Intermediate filaments in rat ovarian surface epithelial cells: changes with neoplastic progression in culture
Biochem. Cell Biol. Downloaded from www.nrcresearchpress.com by YORK UNIV on 11/10/14 For personal use only.
ANNE. HORNBY,' JIE PAN,AND NELLYAUERSPERG~ Department of Anatomy, University of British Columbia, Vancouver, B.C., Canada V6G 2W6 Received July 26, 1991 HORNBY,A. E., PAN, J., and AUERSPERG, N. 1992. Intermediate filaments in rat ovarian surface epithelial cells: changes with neoplastic progression in culture. Biochem. Cell Biol. 70: 16-25. Interrelationships between neoplastic progression and the expression of intermediate filaments were examined in primary cultures, immortal lines, and Kirsten murine sarcoma virus (KiMSV) transformed lines of rat ovarian surface epithelial (ROSE) cells. Immunofluorescence microscopy revealed abundant keratin filaments in all cells of primary cultures. In immortal, nontumorigenic lines, keratin filaments were detected in fewer cells, in smaller numbers, and in microscopically altered forms. The percentage of keratin-positive cells ranged from 4 to 54%. Its expression was inversely proportional to cell density. Keratin expression was similar in the two immortal lines, although one had retained a monolayered epithelial growth pattern resembling primary cultures, while in the other the growth pattern of the cells was more atypical. The two KiMSV-transformed lines were previously shown to produce tumors in vivo that resemble human ovarian endometrioid stromal sarcomas. In spite of this histologic appearance, the proportion of keratin-positive cells in these cells was increased over the immortal lines. Keratin expression was unrelated to cell density, and keratin in most virally transformed cells was limited to few, fine filaments. In thymidine-labelled immortal and virus-transformed cultures stained for keratin, no correlation was found between keratin expression and proliferative activity. The keratin profiles of primary and immortal cultures were identical on Western blots, with subtypes ranging from 52 to 66 kDa. The two virally transformed lines lacked some of the subtypes. Vimentin networks were faint or absent in primary cultures. In the immortal and the virus-transformed lines, neoplastic progression was associated with increasing vimentin expression but with no changes in filament morphology and distribution. The results show that the abnormalities in intermediate filament expression that accompany immortalization do not preclude the retention of a normal epithelial morphology and growth pattern in this cell type. Furthermore, the number of intermediate filaments and their intracellular distribution appear to be altered at an earlier stage in neoplastic progression than those mechanisms that select for specific keratin subtypes, or those that respond to regulation by cell density. Finally, the presence of keratin in the KiMSV-transformed lines examined in this study supports the hypothesis that human ovarian stromal sarcomas can arise in the OSE. Key words: neoplastic transformation, keratin, vimentin, cytoskeleton, ras oncogene, ovary, tissue culture. HORNBY,A. E., PAN, J., et AUERSPERG, N. 1992. Intermediate filaments in rat ovarian surface epithelial cells: changes with neoplastic progression in culture. Biochem. Cell Biol. 70 : 16-25. Nous avons examink les relations entre la progression nkoplasique et l'expression des filaments intermtdiaires dans des cultures primaires, des ligntes immortelles et des ligntes transformtes par le virus du sarcome murin de Kirsten (KiMSV) des cellules tpithkliales de la surface des ovaires de rate (ROSE). La microscopie par immunofluorescence rkvtle d'abondants filaments de kkratine dans toutes les cellules des cultures primaires. Dans les lignkes immortelles, non tumorigtnes, les filaments de kkratine sont dktectts dans peu de cellules, en plus petit nombre et sous des formes microscopiquement modifikes. Le pourcentage des cellules rkagissant positivement A la ktratine varie de 4 B 54%. Son expression est inversement proportionnelle 11 la densitk cellulaire. L'expression de la ktratine est similaire dans les deux ligntes immortelles; l'une a retenu un profil de croissance kpithtliale monolamellaire ressemblant 11 celui des cultures primaires tandis que dans l'autre, le profil de croissance des cellules est plus atypique. I1 a dkjh tte montrt que les deux lignkes transformkes par le KiMSV produisent des tumeurs in vivo qui ressemblent aux sarcomes stromatiques endomktrioides des ovaires humains. En dtpit de cette apparence histologique, la proportion des cellules ktratine-positives dans ces cellules augmente davantage que dans les lignkes immortelles. I1 n'existe aucune relation entre I'expression de la kkratine et la densitt cellulaire et, dans la plupart des cellules transformkes par le virus, la ktratine est limitte un petit nombre de fins filaments. Dans les cultures immortelles et les cultures transformtes par le virus, marqutes a la thymidine et colorkes pour la ktratine, nous n'avons dtcelk aucune corrklation entre l'expression de la kkratine et l'activitt prolifkratrice. Les profils de la kkratine des cultures primaires et des cultures immortelles sont identiques sur les transferts Western; on y voit des sous-types allant de 52 B 66 kDa. Les deux ligntes transformkes par le virus ne posstdent pas de tels sous-types. Dans les cultures primaires. la rtseaux de vimentine sonr absents ou a peine visibles. Dans les lignks immortelles et les ligntes transformkes par le virus, la progression ntoplasique est associCe h l'expression accrue de la vimentine, mais non aux changements dans la rnorphologie et la distribution des filaments. Les rCsultats montrent que les anomalies de l'expression des filaments intermMiaires qui accompagnent I'irnmortalisation n'empechent pas le maintien d'une morphologie et d'un profil de croiss;mce ipithtliales normalcs dans ce type de cellules. De plus, le nombre de filaments intermMiaires et leur distribution intracelIulaire semblent se modifier plus tat dans la progression ntoplasique que ces mkcanismes qui veillent au choix des sous-types sptcifiques de ktratjne ou ceux qui rkpondent ABBREVIATIONS: OSE, ovarian surface epithelium; ROSE, rat ovarian surface epithelium; BSS, balanced salt solution; EGF, epidermal growth factor; PBS, phosphate-buffered saline; NGS, normal goat serum; SDS-PAGE, sodium dodecyl sulfate - polyacrylamide gel electrophoresis; TBS, 20 mM Tris-HC1 buffer (pH 7.5); TTBS, TBS containing 0.05% Tween-20; KiMSV, Kirsten murine sarcoma virus. 'present address: Department of Medical Genetics, University of British Columbia, 6174 University Boulevard, Vancouver. B.C. V6T 123. ' ~ u t h o rto whom correspondence should be addressed. Printed in Canada / lrnprirnt au Canada
17
HORNBY ET AL.
Biochem. Cell Biol. Downloaded from www.nrcresearchpress.com by YORK UNIV on 11/10/14 For personal use only.
a la rtgulation par la densitt cellulaire. Finalement, la presence de ktratine dans les ligntes transformtes par le KiMSV examintes dans cette ttude confirme I'hypothbe que les sarcomes stromatiques ovariens humains peuvent se produire dans les cellules tpithtliales a la surface des ovaires de rate. Mots clds : progression ntoplasique, ktratine, vimentine, cytosquelette, oncogene ras, ovaire, culture de tissus. [Traduit par la rtdaction]
Introduction The shapes of cells and their contacts with the environment influence and are influenced by the characteristics of the intermediate filaments and, ultimately, they affect cellular growth and differentiation (Ben Ze'ev 1986). The intermediate filaments keratin and vimentin belong to a group of widespread, highly polymorphic, and insoluble cytoskeletal proteins (Goldman et al. 1985; Steinert and Roop 1988). Their precise role in the formation and maintenance of cell shapes and cytoplasmic compartmentization is still unclear. It has been suggested on the basis of their complex distribution among tissues that their functions may be differentiation related rather than "housekeeping" or universal (Robson 1989; Stewart 1990). Transformation to malignancy is accompanied by profound qualitative and quantitative alterations in the expression of several cytoskeletal components, including keratin and vimentin. However, it is not clear whether the cytoskeletal alterations are causes or consequences of specific steps in the carcinogenetic process. To investigate these interrelationships, we have examined the distribution of intermediate filaments at different stages during the neoplastic progression of rat ovarian surface epithelial (ROSE) cells. The expression of keratin and vimentin was compared in normal ROSE in primary culture, and in a series of ROSE-derived cell lines that express different degrees of neoplastic transformation. Two immortal lines arose by spontaneous transformation from primary ROSE cultures. Of these, one line (ROSE 239) is nonturnorigenic and resembles normal cells in primary culture in its cellular morphology, its epithelial monolayered growth pattern, and its growth inhibition by crowding. The second immortal line (Rose 199) is also nontumorigenic. However, its cells shed and multilayer in response to crowding and thus represent a more advanced stage within the immortal/nontumorigenic phase of neoplastic progression (Adams and Auersperg 1985). The distinctive characteristics of these lines have remained stable over many passages in culture. Two additional, fully transformed lines 03197-15A and V197-10B) were obtained by infection of primary ROSE cultures with Kirsten murine sarcoma virus (KiMSV). In addition to immortality, these lines expressed the culture characteristics of fully malignant cells, as well as tumorigenicity (Adams and Auersperg 1981). The ovarian surface epithelium (OSE) is the modified mesothelium that covers the surface of the ovary. OSE is of considerable clinical significance because it gives rise to most human ovarian carcinomas (Nicosia and Nicosia 1988). In spite of its importance, successful transformation to malignancy of human OSE under experimental conditions has not been reported. To our knowledge, the immortal and KiMSV-transformed ROSE lines described here represent the only model currently available for a comparison of stages in experimentally induced neoplastic transformation of OSE from any species. The types of intermediate filaments present in normal OSE vary among species. In the rat, ROSE cells in vivo are keratin positive but lack vimentin (Czernobilsky 1985). In humans, keratin is also present but
vimentin is coexpressed in varying proportions of the cells (Dabbs and Geisinger 1988; Benjamin et al. 1987; Moll et al. 1983a, 19833; Vide et al. 1988; Czernobilsky 1985). Human OSE-derived carcinomas tend to be keratin positive (Czernobilsky et al. 1984), while the expression of vimentin varies and seems to depend on the histotype and the degree of differentiation of these neoplasms. The effects of neoplastic transformation on the intermediate filaments of ROSE have not been described. The results of this study suggest that major alterations in the distribution, staining intensity, and morphology of keratin filaments take place at an early stage in neoplastic progression, while ROSE cells are immortalized but remain morphologically normal and nontumorigenic. The tissue-specific keratin subtypes, as shown by Western blotting, changed only in the tumorigenic, fully transformed lines. In contrast to the complex and variable alterations in keratin expression, vimentin expression increased gradually and uniformly in all cells with progression to malignancy. The results suggest that the mechanisms that regulate intermediate filament expression change with immortalization, and again with progression to the fully malignant phenotype. Experimental procedures Cells and culture methods
Primary cultures of ROSE cells were grown on glass coverslips as described previously (Adams and Auersperg 1981)from explants of 2- to 3-month-old rat ovaries. The origin of the immortalized nontumorigenic cell lines, ROSE 199 and ROSE 239, and of the KiMSV-transformed tumorigenic ROSE-derived cell lines, V197-15A and V197-10B, has been described (Adams and Auersperg 1981). All cells were maintained in Wayrnouth medium MB 752/1, with 100 IU of penicillin G and 10 pg streptomycin/mL, supplemented with 25% fetal bovine serum for primary cultures, and 10% fetal bovine serum for the cell lines (Grand Island Biological Co., Grand Island, NY). The cell lines were subcultured with 0.125% trypsin (1:250) in calcium and magnesium-free Hanks' balanced salt solution (BSS) (GIBCO). For immunofluorescence and autoradiography, immortalized and transformed cells were grown on glass coverslips for 3 days and primary ROSE explants were cultured for 1 week. Fibroblast controls were cultured from rat thigh muscle fascia and treated like the primary ROSE explants. Immunofluorescence microscopy
Cells grown on coverslips were rinsed in phosphate-buffered saline (PBS), then fixed in - 20°C methanol for 20 min, and stored in PBS at 4°C. They were stained for immunofluorescence by the method of O'Guin et al. (1985), after which the coverslips were mounted on glass slides with gelvatol (Rodrighuez and Deinhardt 1960). Fibroblasts served as positive controls for vimentin and as negative controls for keratin. Antibodies used were the following: rabbit antiserum prepared against human epidermal (callus) keratin ("RAK") (Sun and Green 1978);goat anti-rabbit (heavy and light chain specific) IgG fraction. FITC-conjugated (Miles Laboratories Inc., Elkhart, IN); mouse monoclonal anti-vimentin antibody (MAV) (Labsystems, Helsinki, Finland); FITC-conjugated goat anti-mouse IgG (GAM) (Meloy Laboratories, Inc., Springfield, Virginia); rhodamine-conjugated goat anti-rabbit IgG (Miles Laboratories, Inc., Elkhart, IN). The "RAK" wide-spectrum, nonsubtype-specific antibody to human keratin was chosen because
Biochem. Cell Biol. Downloaded from www.nrcresearchpress.com by YORK UNIV on 11/10/14 For personal use only.
18
BIOCHEM. CELL
BIOL. VOL.
70. 1992
cellulose was incubated in TTBS containing biotinylated goat antirabbit Ig G + M (Clontech, Palo Alto, CA) at room temperature for 1 h. The filter was transferred to TTBS containing the avidinbiotinylated horseradish peroxidase complex (Clontech, Palo Alto. CA) and incubated at room temperature for 30 min with gentle agitation. After further washing, the keratin bands were demonstrated with 4-chloro-1-naphthol, prepared fresh by dissolving 60 mg in 20 mL of ice-cold methanol and mixing with 100 mL of TBS containing 30 mg urea hydrogen peroxide. The reaction was stopped in distilled water and the bands were photographed. The nitrocellulose was later stained with amido black (0.1% in methanol - acetic acid - water, 45:10:45) to locate molecular weight markers. In addition to the molecular weight markers, the known keratin profile of the human carcinoma cell line, C-41, (Auersperg et al. 1989), was used as a reference.
FIG. 1. (a) Morphology and (6) keratin expression of ROSE cells in primary culture. (a) Phase microscopy. x 170. (b) Keratin. Immunofluorescence microscopy. x 220. it is known to react with rat keratins. To our knowledge, no wide spectrum antibodies directed against rat keratins (either mono- or poly-clonal) are available. Furthermore, subtype-specific antikeratin antibodies, such as the AE-1 and AE-3 antibodies used widely for human material, react with rat cells by immunofluorescence, but not on Western blots (unpublished observations). Some cultures were double-stained to determine if both types of intermediate filaments were present in the same cell, and to compare their distribution within cultures and within individual cells. As controls, cultures were stained with a second antibody only, and with the preimmune serum replacing the first antibody. The cells were examined with a Zeiss photomicroscope 11, equipped with epifluorescence optics. Autoradiography For each cell type, one group of cultures was treated with 1 pCi [ 3 ~ ] t h y m i d i n e / m(methyL3H, ~ 78.0 Ci/mmol; 1 Ci = 37 GBq) (Dupont NEN Research Products) for 2 h, then fixed in - 20°C methanol. A second group of cultures was treated with 1 pCi [3~]thymidine/mmolfor 24 h, then fixed in -20°C methanol. Subsequently, all [3~]thymidine-treatedcells were stained for keratin by the procedure described above. The coverslips were then mounted with glycerine jelly on glass slides, cell side up, keeping the cells moist with PBS at all times. The slides were dipped in emulsion, allowed to dry, then stored in a light-proof box at 4°C for 1 week. They were developed with Dl9 (Kodak) and fixed with regular Kodak fixer. Then, the coverslips were removed from the slides, rinsed with PBS, and mounted cell-side down on clean glass slides with gelvatol. The slides were examined blindly. They were scanned visually to define areas with the most representative (common) staining pattern. Fields within these areas were photographed and used for analysis. Immunoblotting Preparations of enriched keratins (Achtstaetter et al. 1986) were obtained from cultures that were confluent but not overgrown. They were subjected to SDS-PAGE in 8.5% minigels (Laemmli 1970) using the high ionic strength buffer system of Thomas and Kornberg (1975). Enriched keratin (20-30 pg, as estimated by OD?, .,)was loaded in each lane. After electrophoretic transfer to nitrocellulose (Towbin et al. 1979), the keratins were tested with the rabbit polyclonal antibody RAK to human callus (Sun and Green 1978). The reasons for this choice of antibody are described under Immunofluorescence microscopy. The nitrocellulose was blocked with TTBS plus 3% BSA and 1% NGS for 1 h at room temperature and then incubated with RAK antibody (1:1000) in TTBS plus 1% BSA overnight at 4°C. After washing, the nitro-
Results ROSE cells in primary cultures were easily recognized by their distinct cobblestone morphology- d dams and Auersperg 1981). Keratin filaments were abundant throughout the cytoplasm of all these cells, often at a higher concentration surrounding the nucleus (Figs. 1 and 2). Only few cells, at the periphery of colonies, expressed low levels of vimentin, but most cells were vimentin negative (Fig. 3). The cell line ROSE 239 is immortal, but it is indistinguishable from low passage ROSE cultures in cell morphology and intercellular organization, and it remains monolayered when crowded (Auersperg et al. 1991). These cells had a variable and markedly reduced keratin content compared with the primary ROSE cells. Only about 5-30% of cells expressed keratin, which ranged from one or a few granules or short fibrils in the majority of keratin-positive cells, to near-normal staining in some (Tables 1 and 2; Figs. 2 and 6). All ROSE 239 cells were vimentin positive (Fig. 3). The second immortal cell line, ROSE 199, is more atypical than ROSE 239 (Auersperg et al. 1991). When sparse, the cells are monolayered and epithelial, but in densely grown cultures, they continue to proliferate rapidly, forming multilayers and ridges (Adams and Auersperg 1985). Up to 50% of these cells in sparse cultures, and up to 95% in crowded cultures, were keratin negative or contained very sparse granules. The remainder contained more extensive, short fibrils (Tables 1 and 2; Figs. 2 and 6). Small groups of intensely keratin-positive ROSE 199 cells were located on the multilayered ridges that appeared in crowded cultures of this line (Fig. 4). All ROSE 199 cells contained vimentin filaments, with a staining pattern resembling ROSE 239 cultures (Fig. 3). In the KiMSV-transformed lines, the proportion of keratin-positive cells exceeded that of the immortal lines (Tables 1 and 2), but staining intensities were generally reduced. V197-15A cultures contained two distinct cell types (Figs. 2 and 6). There were small round cells, some with long extensions; these cells stained intensely for keratin in the perinuclear region. However, most cells were large, flat, and stained faintly or not at all. The V197-10B cells were large, flat, and more uniform. Most of the cells that stained positively in this line contained only fine, weakly staining keratin filaments (Figs. 2 and 6). All cells in both transformed lines contained extensive fibrillar networks of vimentin, which in staining intensity and distribution resembled the vimentin pattern of fibroblasts (Fig. 3). In both immortal lines, keratin expression was inversely proportional to cell density (Tables 1 and 2). There were some exceptions to this overall pattern. This was evident at the
Biochem. Cell Biol. Downloaded from www.nrcresearchpress.com by YORK UNIV on 11/10/14 For personal use only.
HORNBY ET AL.
19
FIG. 2. Keratin in ROSE cells. Immunofluorescence microscopy. (a) Primary culture. All cells are strongly fluorescent. (6)Line ROSE V197-15A. (c) Line ROSE 239. Note that the whole field is covered with cells. The fluorescence of 3 cells approaches that of primary cultures; in the others, fluorescence is either reduced or absent. ( d ) Line ROSE V197-10B. Arrowheads point to fine filaments. (e) Line ROSE 199. An area with predominantly keratin-positive cells. ( f ) Normal serum control to Fig. 2e. x 300.
periphery of colonies where the keratin staining of positive cells was particularly intense, but there were also some keratin-negative cells. Similarly, in areas of uniform cell density within colonies, keratin-positive and -negative cells sometimes occurred side by side in distinct groups (Fig. 2b). Finally, the multilayered ridges formed by ROSE 199 cells in very densely grown areas were intensely keratin positive. There was no clear-cut interrelationship between density and intermediate filament expression in the virally transformed lines. Double staining showed different distributions of keratin and vimentin within individual cells (data not
shown). Since the immortal and transformed cells were 100% vimentin positive, keratin and vimentin were coexpressed in all keratin-positive cells. Immunoblotting revealed identical keratin subtypes, of 53, 55, 57, 60, 62, and 66 kDa, in primary cultures and in both immortal lines (Fig. 5). The virus-transformed line V197-15A lacked the 53-kDa keratin and line V197-1OB lacked the 53-, 60-, 62-, and 66-kDa keratins (Fig. 5). To determine whether the variation in keratin content among the immortal and transformed ROSE cells was related to their proliferative activity, cells were first treated
Biochem. Cell Biol. Downloaded from www.nrcresearchpress.com by YORK UNIV on 11/10/14 For personal use only.
20
BIOCHEM. CELL BIOL. VOL. 70,
1992
FIG. 3. Vimentin in ROSE cells. Immunofluorescence microscopy. (a) Primary culture. (b) Line ROSE V197-15A. (c) Line ROSE 239. (d) Line ROSE V197-10B. (e) Line ROSE 199. ( f ) Fibroblasts. x 300.
with tritiated thymidine either for 2 h to identify cells in S phase (Table l), or for 24 h to identify cycling populations (Table 2), and were then stained for keratin (Fig. 6). The tables list the percentage of cells that are either keratin positive or thymidine labelled, as well as the percentage that are positive for both markers. A comparison of the observed percentage of cells that expressed keratin and also incorporated thymidine to the percentage expected if keratin expression and thymidine incorporation occurred independently, demonstrated that these characteristics were not interdependent in the immortal lines. Microscopic analysis revealed one exception to this conclusion: the highly keratin-
positive cells that covered the ridges in dense ROSE 199 cultures were consistently nonproliferative. As these structures are quite rare, they did not affect the statistical data based on large cell numbers (Tables 1 and 2). In the virustransformed lines, the relationships among keratinexpressing cells that were thymidine labelled for 2 h resembled those of the immortal lines, although after the 24-h pulse, the percentage of keratin-positive cells that were dividing was less than expected. Analysis of these preparations confirmed the visual impression of an inverse relationship between cell density and keratin expression in the immortal lines, but not in the virus-transformed lines.
21
Biochem. Cell Biol. Downloaded from www.nrcresearchpress.com by YORK UNIV on 11/10/14 For personal use only.
HORNBY ET AL.
FIG. 4. Line ROSE 199. Keratin-positive ridge next to keratin-poor monolayer. (a) Phase microscopy. (b) Immunofluorescence microscopy. x 300.
Discussion In our study, all the cell types examined, with the exception of cells in primary culture, were vimentin positive and showed similar patterns of vimentin filament distribution. The staining intensity increased from primary to immortalized to virally transformed cells, but was fairly uniform among cells at the same stage of neoplastic progression. Thus, at the level of immunofluorescence microscopy, neoplastic progression and major changes in cell morphology and growth pattern did not alter the intracellular vimentin distribution. but did cause auantitative changes in the expression o f this type of- intermediate filament. Interestingly, it has been proposed that vimentin may be a DNA-binding protein and that the polymerized cytoplasmic filaments may be its storage form (Traub 1985). In this context, the increase in vimentin filaments observed during neoplastic conversion in our system, as well as clinical correlations with cancer progression (Cattoretti et al. 1988; Sommers et al. 1989), raise the possibility that some of the changes observed may have a gene regulatory rather than a structural basis. In contrast to the relative uniformity of vimentin expression, neoplastic progression in ROSE cells was associated with a striking intercellular variability in keratin expression at the microscopic level, which could not be attributed to variations in proliferative activity, cell shape, or intercellular organization. The immortalized cell lines, ROSE 199 and ROSE 239, contained fewer keratin-positive cells than the primary cultures. The keratin staining in the positive cells was reduced in intensity and had an altered microscopic appearance. It is particularly interesting that the two immortal cell lines expressed similar abnormalities in keratin expression microscopically as well as identical keratin subtypes by Western blot, though the two lines have very different growth characteristics: line ROSE 239, but not line ROSE 199, resembles primary OSE cultures morphologically. Thus, the maintenance of the epithelial phenotype that characterizes ROSE in primary culture does not require the presence of the normal complement of keratin filaments and persists in the presence of amplified vimentin. It was surprising that the profound microscopic defects in keratin
kDa
1
2
3
4
5
6,,
,66-
@: 53
w
,3 -48 44 -40
-
b
FIG. 5. Western blot showing keratin profile of ROSE cells (lanes 1-5). Lane 1, ROSE in primary culture; lane 2, ROSE 239; lane 3, ROSE 199; lane 4, ROSE V197-15A; lane 5, ROSE V197-10B; lane 6, human C-41 cells, known to express 40-, 44-,48-, 53-, 55-, and 58-kDa keratins (Auersperget al. 1989). were used as a molecular mass marker. Partially enriched keratin, (20-30 pg, estimated by OD,,, m) was loaded in each lane. Molecular mass markers at 36, 44, and 65 kDa are indicated by arrowheads.
expression observed in the two immortal cell lines coincided with the expression of all keratin subtypes detected in primary cultures, as indicated by immunoblotting. It is impossible to distinguish among the many abnormalities that may form the basis for the reduced keratin fluorescence and altered patterns of staining in these cells. Among them are changes in the amounts or ratios of keratin subunits available for polymerization. Alternatively, the defects could result from a loss of (conformational) epitopes, e.g., through changes in keratin monomers at the molecular level, that would interfere with normal filament assembly and prevent detection by immunofluorescence techniques (Miller et al. 1991). In addition, altered nonkeratinous proteins might interfere with the normal cytoplasmic distribution and (or) antigenicity of these intermediate filaments. In the virally transformed cell lines, the proportion of keratin-positive cells was increased compared with the immortalized cell lines, though keratin staining per cell was generally reduced. A similar increase in the proportion of keratin-expressing cells occurs in tumorigenic derivatives of
Biochem. Cell Biol. Downloaded from www.nrcresearchpress.com by YORK UNIV on 11/10/14 For personal use only.
22 BIOCHEM. CELL BIOL. VOL. 70, 1992
23
HORNBY ET AL.
TABLE1. Percentage of ROSE cells positive for keratin and (or) [3~]thymidine (['~lth~midine label for 2 h)
TABLE 2. Percentage of ROSE cells positive for keratin and (or) ['~lthymidine(['Hlthymidine label for 24 h)
K + 'H
K + 3~ Line
No. of cells/field -
ROSE 239
Biochem. Cell Biol. Downloaded from www.nrcresearchpress.com by YORK UNIV on 11/10/14 For personal use only.
ROSE 199 V197-10B V197-15A
-
K
'H
25 31 11 46 21 100 82 42 81 78 53
23 27 17 27 39 16 24 23 22 13 8
Actual Exp. Neg.
Line
No. of cells/field
ROSE 239
< 100 200-400 600-700 < 100 200-400 800-900 < 100 < 100 300-400 < 100 < 100 < 100
K
'H
Actual Exp. Neg.
-
100-200 200-400 600-700 200-400 700-800 < 100 200-300 200-300 < 100 < 100 300-400
5 6 2 9 6 16 14 4 14 0 1
6 8 2 12 8 16 20 10 18 10 4
58 51 76 35 46 0 9 40 11 9 39
NOTE:Cultures of immortal, nontumorigenic cell lines (ROSE 239 and ROSE 199) and of KiMSV-transformed, tumorigenic~elllines (V197-10B and V197-15A) were labelled with ['Hlthymidine for 2 h and then stained by immunofluorescence for keratin. In representative microscopic fields, the total number of cells, and the cells positive for keratin (K), ['H]thymidine ('H), keratin and ['Hlthymidine (K+ 'H), and neither marker (Neg.) were counted and ranked by cell density. (Note that the K and 'H categories include the cells that are listed under K + 'H, in addition to the cells expressing only one marker). The column headed Actual refers to the percentage of cells observed in the K + 'H category; the column headed Exp. (expected) refers to the percent of cells in the K + 'H category, expected on the basis of calculations from the observed K and 'H values (expected = Yo 'H positive x % keratin positive / 100). Results are the means of four microscopic fields.
nontumorigenic rat liver lines (Troyanovsky et al. 1986), and increases in differentiation as a result of KiMSV-mediated transformation have been observed in other types of cells (Auersperg et al. 1990). The altered appearance of keratin filaments in the transformed lines may be related to the results of the Western blot analyses, in which there appeared to be a loss of specific keratin subtypes compared with primary and immortal cultures. The available data do not allow us to distinguish possible changes at the level of gene regulation from alterations in antigenicity based on mutations as the basis for the observed abnormalities in filament numbers and subtype losses. Together, the data suggest that immortalization affects the uniformity of keratin expression among cells, the physical characteristics of keratins, and possibly, the amounts of keratin present, but that the expression of tissue-specific keratin subtypes by ROSE cells is maintained until further steps in neoplastic progression take place. The relatively high molecular masses of ROSE-derived keratins differ from the keratin profile of human simple epithelia, which usually ranges from 40 to 55 kDa. It is possible that this is a difference between species, because a similar discrepancy has been observed between liver keratins of humans and rat (Moll et al. 1982; Langbein and Neupert 1986). In two analyses of rat mesothelial keratins (MacKay et al. 1990; Troyanovsky et al. 1986), molecular masses were reported to range from 40 to 55 kDa; however, it is not clear whether the antibodies used in these studies recognize the same subtypes as the antibody used here. Many epithelia that are vimentin negative in vivo acquire this cytoskeletal component in culture (Connell and Rhein-
ROSE 199 V197-10B V197-15A
NOTE:Cultures of immortal, nontumorigenic cell lines (ROSE 239 and ROSE 199) and of KiMSV-transformed, tumorigenic cell lines (V197-10B and V197-ISA) were labelled with ['H]thymidine for 24 h and then stained by irnrnunofluorescence for keratin. In representative microscopic fields, the total number of cells, the cells positive for keratin (K), ['H~thymidine ('H), keratin and ['~lthymidine(K + 'H), and neither marker (Neg.) were counted and ranked by cell densit . (Note that the K and 'H categories include the cells that are listed under K + H, in addition to the cells expressing only one marker). The column headed Actual refers to the percentage of cells observed in the K + 'H category; the column headed Exp. (expected) refers to ed on the basis of calculations from the percent of cells in the K + 'H category, ex .. x % keratin positive / 100). the observed K and 'H values (expected = % H poslhve Results are the means of four microscopic fields.
Y .
F
wald 1983; Franke et al. 1981; Ben Ze'ev 1985; Langbein and Neupert 1986). Our results show that this group of tissues includes rat ovarian surface epithelial cells. Under certain conditions, keratin and vimentin vary in a reciprocal fashion; in cultures of human mesothelial cells (Connell and Rheinwald 1983), vimentin expression is enhanced and keratin expression reduced in response to epidermal growth factor (EGF), concommittantly with shape changes and enhanced proliferation. In a series of cell lines, Ben Ze'ev (1985) demonstrated that the expression of vimentin decreased with increasing cell density and that the more immediate signals for this response were density-dependent changes in cell shape. On the other hand, keratin expression in MDCK cells was proportional to cell density, and the signal here appeared to involve intercellular contact rather than shape changes. In ROSE cultures, keratin was expressed in all primary cells independently of density, but it was reduced at high densities in the immortal cell lines. The difference between this response and that of other epithelial cell types (Ben Ze'ev 1985; Troyanovsky et al. 1986) may be related to differences in the nature of the intercellular contacts between, for example, the highly polar MDCK cells and the less polar ROSE lines. The lack of a response to density by the virally transformed ROSE cells is in keeping with the increased autonomy of fully transformed cells. Keratin content was inversely proportional to cell density in all regions of immortal lines except in the multilayered ROSE 199 ridges. Possibly, cells in these specialized regions
FIG. 6. Relationship of keratin expression to thymidine incorporation. (a and b) ROSE 239. (c and d) ROSE V197-10B. (e and f ) ROSE 199. (g and h) ROSE V197-15A. (a, c, e, and g) Immunofluorescencemicroscopy. Autoradiographs after [H3]thymidinelabelling for 2 (b and d) and 24 h (f and h). Large arrowheads, proliferating cells that are keratin positive; small arrowheads, proliferating cells that are keratin negative. In the field shown, essentially all V197-10B cells are keratin positive (c); most proliferating V197-15A cells are keratin negative (g and h). x 170.
Biochem. Cell Biol. Downloaded from www.nrcresearchpress.com by YORK UNIV on 11/10/14 For personal use only.
24
BIOCHEM. CELL BIOL. VOL. 70, 1992
express different subtypes of keratins, as has been described in similar regions of hepatocyte cultures (Troyanovsky et al. 1986). Other minor inconsistencies in the inverse correlation between density and keratin expression by immortal ROSE cells resemble the unexplained heterogeneities in keratin patterns that have been reported in cultures of rat liver cells (Langbein 1986). In our system, such variation seemed unrelated to proliferative activity. In transformed rat liver lines, local changes in cell-substratum adhesion have been suggested as a possible explanation (Troyanovsky et al. 1986). Histologically, tumors produced by the two KiMSVtransformed cell lines examined here resembled human endometrioid stromal sarcomas (Adams and Auersperg 1981), which are a minor but nevertheless important subclass of the "Common Epithelial Tumors" category in the WHO classification of ovarian tumors (Scully 1977). In the past, it has been difficult to decide whether the mesodermal phenotypic characteristics of these neoplasms are due to the multipotential character of the mesodermally derived OSE, or to an origin in stromal cells (Adarns and Auersperg 1981). Our demonstration of keratin in the V197-10B and V19715A lines supports the former hypothesis. The alterations in intermediate filaments that accompanied KiMSVmediated transformation of ROSE cells, viz, a drastic reduction and qualitative alterations in keratin expression and increased vimentin expression, differ from the changes reported in many of the human ovarian serous and mucinous carcinomas, which tend to retain the keratin subtypes of normal OSE and express variable amounts of vimentin (Czernobilsky et al. 1984). The phenotype of the virally transformed rat cells may be related to the oncogenic agent used in this study. Ras is one of the most commonly activated protoncogenes in human carcinomas. However, its retroviral form, KiMSV, has a propensity to form sarcomas, a phenotype that would be compatible with a low keratin high vimentin profile. The results of this study demonstrate that normal keratin expression is not necessary to maintain the normal morphology and growth pattern of OSE cells. They show further that keratin expression is drastically altered by immortalization and viral transformation. The variability of these changes supports the impression gained from clinical histopathology that keratin expression may have little direct influence on neoplastic progression. On the other hand, the progressive and relatively uniform increase in vimentin expression with neoplastic progression raises the possibility that a more direct interrelationship between these processes may exist. Acknowledgements This research was supported by grants and a research associateship to N. Auersperg from the National Cancer Institute of Canada and the Medical Research Council of Canada. We thank Dr. Sigrid E. Myrdal and Ms Vera Lee for their advice and help. Achtstaetter, T., Hatzfeld, M., Quinland, R.A., Parmelee, D.C., and Franke, W.W. 1986.Separation of cytokeratin polypeptides by gel electrophoretic and chromatographic techniques and their identification by immunoblotting. Methods Enzymol. 134: 355-371.
Adams, A.T., and Auersperg, N. 1981.Transformation of cultured rat ovarian surface epithelial cells by Kirsten murine sarcoma virus. Cancer Res. 41: 2063-2072. 1985. A cell line, ROSE 199, derived from normal rat ovarian surface epithelium. Exp. Cell Biol. 53: 181-188. Auersperg, N., Kruk, P.A., MacLaren, I.A., Watt, F.M., and Myrdal, S.E. 1989. Heterogeneous expression of keratin, involucrin, and extracellular matrix among subpopulations of a poorly differentiated human cervical carcinoma: possible relationships to patterns of invasion. Cancer Res. 49: 3007-3014. Auersperg, N., MacLaren, I.A., and Bissell, M.J. 1990. V-K-ras transformation induces reversion to an earlier developmental form in adult rat adrenal cells. Differentiation, 43: 29-36. Auersperg, N., MacLaren, I.A., and Kruk, P.A. 1991. Ovarian surface epithelium: autonomous production of connective tissuetype extracellular matrix. Biol. Reprod. 44: 717-724. Benjamin, E., Law, S., and Bobrow, L.G. 1987. Intermeditite filaments cytokeratin and vimentin in ovarian sex cord-stromal tumours with correlative studies in adult and fetal ovaries. J. Pathol. 152: 253-263. Ben Ze'ev, A. 1985. Cell-cell interaction and cell configuration related control of cytokeratins and vimentin expression in epithelial cells and in fibroblasts. Ann. N.Y. Acad. Sci. 455: 597-613. 1986.The relationship between cytoplasmic organization, gene expression and morphogenesis. TIBS, 11: 478-481. Cattoretti, G., Andreola, S., Clemente, C., D'amato, L., and Rilke. F. 1988.Vimentin and p53 expression on epidermal growth factor receptor-positive, oestrogen receptor-negative breast carcinomas. Br. J. Cancer, 57: 353-357. Connell, N.D., and Rheinwald, J.G. 1983. Regulation of the cytoskeleton in mesothelial cells: reversible loss of keratin and increase in vimentin during rapid growth in culture. Cell, 34: 245-253. Czernobilsky, B. 1985.Co-expression of cytokeratin and vimentin filaments in mesothelial, granulosa and rete ovarii cells of the human ovary. Eur. J. Cell Biol. 37: 175-190. Czernobilsky, B., Moll, R., and Franke, W.W. 1984.Intermediate filaments of normal and neoplastic tissues of the female genital tract with emphasis on problems of differential tumor diagnosis. Pathol. Res. Pract. 179: 31-37. Dabbs, D. J., and Geisinger, K.R. 1988.Common epithelial ovarian tumors: immunohistochemical intermediate filament profiles. Cancer, 62: 368-374. Franke, W.W., Mayer, D., Schmid, E., Denk, H., and Borenfreund, E. 1981. Differences of expression of cytoskeletal proteins in cultured rat hepatocytes and hepatoma cells. Exp. Cell Res. 134: 345-365. Goldman, R., Goldman, A., Green, K., Jones, J., Lieska, N., and Yang, H.-Y. 1985. Intermediate filaments: possible functions as cytoskeletal connecting links between the nucleus and the cell surface. Ann. N.Y. Acad. Sci. 455: 1-17. Laemmli, U.K. 1970. Cleavage of structural proteins during the assembly of the head of bacteriophage T4. Nature (London), 227: 680-685. Langbein, L., and Neupert, G. 1986. Modulation of expression of intermediate filaments during the development of established rat liver cell lines. Acta Histochem. 80: 149-158. MacKay, A.M., Tracy, R.P., and Craighead, J.E. 1990. Cytokeratin expression in rat mesothelial cells in vitro is controlled by the extracellular matrix. J. Cell Sci. 95: 97-107. Miller, R.K., Vikstrom, K., and Goldman, R.D. 1991. Keratin incorporation into intermediate filaments is a rapid process. J. Cell Biol. 113: 843-855. Moll, R., Franke, W.W., and Schiller, D.L. 1982. The catalog of human cytokeratins: patterns of expression in normal epithelia, tumors and cultured cells. Cell, 31: 11-24. Moll, R., Krepler, R., and Franke, W.W. 1983a. Complex
-
Biochem. Cell Biol. Downloaded from www.nrcresearchpress.com by YORK UNIV on 11/10/14 For personal use only.
HORNBY ET AL.
cytokeratin polypeptide patterns observed in certain human carcinomas. Differentiation, 23: 256-269. Moll, R., Levy, R., Czernobilsky, B., Hohlweg-Majert. P.. Dallenbach-Hellweg, G., and Franke, W. W. 1983b. Cytokeratins of normal epithelia and some neoplasms of the female genital tract. Lab. Invest. 49: 599-610. Nicosia, S.V., and Nicosia, R.F. 1988. Neoplasms of the ovarian mesothelium. In Pathology of human neoplasms. Edited by H.A. Azar. Raven Press, Ltd., New York. pp. 435-486. O'Guin, W.M., Schermer, A., and Sun, T.-T. 1985. Immunofluorescence staining of keratin filaments in cultured epithelial cells. J. Tissue Cult. Methods. 9: 123-128. Robson, R.M. 1989. Intermediate filaments. Curr. Opin. Cell Biol. 1: 36-43. Rodrighuez, J., and Deinhardt. F. 1960. Preparation of a semipermanent mounting medium for fluorescent antibody studies. Virology, 12: 3 16-320. Scully. R.E. 1977. Ovarian tumors. Am. J. Pathol. 87: 686-720. Sommers, C.L., Walker-Jones, D., Heckford, S.E., Worland, P., Valverius, E., Clark, R., McCormick, F., Stampfer, M., Abularach, S., and Gelmann, E.P. 1989. Vimentin rather than keratin expression in some hormone-independent breast cancer cell lines and in oncogene-transformed mammary epithelial cells. Cancer Res. 49: 4258-4263.
25
Steinert, P.M., and Roop, D.R. 1988. Molecular and cellular biology of intermediate filaments. Annu. Rev. Biochem. 57: 593-625. Stewart, M. 1990. Intermediate filaments: structure, assembly and molecular interactions. Curr. Opin. Cell Biol. 2: 91-100. Sun, T.-T., and Green, H. 1978, Immunofluorescent staining of keratin fibers in cultured cells. Cell, 14: 469-476. Thomas, J.O., and Kornberg, R.D. 1975. An octamer of histones in chromatin and free in solution. Proc. Natl. Acad. Sci. U.S.A. 72: 2626-2630. Towbin, H., Staehelin, T., and Gordon, J. 1979. Electrophoretic transfer of proteins from polyacrylamide gels to nitrocellulose sheets: procedure and some applications. Proc. Natl. Acad. Sci. U.S.A. 76: 4350-4354. Traub, P. 1985. Are intermediate filament proteins involved in gene expression? Ann. N.Y. Acad. Sci. 455: 68-78. Troyanovsky, S.M., Bannikov, G.A., Montesano, R., and Vasiliev, J.M. 1986. Density-dependent expression of keratins in transformed rat liver cell lines. Cell Biol. Int. Rep. 10: 263-270. Viale, G., Gambacorta, M., Dell'Orto, P., and Coggi, G. 1988. Coexpression of cytokeratins and vimentin in common epithelial tumours of the ovary: an immunocytochemical study of eightythree cases. Virchows Archiv. A, Pathol. Anat. 413: 91-101.
Biochem. Cell Biol. Downloaded from www.nrcresearchpress.com by YORK UNIV on 11/10/14 For personal use only.
ANNE. HORNBY,' JIE PAN,AND NELLYAUERSPERG~ Department of Anatomy, University of British Columbia, Vancouver, B.C., Canada V6G 2W6 Received July 26, 1991 HORNBY,A. E., PAN, J., and AUERSPERG, N. 1992. Intermediate filaments in rat ovarian surface epithelial cells: changes with neoplastic progression in culture. Biochem. Cell Biol. 70: 16-25. Interrelationships between neoplastic progression and the expression of intermediate filaments were examined in primary cultures, immortal lines, and Kirsten murine sarcoma virus (KiMSV) transformed lines of rat ovarian surface epithelial (ROSE) cells. Immunofluorescence microscopy revealed abundant keratin filaments in all cells of primary cultures. In immortal, nontumorigenic lines, keratin filaments were detected in fewer cells, in smaller numbers, and in microscopically altered forms. The percentage of keratin-positive cells ranged from 4 to 54%. Its expression was inversely proportional to cell density. Keratin expression was similar in the two immortal lines, although one had retained a monolayered epithelial growth pattern resembling primary cultures, while in the other the growth pattern of the cells was more atypical. The two KiMSV-transformed lines were previously shown to produce tumors in vivo that resemble human ovarian endometrioid stromal sarcomas. In spite of this histologic appearance, the proportion of keratin-positive cells in these cells was increased over the immortal lines. Keratin expression was unrelated to cell density, and keratin in most virally transformed cells was limited to few, fine filaments. In thymidine-labelled immortal and virus-transformed cultures stained for keratin, no correlation was found between keratin expression and proliferative activity. The keratin profiles of primary and immortal cultures were identical on Western blots, with subtypes ranging from 52 to 66 kDa. The two virally transformed lines lacked some of the subtypes. Vimentin networks were faint or absent in primary cultures. In the immortal and the virus-transformed lines, neoplastic progression was associated with increasing vimentin expression but with no changes in filament morphology and distribution. The results show that the abnormalities in intermediate filament expression that accompany immortalization do not preclude the retention of a normal epithelial morphology and growth pattern in this cell type. Furthermore, the number of intermediate filaments and their intracellular distribution appear to be altered at an earlier stage in neoplastic progression than those mechanisms that select for specific keratin subtypes, or those that respond to regulation by cell density. Finally, the presence of keratin in the KiMSV-transformed lines examined in this study supports the hypothesis that human ovarian stromal sarcomas can arise in the OSE. Key words: neoplastic transformation, keratin, vimentin, cytoskeleton, ras oncogene, ovary, tissue culture. HORNBY,A. E., PAN, J., et AUERSPERG, N. 1992. Intermediate filaments in rat ovarian surface epithelial cells: changes with neoplastic progression in culture. Biochem. Cell Biol. 70 : 16-25. Nous avons examink les relations entre la progression nkoplasique et l'expression des filaments intermtdiaires dans des cultures primaires, des ligntes immortelles et des ligntes transformtes par le virus du sarcome murin de Kirsten (KiMSV) des cellules tpithkliales de la surface des ovaires de rate (ROSE). La microscopie par immunofluorescence rkvtle d'abondants filaments de kkratine dans toutes les cellules des cultures primaires. Dans les lignkes immortelles, non tumorigtnes, les filaments de kkratine sont dktectts dans peu de cellules, en plus petit nombre et sous des formes microscopiquement modifikes. Le pourcentage des cellules rkagissant positivement A la ktratine varie de 4 B 54%. Son expression est inversement proportionnelle 11 la densitk cellulaire. L'expression de la ktratine est similaire dans les deux ligntes immortelles; l'une a retenu un profil de croissance kpithtliale monolamellaire ressemblant 11 celui des cultures primaires tandis que dans l'autre, le profil de croissance des cellules est plus atypique. I1 a dkjh tte montrt que les deux lignkes transformkes par le KiMSV produisent des tumeurs in vivo qui ressemblent aux sarcomes stromatiques endomktrioides des ovaires humains. En dtpit de cette apparence histologique, la proportion des cellules ktratine-positives dans ces cellules augmente davantage que dans les lignkes immortelles. I1 n'existe aucune relation entre I'expression de la kkratine et la densitt cellulaire et, dans la plupart des cellules transformkes par le virus, la ktratine est limitte un petit nombre de fins filaments. Dans les cultures immortelles et les cultures transformtes par le virus, marqutes a la thymidine et colorkes pour la ktratine, nous n'avons dtcelk aucune corrklation entre l'expression de la kkratine et l'activitt prolifkratrice. Les profils de la kkratine des cultures primaires et des cultures immortelles sont identiques sur les transferts Western; on y voit des sous-types allant de 52 B 66 kDa. Les deux ligntes transformkes par le virus ne posstdent pas de tels sous-types. Dans les cultures primaires. la rtseaux de vimentine sonr absents ou a peine visibles. Dans les lignks immortelles et les ligntes transformkes par le virus, la progression ntoplasique est associCe h l'expression accrue de la vimentine, mais non aux changements dans la rnorphologie et la distribution des filaments. Les rCsultats montrent que les anomalies de l'expression des filaments intermMiaires qui accompagnent I'irnmortalisation n'empechent pas le maintien d'une morphologie et d'un profil de croiss;mce ipithtliales normalcs dans ce type de cellules. De plus, le nombre de filaments intermMiaires et leur distribution intracelIulaire semblent se modifier plus tat dans la progression ntoplasique que ces mkcanismes qui veillent au choix des sous-types sptcifiques de ktratjne ou ceux qui rkpondent ABBREVIATIONS: OSE, ovarian surface epithelium; ROSE, rat ovarian surface epithelium; BSS, balanced salt solution; EGF, epidermal growth factor; PBS, phosphate-buffered saline; NGS, normal goat serum; SDS-PAGE, sodium dodecyl sulfate - polyacrylamide gel electrophoresis; TBS, 20 mM Tris-HC1 buffer (pH 7.5); TTBS, TBS containing 0.05% Tween-20; KiMSV, Kirsten murine sarcoma virus. 'present address: Department of Medical Genetics, University of British Columbia, 6174 University Boulevard, Vancouver. B.C. V6T 123. ' ~ u t h o rto whom correspondence should be addressed. Printed in Canada / lrnprirnt au Canada
17
HORNBY ET AL.
Biochem. Cell Biol. Downloaded from www.nrcresearchpress.com by YORK UNIV on 11/10/14 For personal use only.
a la rtgulation par la densitt cellulaire. Finalement, la presence de ktratine dans les ligntes transformtes par le KiMSV examintes dans cette ttude confirme I'hypothbe que les sarcomes stromatiques ovariens humains peuvent se produire dans les cellules tpithtliales a la surface des ovaires de rate. Mots clds : progression ntoplasique, ktratine, vimentine, cytosquelette, oncogene ras, ovaire, culture de tissus. [Traduit par la rtdaction]
Introduction The shapes of cells and their contacts with the environment influence and are influenced by the characteristics of the intermediate filaments and, ultimately, they affect cellular growth and differentiation (Ben Ze'ev 1986). The intermediate filaments keratin and vimentin belong to a group of widespread, highly polymorphic, and insoluble cytoskeletal proteins (Goldman et al. 1985; Steinert and Roop 1988). Their precise role in the formation and maintenance of cell shapes and cytoplasmic compartmentization is still unclear. It has been suggested on the basis of their complex distribution among tissues that their functions may be differentiation related rather than "housekeeping" or universal (Robson 1989; Stewart 1990). Transformation to malignancy is accompanied by profound qualitative and quantitative alterations in the expression of several cytoskeletal components, including keratin and vimentin. However, it is not clear whether the cytoskeletal alterations are causes or consequences of specific steps in the carcinogenetic process. To investigate these interrelationships, we have examined the distribution of intermediate filaments at different stages during the neoplastic progression of rat ovarian surface epithelial (ROSE) cells. The expression of keratin and vimentin was compared in normal ROSE in primary culture, and in a series of ROSE-derived cell lines that express different degrees of neoplastic transformation. Two immortal lines arose by spontaneous transformation from primary ROSE cultures. Of these, one line (ROSE 239) is nonturnorigenic and resembles normal cells in primary culture in its cellular morphology, its epithelial monolayered growth pattern, and its growth inhibition by crowding. The second immortal line (Rose 199) is also nontumorigenic. However, its cells shed and multilayer in response to crowding and thus represent a more advanced stage within the immortal/nontumorigenic phase of neoplastic progression (Adams and Auersperg 1985). The distinctive characteristics of these lines have remained stable over many passages in culture. Two additional, fully transformed lines 03197-15A and V197-10B) were obtained by infection of primary ROSE cultures with Kirsten murine sarcoma virus (KiMSV). In addition to immortality, these lines expressed the culture characteristics of fully malignant cells, as well as tumorigenicity (Adams and Auersperg 1981). The ovarian surface epithelium (OSE) is the modified mesothelium that covers the surface of the ovary. OSE is of considerable clinical significance because it gives rise to most human ovarian carcinomas (Nicosia and Nicosia 1988). In spite of its importance, successful transformation to malignancy of human OSE under experimental conditions has not been reported. To our knowledge, the immortal and KiMSV-transformed ROSE lines described here represent the only model currently available for a comparison of stages in experimentally induced neoplastic transformation of OSE from any species. The types of intermediate filaments present in normal OSE vary among species. In the rat, ROSE cells in vivo are keratin positive but lack vimentin (Czernobilsky 1985). In humans, keratin is also present but
vimentin is coexpressed in varying proportions of the cells (Dabbs and Geisinger 1988; Benjamin et al. 1987; Moll et al. 1983a, 19833; Vide et al. 1988; Czernobilsky 1985). Human OSE-derived carcinomas tend to be keratin positive (Czernobilsky et al. 1984), while the expression of vimentin varies and seems to depend on the histotype and the degree of differentiation of these neoplasms. The effects of neoplastic transformation on the intermediate filaments of ROSE have not been described. The results of this study suggest that major alterations in the distribution, staining intensity, and morphology of keratin filaments take place at an early stage in neoplastic progression, while ROSE cells are immortalized but remain morphologically normal and nontumorigenic. The tissue-specific keratin subtypes, as shown by Western blotting, changed only in the tumorigenic, fully transformed lines. In contrast to the complex and variable alterations in keratin expression, vimentin expression increased gradually and uniformly in all cells with progression to malignancy. The results suggest that the mechanisms that regulate intermediate filament expression change with immortalization, and again with progression to the fully malignant phenotype. Experimental procedures Cells and culture methods
Primary cultures of ROSE cells were grown on glass coverslips as described previously (Adams and Auersperg 1981)from explants of 2- to 3-month-old rat ovaries. The origin of the immortalized nontumorigenic cell lines, ROSE 199 and ROSE 239, and of the KiMSV-transformed tumorigenic ROSE-derived cell lines, V197-15A and V197-10B, has been described (Adams and Auersperg 1981). All cells were maintained in Wayrnouth medium MB 752/1, with 100 IU of penicillin G and 10 pg streptomycin/mL, supplemented with 25% fetal bovine serum for primary cultures, and 10% fetal bovine serum for the cell lines (Grand Island Biological Co., Grand Island, NY). The cell lines were subcultured with 0.125% trypsin (1:250) in calcium and magnesium-free Hanks' balanced salt solution (BSS) (GIBCO). For immunofluorescence and autoradiography, immortalized and transformed cells were grown on glass coverslips for 3 days and primary ROSE explants were cultured for 1 week. Fibroblast controls were cultured from rat thigh muscle fascia and treated like the primary ROSE explants. Immunofluorescence microscopy
Cells grown on coverslips were rinsed in phosphate-buffered saline (PBS), then fixed in - 20°C methanol for 20 min, and stored in PBS at 4°C. They were stained for immunofluorescence by the method of O'Guin et al. (1985), after which the coverslips were mounted on glass slides with gelvatol (Rodrighuez and Deinhardt 1960). Fibroblasts served as positive controls for vimentin and as negative controls for keratin. Antibodies used were the following: rabbit antiserum prepared against human epidermal (callus) keratin ("RAK") (Sun and Green 1978);goat anti-rabbit (heavy and light chain specific) IgG fraction. FITC-conjugated (Miles Laboratories Inc., Elkhart, IN); mouse monoclonal anti-vimentin antibody (MAV) (Labsystems, Helsinki, Finland); FITC-conjugated goat anti-mouse IgG (GAM) (Meloy Laboratories, Inc., Springfield, Virginia); rhodamine-conjugated goat anti-rabbit IgG (Miles Laboratories, Inc., Elkhart, IN). The "RAK" wide-spectrum, nonsubtype-specific antibody to human keratin was chosen because
Biochem. Cell Biol. Downloaded from www.nrcresearchpress.com by YORK UNIV on 11/10/14 For personal use only.
18
BIOCHEM. CELL
BIOL. VOL.
70. 1992
cellulose was incubated in TTBS containing biotinylated goat antirabbit Ig G + M (Clontech, Palo Alto, CA) at room temperature for 1 h. The filter was transferred to TTBS containing the avidinbiotinylated horseradish peroxidase complex (Clontech, Palo Alto. CA) and incubated at room temperature for 30 min with gentle agitation. After further washing, the keratin bands were demonstrated with 4-chloro-1-naphthol, prepared fresh by dissolving 60 mg in 20 mL of ice-cold methanol and mixing with 100 mL of TBS containing 30 mg urea hydrogen peroxide. The reaction was stopped in distilled water and the bands were photographed. The nitrocellulose was later stained with amido black (0.1% in methanol - acetic acid - water, 45:10:45) to locate molecular weight markers. In addition to the molecular weight markers, the known keratin profile of the human carcinoma cell line, C-41, (Auersperg et al. 1989), was used as a reference.
FIG. 1. (a) Morphology and (6) keratin expression of ROSE cells in primary culture. (a) Phase microscopy. x 170. (b) Keratin. Immunofluorescence microscopy. x 220. it is known to react with rat keratins. To our knowledge, no wide spectrum antibodies directed against rat keratins (either mono- or poly-clonal) are available. Furthermore, subtype-specific antikeratin antibodies, such as the AE-1 and AE-3 antibodies used widely for human material, react with rat cells by immunofluorescence, but not on Western blots (unpublished observations). Some cultures were double-stained to determine if both types of intermediate filaments were present in the same cell, and to compare their distribution within cultures and within individual cells. As controls, cultures were stained with a second antibody only, and with the preimmune serum replacing the first antibody. The cells were examined with a Zeiss photomicroscope 11, equipped with epifluorescence optics. Autoradiography For each cell type, one group of cultures was treated with 1 pCi [ 3 ~ ] t h y m i d i n e / m(methyL3H, ~ 78.0 Ci/mmol; 1 Ci = 37 GBq) (Dupont NEN Research Products) for 2 h, then fixed in - 20°C methanol. A second group of cultures was treated with 1 pCi [3~]thymidine/mmolfor 24 h, then fixed in -20°C methanol. Subsequently, all [3~]thymidine-treatedcells were stained for keratin by the procedure described above. The coverslips were then mounted with glycerine jelly on glass slides, cell side up, keeping the cells moist with PBS at all times. The slides were dipped in emulsion, allowed to dry, then stored in a light-proof box at 4°C for 1 week. They were developed with Dl9 (Kodak) and fixed with regular Kodak fixer. Then, the coverslips were removed from the slides, rinsed with PBS, and mounted cell-side down on clean glass slides with gelvatol. The slides were examined blindly. They were scanned visually to define areas with the most representative (common) staining pattern. Fields within these areas were photographed and used for analysis. Immunoblotting Preparations of enriched keratins (Achtstaetter et al. 1986) were obtained from cultures that were confluent but not overgrown. They were subjected to SDS-PAGE in 8.5% minigels (Laemmli 1970) using the high ionic strength buffer system of Thomas and Kornberg (1975). Enriched keratin (20-30 pg, as estimated by OD?, .,)was loaded in each lane. After electrophoretic transfer to nitrocellulose (Towbin et al. 1979), the keratins were tested with the rabbit polyclonal antibody RAK to human callus (Sun and Green 1978). The reasons for this choice of antibody are described under Immunofluorescence microscopy. The nitrocellulose was blocked with TTBS plus 3% BSA and 1% NGS for 1 h at room temperature and then incubated with RAK antibody (1:1000) in TTBS plus 1% BSA overnight at 4°C. After washing, the nitro-
Results ROSE cells in primary cultures were easily recognized by their distinct cobblestone morphology- d dams and Auersperg 1981). Keratin filaments were abundant throughout the cytoplasm of all these cells, often at a higher concentration surrounding the nucleus (Figs. 1 and 2). Only few cells, at the periphery of colonies, expressed low levels of vimentin, but most cells were vimentin negative (Fig. 3). The cell line ROSE 239 is immortal, but it is indistinguishable from low passage ROSE cultures in cell morphology and intercellular organization, and it remains monolayered when crowded (Auersperg et al. 1991). These cells had a variable and markedly reduced keratin content compared with the primary ROSE cells. Only about 5-30% of cells expressed keratin, which ranged from one or a few granules or short fibrils in the majority of keratin-positive cells, to near-normal staining in some (Tables 1 and 2; Figs. 2 and 6). All ROSE 239 cells were vimentin positive (Fig. 3). The second immortal cell line, ROSE 199, is more atypical than ROSE 239 (Auersperg et al. 1991). When sparse, the cells are monolayered and epithelial, but in densely grown cultures, they continue to proliferate rapidly, forming multilayers and ridges (Adams and Auersperg 1985). Up to 50% of these cells in sparse cultures, and up to 95% in crowded cultures, were keratin negative or contained very sparse granules. The remainder contained more extensive, short fibrils (Tables 1 and 2; Figs. 2 and 6). Small groups of intensely keratin-positive ROSE 199 cells were located on the multilayered ridges that appeared in crowded cultures of this line (Fig. 4). All ROSE 199 cells contained vimentin filaments, with a staining pattern resembling ROSE 239 cultures (Fig. 3). In the KiMSV-transformed lines, the proportion of keratin-positive cells exceeded that of the immortal lines (Tables 1 and 2), but staining intensities were generally reduced. V197-15A cultures contained two distinct cell types (Figs. 2 and 6). There were small round cells, some with long extensions; these cells stained intensely for keratin in the perinuclear region. However, most cells were large, flat, and stained faintly or not at all. The V197-10B cells were large, flat, and more uniform. Most of the cells that stained positively in this line contained only fine, weakly staining keratin filaments (Figs. 2 and 6). All cells in both transformed lines contained extensive fibrillar networks of vimentin, which in staining intensity and distribution resembled the vimentin pattern of fibroblasts (Fig. 3). In both immortal lines, keratin expression was inversely proportional to cell density (Tables 1 and 2). There were some exceptions to this overall pattern. This was evident at the
Biochem. Cell Biol. Downloaded from www.nrcresearchpress.com by YORK UNIV on 11/10/14 For personal use only.
HORNBY ET AL.
19
FIG. 2. Keratin in ROSE cells. Immunofluorescence microscopy. (a) Primary culture. All cells are strongly fluorescent. (6)Line ROSE V197-15A. (c) Line ROSE 239. Note that the whole field is covered with cells. The fluorescence of 3 cells approaches that of primary cultures; in the others, fluorescence is either reduced or absent. ( d ) Line ROSE V197-10B. Arrowheads point to fine filaments. (e) Line ROSE 199. An area with predominantly keratin-positive cells. ( f ) Normal serum control to Fig. 2e. x 300.
periphery of colonies where the keratin staining of positive cells was particularly intense, but there were also some keratin-negative cells. Similarly, in areas of uniform cell density within colonies, keratin-positive and -negative cells sometimes occurred side by side in distinct groups (Fig. 2b). Finally, the multilayered ridges formed by ROSE 199 cells in very densely grown areas were intensely keratin positive. There was no clear-cut interrelationship between density and intermediate filament expression in the virally transformed lines. Double staining showed different distributions of keratin and vimentin within individual cells (data not
shown). Since the immortal and transformed cells were 100% vimentin positive, keratin and vimentin were coexpressed in all keratin-positive cells. Immunoblotting revealed identical keratin subtypes, of 53, 55, 57, 60, 62, and 66 kDa, in primary cultures and in both immortal lines (Fig. 5). The virus-transformed line V197-15A lacked the 53-kDa keratin and line V197-1OB lacked the 53-, 60-, 62-, and 66-kDa keratins (Fig. 5). To determine whether the variation in keratin content among the immortal and transformed ROSE cells was related to their proliferative activity, cells were first treated
Biochem. Cell Biol. Downloaded from www.nrcresearchpress.com by YORK UNIV on 11/10/14 For personal use only.
20
BIOCHEM. CELL BIOL. VOL. 70,
1992
FIG. 3. Vimentin in ROSE cells. Immunofluorescence microscopy. (a) Primary culture. (b) Line ROSE V197-15A. (c) Line ROSE 239. (d) Line ROSE V197-10B. (e) Line ROSE 199. ( f ) Fibroblasts. x 300.
with tritiated thymidine either for 2 h to identify cells in S phase (Table l), or for 24 h to identify cycling populations (Table 2), and were then stained for keratin (Fig. 6). The tables list the percentage of cells that are either keratin positive or thymidine labelled, as well as the percentage that are positive for both markers. A comparison of the observed percentage of cells that expressed keratin and also incorporated thymidine to the percentage expected if keratin expression and thymidine incorporation occurred independently, demonstrated that these characteristics were not interdependent in the immortal lines. Microscopic analysis revealed one exception to this conclusion: the highly keratin-
positive cells that covered the ridges in dense ROSE 199 cultures were consistently nonproliferative. As these structures are quite rare, they did not affect the statistical data based on large cell numbers (Tables 1 and 2). In the virustransformed lines, the relationships among keratinexpressing cells that were thymidine labelled for 2 h resembled those of the immortal lines, although after the 24-h pulse, the percentage of keratin-positive cells that were dividing was less than expected. Analysis of these preparations confirmed the visual impression of an inverse relationship between cell density and keratin expression in the immortal lines, but not in the virus-transformed lines.
21
Biochem. Cell Biol. Downloaded from www.nrcresearchpress.com by YORK UNIV on 11/10/14 For personal use only.
HORNBY ET AL.
FIG. 4. Line ROSE 199. Keratin-positive ridge next to keratin-poor monolayer. (a) Phase microscopy. (b) Immunofluorescence microscopy. x 300.
Discussion In our study, all the cell types examined, with the exception of cells in primary culture, were vimentin positive and showed similar patterns of vimentin filament distribution. The staining intensity increased from primary to immortalized to virally transformed cells, but was fairly uniform among cells at the same stage of neoplastic progression. Thus, at the level of immunofluorescence microscopy, neoplastic progression and major changes in cell morphology and growth pattern did not alter the intracellular vimentin distribution. but did cause auantitative changes in the expression o f this type of- intermediate filament. Interestingly, it has been proposed that vimentin may be a DNA-binding protein and that the polymerized cytoplasmic filaments may be its storage form (Traub 1985). In this context, the increase in vimentin filaments observed during neoplastic conversion in our system, as well as clinical correlations with cancer progression (Cattoretti et al. 1988; Sommers et al. 1989), raise the possibility that some of the changes observed may have a gene regulatory rather than a structural basis. In contrast to the relative uniformity of vimentin expression, neoplastic progression in ROSE cells was associated with a striking intercellular variability in keratin expression at the microscopic level, which could not be attributed to variations in proliferative activity, cell shape, or intercellular organization. The immortalized cell lines, ROSE 199 and ROSE 239, contained fewer keratin-positive cells than the primary cultures. The keratin staining in the positive cells was reduced in intensity and had an altered microscopic appearance. It is particularly interesting that the two immortal cell lines expressed similar abnormalities in keratin expression microscopically as well as identical keratin subtypes by Western blot, though the two lines have very different growth characteristics: line ROSE 239, but not line ROSE 199, resembles primary OSE cultures morphologically. Thus, the maintenance of the epithelial phenotype that characterizes ROSE in primary culture does not require the presence of the normal complement of keratin filaments and persists in the presence of amplified vimentin. It was surprising that the profound microscopic defects in keratin
kDa
1
2
3
4
5
6,,
,66-
@: 53
w
,3 -48 44 -40
-
b
FIG. 5. Western blot showing keratin profile of ROSE cells (lanes 1-5). Lane 1, ROSE in primary culture; lane 2, ROSE 239; lane 3, ROSE 199; lane 4, ROSE V197-15A; lane 5, ROSE V197-10B; lane 6, human C-41 cells, known to express 40-, 44-,48-, 53-, 55-, and 58-kDa keratins (Auersperget al. 1989). were used as a molecular mass marker. Partially enriched keratin, (20-30 pg, estimated by OD,,, m) was loaded in each lane. Molecular mass markers at 36, 44, and 65 kDa are indicated by arrowheads.
expression observed in the two immortal cell lines coincided with the expression of all keratin subtypes detected in primary cultures, as indicated by immunoblotting. It is impossible to distinguish among the many abnormalities that may form the basis for the reduced keratin fluorescence and altered patterns of staining in these cells. Among them are changes in the amounts or ratios of keratin subunits available for polymerization. Alternatively, the defects could result from a loss of (conformational) epitopes, e.g., through changes in keratin monomers at the molecular level, that would interfere with normal filament assembly and prevent detection by immunofluorescence techniques (Miller et al. 1991). In addition, altered nonkeratinous proteins might interfere with the normal cytoplasmic distribution and (or) antigenicity of these intermediate filaments. In the virally transformed cell lines, the proportion of keratin-positive cells was increased compared with the immortalized cell lines, though keratin staining per cell was generally reduced. A similar increase in the proportion of keratin-expressing cells occurs in tumorigenic derivatives of
Biochem. Cell Biol. Downloaded from www.nrcresearchpress.com by YORK UNIV on 11/10/14 For personal use only.
22 BIOCHEM. CELL BIOL. VOL. 70, 1992
23
HORNBY ET AL.
TABLE1. Percentage of ROSE cells positive for keratin and (or) [3~]thymidine (['~lth~midine label for 2 h)
TABLE 2. Percentage of ROSE cells positive for keratin and (or) ['~lthymidine(['Hlthymidine label for 24 h)
K + 'H
K + 3~ Line
No. of cells/field -
ROSE 239
Biochem. Cell Biol. Downloaded from www.nrcresearchpress.com by YORK UNIV on 11/10/14 For personal use only.
ROSE 199 V197-10B V197-15A
-
K
'H
25 31 11 46 21 100 82 42 81 78 53
23 27 17 27 39 16 24 23 22 13 8
Actual Exp. Neg.
Line
No. of cells/field
ROSE 239
< 100 200-400 600-700 < 100 200-400 800-900 < 100 < 100 300-400 < 100 < 100 < 100
K
'H
Actual Exp. Neg.
-
100-200 200-400 600-700 200-400 700-800 < 100 200-300 200-300 < 100 < 100 300-400
5 6 2 9 6 16 14 4 14 0 1
6 8 2 12 8 16 20 10 18 10 4
58 51 76 35 46 0 9 40 11 9 39
NOTE:Cultures of immortal, nontumorigenic cell lines (ROSE 239 and ROSE 199) and of KiMSV-transformed, tumorigenic~elllines (V197-10B and V197-15A) were labelled with ['Hlthymidine for 2 h and then stained by immunofluorescence for keratin. In representative microscopic fields, the total number of cells, and the cells positive for keratin (K), ['H]thymidine ('H), keratin and ['Hlthymidine (K+ 'H), and neither marker (Neg.) were counted and ranked by cell density. (Note that the K and 'H categories include the cells that are listed under K + 'H, in addition to the cells expressing only one marker). The column headed Actual refers to the percentage of cells observed in the K + 'H category; the column headed Exp. (expected) refers to the percent of cells in the K + 'H category, expected on the basis of calculations from the observed K and 'H values (expected = Yo 'H positive x % keratin positive / 100). Results are the means of four microscopic fields.
nontumorigenic rat liver lines (Troyanovsky et al. 1986), and increases in differentiation as a result of KiMSV-mediated transformation have been observed in other types of cells (Auersperg et al. 1990). The altered appearance of keratin filaments in the transformed lines may be related to the results of the Western blot analyses, in which there appeared to be a loss of specific keratin subtypes compared with primary and immortal cultures. The available data do not allow us to distinguish possible changes at the level of gene regulation from alterations in antigenicity based on mutations as the basis for the observed abnormalities in filament numbers and subtype losses. Together, the data suggest that immortalization affects the uniformity of keratin expression among cells, the physical characteristics of keratins, and possibly, the amounts of keratin present, but that the expression of tissue-specific keratin subtypes by ROSE cells is maintained until further steps in neoplastic progression take place. The relatively high molecular masses of ROSE-derived keratins differ from the keratin profile of human simple epithelia, which usually ranges from 40 to 55 kDa. It is possible that this is a difference between species, because a similar discrepancy has been observed between liver keratins of humans and rat (Moll et al. 1982; Langbein and Neupert 1986). In two analyses of rat mesothelial keratins (MacKay et al. 1990; Troyanovsky et al. 1986), molecular masses were reported to range from 40 to 55 kDa; however, it is not clear whether the antibodies used in these studies recognize the same subtypes as the antibody used here. Many epithelia that are vimentin negative in vivo acquire this cytoskeletal component in culture (Connell and Rhein-
ROSE 199 V197-10B V197-15A
NOTE:Cultures of immortal, nontumorigenic cell lines (ROSE 239 and ROSE 199) and of KiMSV-transformed, tumorigenic cell lines (V197-10B and V197-ISA) were labelled with ['H]thymidine for 24 h and then stained by irnrnunofluorescence for keratin. In representative microscopic fields, the total number of cells, the cells positive for keratin (K), ['H~thymidine ('H), keratin and ['~lthymidine(K + 'H), and neither marker (Neg.) were counted and ranked by cell densit . (Note that the K and 'H categories include the cells that are listed under K + H, in addition to the cells expressing only one marker). The column headed Actual refers to the percentage of cells observed in the K + 'H category; the column headed Exp. (expected) refers to ed on the basis of calculations from the percent of cells in the K + 'H category, ex .. x % keratin positive / 100). the observed K and 'H values (expected = % H poslhve Results are the means of four microscopic fields.
Y .
F
wald 1983; Franke et al. 1981; Ben Ze'ev 1985; Langbein and Neupert 1986). Our results show that this group of tissues includes rat ovarian surface epithelial cells. Under certain conditions, keratin and vimentin vary in a reciprocal fashion; in cultures of human mesothelial cells (Connell and Rheinwald 1983), vimentin expression is enhanced and keratin expression reduced in response to epidermal growth factor (EGF), concommittantly with shape changes and enhanced proliferation. In a series of cell lines, Ben Ze'ev (1985) demonstrated that the expression of vimentin decreased with increasing cell density and that the more immediate signals for this response were density-dependent changes in cell shape. On the other hand, keratin expression in MDCK cells was proportional to cell density, and the signal here appeared to involve intercellular contact rather than shape changes. In ROSE cultures, keratin was expressed in all primary cells independently of density, but it was reduced at high densities in the immortal cell lines. The difference between this response and that of other epithelial cell types (Ben Ze'ev 1985; Troyanovsky et al. 1986) may be related to differences in the nature of the intercellular contacts between, for example, the highly polar MDCK cells and the less polar ROSE lines. The lack of a response to density by the virally transformed ROSE cells is in keeping with the increased autonomy of fully transformed cells. Keratin content was inversely proportional to cell density in all regions of immortal lines except in the multilayered ROSE 199 ridges. Possibly, cells in these specialized regions
FIG. 6. Relationship of keratin expression to thymidine incorporation. (a and b) ROSE 239. (c and d) ROSE V197-10B. (e and f ) ROSE 199. (g and h) ROSE V197-15A. (a, c, e, and g) Immunofluorescencemicroscopy. Autoradiographs after [H3]thymidinelabelling for 2 (b and d) and 24 h (f and h). Large arrowheads, proliferating cells that are keratin positive; small arrowheads, proliferating cells that are keratin negative. In the field shown, essentially all V197-10B cells are keratin positive (c); most proliferating V197-15A cells are keratin negative (g and h). x 170.
Biochem. Cell Biol. Downloaded from www.nrcresearchpress.com by YORK UNIV on 11/10/14 For personal use only.
24
BIOCHEM. CELL BIOL. VOL. 70, 1992
express different subtypes of keratins, as has been described in similar regions of hepatocyte cultures (Troyanovsky et al. 1986). Other minor inconsistencies in the inverse correlation between density and keratin expression by immortal ROSE cells resemble the unexplained heterogeneities in keratin patterns that have been reported in cultures of rat liver cells (Langbein 1986). In our system, such variation seemed unrelated to proliferative activity. In transformed rat liver lines, local changes in cell-substratum adhesion have been suggested as a possible explanation (Troyanovsky et al. 1986). Histologically, tumors produced by the two KiMSVtransformed cell lines examined here resembled human endometrioid stromal sarcomas (Adams and Auersperg 1981), which are a minor but nevertheless important subclass of the "Common Epithelial Tumors" category in the WHO classification of ovarian tumors (Scully 1977). In the past, it has been difficult to decide whether the mesodermal phenotypic characteristics of these neoplasms are due to the multipotential character of the mesodermally derived OSE, or to an origin in stromal cells (Adarns and Auersperg 1981). Our demonstration of keratin in the V197-10B and V19715A lines supports the former hypothesis. The alterations in intermediate filaments that accompanied KiMSVmediated transformation of ROSE cells, viz, a drastic reduction and qualitative alterations in keratin expression and increased vimentin expression, differ from the changes reported in many of the human ovarian serous and mucinous carcinomas, which tend to retain the keratin subtypes of normal OSE and express variable amounts of vimentin (Czernobilsky et al. 1984). The phenotype of the virally transformed rat cells may be related to the oncogenic agent used in this study. Ras is one of the most commonly activated protoncogenes in human carcinomas. However, its retroviral form, KiMSV, has a propensity to form sarcomas, a phenotype that would be compatible with a low keratin high vimentin profile. The results of this study demonstrate that normal keratin expression is not necessary to maintain the normal morphology and growth pattern of OSE cells. They show further that keratin expression is drastically altered by immortalization and viral transformation. The variability of these changes supports the impression gained from clinical histopathology that keratin expression may have little direct influence on neoplastic progression. On the other hand, the progressive and relatively uniform increase in vimentin expression with neoplastic progression raises the possibility that a more direct interrelationship between these processes may exist. Acknowledgements This research was supported by grants and a research associateship to N. Auersperg from the National Cancer Institute of Canada and the Medical Research Council of Canada. We thank Dr. Sigrid E. Myrdal and Ms Vera Lee for their advice and help. Achtstaetter, T., Hatzfeld, M., Quinland, R.A., Parmelee, D.C., and Franke, W.W. 1986.Separation of cytokeratin polypeptides by gel electrophoretic and chromatographic techniques and their identification by immunoblotting. Methods Enzymol. 134: 355-371.
Adams, A.T., and Auersperg, N. 1981.Transformation of cultured rat ovarian surface epithelial cells by Kirsten murine sarcoma virus. Cancer Res. 41: 2063-2072. 1985. A cell line, ROSE 199, derived from normal rat ovarian surface epithelium. Exp. Cell Biol. 53: 181-188. Auersperg, N., Kruk, P.A., MacLaren, I.A., Watt, F.M., and Myrdal, S.E. 1989. Heterogeneous expression of keratin, involucrin, and extracellular matrix among subpopulations of a poorly differentiated human cervical carcinoma: possible relationships to patterns of invasion. Cancer Res. 49: 3007-3014. Auersperg, N., MacLaren, I.A., and Bissell, M.J. 1990. V-K-ras transformation induces reversion to an earlier developmental form in adult rat adrenal cells. Differentiation, 43: 29-36. Auersperg, N., MacLaren, I.A., and Kruk, P.A. 1991. Ovarian surface epithelium: autonomous production of connective tissuetype extracellular matrix. Biol. Reprod. 44: 717-724. Benjamin, E., Law, S., and Bobrow, L.G. 1987. Intermeditite filaments cytokeratin and vimentin in ovarian sex cord-stromal tumours with correlative studies in adult and fetal ovaries. J. Pathol. 152: 253-263. Ben Ze'ev, A. 1985. Cell-cell interaction and cell configuration related control of cytokeratins and vimentin expression in epithelial cells and in fibroblasts. Ann. N.Y. Acad. Sci. 455: 597-613. 1986.The relationship between cytoplasmic organization, gene expression and morphogenesis. TIBS, 11: 478-481. Cattoretti, G., Andreola, S., Clemente, C., D'amato, L., and Rilke. F. 1988.Vimentin and p53 expression on epidermal growth factor receptor-positive, oestrogen receptor-negative breast carcinomas. Br. J. Cancer, 57: 353-357. Connell, N.D., and Rheinwald, J.G. 1983. Regulation of the cytoskeleton in mesothelial cells: reversible loss of keratin and increase in vimentin during rapid growth in culture. Cell, 34: 245-253. Czernobilsky, B. 1985.Co-expression of cytokeratin and vimentin filaments in mesothelial, granulosa and rete ovarii cells of the human ovary. Eur. J. Cell Biol. 37: 175-190. Czernobilsky, B., Moll, R., and Franke, W.W. 1984.Intermediate filaments of normal and neoplastic tissues of the female genital tract with emphasis on problems of differential tumor diagnosis. Pathol. Res. Pract. 179: 31-37. Dabbs, D. J., and Geisinger, K.R. 1988.Common epithelial ovarian tumors: immunohistochemical intermediate filament profiles. Cancer, 62: 368-374. Franke, W.W., Mayer, D., Schmid, E., Denk, H., and Borenfreund, E. 1981. Differences of expression of cytoskeletal proteins in cultured rat hepatocytes and hepatoma cells. Exp. Cell Res. 134: 345-365. Goldman, R., Goldman, A., Green, K., Jones, J., Lieska, N., and Yang, H.-Y. 1985. Intermediate filaments: possible functions as cytoskeletal connecting links between the nucleus and the cell surface. Ann. N.Y. Acad. Sci. 455: 1-17. Laemmli, U.K. 1970. Cleavage of structural proteins during the assembly of the head of bacteriophage T4. Nature (London), 227: 680-685. Langbein, L., and Neupert, G. 1986. Modulation of expression of intermediate filaments during the development of established rat liver cell lines. Acta Histochem. 80: 149-158. MacKay, A.M., Tracy, R.P., and Craighead, J.E. 1990. Cytokeratin expression in rat mesothelial cells in vitro is controlled by the extracellular matrix. J. Cell Sci. 95: 97-107. Miller, R.K., Vikstrom, K., and Goldman, R.D. 1991. Keratin incorporation into intermediate filaments is a rapid process. J. Cell Biol. 113: 843-855. Moll, R., Franke, W.W., and Schiller, D.L. 1982. The catalog of human cytokeratins: patterns of expression in normal epithelia, tumors and cultured cells. Cell, 31: 11-24. Moll, R., Krepler, R., and Franke, W.W. 1983a. Complex
-
Biochem. Cell Biol. Downloaded from www.nrcresearchpress.com by YORK UNIV on 11/10/14 For personal use only.
HORNBY ET AL.
cytokeratin polypeptide patterns observed in certain human carcinomas. Differentiation, 23: 256-269. Moll, R., Levy, R., Czernobilsky, B., Hohlweg-Majert. P.. Dallenbach-Hellweg, G., and Franke, W. W. 1983b. Cytokeratins of normal epithelia and some neoplasms of the female genital tract. Lab. Invest. 49: 599-610. Nicosia, S.V., and Nicosia, R.F. 1988. Neoplasms of the ovarian mesothelium. In Pathology of human neoplasms. Edited by H.A. Azar. Raven Press, Ltd., New York. pp. 435-486. O'Guin, W.M., Schermer, A., and Sun, T.-T. 1985. Immunofluorescence staining of keratin filaments in cultured epithelial cells. J. Tissue Cult. Methods. 9: 123-128. Robson, R.M. 1989. Intermediate filaments. Curr. Opin. Cell Biol. 1: 36-43. Rodrighuez, J., and Deinhardt. F. 1960. Preparation of a semipermanent mounting medium for fluorescent antibody studies. Virology, 12: 3 16-320. Scully. R.E. 1977. Ovarian tumors. Am. J. Pathol. 87: 686-720. Sommers, C.L., Walker-Jones, D., Heckford, S.E., Worland, P., Valverius, E., Clark, R., McCormick, F., Stampfer, M., Abularach, S., and Gelmann, E.P. 1989. Vimentin rather than keratin expression in some hormone-independent breast cancer cell lines and in oncogene-transformed mammary epithelial cells. Cancer Res. 49: 4258-4263.
25
Steinert, P.M., and Roop, D.R. 1988. Molecular and cellular biology of intermediate filaments. Annu. Rev. Biochem. 57: 593-625. Stewart, M. 1990. Intermediate filaments: structure, assembly and molecular interactions. Curr. Opin. Cell Biol. 2: 91-100. Sun, T.-T., and Green, H. 1978, Immunofluorescent staining of keratin fibers in cultured cells. Cell, 14: 469-476. Thomas, J.O., and Kornberg, R.D. 1975. An octamer of histones in chromatin and free in solution. Proc. Natl. Acad. Sci. U.S.A. 72: 2626-2630. Towbin, H., Staehelin, T., and Gordon, J. 1979. Electrophoretic transfer of proteins from polyacrylamide gels to nitrocellulose sheets: procedure and some applications. Proc. Natl. Acad. Sci. U.S.A. 76: 4350-4354. Traub, P. 1985. Are intermediate filament proteins involved in gene expression? Ann. N.Y. Acad. Sci. 455: 68-78. Troyanovsky, S.M., Bannikov, G.A., Montesano, R., and Vasiliev, J.M. 1986. Density-dependent expression of keratins in transformed rat liver cell lines. Cell Biol. Int. Rep. 10: 263-270. Viale, G., Gambacorta, M., Dell'Orto, P., and Coggi, G. 1988. Coexpression of cytokeratins and vimentin in common epithelial tumours of the ovary: an immunocytochemical study of eightythree cases. Virchows Archiv. A, Pathol. Anat. 413: 91-101.
