0013-7227/89/1261-0399$02.00/0 Endocrinology Copyright © 1990 by The Endocrine Society

Vol. 126, No. 1 Printed in U.S.A.

Maintenance of in Vivo-Like Keratin Expression, Sex Steroid Responsiveness, and Estrogen Receptor Expression in Cultured Human Ectocervical Epithelial Cells* GEORGE I. GORODESKI, RICHARD L. ECKERT, WOLF H. UTIAN, AND ELLEN A. RORKE Departments of Physiology and Biophysics (G.I.G., R.L.E.), Dermatology (R.L.E.), and Environmental Health Sciences (E.A.R.), Case Western Reserve University School of Medicine, and the Department of Obstetrics and Gynecology, Mount Sinai Medical Center (W.H.U.), Cleveland, Ohio 44106

epidermis in vivo and cultured epidermal keratinocytes, express very abundant levels of K13. In fact, K13 appears to be a specific marker of ECE cells in the female reproductive tract. When incubated with 10 nM diethylstilbestrol, ECE cell envelope production increased 3-fold, while incubation with 100 nM progesterone decreased envelope formation 3.4-fold. Simultaneous incubation with progesterone antagonized the diethylstilbestrol stimulation. Thus, in vivo-like sex steroid regulation of ECE cell differentiation is maintained in culture. In addition, the cells possess a high affinity, limited capacity binding site for estradiol that has a Kd of 1.2 ± 0.1 nM. This system is likely to provide a useful model for the study of sex steroid regulation of normal ectocervical epithelial cell function. (Endocrinology 126: 399-406, 1989)

ABSTRACT. In the present manuscript we demonstrate that ectocervical epithelial cells (ECE cells) retain a high degree of differentiated function when cultured using feeder layers. We characterize the cultured cells with respect to morphology, expression of cytokeratins, responsiveness to sex steroids, and the presence of estrogen-binding sites. Like ectocervical cells in vivo, the cultured cells display a typical epithelial cell morphology and undergo extensive stratification and envelope (superficial cell) formation. Like the in vivo ectocervical epithelium, the cultured ECE cells express type I cytokeratins K13, K14, K16, K17, and K19 and type II cytokeratins K5 and K6. Under normal culture conditions, however, cytokeratins Kl, K2, K4, Kll, and K15, which are expressed in vivo, are not expressed. An interesting finding is that ECE cells, in contrast to endocervix and

T

HE HUMAN uterine cervix is lined by three types of epithelia: the endocervix, a simple columnar epithelium; the ectocervix, a stratifying squamous nonkeratinizing epithelium; and the transition zone epithelium, a squamous metaplastic transdifferentiated epithelium of the transformation zone between the endo- and ectocervix (1). Although some controversy still exists as to the embryonic origin of the human cervix, most researchers believe that it is derived from the Mullerian duct, which in the embryo is the retroperitoneal invagination of the celomic epithelium (2). Clinical studies have yielded considerable information regarding pathological processes in the cervix, but despite this progress in describing the clinical progression of

cervical disease, little is known about the underlying mechanisms responsible for these changes. Part of the difficulty has been the absence of an adequate human ectocervical epithelial cell culture system. In vivo, ectocervical epithelial cells resemble epidermal keratinocytes in morphology and differentiation. Both types of epithelia are composed of proliferating (basal) and differentiating (spinous and granular) cells, and both cell types form cornified envelopes as their terminal step in differentiation. In the cervix the terminal envelopes are called intermediate or superficial cells (3). The ectocervical epithelium in vivo is regulated by the sex steroid hormones. Estradiol (E2) promotes cell maturation before exfoliation; progesterone (P) has the opposite effect, as it promotes exfoliation before cell maturation (4). The effects of the sex steroids appear to be mediated by specific receptors that have been described in ectocervix in vivo (5). In the present manuscript we demonstrate that cultured ectocervical epithelial cells (ECE cells) express a unique keratin profile that distinguishes them from other

Received June 29,1989. Address requests for reprints to: Dr. Ellen A. Rorke, Department of Environmental Health Sciences, Case Western Reserve University School of Medicine, 2119 Abington Road, Cleveland, Ohio 44106. * This work was supported by grants from the Cystic Fibrosis Foundation (to R.L.E. and E.A.R.). Some of the work was also supported by the Skin Diseases Research Center of Northeast Ohio (NIH, AR-39750) and a Cancer Center Grant (to E.A.R.; NIH, P30-CA43703).

399

HUMAN ECTOCERVICAL CELL CULTURES

400

cell types in the female reproductive tract, express an estrogen receptor-like binding activity, and respond to physiological levels of sex steroids. Our results suggest that this system will be a useful model for the study of human ECE cell function.

Materials and Methods

Endo • 1990 Vol 126* No 1

pellet was extracted sequentially with TE containing 1% Triton X-100 and 10 U/ml DNAase-I, TE containing 1% Triton X100 and 0.6 M KC1, and TE containing 1% Triton X-100. After the last extraction, the samples were centrifuged for 5 min at 15,000 X g to remove excess supernatant, resuspended in sample buffer (11), boiled 5 min, and stored at —20 C. Pellets for twodimensional gel electrophoresis were resuspended in lysis buffer (12) and stored at -20 C.

Preparation of 3T3 feeder cells Murine embryonic 3T3 fibroblasts, grown as previously described (6), were irradiated with 6000 rads Cobalt-60 and plated onto plastic culture dishes at a density of 5 X 104 cells/cm2. Ectoceruical cultures Ectocervical tissue was obtained from women, aged 22-49 yr, undergoing hysterectomy for a variety of clinical reasons. Only histologically normal tissue was used in the present study. The tissue was washed in cold PBS, scraped to remove residual connective tissue, minced into 1-mm squares, and placed onto dishes containing irradiated 3T3 feeders (6). The medium was changed every 48 h, and after 21-28 days the primary cultures were subcultured. The growth medium was a mixture of Dulbecco's Modified Eagle's Medium and Ham's F-12 (3 parts Dulbecco's Modified Eagle's Medium and 1 part Ham's F-12) (7) supplemented with 7.5% fetal calf serum, nonessential amino acids, insulin (5 pg/ ml), cholera toxin (1 X 1O~10 M), transferrin (5 Mg/ml), T 3 (2 X 10~9 M), hydrocortisone (1 X 10"8 M), L-glutamine (2 mM), penicillin (100 U/ml), streptomycin (100 mg/ml), gentamicin (50 Mg/ml), and adenine (1.8 X 10~4 M). For experiments, nearconfluent cultures of third to fifth passage cells were shifted to identical medium, except that the hydrocortisone was eliminated, and the fetal calf serum was replaced with 7% CXDLCDFCS [chelexed (CX), delipidized (DL), and charcoaldextran (CD)-treated fetal calf serum]. Chelex treatment removes calcium (8), delipidization removes retinoids (9), and charcoal dextran removes estrogens (10).

Envelope quantitation Cultures were harvested, counted to determine total cell number, washed with PBS containing 0.5 mM EDTA, resuspended in 1 ml 2% sodium dodecyl sulfate-20 mM dithiothreitol, and boiled for 5 min. Living cells are dissolved by boiling, but the covalently cross-linked envelopes survive and can be counted using the phase contrast microscope. Immunological detection of keratins Whole cell extracts or cytoskeleton preparations were separated by gel electrophoresis on 8.5% or 10% acrylamide gels (11), transfer blotted to nitrocellulose (13), and treated with mouse antikeratin antibodies specific for the type I or type II keratin subfamily or for specific members of each family (14). Antibody binding was detected using 125I-labeled goat antimouse second antibody and exposure on x-ray film. Detection of keratin RNA transcripts Total and poly(A)+ RNA were isolated using the CsCl gradient method (15) and oligo(dT)-cellulose chromatography (16). Equal quantities (1-5 ng) of purified mRNA were fractionated on formaldehyde-containing agarose gels (17) and transferred to Biodyne-A membrane or applied to Biodyne-A using a dot blot manifold as previously described (18). mRNA species encoding keratins K5, K6, K13, and K19 were detected using cDNAs specific for each keratin sequence (19, 20) labeled to a specific activity of 2 x 10~8 dpm/jug DNA by nick translation (21).

^SJMethionine labeling of cultures

Estrogen receptor assay

Cultures were vigorously washed with Hanks' Balanced Salt Solution containing 1 mM EDTA to remove any residual 3T3 feeders, and the ECE cells were then shifted to medium containing l/10th the normal level of methionine and 50 /uCi/ml [35S]methionine and incubated for 14-16 h before harvest.

Scatchard analysis (22) was performed on total cell extracts using [3H]E2 as the tracer and hydroxylapatite (HAP) to separate bound and free ligand (10). Confluent cultures of ECE cells were harvested, resuspended in iced phosphate buffer (5 mM sodium phosphate, pH 7.4, at 4 C, 10 mM thioglycerol, 10 mM sodium molybdate, and 10% glycerol), homogenized in a Dounce homogenizer (B-pestle; Kontes Co., Vineland, NJ), and extracted for 60 min at 4 C by the addition of 3 vol high salt buffer (10 mM Tris-HCl, pH 8.5 at 4 C, 1.5 mM EDTA, 10 mM thioglycerol, 10% glycerol, 10 mM sodium molybdate, and 0.8 M KC1). The extracts were then centrifuged at 180,000 X g for 30 min at 4 C, and the supernatant (total cell extract) was diluted with 3 vol iced phosphate buffer and used immediately. To assess total and nonspecific binding, 200 /x\ of extract was incubated for 4 h at 30 C with [3H]E2 at concentrations ranging from 1 x 10"11 M to 6 x 10"8 M in the presence or absence of a 100-fold excess of diethylstilbestrol (DES). These

Cell extracts and cytoskeletal preparations Cells were harvested, resuspended in 1 ml growth medium, and counted in a hemacytometer. Viability was determined using trypan blue exclusion and was generally greater than 95%. After counting, the cells were washed with PBS to remove serum proteins. For whole cell extracts, the cells were resuspended in sodium dodecyl sulfate sample buffer, boiled for 5 min, and stored at -20 C until use. For preparation of cytoskeletal proteins, the cells were homogenized in 10 mM TrisHC1 (pH 7.5) containing 1 mM EDTA (TE) and pelleted. The

HUMAN ECTOCERVICAL CELL CULTURES

401

conditions (4 h at 30 C) measure total (empty plus filled) receptor sites (10). The tubes were cooled to 4 C, and an aliquot was removed and counted to accurately determine the final [3H]E2 concentration. The remaining sample was treated with HAP for 30 min at 4 C to adsorb the hormone-receptor complex, the HAP was washed four times with phosphate buffer to remove residual free [3H]E2, and the HAP pellet was counted in scintillation fluid as previously described (10). Specific binding was determined by subtracting nonspecific from total binding at each concentration. The results were plotted according to the equation of Scatchard (22). In some experiments incubations were included at 0-4 C for 20 h to determine the fraction of filled and empty sites (10). These values agreed with the values obtained at 30 C, indicating that essentially all of the receptor sites were empty. Specific receptor binding is expressed per mg protein. Protein concentrations were determined using the Bio-Rad protein assay (Bio-Rad, Richmond, CA).

Results Within 2 weeks after the initial seeding of ectocervical tissue, colonies of tightly packed cells were observed growing outward from each explant (Fig. la). These cells display a characteristic cuboidal epithelial cell morphology and large centrally located nuclei. They also appear to be tightly linked, presumably by desmosomes, since sheets of released cells are frequently observed in the culture medium. Mitotic figures could be identified in cells at the growing outer edge of the colony during colony expansion. The ectocervical ECE cell cultures are free of nonepithelial cell types, as judged by immunofluorescence with an anti-vimentin antibody (not shown). Once individual colonies fuse, the cultures spontaneously stratify, and individual cells and cell aggregates are released from the surface layer. This process of stratification and release proceeds spontaneously with or without serum. The released structures are cornified envelopes, the terminal product of ECE cell differentiation, as confirmed by their characteristic morphology (Fig. lb) and stability when boiled in the presence of reducing and denaturing agents. We frequently observe, as shown in Fig. lb, the presence of nuclei in the released squames. This is reminiscent of the situation in vivo, where ectocervical envelopes, clinically termed superficial cells, retain visible portions of the nuclei (3). To confirm that the cultured cells are indeed ECE cells and not endocervical epithelial cells or epidermal keratinocytes, we have used one- and two-dimensional gel electrophoresis as well as reaction with specific antibodies to characterize the keratins expressed in ECE cells. The keratins comprise a family of 20 proteins that compose the intermediate filaments in epithelial cells (23). Expression of members of the keratin family is tissue and cell type specific (24). Thus, the pattern of keratin expression in cultured cells is a useful indicator of tissue

FIG. 1. ECE cell colony formation and envelope production, a, The border of a colony growing outward from a slice of ectocervical tissue. Note the cuboidal epithelial cell morphology of the cells. The amorphous translucent region is an area of extensive stratification and release from the dish, b, A single cornified envelope (left) and an aggregate comprised of many envelopes (right). Many of the squames retain pyknotic nuclei resembling those present in the envelopes produced in vivo (3).

of origin (25). Shown in Fig. 2 is a two-dimensional gel fractionation of the cytoskeletal fraction isolated from [35S]methionine-labeled ECE cell cultures and from cultured epidermal keratinocytes. As previously reported (23-26), the keratinocyte cultures express the type II keratins K5 and K6 and type I keratins K14, K16, and K17 (Fig. 2b). When grown under identical conditions, the ECE cells express these same keratins, but in addition express K13 and K19 (Fig. 2a). This difference in expression of K13 and K19 was further confirmed by immunoblotting extracts from ECE cells and epidermal keratinocytes using specific antibodies. The immunological data confirm that ECE cells express much higher levels of K13 and K19 than keratinocytes (Fig. 3). This difference appears to be a tissue-specific difference in expression that is retained when the cells are placed into culture. Table 1

HUMAN ECTOCERVICAL CELL CULTURES

402

Endo • 1990 Voll26»Nol

K19 f ^ , Jt

\

:?%.-

.•*

K13 FIG. 3. Immunological detection of K13 and K19 in cultured human epidermal keratinocytes and ECE cells. Confluent cultures of human epidermal keratinocytes or ECE cells were harvested, counted, and resuspended at 5000 cells/^l in sample buffer. Twenty microliters of each extract were fractionated on an 8% acrylamide gel in paired sets of lanes. The fractionated proteins were blotted to nitrocellulose and incubated with primary antibody and secondary 125I-labeled antibody as described in the text. K13, Incubated with anti-Kl3 antibody; K19, incubated with anti-K19 antibody; KER, keratinocyte.

- « - NEPHGE FIG. 2. Cytokeratin expression in cultured epidermal keratinocytes and ECE cells. Confluent cultures of epidermal keratinocytes and ECE cells were labeled overnight with [35S]methionine. Cytoskeletal extracts were prepared and fractionated on a two-dimensional non-equilibrium pH gradient electrophoresis (NEPHGE) gel (12). The gels were then fluorographed and exposed on x-ray film. Each keratin is indicated by number. The arrowheads indicate keratin aggregates formed during isoelectric focusing (14). The Coomassie staining pattern was identical to the fluorographic pattern (not shown), a, ECE cells; b, keratinocytes. In the non-equilibrium pH gradient electrophoresis (NEPHGE) dimension the arrow points toward the basic end of the gel.

compares the cytokeratin expression in various epithelia both in vivo (14, 23, 27-29) and in vitro. A major difference between ectocervix, endocervix, and epidermis is the difference in expression of K13 and K19. Our data suggest that K13 is a specific marker of ectocervical cell differentiation and that K13 expression is retained by ECE cells in culture (27, 29, 30). We have also monitored for the presence of specific transcripts encoding K5, K6, K13, and K19 using cloned cDNA sequences specific for each keratin (19). The relative size of the transcripts (K5 and K6, 2350 bases; K13, 2075 bases; K19,1580 bases) is essentially identical to that observed in human epidermal keratinocytes (Fig. 4) (19). The ectocervix in vivo is a target tissue for the sex steroids estrogen and P (1, 4, 31). Since steroid hormone responses in most tissues studied are mediated by specific receptors (32, 33), we next determined whether a specific E2-binding site was present in ECE cells. Scatchard

TABLE 1. Comparison of epidermal, endocervical and ectocervical cell keratin composition

In Vivo #

Cell Culture

MW Ecto* Endo*

Kl K2 K3 K4 K5 K6 K7

63 59 58 56 54

K8

52.5

Epider* v

ECE Cell „ ,,

Cultures«

Epidermal Keratinocyte

„ , / Cultures «:

67

65.5

K9

64

K10 Kll K12 K13 K14 K15 K16 K17 K18 K19

56.5 56

55 54 50 50 48 46 45 40

Information utilized in this table was compiled from references 27 (*), 23 (#), 14 (+) and the present experiments («:). Relative keratin content is indicated by ++, highly abundant; +, abundant; (+), minor species; —, not detected. Comparison of relative keratin expression is

valid within columns but not between columns, except for the epidermal keratinocyte and ECE cell cultures («:) where all intra- and intercolumn comparisons are valid. Kl through K8 are type II keratins, K9 through K19 are type I keratins.

403

HUMAN ECTOCERVICAL CELL CULTURES

0 200 400 600 3 Fmol Bound [ H]E2 (x10~3) FlG. 5. Scatchard analysis of estrogen-binding activity in ECE cells. The correlation coefficient for this line is 0.93, and the number of binding sites is 3.0 fmol/mg protein.

FlG. 4. Detection of specific transcripts encoding cytokeratins. Poly(A)+ RNA was prepared from confluent cultures of ECE cells. Five micrograms were fractionated in each of four lanes on a formaldehydecontaining denaturing agarose gel (17). The fractionated RNA was transferred to Biodyne-A membrane and irreversibly immobilized with UV irradiation as previously described (18). Lanes K19, K13, K6, and K5 were hybridized with plasmids containing cDNAs encoding keratins 19, 13, 6, and 5 (19, 20), respectively. Specific hybridization was visualized by exposure of the blots on x-ray film. RNA markers are listed in kilobases. K5 appears smeared because K5 is an abundant mRNA.

analysis of extracts prepared from cells derived from four separate cervical samples revealed a single class of specific binding sites with an affinity of 1.2 ± 0.1 nM (n = 4) for E2. The Scatchard plot in Fig. 5 shows a Kd of 1.1 nM. The binding activity ranged from 3-20 fmol/mg protein. One of the effects of sex steroids in vivo is to modulate the degree of differentiation of the ectocervical epithelium (1, 34). This regulation manifests itself as a change in the release of superficial cells (envelopes) as a function of the stage within the reproductive cycle (1, 3). We, therefore, tested whether estrogens and progestins modulate the production of envelopes (Fig. lb) by cultured ECE cells. Confluent cultures were shifted to steroidfree medium and then treated with physiological levels of DES (10 nM), a potent synthetic estrogen, or progesterone (100 nM, P) and the production of cornified envelopes was monitored. DES treatment increased enve-

Control 100 nM 10 nM 100 nM P P DES + 10 nM DES

FIG. 6. Estrogen and progestin regulation of ECE cell cornified envelope formation. Confluent cultures growing in 8-cm2 wells were shifted to experimental growth medium and treated with vehicle (0.1% ethanol), 10 nM DES, 100 nM P, or 10 nM DES plus 100 nM P for 7 days. Fresh hormone-containing medium was added on alternate days. The cells were then harvested and envelopes counted as described in the text. The results are plotted as the mean ± SEM (n = 6).

lope production nearly 3-fold from 24 ± 3 x 103 to 68 ± 7 x 103 envelopes/105 cells. (Fig. 6). P decreased envelope production approximately 3.4-fold to 7 ± 1 X 10~3 envelopes/105 cells. In addition, 100 nM P was able to antagonize the increase in envelope production promoted by 10 nM DES (26 ± 4 x 103 envelopes/105 cells). These changes mimic the changes observed in vivo during the menstrual cycle (1, 3).

Discussion In recent years numerous studies have been aimed at understanding the function of various cell types within

404

HUMAN ECTOCERVICAL CELL CULTURES

the female reproductive tract (35, 36). A major problem is always the availability of a well characterized cell culture model system where the cells retain to a large extent their differentiated properties. In the present study we have used irradiated 3T3 cells as a feeder layer for the culture of human ECE cells. The 3T3 feeder layer system has been extensively used to culture other types of epithelial cells and are thought to assist epithelial cell growth by providing an extracellular surface for cell attachment and releasing factors that enhance growth (6,37). Feeder layers have previously been used to culture human ectocervical cells (38, 39); however, we have modified the culture conditions. First, our concentration of hydrocortisone is 4 ng/ml. In our experience this concentration is sufficient to maintain the cells, while higher concentrations tend to cause premature colony stratification and termination of growth. Second, we include insulin, transferrin, cholera toxin, and T 3 in the growth medium, hormones that are known to be essential for optimal epidermal keratinocyte growth (40). A major objective of these studies is to validate this cell culture system as a model for the study of regulation of ECE cell function by hormones. Development of such a system first requires positive identification that the cells are derived from the ectocervix and not from the endocervix or epidermis. We used several approaches to confirm the ectocervical nature of our cultures. As expected, the cultured cells did not react with antivimentin antibodies, vimentin being a specific marker of mesenchymal (fibroblast) cell types (41), but did react with antikeratin and antiinvolucrin antibodies (not shown) This indicates that the cells are epithelial, as keratins are only expressed in abundance in epithelial cells (41). Moreover, involucrin is a specific marker of surface epithelial cells (42, 43) and has been detected in vivo (44) in human ectocervix and in vitro in human ectocervical cells (Gorodeski, G. I., E. A. Rorke, and R. L. Eckert, unpublished). Keratins comprise a complex family of peptides, each encoded by a specific gene (45). They are expressed in a cell type-specific manner in all epithelial cells (24). Because keratin expression frequently varies in different epithelial cell types within an organ, keratin gene expression can be used as a fingerprint to ascertain the source of cultured cells (24, 25). To further confirm that our cultured cells are ectocervical, we determined which members of the keratin family are expressed in vitro for comparison to the pattern of expression previously reported in vivo (14, 23, 27). The keratin expression pattern in ectocervix is complex. Studies of in vivo expression describe expression of acidic (type I) keratins Kl, K2, K4, K5, and K6 and basic (type II) keratins K l l , K13, K14, K15, K16, K17, and K19 (23, 27-29). As shown in Table 1, only ectocer-

Endo • 1990 Vol 126 • No 1

vix, both in vivo and in vitro, express appreciable levels of K13. K13 is not expressed in epidermis, endocervix, or cultured epidermal keratinocytes (23, 27-29). Thus, K13 appears to be a marker of human ectocervical call differentiation. Our results, indicating the presence of K13 is ectocervical cells differ from those of Woodworth et al. (46), who found only low levels of K13 in their ectocervical cultures. We find that K13 is one of the major keratins expressed. This finding is supported by our identification of a specific transcript hybridizing to our K13-specific cDNA clone (19). The difference may be related to the fact that Woodworth et al. grew their cells in a defined serum-free medium in the absence of feeder support (46). If this is so, it appears that serumcontaining culture medium and the presence of feeder cells promote retention of an in uiuo-like differentiated phenotype, as measured by K13 expression. Based on these results, we propose that K13 is a marker of ECE cell differentiation. K19 is found in a variety of epithelial tissues and is a major keratin in many simple epithelia, but is only a minor component in stratifying epithelia (20, 23, 24). It is expressed at much higher levels in endocervix than ectocervix in vivo (27, 29) and is absent from epidermis (24). Thus, the pattern of K19 expression, endocervix > ectocervix > epidermis, supports the conclusion that our epithelial cell cultures are indeed derived from ectocervix. K6 and K16 were also expressed in our ectocervical cultures. These keratins are expressed in cultured human epidermal keratinocytes and in a variety of epidermal hyperproliferative diseases in vivo (47). Because their expression is associated with rapid in vitro and in vivo cell growth, K6 and K16 have come to be known as the hyperproliferative keratins (24). In vivo, K6 and K16 are present in normal ectocervix and in the squamous metaplastic epithelium of the transformation zone (the junction between the endo- and ectocervix) (29). This may be due to the tendency of the ectocervix to turnover and remodel rapidly during specific stages of the menstrual cycle (3, 4). In spite of the fact that keratins Kl, K2, and K4 are expressed in ectocervix in vivo, it is not surprising that these are not expressed in cultured ECE cells. These keratins are markers of terminal differentiation and are frequently not expressed in cultured cells. A precedent for this observation is provided by the cultures of human epidermal keratinocytes that do not express Kl or K2, although these keratins are expressed in epidermis in vivo (48). The cultured ECE cells also do not express cytokeratins K l l and K15, which are expressed in vivo. Thus, the keratin profile indicates beyond doubt that these cells are ECE cells. To be a useful system for the study of hormonal

HUMAN ECTOCERVICAL CELL CULTURES regulation of ECE cell differentiation, the cells must respond to physiological agents known to modulate ECE cell function in vivo. Two groups have previously described methods of growing cells derived from human ectocervix (38, 46). While these reports described conditions for growing the cells, the researchers did not test the cultures for responsiveness to physiological concentrations of sex steroids. The major hormones that regulate the female reproductive tract are E2 and P. In vivo, the ectocervical epithelium responds to changes in plasma levels of E2and P with changes in the architecture of the tissue and differentiation of the cells, including a change in the degree of desquamation (envelope release) from the surface of the ectocervix (3). During the proliferative phase of the cycle at the peak of E2 secretion, the ectocervical epithelium thickens, and the cells exfoliate in the form of superficial cells. During the luteal phase, under the influence of P, the ectocervical epithelium thins, and the rate of exfoliation drops (3). Thus, the ECE cells described in this study mimmic the in vivo ectocervix, since E2 promotes differentiation to more mature envelope forms, while P inhibits envelope maturation and stratification of the cultured ECE cells. In addition, the effects of DES are antagonized by P. It is worth noting that these cultures stratify and release envelopes spontaneously. This is in contrast to the situation in vivo, where the ectocervical and vaginal epithelial cells mature only in the presence of E2. Moreover, while many of the envelopes produced in culture contain pyknotic nuclei and thus resemble the in vivo superficial cell, most envelopes lacked nuclei. Enucleated envelopes are seen only rarely clinically, and their presence is associated with diseases such as hyperestrogenism, vitamin A deficiency, and cancer (34). The enucleated envelopes in vitro, therefore, may represent a more mature state of differentiation that is normally not reached in vivo. This suggests that ectocervical cells in vivo are regulated by other agents in addition to the sex steroids that inhibit the tendency of these cells to spontaneously desquamate. A possible class of agents that may play a role in the regulation of ECE cell differentiation is the retinoids. These compounds alter the differentiation of many different types of cells and tend to suppress the differentiation of stratifying surface epithelial cells (20, 26). We have identified a saturable, high affinity, low capacity E2-binding site with a Kd of 1.2 ±0.1 nM. The Kd is in the range of the Kd of the E2 receptor reported in a variety of E2-responsive tissues and cell types, including the ectocervix in vivo (49, 50). The binding site capacity for E2 (3-20 fmol/mg protein) is low compared to that of some E2-responsive cells, but is in the range found in the lower reproductive tract in women (5). The biological responsiveness of these cultures to DES concentrations

405

close to the K= of the estrogen receptor suggests that these effects are physiological and are likely to be estrogen receptor mediated. These results indicate that cultured ECE cells retain their responsiveness to the physiologically important sex steroid hormones and suggest that the regulation by E2 is mediated by a specific receptor. In conclusion, we have established conditions for growing human ECE cells and proven, using cytokeratin typing, that these cells are indeed ECE cells. These cells are phenotypically distinct from epidermal keratinocytes and endocervical cells; they retain a high degree of differentiated function and can be regulated by physiological levels of the appropriate sex steroid hormones. This system is likely to be useful for future studies on the effects of a variety of agents, including the sex steroids, on the female ectocervix.

Acknowledgment The authors wish to thank Ms. Sandy Hufeisen for her expert technical assistance.

References 1. Ferenczy A, Winkler B 1987 Anatomy and histology of the cervix. In: Kurman RJ (ed) Blaustein's Pathology of the Female Genital Tract, ed 3. Springer-Verlag, New York, p 141 2. Parmley T 1987 Embryology of the female genital tract. In: Kurman RJ (ed) Blaustein's Pathology of the Female Genital Tract, ed 3. Springer-Verlag, New York, p 1 3. Sedlis A, Chen P 1987 Cytology. In: Droegemueller W, Sciarra JJ (eds) Gynecology and Obstetrics. Harper and Row, Philadelphia, vol 1:1 4. Singer A 1975 The uterine cervix from adolescence to the menopause. Br J Obstet Gynecol 82:81 5. Sanborn BM, Kuo HS, Held B 1978 Estrogen and progesterone binding sites concentrations in human endometrium and cervix throughout the menstrual cycle and in tissues from women taking oral contraceptives. J Steroid Biochem 9:951 6. Rheinwald JG, Green H 1975 Serial cultivation of strains of human epidermal keratinocytes: the formation of keratinizing colonies from single cells. Cell 6:331 7. Wu YJ, Parker LM, Binder NE, Beckett MA, Sinard JH, Griffiths CT, Rheinwald JG 1982 The mesothelial keratins: a new family of cytoskeletal proteins identified in cultured mesothelial cells and nonkeratinizing epithelia. Cell 31:693 8. Brennan JH, Mansky J, Roberts G, Lichtman MA 1975 Improved methods for reducing calcium and magnesium concentrations in tissue culture medium: application to studies of lymphoblast proliferation in vitro. In Vitro 11:354 9. Rothblat GH, Arborgast LY, Ouellett L, Howard BV 1976 Preparation of delipidized serum protein for use in cell culture systems. In Vitro 12:554 10. Eckert RL, Katzenellenbogen BA 1982 Effects of estrogens and antiestrogens on estrogen receptor dynamics and the induction of progesterone receptor in MCF-7 human breast cancer cells. Cancer Res 42:139 11. Laemmli UK 1970 Cleavage of structural proteins during the assembly of the head of the bacteriophage T4. Nature 227:680 12. O'Farrell PZ, Goodman HM, O'Farrell, PH 1977 High resolution two-dimensional electrophoresis of basic as well as acidic proteins. Cell 12:1133 13. Southern EM 1975 Detection of specific sequences among DNA fragments separated by gel electrophoresis. J Mol Biol 98:503

406

HUMAN ECTOCERVICAL CELL CULTURES

14. Eichner R, Bonitz P, Sun TT 1984 Classification of epidermal keratins according to their immunoreactivity, isoelectric point, and mode of expression. J Cell Biol 98:1388 15. Maniatis T, Fritsch EF, Sambrook J 1982 Molecular Cloning-A Laboratory Manual. Cold Spring Harbor Laboratory, Cold Spring Harbor, New York 16. Aviv H, Leder P 1972 Purification of biologically active globin messenger RNA by chromatography on oligothymidylic acid-cellulose. Proc Natl Acad Sci USA 69:1408 17. Boedtker H 1971 Conformation independent molecular weight determinations of RNA by gel electrophoresis. Biochim Biophys Acta 240:448 18. Church GM, Gilbert W 1984 Genomic sequencing. Proc Natl Acad Sci USA 81:1991 19. Eckert RL, Green H 1984 Cloning of cDNAs specifying vitamin A responsive human keratins. Proc Natl Acad Sci USA 81:4321 20. Gilfix BM, Eckert RL 1985 Coordinate control by vitamin A of keratin gene expression in human keratinocytes. J Biol Chem 260:14026 21. Rigby PWJ, Dieckman M, Rhodes C, Berg P 1977 Labeling deoxyribonucleic acid to high specific activity in vitro by nick translation with DNA polymerase I. J Mol Biol 113:237 22. Scatchard G 1949 The attraction of proteins for small molecules and ions. Ann NY Acad Sci 51:660 23. Moll R. Franke WW, Schiller DL, Geiger B, Krepler R 1982 The catalog of human cytokeratins: patterns of expression in normal epithelia tumors and cultured cells. Cell 31:11 24. Sun TT, Tseng SCG, Huang AJW, Cooper D, Schermer A, Lynch MH, Weiss R, Eichner R 1985 Monoclonal antibody studies of mammalian epithelial keratins. Ann NY Acad Sci 455:307 25. Cooper D, Schermer A, Sun TT 1985 Classification of human epithelia and their neoplasms using monoclonal antibodies to keratins. Strategies, applications and limitations. Lab Invest 52:243 26. Eckert RL, Rorke EA 1989 Molecular biology of keratinocyte differentiation. Environ Health Perspect 80:109 27. Moll R, Levy R, Czernobilsky B, Hohlweg-Majert P, DallenbachHellweg G, Franke WW 1983 Cytokeratins of normal epithelia and some neoplasms of the female genital tract. Lab Invest 49:599 28. Dixon IS, Stanley MA 1984 Immunofluorescent studies of human cervical epithelia in vivo and in vitro using antibodies against specific keratin components. Mol Biol Med 2:37 29. Gigi-Leithner 0, Geiger B, Levy R, Czernobilsky B 1986 Cytokeratin expression in squamous metaplasia of the human uterine cervix, Differentiation 31:191 30. Nischt R, Roop DR, Mehrel T, Yuspa SH, Rentrop M, Winter H, Schweizer J 1988 Aberrant expression during two-stage mouse skin carcinogenesis of a type 147-kDa keratin, K13, normally associated with terminal differentiation of internal stratified epithelia. Mol Carcinogenesis 1:96 31. Gorodeski IG, Geier A, Lunenfeld B, Beery R, Bahary CM 1987 Progesterone (P) receptor dynamics in estrogen primed normal human cervix following P injection. Fertil Steril 47:108

Endo • 1990 Voll26«Nol

32. Gorski J, Gannon F1976 Current models of steroid hormone action: a critique. Annu Rev Physiol 38:425 33. Welshons WV, Leiberman ME, Gorski J 1984 Nuclear localization of unoccupied estrogen receptors. Nature 307:747 34. Novak ER, Woodruff JD 1979 Novak's Gynecologic and Obstetric Pathology, ed 8. Saunders, Philadelphia, p 82 35. Sengupta J, Given RL, Carey JB, Weitlauf HM 1987 Primary culture of mouse endometrium on floating collagen gels: a potential in vitro model for implantation. Ann NY Acad Sci 476:75 36. Leavitt WW 1987 Cell and molecular biology of the uterus. In: Leavitt WW (ed) Advances in Experimental Medicine and Biology. Plenium Press, New York, vol 230:1 37. Green H, Rheinwald JG, Sun TT 1977 Properties of an epithelial cell type in culture: the epidermal keratinocyte and its dependence on products of fibroblasts. In: Revel JP, Henning U, Fox CF (ed) Cell Shape and Surface Architecture. Liss, New York, p 493 38. Stanley MA, Parkinson EH 1979 Growth requirements of human cervical epithelial cells in culture. Int J Cancer 24:407 39. Stanley MA, Dahlenburg H 1984 The effect of choleragen and epidermal growth factor on proliferation and maturation in vitro of human ectocervical cells. In Vitro 20:144 40. Tsao MC, Walthall BJ, Ham RG 1982 Clonal growth of normal human epidermal keratinocytes in a defined medium. J Cell Physiol 110:219 41. Lazarides E 1982 Intermediate filaments: a chemically heterogeneous, developmentally regulated class of proteins. Annu Rev Biochem 51:219 42. Rice RH, Green H 1977 The cornified envelope of terminally differentiated human epidermal keratinocytes consists of crosslinked protein. Cell 11:417 43. Rice RH, Green H 1979 Presence in human epidermal cells of a soluble protein precursor of the cross-linked envelope: activation of the cross-linking by calcium ions. Cell 18:681 44. Sassoon AF, Said JW, Nash G, Shintaku IP, Banks-Schlegel S 1985 Involucrin in intraepithelial and invasive cell carcinomas of the cervix: an immunohistochemical study. Hum Pathol 16:467 45. Steinert PM, Roop DR 1988 Molecular and cellular biology of intermediate filaments. Annu Rev Biochem 57:593 46. Woodworth CD, Bowden PE, Doniger J, Pirisi L, Barnes W, Lancaster WD, DiPaolo JA 1988 Characterization of normal human exocervical epithelial cells immortalized in vitro by papillomavirus types 16 and 18 DNA. Cancer Res 48:4620 47. Weiss RA, Eichner R, Sun T-T 1984 Monoclonal antibody analysis of keratin expression in epidermal diseases: a 48-and 56-kdalton keratin as molecular markers for hyperproliferative keratinocytes. J Cell Biol 98:1397 48. Nelson WG, Sun TT 1983 The 50- and 58-kdalton keratin classes as molecular markers for stratified squamous epithelia: cell culture studies. J Cell Biol 97:244 49. Katzenellenbogen BS 1980 Dynamics of steroid hormone receptor action. Annu Rev Physiol 42:17 50. Sanborn BM, Held B, Hsu SH 1975 Specific estrogen binding proteins in human cervix. J Steroid Biochem 6:1107