Regulation of apoptosis in the female reproductive system
Prev	Chapter 2. Review of the literature	Next

2.3. Ovarian function

2.3.1. Ovarian development

It is thought that in the human embryo, primordial germ cells originate from embryonic ectoderm (Buehr 1997). They migrate from their original location and during the fourth week of development, germ cells are visible in the wall of the yolk sac. Shortly thereafter, germ cells migrate one final time to the developing gonad where they arrive during the fifth or sixth week after fertilization (Sadler 1990). The germ cells express a tyrosine kinase receptor, c-kit, on their surface (Manova & Bachvarova 1991). Its ligand, stem cell factor or kit ligand, is expressed by the cells along the germ cell migratory pathway (Matsui et al. 1990). The interaction between c-kit and its ligand seems to be fundamental for germ cell migration. Today it is known that in mice c-kit and kit ligand are encoded by white spotting and Steel loci respectively. White spotting or Steel mutant mice have ovaries with very few, if any, germ cells (Coulombre & Russel 1954, Bennet 1956, Mintz & Russel 1957). In addition, other genetic factors are known to be essential for gonadal development (Table 1). Wilms’ tumour gene, WT-1, is essential for ovarian development. Without WT-1 action, germ cells migrate normally, but the urogenital ridge fails to develop, leading to gonadal and kidney agenesis (Kreidberg et al. 1993). It is well known that the male differentiation pathway is triggered by Y chromosome-encoded testis-determining factor (SRY) (Gubbay et al. 1990, Koopman et al. 1991). Recent observations demonstrate that the female developmental pathway requires active participation of the Wnt-4 signalling pathway. In wnt-4 deficient mice the Müllerian ducts are absent and the females are masculinized (Vainio et al. 1999).

The human ovary begins to differentiate morphologically during the 11–12th weeks of development. Before this time the germ cells have already started increasing their number through mitotic cell divisions (reviewed by Peters 1970, Hirshfield 1991). Already at this age, large numbers of germ cells are culled from the developing ovary. This process is termed germ cell attrition and it is probable that this process continues in the ovary until menopause (reviewed by Kaipia & Hsueh 1997, Morita & Tilly 1999). Evidence from rodent ovaries suggests that this degeneration in the fetal gonad occurs via apoptosis (Coucouvanis et al. 1993). This phenomenon seems to be universal, since all vertebrate species that have been examined to date are born with much fever oocytes than their maximum number during early development (Beaumont & Mandl 1961, Borum 1961, Baker 1963, Forabosco et al. 1991).

Table 1. Mouse models of ovarian failure.

Transgenic/mutant mouse	Ovarian phenotype	Reference
c-kit deficiency	Loss of germ cells due to defect in migration and proliferation	(Coulombre & Russel 1954, Mintz & Russel 1957)
Kit ligand deficiency	Loss of germ cells due to defect in migration and proliferation	(Bennet 1956)
Zfx knockout	Loss of germ cells due to defect in proliferation	(Luoh et al. 1997)
Atm knockout	Loss of germ cells	(Barlow et al. 1996)
Dazla knockout	Loss of germ cells	(Ruggiu et al. 1997)
WT-1 knockout	Failure of gonadal development	(Kreidberg et al. 1993)
GDF-9 knockout	Arrest of folliculogenesis at primary stage	(Dong et al. 1996)
IGF-I knockout	Arrest of folliculogenesis before antral follicle stage	(Baker et al. 1996)
FSH-β subunit knockout	Arrest of folliculogenesis at preantral stage	(Kumar et al. 1997)
ER_ knockout	Infertility due to failure to ovulate	(Lubahn et al. 1993)
Wnt-4 knockout	Lowered germ cell number, masuclinization	(Vainio et al. 1999)

After the 10th week, epithelial precursors of human granulosa cells begin to form the very first follicular structures. Initially, the pre-granulosa cells arrange themselves in structures that encircle several oogonia. This loose net starts to tighten and granulosa cells envelop most oocytes by the 24th week and all at birth. The number of oogonia increases steadily, until by means of an unknown signal they begin to enter meiosis (Sadler 1990). The initiation of meiosis denotes the transformation of the oogonia into an oocyte. This process starts at around week 15 and the number of oocytes reaches its maximum of 7 million around the 20th week (Baker 1971). The meiosis of oocytes is arrested at prophase of the first meiotic division. Meiotic division is completed just prior to ovulation, and in humans this arrest can be 50 years long (Mira 1998). When entering meiosis oocytes lose their ability to increase their numbers through mitosis. This inclines the balance in favour of germ cell attrition and from this moment onwards, the number of oocytes begins to inevitably decline (Baker 1963). Unlike spermatogenesis in males, where the spermatogonia act as stem cells and constantly divide to produce gametes, the ovary has a finite supply of oocytes that is determined at birth (reviewed in Morita & Tilly 1999).

At birth, all oocytes are surrounded by a single layer of flat granulosa cells, which together with the oocyte constitute the primordial follicle. The number of primordial follicles has been estimated to be anywhere from 266,000 to 2,000,000 at the time of birth (Block 1953, Baker 1971, Forabosco et al. 1991, Gougeon et al. 1994). These follicles comprise the pool of resting follicles that is the basic factor in determining postnatal ovarian life span (Gougeon 1996).

Figure 8. Germ cell attrition and follicular atresia according to Kaipia & Hsueh 1997. Germ cells migrate to the ovary during early embryonic development. Their number increases through mitotic divisions but most of the oocytes formed during development do not survive to the time of birth.

2.3.1.1. Postnatal trends

After birth the attrition of oocytes continues, although the process is attenuated. Throughout childhood the ovary remains endocrinologically inactive, but little by little the number of primordial follicles declines. It has been estimated that at the onset of puberty there are at maximum 200,000 resting follicles left in the human ovary (Baker 1963, Faddy et al. 1992). In the fertile ovary, a resting follicle has two future options, it can either undergo atresia or it can be recruited into the pool of growing follicles. While atresia represents certain death, the fate of growing follicles is only slightly better. Ultimately, during every menstrual cycle only one follicle survives the process, reaches full maturation and ovulates, and only approximately 400 oocytes from the original 7 million will ovulate during the fertile life span (Gougeon 1996). In the end, when all the follicles have been depleted, fertile senescence, i.e. menopause, ensues (Fig. 8).

2.3.2. The adult ovary

Throughout the entire female reproductive lifespan, the ovary relies on the reserve of resting follicles. A small number of follicles continuously leave the resting follicle pool and start growing. The early stages of folliculogenesis proceed very slowly. While the exact time for the follicles to attain the preantral stage is unknown, it has been estimated that in humans the process can take over 300 days. One of the reasons for this is the long doubling time of granulosa cells (~250 hr) (Erickson 2000). After recruitment the follicle begins to increase in size, both by proliferation of granulosa cells and by growth of the oocyte. This process is at least partially guided by a gene for a factor termed growth differentiation factor 9 (GDF-9). GDF-9 is a member of the transforming growth factor β (TGFβ ) superfamily that includes TGFs, bone morphogenetic proteins (BMPs), MIS, activins and inhibin. GDF-9-deficient mice show a block in follicular development at the one layer follicle stage (Dong et al. 1996). The observation that GDF-9 is also expressed in human oocytes during early folliculogenesis suggests that the gene also plays a major role in the human ovary (Aaltonen et al. 1999). It has also been proposed that proper interaction between c-kit and kit ligand is required for this process.

Figure 9. Folliculogenesis and classification of growing follicles in the human ovary according to Gougeon (1996). Growing follicles enter class 2 usually in the late luteal phase, class 3 between late luteal and early follicular phases, class 4 during late follicular phase and become recruitable class 5 follicles during late luteal phase.

When the follicles reach the size where they contain three to six layers of granulosa cells, the surrounding connective tissue stratifies and differentiates into two parts. The outer part, the theca externa, is basically similar to the stromal tissue surrounding it. In the inner part, the theca interna, stromal precursor cells differentiate into epithelioid cells. At this stage, the follicle is defined as a preantral follicle. Already at this stage, a considerable proportion of growing follicles fail to survive (Fig. 9), and they degenerate through a process termed follicular atresia. Observations in humans and in animals suggest that apoptosis is the mechanism of follicular atresia (Kaipia & Hsueh 1997).

After the primary follicle stage, gonadotrophins especially FSH, are increasingly important in sustaining follicular growth (reviewed by Hirshfield 1991, Zeleznik 2001, Adashi 1994).

2.3.2.1. Hormonal regulation of follicular growth

After the initial stages, follicular growth and development are brought about by the combined action of FSH and LH on the follicular cells. FSH and LH bind to their specific receptors on the surface of granulosa and theca interstitial cells, respectively. The activation of FSHR and LHR stimulates mitosis and differentiation responses in granulosa and theca interna cells (Gougeon 1996). In addition, gonadotrophins have two major endocrine effects. The first is that FSH and LH action stimulate the production of estradiol specifically in the dominant follicle (Erickson 2000). According to the two-cell-two-gonadotrophin theory, both gonadotrophins and both granulosa and theca interna cells have a specific task in this process (Fig. 10). Another endocrine response is the marked increase in production of inhibin by FSH (Groome et al. 1994).

Figure 10. Two-cell-two-gonadotrophin theory. LH induces androstenedione synthesis in theca cells. Driven by the FSH stimulus, granulosa cells process androstenedione into estrone which is further converted into estradiol by type 1 17HSD enzyme.

After the preantral stage, follicular fluid begins to accumulate in the follicle, expanding it relatively rapidly (Fig. 9). The dominant follicle is selected from a cohort of stage 5 follicles (Fig. 9) and eventually, the whole hormone-dependent stage of follicular growth can take approximately 40–50 days to complete (Gougeon 1996). Only one follicle is selected at a time, and the fate of all remaining growing follicles is atresia. The first evidence that apoptosis is responsible for this process was gathered from work on rodents (Tilly et al. 1991, Tilly et al. 1992). Similarly, it was discovered that the process is hormonally controlled. Estrogens and gonadotrophins inhibit and androgens induce ovarian cell death (Chun et al. 1994, Billig et al. 1993, Yuan & Giudice 1997).

It has been estimated that up to 30 years of age, the loss of follicles (and oocytes) in the human ovary is mainly due to atresia of resting follicles. After reaching a critical threshold in the number of follicles (estimated at 25,000), the recruitment of growing follicles is increased twofold and thereafter the loss of follicles is mainly due to atresia of growing follicles (Gougeon 1996, Erickson 2000).

2.3.3. Regulation of ovarian apoptosis

Apoptosis has been found to occur in three different ovarian cell types: oocytes, granulosa cells and luteal cells. Animal studies have indicated that oocyte apoptosis is most profound during embryonal development (reviewed in Morita & Tilly 1999). The peak number of oocytes in mouse ovaries is reached on day 13.5 (E13.5) of gestation, when some of the oocytes have already abandoned the mitotic cell cycle and entered the first phases of meiotic divisions (Byskov 1986). Apoptosis has been confirmed to be the mechanism of germ cell atresia in animals (Pesce & De Felici 1994, Ratts et al. 1995, Morita et al. 1999). However, no reports confirming these results in humans have been published. Furthermore, while oocyte apoptosis has clearly been established to affect neonatal ovarian development in several species, it is not clear what physiological role oocyte apoptosis plays in adult life. In vitro and in vivo experiments have shown that DNA damaging agents, such as cyclic polyaromatic hydrocarbons (PAHs), and cytostatics induce oocyte apoptosis in adult mice (Perez et al. 1997a, Matikainen et al. 2001a). Based on these results, it is clear that oocytes of the adult ovary have retained their ability to undergo apoptosis under a pathological insult.

Transgenic and knockout animal models have shed some light on the issue of how oocyte cell death is regulated (Table 2). Bcl-2 knockout mice have been found to have significantly fewer oocytes than similar age-matched wild type animals (Ratts et al. 1995). Bcl-2 can also be detected in E12.5 and E15.5 wild type mouse oocytes (Felici et al. 1999). Another anti-apoptotic Bcl-2 family member, Bcl-X_L, also seems to play a role in determining germ cell fate. Although Bcl-X_L knockout mice die at a very early stage of fetal development, there is a clear loss of germ cells at E11.5 compared with wild type animals, and this loss can be prevented in vitro by transfection with the Bcl-X_L gene (Watanabe et al. 1997). Its role in oocyte cell death is supported by the finding that the gene is expressed in murine oocytes (Jurisicova et al. 1998). Mice deficient in the pro-apoptotic Bcl-2 family member Bax showed a normal number of oocytes at birth, but at the onset of puberty, the follicle pool was threefold in size when compared with age-matched wild type mice (Perez et al. 1999).

During folliculogenesis, apoptosis has been mainly observed in the granulosa cells of the growing ovaries (Yuan & Giudice 1997). This process seems to be principally controlled by hormones. Testosterone induces and estradiol inhibits granulosa cell apoptosis in rodent ovaries (Chun et al. 1994). Gonadotrophins have also been shown to be survival factors for human granulosa cells in vitro (Chun et al. 1994, Jablonka-Shariff et al. 1996). Furthermore, the high androgen/estrogen ratio in the follicular fluid seems to predispose human granulosa cells to cell death (Yuan & Giudice 1997). Granulosa cell apoptosis is mediated by the Bcl-2 family of cell death-regulating factors (Tilly et al. 1995a, reviewed in Hsu & Hsueh 2000). Apaf-1 is also expressed in granulosa cells and it is activated by Bcl-2 family-mediated release of cytochrome c (Robles et al. 1999). In addition, the tumour suppressor genes TP53 and WT1 might have a role in regulating granulosa cell apoptosis in rodents (Tilly et al. 1995b) and in humans (Makrigiannakis et al. 2000).

There is accumulating evidence that apoptosis plays a part in CL regression in animals (Bruce et al. 2001, Zheng et al. 1994) and in humans (Shikone et al. 1996, Sugino et al. 2000). Recent observations also reflect the expression of apoptosis-regulating members of the Bcl-2 family, Bcl-2, Bcl-X and Bax, in the human CL (Rodger et al. 1995, Rodger et al. 1998, Sugino et al. 2000). Furthermore, experiments with cultured granulosa luteal cells suggest that caspases 3 and 9 are activated in staurosporin-induced apoptosis in luteinized human granulosa cells (Khan et al. 2000).

Table 2. Transgenic and knockout animal models for abnormal ovarian cell death.

Transgenic/mutant mouse	Ovarian phenotype	Reference
Bcl-2 knockout	Loss of oocytes	(Ratts et al. 1995)
Bcl-2 transgene	Loss of ovarian apoptosis, enhanced folliculogenesis, teratomas	(Hsu et al. 1996)
Bcl-X_L knockout	Loss of germ cells	(Watanabe et al. 1997)
Bax knockout	Normal at birth, increased number of oocytes at puberty and during fertile life	(Perez et al. 1999)
Caspase-2 knockout	Increased number of primary follicles	(Bergeron et al. 1998)
Caspase-3 knockout	Normal number of primary follicles	(Matikainen et al. 2001b)

Prev	Home	Next
Common mechanisms of apoptosis	Up	Endometrial function