|[Frontiers in Bioscience 1, d1-11, January 1, 1996]|
THE MOLECULAR BASIS OF OVARIAN CELL DEATH DURING GERM CELL ATTRITION, FOLLICULAR ATRESIA, AND LUTEOLYSIS
The Vincent Center for Reproductive Biology, Department of Obstetrics and Gynecology, Massachusetts General Hospital/Harvard Medical School,Boston, MA 02114 USA
Received 12/14/95; Accepted 12/29/95; On-line 1/1/96
Despite the fact that the regulation of cell death has been extensively studied in several extragonadal tissues, only recently have reproductive biologists realized that the cyclic nature of germ cell development in the ovary and testis is also very tightly linked to the process of apoptosis. At one of the earliest points during embryonic ovarian development, when oogonia are undergoing mitosis to increase the total germ cell pool, large numbers of these cells are lost through degeneration [62-64]. Additionally, oogonia that escape death during mitotic division may still be lost via apoptosis during the first phases of meiosis when the diploid oogonia mature into haploid oocytes [62,64-66]. Although the reasons for the massive destruction of oogonia and oocytes remains a mystery, it has been estimated that the fate of up to two-thirds of the ovarian germ cell pool present during embryogenesis is degeneration. Moreover, it is now known that oogonia and oocytes which degenerate in vivo or in vitro exhibit many of the morphological and biochemical criteria associated with an apoptotic event. These features include chromatin margination and fragmentation with the resultant nuclear pyknosis, cellular condensation, and induction of cell death markers such as tissue transglutaminase [67-69].
From a regulatory standpoint, somatic cell-derived growth factors, most notably the product of the Steel gene (stem cell growth factor/SCF or kit ligand), appear to be critically important for the maintenance of germ cell survival. Data which strongly support this concept come from both in vivo (naturally-occurring mutations in the genes encoding SCF or the SCF receptor, c-kit) and in vitro (cultures of primordial germ cells) analyses. It has been known for almost forty years that mutations in the SCF and c-kit genes lead to gonadal dysgenesis and sterility . The fundamental role of SCF in germ cell survival has been further substantiated by studies of SCF effects on primordial germ cells cultured in vitro. As anticipated from the genetic mutation studies, the presence of SCF prevents apoptosis which occurs in germ cells deprived of tropic hormone support . Interestingly, SCF, although critical for germ cell survival, does not by itself markedly alter germ cell proliferation, suggesting that the actions of SCF in the developing ovary must be complemented by the presence of additional growth factors.
The intracellular pathways which mediate the anti-apoptotic effects of SCF (and other factors?) in germ cells are not well-understood. A recent report has linked activation of the retinoic acid receptor pathway to enhanced germ cell proliferation , suggesting that downstream genes which are targets for transcriptional regulation by the ligand-activated retinoic acid receptor may be important components of oogonium and oocyte survival. The identity of these genes have yet to be elucidated. However, recent findings obtained from analysis of ovaries collected from mice possessing a targeted disruption of the bcl-2 gene have revealed that a loss of BCL-2 bioactivity is associated with a compromised endowment of oocytes and primordial follicles . Based on the similar phenotypes observed in ovaries of mice which lack SCF, c-kit or BCL-2, it has been proposed that SCF mediates germ cell survival, at least in part, via enhanced expression of bcl-2 . Concrete proof of this hypothesis, however, awaits further testing.
Once follicles are established in the ovary during the peri-natal period, germ cell loss occurs primarily as an indirect consequence of atresia of follicles not selected for ovulation. In all species studied thus far, the initiation of apoptosis in granulosa cells is one of the earliest signs of follicular demise [reviewed in 73]. The occurrence of apoptosis in granulosa cells of atretic follicles has been documented by both morphological [74,75] and biochemical criteria [76,77], and the use of DNA oligonucleosomes to identify apoptotic cells has recently paved the way for a large number of studies concerning the regulation of atresia (see below). Data derived from these efforts have demonstrated that, in vivo, pituitary-derived gonadotropins (follicle-stimulating hormone or FSH, luteinizing hormone or LH) are the primary endocrine factors responsible for inhibiting apoptosis in granulosa cells of developing follicles [78,79]. These findings have been confirmed and extended through analysis of apoptosis in granulosa cells of ovarian follicles incubated in vitro as a model for elucidating the events associated with atresia [80,81]. Moreover, the ability of FSH and LH to suppress apoptosis and atresia likely involves augmented signaling through intrafollicular growth factor pathways [78,80], which in turn may be tied to progesterone as a final autocrine mediator . It should also be pointed out that gonadotropin-independent pathways probably contribute to the maintenance of granulosa cell survival [83,84], collectively demonstrating the complexity of events which surround the fate of any given follicle. The involvement of multiple hormonal signaling pathways in apoptosis regulation has been reported for granulosa cells of the avian ovary as well [83,85]. These findings suggest that, regardless of species, the final step which either activates or represses widespread apoptosis during atresia is dependent upon the prior interpretation of many extracellular signals.
The intracellular effectors responsible for the regulation of apoptosis in granulosa cells are no less complex. Studies towards this endpoint have, however, been greatly facilitated by the wealth of information available in the literature concerning the genes involved in apoptosis in cells of extragonadal tissues (see previous section and Table 1). Using these data as a foundation, initial investigations centered on the potential role of BCL-2 and related factors in deciding the fate of rat granulosa cells during follicular development . Results from these studies indicated that increased expression of the bax death susceptibility gene coincides with the induction of apoptosis in granulosa cells during atresia both in vivo and in vitro . The role of BAX in mediating granulosa cell demise has been reinforced by histological analysis of ovaries collected from mice deficient in functional BAX protein . These data have indicated that "knock-out" of the bax gene renders granulosa cells resistant to the normal induction of apoptosis in follicles destined for atresia (e.g. possessing degenerative oocytes) . These findings provide concrete proof that this cell death factor plays a fundamental role in regulating apoptosis in granulosa cells. Expression of the bcl-2 and bcl-x genes in granulosa cells of rodent and avian ovaries has also been reported [79,85,87], albeit the role of the proteins encoded by these genes in determining the fate of granulosa cells remains to be established.
In keeping with the proposal that apoptosis proceeds in many tissues via a universal pathway, a follow-up study was conducted to complement the data derived from analysis of bcl-2-related genes in the ovary. This series of experiments reported on the potential role of reactive oxygen species as a trigger for apoptosis in granulosa cells during atresia . Data derived from these investigations demonstrated that the ability of FSH to suppress apoptosis in granulosa cells of rat follicles in vitro could be mimicked by addition of anti-oxidant enzymes to the culture medium. Although these data provide the first evidence that reactive oxygen species may activate the apoptotic pathway in granulosa cells, it remained unclear how large enzymes such as superoxide dismutase and catalase, when added exogenously, could convey protection from reactive oxygen species within the cell. However, recent experiments with cultured cells have demonstrated that addition of exogenous catalase to the culture medium does in fact dramatically raise the intracellular levels of catalase activity , thus supporting the concept that prolonged oxidative stress may be one determinant of granulosa cell demise . Moreover, expression of factors such as superoxide dismutase (secreted and mitochondrial) in the immature rat ovary is markedly increased in response to exogenous gonadotropin treatment in vivo . Collectively, these data suggest that not only do gonadotropins enhance expression of anti-oxidant genes in the ovary, but that the defense enzymes encoded by anti-oxidant genes (e.g. superoxide dismutase, catalase) are as effective as gonadotropins in the suppression of granulosa cell apoptosis. However, more work is needed to characterize the pathways activated by reactive oxygen species in granulosa cells, and to elucidate how these pathways then converge with other signaling events to activate apoptosis during atresia.
The role that transcriptional regulators play in the process of granulosa cell apoptosis has primarily centered on the p53 tumor suppressor protein. Two reports are currently available in the literature, one of which documents by immunocytochemical procedures the nuclear accumulation of p53 in rat granulosa cells destined for apoptosis . By comparison, p53 is absent in ovaries collected from gonadotropin-stimulated rats, consistent with the fact that this cell death inducer is only present in ovarian cells during episodes of apoptosis . In support of these data, over-expression of p53 in rat granulosa cells has been reported to trigger a rapid onset of apoptosis , confirming that the actions of p53 in the ovary are indeed tied to the regulation of cell death. Although nothing is known of how nuclear accumulation of p53 in granulosa cells leads to apoptosis, p53 is detectable only in follicle populations that concomitantly express high levels of bax (e.g. atretic follicles) [see 79 and 89]. These data therefore suggest that the bax death gene may in fact be a target for p53 transactivation in the ovary, as has been reported for other cell types [43,45]. To date there exists only a single preliminary observation describing the relationship between c-myc expression and apoptosis in granulosa cells . These data indicate that the levels of c-myc are highest in populations of avian granulosa cells in follicles that still retain the capacity to undergo atresia . Although these findings are consistent with the proposed role of c-myc in apoptosis induction [41,42], the precise functions of c-myc in granulosa cells during atresia require further investigation.
Only recently has the role of cytoplasmic proteases, specifically members of the ICE gene family, in the demise of granulosa cells been assessed . Following isolation of cDNAs encoding rat ICE, Ich-1 and CPP32, it was determined that levels of Ich-1 and CPP32 mRNA in the ovary are reduced following in vivo gonadotropin stimulation . By comparison, extremely low levels of ICE mRNA are detectable in ovarian homogenates and expression of this protease is not altered by gonadotropin treatment. Along these same lines, ICE activity is undetectable in either healthy or atretic ovarian antral follicles of the rat, collectively indicating that ICE per se is probably not involved in the final life-and-death decision making process in granulosa cells . It should also be noted that one of the primary products of ICE activity, namely active interleukin-1ß (IL1ß), has been reported to slightly but non-significantly increase basal rates of apoptosis in rat follicles cultured in vitro. Moreover, this cytokine does not antagonize (nor potentiate) tropic hormone-supported survival of granulosa cells in follicles in vitro , further substantiating that ICE itself is not likely involved in the survival or death of granulosa cells.
Although another recent report has suggested that IL1ß prevents apoptosis in granulosa cells of cultured rat follicles , interpretation of these findings is difficult for several reasons. First, the ICE/IL1ß pathway is primarily linked to the induction of cell death . Moreover, the source of ICE activity required for generation of IL1ß from the precursor proenzyme is unknown as ICE activity is below detectable limits in rat follicles at the stage of development used for those studies . Lastly, previous work documents a cytotoxic effect of IL1ß in rat ovarian cell dispersates , and these data are similar to the slight but non-significant increase in apoptosis induced by IL1ß in rat ovarian follicles reported in a separate study . In any case, enhanced intrafollicular activity of other members of the ICE protease family (e.g. ICH-1, CPP32) may be responsible for such endpoints as endonuclease activation  and disrupted messenger RNA processing  in granulosa cells during the initiation and progression of apoptosis.
Lastly, the development of a CL following ovulation of a follicle is required for the maintenance of pregnancy [95,96]. The CL functions primarily as a site for the massive synthesis of progesterone which is required for the maintenance of the uterine endometrial lining. If pregnancy does not occur, the CL predictably regresses at a specific point in the estrous or menstrual cycle [95,96], and the ensuing loss of steroid hormone support to the uterus leads to apoptosis in endometrial cells . Although several studies concerning cell death during luteal regression have been published, the precise role of apoptosis in the fate of the CL remains unclear. Regression of the CL occurs in two phases, the first of which is associated with the loss of progesterone synthesizing capacity. This process, known as functional regression, occurs prior to any discernible morphological changes in luteal cell integrity and is likely a reversible step if sufficient luteotropic support is provided [95,96]. In contrast, the second phase of luteolysis, termed structural regression, is probably not reversible. Structural regression of the CL occurs well after the initial decline in steroid output, and is most likely the point at which apoptosis comes into play [98-101]. Data to support this contention are derived from analysis of both natural and induced luteolysis in many species, with the common link being the identification of apoptosis in structurally regressing luteal tissue [98-101].
To date, two key hormones appear to directly regulate the process of cell death during luteal regression: the luteolysin prostaglandin F2alpha (PGF2alpha), and the luteotropin chorionic gonadotropin (CG). Consistent with the themes presented throughout this review, ample evidence exists which link the actions of both PGF2alpha and CG to the regulation of reactive oxygen species in the CL. For example, the immediate cellular response to PGF2alpha following interaction of the hormone with its receptor in luteal cells is believed to be the generation of free radical species . Luteal cells exposed to an environment rich in reactive oxygen species in vitro exhibit a markedly impaired ability to synthesize progesterone [103; reviewed in 104, 105], one of the hallmark features of functional luteolysis in vivo. Furthermore, prolonged oxidative stress has been suggested as a trigger for apoptosis during structural luteolysis, a hypothesis supported by several lines of investigation . On the other hand, treatment of rats with human CG has been reported to increase the expression of anti-oxidant enzymes in the CL , whereas a lack of CG support in the human CL may predispose the luteal cells to increased oxidative stress . These data, taken with the fact that human CG can directly suppress the occurrence of apoptosis in luteal tissue , strongly support that changes in the redox state of luteal cells may be one of the primary determinants of the lifespan of the CL [104,105]. Although activation of other intracellular effector pathways, such as those involving changes in bcl-2 and c-myc expression, have been proposed as potential mechanisms underlying luteolysis [109-111], additional studies are required to clarify the role of these signaling events in determining the fate of luteal cells.