-fixed paraffin-embedded tissue sections (see
above) was carried out as described (Bartek et al., 1985a) with
peroxidase-conjugated rabbit antiserum to mouse immuno-
globulins (Dako, Copenhagen, Denmark, diluted 1:50) as the
second antibody, diaminobenzidine (Sigma, London, UK) as
chromogen and Haematoxylin to counterstain nuclei. In some
experiments, the dewaxed sections were treated with 0.1%
trypsin for 10--20 min at 37 ° C before staining.
Biotin--streptavidin immunohistochemistry
Two immunoperoxidase methods based on the biotin--strepta-
vidin system were used in this study.
1. The streptavidin-biotin bridge technique employing
consecutively biotinylated sheep antiserum against mouse
immunoglobulin, streptavidin bridge and biotinylated
horseradish peroxidase.
2. The most sensitive method using pre-formed strepta-
vidin-biotinylated peroxidase complexes after the bioti-
nylated sheep antiserum to mouse immunoglobulins.
All reagents were supplied by Amersham Intemational (Amer-
sham, UK) and the procedures were performed essentially as
recommended by the manufacturer.
Immunohistochemical staining using fl-galactosidase label
Tissues containing cells with brown pigmentation were exam-
ined with an immunohistochemical technique based on //-
galactosidase-conjugated donkey anti-mouse immunoglohulin
antiserum (Farmitalia Carlo Erba, Milan, Italy). The procedure
was performed as recommended by the manufacturer using beta-
Histogal-M staining kit and nuclei were counterstained with
Carmine Red.
Results
The results of immunohistochemical staining of a variety
of normal human tissues with monoclonal antibodies
BA16, BA17 and A53-B/A2 are summarized in Table 1.
The tissues listed in Table 1 were examined as frozen
sections with all three antibodies in both indirect
immunofluorescence and indirect immunoperoxidase tech-
niques. Samples of most tissues were also fixed in
methacarn, embedded in paraffin and dewaxed sections
stained with the antibodies though the staining intensity
obtained with BA16 is weaker than that of BA17 and
A53-B/A2 on tissue sections processed in this way.
No significant reaction could be observed with any of
the antibodies in nonepithelial tissues. In agreement with
our previous studies on human breast tissues and various
Keratin I9 detection in human tissues
epithelia-derived tumours (Bartek et al., 1985a, b, 1986) we
observed three different staining patterns, i.e. completely
negative, homogeneously positive and heterogeneous
among different human epithelial tissues. Some examples
of the staining patterns are illustrated in Fig. 1 (immuno-
fluorescence) and Figs. 2 and 3 (immunoperoxidase).
As far as epithelial tissues are concerned, it is obvious
from the data presented in Table I that stratified squamous
epithelium of the epidermis (Figs. la, b), sebaceous glands,
liver hepatocytes (Fig. 2c), some testicular cell types and
the cells of some endocrine glands are completely negative
when stained with the antibodies specific for keratin 19.
A homogeneously positive reaction (i.e. no obviously
negative cells) were seen with a large number of epithelial
cell types including many ductal and glandular epithelia.
The full list is shown in Table I and some examples of this
type of staining are shown in Figs. 1 and 2. Special cell
types staining positively with the antibodies include
Merkel cells found in some stratified squamous epithelia
and epithelial cells in the taste buds of the tongue.
In several glandular epithelial tissues the antibodies
stained only some cells whereas the rest of the epithelium
was negative. As can be seen from Table I this hetero-
geneous staining pattern (Type I) was observed among
luminal cells of the mammary gland (Figs. lh;3h, i), prostate
(Fig. 2i), lung alveoli, epididymis (Fig. li), thyroid (Fig. 2h),
kidney, secretory portions of the salivary, oesophageal,
tracheal and bronchial glands, stomach glands and inside
the intestinal crypts. On the other hand, epithelium of the
ductal portions of some of these glands stained homo-
geneously positive (see Table 1). The proportion of
positively stained and unstained cells in the above
mentioned glandular epithelia varied considerably among
different individuals, different areas of the same organ or
even neighbouring structures of the same area. Monoclon-
al antibody BA16 stained frozen sections of these tissues
more strongly than paraffin sections while antibodies
BA17 and A53-B/A2 gave generally better results on
paraffin sections. Despite these differences the overall
heterogeneous pattern was very similar with all three
antibodies. The intensity of staining was usually weaker in
the thyroid, intestinal and some kidney epithelia making
the evaluation of heterogeneity more difficult in these
tissues.
A heterogeneous staining pattern (Type II) different
from that seen in glandular tissues (Type I) was found in
non-keratinizing stratified squamous epithelia represented
by the tongue, oesophagus, exocervix and vaginal fomix.
In these multilayered epithelia, the basal layer was
generally the most strongly stained one (Figs. 3a,b,c,e,f),
although variation of the staining intensity was seen
among the basal cells. Suprabasal layers of these epithelia
could be completely negative (Figs. 3a,b,e), or positively
stained to various degrees (Figs. 3c,f) in some areas even
equal to that of the basal layer (Fig. 3g). The pattern
observed varied remarkably depending on the fixation
method and particular antibody used as well as on the
567
donor and the area of the same organ. Thus antibody
A 5 3 - B / A 2 showed the staining pattern with strongest
positivity in the basal layer and variable intensity of the
suprabasal staining as described above without any
obvious differences between sections of frozen and
methcam-fixed paraffin-embedded tissues. BA16 stained
only some cells of the basal layer sometimes with long
stretches of completely negative epithelium even on
frozen sections. The patterns seen with antibody BA17
were intermediate to those of A53-B/A2 and BAI6 in that
it was similar to A53-B/A2 on paraffin sections while
being usually restricted to basal layer (thus being closer to
BA16) on frozen sections (see Figs. 3a,b,c,e,f,g).
Differences similar to those observed on non-kerati-
nized squamous epithelia of the tongue, oesophagus and
exocervix were also found in the hair follicle epithelium.
Thus antibody A53-B/A2 stained a subpopulation of basal
cells of the outer hair root sheath and corresponding
groups of basal cells were also positive with BAIT on
parallel sections of methacam-fixed paraffin-embedded
tissues. Unlike in squamous non-keratinized epithelia, no
suprabasal positivity was seen with either antibody in hair
follicles. On frozen sections BA17 was consistently
negative on hair follicle epithelia as was BA16 antibody on
sections processed in either way. In order to avoid
interference of melanin pigmentation with the diamino-
benzidine reaction product in the immunoperoxidase
reaction, we confirmed these data using a ]/-galactosidase-
conjugated second antibody in an indirect immunostaining
method resulting in a blue-colour reaction product (data
not shown).
To exclude the possibility that the heterogeneous
staining pattern seen in many tissues is due to low
sensitivity of the indirect immunoperoxidase method used
we employed two biotin-streptavidin-based techniques as
well and compared the staining results on parallel sections.
The mosaic pattern of positive and negative cells was
found again, regardless of t h e staining method used
(Figs. 3h,i). In another series of experiments the results
obtained with untreated sections were compared with
those after pretreatment with 0.1% trypsin for 10, 15 or
20 minutes at 37 ° C. The staining intensity of BA16 was
somewhat increased after proteolytic pretreatment while
the results obtained with antibodies A53-B/A2 and BA17
remained more or less unaffected or the intensity could be
even reduced (with BA17). The original mosaic pattern
remained unchanged with all three antibodies despite the
minor variations of the staining intensity mentioned
above.
Discussion
The present immunohistochemical study was undertaken
to define keratin 19 distribution at the single cell level in a
broad range of normal human tissues. This kind of
monospecific monodonal antibody-based analysis was
stimulated by both available biochemical data on
568 B A R T E K et al.

Table 1. Staining patterns of monoclonal antibodies to keratin 19 in normal human tissues.

Staining pattern

Tissue Negative Homogeneous Heterogeneous

(4-20 samples of each tissue (no positive (no obvious
from different donors) cells) negative cells) Type 11 Type II 2

Interfollicular epidermis +
Sebaceous gland +
Hair follicle +
Sweat gland
secretory cells (acini)3 +
myoepithelium +
Merkell cells4 +
Mammary gland
luminal cells +
myoepithelium5 +
Endometrium
surface mucosa +
glands +
Endocervix
surface mucosa +
glands +
Exocervix, at endocervical junction +
Exocervix, major portion +
Vaginal fomix +
Ovary
surface mesothelium +
corpus luteum +
corpus albicans +
Fallopian tube epithelium +
Placenta
syncitiotrophoblast +
cytotrophoblast 3 +
amnion +
Umbilical cord surface epithelium +
Prostate
luminal cells +
basal cells +
Testes
seminiferous tubules +
Leydig cells +
rete epithelium +
ductuli efferentes +
Epididymal tubules +
Kidney
Bowman's capsule +
proximal tubules +
distal tubules +
collecting tubules +
Urinary bladder epithelium +
Liver
hepatocytes +
bile duct epithelium +
Gall bladder epithelium +
Tongue
squamous epithelium +
taste buds +
glands--secretory cells +
glands-ducts +
Keratin 19 detection in human tissues 569

Table 1---continued

Staining pattern

Tissue Negative Homogeneous Heterogeneous

{4-20 samples of each tissue (no positive (no obvious
from different donors) cells) negative cells) Type 11 Type II 2

Oesophagus
squamous epithelium +
submucosal glands +
Stomach
surface mucosa +
glands -t-
Small bowel
surface mucosa +
crypts -I-
Large bowel
surface mucosa +
crypts +
Pancreas
ducts +
acini +
endocrine +
Salivary gland
secretory portion _i_
ducts +
Thyroid epithelium +
Adrenal gland +
Trachea
surface mucosa +
glands +
Lung
bronchial mucosa +
bronchial glands +
alveoli +
pleura-mesothelium +
Smooth muscle +
Striated muscle +
Myokard +
Cartilage +
Blood vessels +
Brain tissue +
Cerebellum +
Stroma of all organs +
Spleen +
Thymus
Hassal's corpuscles 6 +
epithelial cells4 +
other cells +
Lymph nodes +

1Between 5 and 90% unstained cells, variable throughout the tissues and between individuals.
2The basal layer most strongly stained, variable suprabasal staining. See Results section for details.
3In ducts of sweat gland, basal cells show negative staining with all three antibodies. Luminal cells are negative
with BAI6 and show weak positive staining with BAI7 and A53-B/A2.
4Existence of a minor subpopulation of negative cells cannot be excluded.
5In large ducts weak staining of some basal cells can be seen.
6Staining intensity was reduced towards the centre in some Hassal's corpuscles due to hyalinization.
570
keratin 19 presence in some tissues (see Moll el al., 1982a;
Quinlan et al., 1985 for review) and our previous work
concerning the heterogeneity of expression of this 40 kDa
keratin in human breast tissues, tumours of various
histogenesis as well as cultured cells (Bartek et al., 1985a,b,
1986). Similar cell type heterogeneity of cytokeratin
expression in some epithelia and tumours was recently
described by van Muijen et al. (1986) using monoclonal
antibodies specific to keratins 4 and 13.
Using three monoclonal antibodies BA16, BA17 and
A53-B/A2 specific for human keratin 19 we found good
correlation between the staining patterns observed and
two-dimensional gel analyses of human tissues reviewed
by Moll et al. (I982a) and Quinlan et al. (1985). In general
we found homogeneous positivity and/or heterogeneous
staining (Type I) in epithelial tissues containing enough
keratin 19 to be detected biochemically as opposed to
complete negativity of staining in tissues lacking this
keratin according to two-dimensional gel analyses (Moll et
al., 1982a; Quinlan et al., 1985; Achtstatter et al., 1985).
The only discrepancy between the two sets of data
concerns hair follicle epithelium (Moll et al., 1982b) with
the minor subpopulation of positive cells possibly
representing too little keratin 19 to be detected biochemi-
cally.
Furthermore, our immunohistochemical data provide
two possible explanations for small or variable amounts of
keratin 19 detected in some epithelial tissues (Moll et al.,
1982a; Quinlan et al., 1985). Firstly, we interpret the
mosaic pattern of positive and negative cells (Type I
heterogeneity) in some glandular epithelia (e.g. prostate) as
the coexistence of two subpopulations of cells only one of
which expresses keratin 19, thus resulting in lower
amounts of the 40 kDa keratin polypeptide detected in
tissue homogenates. Secondly, the complex and variable
heterogeneous staining pattern seen in non-keratinizing
squamous epithelia (Type II heterogeneity) with generally
strong positivity of the basal layer and variable intensity of
the suprabasal cells suggests that both qualitative and
quantitative differences among epithelia of various donors
and different areas of the tissue are responsible for
variability detected by biochemical methods in samples of
these tissues (Quinlan et al., 1985).
While false negative staining results due to masking of
keratin epitopes have been reported (Woodcock-Mitchell
et al., 1982; Franke et al., 1983) we prefer the above-
BARTEK et al.
mentioned interpretation of the Type I heterogeneity
because the same staining pattern was seen on parallel
sections with all three monoclonal antibodies recognizing
different antigenic determinants, and because Western blot
analysis of various cell and tissue-lysates always correlated
with the staining data (Bartek et al., 1985a, b, 1986). Our
interpretation is further supported by the fact that both
protease treatment, known to be able to unmask some
keratin determinants on paraffin sections (Debus el al.,
1982; Makin et al., 1984), and application of biotin--strep-
tavidin-based techniques failed to induce any positivity of
the glandular epithelia negative in the standard indirect
immunoperoxidase method.
In a previous paper we have suggested that the keratin
19-negative cells seen in normal breast terminal ductal
lobular units and in benign breast tumours may represent a
subclass of lumenal cells which have a higher proliferative
potential than the keratin 19 positive cells, which however
may arise from them (Bartek el al., 1985a,b). The fact that
this type of mosaic pattern of positive and negative cells is
seen in other normal tissues in areas where stem cells might
be present (e.g. in the crypts of the small and large bowel)
suggests that keratin 19 expression may indeed be related
to proliferative potential in a variety of tissues. Such a
concept can be experimentally tested using a combination
of autoradiography and immunohistochemistry and
studies are under way to do this using breast tissue.
Interpretation such as that presented for Type I
heterogeneity is obviously not applicable for the staining
patterns observed in non-keratinized squamous epithelia
(Type II heterogeneity) and possibly the hair follicle
epithelium as well. Besides considerable variability due to
different sources of tissues, two striking phenomena were
observed which were not found in any other epithelial
tissues. Firstly, there were clear differences between
staining patterns of antibody BA17 on frozen sections
(mainly restricted to basal layer) and sections of metha-
cam-fixed paraffin-embedded tissues (often suprabasal
staining of various intensity in addition to positive basal
cells). Secondly, in contrast to glandular epithelia (some-
times present on the same section) considerably fewer cells
were positive with BA16 than with the remaining two
antibodies in stratified squamous non-keratinized epithelia.
One possible explanation of the major differences
between the staining patterns in glandular epithelia as
opposed to squamous non-keratinized epithelia could be

Fig. 1. Indirect immunofluorescenceon human tissues using (a-d, i) antibody BA17 and (e-h) BA16. (a,b) Interfollicular
epidermis, (c) gall bladder, (d) Fallopian tube, (e) eccrine sweat gland, (f) urinary bladder, (g) colon: surface mucosa and upper
parts of several crypts, (h) mammary gland: terminal ductal lobular unit, (i) epididymis ductuli efferentes. (a-d, i) Methacam-fixed
paraffin-embeddedsections; (e-h) frozen sections. (a-i) x 160.
Fig. 2. Indirect immunoperoxidase staining of human tissues with (b, d-g, i) antibody BA17 and (a,c,h) BA16. (a) Endometrium,
(b) pancreas (positive ductal epithelium as opposed to negative acini), (c) liver (negative hepatocytes in contrast to
homogeneously positive small bile duct), (d) thymus, (e) placenta: cytotrophoblast, (f) tracheal surface epithelium, (g) gastric
mucosa, (h) thyroid gland: only few epithelial cells are positive, (i) prostate. (d-i) Methacarn-fixed paraffin-embeddedtissue
sections; (a--c)frozen sections. (a-i) x 260.
Keratin 19 detection in human tissues 571

' ,~ ~ - ta ~

d " -~~j e

,p °;,~

h
572 BARTEK et al.
Keratin 19 detection in human tissues 573

1 ?

~ , ~ • ~ ~ ~. ~ •, ~ k L ~¸

2-" 2 :_

Fig. 3. (a-h) Indirect immunoperoxidase and (i) pre-formed streptavidin-biotinylated peroxidase complexes techniques using
(b-i) BA17 and (a) BA16 as primary antibodies. (a--c) Oesophagus, (d) transition zone of endo- and exocervia, (e,f) exocervix, (g)
tongue: homogeneously positive surface squamous nonkeratinized epithelium and heterogeneous glands, (h,i) parallel sections of
human breast tissue showing the same mosaic pattern regardless of the staining method used. (c-i) Methacarn-fixed paraffin-
embedded tissue sections; (a,b) frozen sections. (a,b) x 160, (c~f,h,i) x 260, (g) x 100.
574
associated with corresponding differences in keratins
which form naturally occurring complexes with keratin 19.
Thus while keratins 7 and 8 form keratin pairs by
association with keratin 19 in simple epithelia, it is
keratin 5 which is the partner of keratin 19 in non-
keratinizing squamous epithelia (Quinlan et al., 1985).
Different keratin partners can form sterically different
complexes, one consequence of which could be differential
masking of antigenic determinants in these two kinds of
epithelia. In other words, while the epitopes recognized by
all three antibodies are available in simple glandular
epithelia, only the A53-B/A2 determinant is not masked
in non-keratinized squamous epithelia thus giving rise to
positive staining regardless of the fixation method used.
The epitope recognized by BA17 would be then masked
(especially in suprabasal layers) on frozen sections while
being unmasked by methacam fixation, thus resulting in
positivity of tissues processed in this way. This explana-
tion is supported by suprabasal positivity of BA17 seen
after methacam fixation of frozen sections as compared
with standard methanol-acetone fixation (Bartek & Bart-
kova, unpublished observation, 1986).
The observation that fewer cells are stained by BA16
than by either BA17 or A53-B/A2 could be due to
'permanent masking' of the BA16 target epitope by
association with other molecules (possible keratin 5) or to
some post-translational modification of the target determi-
nant occurring only in some epithelia. Both of these
explanations are possible since masking of epitopes in
intermediate filaments has been demonstrated (Franke et
al., 1983, 1984) and Virtanen and colleagues have
characterized a monclonal antibody which can distinguish
between unmodified and post-translationally phosphory-
lated vimentin (Virtanen, personal communication, 1986).
The possibility that the antibodies are cross-reacting with a
keratin other than keratin 19 in non-keratinizing squamous
epithelium is unlikely.
Other investigators have demonstrated the usefulness
of anti-keratin monoclonal antibodies in many applications
including the studies of epithelial cell lineages, differentia-
tion and malignancy (Tseng el al., 1982; Debus el al., 1982;
Taylor-Papadimitriou et al., 1983; Ramaekers et al., 1983;
Cooper et al., 1985; Lane et al., 1985; Taylor-Papadimitriou
et al., 1985; Bartek et al., 1985a, b; van Muijen el al., 1986;
Bartek et al., 1986). We believe our present work on
differential expression of the human keratin 19 in normal
tissues will contribute to such studies using anti-keratin 19
monospecific monoclonal antibodies as a tool.
Conclusion
Immunohistochemical examination of normal human
tissues reveals hitherto unrecognized complexity of
keratin 19 expression as detected with three monoclonal
antibodies monospecific for keratin 19. Thus, although
some simple epithelia show a homogeneously positive
reaction with all three antibodies and some other epithelia
B A R T E K et al.
(e.g. epidermis, hepatocytes) show complete negativity,
heterogeneity of staining can also be observed. Where the
heterogeneity is the same for all three antibodies it is
characterized by a mosaic pattern made up of strongly
positive a n d completely negative cells. Morphologically
,similar epithelial cells in these tissues (usually simple
glandular epithelia) are therefore subclassified into cells
expressing keratin 19 and cells not expressing keratin 19,
and these two subclasses may have other phenotypic and
functional differences. Where the heterogeneity is less
clear cut and is not the same for each antibody, (in some
non-keratinizing squamous epithelia) it seems likely that it
reflects differences in masking of the target epitope, or in
post translational modifications in the keratin polypeptide.
Acknowledgements
The authors are grateful to Dr U. Karsten for his generous
gift of the A53-B/A2 antibody, to Jana Hanykova, Jirina
Sevcikova and Leona Hladka for doing tissue sections, and
to Liz Eaton for typing the manuscript.
References
ACHTSTATTER, T., MOLL,R., MOORE, B. & FRANKE,W. W.
(1985) Cytokeratin polypeptide patterns of different
epithelia of the human male urogenital tract:
immunofluorescence and gel electrophoretic studies. J.
Histochem. Cytochem. 33, 415-26.
BARTEK,J., VIKLICKY,V., FRANEK,F., ANGELISOVA,P.,
DRABER, P., JAROSIKOVA,T., NEMEC, M. & VERLOVA,
H. (1982) Monoclonal antibodies against transferrin.
Precipitating mixtures and lack of interspecies cross-
reactivity. Immunol. Left. 4, 231-5.
BARTEK,J., DURBAN, E. M., HALLOWES,R. C. & TAYLOR-
PAPADIMITR1OU, J. (1985a) A subclass of luminal
epithelial cells in the human mammary gland defined l~y
antibodies to cytokeratins. J. Cell Sci. 75, 17-33.
BARTEK,J., TAYLOR-PAPADIMITRIOU,J., MILLER,N. &
MILLIS,R. (1985b) Patterns of expression of keratin 19
as detected with monoclonal antibodies in human breast
tissues and tumours. Int. ]. Cancer 36, 299-306.
BARTEK,]., KOVARIK,J., LAUEROVA,L. & MUNZAROVA, M.
(1985c) Monoclonal antibody to intermediate filaments
of cytokeratin type. I. Drug studies and reactivity with
cultured cells and tissue sections. Folia biol., Praha 31,
1-8.
BARTEK,J., BARTKOVA,J., SCHNEIDER,J., TAYLOR-
PAPADIMINTRIOU, J., KOVARIK,J. & REITHAR,A.
(1986) Expression of monoclonal antibody-defined
epitopes of keratin 19 in human tumours and cultured
cells. Eur. ]. Cancer Clin. Oncol. (in press).
COOPER, D., SCHERMER,A. & SUN, T. T. (1985) Classification
of human epithelia and their neoplasms using
monoclonal antibodies to keratins: strategies,
applications, and limitations. Lab. Invest. 52, 243-56.
DEBUS, E., WEBER,K. & OSBORN, M. (1982) Monoclonal
cytokeratin antibodies that distinguish simple from
stratified squamous epithelia: characterization on human
tissues. EMBO ]. 1, 1641-7.
K e r a t i n 19 d e t e c t i o n in h u m a n t i s s u e s
FRANKE, W. W., SCHMID, E., WELLSTEED, J., GRUND, C.,
GIGI, O. & GEIGER, B. (1983) Change of cytokeratin
organization during the cell cycle: selective masking of
an immunologic determinant in PtK 2 cells. ]. Cell Biol.
97, 1255--60.
FRANKE, W. W., GRUND, C., KUHN, C., LEHTO, V.-P. &
VIRTANEN, I. (1984) Transient change of organization of
vimentin filaments during mitosis as demonstrated by a
monoclonal antibody. Exp. Cell Res. 154, 567-80.
KARSTEN, U., PAPSDORF, G., ROLOFF, G., STOLLEY, P., ABEL,
H., WALTHER, I. & WEISS, H. (1985) Monoclonal anti-
cytokeratin antibody from a hybridoma clone generated
by electrofusion. Eur. J. Cancer Clin. Oncol. 21, 733-40.
LANE, E. B. (1982) Monoclonal antibodies provide specific
intramolecular markers for the study of epithelial
tonofilament organization. ]. Cell Biol. 92, 665-73.
LANE, E. B., BARTEK, J., PURKIS, P. E. & LEIGH, I. M. (1985)
Keratin antigens in differentiating skin. Ann. N.Y. Acad.
Sci. 455, 241-58.
LAZARIDES, E. (1980) Intermediate filaments as mechanical
integrators of cellular space. Nature, Lond. 283, 249-56.
MAKIN, C. A., BOBROW, L. G. & BODMER, W. F. (1984)
Monoclonal antibody to cytokeratin for use in routine
histopathology. ]. Clin. Pathol. 37, 975-83.
MOLL, R., FRANKE, W. W., SCHILLER, D. L., GEIGER, B. &
KREPLER, R. (1982a) The catalog of human cytokeratins:
patterns of expression in normal epithelia, tumors and
cultured cells. Cell 31, 11-24.
MOLL, R., MOLL, I. & WIEST, W. (1982b) Changes in the
pattern of cytokeratin polypeptides in epidemis and hair
follicles during skin development in human fetuses.
Differentiation 23, 170--8.
OSBORN, M. & WEBER, K. (1982) Intermediate filaments: cell-
type-specific markers in differentiation and pathology.
Cell 31, 303--6.
QUINLAN, R. A., SCHILLER, D. L., HATZFELD, M.,
ACHTSTATTER, T., MOLL, R., JORCANO, J. L., MAGIN,
575
T. M. & FRANKE, W. W. (1985) Patterns of expression
and organization of cytokeratin intermediate filaments.
Ann. N.Y. Acad. Sci. 455, 282-306.
RAMAEKERS, F. C. S., HUYSMANS, A., MOESKER, O., KANT,
A., JAP, P., HERMAN, C. & VOOIJS, P. (1983)
Monoclonal antibody to keratin filaments, specific for
glandular epithelia and their turnouts. Use in surgical
pathology. Lab. Invest. 49, 353--61.
TAYLOR-PAPADIMITRIOU, J., LANE, E. B. & CHANG, S. E.
(1983) Cell lineages and interactions in neoplastic
expression in the human breast. In Understanding Breast
Cancer, Clinical and Laboratory Concepts (edited by RICH,
M. A., HAGER, J. C. and FURMANSKI, P.), pp. 215-46.
New York, Basel: Marcel Dekker.
TAYLOR-PAPADIMITRIOU, J., BARTEK, J., DURBAN, E. M.,
BURCHELL, J., HALLOWES, R. C. & LANE, E. B. (1985)
Monoclonal antibodies in the study of cell lineage,
differentiation and malignancy in the human breast. In
Monoclonal Antibodies and Breast Cancer (edited by
CERIANI, R. L.), pp. 60--79. Boston: Martinus Nijhoff.
TSENG, S. C. G., JARVINEN, M. J., NELSON, W. G., HUANG, J.
w., WOODCOCK-MITCHELL, J. & SUN, T. T. (1982)
Correlation of specific keratins with different types of
epithelial differentiation: Monoclonal antibody studies.
Cell 30, 361-72.
VAN MUIJEN, G. N. P., RUITER, D. J., FRANKE, W. W.,
ACHTSTATTER, T., HAASNOOT, W. H. B., PONEC, M. &
WARNAAR, S. O. (1986) Cell type heterogeneity of
cytokeratin expression in complex epithelia and
carcinomas as demonstrated by monoclonal antibodies
specific for cytokeratin Nos. 4 and 13. Exp. Cell Res.
162, 97-113.
WOODCOCK-MITCHELL, J., EICHNER, R., NELSON, W. G. &
SUN, T.-T. (1982) Immunolocalization of keratin
polypeptides in human epidermis using monoclonal
antibodies. J. Cell Biol. 95, 580-8.