This Article Abstract Full Text (PDF) Purchase Article View Shopping Cart Alert me when this article is cited Alert me if a correction is posted Citation Map Services Email this article to a friend Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Download to citation manager Request Copyright Permission Citing Articles Citing Articles via HighWire Citing Articles via Google Scholar Google Scholar Articles by Amsterdam, A. Articles by Selvaraj, N. Search for Related Content PubMed PubMed Citation Articles by Amsterdam, A. Articles by Selvaraj, N.

Control of Differentiation, Transformation, and Apoptosis in Granulosa Cells by Oncogenes, Oncoviruses, and Tumor Suppressor Genes¹

Department of Molecular Cell Biology, The Weizmann Institute of Science, Rehovot 76100, Israel

	Abstract

Top Abstract I. Introduction II. Protooncogene Expression and... III. Tumor Suppressor Genes,... IV. Immortalization of Primary... V. Mechanism of Induction... VI. Ovarian Cancer VII. Conclusions VIII. Future Directions References

 

I. Introduction

II. Protooncogene Expression and Follicular Cell Development

A. Expression of myc, jun, and fos in granulosa cells

B. Expression of c-kit protooncogene and its ligand steel/kit in the ovary

III. Tumor Suppressor Genes, Death Genes, and Survival Genes in Granulosa Cells

A. Expression of Fas antigen and its ligand in the ovary

B. Modulation of gene expression

IV. Immortalization of Primary Granulosa Cells by Oncogenes and Oncoviruses

A. Establishment of immortalized granulosa cells expressing gonadotropin receptors

V. Mechanism of Induction of Differentiation in Oncogene-Transformed Cells

A. Expression of adrenal 4-binding protein/steroidogenic factor-1

B. Expression of steroidogenic acute regulatory protein

C. Expression of sterol carrier protein 2 and the peripheral benzodiazepine receptor

D. Induction of steroidogenesis in immortalized granulosa cells

E. Expression of inhibin, activin, and follistatin

F. Involvement of the cytoskeleton in granulosa cell differentiation, transformation, and programmed cell death

VI. Ovarian Cancer

A. Endocrine factors in ovarian cancer

B. Tumors of ovarian granulosa cells

C. The role of protooncogenes and tumor suppressor genes in ovarian cancer

VII. Conclusions

VIII. Future Directions

	I. Introduction

Top Abstract I. Introduction II. Protooncogene Expression and... III. Tumor Suppressor Genes,... IV. Immortalization of Primary... V. Mechanism of Induction... VI. Ovarian Cancer VII. Conclusions VIII. Future Directions References

 
GRANULOSA cells are stimulated to grow, to differentiate, and to luteinize by endocrine, paracrine, and autocrine factors (1, 2, 3). In spite of having several hundred thousands of follicles in the mammalian ovary, in numerous primordial and primary stages of development, only very few in each cycle will fully mature, while others will be eliminated by atresia, a process that exhibits both the biochemical and the morphological features of programmed cell death (4, 5, 6, 7). Moreover, in each reproductive cycle, a new corpus luteum will be formed, and the old one will degenerate in a process called luteolysis, which is by nature an apoptotic process (8, 9, 10, 11, 12).

Most of the autocrine or paracrine factors, such as steroids, gonadal peptides, and growth factors, which modulate granulosa cell differentiation, exert their biological effects in a paradoxical manner: in early stages of follicular development, they are believed to be mitogenic while at later stages they enhance granulosa cell differentiation and luteinization in a coordinated manner with gonadotropin-cAMP-generated signals (13, 14, 15). The principal effective factors are those that are involved in modulation of tyrosine kinase signaling, such as insulin (16, 17), insulin-like growth factors (IGFs) (18, 19, 20), epidermal growth factor (EGF) (21, 22, 23), fibroblast growth factor (FGF) (24), transforming growth factor- (TGF) (21, 25), TGFß (26, 27, 28, 29), and PRL (30, 31, 32). Some of their effects could be exerted by activation and modulation of protooncogenes and tumor suppressor genes such as RAS, p53, WAF-1, c-myc, c-jun, and c-fos, which upon mutation can induce tumorigenesis (33, 34, 35, 36, 37). Since oncogenes and oncoviruses have the potential to immortalize normal cells, successful attempts were made in the last decade to immortalize granulosa cells while preserving their differentiation potential (38, 39, 40, 41, 42, 43, 44). These immortalized cells undergo biochemical and morphological changes that closely resemble changes that normal granulosa cells undergo during follicular growth differentiation and luteinization. Therefore, these cell models are extensively used for the study of granulosa cell growth, differentiation, and induction or prevention of programmed cell death (41, 42, 43, 44, 45).

In this review, we shall focus on the effects of protooncogenes, oncogenes, oncoviruses, and tumor suppressor genes on granulosa cell differentiation and death and will discuss their potential role in the development of healthy and atretic follicles. Moreover, we shall discuss implication of these genes in the development of ovarian malignancies.

	II. Protooncogene Expression and Follicular Cell Development

Top Abstract I. Introduction II. Protooncogene Expression and... III. Tumor Suppressor Genes,... IV. Immortalization of Primary... V. Mechanism of Induction... VI. Ovarian Cancer VII. Conclusions VIII. Future Directions References

 
The protooncogenes c-fos and jun are members of the AP-1 family of transcription factors (46, 47). Dimerization between fos and jun or between members of the jun family is necessary for binding to the AP-1 site in the enhancer region of target genes, to regulate their transcription (48, 49). In normal cells, the consistent correlation between the early stages of proliferation and expression of the protooncogenes fos, jun, and myc clearly suggests that these protooncogenes function as essential mediators of the biochemical pathways that regulate proliferation, and that their corresponding oncogenic forms may act via sustained perturbation of normal growth control mechanisms. In addition, jun and fos proteins dimerize with some members of the cAMP response element binding protein (CREB) family. The jun/CREB heterodimer, but not the fos/CREB heterodimer, binds efficiently to the cAMP response elements (CRE) (50). The ability of jun to interact with cAMP signaling pathways by forming jun/CREB dimers thus provides an additional group of transcription factors.

A. Expression of myc, jun, and fos in granulosa cells
The synthesis of the protooncogenes of the AP-1 family in the mammal is triggered by growth factors such as insulin, EGF, FGF, and IGF-I, which are synthesized by various follicular components, e.g., the oocyte, granulosa cells, thecal and the surrounding stromal cells, and therefore can form an autocrine loop that controls granulosa cell proliferation and differentiation (36, 51). Moreover, c-myc, c-jun, jun-d, and c-fos are elevated during granulosa cell proliferation and differentiation (52, 53). PMSG administration to immature rats, or perifusion of rat ovaries with FSH and insulin, increase the expression of c-myc and c-fos mRNA and protein before an increase in DNA synthesis (51, 53). FSH elevates c-fos mRNA levels in rat granulosa cells via the protein kinase C (PKC)-dependent pathway (54). The PKC inhibitor staurosporine was able to block FSH-induced c-fos mRNA expression, whereas specific inhibitors of cAMP- and cGMP-dependent protein kinases had minimal effect on the gonadotropin-induced c-fos mRNA levels (54).

Recent data obtained from intact rat ovaries implicate differential expression of IGF-I, c-jun, and c-fos in granulosa cell proliferation, differentiation, and programmed cell death (37). Granulosa cell DNA synthesis was strictly correlated with the presence of IGF-I and the absence of c-fos and c-jun (37). In contrast, both c-fos and c-jun were detected in luteinized granulosa cells where IGF-I mRNA was undetectable (37). Expression of c-jun in the absence of c-fos was a characteristic feature of granulosa cells in atretic follicles (37). In cultured rat granulosa cells, the messages for c-fos and c-jun were induced by acute gonadotropin, (Bu)₂cAMP, or phorbol ester treatment (55). Because estradiol can regulate the expression of c-fos and c-jun genes in other systems (56, 57, 58, 59), it is reasonable to believe that these genes could be modulated in granulosa cells by estradiol and by growth factors regulating granulosa cell proliferation (13, 21, 60, 61).

Some of the signals for growth and differentiation are associated with the ras protooncogene (35, 62). Therefore, it is important to examine the expression of different members of this protooncogene family during granulosa cell growth differentiation and luteinization. There are initial indications that the expression of the Ras protein is elevated in rat antral follicles and corpora lutea compared with preantral follicles (33). Although the possible role of Ras in steroidogenesis is not yet understood, it is interesting that immortalized granulosa cells transfected with Harvey-ras (Ha-ras) or Kirsten-ras (Ki-ras) are able to preserve high steroidogenic capacity (15, 33, 35, 44, 62).

B. Expression of c-kit protooncogene and its ligand steel/kit in the ovary
An interesting example of cross-talk between follicular cells and the oocyte is suggested to take place via the interaction between the c-kit protooncogene receptor tyrosine kinase and its ligand steel/Kit Ligand (KL) (63). Several studies have shown the expression of c-kit in mouse oocytes (63, 64, 65, 66, 67) and in theca interna cells (64, 67, 68), whereas the ligand KL was localized in granulosa cells (63, 65, 67, 68).

In mouse oocytes, the expression of c-kit is first observed at the diplotene stage close to the time of birth (66, 67). It is maintained in primordial oocytes, accumulates through oocyte growth, and persists through oocyte maturation (66, 67). During ovulation and resumption of meiosis, its expression declines. In one-cell embryos, the c-Kit protein is still observed, while it is undetectable in embryos of four-cell, eight-cell, and morula stage (66). These observations suggest that c-Kit protein may play a significant role in meiotic arrest, oocyte growth, and oocyte maturation.

In human ovaries, c-kit was detected in oocytes (69), and the message for KL was expressed in the granulosa cells (70). The interaction between c-Kit receptor and the ligand steel/KL has been suggested to be involved in embryogenesis as well as in carcinogenesis through a paracrine loop (71). During embryogenesis, c-kit is expressed in primordial germ cells, whereas KL is found along the migratory pathway toward the genital ridge (65, 72, 73). KL expression is obligatory for early folliculogenesis (74, 75) as well as for the survival and proliferation of primordial germ cells in culture (76, 77, 78). In in vitro cultures of mouse primordial germ cells, death occurs with the hallmark of programmed cell death or apoptosis (79, 80), while KL promote primordial germ cell survival by suppressing apoptosis (79). Moreover, granulosa cell-extracted KL was shown to be a potent inducer of mouse oocyte growth in vitro (81).

Increase in KL levels in granulosa cells of antral follicles was observed after hCG administration to mice (67). In contrast, hCG down-regulated c-kit mRNA in the thecal cells, although it did not affect its expression in oocytes (67). Interestingly, in cultured human granulosa-lutein cells, KL transcript levels were rapidly decreased by gonadotropin in a time- and dose-dependent manner (70). Thus, it appears that KL is hormonally regulated in granulosa cells in a species-specific manner. However, the mechanism by which KL exerts its effect on oocyte maturation has not yet been resolved. Further studies are required to clarify the biological significance of granulosa cell KL formation and its possible interaction with the c-kit protooncogene product localized in the oocyte.

	III. Tumor Suppressor Genes, Death Genes, and Survival Genes in Granulosa Cells

Top Abstract I. Introduction II. Protooncogene Expression and... III. Tumor Suppressor Genes,... IV. Immortalization of Primary... V. Mechanism of Induction... VI. Ovarian Cancer VII. Conclusions VIII. Future Directions References

 
Granulosa cells express tumor suppressor genes, such as p53 (34, 82, 83), Wilm’s tumor suppressor gene (WT-1) (34), death genes such as APO-1/Fas (84, 85), and survival genes such as dad-1 (86) and bcl-2 (87). It was recently demonstrated that p53 is highly expressed in apoptotic granulosa cells (34). In extragonadal systems, it was shown that p53 inhibits the expression of bcl-2, a cell survival gene, concomitantly with increasing the expression of Bax, a Bcl-2-related protein that accelerates apoptosis (88, 89, 90). In the rat ovary, it was recently demonstrated that gonadotropin-induced prevention of granulosa cell apoptosis and follicular atresia is associated with a marked reduction in Bax expression (87). WT-1 is a recessive oncogene that encodes a putative transcription factor implicated in nephrogenesis during kidney development (91). It is expressed at high levels in granulosa and epithelial cells of ovaries, Sertoli cells of testis, and in the uterine wall, in addition to the glomerulii of the kidney (91, 92). In the ovary, expression of the WT-1 gene has been shown to be regulated by gonadotropin (34). Physical and functional interaction between WT-1 and p53 proteins was demonstrated in extragonadal cells (93). When these observations were combined, it was suggested that p53 and WT-1, possibly by interacting with each other and regulating bcl-2 and related genes, may play a major role in controlling granulosa cell death during follicular atresia.

A. Expression of Fas antigen and its ligand in the ovary
The Fas antigen, a transmembrane receptor that can trigger apoptosis in a variety of tumor and hematopoietic cells, was detected by RT-PCR and by flow cytometry in human granulosa-lutein cells (84). Anti-Fas antibody induced apoptosis in granulosa-lutein cells pretreated with interferon gamma (84). It was found recently that the Fas antigen is expressed in degenerating oocytes of atretic primordial and primary follicles of human ovary, while the degenerating granulosa cells at various stages of atresia as well as regressing corpora lutea showed enhanced expression of the Fas antigen (85). Furthermore, substantial expression of the Fas antigen was found in oocytes of primordial and primary follicles of infant and adult human ovaries compared with its decreased expression in oocytes of the more developed follicles, suggesting that Fas antigen expression may play a role in regulating the development of follicles in the human ovary (85). Fas antigen was also localized in granulosa cells of secondary and tertiary follicles at an early stage of atresia but not in healthy follicles of the rat ovary (94). Interestingly, the Fas ligand was localized in the oocytes of developing follicles in the rat (94). Localization of Fas in granulosa cells and Fas ligand in the oocytes of certain follicles that undergo atresia suggests a possible mode of cross-talk between the oocytes and the surrounding granulosa cells, which leads to ovarian atresia (94).

B. Modulation of gene expression
Our knowledge of the effect of tumor suppressor genes and survival genes in granulosa cell growth, differentiation, and death in the normal ovary is very limited; however, three recent approaches can shed some light on such processes. One approach is to knock out specific genes in transgenic animals. The second is to overexpress specific genes in transgenic animals, and the third is to transfect primary or immortalized granulosa cells in vitro with tumor suppressor genes.

An elegant example of the first approach is the knockout of p53 and inhibin- in mice (95). Inhibin is a dimeric protein secreted by the granulosa cell in the ovary that functions as an inhibitor of FSH secretion (96). Inhibin- knockout mice invariably develop gonadal sex cord-stromal tumors, suggesting that inhibin can function as a tumor suppressor protein (97). However, gonadal tumor cells from inhibin--deficient mice multiplied poorly, although the cells from mice deficient for both inhibin- and p53 proliferated rapidly (95). These data suggest an interesting cross-talk between p53 and inhibin in the regulation of granulosa cell proliferation. In other experiments, knockout of bcl-2 gene expression reduced the number of oocytes and primordial follicles in the ovary (98). In another study, the ovaries of bax-deficient mice displayed relatively normal oocyte development and follicular formation; however, a marked accumulation of unusual atretic follicles containing numerous atropic granulosa cells that failed to undergo apoptosis were also observed (99). These studies indicate that granulosa cell apoptosis could be regulated by expression of bcl-2-related genes.

In the second approach, targeted overexpression of Bcl-2 in the ovary was achieved by using mouse inhibin- gene promoter. Overexpression of Bcl-2 protein in the ovary led to decreased follicular cell apoptosis, enhanced folliculogenesis, and increased susceptibility to ovarian germ cell tumorigenesis (100). Bcl-2 overexpression was observed only in the somatic cells. Enhanced somatic cell survival appears, therefore, to increase the susceptibility to the formation of ovarian teratoma. However, the exact mechanism by which overexpression of Bcl-2 cells in somatic follicular cells leads to increased germ cell tumorigenesis is currently not understood.

In the third approach, granulosa cells were transfected with a temperature-sensitive mutant of p53 (Val135p53) (82). At 37 C, the temperature-sensitive mutant of p53 is unable to bind DNA; at 32 C it exerts its wild type tumor suppressor activity, since it can bind cellular DNA and induce the WAF-1/CIP-1 gene (82). Cells cotransfected with Simian virus 40 DNA (SV40), Ha-ras, and the p53 temperature-sensitive mutant proliferate rapidly at 37 C but their growth is completely arrested at 32 C (82). Moreover, it was shown that the antiproliferative effect of p53 is due to the activation of the WAF-1/CIP-1 gene, known to be a target gene for p53 in other cell types as well (101, 102, 103). Therefore, p53 may play a role in growth arrest also in normal granulosa cells. The temperature shift of growth of these cells to 32 C stimulated rapid apoptosis, only if cells were pretreated with forskolin, which elevates intracellular cAMP and up-regulates the P450 side chain cleavage enzyme system (82). This suggests that the wild type p53- and cAMP-generated signals may cooperate in inducing apoptosis in normal granulosa cells. It was recently demonstrated cytochemically that apoptotic cells in antral follicles express a high level of the wild type p53 (34). Mutation of the p53 gene can lead to neoplastic transformation, as was evident in p53 knockout mice (104). It was demonstrated that p53 mutation is involved in ovarian cancer originating from ovarian epithelial cells, although it was not yet proven to initiate the epithelial cell transformation (105).

	IV. Immortalization of Primary Granulosa Cells by Oncogenes and Oncoviruses

Top Abstract I. Introduction II. Protooncogene Expression and... III. Tumor Suppressor Genes,... IV. Immortalization of Primary... V. Mechanism of Induction... VI. Ovarian Cancer VII. Conclusions VIII. Future Directions References

 
Cellular and viral oncogenes are usually defined on the basis of their ability to elicit neoplastic transformation (106, 107, 108, 109) but also have been implicated in the control of differentiation (110, 111, 112).

During the last two decades, several attempts were made to immortalize granulosa cells by oncogene and oncovirus transfection while keeping their steroidogenic potential (38, 39, 40, 41, 42, 43, 44). A long-term, steroid-producing rat granulosa cell culture was obtained by fusing hypoxanthine guanine phosphoribosyl transferase-deficient SV40-transformed granulosa cells with freshly prepared rat granulosa cells using Sendai virus (38). These cells produced modest, but significant, amounts of progesterone in response to prostaglandin E₂, cholera toxin, (Bu)₂cAMP, and 2-chloroadenosine. Attempts to transform primary cultures of rat ovarian granulosa cells with Kirsten murine sarcoma virus (KiMSV) led only to the formation of transiently transformed foci (113). When KiMSV was supplemented with EGF, focus formation was greatly enhanced, and two permanently transformed lines that produced low levels of 20-dihydroxyprogesterone were obtained (40, 113).

Because rat cells are nonpermissive hosts for SV40 multiplication, the virus DNA can integrate permanently into the granulosa cells genome; the transfected cells will express some viral proteins without the ability to form new generations of viruses (39, 114). The transforming factor in this virus is the large T antigen, which can immortalize primary cells (115, 116, 117). The transforming potential of SV40 T antigen lies in its capacity to bind and inactivate the retinoblastoma tumor suppressor gene product and p53, both of which regulate the proliferation of normal cells (118, 119, 120, 121).

Several groups have transfected rat granulosa cells with SV40 DNA to yield permanent lines, but there have been discrepant reports on their ability to produce steroid hormones. In one case, a rat granulosa cell line, established by SV40 DNA transfection, showed enhanced synthesis of both progesterone and estradiol upon treatment with forskolin and cholera toxin (114). Another study showed production of higher levels of progesterone by a SV40-transformed granulosa cell line in response to cAMP analogs (122). Enhanced expression of cytochrome P450 side chain cleavage (P450 scc) mRNA was evident upon treating the transformed granulosa cells with 8-Br-cAMP for 24 h (122). Recently, human granulosa-lutein cells were immortalized with SV40 large T antigen (123), and some of the lines responded to 8 Br-cAMP, forskolin, or cholera toxin by secretion of progesterone. However, they showed an inconsistent response to hCG and no response to FSH stimulation (123). In contrast, several other groups reported no or extremely low levels of steroid hormone biosynthesis by SV40 transformation of granulosa cells (38, 44, 124, 125). No detectable expression of the P450 scc enzyme system and the steroidogenic factor, SF1/Ad4BP, or the recently cloned steroidogenic acute regulatory protein (StAR), was observed in these cells (126, 127, 128, 129). Interestingly, the SV40 transformed granulosa cells were able to express the sterol carrier protein 2 (130), the peripheral benzodiazepine receptor (131), IGF-I and its receptors (124), and follistatin (132).

A high steroidogenic potential of immortalized granulosa cells was maintained by cotransfection of rat granulosa cells with SV40 DNA and the Ha-ras oncogene (133). Such a transfection yielded rapidly growing cells that upon cAMP stimulation, subsequent to a lag period, produced progesterone and 20-dihydroprogesterone but failed to produce estradiol (133).

A human ovarian granulosa tumor cell line, which was able to produce estrone and estradiol, was established by long-term culture of human granulosa cells (134). Long-term culture of thecal tumor cells did not yield a permanent cell line (134). Recently, an ovarian thecal-like tumor cell culture model system was developed from an ovarian tumor (135); these cells produced excessive amounts of androgen. In this cell culture model, activation of the protein kinase A (PKA) pathway increases the expression of 3ß-hydroxysteroid dehydrogenase, cytochrome 17-hydroxylase P450 (P450 17), and P450 scc (136). However, by simultaneous activation of the PKA and PKC pathways, progesterone biosynthesis was enhanced while androstenedione production and the levels of mRNA for P450 17 and P450 scc were decreased (136). Another approach was the immortalization of human granulosa cells with the papilloma viruses E6 and E7 (43). These immortalized cells were able to produce significant amounts of estradiol and progesterone in response to forskolin and (Bu)₂cAMP but not to FSH or LH.

A. Establishment of immortalized granulosa cells expressing gonadotropin receptors
Tumor cells grow continuously both in vitro and in vivo; attempts were made in the past, therefore, to isolate steroidogenic tumor cells to establish new cell lines. Attempts were successful for Leydig (137) and adrenal cells (138) but not for granulosa cells. Adrenal tumor cell lines (e.g., Y1) lack ACTH receptors, while the Leydig tumor cell lines (e.g., MA-10) express small numbers of LH receptors. Nevertheless, both lines were extensively used for the study of cellular and molecular mechanisms of steroidogenesis (137, 138).

Since receptors for gonadotropin are generally lost upon cell transformation, immortalized nonsteroidogenic cells, such as CHO and embryonic kidney cells, were transfected with plasmids expressing either the LH/CG or FSH receptors. However, only the initial interaction between gonadotropins and their receptors and coupling to adenylyl cyclase could be studied because such cells do not express regulatory proteins and steroidogenic enzymes (139, 140, 141, 142).

To restore the steroidogenic response to gonadotropins in immortalized cells, LH/CG or FSH receptor expression plasmids were prepared by introducing the complete coding region of LH/CG or FSH receptor cDNAs (143, 144) into an SV40 early promoter-based eukaryotic expression vector. Granulosa cells from rat preovulatory follicles transfected with gonadotropin receptor expression plasmid, together with SV40 DNA and the Ha-ras oncogene (145, 146), expressed about 5–10 times more receptors than primary rat granulosa cells from preovulatory follicles. The recombinant rat LH or FSH receptor molecules expressed in these cells exhibit similar affinities to their hormones as in parental granulosa cells (145, 146). These cell lines responded well to LH or FSH stimulation by cAMP formation as well as progesterone and 20-dihydroprogesterone biosynthesis. The LH-responsive cell lines responded well to both hLH and hCG, but not to FSH (145, 147). These cells showed a dose-dependent increase of both progesterone and 20-dihydroprogesterone in response to hCG (Fig. 1). The FSH-responsive cell lines responded well to rat, ovine, and bovine FSH but not to LH or hCG (Fig. 2). The steroidogenic response of these cell lines was found to be comparable to that of primary granulosa cells (Fig. 2) and, thus, could be a useful system for gonadotropin bioassay in human sera (147). Luteinized granulosa cell lines were established recently from transgenic mice produced by targeting the expression of SV40 large T antigen into gonads using inhibin -subunit promoter (148). These cells possessed high-affinity LH receptor and secreted progesterone and estrogen in response to hCG and FSH, respectively.

View larger version (24K): [in this window] [in a new window]   Figure 1. Concentration dependence of the formation of progesterone () and 20-hydroxy-4-pregnan-3-one (20-OH-progesterone, •). Granulosa cells triple transfected with SV40 DNA, Ha-ras oncogene, and LH/CG receptor (GLHR-15) were stimulated with hCG for 48 h. Data are means ± SEM [Adapted from B.S. Suh et al.: J Cell Biol 119:439–450, 1992 (145) by copyright permission of The Rockefeller University Press.]

View larger version (38K): [in this window] [in a new window]   Figure 2. Progesterone production in primary granulosa cells and a granulosa cell line triply transfected with SV40 DNA, Ha-ras oncogene, and FSH receptor (GFSHR-17), stimulated by gonadotropins. GFSHR-17 cells and primary cultures obtained from PMSG-treated immature rats were stimulated for 48 h at 37 C. oLH and hCG did not cause any significant rise of progesterone production in GFSHR-17 cells above the basal levels (<0.02 ng/5 x 105 cells). Data are means ± SEM (n = 3). Stimulated cultures are significantly higher than controls. *, P < 0.01; **, P < 0.001. [Data reproduced from I. Keren-Tal et al.: Mol Cell Endocrinol 95:R1-R10, 1993 (146) with kind permission from Elsevier Science Ireland Ltd., Bay 15K, Shannon Industrial Estate, Co. Clare, Ireland.]

Spontaneous immortalization of granulosa cells has been demonstrated (152, 153). Repeated subculture of primary bovine granulosa cells at high density yielded cell lines that synthesize estradiol in response to FSH (152). (Bu)₂cAMP and FSH decreased the message and the protein for fibronectin in these cell lines (152). A spontaneously immortalized rat granulosa cell line with constitutive expression of p53 was described (153). Although these cells were positive for P450 scc staining, no pregnenolone, progesterone, or estradiol were detectable when they were stimulated (153). These cells were characterized by an undifferentiated phenotype with prominent intermediate filaments, desmosomes, and gap junctions. Transfection of these immortalized cells with SV40 DNA caused reduced intercellular communication, compared with the parental immortalized cells, which suggests a progressive loss of functional communication during multistep transformation of granulosa cells (153). Establishment of steroidogenic rat granulosa cells expressing the temperature-sensitive mutant of p53 made it possible to investigate the role of a tumor suppressor gene in growth arrest and induction of apoptosis in a well defined system, as discussed previously (82). Thus, transfection of granulosa cells with oncogenes, oncoviruses, and tumor suppressor genes provides experimental models in which one can examine systematically the modulation in expression of proteins associated with the steroidogenic apparatus as well as with other specific markers of the differentiated phenotype of granulosa cells.

	V. Mechanism of Induction of Differentiation in Oncogene-Transformed Cells

Top Abstract I. Introduction II. Protooncogene Expression and... III. Tumor Suppressor Genes,... IV. Immortalization of Primary... V. Mechanism of Induction... VI. Ovarian Cancer VII. Conclusions VIII. Future Directions References

 
Induction of differentiation and steroidogenesis in the mammalian ovary involves a sequential change in the expression of specific genes such as those coding for Ad4BP/SF1, StAR protein, and the steroidogenic enzymes of the P450 scc enzyme system (154, 155, 156). Because induction of steroidogenesis and luteinization in the normal ovary is not homogeneous throughout the follicle population and even among the different granulosa cells of the same follicle, it is sometime difficult to correlate the cellular and the molecular events with biochemical analysis of steroidogenesis. One advantage of the immortalized granulosa cells is that they are a homogeneous population of cells in which induction of steroidogenesis and other differentiation processes can be synchronized. Therefore such cells can serve as a useful tool with which to study the regulation of steroidogenic granulosa cells.

A. Expression of adrenal 4-binding protein/steroidogenic factor-1
The promoter regions of all steroidogenic P450 genes contain regulatory elements that have similar AGGTCA motifs. These motifs interact with a common DNA-binding protein, alternatively designated adrenal-4 binding protein (Ad4BP) or steroidogenic factor 1 (SF-1) (157, 158, 159, 160, 161, 162, 163, 164). A 51 kDa protein which binds to the Ad4 site was purified and the corresponding cDNA clone was isolated (159, 165, 166). The nucleotide sequence of the cDNA revealed that this protein, which has a zinc finger domain and a putative ligand binding/dimerization domain, is an orphan member of the steroid/thyroid hormone receptor superfamily (167). All steroidogenic tissues examined (adrenal, ovary, testis, placenta, adipocyte, and brain) express Ad4BP mRNA (165, 166). In situ hybridization (168) and immunohistochemical staining (169, 170) of the adrenal glands, testes, and ovaries of adult rat or mice localized Ad4BP expression to the specific steroid hormone-producing cells in the tissues, i.e., adenocortical cells in the adrenal gland, Leydig cells in the testis, and granulosa and theca cells in the ovary. Expression of Ad4BP was reported recently in human granulosa-lutein cells (171). The essential role of Ad4BP in governing the steroidogenic cell-specific expression of P450 genes was confirmed by a functional study using an Ad4BP expression vector (155, 172).

Targeted disruption of Ad4BP/SF1 gene resulted in mice lacking adrenal glands and gonads (173). Male and female Ad4BP null mice had female internal genitalia despite complete gonadal agenesis (173). In normal male sex differentiation, Sertoli cells in the embryonic testes are required to produce Mullerian inhibiting substance (MIS), a critical gonadal hormone that mediates duct regression (174, 175). Ad4BP regulates MIS expression in vivo and participates directly in the process of mammalian sex determination (176). Thus, knockout of Ad4BP possibly results in ablation of MIS expression during embryogenesis in male gonads, leading to the development of female internal genitalia.

These and other studies suggest a role for Ad4BP in regulating the genes essential for gonadal development and sexual differentiation in mammalian embryos (170, 173, 176, 177). In these studies, expression of Ad4BP in cells of the steroidogenic tissues was found to precede the expression of steroidogenic P450 side chain cleavage enzyme system.

Ad4BP inhibits the proliferative response of rat follicular granulosa cells to mitogens (178). Ad4BP/SF1 expression is rapidly and transiently expressed in response to an ovulatory dose of hCG in PMSG-primed immature rat (179). Similarly a transient decrease in P450arom is observed during this period. In contrast, the expression of P450 scc increased after the LH surge (179). c-myc Gene expression and incorporation of BrdU in granulosa cell, a marker for active DNA synthesis, also increased in response to hCG (179). This study suggests that hCG depresses Ad4BP expression, while increasing DNA synthesis and c-myc expression. However, it is not clear how Ad4BP can repress granulosa cell DNA synthesis.

Ad4BP was found to be expressed in steroidogenic adrenal tumor Y-1 cells and testicular tumor Leydig MA-10 cells (172, 177) and R2C cells (161). Expression of Ad4BP was observed only in sex-cord tumor cells that were positive for steroidogenic enzymes, but not in nonsteroidogenic tumor cells (180). Recently, it was demonstrated that only steroidogenic cell lines cotransfected with SV40 and Ha-ras express Ad4BP/SF1, whereas nonsteroidogenic granulosa cell lines (transfected by SV40 alone) have completely lost the expression of this transcription factor (128). Moreover, in lines that demonstrate cAMP-induced steroidogenesis, the level of Ad4BP expression was maximal even in nonstimulated cells that proliferate rapidly while exhibiting only traces of steroidogenic activity. This correlates well with the constitutive expression of Ad4BP in human granulosa-lutein cells (171). The data support the view that Ad4BP/SF1 expression is an intrinsic and specific property of cells that determines its steroidogenic ability. Furthermore, the data suggest that Ad4BP expression is required, but not sufficient, for active steroidogenesis.

Several researchers reported that SV40-induced transformation resulted in a dramatic reduction of the steroidogenic activity of granulosa cells (38, 39, 44, 124, 125). The recent work cited above demonstrated that SV40-induced transformation eliminated the expression of Ad4BP, while coexpression of Ha-ras and SV40 could override this deficiency (128). The mechanism by which different oncoproteins can regulate expression of such an essential component of the steroidogenic machinery in opposite directions remains to be elucidated.

B. Expression of steroidogenic acute regulatory protein
The rate-limiting enzymatic step in adrenal and gonadal steroid production, in response to tropic hormone stimulation, is the conversion of cholesterol to pregnenolone (181, 182). This enzymatic reaction is catalyzed by the cytochrome P450 scc system and its ancillary electron transport proteins, adrenodoxin and adrenodoxin reductase (CSSC system), located on the matrix side of the inner mitochondrial membrane (183, 184). Mobilization of the substrate cholesterol to the inner mitochondrial membrane and the CSCC system is a crucial step in this biochemical process (185, 186). In addition, the acute production of steroid hormone depends on a rapidly synthesized, cycloheximide-sensitive, and highly labile protein that appears in response to tropic hormones and transfers cholesterol to the inner mitochondrial membrane (187, 188, 189, 190, 191).

A protein of 30 kDa was observed to be synthesized in response to tropic hormones or cAMP analogs in adrenal (192, 193), ovary (194), and MA-10 mouse Leydig tumor cells (195). This protein is derived from a larger 37-kDa precursor in all the steroidogenic cell types (195, 196) and may require phosphorylation on a threonine residue for its activity (197). MA-10 cells deficient in protein kinase A do not express this protein (198). Recently, this 30-kDa protein was purified and its cDNA was cloned from MA-10 cells (199); it was named the steroidogenic acute regulatory StAR protein (199). Expression of the StAR cDNA in transiently transfected MA-10 cells resulted in increased steroidogenesis in the absence of hormone stimulation (199). A cDNA for StAR isolated from a human adrenal library showed a deduced amino acid sequence that was 87% identical to the mouse sequence (200). Perhaps the most striking evidence for the function of StAR in cholesterol transport and steroidogenesis was observed in patients with lipoid congenital adrenal hyperplasia, a condition characterized by deficiency in adrenal and gonadal steroid production despite a normal CSSC enzyme system (201). The cause of this disease, in two patients, is a nonsense mutation in StAR resulting in truncation of the StAR protein by 93 or 28 amino acids, which leads to a defective cholesterol transport mechanism (201). Coexpression of StAR cDNA with the CSCC system in COS1 cells resulted in an 8-fold increase in pregnenolone production with cholesterol as a substrate, whereas the mutant StAR was inactive; the need for StAR activity could be circumvented by using freely diffusable 20-hydroxycholesterol as a substrate for steroidogenesis (201). Therefore, StAR appears to play a key role in cholesterol delivery to the inner mitochondrial membrane for the enzymatic action of the CSSC system, which is the rate-limiting enzymatic step in steroidogenesis (156).

It was recently demonstrated that FSH and IGF-I interact synergistically to induce expression of the StAR message and protein in immature porcine granulosa cells (202). Basic FGF, either free or sequestered in a native basement membrane, was found to increase the level of StAR protein in rat preovulatory granulosa cells (203). These findings suggest a novel mechanism of cross-talk between gonadotropins/cAMP-mediated signals and tyrosine kinase signals induced by growth factors in stimulation of granulosa cell steroidogenesis.

In a recent paper it was demonstrated that StAR mRNA is expressed in rat granulosa cells, transformed by SV40 DNA and Ha-ras oncogene, which preserve their steroidogenic potential (129). In contrast, cells transformed with SV40 DNA alone that lost their steroidogenic capacity did not express the StAR message. This implies that expression of the StAR gene is obligatory to the steroidogenic activity not only in normal steroidogenic cells (156, 201, 204) but also in oncogene-transformed cells. In addition, it is possible that the Ras protein is important for the preservation of differentiation in immortalized granulosa cells. However, in SV40-Ha-ras-transformed fibroblasts, StAR expression was undetected (129), suggesting that Ras protein by itself was not sufficient to induce StAR expression in transformed cells that did not originate from steroidogenic tissues.

Using immortalized granulosa cells expressing receptors to LH/CG, FSH, or the ß₂-adrenergic receptor, it was possible to demonstrate that expression of StAR mRNA, and its regulation by agents elevating cAMP levels such as catecholamines, gonadotropins, and forskolin, can be preserved in transformed rat granulosa cells (129). The sequence of the partial cDNA isolated from a granulosa cell line expressing FSH receptor demonstrated a high degree of homology with the corresponding region of StAR sequence from the mouse and human cDNA (Fig. 3). Such systems can serve as a useful tool with which to study the regulation of the StAR gene by endocrine factors as well as by oncogenes used to immortalize these cells, which may also play an important role in ovarian malignancies.

View larger version (58K): [in this window] [in a new window]   Figure 3. Comparison of the partial rat StAR cDNA sequence with the corresponding region of the StAR cDNA sequence of mouse and human. The rat sequence was obtained from the partial cDNA isolated from the granulosa cell line triply transfected with SV40 DNA, Ha-ras oncogene, and FSH receptor plasmid (GFSHR-17). [Adapted from N. Selvaraj et al.: Mol Cell Endocrinol 123:171–177, 1996 (129) with kind permission from Elsevier Science Ireland Ltd., Bay 15K, Shannon Industrial Estate, Co. Clare, Ireland.]

In rat ovary, SCP₂ mRNA expression was found in granulosa and thecal cells as well as in corpora lutea (130). Gonadotropins, which promote follicular growth and luteinization, increased the ovarian content of SCP₂ mRNA along with an increase in cytochrome P450 scc mRNA (130). Using the steroidogenic rat granulosa cells, cotransfected with SV40 and the Ha-ras oncogene, 8-Br-cAMP was found to increase SCP₂ mRNA and protein levels within 24 h of treatment (130); P450 scc mRNA was also induced, whereas actin mRNA levels were not affected. The 8-Br-cAMP stimulation of SCP₂ mRNA accumulation was completely inhibited by actinomycin D or cycloheximide. The cAMP analog also increased SCP₂ mRNA levels in a nonsteroidogenic rat granulosa cell line transfected with SV40 DNA alone (130). Thus, it seems that stimulation of SCP₂ expression in ovarian cells is mediated, at least in part, by cAMP, by a mechanism requiring ongoing RNA and protein synthesis. SCP₂ gene expression, however, is not obligatorily coupled to steroidogenic activity, as cAMP analogs can increase SCP₂ mRNA in transformed ovarian granulosa cell lines incapable of synthesizing steroid hormones (130).

PBR has recently been shown to be expressed in steroidogenic cells of the adrenal medulla (207, 208, 209) and MA-10 Leydig tumor cells (210). It was suggested that the receptor molecules are localized mainly in the mitochondrial outer membrane (207). This receptor may stimulate cholesterol import into mitochondria (208) and thus accelerate the conversion of cholesterol to pregnenolone, which is the limiting step in the biosynthesis of steroid hormones. In the ovary, both central and peripheral receptor types exist in tissue homogenates of normal and cancerous tissues (211, 212).

A high content of the PBR was found in SV40/Ha-ras transformed granulosa cells, and a lower content was found in granulosa cells transformed with SV40 alone (131). The number of PBR was found to increase in cAMP-stimulated cells. It was also demonstrated that, both in normal cells as well as in transformed steroidogenic granulosa cells, a benzodiazepine agonist dramatically elevates progesterone production (131). These data support a possible role of the PBR in ovarian steroidogenesis. Because the expression of SCP₂ (130) and PBR (131) were evident both in SV40-transformed cells and SV40/Ha-ras-transformed cells, it can be concluded that the expression of these proteins is less sensitive to SV40 transformation than the expression of SF-1/Ad4BP (128), StAR (129), and P450 scc enzymes (126, 127), which are not expressed in cells that were transformed with SV40 DNA alone.

D. Induction of steroidogenesis in immortalized granulosa cells
The induction of steroidogenesis in granulosa cells is initiated by the gonadotropic hormones acting directly on these cells. Gonadotropins bind to cell surface receptors and activate intracellular signaling systems including adenylate cyclase (2, 213, 214, 215). Their inductive effects can be mimicked by stimulation with cAMP, suggesting that this is the principal intracellular messenger of the gonadotropins (3, 216, 217). In primary granulosa cells, the CSCC enzyme system can be induced by gonadotropins (218, 219, 220, 221). Increase in the levels of these enzymes results from enhanced transcription of their genes (218, 222, 223, 224, 225). Interestingly, transfection of granulosa cells with SV40 alone knocks out almost completely the expression of the P450 scc enzyme system, whereas cotransfection of the cells with SV40+Ha-ras or Ki-ras preserves the potential of the cells to express these enzymes in a cAMP-dependent manner (126, 150). This implies that the P450 scc enzyme system may be sensitive to viral transfection and that Ras protein may be essential for the induction of their expression by a mechanism that is not yet understood.

One of the characteristics of SV40-ras-transformed granulosa cells is that when they are cultured in the absence of stimulants, which elevate intracellular cAMP, they proliferate very rapidly, show extremely low expression of the steroidogenic enzymes, and release very small quantities of progesterone. In contrast, upon stimulation with gonadotropic hormones, after a lag period of 6–12 h, the cells produce high levels of progesterone. Steroid hormone production in response to gonadotropin is 100 times higher than in nonstimulated cells, in the range of progesterone production of highly luteinized primary cells. This unique feature of the cells permits a detailed analysis of the induction kinetics of the steroidogenic enzymes in a homogeneous cell system, compared with the heterogenous population of the granulosa cells in the intact follicle or in primary cultures (226). Such studies showed that the induction of P450 scc is significantly slower than that of adrenodoxin (126, 226) (Fig. 4). Nevertheless, the individual components of the CSCC system, i.e., the cytochrome P450 scc, adrenodoxin, and adrenodoxin reductase are uniformly incorporated into all mitochondria of the steroidogenic cells and localized in the inner face of the mitochondrial cristae (126) (Fig. 5).

View larger version (35K): [in this window] [in a new window]   Figure 4. Correlation of progesterone synthesis and induction of the cytochrome P450 scc enzyme system in a granulosa cell line transfected with SV40 DNA and Ha-ras oncogene (POGRS-1). The top panel shows autoradiography of Western blots of two different gels containing either purified adrenodoxin or P450 scc and samples of cell protein isolated at the indicated times after addition of fresh medium containing 1 mM 8-Br-cAMP. In each gel the standards were 0.25, 0.5, 1, and 2 pmol of the indicated protein purified from bovine adrenal cortex. The quantification of the enzymes is based on densitometric scanning of the Western blots. Bottom panel, Induction of adrenodoxin, cytochrome P450 scc (•), and progesterone () synthesis in the POGRS-1 cell line after stimulation with 8-Br-cAMP. [Reproduced with permission from I. Hanukoglu et al.: J Cell Biol 111:1373–1381, 1990 (126) by copyright permission of The Rockefeller University Press.]

View larger version (156K): [in this window] [in a new window]   Figure 5. Morphology of SV40-Ha-ras transfected immortalized granulosa cells expressing the LH/CG receptor (A) and localization of the steroidogenic enzyme adrenodoxin by immunofluorescence (B) and by immuno-cryo electron microscopy (C). Cells were incubated for 48 h with hCG for induction of steroidogenesis. A, Ultrastructure of part of a cell. The cytoplasm is rich in mitochondria (m). The endoplasmic reticulum (er) is well developed. B, Cells were stained with rabbit anti-adrenodoxin antibodies and rhodamine-goat anti-rabbit IgG and visualized in the fluorescent microscope. Mitochondria (arrowheads) are intensively labeled throughout the cytoplasm, leaving the nucleus (N) unstained. C, Ultrathin cryosection stained with rabbit anti-adrenodoxin and goat anti-rabbit IgG coupled to 15 nm gold particles. The majority of the gold particles are located on the mitochondrial inner membrane (arrowheads). [Reproduced with permission from A. Amsterdam and D. Aharoni: Micros Res Tech 27:108–124, 1994 (45). © 1994 John Wiley & Sons, Inc.]

View larger version (38K): [in this window] [in a new window]   Figure 6. Effect of dexamethasone on progesterone (A) and cAMP (B) formation in FSH-stimulated immortalized rat granulosa cells triply transfected with SV40 DNA, Ha-ras oncogene, and FSH receptor. cAMP in the culture medium was measured after 30 min incubation with oFSH (1.6 nM). Progesterone production was measured 24 h after FSH stimulation. (For experimental details see Refs. 128 and 146.)

Activin regulates FSH-stimulated progesterone production by rat granulosa cells in a developmentally related manner (239). In nondifferentiated granulosa cells, activin enhances the response to FSH, but in differentiated cells, it is inhibitory (239). FSH-stimulated expression of P450 scc mRNA was enhanced by combined treatment of nondifferentiated granulosa cells with activin and FSH (239). However, activin had no consistent effect on FSH-stimulated expression of 3ß-hydroxysteroid dehydrogenase mRNA in nondifferentiated cells (239). In differentiated granulosa cells, both mRNAs were suppressed by more than 50% in the presence of activin (239). In cultured human granulosa-lutein cells, activin inhibited both progesterone and estrogen biosynthesis (240, 241, 242). Although inhibin had no effect on steroid production by human granulosa-lutein cells, it induced androgen synthesis in thecal cells (241, 242, 243). Activin-A was shown to stimulate locally the synthesis of ß_B-subunit mRNA in human granulosa-lutein cells by an autocrine or paracrine mechanism (244). In addition, TGF-ß₁ and ß₂ enhanced inhibin-A and activin-ß_B subunit mRNA levels in cultured human granulosa-lutein cells (245). Activation of PKA and PKC by 8-Br-cAMP and phorbol ester resulted in differential responses in the steady-state levels of inhibin/activin- and ß_A subunit and follistatin mRNAs in human granulosa-lutein cells (246).

In developing granulosa cells, activin promotes cell proliferation (247). In the presence of activin, but not inhibin, FSH stimulated DNA synthesis in granulosa cells isolated from immature rat ovaries (247). Proliferation of Sertoli cells (248), human granulosa lutein cells (240), and a sex-cord tumor cell line, derived from a mouse deficient for inhibin- and p53 (95), were enhanced by activin in vitro. In situ hybridization in lamb ovary showed a sequential appearance and disappearance of message for follistatin and inhibin/activin during follicular maturation and atresia (249). Interestingly, expression of these messages was much higher in the granulosa cells located in proximity to the oocyte (cumulus), compared with more distant cells of the membrana granulosa (249). Moreover, during follicular atresia, the mRNA levels in the granulosa cells declined and finally disappeared as atresia progressed, persisting only in the cumulus cells (249).

Activin A was reported to induce apoptotic cell death of myelomas (250). Overexpression of Bcl-2 suppressed activin-induced apoptosis in the B cell hybridoma cell line (251). However, the possible cross-talk between Bcl-2 and activin in regulating granulosa cell apoptosis has not been established.

The protooncogene c-kit is present in the mouse oocyte, whereas its ligand steel/KL was localized in the granulosa cells (81). Activin A reduced the expression of c-kit mRNA in murine erythroleukemia cells (252). Therefore, one cannot exclude the possibility that ovarian activin produced by the granulosa cells could have a paracrine effect on the modulation of c-kit present in the oocyte, thereby modulating oocyte maturation.

Ovarian epithelial tumors and granulosa cell tumors secrete inhibin, and the circulating level of inhibin has been suggested as a marker for these tumors (253, 254). Knock-out of the inhibin- gene led to the formation of sex-cord tumors in mice, suggesting that inhibin acts as a tumor suppressor protein (95).

Although follistatin has recently been shown to be expressed in a number of different tissues, the granulosa cells are a major production site (236, 255). Follistatin protein production by primary rat and bovine granulosa cells was shown to be regulated by FSH and cAMP, but not by LH (256, 257). However, in primary porcine granulosa cells, LH had a stimulatory effect on follistatin gene expression (258). This difference may be related to the stage of differentiation of the granulosa cells, which might differ in their response to gonadotropins according to the presence or absence of receptors to LH and FSH.

A recent study on the regulation of follistatin gene expression was undertaken in four different rat granulosa cell lines, transfected with SV40 DNA alone, or with SV40 DNA and Ha-ras oncogene, which lacked or expressed LH or FSH receptors (132). All the cell lines expressed follistatin mRNA, which could be regulated by forskolin. In cell lines expressing either LH or FSH receptors, follistatin was elevated by stimulation of the appropriate gonadotropins (132) (Fig. 7). Activation of PKC by phorbol ester also stimulated follistatin mRNA (Fig. 7), as in primary granulosa cells (132, 258, 259). This suggests that follistatin gene expression is regulated by multiple signal transduction pathways in granulosa cells. Moreover, follistatin, which is predominantly expressed in normal granulosa cells, is maintained subsequent to oncogene transformation and therefore can serve as a potential marker for granulosa cell tumors.

View larger version (20K): [in this window] [in a new window]   Figure 7. Effects of FSH, forskolin (Fk), and 12-O-tetradecanoylphorbol-13-acetate (TPA) on follistatin mRNA accumulation in granulosa cell line triply transfected with SV40 DNA, Ha-ras oncogene, and FSH receptor (GFSHR-17) (top panel). Cells were treated for 6 h with FSH (1.6 nM), Fk (50 µM), or TPA (50 nM) or with FSH in the presence of forskolin or TPA. Duplicate cultures of each treatment group were analyzed by Northern blot hybridization to 32P-labeled follistatin and GAPDH cDNA probes. Dose-response stimulation of follistatin gene expression and progesterone accumulation by FSH in the GFSHR-17 cell line is shown in the bottom panel. Cells were treated with FSH (0.004–4 nM) for 6 h. A, Duplicate cultures of each treatment group were analyzed by Northern blot hybridization to 32P-labeled follistatin and GAPDH cDNA probes. B, Progesterone concentrations (picograms per 0.5 ml) in culture medium stimulated by different concentrations of FSH. The densitometric scanning of the Northern blot is expressed as mean densitometric units ± SEM for follistatin relative to GAPDH mRNA, adjusted to a value of 100% for the control cells. [Adapted with permission from L. Shukovski et al.: Endocrinology 136:2889–2895, 1995 (132). © The Endocrine Society.]

Rearrangement of the cytoskeleton and down-regulation of actin and actin-binding proteins is a characteristic of granulosa cells as well as other steroidogenic cells like Leydig and adrenal cells, which exhibit high levels of steroidogenesis (33, 260, 261, 262, 263, 264, 265, 266). Steroidogenic granulosa cell lines transformed by SV40 DNA and the Ha-ras oncogene also show poor organization of the actin cytoskeleton and extremely low expression of tropomyosin 2 and 3, in contrast to cells transformed with SV40 alone, which demonstrate high expression and organization of the actin cytoskeleton including high expression of tropomyosin 2 and 3 (214, 267) (Fig. 8). SV40-transformed cells showed a low tumorigenic capacity when injected into nude mice, and even if stimulated by cAMP, only traces or no steroidogenic activity was evident (41, 268). In contrast, SV40/Ha-ras transformed cells show high tumorigenic activity and metastatic spread when injected into nude mice (Fig. 9). However, upon cAMP stimulation they become highly steroidogenic (41, 268). These observations suggest that the Ras protein plays an important role in down-regulation of the actin cytoskeleton, which leads to enhanced proliferation on the one hand and to enhanced steroidogenesis on the other, in cAMP-stimulated cells. Modulation of the expression of the actin cytoskeleton, which is probably involved with Ras expression, is therefore important both for differentiation and proliferation of oncogene-transformed granulosa cells.

View larger version (60K): [in this window] [in a new window]   Figure 8. Tropomyosin expression in SV40-transformed granulosa cell line (left) compared with SV40-Ha-ras-transformed granulosa cell line (right). The cells were labeled for 1 h with [35S]methionine, and the proteins were analyzed by two-dimensional gel electrophoresis. a, Actin; v, vinculin; closed arrowhead, -actinin; 1,2,3,4,5, tropomyosin isoforms; c, PCNA/cyclin; , -tubulin; ß, ß-tubulin. The inset shows the tropomyosin area of the autoradiogram exposed for a longer time to visualize the low but detectable levels of tropomyosin 1 synthesis. Note that the expression of tropomyosin 2 and 3 are evident in SV40-transformed cells (left) but not detected in SV40-Ha-ras transformed cells (right). Actin-binding proteins vinculin (v) and -actinin (closed arrowhead) are higher in SV40-transformed cells compared with SV40-Ha-ras-transformed cells. [Modified with permission from G. Baum et al.: Dev Biol 142:115–128, 1990 (267).]

View larger version (148K): [in this window] [in a new window]   Figure 9. Granulosa cell tumor development and metastatic spread in nude mice after injection of cells transformed by SV40 DNA and Ha-ras oncogene. A, Section of a solid tumor in the back of a nude mouse 2 weeks after subcutaneous injection of 106 cells. Cells are tightly packed, with a coffee bean-shaped nuclei. A high incidence of mitotic figures is visible (arrowheads). B-D, Metastatic spread in lung (B), kidney (C), and ovary (D) 3 weeks after iv injection of 2.5 x 105 cells. Growing foci of transformed cells (asterisks) are visible in close vicinity to lung alveoli in the respiratory tissue (a in panel B), to convoluted tubules (c in panel C) in the renal lobule of the kidney, and to thecal cells (t) of a large preantral ovarian follicle. g, Normal granulosa cells; o, cavity of oocyte location. Bar = 50 µm. [Reproduced with permission from B. S. Suh et al.: Endocrinology 131:526–532, 1992 (268). © The Endocrine Society.]

Taxol, a drug that affects microtubule organization, is in common use in chemotheraphy of ovarian cancer (270, 271, 272), suggesting that the organization of microtubules is an important factor in proliferation of ovarian tumor cells.

The actin cytoskeleton also appears to play a role in the induction of apoptosis in granulosa cells cotransfected with SV40 + Ha-ras and the temperature-sensitive mutant of p53. Apoptosis in these cells is induced by stimulating the cells with forskolin and shifting the temperature of cell growth from 37 C to 32 C, which leads to the manifestation of the wild type p53. This is accompanied by rearrangement of actin filaments to form a spherical network that separates the bulk of the cells from apoptotic blebs (273). Thus the transformed steroidogenic cells do not lose their steroidogenic organelles such as mitochondria, lipid droplets, and smooth ER. This compartmentalization of the steroidogenic organelles around the perinuclear region allows ongoing and even enhanced steroidogenesis in the apoptotic cell until total cell collapse (273) (Fig. 10). Early observations indicated that there is a temporal elevation of steroidogenesis upon induction of atresia in the intact rat ovary (274, 275). The rearrangement of the actin cytoskeleton and clustering of the steroidogenic organelles may be responsible for this phenomenon even in the intact follicles.

View larger version (131K): [in this window] [in a new window]   Figure 10. Electron micrograph of apoptotic granulosa cells transfected with SV40 DNA Ha-ras oncogene and a temperature-sensitive p53 (GTS-5). The cells were stimulated with forskolin for 24 h at 32 C. A, Note condensation of chromatin (arrow) in the nucleus and numerous apoptotic blebs at the periphery of the cell (asterisks), which are devoid of mitochondria and lipid droplets; mitochondria (m), are concentrated in the perinuclear region. Bar, 2 µm. B, Intensive clustering of steroidogenic organelles in a GTS-5 cell. Note initial condensation of chromatin in the nuclear periphery (arrow), apoptotic cytoplasmic blebs (asterisk), a massive aggregation of mitochondria (m), lipid droplets (d), Golgi complexes (g), and small vesicles (s) of smooth membrane characteristic of highly steroidogenic cells. The apoptotic blebs are devoid of mitochondria and lipid droplets. [Adapted with permission from I. Keren-Tal et al.: Exp Cell Res 218:283–296, 1995 (82).]

View larger version (107K): [in this window] [in a new window]   Figure 11. Translocation of proteasomes, mitochondria, and actin cytoskeleton during apoptosis in immortalized granulosa cells expressing a temperature-sensitive mutant of p53. Upper set of images, Nonapoptotic cells stained with rabbit antibodies to proteasomes (A) or rabbit adrenodoxin (a marker for the steroidogenic mitochondria) (B) and with rhodamine phalloidine (B'). Proteasomes are distributed both in the nucleus and the cytoplasm whereas mitochondria are distributed in the cytoplasm and cell processes delineated by actin cables. Apoptotic cells (C) stained with anti-proteasome antibodies (C') show a high concentration of proteasomes in apoptotic blebs leaving the nucleus free of proteasomes.