| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Department of Obstetrics and Gynecology (H.-F.H., P.C.K.L.), Zhejiang University School of Medicine, Zhejiang 310016, China; Department of Zoology (A.S.T.W.), University of Hong Kong, Hong Kong; and Department of Obstetrics and Gynecology (J.-H.C., P.C.K.L.), University of British Columbia, Child and Family Research Institute, Vancouver, British Columbia, Canada V6H 3V5
Correspondence: Address all correspondence and requests for reprints to: Peter C. K. Leung, Ph.D., Department of Obstetrics and Gynecology, University of British Columbia, 2H-30, 4490 Oak Street, Vancouver, British Columbia, Canada V6H 3V5. E-mail: peleung{at}interchange.ubc.ca
 |
Abstract |
|---|
 Top Abstract I. IntroductionII. Biology of Ovarian...III. Gonadotropin TheoryIV. Action of Gonadotropins...V. Clinical ApplicationsVI. Conclusion and Future...References |
|---|
|
I. Introduction |
|---|
|
TopAbstractI. IntroductionII. Biology of Ovarian...III. Gonadotropin TheoryIV. Action of Gonadotropins...V. Clinical ApplicationsVI. Conclusion and Future...References |
|---|
|
II. Biology of Ovarian Epithelial Cancer (OEC) |
|---|
|
TopAbstractI. IntroductionII. Biology of Ovarian...III. Gonadotropin TheoryIV. Action of Gonadotropins...V. Clinical ApplicationsVI. Conclusion and Future...References |
|---|
The etiology of ovarian cancer remains poorly understood. Epidemiological studies have failed to yield a consensus regarding the contribution of chemical carcinogens to the development of ovarian cancer. Recent studies have linked tobacco products/smoking to an increased incidence of the mucinous type of ovarian cancer (6, 7, 8). In addition, talc (9, 10), asbestos (11), and alcohol (12) have been sometimes associated with an increased risk. However, these issues are limited by inconsistent data and/or the lack of supportive animal models. Other factors include a family history (13). In addition, hormonal and/or reproductive factors such as nulliparity and infertility have been consistently recognized as risk factors of ovarian cancer, whereas pregnancy, oral contraceptive (OC) use, hysterectomy, and tubal ligation have protective effects on ovarian cancer (14). Women suffering from infertility have about a 2-fold higher ovarian cancer risk than does the general population (14). Ovulation induction agents used in the treatment of infertility (clomiphene or gonadotropins) may also be a risk factor for ovarian cancer (15, 16, 17, 18, 19). Interestingly, each additional pregnancy after the first also decreases the risk 10–16% (20, 21). The reduction of ovarian cancer risk in women ever taking OC compared with nonusers is about 30% (22, 23, 24), and tubal ligation and hysterectomy, without oophorectomy, also reduced the risk of ovarian cancer. This is presumably due to decreased exposure of the ovary to potential carcinogen factors and/or inflammation (25, 26, 27).
There are many types of tumors that can start in the ovaries, including germ cell tumors, sex cord-stromal tumors, and epithelial tumors. Ovarian germ cell tumors develop from primitive germ cells of the embryonic gonad, and the most common germ cell malignancies are dysgerminoma, teratoma, and yolk sac tumor. Sex cord-stromal neoplasms arise from various cell types in gonadal stroma and sex cords. Granulosa cell tumors (GCT) are the most prevalent tumor in this group, which constitutes of approximately 5% of all ovarian cancers. In addition to estrogen synthesis (28) and secretion of the gonadal peptide hormones (29), GCT has been found to express gonadotropin receptor (30, 31) and response to gonadotropins (32, 33, 34). In this regard, it has been speculated that gonadotropin may be involved in the pathogenesis of GCT. Approximately 90% of malignant ovarian tumors are epithelial tumors and arise from the OSE. Interestingly, the simple primitive OSE with mesenchymal features acquires the characteristics of Müllerian epithelium as it develops to malignancy (35). In fact, based on the morphological, functional, and antigenic resemblance to epithelium of Müllerian ducts, this disease is divided into a number of subtypes, including serous (fallopian tube), mucinous (endocervical-like), and endometrioid (endometrium-like) tumors (35, 36).
B. Ovarian surface epithelium (OSE)
The OSE is a single layer of flat-to-cuboidal epithelial cells covering the ovary (37). Studies on the ovary have mainly focused on the other cell types such as granulosa and theca cells, which play a critical role in folliculogenesis and steroidogenesis. The OSE was among the least studied part of the ovary due to its inconspicuous histological appearance and apparent lack of significant functions. Interest in the OSE began when it became apparent that OSE might be the tissue of origin of the OEC (38, 39). However, the cellular and molecular mechanisms by which OSE cells undergo tumor formation and neoplastic progression are still not well understood.
The OSE is separated from the hormone/growth factor-producing stroma by the collagenous tunica albuginea and a basement membrane. Ovulation and aging-related trapping of OSE fragments result in surface invaginations (clefts) and inclusion cysts in the ovarian cortex (40). Numerous studies have provided direct evidence bearing on the frequent metaplasia and neoplastic conversion of the clefts and inclusion cysts, possibly because of the aberrant exposure to the hormone/growth factor-rich stromal microenvironment (41, 42, 43, 44). Thus, the OSE trapped into the ovarian cortex has been generally considered as a neoplastic progression-prone site for OEC. However, whether the surface and/or cyst epithelial cells directly progress to a malignant neoplasm is still not clear. Ness and Cottreau (11) have suggested that it is an inflammatory microenvironment such as cell damage, oxidative stress, and elevations of cytokines and prostaglandins, rather than the trapping of the OSE into the stroma, that could mediate the mutagenesis induced by ovarian ovulation. Repetitive loss of the basement membrane, during ovulation, has been implicated as an early event in the preneoplastic transformation of OSE. This may be mediated through enhanced survival of the epithelial cells and/or preferential selection of apoptosis-resistant cells (45, 46, 47).
More recently, there is the "stem cell niche" concept for cancer development (48). Somatic stem cells appear to reside within a specialized area, or niche, where they could remain quiescent until activation by an injury or other stimulation. Adult OSE cells have the stem cell property of self-renewal, which makes them candidates for this new concept. It is also noteworthy that the OSE is closer in its differentiation state to the pluripotent or multipotent mesodermal embryonic precursor cells than other epithelial derivates (49, 50). As discussed above, OSE is a simple, rather primitive epithelium with some stromal features, but as it progresses to malignancy it loses stromal characteristics and acquires the characteristics of the complex glandular epithelial phenotypes of Müllerian duct-derived tissues. This retention of pluripotentiality may be accompanied by greater proliferative capacity with reduced induction of apoptotic cell death pathways and therefore perhaps enhanced the susceptibility to neoplastic transformation. Many signal molecules have been shown to be involved in regulation of stem cell behavior, including hormones/growth factors, Wnt, Notch, and sonic hedgehog, in various organ systems (51, 52, 53, 54, 55). In the ovary, gonadotropins are involved in the regulation of cell proliferation, apoptosis, and differentiation. Whether they would play any role in the OSE stem cell niche remains to be elucidated and is an area for further research.
|
III. Gonadotropin Theory |
|---|
|
TopAbstractI. IntroductionII. Biology of Ovarian...III. Gonadotropin TheoryIV. Action of Gonadotropins...V. Clinical ApplicationsVI. Conclusion and Future...References |
|---|
The incessant ovulation hypothesis, which was initially proposed by Fathalla (56) in 1971 and later extended by other researchers (57), is that repeated trauma during ovulation leads to an increased exposure of the OSE to genetic abnormalities and/or other risk factors. Indeed, early menarche, late menopause, and nulliparity, all of which have more ovulation episodes, increase the risk of developing ovarian cancer. On the other hand, conditions in which ovulation is suppressed, such as multiple pregnancies and prolonged breastfeeding, have been reported to lower the risk to develop ovarian cancer (58, 59, 60). Although this incessant ovulation theory reconciles with certain epidemiological and experimental data, the underlying mechanism remains poorly understood. Moreover, it fails to provide an explanation for the decreased ovarian cancer risk associated with twin pregnancies (61), the increased risk of ovarian cancer among women with obesity (62, 63, 64, 65), and polycystic ovary syndrome (PCOS) (66) Alternative complementary, not mutually exclusive, hypotheses of the incessant ovulation theory have been proposed. The inflammation hypothesis was advocated based on an increased incidence of ovarian cancer among individuals with pelvic inflammatory disease along with the interesting possibility that hysterectomy and tubal ligation could prevent ovarian cancer (12). Endometriosis, which causes a marked local inflammatory reaction, has been linked to ovarian cancer, in particular endometrioid and clear cell histotypes, in a variety of epidemiological and clinical studies. It is also associated with decreased fertility and general endocrine dysfunction, both of which may be important for ovarian cancer development (67, 68). During postovulatory repair, OSE cells have the ability to undergo epithelial-mesenchymal transition (EMT) to modulate to a fibroblast-like form. This EMT might also function as a hemostatic mechanism to accommodate OSE cells that become trapped within the ovary at ovulation, which would allow them to become incorporated into the ovarian stroma as stromal fibroblasts. An inability to undergo EMT would preserve the epithelial forms within the ovarian stroma, and according to the stromal hypothesis (69), the repeated stimulation of the entrapped epithelium by stromal estrogens may result in increased proliferation and transformation. This suggests that stromal cell-epithelial cell interactions can influence ovarian tumor development (70, 71, 72). Cramer and Welch (73) further extended this concept by illustrating the nature of the proposed interplay between gonadotropins and estrogens. They suggest that it is the excessive gonadotropin secretion, which ultimately leads to increased estrogenic stimulation of the OSE, that is responsible for the increased risk of ovarian cancer.
A growing body of evidence indicates that several key reproductive hormones can influence the incidence and/or growth characteristics of ovarian cancer. Expression of these hormone receptors has been reported in OSE and ovarian cancer (Table 1
). This, together with their roles in the physiology of normal and neoplastic OSE cells summarized in our recent review (74), is in line with the hormonal carcinogenesis hypothesis, suggesting that the endocrine factors that control the normal growth of target organs can also provide suitable conditions for neoplastic transformation. Unlike external risk factors such as diet and smoking, endogenous hormone levels are not easily modified; thus, it is important to understand the molecular changes induced by endocrine factors that might have a positive or negative association with neoplastic transformation in ovarian cancer. There have been two main hormonal hypotheses of ovarian tumorigenesis. One is an androgen/progesterone hypothesis stating that higher levels of androgens, which are increased in menopausal or obese women and also seen among women with PCOS, were associated with an increased risk of ovarian cancer, whereas progesterone can protect against ovarian cancer (75). The gonadotropin hypothesis has also been proposed as an underlying mechanism to ovarian cancer, in that excessive levels of gonadotropins, related to the surge occurring during ovulation and the loss of gonadal negative feedback for menopause and premature ovarian failure (73, 76, 77), may play a role in the development and progression of OEC. This theory would also explain the decreased risk of ovarian cancer associated with pregnancy and OC use, which results in reduced exposure to gonadotropins due to the steroid feedback on the pituitary.
|
). The onset of menopause, which happens at approximately age 51 yr, involves changes in gonadotropin levels as a result of cessation of ovarian function and menstrual cycle. The complete cessation of ovarian function results in the loss of negative feedback of ovarian steroids on gonadotropins. In 2–3 yr after menopause, gonadotropin levels are particularly high, such that concentrations of FSH and LH reach a peak of 10–20 times (50–100 mIU/ml) and 3–4 times (20–50 mIU/ml) the values recorded during the proliferative phase of the menstrual cycle, respectively, and after which there is a gradual but slight decline in both gonadotropins (78, 79). Thus, in support of the gonadotropin theory, the incidence of ovarian cancer climbs dramatically around the age at which most women reach menopause. In the United States, the mean age of incidence of ovarian cancer is 57–59 yr, with 50% of all cases occurring after age 65, and age-specific incidence peaks in the mid-70s (80, 81). In fact, 85–90% of incidence occurs in peri- or postmenopausal women, whereas only 10–15% of OEC appears in premenopausal women.
|
Use of hormone replacement therapy (HRT) is another source for exogenous steroids; however, it does not seem to support a role of gonadotropin stimulation on ovarian cancer development. Based on the results from the Women’s Health Initiative, women who received estrogen plus progestin had an increased risk of developing invasive ovarian cancer compared with those assigned the placebo [hazard ratio, 1.6; 95% confidence interval (CI), 0.77–3.24] (85). Other epidemiological findings concerning HRT use and ovarian cancer risk are equivocal (15, 16, 17, 85, 86, 87). It is suggested that the risk for ovarian cancer may also vary with the type of HRT as well as with the duration of use and mode of administration and generally, long-term use of HRT is likely associated with an increased risk (15, 16, 85, 86, 87). The reason why the use of HRT, compared with OC, shows adverse or no effect on ovarian carcinogenesis is still poorly understood. Estrogen levels might explain this discrepancy because the mitogenic effect of estrogen has been well demonstrated in OSE and ovarian cancer (88, 89). In this regard, OC has been shown to decrease endogenous estrogen release, possibly by lowering of basal LH and FSH levels (75). It is also noteworthy that HRT is commonly prescribed for menopausal women who are insensitive to negative feedback regulation of gonadotropin secretion by steroids (90).
The ovarian cancer risk associated with ovulation induction treatment used for infertile women at in vitro fertilization processes has been considered direct evidence for the gonadotropin theory because the ovary is directly exposed to the superphysiological levels of gonadotropins for multiple follicular developments during the controlled ovarian hyperstimulation. Numerous researchers in various population pools have studied whether the superovulation with high gonadotropins is an independent risk factor for ovarian cancer. Two early studies, which prompted discussion of this issue, have suggested strong association between ovulation induction and ovarian cancer incidence. A case-control study by Whittemore et al. (17) was reported in 1992. They collected data from 2197 patients with ovarian cancer and 8893 controls in 12 U.S. case-control studies conducted from 1956 to 1986. In infertile women who had used fertility drugs, a 2.7-fold increased risk of ovarian cancer was observed; however, information on history of infertility and OC use was not provided. In 1994, Rossing et al. (18) reported the risk of ovarian tumors in a cohort of 3837 women tested for infertility between 1974 and 1985 in Seattle with 11 developed malignancies. The risk of invasive ovarian cancer development compared with that in the general population was 2.5 overall (95% CI, 1.3 to 4.5), 1.5 (95% CI, 0.4 to 3.7) in invasive ovarian cancer, and 3.3 (95% CI, 1.1–7.8) in borderline tumors. Women who used clomiphene citrate had a relative risk 2.3 times (95% CI, 0.5 to 11.4) greater than that of infertile women with no use. The risk was most prominent in women with long-term use of clomiphene citrate (12 or more cycles with a relative risk of 11.1% (95% CI, 1.5–82.3). Although this study was relatively well controlled for confounding factors such as OC use, infertility history, and type of fertility drug, it has been criticized for the small number of tumors and no discrimination between epithelial and nonepithelial tumors. Since then, most studies on this issue have observed either a weak association or no association between the infertility treatment and ovarian cancer risk (14, 19, 84, 91, 92, 93, 94, 95, 96, 97). The principal findings in these reports are well summarized in several recent reviews (98, 99, 100). Discrepancies among the findings are seemingly caused by differences in review methods. For example, studies that have limited or no information of the specific types of fertility drugs used failed to show an association between ovulation induction and ovarian cancer incidence (94, 95, 96). In contrast, other studies in which such information was available have shown a mild increase in ovarian cancer incidence with the use of human menopausal gonadotropin (97, 101). In general, epidemiological studies on fertility drug use and risk of ovarian cancer are hampered by problems, such as small study size, short follow-up time, and low prevalence of infertility and fertility drug use. In addition, most of the cohort studies have been terminated before the mean age of ovarian cancer incidence. Thus, the involvement of infertility treatment with ovarian cancer incidence is likely to be underestimated. Although the currently available findings are inconclusive, they are generally reassuring and support the need for well-designed studies including a larger population study, longer follow-up time, and a consideration of histological types and related factors such as OC use.
When investigating the pathogenesis of ovarian cancer, the gonadotropin theory cannot reconcile all the epidemiological observations. For example, the observations of the strong protective effect from a twin pregnancy or the nonprotective effect of HRT, which reduce to some extent the gonadotropin levels of menopausal women, are unlikely to be consistent with the theory. This lack of reconciliation is not unique to this theory alone, and this trait is shared by other hypotheses put forth by other investigators. Again, further research is required.
B. Animal models
Several animal studies initially drew attention to the possible involvement of gonadotropins in ovarian tumorigenesis. In 1944, Biskind and Biskind (76) reported a high incidence of ovarian tumors in rats whose ovaries were autotransplanted to the spleen. However, they failed to observe the formation of the ovarian tumors when one ovary was left intact or when the ovary was autotransplanted in previously hypophysectomized animals (102). This tumorigenesis has been attributed to elevated pituitary gonadotropins due to the deactivation of estrogen in the liver and the consequent depletion of negative feedback of estrogen on the pituitary. Since then, the development of ovarian tumors has been reported in several transgenic or knockout animal models that exhibit hypergonadotropism with high levels of circulating FSH and LH similar to the postmenopausal state in women (103, 104, 105, 106, 107). For instance, mice transgenic for the mouse inhibin 
promoter/Simian virus 40 (SV40) Tag fusion gene developed ovarian tumors (107). However, it should be pointed out that this animal model and as well as others, including Biskind’s model, developed nonepithelial ovarian tumors arising from stromal cells or granulosa cells. These cells were considered to have little relevance to explain the tumorigenic mechanism of the OEC. On the other hand, more recent animal studies that have focused on OSE’s proliferation or the factors involved in this process (108, 109, 110, 111) have demonstrated that administration of gonadotropins for ovulation induction significantly stimulates proliferation in the OSE. This lends support to the gonadotropin hypothesis for OEC.
One study by Davies et al. (109) monitored the proliferative activity of the OSE cells during the lifetime of the mice. The OSE cells were scarcely proliferative in adult life when compared with the prenatal period. The OSE cells of nonpregnant, pregnant, and lactating mice showed little proliferative activity, but they significantly proliferated after the administration of pregnant mare serum gonadotropin or pure recombinant gonadotropins. These data suggest that gonadotropins may play a role in the proliferation of the OSE as well as their classical role in ovarian folliculogenesis (109). Similar patterns of increased proliferation were observed in rat and rabbit OSE after administration of gonadotropins. With regard to the site and duration of proliferation, it has been demonstrated that the increased proliferation of the OSE cells occurs adjacent to the ovulatory site and the corpus luteum, and then it persists throughout most of the postovulatory period. In some studies, proliferation of the OSE cells was also detectable in nonovulatory sites after ovulation (108, 110, 111). Stewart et al. (108) further examined the effect of rupture, estrogen, and superovulation with gonadotropins on the OSE cells in two groups of rats, with and without surgically induced injury to the ovary. Here, the administration of pregnant mare serum gonadotropin and hCG led to a more than 10-fold increase in proliferation of rat OSE cells, and the effect persisted up to 40 h. In contrast, surgical injury mimicking the rupture of ovulation induced just a 5-fold increase in proliferation, and the effect was relatively transient. It is noteworthy that surgery failed to alter the response of the OSE to the gonadotropin treatment, whereas it doubled the effect of estradiol alone on OSE proliferation; thus, estrogen, but not gonadotropins, may be involved in the regulation of the wound-healing process of the OSE. Based on the results, the author postulated that hormones related to infertility treatments and HRT as well as injury- or ovulation-induced rupture of the OSE significantly stimulated proliferation of OSE cells. Consequently, this could contribute to ovarian tumorigenesis via proliferation-stimulated DNA mutations and/or a rapid selection of OSE cells with accumulated mutations (108). Experimental evidence to support the susceptibility of OSE cells to mutagenic events during mitosis is provided by studies showing that rat and mouse OSE cells, which have been repeatedly subcultured in vitro, elicits malignant transformation and the ability to form tumors in nude mice (112, 113).
One important finding published by Celik et al. (114) is that gonadotropins may associate with dysplastic morphological transformation of the OSE. Histopathological examinations in three groups of rats were performed, including one, three, and six cycles of ovulation induction groups at the 12th month after treatment ended. Malignant ovarian lesion did not develop in any of the three groups after ovulation induction; there was, however, epithelial dysplasia such as inclusion cysts, tufts, and epithelial stratification when the number of induction cycles increased. Because inclusion cysts and tufts are sites prone to develop ovarian cancer, these data seem to support the possibility that gonadotropin-induced superovulation and proliferation of OSE stimulate ovarian cancer development by increasing epithelial dysplasia in the ovary (114). Direct evidence for gonadotropins’ contribution to the initiation of experimental OEC has yet to be established. One problem here is that in contrast to humans, OSE-derived carcinomas are essentially nonexistent in rodents (35). In fact, the structure, physiology, and differentiation differ between human and rodent OSE (115, 116, 117). Thus, specially designed studies using the human surface OSE and/or inclusion cyst are necessary to explore the direct mechanism for the contribution of gonadotropins to ovarian epithelial tumorigenesis.
Most research has focused on the role of gonadotropins in the initial transformation of normal OSE to neoplastic OSE. More recently, the possible involvement of gonadotropins in tumor progression was examined in several animal models. Stewart et al. (118) developed a rat model using a carcinogen, DMBA (7, 12-dimethylbenz(a)anthracene), that closely resembled the histopathology, pattern of metastasis, and mutations of Tp53 and Ki-Ras in human OEC. Interestingly, the addition of gonadotropins after treatment with DMBA considerably increased lesion severity in this animal model. This suggested that gonadotropin stimulation may have played a contributing role in the neoplastic progression of the ovarian lesions, presumably due to the increased rate of OSE cell proliferation and their effects on the underlying stroma (118). Elevated levels of gonadotropins are also related to increased angiogenesis of ovarian cancer, probably through the regulation of vascular endothelial growth factor (VEGF). For example, ovariectomy-induced or directly administrated gonadotropins stimulated the progression of human epithelial ovarian tumors in mouse models. Magnetic resonance imaging data showed increased neovascularization of human ovarian carcinoma spheroids in ovariectomized mice, and treatment with gonadotropins up-regulated VEGF in vitro and in vivo (119). As for the mechanism of gonadotropic stimulation, it was suggested that the excessive levels of gonadotropins acts in the stroma compartment, rather than in OSE cells, as a stimulatory factor for steroidogenesis. Consequently, the steroidogenic stromal cells secrete increased levels of estrogen, which may play a mitogenic and/or carcinogenic role in the OSE cells in a paracrine manner.
Recently, data have been rapidly accumulating to demonstrate that gonadotropin receptors, FSH receptor (FSHR) and LH receptor (LHR), are expressed in the OSE as well as their classical expression sites such as granulosa and theca cells. Furthermore, when FSH and LH bind to their respective receptors, they activate various key signaling transduction pathways, resulting in a number of biological effects. These findings brought the gonadotropin theory into a new era, where the direct effect of gonadotropins in the OSE cells should be considered to be a legitimate candidate to evaluate the association of gonadotropins with ovarian tumorigenesis, as well as their indirect effects on other ovarian cells such as granulosa and theca cells. Furthermore, expression of gonadotropin receptors in ovarian cancer cells strongly suggests a potential role of gonadotropins in various aspects of ovarian cancer progression, including metastasis and chemoresistance.
C. Gonadotropin levels in patients with OEC
Several groups have studied an association of prediagnostic serum gonadotropin levels with the risk for OEC. For example, no association between serum FSH and ovarian cancer risk (120, 121) or even reverse correlation has been observed (122, 123). Similarly, data on circulating LH levels suggested that neither wild-type LH level, nor its variant containing two missense point mutations in its ß-subunit, contributes to the ovarian cancer risk (121, 124). The free ß-subunit of hCG possibly produced from the tumor is generally increased in the serum of women with ovarian cancer, in contrast to the pituitary gonadotropins (125, 126, 127). Its prognostic value has also been evaluated (128). A recent study that demonstrated the hCG-ß gene expression in ovarian cancer tissues, in contrast to the lack of its expression in noncancerous tissues, proposed the possibility that the high level of circulating hCG-ß is directly produced from ovarian cancer tissue (129). However, it is still unknown whether hCG-ß plays a physiopathological role in OEC. Interestingly, serum LH levels are associated with BRCA1 mutation status. LH levels in the follicular phase among young healthy BRCA1 mutation carriers was drastically elevated compared with noncarriers from BRCA1 families, suggesting that high levels of LH may be a part of the BRCA1 phenotype (130). In contrast, no significant difference among serum FSH levels in noncarriers, BRCA1 mutation carriers, and women from non-BRCA1/2 families was found.
The observations of serum levels of gonadotropins in women with ovarian cancer are inconsistent and in some cases appear to be contradictory to the gonadotropin hypothesis of ovarian epithelial carcinogenesis. In contrast, gonadotropin levels in ovarian cyst and peritoneal fluid appear to be associated with ovarian cancer incidence. For instance, Kramer et al. (121) found high levels of gonadotropin in cysts of serous ovarian cancer and low or absent gonadotropin levels in benign ovarian tumors. There was no significant difference in the serum FSH and LH concentrations between patients with malignant tumors and nonneoplastic cysts. LH levels in peritoneal fluid and ovarian cyst fluid from tumor aspirates was considerably higher in patients with ovarian cancer or borderline ovarian tumor than in patients with benign cysts or tumors (131, 132). The concentrations of both gonadotropins in ovarian cancer fluid were greater than in fluid from borderline tumors, benign tumors, and functional cysts of the ovary, and the serum/tumor fluid ratio was lowest in ovarian cancer (133). The observation of high levels of FSH and LH in cyst fluid of malignant ovarian tumors is consistent with the gonadotropin theory. This information suggests that FSH and LH might play a role in ovarian cancer development; however, further larger-scale experiments are necessary to confirm the relationship of the high levels of ovarian and peritoneal gonadotropins to the risk for ovarian cancer.
D. Expression of gonadotropin receptors in OSE and OEC cells
Gonadotropin receptors are G protein-coupled receptors with seven- transmembrane domains. In fact, they contain a large extracellular domain that is necessary for binding of their heterodimeric ligands, a transmembrane domain including three extracellular and intracellular loops, and a cytoplasmic domain binding to G protein. The FSHR is expressed by granulosa cells, whereas LHR is expressed by theca cells in early developing follicles and by both theca and granulosa cells during the later stages of follicular development. In addition to their primary target site, both FSHR and LHR have been found to be expressed in normal OSE cells and ovarian tumors (134, 135, 136, 137, 138, 139, 140, 141) (Table 2) with 80% of cystadenomas, 71% of borderline tumors, and 40% of epithelial ovarian carcinomas being LHR/hCGR positive. The patients with LHR-positive cancer had a better prognosis compared with those without LHR; yet there was no significant correlation between the receptor expression and the clinical stage of the disease (138). Additionally, Lu et al. (140) examined the LHR expression in tumor epithelium and tumor stroma separately. The positivity of LHR declined from precursor lesions, to benign and borderline tumors, to invasive ovarian cancer in both tumor epithelium and stroma. This information suggests a role for LH in early tumorigenesis via direct stimulation of the epithelium and/or indirect activation of stromal cells (140). Zheng et al. demonstrated the mRNA expression of FSHR in 100% of epithelial inclusions, 100% of cystadenomas, 94% of borderline tumors, and 60% of the carcinomas, and, although the result was not statistically significant, negative correlation between the positivity of FSHR and carcinoma grade was observed. This suggests that FSH may be more important in early tumorigenesis (134). Recent studies have suggested a relationship between the expression of gonadotropin receptors and ovarian cancer development (90, 140, 142). Indeed, semiquantitative RT-PCR data demonstrated that the relative mRNA levels of FSHR in ovarian cancer cells were higher than those in primary cultures of human OSE cells or in immortalized cell lines, whereas an opposite expression pattern was observed for LHR mRNA (90). These data are consistent with this recent observation that used quantitative real-time PCR. FSHR mRNA was significantly enhanced in the following order: normal OSE, benign tumors, low malignant potential (LMP) tumors, and invasive ovarian cancer (143). Furthermore, FSHR levels increased from presumed precursor lesions [ovarian epithelial inclusions (OEIs)], to benign ovarian epithelial tumor (OET), and to borderline OETs, whereas its levels decreased from borderline OETs to ovarian carcinomas (142). Together, these observations suggest that high levels of FSHR in the ovarian epithelium may play a role in tumorigenesis of ovarian cancer, especially early in the process. To investigate the role of FSHR in ovarian cancer development, we overexpressed the receptor in SV40 Tag immortalized OSE (IOSE) cells, which are preneoplastic and nontumorigenic. It should be noted that overexpression of FSHR resulted in activation of ERK1/2 MAPK and an increased expression of potential oncogenes including epidermal growth factor receptor (EGFR), c-myc, and HER-2/neu, in IOSE cells. In addition, the overexpression of FSHR accelerated cell proliferation in these cells. These results may support a pivotal role of FSHR in ovarian cancer development in terms of neoplastic conversion and growth potential (144).
|
-subunit of adenylyl cyclase-binding G proteins. This information suggests that there is no relation of genetic mutations of the receptors to ovarian tumorigenesis (153). The clinical findings have demonstrated that the ovarian response to FSH stimulation is associated with the FSHR genotype (152, 154, 155, 156). For example, two nonsynonymous SNPs of the FSHR gene, which occur at Ala307Thr and Ser680Asn in the coding region of the gene, have been identified. Among 202 ovarian cancer patients and 266 age-matched cancer-free controls, the SNP carriers have shown an increased susceptibility to develop serous and mucinous types of ovarian cancers, but not endometrioid or clear cell types (156). The functional aspects of these SNPs in ovarian carcinogenesis require further study. |
IV. Action of Gonadotropins in OSE and OEC Cells |
|---|
|
TopAbstractI. IntroductionII. Biology of Ovarian...III. Gonadotropin TheoryIV. Action of Gonadotropins...V. Clinical ApplicationsVI. Conclusion and Future...References |
|---|
).
|
A number of findings from gene expression profiling in gonadotropin-stimulated granulosa cells have suggested a possible involvement of the indirect effect of gonadotropins on granulosa cells. Gene transcripts for growth factors such as epiregulin and amphiregulin, whose binding to EGFR and/or ERB4 stimulates oncogenic/mitogenic signaling pathways, were markedly elevated in human granulosa cells after LH and FSH treatment along with the predicted amplification of steroidogenesis-related genes, such as StAR, cytochrome P450scc, and aromatase, in several cancers including ovarian cancer (158). These data are consistent with a previous finding for up-regulation of the epiregulin in granulosa layer after the administration of hCG to the rat (159). This increase in growth factors by gonadotropins may be associated with an enhanced risk not only for GCT but also for OEC due to the excessive exposure of trapped OSE to the growth factors secreted from granulosa cells.
This speculation for the role of gonadotropins in the growth factor milieu of the ovary is further supported by the findings in OSE treated with FSH or LH/hCG. In bovine OSE cells, gonadotropins increase the expression of growth factors such as keratinocyte growth factor (KGF), hepatocyte growth factor (HGF), and kit ligand, which possibly act as autocrine and/or paracrine mitogenic factors (136). Similarly, both hormones increased mRNA expression of EGFR in the bovine OSE cells, whereas hCG augmented the expression of TGF, a ligand for EGFR; yet, this was not found with FSH. Collectively, these data suggest the possible association of gonadotropins with a growth factor-receptor tyrosine kinase system in ovarian cells including the granulosa and the OSE cells, resulting in a more proliferative microenvironment for these cells (160).
B. Proliferation
FSH and LH/hCG treatment appears to regulate the growth of normal OSE, IOSE, and some ovarian cancer cells in vitro. Although administration of gonadotropins certainly stimulates proliferation of the OSE cells in most animal models (108, 109, 110, 111), the reports on the association of gonadotropins with the proliferation of OSE and ovarian cancer cells have been inconclusive. The following collection of published studies illustrates this point effectively. FSH possibly increases proliferation in the OSE and cancer cells (90, 134, 136, 141, 143, 161, 162, 163, 164, 165), yet in other studies, no association (166) and even reduced proliferation have been observed (167, 168). Some studies have reported the stimulatory effect of LH/hCG on cell growth (90, 136, 161, 163, 164, 165, 169, 170, 171, 172, 173). And again in other studies, its inhibitory effect was noticed (134, 174) and no association (166, 167) was observed. These studies were performed with highly purified or recombinant gonadotropins, indicating negligible contamination. Thus, the reason for the discrepancy in results is not known, but it may be that variations in hormone concentration or the culture density used in different studies could be affecting the outcomes (169, 170, 175). It is also possible that variations in the expression of the gonadotropin receptors or their downstream signaling molecules between cultures of OSE derived from different women may explain an inconsistent response to gonadotropins (165).
The mechanism by which gonadotropins regulate the growth of OSE and ovarian cancer cells has been controversial. A few studies suggested that there may be a relationship between gonadotropins and other hormones that have been known to be mitogenic in ovarian cancer. For instance, Kraemer et al. (163) found that OVCAR-3 has steroidogenic activity and produces estrogen responding to hCG and FSH. Because of the mitogenic effect of estradiol in ovarian cancer cells, the gonadotropin-stimulated secretion of estrogen led to growth stimulation in OVCAR-3 cells. This role of gonadotropins in steroid metabolism in OEC supports a previous finding by Wimalasena et al. (176) that about 30% of primary cystadenocarcinoma cultures respond with increased estradiol secretion to hCG and FSH. One report by Zheng et al. (177) indicated the involvement of unopposed production of activin A, which stimulates the proliferation, in an increased proliferation by FSH in ovarian cancer cells. Moreover, recent studies have suggested interactions between gonadotropins and growth factors and/or cytokines. Indeed, combined treatments of hCG with estradiol may regulate the growth response of OEC cells through a mechanism involving IGF-I (173). This finding is further supported by a recent report proposing the possible association of IGF-I and intergrin ß1 with anchorage-dependent and -independent growth of OSE cells by LH (170). In addition, FSH and hCG stimulated steady-state mRNA levels of KGF, HGF, and kit ligand in bovine OSE cells (136). Moreover, FSH-, LH-, and estrogen-stimulated IOSE cell proliferation involves the IL-6/signal transducer and activator of transcription-3 (STAT3) signaling pathway (90, 178).
With regard to the signaling pathways, it is generally accepted that the effects of gonadotropins on classical target cells such as granulosa and Leydig cells are mediated by the protein kinase A (PKA) pathway, such that activation of adenylyl cyclase by the stimulatory G protein, Gs, is followed by a rapid increase in cAMP and a subsequent activation of PKA (179). Some studies have proposed that the PKA pathway is also involved in gonadotropin-induced proliferation of OSE and ovarian cancer cells. For example, a specific PKA inhibitor, H89, completely abolished gonadotropin-stimulated cell growth via the IL-6/STAT3 pathway in human OSE and ovarian cancer (90, 178). Interestingly, growing evidence shows that FSHR and LHR can activate a number of additional cellular signaling pathways: protein kinase C (PKC), phosphatidylinositol 3-kinase (PI3K), and MAPK in granulosa, OSE, and ovarian cancer cells (180, 181, 182, 183, 184, 185, 186, 187, 188, 189, 190). Treatment of OEC cells with FSH significantly increased the levels of PKC mRNA and protein, suggesting that the stimulation of PKC transcription is involved in FSH-induced cell proliferation (162).
Our findings have also demonstrated that after treatment with FSH, phosphorylation of ERK1/2 and cell growth were significantly increased in preneoplastic IOSE-29 and neoplastic IOSE-29EC cells (191). FSH-induced activation of ERK1/2 and proliferation were completely abolished in the presence of PD98059, suggesting that the growth-stimulatory effect of FSH is mediated by ERK1/2. In addition, treatment with FSH significantly phosphorylated a transcription factor, Elk-1, a recognized downstream target of ERK1/2. The fact that FSH did not stimulate basal cAMP levels in OSE cells, but was observed in human granulosa cells, is a noteworthy observation (141).
More recently, we reported that gonadotropins up-regulate EGFR levels through activation of ERK1/2 and PI3K in human IOSE cells (192) (Fig. 2). Treatment of IOSE-80PC cells with FSH and LH resulted in a significant increase in EGFR mRNA at 24 h and EGFR protein at 48 h, whereas only a mild increase was observed in OVCAR-3 cells. In addition, IOSE-80PC cells treated with gonadotropins and EGF exhibited an additive stimulation of mitogenesis. FSH and LH significantly increased the activities of various kinases at 5–10 min. Pretreatments with LY294002 (an inhibitor of PI3K) or PD98059 (an inhibitor of MEK1) partially blocked gonadotropin-induced up-regulation of EGFR mRNA in IOSE-80PC cells. Our results suggest that FSH and LH use alternate mechanisms to increase EGFR mRNA, such that the former increases EGFR gene transcription, whereas the latter enhances EGFR mRNA stability. Taken together, these results suggest that FSH- and LH-induced activation of the ERK1/2 and PI3K/protein kinase B (Akt) pathways may affect cell growth by means of a direct mitogenic effect or by interaction with a growth factor system(s), such as EGFR, in preneoplastic OSE cells.
|
s-coupled ß2-adrenergic receptor increased cAMP levels and stimulated integrin-medicated cell adhesion via Epac and Rap1 (198). Furthermore, the identification of a binding partner for FSHR, APPL1, provides a potential link between G protein-coupled receptor activation and PI3K signaling. Specifically, APPL1 is an adaptor that binds to inactive AKT2 and the p110 catalytic subunit of PI3K and the result of their interaction leading to active AKT2 and p110 (199).
C. Apoptosis
The overall 5-yr survival rate for advanced ovarian cancer patients is low (20–30%) due to chemoresistance to conventional cytotoxic drugs such as cisplatin in the primary or recurrent tumors that results in treatment failure (200, 201). Compared with the effects of gonadotropins on proliferation of OSE and OEC cells, little is known about whether gonadotropins play a role in chemosensitivity of ovarian cancer. During ovarian folliculogenesis, gonadotropins—especially FSH—changed the expression of antiapoptotic proteins such as Bax (202) and XIAP in granulosa cells and inhibited apoptosis (203). In ovarian cancer, limited data that have examined the link between gonadotropins and regulation of apoptosis are available. In OVCAR-3 cells bearing LHR, LH remarkably inhibited cisplatin-induced apoptosis by increasing the expression of IGF, but not Bcl-2 family proteins such as Bcl-2 and Bax (172). The TNF receptor superfamily member 6 (Fas) system is well known to mediate the cisplatin cytotoxicity. LH interfered with the Fas-induced apoptosis in HEY and CaOV-3 cells (204). Similarly, a study found that pretreatment of OVCAR-3 cells with FSH significantly reduced cisplatin-induced apoptosis and cell cycle arrest using various methods such as MTT, TUNEL assay, and flow cytometry. Furthermore, this inhibitory effect is mediated by overexpression of an inhibitor of apoptosis protein family survivin and down-regulation of caspase activity (205, 206). These data indicate that gonadotropins may be involved in chemoresistance, which is a main hurdle for ovarian cancer management. Accordingly, some reports revealed the antiapoptotic effect of gonadotropins in OSE cells (165, 172).
D. Invasion and angiogenesis
The adhesion and angiogenesis of ovarian cancer play a crucial role in peritoneal metastatic dissemination and nutrition supply, resulting in the progression of ovarian cancer via invasion and metastasis. Gonadotropins enhance tumor angiogenesis and adhesion in ovarian cancer cells (119, 207, 208, 209). Schiffenbauer et al. (119) found that human OEC progressed faster with increased neovascularization in ovariectomized mice as a result of elevated FSH and LH levels, which promoted the expression of VEGF in a dose-dependent manner in vitro. Neovascularization was significantly increased both in vitro and in vivo by hCG (207). Gonadotropins up-regulated VEGF expression in both malignant and borderline tumor; however, the effect of gonadotropins was significantly stronger in malignant tumor than in borderline tumor (209). In addition to angiogenesis and vascular remodeling, treatment with gonadotropins increased integrin v and CD44, the cell surface hyaluronan receptor, resulting in enhanced adhesion of the cancer cells to culture plates coated with hyaluronan, fibronectin, as well as a thrombin-derived RGD containing peptide (208). One significant finding we discovered was that treatment with FSH or LH significantly increased the invasion of ovarian cancer cells, including BG-1, CaOV-3, and SKOV-3 cells (210) (Fig. 3). Treatment of SKOV-3 cells with FSH or LH enhanced the net balance of metalloproteinase (MMP)-2 and -9/ tissue inhibitor of metalloproteinase (TIMP)-1 and -2, resulting in enhanced proteolytic activities. In addition to the above finding, we demonstrated that gonadotropins induced an increase in SKOV-3 invasiveness via the activation of PKA and PI3K signaling pathways. No association was observed between gonadotropins and urokinase-type plasminogen activator (210, 211).
|
|
V. Clinical Applications |
|---|
|
TopAbstractI. IntroductionII. Biology of Ovarian...III. Gonadotropin TheoryIV. Action of Gonadotropins...V. Clinical ApplicationsVI. Conclusion and Future...References |
|---|
subunit has not been associated with ovarian cancer (125). Interestingly, high LH levels, but not FSH and estradiol, were found in ovarian cancer-prone women carrying BRCA1 mutations (130). It remains to be determined whether elevated LH levels are part of the BRCA1 phenotype or may be manipulated to reduce cancer risk; however, clinically, LH might be helpful in the early detection of ovarian cancer. An immunological variant of LH has been described and attributed to two point mutations (Trp8Arg and Ile15Thr) in gene coding the LH ß-subunit. Although there is no evidence that the LH-ß variant increases risk for ovarian cancer, it has an impact on ovarian function and LH concentrations (124, 214). Reimer et al. (216) were the first to measure gonadotropin concentrations in tumor fluid and found significantly lower FSH (0.45 IU/liter) and LH (0.55 IU/liter) concentrations in functional cysts compared with neoplastic cysts (FSH, 10.2 IU/liter; LH, 2.4 IU/liter). Furthermore, it has been demonstrated that FSH and LH levels in the cyst and peritoneal fluid from patients with ovarian cancer or borderline ovarian tumors were much higher as compared with patients with benign cysts or tumors; and yet there was no significant difference in the serum FSH and LH concentrations, suggesting a lower serum to tumor ratio in malignant and borderline cancer cases as compared with the cysts (121). Therefore, the ovarian and peritoneal gonadotropin patterns might provide yet another strategy for diagnosis of OEC, and at the same time, substantiate the use of GnRH analogs with the aim of lowering gonadotropin levels for the management of ovarian cancer.
B. Hormone therapy
Surgical removal of most, if not all of the cancer, followed by anticancer chemotherapy is the routine treatment for ovarian carcinoma. However, various hormone therapies, such as antiestrogens, progestins, GnRH analogs, and antiandrogens have been tested in recurrent and/or chemoresistant ovarian cancer patients. GnRH analogs including both agonist (triptorelin and leuprolide) and antagonist agents (cetrorelix and flutamide) reduce gonadotropin secretion from the pituitary through eventual desensitization of the pituitary receptors. Several clinical studies upon the gonadotropin theory were carried out using these hormonal analogs. Initially, a number of phase I and phase II clinical trials suggested that GnRH analogs may result in relevant response and/or disease stabilization in 10–30% of patients with refractory ovarian cancer (217). Despite these encouraging results, the benefit of using GnRH analogs as a second-line therapy of advanced ovarian cancer is debatable. In several recent studies, the GnRH agonists, such as leuprolide and triptorelin, were used as a second-line therapy (218). Generally some, but not all, agonists showed modest efficacy in patients with platinum-refractory ovarian cancer (219, 220, 221). The GnRH antagonist cetrorelix was also tested in a phase II study. Of 17 patients who received GnRH antagonist cetrorelix 10 mg sc daily, three had partial remissions lasting 9, 16, and 17 wk with minor toxicity, but incurred potential anaphylactic reactions (222). This is likely consistent with a observation that cetrorelix, alone or in combination with the bombesin/gastrin-releasing peptide antagonist, effectively inhibited the growth of ES-2 cells engrafted nude mice (222, 223).
The combination therapy of another GnRH agonist goserelin with antiestrogenic tamoxifen or conventional platinum regimens has shown promise in the treatment of the malignancy. Patients with recurrent ovarian cancer that were treated with oral tamoxifen 20 mg twice daily and sc goserelin 3.6 mg once a month until disease showed progression. About 50% of patients exhibited an "endocrine response," including patients with complete response (3.8%), partial responses (7.7%), and stable disease (38.5%) (224). In a study that tested the combination effect of chemotherapy, radiotherapy, and goserelin, the chemotherapy was effective only with goserelin treatment or radiotherapy. In the goserelin treatment group, the 5-yr survival rate rose 54% in contrast to only 19% in the nontreatment group. Of the goserelin treatment group, patients surviving at least 5 yr or in complete remission at the time of this study showed significantly lower FSH levels in serum (225). In addition to lowering the circulating levels of gonadotropins, the antiproliferative and/or apoptosis-inducing effect of GnRH analogs via their receptor in the ovarian cancer cells has been implicated as their supplementary mechanism of action (224). In fact, GnRHR is expressed in about 80% of ovarian cancer and its ligands. Both GnRH I and GnRH II have been shown to inhibit cell growth (226, 227, 228, 229, 230). The conclusion drawn from these studies is that hormonal therapies including the GnRH analogs may be beneficial because of their relatively lower toxicity, fewer side effects, and higher tolerability.
C. Drug development
A number of researchers have attempted to increase the specificity of conventional cytotoxic drugs based on the common expression of gonadotropin receptors in ovarian cancer cells. Despite a general prolonged survival rate by conventional chemotherapies that contain platinum, taxanes, and anthracyclines for ovarian cancer, there has been limited use of chemotherapy regimens due to significant side effects and a high recurrence rate. Cytostatics conjugated with hCG have significantly increased the toxicity in LHR-expressing breast (231) and prostate cancer cells (232, 233). Treatment with hCG-conjugated doxorubicin for 2 h induced more than an 8-fold increase in cytotoxicity compared with unconjugated doxorubicin in ovarian cancer. This effect showed prolonged antiproliferative action up to 168 h rather than an acute cytotoxic action (234). The conjugate-induced cytotoxicity in ovarian cancer cells was likely dependent on LHR expression levels (235). A similar study on lytic peptides conjugated to hCG-ß has shown that the conjugates selectively destroy OVCAR-3 cells in vitro and OVCAR-3 cells engrafted in nude mice models in vivo (236). In a group of animals treated by hecate-hCG-ß, tumor volume expressed as a percentage of increase was 199 ± 18.57% when compared with control animals (263.0 ± 21.72%). The expression of LHR was increased when estradiol was present (236), resulting in a positive correlation between steroids in the culture medium and sensitivity of the OVCAR-3 cells to the hecate-hCG-ß. Together, these findings suggest that cytotoxic drugs conjugated with gonadotropins may be used for treating gonadotropin receptors-bearing OEC.
|
VI. Conclusion and Future Directions |
|---|
|
TopAbstractI. IntroductionII. Biology of Ovarian...III. Gonadotropin TheoryIV. Action of Gonadotropins...V. Clinical ApplicationsVI. Conclusion and Future...References |
|---|
). A deeper and comprehensive analysis with a large population study that provides a longer follow-up and includes a consideration of histological types and related factors such as OC use should provide a better understanding of how gonadotropins influence cancer risk. Supporting evidence of ovarian epithelial tumorigenesis from animal models with targeted manipulation of gonadotropin secretion or action would be particularly desirable, but it has yet to be established and is an area for future research. The reported observations here on gonadotropins’ role in OEC, including proliferation, apoptosis, adhesion, invasion, and angiogenesis, not only confirm their importance in the cancer process, but also may prove to be an integral part of the translation into clinical applications to eliminate primary and metastatic ovarian cancers. Additionally, the role of gonadotropin in ovarian cancer development should be further explored in terms of its possible involvement with stromal-epithelial interaction and regulation of OSE stem cell self-renewal and differentiation. The recent cDNA microarray analyses and characterization in the molecular mechanisms of gonadotropin signaling have indicated the effects of gonadotropins on the regulation of normal OSE and OEC cell growth, survival, and metastasis and may involve other growth factors mediated by a complex array of intracellular signaling pathways (Fig. 4). These studies provide a great deal of novel information on understudied mechanisms involved in dysregulation of gonadotropin functions. Further investigation may establish a new paradigm of therapy for targeting ovaria
