In VitroCell.Dev.Biol.~Anima135:339-345,June 1999 © 1999Societyfor In VitroBiology I071-2690/99 $05.00+ 0.00

CHARACTERIZATION OF A CELL LINE ESTABLISHED FROM DIETHYLSTILBESTROLI N D U C E D RENAL TUMORS IN SYRIAN HAMSTERS 1 GUY LAURENT,2 DENIS NONCLERCQ,FABRICE JOURNE, REGINE BROHEE, GERARD TOUBEAU, PAUL FALMAGNE,AND JEANINE-ANNE HEUSON-STIENNON

Laboratory of Histology and Experimental Cytology, Faculty of Medicine and Pharmacy (G. L., D. N., R. B., G. T., J.-A. H.-S.) and Laboratory of Biological Chemistry, Faculty of Sciences (E J., P. E), Universit~de Mons-Hainaut, Mons, Belgium (Received 19 October 1998; accepted 4 February 1999)

SUMMARY This article describes HKT-1097, a new cell line established from renal tumors induced by the protracted administration of diethylstilbestrol (DES) to male Syrian golden hamsters. Cell culture was initiated from tumor samples obtained from two 14-mo.-old animals which had undergone exposure to DES for a period of 11 mo. The HKT-1097 cell line was characterized between Passages 16 and 22 with respect to cell morphology, growth properties, karyology, and the presence of estrogen receptors. Moreover, immunostaining with a panel of antisera was performed to identify the cytological profile of the cell line and establish a parallel with tumor tissue in vivo. HKT-1097 cells are fibroblastoid; their most distinctive feature is that they exhibit strikingly long processes. The HKT-1097 cell line grows as a monolayer with a tendency toward a less stringent density-dependent inhibition of growth. The modal chromosome number is 44, but more than 50% of the cells are aneuploid, suggesting a substantial degree of karyotype instability. HKT-1097 cells express estrogen receptors. They contain immunoreactive vimentin and desmin, but appear negative upon cytokeratin irnmunostaining. In addition, these ceils express glial fibrillary acidic protein and other markers of the neuroeetodermal lineage, but lack neurofilament protein. Insofar as the same lineage markers have been demonstrated in DES-induced Syrian hamster kidney tumors (SHKT), we conclude that HKT-1097 cells retain some of the original tumor cell phenotype. The current observations suggest that estrogen-induced SHKT derive from the renal interstitium and point to an involvement of neuroeetodermal cells in the development of these neoplasms.

Key words: HKT-1097; hormonal carcinogenesis; estrogen; kidney; intermediate filaments. onstration of various cell lineage markers in DES-induced SHKT point to a major involvement of undifferentiated interstitial cells in the process of tumorigenesis (4,22). The use of estrogen-induced SHKT as an experimental model of renal or hormonal carcinogenesis is hampered by the duration of estrogen treatment which has to be administered, as well as by the small amounts of tumor tissue samples that can generally be recovered at necropsy, even though long-term administration of an estrogen such as DES leads to the appearance of renal neoplasms in virtually 100% of cases (21). Thus, further exploration of this model would benefit from further in vitro investigations on tumor cell cultures. So far there have been few attempts to develop cell lines from estrogeninduced SHKT. This paper describes a new cell line established from DES-induced SHKT.

INTRODUCTION Renal tumors induced by administration of diethylstilbestrol (DES) or other estrogens to male Syrian golden hamsters, originally described by Kirkman and Bacon (14) almost 50 yr ago, remain a reference model in experimental studies on hormonal carcinogenesis (18). In DES-treated animals, distinguishable kidney tumors start appearing approximately 6 too. after the beginning of estrogen exposure. It still remains unclear whether tumorigenesis results from genotoxie properties of estrogen metabolites (18) or must rather be attributed to epigenetie mechanisms related to chronic tubular injury (16). In spite of the considerable work devoted to the histogenesis of estrogen-induced Syrian hamster kidney tumors (SHKT), the issue still appears incompletely resolved. The morphological analysis of preneoplastic alterations in hamster kidneys exposed to estrogen suggests that renal tumors primarily derive from proximal tubules (10). On the other hand, data collected from immunocytoehemical dem-

MATERIALSAND METHODS

Animal treatment and tissue sampling. In vivo experiments were performed on male Syrian golden hamsters (Mesocricetus auratus) in compliance with the guidelines specified by the Belgian Ministry of Trade and Agriculture. Renal carcinogenesis was induced by chronic exposure to diethylstilbestrol (DES) (Sigma-Aldrich, Bornem, Belgium). DES treatment was started on 3mo.-old animals, following a procedure detailed in previous publications (20.21). Briefly, implants were prepared by filling 2.5-cm-long segments of Silastie ® tubing (Dow Coming, Midland, MI) with 25 mg of DES, and sealing them at both extremities with medical grade silicon glue. These implants

1The cell line described in this article has been deposited with the European Collection of Cell Cultures with the accession number 98061003. 2To whom correspondence should be addressed at Laboratory of Histology and Experimental Cytology, Faculty of Medicine and Pharmacy, Pentagone 1B, Universit~ de Mons-Hainaut, 6 avenue du Champ de Mars, B7000 Mons, Belgium. E-mail: [email protected] 339

340

LAURENT ET AL.

were inserted subcutaneously in the shoulder region of anesthetized animals and renewed every 2 mo. to maintain a sufficient blood level of DES. Under these experimental conditions, DES concentration in the blood was approximately 13 ng/ml (21). Necropsies of renal tumors were made from two 14mo.-old individuals which had been exposed to DES for 11 mo. In both cases, tumors were clearly apparent upon laparotomy due to their whitish appearance and their sizes which ranged between 0.1 and 0.3 cm z. Tumor tissue samples were excised by sharp dissection under aseptic conditions and immediately immersed in sterile Dulbecco's phosphate-buffered saline (DPBS). Tissue disaggregation and initiation of cell cultures. Tumor necropsies were minced with scissors in DPBS and transferred to 12-well plates containing a solution of 3 mg Clostridium histolyticum collagenase type I (Life Technologies, Merelbeke, Belgium) per ml in Dulbecco's modified Eagle medium (DMEM) (see below) prewarmed at 37 ° C. Tissue fragments in suspension were incubated for 15 rain on a shaker set at low speed. The process was repeated twice; the medium was replaced with fresh enzyme solution between each run. At the end of the procedure, tissue fragments were further disaggregated by several passages through the 1-ml tip of a micropipette and plated in 25-cm=flasks containing DMEM (BioWhittaker Europe, Verviers. Belgium) supplemented with 10% fetal bovine serum, 25 mM N-2-hydroxyethylpiperazine-N'-2-ethanesulfonic acid (HEPES), 2 mM glutamine, 100 U penicillin per ml, 100 p~g streptomycin per ml, 0.25 p.g amphotericin B per ml, 100 ng DES per ml, and 5 ng a-melanocyte stimulating hormone (a-MSH) (Biogenesis, Poole, UK) per ml. Cell cultures were maintained under conditions detailed below. Evidence of fast and steady growth was seen 6 mo. after initiation of cell cultures. During that period of adaptation to in vitro conditions, cells grew slowly and were passed every 3-4 wk. Maintenance of cell cultures. The cell line established from DES-induced SHKT and described in this paper was christened HKT-1097. It was characterized between passages 16 and 22. Cell cultures were performed at 37 ° C in a 5% CO2, humidified cell incubator. For routine propagation, ceils were cultured in 75-cm2 flasks containing DMEM with supplements as indicated above. DES and c~-MSH have been included in the formulation of the medium since culture initiation in order to mimic the hormonal environment prevailing in vivo at the time of renal carcinogenesis. Indeed, DES treatment not only induces kidney tumors but also causes a substantial elevation of serum ct-MSH, and the latter might contribute to the development of renal neoplasms (15). However, control experiments have shown that cell growth can be maintained without these additives. Cells were subcuhured once a wk at a density of 0.5 X 104 cells/cm2, with one or two renewals of the culture medium between two consecutive passages. For passages and measurement of growth, ceils were rinsed with DPBS and isotonic buffered EDTA solution (Versene), and dislodged from the vessel bottom by treatment with trypsin/ EDTA solution. Concentrations of cells in suspensions were measured in a model Z1 cell counter (Coulter Electronics, Luton, UK). All culture reagents other than DMEM were from Life Technologies (Merelbeke, Belgium). Chromosome preparation. Cells were plated at a density of 104 cells/cm2 in 75-cm2 flasks. Six d after seeding, metaphase arrest was induced by addition of colcemid (Boehringer Mannheim Belgium, Brussels) to the culture medium at a final concentration of 10 7 M. Cells were trypsinized after 6 h of exposure to colcemid and chromosome spreads were prepared following the procedure described by Freshney (9), except that for chromosome staining, Giemsa's stain was replaced by Pappenbeim's panchrome. Western blot analysis of cell proteins. Cell monolayers were rinsed with DPBS and incubated with Versene for 5 min at 37 ° C. The ceils were harvested with a cell scraper and suspended by vigorous pipetting. Cell concentration was determined and the cell suspensions were centrifuged for 15 min at 220 × g. The pellets were rinsed with DPBS and the suspensions were centrifuged again. The supernatants were discarded and the cell pellets were immediately frozen at - 80 ° C. The cells were lysed in 100 nsl//phosphate buffer, pH 7.4, containing 1% Nonidet P40, 0.1% sodium dodecyl sulfate (SDS), 1% sodium deoxycholate and a freshly added cocktail of protease inhibitors (Complete TM, Roche Diagnostics Belgimn, Brussels, Belgium), with a Potter-Elvehjem tissue grinder equipped with a teflon pestle. The cell lysates were centrifuged for 15 min at 13 000 × g and the protein content in the supernatants (cell extracts) was measured with the BCA assay kit (Pierce, Rockford. IL). Protein aliquots (20 gg) from the cell extracts were separated under reducing conditions by SDS-polyacrylamide gel electrophoresis in 10% polyacrylamide gels and then transferred by electroblotting to nitrocellulose membranes. Biotinylated molecular weight markers (Sigma Chemical Co., St. Louis, MO) were used for calibration. Nonspecific binding sites on the mem-

TABLE 1 PRIMARY ANTISERA TESTED ON HKT-1097 CELLS" Antiserum

Origin

Anti-mouse estrogen receptor b Anti-porcine vimentin Anti-human desmin Anti-cow GFAP

Rabbit, Mouse, Mouse, Rabbit,

Anti-human cytokeratinc (clone MNFll6)

Mouse, monoclonal

1:20

Rabbit, polyclonal Rabbit, polyclonal Rabbit, polyclonal

1:300 Prediluted 1:100

Mouse, monoclonal Rabbit, polyclonal

1:40 1:50

Anti-human PGP 9.5 Anti-human NSE Anti-cow S100 protein Anti-human neurofilament protein (clone 2 F l l ) Anti-human TGF-alpha

polyclonal monoclonal monoclonal polyclonal

Dilution 1:100 1:20 1:20 1:200

~All antisera were checked for interspecies cross-reactivity by examining the pattern of immunostaining in tissue sections of brain and intestine of Syrian hamsters. bDirected against a polypeptide comprising residues 580-599 at the carboxy terminus of the mouse estrogen receptor. ~Directed against keratins 5, 6, 8, 17, according to information from the supplier.

branes were blocked with 5% nonfat milk in 10 mM Tris, 150 mM NaC1, pH 8.0 (Tris buffered saline) containing 0.05% Tween-20. The membranes were incubated overnight with polyclonal rabbit anti-estrogen receptor (ER) antibody (Santa Cruz Biotechnology, Santa Cruz, CA). This was followed by exposure to a donkey anti-rabbit IgG conjugated with horseradish peroxidase (Amersham Belgium, Gent, Belgium). Finally, the immunoblots were visualized with BM Chemiluminescence Blotting Substrate (Roche Diagnostics Belgium. Brussels, Belgium). The specificity of immunoblotting was demonstrated by preincubating the primary antibody in the presence of the peptide which had been used as an immuuogen (Santa Cruz Biotechnology) (data not shown). Cell morphology and irnmunocytochemistry. Sterile round glass coverslips (12-mm diameter) placed in 12-well plates were seeded with ceils at a density of 0.5-1 X 104 cells/cm2. Four d after plating, cell monolayers were rinsed with DPBS and fixed for 15 min with ice-cold Carnoy's mixture. The fixative was discarded and replaced by 70% ethanol. The cultures were kept at 4 ° C in ethanol for a maximum of 16 h until histological or immunocytochemical staining. For histological staining, we used a combination of May-Grttnwald's and Pappenheim's solutions following routine procedure. For immunocytochemistry, fixed cell cultures were rinsed several times with phosphate-buffered saline (PBS) (0.04 M Na2HPQ, 0.01 M KHzPO,, 0.12 M NaC1, pH 7.2), and preincubated for 20 min in PBS containing 5% normal goat serum and 0.05 M NH4C1 to prevent nonspecific adsorption of immunoglobulins during immunostaining. Thereafter, the coverslips were inverted and layered face down on drops of primary antisera distributed on parafilm. A similar method of incubation was applied during subsequent steps of the immunostaining protocol. The cultures were incubated for 1 h in the presence of the primary antibodies. The primary antisera used for cell characterization are listed in Table 1. Anti-estrogen receptor, anti-PGP 9.5 and anti-TGF-a came from Santa Cruz Biotechnology, Biogenesis (Poole, UK) and Pepro Tech (Rocky Hill, NJ), respectively. All other primary antisera were from Dakopans (Glostrup, Denmark). After exposure to the primary antibody, the cell monolayers were incubated for 30 rain with a dextran polymer conjugated with both peroxidase and antibodies raised against rabbit or mouse immunoglobulins (EnVision TM, Dakopatts) (32). This immunostaining procedure was prefen'ed to standard avidin-biotin techniques since it proved to be more sensitive and gave better signals for cell examination and microphntography (24). Exposure to the dextran polymer conjugate was followed by a 30min incubation with rabbit anti-peroxidase antiserum (Laboratory of Hol~ monology, Marloie, Belgium). Cell preparations were finally exposed for 30 min to swine anti-rabbit immunoglobuliu antibodies conjugated with fluorescein isothiocyanate (FITC) (Dakopatts), or to biotiuylated swine anti-rabbit

341

RENAL TUMOR-DERIVED CELL LINE

immunoglobulin antibodies (Dakopatts). In the latter case, immunostaining was achieved by incubating the preparations with Texas Red-labeled streptavidin (Sigma-Aldrich, Bornem, Belgium) for 30 min. After final rinses in PBS and distilled water, the coverslips were mounted on microscope slides with Mowiol mountant. Immunostained cell preparations were examined on a Leitz Orthoplan fluorescence microscope equipped with a Ploem system for epiillumination. Wavelengths of 490 nm and 596 nm for excitation and 525 nm and 615 nm for emission were used for the observation of FITC and Texas Red, respectively.

RESULTS When growing at low density, HKT-1097 cells were generally fibroblast-like, although the frequent occurrence of multipolar or polygonal cells conferred to cultures a polymorphic appearance (Fig. 1 A,B). Of note, dipolar or muhipolar cells often exhibited exquisitely long processes (Fig. 1 B) which became particularly obvious after immunostaining with antisera raised against intermediate filaments (Fig. 5 D). Occasionally, such processes seemed to connect neighboring cells. At confluency, fibroblast-like cells formed a monolayer, often disposing themselves in parallel arrays around polygonal cells (Fig. 1 C). The monolayers tended to become more tightly packed when the saturation density of the cultures increased, but cells never grew as muhilayers. Typical growth curves of HKT-1097 cells are illustrated in Fig. 2. After seeding, a short (approx. 27 h) lag phase was followed by exponential growth characterized by a population doubling time of 15 h. Amazingly, with an increasing number of passages the duration of exponential growth lengthened without apparent change in the population doubling time. Thus, cells reached the plateau phase at higher saturation densities (e.g., 15 × 104 cells/cm 2 instead of 6 X 104 cells/cm 2) (Fig. 2). Such behavior indicated that from generation to generation (i.e., over six passages, corresponding to approximately 20 population doublings) HKT-1097 cells became progressively less sensitive to density-dependent inhibition of growth. Of note, we found recently that HKT-1097 ceils could proliferate on the surface of non cell culture-treated petri dishes wherein they reached saturation densities equivalent to those observed in cell culture vessels (data not shown). Similar findings have been reported for H301, another cell line previously established from DES-induced SHKT (28).

FIG. 1. Appearance of HKT-1097 cells under light microscopy. A, Histological staining; B,C, Hoffmann modulation contrast. Subconfluent cells (A,B) with a fibroblast-like morphology often exhibit long, slender processes. Occasionally, such processes seem to connect neighboring cells. At confluency (C), cells frequently assume a parallel orientation in the monolayer. Bars: A, 50 ~tm; B,C,, 60 IXm.

200

Passage 22

Passage 16

? 150

1 O0

50

2

4

6

,'o

I e/

. 2

Time after ~lating (days)

.

. 4

. 6

8

FIG. 2. Growth curves of HKT-1097 cells. Ceils were plated at a density of 104 cells/cm2 in 12-well plates and harvested by trypsinization at different time points after seeding. Cell cultures were fed fresh medium at days 1 and 4. Vertical bars indicate standard deviations of four determinations.

342

LAURENT ET AL. Since HKT-1097 cells derive from estrogen-induced tumors, they should be expected to contain ERs. As shown in Fig. 4 A, Western blot analysis of the cell proteins detected the presence of a 67-kDa protein with ER immunoreactivity. This result was confirmed by cell immunostaining with the same anti-ER antiserum (Fig. 4 B). The resuhs of cell immunostaining for other markers are illustrated in Fig. 5. Immunostained cell preparations with the best signal-tonoise ratio were selected for illustration, even though similar results were obtained whichever fluorescent label (FITC-conjugated antirabbit immunoglobulins or Texas Red-labeled streptavidin) was used to visualize the presence of immunoreactive material. The presence of TGF-c~ immunoreactivity in HKT-1097 cells was detected by immunocytochemical staining with anti-TGF-c~ antiserum (Fig. 5 B). Virtually all cells were immunostained after application of anti-vimentin antiserum (Fig. 5 C), whereas anti-desmin (Fig. 5 D) and anti-glial fibrillary acidic protein (GFAP) antisera (Fig. 5 E) decorated the cytoskeleton in only a subset of the cell population. By contrast, no immunostaining was observed with either anti-cytokeratins or anti-neurofilament antiserum (data not shown). Cell expression of PGP 9.5 (Fig. 5 F), NSE (Fig. 5 G) and S100 (Fig. 5 H) was also demonstrated by using the corresponding antisera. In that case, the signal was often weaker, as compared to that generated by antisera raised against intermediate filaments. No immunostaining was observed in negative controls in which the primary antiserum had been omitted or replaced by nonimmune rabbit serum (Fig. 5 A). DISCUSSION

FIG. 3. Metaphase spreads of HKT-1097 cells prepared as detailed in Methods. A, Euploid cell (n = 44); B, Hypodiploid cell (n = 35); C, hypotetraploid cell (n = 78). Bar, 10 p.m.

The distribution of chromosome numbers was evaluated in 63 metaphase cells. The modal chromosome number was 44 (Fig. 3 A), occurring in 41% of ceils. Aneuploid cells were mostly hypodiploid (Fig. 3 B), although metaphases ranging from hyperdiploidy to hypotetraploidy (Fig. 3 C) were also encountered. Polyploidy (n > 88) was present in approximately 2% of all cases.

1

58.1 Idea-

2

3

The present paper reports the description of HKT-1097, a cell line recently established from DES-induced SHKT. HKT-1097 cells were characterized with respect to morphology, karyology, growth properties, and the expression of ER. Immunocytochemical staining with a panel of antisera also enabled us to identify the relationship between this new cell line and DES-induced SHKT growing in vivo. Chromosome analysis of HKT-1097 ceils disclosed a substantial degree of aneuploidy. As shown in previous work (17), the chromosome content of kidney cells isolated from untreated Syrian hamsters shows little departure from the normal Mesocricetus auratus karyotype (2n = 44). By contrast, treatment with DES leads to a marked elevation of the frequency of near diploid and near tetraploid cells in renal tissue (17). Since HKT-1097 cells derive from tumor tissue, it is difficult to ascertain whether the karyotype instability seen in this cell line is a direct consequence of exposure to DES (30) or is a manifestation of the transformed phenotype. Anyway, it is worth noting that chromosome imbalance has been put forward along with

FIG. 4. Demonstration of estrogen receptors in HKT-1097 cells by A, Western blotting, and B, immunofluorescence. A, Cell proteins were extracted, separated by SDS-PAGE, and submitted to immunoblot analysis, as detailed in Methods. Lane 1, biotinylated molecular weight markers; lane 2, 20-~tg proteins from HKT-1097 cells; lane 3, 2-ng recombinant human estrogen receptor, used as a reference. B, Cells were fixed, and immunostained with the same antibody as that used for immunoblotring. Bar, 20 p.m.

RENAL TUMOR-DERIVED CELL LINE

FIG. 5. HKT-1097 immunostaining for TGF-a (B), vimentin (C), desmin (D), GFAP (E), PGP 9.5 (F), NSE (G), and S100 (H). Note the fibrillar appearance of immunostaining patterns produced by antisera decorating intermediate filaments (C-E) and the more granular appearance of immunostaining obtained after application of other antisera. In D, conspicuous cell processes are highlighted by anti-desmin antiserum. A, Control preparation in which nonimmune rabbit serum was substituted for the primary antiserum. Only nuclei can be distinguished, due to their autofluorescence visible in the fluorescein channel. A,C,G: FITC labeling; B,D-F,H: Texas Red labeling. Bars: C,D, 40 ~m; A,B,E-H, 20 gm.

regenerative hyperplasia as a major factor contributing to estrogen carcinogenesis in the hamster kidney (16), To our knowledge, there are in the literature only two reports of long-term cell cultures having been established from estrogen-induced SHKT. In the early study of Sirbasku and Kirkland (28), animal passage was alternated with in vitro tumor cell culture to induce the selection of a permanent cell line that the authors designated H301. More recently, Goazalez et al. (11) have used a serum-free medium of optimized composition for the primary culture and propagation of tumor ceils from SHKT induced by DES. However, those cell lines described in previous publications differ from the one characterized in this study. Whereas HKT-1097 cells grow as a monolayer, the tumor cell cultures derived by Gonzalez et al. (11) formed muhilayer colonies which eventually developed into viable anchorage-independent spheroids. Moreover, these cultures of tumor cells could not be maintained for an extended period of time since they became senescent and stopped growing after 6 mo. This is not the

343

case for HKT-1097 cell line which was established 15 mo. ago and has successfully been through more than 35 passages since its initiation. In this respect, HKT-1097 more resembles the H-301 cell line which has undergone immortalization and which after its derivation has been used as an in vitro model in several studies on estrogen carcinogenesis (e.g., 2,29). On the basis of morphology, however, HKT-1097 ceils differ from H-301 cells since the latter show an epithelial morphology (28) whereas the former are rather fibroblastoid. Nonetheless, HKT-1097 cells share with H-301 the ability to grow on the surface of non cell culture-treated petri dishes. According to Sirbasku and Kirkland (28), normal hamster kidney fibroblasts lack this property. The fact that the same type of tumor gave rise to dissimilar cell lines is not all that surprising considering the peculiar pleomorphism of estrogen-induced SHKT. These tumors were originally classified as adenomas or adenoeareinomas (14), hence the hypothesis that they originate from proximal tubule dysplasias (10). Accordingly, the initial selection of epithelial cells to produce the H-301 cell line was based on the assumption that estrogen-induced SHKT are primarily adenoeareinomas (28). However, more recent work relying on thorough histological examination as well as on immunocytoehemieal staining with a panel of antisera used in routine diagnostics has demonstrated in DES-induced SHKT a variety of histological patterns (blastemal, sareomatoid, and neuroeetodermal) besides the epithelial morphology (4). These observations, as well as studies based on histochemistry (12) or the immunolocalization of estrogen receptors (1) have led investigators to consider for estrogen-induced SHKT an interstitial-mesenchymal origin rather than a tubular one. Furthermore, HKT-1097 cells exhibit ERs, whereas this protein is not expressed in the tubular epithelium of the hamster kidney (1). In view of the complex and still debated histogenesis of estrogeninduced SHKT, immunocytochemistry was used for collecting more information about the phenotype of the HKT-1097 cells. These cells show positive immunostaining after application of anti- TGF-ot antiserum. TGF-c~ is a poor indicator of the cell lineage, insofar as this growth factor is widely distributed among different tissues i~xespective of the cell type (3). Nevertheless, previous work of our group has demonstrated strong immunostaining for TGF-ct in a subpopulation of cells of DES-induced SHKT (33). Thus, the fact that HKT-1097 cells express TGF-o~ indicates that this cell line has retained the phenotype exhibited by tumor ceils in vivo. Antisera directed against intermediate filaments such as vimentin, desmin, and cytokeratins are currently used as diagnostic tools to determine cell lineage, particularly in neoplasms (23). HKT-1097 cells show positive vimentin immunostaining. Vimentin has been reported to be a marker of nonmuscle ceils of a mesenchymal type (8). However, this intermediate filament has been demonstrated in a variety of cell lines (23), including epithelial ceils such as NRK-52E and LLC-PKt derived from renal tubules (unpublished observations). On the other hand, the absence of cytokeratin expression by HKT-1097 cells argues against an epithelial origin insofar as eytokeratins are found in most, if not all epithelial cells (5,23). Some but not all HKT-1097 cells exhibit strong desmin immunostaining. Desmin typifies muscle cells and is also present in podoeytes and/or mesangial cells of renal corpuscles (13,34), whereas it is not found in renal fibroblasts (6). Few cell lines express desmin, although this intermediate filament is observed in the BHK-21 cell line established from the hamster kidney (31). Once again, desmin expression by HKT-1097 cells points to a nonepithelial origin.

344

LAURENT ET AL.

Finally, HKT-1097 cells were tested for the presence of two intermediate filaments considered to be neuroectodermal m a r k e r s - neurofilament and glial fibrilla13, acidic protein (GFAP)--as well as for the expression of other proteins associated with the neuroectodermal lineage. While neurofilament immunostaining was negative, immunocytochemical staining detected the presence of PGP 9.5, a marker of neuronal tissue and primitive neuroectodermal tumors (7). Besides, HKT-1097 expressed GFAP, a constituent of the cytoskeleton of astrocytes (27). They were also found to contain neuronspecific enolase and S-100 protein, two other markers of glial cells (25,26). The immunocytochemical demonstration of vimentin, desmin, neuron-specific enolase, and S-100 protein in estrogen-induced SHKT has been reported previously (4,12), and has been confirmed by recent studies of our group. In addition, we have also found that these tumors express PGP 9.5 and GFAP. In this context, S-100 protein and GFAP seem to be the most reliable markers of tumor cells since they are not detected in nonneoplastic renal parenchyma of DES-treated animals (manuscript in preparation). Altogether data obtained by protein immunoblotting and cell immunostaining indicate a close relationship between HKT-1097 cells and the DESinduced SHKT from which they derive, they also lend further support to the hypothesis of an interstitial origin for DES-induced SHKT, as already suggested by previous studies (19), and point to involvement of neuroectodermal cells in the development of these neoplasms. ACKNOWLEDGMENTS This study received financial support from the Belgian Fund for Medical Scientific Research (Grant 3.4551.86). G. Laurent and R. Broh6e are Senior Research Associate and Research Assistant, respectively, of the National Fund for Scientific Research (Belgium). E Journ6 was the recipient of a fellowship (n° 7.4522.97, <<: T616vie 1997 > > ) from the National Fund for Scientific Research. Recombinant human ER (KARO-BIO, Sweden) was a kind gift from Doctor G. Leclercq (Laboratoire J.-C. Heuson de Canc4rologie Mammaire, Service de M6decine Interne, Institut Jules Bordet, Brussels). The expert technical assistance of J. Noal is gratefully acknowledged. REFERENCES 1. Bhat, H. K.; Hacker, H. J.; Bannasch, P.; Thompson, E. A.; Liehr, J. G. Localization of estrogen receptors in interstitial cells of hamster kidney and in estradioMnduced renal tumors as evidence of the mesenchymal origin of this neoplasm. Cancer Res. 53:5447-5451; 1993. 2. Bursch, W.; Liehr, J. G.; Sirbasku, D. A.; Putz, B.; Taper, H.; SchuheHermann, R. Control of cell death (apoptosis) by diethylstilbestrol in an estrogen-dependent kidney tumor. Carcinogenesis 12:855~60; 1991. 3. Christensen, M. E.; Poulsen, S. S. Immunolocalization of transforming growth factor alpha in normal human tissue. Histochem. Cell Biol. 105:391-400; 1996. 4. Cort6s-Vizcafno, V.; Peydr6-Olaya, A.; Llombart-Bosch, A. Morphological and immunohistochemical support for the interstitial cell origin of oestrogen-induced kidney tumours in the Syrian golden hamster. Carcinogenesis 15:2155-2162; 1994. 5. Corwin, D. J.; Gown, A. M. Review of selected lineage-directed antibodies useful in routinely processed tissues. Arch. Pathol. Lab. Med. 113:645~552; 1989. 6. Ennulat, D.; Steffens, W. L.; Brown, S. A. Desmin expression in mesangial cells and fibroblasts in vitro. In Vitro Cell. Dev. Biol. 34:450454; 1998. 7. Ermisch, B.; Schwechheimer, K. Protein gene product (PGP) 9.5 in diagnostic (neuro-) oncology. An immunomorphological study. Clin. Neuropathol. 14:130-136; 1995. 8. Franke, W. W.; Schmid, E.; Osborn, M.; Weber, K. Different intermediatesized filaments distinguished by immunofluorescence microscopy. Proc. Natl. Acad. Sci. USA 75:5034-5038; 1978.

9. Freshney, R. Culture of animal cells. A manual of basic technique. 3rd ed. New York: Wiley-Liss; 1994:205-206. 10. Goldfarb, S.; Pugh, T. D. Morphology and anatomic localization of renal microneoplasms and proximal tubule dysplasias induced by four different estrogens in the hamster. Cancer Res. 50:113-119; 1990. 11. Gonzalez, A.; Obedey, T. D.; Sehultz, J. L.; Ostrom, J.; Li, J. J. In vitro characterization of estrogen induced Syrian hamster renal tumors: comparison with an immortalized cell line derived from diethylstilbestrol-treated adult hamster kidney. In Vitro Cell. Dev. Biol. 29A:562-573; 1993. 12. Hacker, H. G.; Bannasch, P.; Liehr, J. G. Histochemical analysis of the development of estradiol-induced kidney tumors in male Syrian hamsters. Cancer Res. 48:971-976; 1988. 13. Holthofer, H.; Sainio, K.; Miettinen, A. The glomerular mesangium: studies of its developmental origin and markers in vivo and in vitro. Acta Pathologica Microbiologica et lmmunologica Scandinavica (APMIS) 103:354-366; 1995. 14. Kirkman, A.; Bacon, R. L. Malignant renal tumors in male hamsters (Cricetus auratus) treated with estrogen. Cancer Res. 10:122-124; 1950. 15. Kurl, R. N.; Naylor, P. H.; Loring, J. M.; Todd, R. B.; Villee, C. A. Acute and chronic effects of diethylstilbestrol on serum levels of alpha melanocyte stimulating hormone and prolactin in the hamster. In Vivo 4:89-92; 1990. 16. Li, J. J. Estrogen carcinogenesis in hamster tissues: update 1993. Endocrine Rev. 1:94-95; 1993. 17. Li, J. J.; Gonzalez, A.; Bm~erjee, S.; Banerjee, S. K.; Li, S. A. Estrogen carcinogenesis in the hamster kidney: role of cytotoxicity and cell proliferation. Environ. Health Perspect. 101 (Suppl. 5):259-264; 1993. 18. Liehr, J. G. Hormone-associated cancer: mechanistic similarities between human breast cancer and estrogen-induced kidney carcinogenesis in hamsters. Environ. Health Perspect. 105 (Suppl. 3):565-569; 1997. 19. Llombart-Bosch, A.; Peydro-Olaya, A.; Cremades, A.; Cortes, V.; Gregori, M. A. Interstitial cell origin of estrogen-induced kidney tumors in the Syrian golden hamster. In: Li, J. J.; Li, S. A.; Nandi, S.; Gustafsson, J.-A.; Sekely, L. I., ed. Hormonal carcinogenesis If. New York: Springer-Verlag; 1996:185-192. 20. Nonclercq, D.; Reverse, D.; Toubeau, G.; Beckers, J.-E; Sulon, J.; Laurent, G.; Zanen, J.; Heuson-Stiennon, J.-A. In situ demonstration of germinal cell apoptosis during diethylstilbestrol-induced testis regression in adult male Syrian hamster. Biol. Reprod. 55:1368-1376; 1996. 21. Nonclercq, D.; Toubeau, G.; Wattiez, R.; Laurent, G.; Bernard, A.; Journ6, E; Falmagne, P.; Heuson-Stiennon, J.-A. Sublethal alterations and sustained cell proliferation associated with the diethylstilbestrolinduced renal carcinogenesis in male Syrian golden hamster. Eur. J. Morphol. 36:83-96; 1998. 22. Oberley, T. D.; Gonzales, A.; Lauchner, L. J.; Oberley, L. W.; Li, J. J. Characterization of early kidney lesions in estrogen-induced tumors in the Syrian hamster. Cancer Res. 51:1922-1929; 1991. 23. Osborn, M.; Weber, K. Tumor diagnosis by intermediate filament typing: a novel tool for surgical pathology. Lab. Invest. 48:372-394; 1983. 24. Polack, J. M.; Van Noorden, S. Introduction to immunocytochemistry. 2nd ed. Oxford: BIOS Scientific Publishers; 1997:63-71. 25. Rickmann, M.; Wolff, J. R. Modifications of S100-protein immunoreactivity in rat brain induced by tissue preparation. Histochem. Cell Biol. 103:135-145; 1995. 26. Royds, J. A.; Ironsidc, J. W.; Taylor, C. B.; Graham, D. I.; Timperley, W. R. An immunohistochemical study of glial and neuronal markers in primary neoplasms of the central nervous system. Acta Neuropathol. 70:320-326; 1986. 27. Rutka, J. T.; Murakami, M.; Dirks, P. B.; Hubbard, S. L.; Becker, L. E.; Fukuyama, K.; Jung, S.; Tsugu. A.; Matsuzawa, K. Role of glial filaments in cells and tmnors of glial origin: a review. J. Neurosurg. 87:420-430; 1997.

RENAL TUMOR-DERIVED CELL LINE 28. Sirbasku, D. A.; Kirkland, W. L. Control of cell growth. IV. Growth properties of a new cell line established from an estrogen-dependent kidney tumor of the Syrian hamster. Endocrinology 98:1260-1272; 1976. 29. Soto, A. M.; Bass, J. C.; Sonnenschein, C. Estrogen-sensitive proliferation pattern of cloned Syrian hamster kidney tumor ceils. Cancer Res. 48:3676-3680; 1988. 30. Tsutsui, T.; Taguchi, S.; Tanaka, Y.; Barrett, C. 1713-Estradiol, diethylstilbestrol, tamoxifen, toremifene, and ICI 164,384 induce morphological transformation and aneuploidy in cultured Syrian hamster embryo ceils. Int. J. Cancer. 70:188-193; 1997. 31. Tuszynski, G. P.; Frank, E. D.; Damsky, C. H.; Buck, C. A.; Warren, L. The detection of smooth muscle desmin-like protein in BHK 21/C13 fibroblasts. J. Biol. Chem. 254:6138-6143; 1979.

345

32. Vyberg, M.; Nielsen, S. Dextran polymer conjugate two-step vizualisation system for immunohistochemistry--a comparison of Envision + with two three-step avidin-biotin techniques. Appl. Immunohistochem. 6:3-10; 1998. 33. Wattiez, R.; Nonclercq, D.; Journr, F.; Toubeau, G.; Zanen, J.; Falmagne, P.; Heuson-Stiennon, J.-A. Involvement of transforming growth factoroL and its receptor in a model of DES-induced renal carcinogenesis in the Syrian hamster. Carcinogenesis 17:1615-1622; 1996. 34. Yaoita, E.; Yamamoto, T.; Takashima, N.; Kawasaki, K.; Kawachi, H.; Shimizu, F.; Kihara, I. Visceral epithelial cells in rat glomerular cell culture. Eur. J. Cell Biol. 67:136-144; 1995.