Coexpression of specific acid and basic cytokeratins in teratocarcinoma-derived fibroblasts treated with 5-azacytidine

DEVELOPMENTAL

BIOLOGY

110,47-52

(19%)

Coexpression of Specific Acid and Basic Cytokeratins in Teratocarcinoma-Derived Fibroblasts Treated with 5-Azacytidine MICHEL Centre International

DARMON

de Recherches Dermatologiques,

Sophia Antipolis,

Received June 29, 1984; accepted in revised form

February

06565 Valbonne, France 5, 1985

Epithelial cells always coexpress acidic and basic keratin polypeptides. Mesenchymal cells, which do not normally contain keratins, can be induced by the inhibitor of DNA methylation 5-azacytidine to synthesize the basic keratin Endo A. In the present paper we show that the acidic keratins Endo B and Endo C can also be induced by 5-azacytidine in teratocarcinoma-derived fibroblasts. Furthermore, individual cells in which Endo B and/or Endo C keratins are found, always coexpress the basic polypeptide Endo A. Other cytokeratins are not or very rarely found. Interestingly, Endo A, B, and C are usually associated in viva and are known to be the first keratin polypeptides 0 1985 Academic Press, Inc. appearing during the development of the mouse embryo.

tionally associated. Actually, these signals could be elements of more general control mechanisms since the composition in keratin polypeptides of epithelial cell types plays probably some role in their morphological differentiation. Endo A (Oshima, 1981, 1982) also called cytokeratin A (Franke et aZ., 1981) is a basic keratin (pK, 6.4) having a M, of 55 in the mouse. In all situations so far examined it is associated with the acidic keratin polypeptide Endo B of pK, 5.4, and M, 50 (Oshima, 1981, 1982), also called cytokeratin D (Franke et ah, 1981). Endo A and Endo B are the first keratin polypeptides appearing in the embryo; they appear in the morula as early as it becomes possible to allocate cells to the trophectodermal lineage; these proteins are then synthesized in large amounts in trophectoderm and both visceral and parietal endoderm, but are not present in the inner cell mass (Brtilet et al., 1980; Kemler et al., 1981a,b; Jackson et al., 1981; Franke et ah, 1981; Oshima et ah, 1983; Lane et al, 1983; Boller et al., 1983; Lehtonen et al., 1983). Antibodies specific for these polypeptides, both monoclonal and polyclonal have been produced (Oshima, 1981, 1982; B&et et al., 1980; Kemler et al., 1981a,b). In the present work the monoclonal antibody (MAb) TROMA-I was used to detect Endo A, and a rabbit polyclonal antibody (PAb) to detect Endo B. Additionally, Kemler et al. (1981a) described another cytoskeletal polypeptide appearing during early mouse development. This polypeptide recognized by TROMAIII MAb is frequently, but not always associated with Endo A and Endo B (Kemler et al, 1981a; Boller and Kemler, 1983). For instance, it is present in trophectoderm and parietal endoderm, but not in visceral endoderm (Boller and Kemler, 1983). This cytoskeletal

INTRODUCTION

The intermediate filaments (Lazarides, 1980; Osborn et ah, 1981) of the various epithelial cell types are composed of a specific subset of keratin polypeptides (Franke et al., 1978; Moll et al., 1982; Steiner et al., 1983). These cytoskeletal proteins are the products of two multigene families (I and II), coordinately conserved throughout vertebrate evolution (Fuchs et al., 1981; Schiller et ah, 1982; Tseng et ah, 1982; Hanukoglu and Fuchs, 1983; Fuchs and Marchuk, 1983). Type I keratins are acidic (pK, 4.5-5.5) whereas type II keratins are more basic (pK, 6.5-7.5). All epithelial cells express coordinately products of both families (Schiller et al., 1982; Tseng et ah, 1982; Moll et aZ., 1982; Steiner et aZ., 1983; Fuchs et ah, 1981; Kim et al, 1983). Since at least two keratins are known to be required for filament reconstitution (Steiner et al., 1976; Lee and Baden, 1976), the two subfamily concept may be of importance for filament formation. Moreover if one considers the keratin patterns of different epithelia, certain rules of association seem to prevail. Particularly, keratin polypeptides seem to appear during differentiation as specific acid-basic couples (WoodcockMitchell et aZ., 1982; Moll et aZ., 1982; Kim et al, 1983; Oshima et ah, 1983; Sun et al., 1984). A classical example is the coordinated expression of a first 50-kDa acid-58-kDa basic couple in basal keratinocytes and the appearance of a second 56-kDa acid-67-kDa basic couple during subsequent keratinocyte differentiation (Woodcock-Mitchell et ah, 1982; Fuchs, 1983). Regulatory signals must thus exist to ensure in the two keratin multigene families the coordinated activation of the pair of genes coding for the proteins which are func47

0012-1606185 $3.00 Copyright All rights

0 1985 by Academic Press, Inc. of reproduction in any form resewed.

DEVELOPMENTALBIOLOGY

48

protein copurifies with Endo A and Endo B when intermediate filaments are extracted from trophectodermal cells (Kemler et aL, 1981a). Moreover, an intermediate filament network can be decorated by TROMAIII MAb only in epithelial cells (Kemler et al., 1981a). This intermediate filament protein can thus be considered to be a cytokeratin. It is acidic (pK, 5.6) and has 45 (R. Kemler personal coman M, of approximately munication). In this paper we call this keratin polypeptide Endo C. The murine keratins Endo A and B correspond to the human keratins 8 and 18, respectively (Moll et al, 1982), but it is not clear to which human keratin Endo C is related. The description of keratin patterns in embryonic and adult tissues or else in culture systems where spontaneous differentiation occurs is useful to determine what keratin polypeptides are usually coordinately expressed. However, to get a further insight in the regulatory mechanisms involved, we decided to try to induce the synthesis of keratin polypeptides in cells which do not normally contain keratins, such as mesenchymal cells. We used an agent acting directly at the gene level by inhibiting DNA methylation, 5azacytidine (Constantinides et ak, 1977,1978). We found that when teratocarcinoma-derived fibroblasts (1246 cell line) are treated with 5-azacytidine (5-AzaC) a certain fraction of the cells expresses the cytokeratin Endo A (Darmon et aZ., 1984) and the corresponding messenger RNA (M. Vasseur and M. Darmon, unpublished data). The goal of the work here reported was to determine whether the keratin polypeptides usually associated with Endo A in natural systems (Endo B and Endo C) are also synthesized under such artificial conditions of induction. MATERIALS

AND

METHODS

Cell lines. 1246 Fibroblasts were grown and treated with 5-AzaC as described previously (Darmon et al., 1984). Antibodies. TROMA-I, TROMA-III, LE61 and KLl, KGS-13, AEl, and AE3 MAbs have been described previously (Brtilet et ak, 1980; Kemler et al, 1981; Lane, 1982; Gigi et al., 1982; Tseng et al., 1982; Viac et al., 1983). Anti-End0 B PAb have been described previously (Oshima, 1982). Immunofluorescence and cell counting were performed as described (Darmon et al., 1984) except in the case of the double and triple labeling using monoclonal antibodies TROMA-I and TROMAIII which could not be performed following classical procedures but had to be done in two successive steps separated by washing at pH 2.2 to eliminate the first MAb. Pictures of the same area of the dish were taken both after the labeling with the first and the second monoclonal antibody. Immune blots were performed

VOLUMEHO,1985

as described (Darmon et al., 1984) using the procedure of Towbin et al. (1979) on high salt-insoluble extracts (Winter et ah, 1980). RESULTS

AND

DISCUSSION

1246 Fibroblasts (Darmon et al, 1984) were treated with 8 X 10e6 M 5-AzaC for 24-48 hr, then refed with fresh medium for 48 hr, fixed, and stained with the antibodies to be used; positive cells were scored as previously described (Darmon et ak, 1984). We found that the treated cultures not only contained cells recognized by TROMA-I MAb as previously reported, but also cells recognized by anti-End0 B PAb and TROMA-III MAb (see Fig. 1). Immune blots were performed to confirm that these changes actually corresponded to the induction of Endo A, B and C keratins. High salt-insoluble proteins, prepared from approximately 4 X lo7 cells treated with 8 X 10m6M 5-AzaC, showed bands of 55 (Endo A), 50 (Endo B) and 43 (Endo C) kDa recognized, respectively, by TROMA-I, anti-End0 B, and TROMA-III antibodies (see Fig. 2). No reaction could be detected with extracts prepared from control cultures. Although the percentage of positive cells varied from one experiment to another (5 to 10% Endo A+ cells) the ratio of the number of Endo B+ cells or Endo C+ cells to the number of Endo A+ was fairly constant and of approximately 60-80s. On the other hand very few cells (if any) could be labeled with reagents recognizing basic and acidic keratin polypeptides different from Endo A, B, or C, such as LE61 (Lane, 1982), KLl (Viac et ah, 1983), KG8-13 (Gigi et aZ., 1982), AEl and AE3 (Tseng et aZ., 1982) MAbs, or anti-total epidermal keratins PAb (Darmon et ab, 1984). This fact is not surprising since keratinocyte clones could be derived from 5-azacytidine-treated 1246 cells at an extremely low frequency (Darmon et ah, 1984). Double labeling with TROMA-I MAb and anti-End0 B PAb indicated that all the Endo B+ cells were also Endo A+. However, some Endo A+ cells (approximately one fourth) were found to be Endo B-. An example of such a double labeling is reported in Figs. la and b. This observation also points out that among all the possible combinations of Endo A, B, and C in the same cells, the two classes of cells A-B’C and A-B’C? contain very few elements if any (see Table 1). Double labeling with TROMA-I and TROMA-III MAbs could not be performed following classical procedures but had to be done in two successive steps (TROMA-III, then TROMA-I) separated by washing at pH 2.2 to eliminate the first MAb. Pictures of TROMAIII+ cells were taken before the washing and pictures of the same area of the dish after the second labeling

MICHEL DARMON

Goexpression

of Acid

and Basic

Keratins

FIG. 1. Staining of Endo A, Endo B, and Endo C cytokeratins by double immunofluorescence in 1246 fibroblasts treated with 8 X 10m6M 5-AzaC. Endo A was recognized by TROMA-I MAb, Endo B by specific rabbit PAb, and Endo C by TROMA-III MAb. Double labeling with TROMA-I (rat) MAb and anti-End0 B (rabbit) PAb was performed with anti-rat Ig coupled to rhodamine and anti-rabbit Ig coupled to fluorescein. The same fluorescent conjugates were used to perform double labeling with TROMA-III (rat) MAb and anti-End0 B PAb. (a, b) Double labeling with TROMA-I and Endo B; all Endo B+ cells (b) are Endo A’ (a) but some Endo A+ cells are Endo B-. (c, d) Double labeling with anti-End0 B PAb and TROMA-III, two Endo B’ cells (c) are Endo C (d); one Endo C+ cell is Endo B-; one cell is Endo B+ Endo C+. (e, f) Same conditions as (c, d) two Endo C+ cells (f) are Endo B- (e). (g, h) Successive labelings of the same area of the dish with TROMA-III and anti-rat Ig coupled to rhodamine (g) then TROMA-I and anti-rat Ig coupled to fluorescein (h); elimination of the first sandwich was done by washes with 0.2 M glycine-HCl, pH 2.2; appropriate controls were performed to check that the washes eliminated not only the anti-rat Ig coupled to rhodamine but also TROMA-III MAb. All Endo C’ cells (g) are Endo A+ (h) but some Endo A+ cells are Endo C-. Magnification: (a, b, g, h) X800; (c, d, e, f) X750.

with TROMA-I (see Figs. lg, h). In these experiments all the Endo C+ cells photographed (about 50) were found to be also Endo A+ but approximately one-third

of Endo A+ cells were found to be Endo C-. This result seems to indicate that cells of the types A-B-C? and A-B+C+ do not exist or are very rare (see Table 1).

50

DEVELOPMENTAL BIOLOGY

VOLUME 110, 1985

cells having already one of the others is 5- to lo-fold greater than in the overall population. (2) As in viva, the presence of the basic keratin Endo A is constant in cells expressing the acidic keratins Endo B and/or 92.5, C. However, contrary to what is found in vivo, some cells expressing only Endo A could be detected. More68, : *,55 i over, in experiments performed only 24 hr after the 0 -,50 end of 5-AzaC treatment, when the first Endo A+ cells 43, x ,,43 appear, no Endo B+ or Endo C+ cells could be detected, which indicates that the synthesis of Endo A precedes i that of Endo B and C. To try to answer the question 25.7, % whether A+B-C- cells contain other acidic polypeptides than Endo B or C, we labeled the cultures with AE1 a b c d e f MAb reported to react with most acidic keratins (Tseng et al, 1982), but obtained negative results. This question FIG. 2. Polyacrylamide gels (10%) and immune blots of 1246 high salt-insoluble extracts. (a, b) TROMA-I MAb; (c, d) anti-End0 B will be solved when an antibody (or a collection of PAb; (e, f) TROMA-III MAb; (b, d, f) 5-AzaC-treated cells; (a, c, e) antibodies) reacting with all acidic keratins of the control cultures. mouse will be available. (3) Contrary to in viva, Endo A is not necessarily associated with Endo B but may Double labeling with anti-End0 B PAb and TROMA- instead be associated with Endo C only. (4) As in vivo, III MAb showed that the association of the Endo B Endo B and C (both acidic) may coexist in the same and Endo C characters was facultative. Examples are cells. Since it is reasonable to assume that 5-AzaC is reported in Figs. lc and d and e and f. Among Endo B+ cells, approximately two-thirds were also Endo C+ incorporated at random in the DNA, the high frequency of associations among Endo A, B, and C cannot be and among Endo C? cells, approximately three-quarters interpreted as the addition of independent events unless were also Endo B+. Thus, although the association the corresponding genes (but not any other keratin Endo B-End0 C is facultative, the frequency of Endo gene) would be particularly inaccessible to DNA-cytoB+ cells among Endo C+ cells or Endo B+ cells among sine methyltransferases (Santi et al., 1983). Moreover, Endo C+ cells is 5- to lo-fold higher than their frequency 5-AzaC is able to induce complete phenotypic converin the total population (Endo B+ or Endo C+ cells are sions of fibroblasts into myoblasts, adipocytes, chonless than 5% of the total population). The results described above suggest that cells expressing acidic keratins (Endo B and/or C) always TABLE 1 coexpress the basic keratin Endo A. To determine ASSOCIATION OF THE MARKERS Endo A, B, AND C IN 1246 whether the reciprocal is also true, triple labeling had FIBROBLASTS TREATED WITH 5-AzaC to be performed. Cells were first stained with TROMAClasses of cells Number of cells Percentage III MAb and anti-End0 B PAb; photographs were taken, and the cells washed at pH 2.2; a staining with A+B+C+ 50 51 TROMA-I MAb was then performed, and the same A+B+C26 26.5 areas of the dishes were photographed. Examples of A+B-C+ 17 17.5 A+B-C5 5 such labelings are reported in Fig. 3. This procedure A-B+C 0 0 allowed to obtain an estimate in the same experiment A-B-C+ 0 0 of the proportions of cells in the various classes (see A-BY+ 0 0 results Table 1). Surprisingly, cells expressing Endo Total 98 100 A, but neither Endo B nor C (A+B-C- cells) could be Total A+ 98 100 detected (see Figs. 3a-c). However, they were not very Total B’ 76 77.5 frequent (5% of total Endo A+). The majority of the Total C+ 67 68.5 cells fell into three classes, A+B+C? (51%), A+B+C Note. Triple labeling was performed as described in Results, and (26.5%), and A+B-C? (17.5%). It was also confirmed in photographs similar to those reproduced in Fig. 3 were analyzed. this experiment that no cells of the classes A-B+C-, Cells of the various classes were scored and the percentage was A-B-C+, or A-BY? could be detected. calculated relative to the number of cells expressing at least one of Several conclusions can be drawn from these obser- the three cytokeratins (equal in the present case to total Endo A+ vations. (1) As in Vito, Endo A, B and C tend to be cells). In this experiment, total Endo A+ cells were 6% of the 1246associated; the frequency of each of these markers in treated cells. 200,

MICHEL DARMON

FIG. 3. Staining detection of Endo A+B+C+ class, two cells belong to the

Coexpressim of Acid and Basic Keratins

of Endo A, Endo B, and Endo C cytokeratins Cf cells; (c, f) detection of Endo A+ cells. (a-c) cells to the A+B+C- class, and three cells to the A+B’C- class and two cells to the A+B-C+ class.

by triple immunofluorescence. (a, d) Detection of Endo B+ cells; (b, e) Are photographs of the same area of the dish; two cells belong to the AfB-C- class. (d-f) Are photographs of the same area of the dish, Four Magnification X750.

drocytes, keratinocytes, and glial cells (Darmon et al., 1984; Constantinides, 1977, 19’78). The wide variety of characters acquired during such conversions almost excludes the possibility of their independent induction. It is therefore more likely that 5-AzaC promotes the activation of a single DNA target or of a few DNA targets as one could interpret from the analysis of dose-response curves (Darmon et al., 1984) and that this activation results in pleiotropic effects. Since these pleiotropic effects eventually lead to phenotypes resembling those observed in vivo, the targets hit by 5-AzaC can be imagined to play a key role in the control of differentiation programs. As far as the control of keratin synthesis is concerned, some working hypotheses can be proposed: (1) keratin genes which are

coexpressed (acid-basic couples) are in tandem on the chromosomes (2) they contain (or are flanked with) similar regulatory sequences recognized by an intracellular inducer (3) the product of one gene induces the activation of the other appropriate gene(s). The availability of probes for the elements of this system such as cDNAs (BrGlet et al., 1982) and polyclonal and monoclonal antibodies (Oshima, 1981, 1982; Brdlet et aZ., 1980; Kemler et al., 1981) should allow to tackle these questions. We thank Dr. R. Kemler and Dr. B. Oshima for the gift of antibodies and for stimulating discussions, and Dr. B. Geiger for the gift of MAb KG&13 and Dr. T. Sun for the gifts of MAbs AEl and AE3. We thank Dr. P. Giacomoni for his criticism of the manuscript and A. Semat for excellent technical assistance.

52

DEVELOPMENTAL BIOLOGY REFERENCES

BOLLER, K., and KEMLER, R. (1983). In vitro differentiation of embryonal carcinoma cells characterized by monoclonal antibodies against embryonic cell markers. In “Cold Spring Harbor Symposium on Cell Proliferation” Vol. 10, pp. 39-49. Cold Spring Harbor, N. Y. BR~?LET,P., BABINET, C., KEMLER, R., and JACOB, F. (1980). Monoclonal antibodies against trophectoderm specific markers during mouse blastocyst formation. Proc. Natl Acad. Sci. USA 77, 4113-4117. BRI?LET, P., and JACOB, F. (1982). Molecular cloning of a cDNA sequence encoding a trophectoderm-specific marker during mouse blastocyst formation. Proc. Nat1 Acad. Sci USA 79, 2328-2332. CONSTANTINIDES, P., JONES, P., and GEVERS, W. (1977). Functional striated muscle cells from non-myoblast precursors following 5-Azacytidine treatment. Nature (London) 267, 364-366. CONSTANTINIDES, P., TAYLOR, S., and JONES, P. (1978). Phenotypic conversion of cultured mouse embryo cells by Aza pyrimidine nucleosides. Dev. BioL 66, 57-71. DARMON, M., NICOLAS, J. F., and LAMBLIN, D. (1984). 5-Azacytidine is able to induce the conversion of teratocarcinoma-derived mesenchymal cells into epithelial cells. EMBO J. 3, 961-967. FRANKE, W., DENK, H., KALT, R., and SCHMID, E. (1981). Biochemical and immunological identification of cyto-keratin proteins present in hepatoeytes of mammalian liver tissue. Exp. Cell. Res. 131, 299318. FRANKE, W., SCHMID, E., OSBORN, M., and WEBER, K. (1978). Different intermediate-sized filaments distinguished by immunofluorescence microscopy. Proc. Natl. Acad. Sci. USA 75, 5034-5038. FUCHS, E. (1983). Evolution and complexity of the genes encoding the keratins of human epidermal cells. J. Inu DermatoL 81, 141s144s. FUCHS, E. V., COPPOCK, S. M., GREEN, H., and CLEVELAND, D. W. (1981). Two distinct classes of keratin genes and their evolutionary significance. Cell 27, 75-84. FUCHS, E., and MARCHUK, D. (1983). Type I and type II keratins have evolved from lower eukaryotes to form the epidermal intermediate filaments in mammalian skin. Proc. Natl. Acad Sci. USA 80, 5857-5861. GIGI, O., GEIGER, B., ESHHAR, Z., MOLL, R., SCHMID, E., WINTER, S., SCHILLER, D., and FRANKE, W. (1982). Detection of a cytokeratin determinant common to diverse epithelial cells by a broadly crossreacting monoclonal antibody. EMBO J. 1, 1429-1437. HANUKOGLU, I., and FUCHS, E. (1983). The cDNA sequence of a type II cytoskeletal keratin reveals constant and variable structural domaines among keratins. Cell 33, 915-924. JACKSON, B. W., GRUND, C., WINTER, S.,FRANKE, W.,~~~~LLMENSEE, K. (1981). Formation of cytoskeletal elements during mouse embryogenesis. II. Epithelial differentiation and intermediate sized in early postimplantation embryos. Dgerentiation 20, 203-215. KEMLER, R., BR~?LET,P., and JACOB, F. (1981a). Monoclonal antibodies as a tool for the study of embryonic development. In “The Immune System,” Vol. 1, pp. 102-109. Karger, Basel. KEMLER, R., BR~~LET,P., SCHNEBELEN, M. T., GAILLARD, J., and JACOB, F. (1981b). Reactivity of monoclonal antibodies against intermediate filament proteins during embryonic development. J. Embyol. Exp. Mwrphol. 64,45-60. KIM, K. H., RHEINWALD, J. G., and FUCHS, E. V. (1983). Tissuespecific epithelial keratins are encoded by two distinct classes of RNAs. Mol Cell Biol. 3, 495-502. LEHTONEN, E., LEHTO, V. P., VARTIO, T., BADLEY, R. A., and VIRTANEN, I. (1983). Expression of cytokeratin polypeptides in mouse oocytes and preimplantation embryos. Dev. Biol. 100, 158-165.

VOLUME 110, 1985 LANE, A. B. (1983). Monoclonal antibodies provide specific intramolecular markers for the study of epithelial tonolilament organization. J. Cell Biol. 92, 665-673. LANE, E. B., HOGAN, B. L. M., KURKINEN, M., and GARRELS, J. I. (1983). Co-expression of vimentin and cytokeratins in parietal endoderm cells of early mouse embryo. Nature (London) 303, 701704. LAZARIDES, E. (1980). Intermediate filaments as mechanical integrators of cellular space. Nature (London) 283, 249-256. LEE, L. D., and BADEN, H. P. (1976). Organization of the polypeptide chains in mammalian keratin. Nature (Lon&m,j 264, 377. MOLL, R., FRANKE, W., SCHILLER, D. L., GEIGER, B., and KREPLER, R. (1982). The catalog of human cytokeratins: Patterns of expression in normal epithelia, tumors and cultured cells. Cell 31, 11-24. OSBORN, M., GEISLER, N., SHAW, G., SHARP, G., and WEBER, K. (1981). Intermediate filaments. Cold Spring Harbor Symp. Quant. Biol. 46, 413-429. OSHIMA, R. G. (1981). Identification and immunopreeipitation of cytoskeletal proteins from murine extra-embryonic endodermal cells. J. Biol. Chem 256, 8124-8133. OSHIMA, R. G. (1982). Developmental expression of murine extraembryonic endodermal cytoskeletal proteins. J. Biol. Chem 257, 3414-3421. OSHIMA, R. G., HOWE, W. E., KLIER, F. G., ADAMSON, E. D., and SHEVINSKY, L. H. (1983). Intermediate filament protein synthesis in preimplantation murine embryos. Deu. Biol. 99, 447-455. SANTI, D., GARRETT, C., and BARR, P. (1983). On the mechanism of inhibition of DNA-cytosine methyltransferases by cytosine analogs. Cell 33, 9-10. SCHILLER, D. L., FRANKE, W. W., and GEIGER, B. (1982). A subfamily of relatively large and basic cytokeratin polypeptides as defined by peptide mapping is represented by one or several polypeptides in epithelial cells. EMBO J. 1, 761-769. STEINER, P. M., IDLER, W. W., and ZIMMERMAN, S. B. (1976). Selfassembly of bovine epidermal keratin filaments in vitro. J. Mol. Biol. 108, 541-547. STEINER, P. M., RICE, R. H., ROOP, D. R., TRUS, B. L., and STEVEN, A. C. (1983). Complete amino-acid sequence of a mouse epidermal keratin subunit and implications for the structure of intermediate filaments. Nature &m&m) 302, 794-800. SUN, T.-T., EICHNER, R., SCHERMER, A., COOPER, D., NELSON, W. G., and WEISS, R. A. (1984). Classification, expression and possible mechanisms of evolution of mammalian epithelial keratins: A unifying model. In “Cancer Cells” Vol. 1, “The Transformed Phenotype,” pp. 69-176. Cold Spring Harbor Laboratory, Cold Spring Harbor, N. Y. TOWBIN, H., STAEHELIN, T., and GORDON, J. (1979). Electrophoretic transfer of proteins from polyacrylamide gels to nitroeellulose sheets: Procedure and some applications. Proc. Natl Acad Sci. USA 76, 4350-4354. TSENG, S. C. G., JARVINEN, M. J., NELSON, W. G., HUANG, J. W., WOODCOCK-MITCHELL, J., and SUN, T. T. (1982). Correlation of specific keratins with different types of epithelial differentiation: Monoclonal antibody studies. C&Z 30, 361-372. VIAC, J., REANO, A., BROCHIER, J., STAQUET, M. J., and THIVOLET, J. (1983). Reactivity pattern of a monoclonal antikeratin antibody (KLi). J. Invest. Dermatol. 81, 351-354. WINTER, H., SCHWEIZER, J., and GOERTTLER, K. (1980). Keratins as markers of malignancy in mouse epidermal tumours. Careinogenesti 1, 391-398. WOODCOCK-MITCHELL, J., EICHNER, R., NELSON, W. G., and SUN, T. T. (1982). Immunolocalization of keratin polypeptides in human epidermis using monoclonal antibodies. J. Cell Biol. 95, 580-588.