Human Reproduction Update Advance Access originally published online on May 25, 2006
Human Reproduction Update 2006 12(5):537-555; doi:10.1093/humupd/dml022
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Ovarian follicle development and transgenic mouse models
Department of Epidemiology and Preventive Medicine, University of Maryland School of Medicine, Baltimore, MD, USA
1 To whom correspondence should be addressed at: Department of Epidemiology and Preventive Medicine, University of Maryland School of Medicine, 660 West Redwood Street, Howard Hall, Room 133, Baltimore, MD 21201, USA. E-mail: jflaws{at}epi.umaryland.edu
 |
Abstract |
|---|
TOP
Abstract
IntroductionOvarian follicle development is a complex process that begins with the establishment of what is thought to be a finite pool of primordial follicles and culminates in either the atretic degradation of the follicle or the release of a mature oocyte for fertilization. This review highlights the many advances made in understanding these events using transgenic mouse models. Specifically, this review describes the ovarian phenotypes of mice with genetic mutations that affect ovarian differentiation, primordial follicle formation, follicular growth, atresia, ovulation and corpus luteum (CL) formation. In addition, this review describes the phenotypes of mice with mutations in a variety of genes, which affect the hormones that regulate folliculogenesis. Because studies using transgenic animals have revealed a variety of reproductive abnormalities that resemble many reproductive disorders in women, it is likely that studies using transgenic mouse models will impact our understanding of ovarian function and fertility in women.
Key words: follicle development / ovary / transgenic mouse models
|
Introduction |
|---|
The ovary is a primary functional organ of the female reproductive system, and it plays two major physiological roles. First, the ovary is responsible for the differentiation and release of a mature oocyte for fertilization (McGee and Hsueh, 2000
). Second, it is responsible for synthesizing and secreting hormones that are essential for follicle development, menstrual/estrous cyclicity and maintenance of the reproductive tract and its function (Hirshfield, 1991). During the past several years, studies using transgenic mouse models have revealed novel information about the genes that control the processes of ovarian development and function. Some of these genes affect the ovary through direct mechanisms, whereas others do so through indirect mechanisms. Therefore, the purpose of this review is to highlight the advances made in understanding these events through the use of transgenic mouse models. Specifically, this review describes the ovarian phenotypes of mice with genetic mutations that directly affect ovarian differentiation, primordial follicle formation, follicular growth, atresia, ovulation and corpus luteum (CL) formation. In addition, this review describes the phenotypes of mice with mutations in a variety of genes that indirectly affect the ovary because they affect the hormones or growth factors that regulate folliculogenesis. |
Mouse models with alterations in ovarian differentiation |
|---|
Differentiation of a functional ovary from a bipotential gonad is a complex process (Tevosian et al., 2002
). The bipotential gonad arises from the coelomic epithelium of the urogenital ridges and initially is indistinguishable in male and female mouse embryos (Cui et al., 2004). During embryonic life, the bipotential gonad undergoes a series of changes, which results in the formation of the ovary. Although relatively little information is known about the factors absolutely required for the differentiation of the ovary, some studies indicate that Pod1 (also called capsulin/epicardin/Tcf21) is an important regulator of gonadal development (Cui et al., 2004). Specifically, deletion of Pod1 disrupts proper differentiation of the ovary, because Pod1-deficient (Pod1–/–) ovarian germ cells resemble germ cells in the testes (Cui et al., 2004). This may be, in part, because of a failure to appropriately down-regulate steroidogenic factor 1 (Sf1) (Hanley et al., 2000; Cui et al., 2004). Sf1 expression in the gonads has been shown to be essential for reproductive function, although wild type (WT) and Sf1–/– ovaries are indistinguishable during embryogenesis and at birth (Jeyasuria et al., 2004).
Forkhead transcription factor 2 (Foxl2) is also an important regulator of ovarian differentiation because Foxl2–/– ovaries express markers of testis determination shortly after birth (Ottolenghi et al., 2005). Other genes, such as Dax1, Wnt4 and follistatin (Fst), are important, but not absolutely required, for ovarian differentiation because their deletion does not result in complete female-to-male sex reversal (Swain et al., 1996; Vainio et al., 1999; Uda et al., 2004; Yao et al., 2004). Instead, deletion of Fst and Wnt4 leads to the appearance of a testis-specific coelomic vessel (Jeays-Ward et al., 2003; Yao et al., 2004). Taken together, these studies suggest that Fst and Wnt4 must be expressed within the ovary to inhibit the testis differentiation pathway and promote the formation of the ovary. To date, little is known about whether WNT and FST play a role in ovarian differentiation in humans. A mutation in WNT4 has been shown in a female patient whose phenotype resembles Wnt4–/– mice (Baison-Lauber et al., 2004). Furthermore, abnormal FST levels have been linked to polycystic ovary syndrome (PCOS), a leading cause of infertility in women (Mason, 2000).
|
Mouse models with alterations in the formation of primordial germ cells |
|---|
A founding population of primordial germ cells (PGCs) is allocated during gastrulation outside of the developing gonad, to which PGCs migrate by embryonic day (ED) 11 (Hirshfield, 1991
; McClellan et al., 2003). The PGCs that reach the gonad, now referred to as oogonia, lose their motile characteristics and undergo extensive mitotic proliferation and apoptosis (Hirshfield, 1991). Many factors are involved in the allocation and expansion of the germ cell population. Recent studies using transgenic mouse models provide evidence that various bone morphogenetic proteins (BMPs) are involved in the allocation of the founder population of PGCs. For example, studies using Bmp4–/– embryos indicate that BMP4 is required for PGC allocation, as Bmp4–/– embryos are devoid of PGCs (Lawson et al., 1999; Ying and Zhao, 2001). Similarly, studies using Bmp8b–/– embryos demonstrate its importance for PGC allocation, although its absolute requirement depends on the genetic strain used (Ying et al., 2000).
BMPs bind to the type 1 activin receptor-like kinase (ALK) receptors ALK2, ALK3 or ALK6 to regulate gene expression (Chen et al., 2004). Alk2–/– embryos have no detectable PGCs, suggesting that a factor induced by BMP4 signalling through Alk2 is essential for the formation of the PGC founder population (Chuva de Sousa Lopes et al., 2004). BMP signalling is mediated, in part, by SMAD5. Interestingly, Smad5–/– embryos have a significant reduction in the size of the founder population of PGCs (Chang and Matzuk, 2001). This fact, together with similar spatiotemporal expression patterns in BMPs and SMAD5, suggests that SMAD5 may be a signalling intermediate for BMP4 and BMP8b during PGC allocation (Chang and Matzuk, 2001; Chuva de Sousa Lopes et al., 2004).
Interactions between PGCs and the extracellular matrix (ECM) are thought to be important for PGC migration (Garcia-Castro et al., 1997). ECM-interacting integrins are expressed by PGCs and are necessary for colonization of the genital ridges, because nearly 70% of PGCs lacking the integrin ß1 subunit are unable to enter the gonad (Anderson et al., 1999). Gap junction communication, either between PGCs themselves or between PGCs and somatic cells, may also be important for colonization of the genital ridges or PGC survival, as connexin (Cx)43-deficient (Gja1–/–) mice have a reduced number of PGCs by ED 11.5 (Juneja et al., 1999).
Stromal cell-derived factor 1 (Sdf1) and its receptor Cxcr4 have also been shown to be important for PGC migration (Ara et al., 2003; Molyneaux et al., 2003b). In both Sdf1–/– and Cxcr4–/– embryos, there is a large reduction in the number of PGCs that enter the gonad, with many PGCs remaining in the hindgut endoderm or mesentery (Ara et al., 2003; Molyneaux et al., 2003b). Naturally occurring mutations in the Steel (Sl) and Dominant white spotting (W) loci, which encode kit ligand KL and c-kit, respectively, lead to germ cell deficiency and improper PGC migration (Buehr et al., 1993). Specifically, studies have shown that some mutations in KLSL lead to a migratory defect, resulting in an absence of PGCs in mutant genital ridges at ED 11.5 in a dose-dependent manner (Zama et al., 2005).
Many factors are involved in PGC survival, including zinc-finger transcription factor (Zfx) and Oct4. Deletion of Zfx results in a 50% reduction in the number of PGCs at ED 11.5 and results in early ovarian failure (Luoh et al., 1997). Similarly, germ cell-specific deletion of Oct4 leads to massive PGC apoptosis around ED 10 and an absence of oocytes in post-natal ovaries (Kehler et al., 2004). T-cell intracellular antigen-1-related (TIAR) protein has been shown to promote the survival and the proliferative activity of PGCs. Specifically, Tiar–/– mice contain substantially fewer PGCs at ED 11.5 compared with WT mice and completely lack oogonia at ED 13.5 (Beck et al., 1998).
A factor known as proliferation of germ (pog) cells is important for maintaining germ cell proliferation (Agoulnik et al., 2002). Specifically, loss of pog cells is responsible for the germ cell-deficient (gcd) phenotype, in which animals are born with a dramatically reduced germ cell population. Therefore, females are infertile because of an absence of germ cells in adults (Agoulnik et al., 2002). This phenotype is also seen in patients with Fanconi anaemia (FANCL), in which patients have an autosomal recessive disorder characterized by cellular hypersensitivity to DNA cross-linking agents. As in the human disease, Fanca–/– mice show hypogonadism (Cheng et al., 2000; Wong et al., 2003). Furthermore, at ED 8.5, Fanca–/– and WT embryos have similar numbers of PGCs, but by ED 11.5, Fanca–/– mice have 50% fewer germ cells than WT embryos (Wong et al., 2003), leading to early ovarian failure due to early depletion of oocytes (Koomen et al., 2002).
Homologues of the Drosophila RNA-binding protein, NANOS, are also thought to be regulators of PGC proliferation (Haraguchi et al., 2003; Tsuda et al., 2003). Whereas NANOS1 and NANOS2 are dispensable for female fertility, NANOS3 is essential, because Nanos3–/– females are infertile (Haraguchi et al., 2003; Tsuda et al., 2003). Normal numbers of PGCs are allocated and seen migrating at ED 9.5 in Nanos3–/– embryos, but there are no PGCs remaining in mutant gonads by ED 12.5, suggesting that Nanos3 is important for PGC maintenance or proliferation after ED 9.5 (Tsuda et al., 2003).
Although PGC proliferation appears to be maintained by extrinsic regulation, the rate of proliferation appears to be regulated by the PGCs themselves (De Felici et al., 2004). Studies using the Pin1–/– model provide evidence for this. PIN1 is a peptidyl-prolyl isomerase that affects protein stability and/or function and is also involved in cell-cycle progression (Joseph et al., 2003). Furthermore, PGCs express PIN1 through migration at least until ED 13.5, and whereas a normal number of PGCs are allocated in Pin1–/– embryos, their population expands very slowly compared with WT embryos (Atchison et al., 2003). The slow rate of expansion in the PGC population of Pin1–/– embryos appears to be because of inefficient cell-cycle progression leading to a lengthening of the cell cycle (Atchison et al., 2003).
Members of the B-cell lymphoma/leukaemia 2 (Bcl2) family have been shown to regulate apoptosis of meiotic germ cells (Morita et al., 1999). Targeted overexpression of Bcl2, an anti-apoptotic member of the Bcl2 family, in germ cells enhances germ cell survival so that females are born with a surfeit of follicles (Flaws et al., 2001). Bclxl is another anti-apoptotic member of the Bcl2 family. Studies indicate that this factor is an important regulator of oocyte apoptosis in the embryo because reduced Bclxl function results in a shortage of follicles at birth due to decreased oocyte survival beginning at ED 13.5 (Rucker et al., 2000). Bax, a pro-apoptotic family member, has been reported not to affect follicle endowment (Perez et al., 1999), which is surprising given that Bax deletion rescues the Bclxl reduced function phenotype (Rucker et al., 2000).
Caspases (Casps) have also been shown to regulate apoptosis in oocytes. Specifically, deletion of Casp11 (also known as Casp4) significantly reduces the number of oocytes present at birth (Morita et al., 2001). Also, studies demonstrate that oocyte loss due to cytokine insufficiency in Casp11–/– mice is dependent on the action of Casp2, because co-deletion of Casp2 restores oocyte numbers to normal (Morita et al., 2001).
|
Mouse models with alterations in primordial follicle formation |
|---|
Oogonia cease dividing at ED 13.5 and enter meiosis to form oocytes. These oocytes are closely associated in clusters called germ cell nests (Gomperts et al., 1994
; Pepling and Spradling, 1998). Germ cell nests break down shortly after birth to allow the formation of primordial follicles, a process involving apoptosis of some of the oocytes within the nests. The surviving oocytes become enveloped by somatic cells surrounding the nest, thereby producing primordial follicles (Figure 1) (Pepling and Spradling, 2001).
|
The primordial follicle population present at birth has long been believed to be finite (Zuckerman, 1951), although this notion has recently been challenged (Johnson et al., 2004, 2005). Establishment of this pool of primordial follicles marks the first stage of follicle development (growth of follicles from the primordial to pre-ovulatory stage), as defined by Hirshfield (1991). The end of reproductive life or ovarian senescence occurs when this pool of primordial follicles is depleted by death or through growth, followed by subsequent ovulation and/or atresia (Hirshfield, 1991). Although some primordial follicles will be stimulated to grow immediately, the majority will remain dormant, perhaps because of inhibitory stimuli, until they receive signals to enter the growing pool (McGee and Hsueh, 2000).
Although the processes of germ cell nest breakdown and primordial follicle formation are not well understood, several genetic models have identified factors important for their regulation. The Dazla gene, in particular, has been shown to be essential for the differentiation of germ cells because disruption leads to reduction in germ cell numbers embryonically and complete absence of follicles and ova in the adult ovary (Ruggiu et al., 1997; McNeilly et al., 2000). An oocyte-specific gene, factor in the germline 
(Figl), is also essential for follicle formation as Figl–/– females are infertile despite the presence of normal numbers of meiotically competent germ cells in embryonic ovaries (Soyal et al., 2000). By post-natal day (PD) 1, Figl–/– oocytes begin to disappear and those that are present are not surrounded by somatic cells, suggesting that Figl is important for regulating interactions between oocytes and granulosa cells (Soyal et al., 2000). Interestingly, Figl has been linked to premature ovarian failure (POF) in women (Pangas and Rajkovic, 2006).
Mice lacking nerve growth factor (NGF; Ngf–/–) also appear to have a deficiency in germ cell nest breakdown, consequently reducing the number of growing follicles in the ovary (Dissen et al., 2001). Tyrosine kinase (Trk) receptors are receptors for NGF and may also play a role in follicle formation (Donovan et al., 1996). Specifically, studies have shown that deletion of TrkB reduces the number of follicles present in neonatal mice because of decreased oocyte survival during the period of follicle formation (Donovan et al., 1996; Spears et al., 2003). There is some evidence that NGF is involved in the pathogenesis of PCOS in women (Bai et al., 2004).
Studies have shown that proteins such as SPO11 [sporulation protein homology (S. cerevisiae)], DMC1 [disrupted meiotic cDNA 1 homologue (human)] and MSH5 [mutS homologue 5 (Escherichia coli)] are also important for the formation of primordial follicles (Di Giacomo et al., 2005). SPO11 is an enzyme required for the introduction of double-strand breaks during meiosis. Spo11–/– ovaries have a reduced number of primordial follicles compared with WT ovaries due to oocyte depletion at or before follicle formation, which is independent of DNA damage arising from unrepaired recombination intermediates (Di Giacomo et al., 2005).
DMC1 is a DNA strand exchange protein that acts on double-strand breaks to catalyze strand invasion into intact homologous products, giving rise to mature recombinant products. Dmc1–/– mice are viable but infertile because of defects in chromosome synapsis (Bannister and Schimenti, 2004). Furthermore, Dmc1–/– ovaries are devoid of follicles, due to the elimination of oocytes before follicle formation (Di Giacomo et al., 2005). Spo11–/–Dmc1–/– double-mutant mice mirror the Spo11–/– phenotype. In these animals, the introduction of double-strand breaks is eliminated, indicating that oocytes die in Dmc1–/– ovaries more rapidly because of a DNA damage-dependent mechanism (Di Giacomo et al., 2005).
MSH5 is another meiosis-specific protein involved in recombination. Msh5–/– mice are sterile as a result of severe gametogenic failure (Di Giacomo et al., 2005). As in Dmc1–/– mice, deletion of Msh5 results in reduced oocyte numbers due to a DNA damage-dependent mechanism. Spo11–/–Msh5–/– double knockouts are also identical to Spo11–/– single mutants (Di Giacomo et al., 2005).
In addition to the above proteins, the atm gene [ataxia telangiectasia-mutated homologue (humans)] is involved in DNA damage checkpoint control and is activated in response to double-strand breaks (Yamada and Coffman, 2005). Atm–/– females are infertile because of abnormal chromosome synapsis and fragmentation (Barlow et al., 1996; Xu et al., 1996; Barlow et al., 1998). There is increased apoptosis of oocytes in embryonic atm–/– ovaries compared with WT ovaries, and shortly after birth atm–/– ovaries are devoid of follicles. Oocyte loss in atm–/– ovaries may be because of unrepaired double-strand breaks introduced by SPO11 as Spo11–/–atm–/– double mutants are also indistinguishable from Spo11-deficient mice (Di Giacomo et al., 2005).
Collectively, transgenic models with mutations in meiotic genes provide evidence that several meiotic genes play an essential role in meiotic chromosome synapsis and gametogenesis (Mandon-Pepin et al., 2002). Their phenotypes are supported clinically, because mutations in DMC1 have been found in patients with POF (Mandon-Pepin et al., 2002). Furthermore, many of the mouse models discussed in this section may shed light on the recent claims that oogenesis continues in the post-natal ovary. Specifically, the deletion of Dazla, Bclxl and Zfx refute these claims as these knockout mice all experience POF, apparently due to prenatal germ cell depletion.
|
Mouse models with alterations in primary follicle development |
|---|
Once in the growing pool, the primordial follicle enlarges because of an increase in size of the oocyte and conversion of the squamous granulosa cells into cuboidal granulosa cells. At this time, the follicle is known as a primary follicle (Figure 2) (Hirshfield, 1991
). One striking characteristic of the primary follicle is the presence of a zona pellucida, which surrounds the oocyte and is maintained throughout growth until the oocyte is ovulated (Rankin et al., 1996, 1999, 2001). The extracellular zona pellucida matrix is composed of three glycoproteins (ZP1, ZP2 and ZP3) (Rankin et al., 1999). These proteins have been shown to affect folliculogenesis in that Zp1–/– mice have loosely organized zonae pellucidae and reduced litter sizes compared with WT mice (Rankin et al., 1999, 2001). Zp2–/– mice have a thin zona pellucida surrounding oocytes and are sterile (Rankin et al., 2001). Zp3–/– mice do not form a zona pellucida and are sterile (Rankin et al., 1996). Interestingly, genetic variations in ZPs have also been associated with fertilization failure in patients undergoing IVF (Mannikko et al., 2005).
|
An oocyte-specific factor, NOBOX, is important for primary follicle formation as Nobox–/– mice have accelerated post-natal oocyte loss, and their follicles do not transition from the primordial stage into the growing follicle pool (Rajkovic et al., 2004). Foxl2 has also been shown to play a role in folliculogenesis at the primary follicle stage. In Foxl2–/– ovaries, granulosa cell differentiation is blocked at the squamous to the cuboidal transition; thus, no primary follicles are formed (Schmidt et al., 2003). Foxo3a has also been identified as being a major activator of follicular development as Foxo3a–/– ovaries have a tremendous increase in the number of early growing follicles with enlarged oocytes at PD 14 (Castrillon et al., 2003). Furthermore, there appears to be an apparent lack of co-ordination between oocytes and granulosa cells in these follicles because oocyte enlargement is not accompanied by a transition of granulosa cells from flattened to cuboidal shape (Castrillon et al., 2003). Thus, the early activation of primordial follicles leads to premature exhaustion of the follicle pool (Castrillon et al., 2003). To our knowledge, there is no clinical evidence that genetic alterations in Foxo3a cause POF in women. Mutations in Foxl2 and NOBOX, however, have been linked to POF (Gersak et al., 2004; Pangas and Rajkovic, 2006).
|
Mouse models with alterations in pre-antral follicle development |
|---|
The primary follicle transitions into a pre-antral follicle as the granulosa cells proliferate to form multiple layers (Figure 2). In addition, the pre-antral follicle acquires an outer layer of thecal cells. At the end of this stage, the pre-antral follicle has several layers of granulosa cells and has acquired an extensive network of gap junctions (Hirshfield, 1991
).
Gap junctions are intercellular membrane channels that allow nutrients, inorganic ions, second messengers and small metabolites to pass from cell to cell (Kidder and Mhawi, 2002). Each of these intercellular channels is composed of CXS, a family of proteins that connect adjacent cells (Kidder and Mhawi, 2002). Studies using mice lacking Cx43 (Gja1–/–) and Cx37 (Gja4–/–) provide evidence that gap junctions are important for follicular development to the pre-antral stage (Simon et al., 1997; Ackert et al., 2001). CX43, the most abundant CX in the ovary, is expressed in the granulosa cells from the start of folliculogenesis. In Gja1–/– mice, folliculogenesis is arrested at the primary stage and the follicles contain meiotically incompetent oocytes (Carabatsos et al., 2000; Ackert et al., 2001). Furthermore, gap junctions are extremely rare in Gja1–/– granulosa cells, although these cells are still able to form gap junctions with WT granulosa cells (Gittens et al., 2003, 2005). This suggests that CX43 is not the only CX expressed during the early stages of folliculogenesis (Gittens et al., 2003, 2005). Whereas Cx43 is required for the earliest stages of development (Juneja et al., 1999), Cx37 has also been shown to be important (Simon et al., 1997; Juneja, 2003). Cx37 is expressed mainly between the oocyte and surrounding granulosa cells (Simon et al., 1997). In Gja4–/– mice, folliculogenesis is arrested at the pre-antral stage and the oocytes are not meiotically competent (Carabatsos et al., 2000).
Glutathione (GSH), which plays a role in protecting cells from oxidative stress and cellular injury, may also be involved in regulating folliculogenesis to the pre-antral follicle stage (Kumar et al., 2000). 
-Glutamyl transpeptidase (GGT) is responsible for converting secreted GSH into cysteinyl-glycine and -glutamic acid (Kumar et al., 2000). Transgenic mice deficient in GGT are infertile because of a block in folliculogenesis at the pre-antral stage and an increased number of degenerating oocytes compared with WT mice (Kumar et al., 2000).
|
Mouse models with alterations in antral follicle development |
|---|
The next phase of folliculogenesis, the antral stage, is marked by the appearance of small fluid-filled spaces, which eventually form a single antral cavity (Figure 2). The follicular fluid that fills the antral cavity contains water, electrolytes, serum proteins and high concentrations of steroid hormones secreted by the granulosa cells (Hirshfield, 1991
). At the antral stage, most of the follicles will undergo atresia, whereas the remaining antral follicles will survive under the influence of FSH and grow to the pre-ovulatory stage (Hirshfield, 1991).
Atresia is an apoptotic process that is highly regulated by pro-apoptotic and anti-apoptotic factors. Members of the Bcl2 family are important for regulating the atretic degradation of antral follicles. Specifically, studies have shown that deletion of Bcl2 reduces healthy follicle numbers and increases abnormal follicle numbers compared with WT ovaries (Ratts et al., 1995). Furthermore, targeted overexpression of Bcl2 to granulosa cells of growing follicles results in reduced apoptosis of granulosa cells, more follicles that spontaneously ovulate when given exogenous pregnant mare’s serum gonadotrophin (PMSG) and larger litter sizes than those of WT mice (Hsu et al., 1996). In addition, deletion of Bax leads to the presence of numerous unusual atretic follicles in adults (Knudson et al., 1995) and reduces the number of atretic immature follicles (Perez et al., 1999). Casps also regulate follicular atresia because Casp3–/– follicles fail to be eliminated by apoptosis (Matikainen et al., 2001).
Other factors such as the aryl hydrocarbon receptor (Ahr) and superoxide dismutase (SOD) have been shown to regulate follicular growth to the antral stage. The Ahr is a ligand-activated transcription factor that functions in mediating the toxicity of various environmental contaminants (Pocar et al., 2005). Although the Ahr can be activated exogenously, the natural endogenous ligand is unknown (Pocar et al., 2005). Studies using Ahr–/– mice have provided insight into its physiological role in reproduction (Pocar et al., 2005). Specifically, Ahr–/– mice have been shown to have reduced fertility due to a decreased number of antral follicles (Benedict et al., 2000), which is due to slow follicular growth and not increased atresia (Benedict et al., 2003).
SOD plays a role in inactivating superoxide radicals during oxidative stress (Matzuk et al., 1998). Sod1–/– ovaries have normal primary and pre-antral follicles but have fewer large antral follicles than WT ovaries (Matzuk et al., 1998). Sod2–/– mice die within 3 weeks of birth because of oxidative mitochondrial injury (Matzuk et al., 1998). Interestingly, ovaries from Sod2-deficient (Sod2–/–) mice transplanted into WT mice show all stages of folliculogenesis and can give rise to offspring (Matzuk et al., 1998). Taken together, these data suggest that SOD1, but not SOD2, is essential for ovarian function (Matzuk et al., 1998). An additional protein, ZP2, may also play a role in regulating antral follicle growth as Zp2–/– ovaries have fewer antral follicles compared with WT ovaries (Rankin et al., 2001).
|
Mouse models with alterations in ovulation |
|---|
After the antral stage, follicles grow to the pre-ovulatory stage (Hirshfield, 1991
). The pre-ovulatory follicle is the only follicle type that is capable of releasing an oocyte for fertilization, upon stimulation by LH (McGee and Hsueh, 2000). After the oocyte is released for fertilization, the remaining theca and granulosa cells differentiate into a structure called the CL (Figure 2) (Hirshfield, 1991; Elvin and Matzuk, 1998; McGee and Hsueh, 2000).
During ovulation, the follicle’s basement membrane ruptures, and a mature oocyte is released for fertilization (Hirshfield, 1991). The remodelling of the follicle after rupture is characteristic of an inflammatory response, and thus, many inflammatory mediators are elevated during the ovulatory process (Bukulmez and Arici, 2000). For example, the free radical gas, nitric oxide (NO), is an inflammatory mediator that is involved in ovulation. NO is synthesized in cells by the enzyme NO synthase (NOS). To date, three isoforms encoding NOSs have been identified: endothelial (eNOS), neuronal (nNOS) and inducible (iNOS) (Moncada et al., 1991). In the mouse ovary, both eNOS and iNOS are present in the oocytes and thecal cells (Mitchell et al., 2004). Studies using eNos–/– mice have demonstrated that they have reduced fertility due to impaired ovulatory efficiency, abnormalities in meiotic maturation, increased oocyte apoptosis and altered estrous cyclicity compared with their WT littermates (Jablonka-Shariff and Olson, 1998; Drazen et al., 1999; Hefler and Gregg, 2002). Studies using iNos–/– mice, however, show that iNos does not alter ovulatory capacity, but it may play a role in fertilization (Yang et al., 2005). Conditional nNos–/– mice are infertile because of ovulation defects as well as altered hormonal regulation (Gyurko et al., 2002).
Macrophage-stimulating factor 1 receptor (Mstr1) or Ron is involved in the regulation of iNOS and the production of NO during inflammatory responses (Hess et al., 2003). Specifically, RON inhibition has been shown to increase NO activity (Chen et al., 1998; Waltz et al., 2001). Transgenic mice deficient in Ron are infertile because of abnormalities in ovulation (Waltz et al., 2001). The ovulation defect in these mice is due to elevated levels of NO caused by increased iNOS levels (Hess et al., 2003).
The cyclooxygenase (COX) pathway, which responsible for in prostaglandin (PG) synthesis, is very similar to the NO pathway in that both pathways are important regulators of inflammatory responses and both have constitutive and inducible isozymes (McGarry et al., 2005). Furthermore, both pathways are able to crosstalk with other pathways. Specifically, there is crosstalk between NO and PGs and/or iNOS and COX (Clancy et al., 2000). PGs are common mediators of many inflammatory responses, including ovulation. Female mice lacking PGE receptor 2 (Ep2) have defects in ovulation due to defects in cumulus–oocyte complex (COC) expansion (Hizaki et al., 1999).
The two COX isozymes responsible for the synthesis of PGs are COX1 and COX2. COX1 is the constitutive form, whereas COX2 is inducible by a variety of factors including growth factors, cytokines, mitogens and tumour promoters (Smith and Langenbach, 2001). In the ovary, LH/hCG induces COX2 in the mural granulosa cells (Matsumoto et al., 2001). Although Cox1–/– mice have normal fertility, except for some defects in parturition, Cox2–/– females are infertile and exhibit abnormalities in ovulation due to PG deficiency (Matsumoto et al., 2001). Furthermore, Cox2–/– and Ep2–/– mice show decreased expression of tumour necrosis factor-induced protein-6 (TNFIP6 or TSG6), which is a hyaluronan-binding protein involved in COC expansion (Ochsner et al., 2003). Interestingly, Tnfip6–/– mice are sterile because of the inability of the cumulus cells to assemble their hyaluronan-rich ECM (Fulop et al., 2003). Collectively, these results indicate that COX2 is a critical stimulator of PG signalling and that COX2-derived PGs interact with their cognate receptor, EP2, to promote ovulation (Matsumoto et al., 2001).
Inter--trypsin inhibitors (ITIs) have been shown to inhibit inflammatory responses by blocking the induction of pro-inflammatory cytokines (Suzuki et al., 2004). Bikunin, commonly referred to as urinary trypsin inhibitor (UTI), is a member of the ITI family (Sato et al., 2001). Bikunin-deficient mice are infertile because of severe ovulation defects (Sato et al., 2001). Granulocyte macrophage colony-stimulating factor (GM-CSF) is also a member of the cytokine family. Studies indicate that Gmcsf–/– mice have normal ovulation rates but decreased numbers of activated macrophages in the stromal and thecal cells during ovulation (Jasper et al., 2000). These results suggest that GM-CSF plays a role in regulating steroidogenesis in the theca by regulating local macrophage populations (Gilchrist et al., 2000).
The plasminogen activator (PA) system plays a role in the proteolytic degradation of the follicle wall at the time of ovulation (Ny et al., 1997) and in the formation/degradation of the CL (Liu et al., 2003). Studies of gonadotrophin-induced ovulation reveal that mice with single deficiencies of tPA, uPA or PAI-1 have normal ovulation, whereas tPA–/–uPA–/– double-mutant mice have impaired ovulation (Ny et al., 1997).
Oocyte maturation is an important event in ovulation (Jamnongjit and Hammes, 2005). During the process of oocyte maturation, the oocyte resumes meiosis and progresses from prophase I to metaphase II (Jamnongjit and Hammes, 2005). Shortly before follicle rupture, the nucleus of the oocyte, or germinal vesicle (GV), undergoes a series of changes that involve GV breakdown (GVBD), chromatin condensation, chromosome segregation and extrusion of the first polar body (Wiersma et al., 1998). Studies using transgenic mouse models indicate that alterations in the factors that regulate these events result in defects in maturation/ovulation as well as other female reproductive phenotypes. For example, phosphodiesterase 3A (PDE3A) is primarily responsible for oocyte cAMP hydrolysis and plays an important role in resuming meiosis in mammalian oocytes. Pde3a–/– mice are completely infertile because their oocytes remain arrested in meiotic prophase I, as demonstrated by the persistence of the nuclear GV, suggesting an inability to resume meiosis (Masciarelli et al., 2004). This supports the notion that cAMP is a maturation-inhibiting factor.
The Ercc1 (excision repair cross complementation group 1) gene is required for successful completion of meiosis in oocytes (Selfridge et al., 2001). Specifically, Ercc1 is essential for the nucleotide excision repair pathway and thus plays a role in homologous recombination, double-strand break repair and the repair of inter-strand DNA cross-links (Selfridge et al., 2001). Ercc1–/– mice die by 3 weeks of age because of a liver problem that can be corrected by use of a liver-specific transgene (Selfridge et al., 2001). Ercc1–/– mice, with the liver-specific transgene, die by 12 weeks of age and are infertile because of a reduced number of oocytes compared with WT mice (Hsia et al., 2003). Furthermore, oocytes that are present often seem to be in the process of degeneration (Hsia et al., 2003).
Notch gene family members are also thought to regulate meiosis (Hahn et al., 2005). The lunatic fringe gene (Lfng) is an important regulator of the Notch-signalling pathway. Lfng–/– mice are infertile and have many aberrant follicles (Hahn et al., 2005). Furthermore, Lfng–/– mice can ovulate when induced, but the oocytes cannot be fertilized because they are incapable of completing meiotic maturation (Hahn et al., 2005). Lfng–/– follicles can undergo cumulus expansion, suggesting a disconnection between cumulus expansion and GVBD, and completion of meiotic maturation (Hahn et al., 2005).
Heat shock factors (Hsfs) are also required for normal progression of meiosis (Christians et al., 2000). HSF1 is required for the development of the embryo to the 2-cell stage as Hsf1–/– females are infertile (Christians et al., 2000). Hsf2–/– females are subfertile and exhibit ovulation defects due to large haemorrhagic follicles with trapped oocytes (Kallio et al., 2002).
Regulators of cell-cycle progression, such as cyclin-dependent kinases (CDKs), cyclin kinase (CK) proteins and G protein-coupled receptor 3 (GPR3), have also been shown to be important for meiotic oocyte maturation (Ortega et al., 2003; Spruck et al., 2003). Specifically, Cdk2–/– mice are infertile, in part, because of improper localization of centromeres and synaptonemal complex protein 3 (SCP3) (Berthet et al., 2003; Ortega et al., 2003). Cdk4–/– mice are also infertile due to impaired progesterone production by the CL (Moons et al., 2002). CK proteins are homologues of CDKs and, thus, are important for oocyte maturation. Ck2–/– mice are sterile because of an arrest of the oocyte at metaphase I (Spruck et al., 2003). GPR3 is endowed with constitutive Gs-signalling activity during the cell cycle (Mehlmann et al., 2004). Gpr3–/– ovaries show a decrease in the percentage of GV-stage oocytes, suggesting that GPR3 is required to maintain meiotic arrest of the oocyte (Ledent et al., 2005). Cytokines may also play an essential role in oocyte maturation. Specifically, GP130 is the shared receptor for members of the IL6 family of cytokines (Molyneaux et al., 2003a). Female mice with a germ cell-specific ablation of Gp130 have fertility defects due to abnormal oocyte maturation/ovulation (Molyneaux et al., 2003a).
Transgenic models with alterations in certain transcription factors also provide evidence that improper gene transcription can alter ovulation. Specifically, the Ahr may be involved in regulating events involved in ovulation as Ahr–/– mice have reduced numbers of CL compared with WT mice (Benedict et al., 2000). Mice deficient in the Zfx, Krox24 (also known as Egr1), do not ovulate and, thus, are infertile (Topilko et al., 1998). Members of the CCAAAT/enhancer-binding protein (CEBP) family could be involved in regulating ovulation, because Cebpb–/– mice lack CL and are completely infertile (Sterneck et al., 1997).
|
Mouse models with alterations in steroid hormones or steroid receptors that control follicle development |
|---|
The process of ovarian follicle development requires the co-ordinated actions of steroid hormones. Steroid hormones also play an important role in the maintenance of reproductive capacity and secondary sex characteristics (Hirshfield, 1991
; Richards, 1994). Steroid hormones produced by the ovary (progestins, androgens and estrogens) are synthesized in a sequential manner by thecal and granulosa cells of follicles (Drummond et al., 2002). The steroidogenesis pathway begins with the precursor cholesterol binding to low-density lipoprotein (LDL) receptors on the plasma membrane of granulosa cells. Steroidogenic acute regulatory (STAR) protein then promotes the transport of cholesterol from the outer membrane to the inner membrane of the mitochondria. This step is considered to be one of the rate-limiting steps in the steroidogenesis pathway.
Once cholesterol reaches the inner membrane of the mitochondria, it undergoes enzymatic cleavage of its side chain by an enzyme known as cytochrome P450 side-chain cleavage (CYP11a1) to form pregnenolone. Pregnenolone is converted to estradiol (E2) via two pathways (Suter, 2004). In the first pathway, pregnenolone is converted to dehydroepiandrosterone (DHEA) by CYP17a1. DHEA is converted to androstenedione (ASD) by 3ß-hydroxysteroid dehydrogenase (3ß-HSD). In the second pathway, pregnenolone is converted to progesterone by 3ß-HSD. Progesterone is then converted to ASD by CYP17a1. Here, the pathways converge, and ASD is converted to testosterone by 17ß-HSD. In the final step, testosterone is converted to E2 by aromatase (Suter, 2004). Once E2 is produced, it can stimulate the growth of ovarian follicles by inducing proliferation of granulosa cells (Suter, 2004).
Several recent studies have used transgenic mouse models to examine the roles of steroidogenic factors in the process of ovarian folliculogenesis, female reproduction and disease. For example, investigators have shown that transgenic mice deficient in Star mimic patients with congenital lipoid adrenal hyperplasia, an autosomal recessive disorder in which patients have impaired adrenal and gonadal steroidogenesis (Miller, 2005). Studies conducted with Star–/– (StARKO) mice reveal that after birth, the animals fail to grow normally and die within 1–2 days after birth because of adrenocortical insufficiency (Caron et al., 1997). Further studies, in which StARKO mice were treated with corticosteroids to keep them alive into adulthood, revealed that StARKO ovaries have impaired folliculogenesis and contain lipid deposits in the stromal cells (Hasegawa et al., 2000). These mice have high levels of CYP11a1 along with decreased levels of progesterone (Hasegawa et al., 2000). Taken together, these studies suggest that Star deficiency has direct consequences on the steroidogenic capacity of the ovary (Caron et al., 1997; Hasegawa et al., 2000).
To our knowledge, there are no transgenic or knockout mouse models with alterations in any of the CYP450 enzymes in the steroidogenic pathway except for the CYP19 (aromatase) knockout (ArKO) model. ArKO mice, which lack the capacity to produce estrogen, are infertile because of impaired folliculogenesis at the antral stage and an inability to ovulate (Fisher et al., 1998). In addition, the antral follicles in ArKO ovaries become haemorrhagic and cystic with advancing age (Britt et al., 2000). The reasons for the abnormal follicle growth in the ArKO model may be because of abnormal levels of hormones including FSH, LH, E2 and testosterone (Britt et al., 2001).
Although there are limited models with alterations in steroidogenic CYP450 enzymes, there are several models with alterations in either steroid hormones or steroid hormone receptors. For example, several studies have been conducted using mice with alterations in the ability to respond to progesterone (Lydon et al., 1995; Conneely et al., 2001). Progesterone plays an important role in the establishment and maintenance of pregnancy. The actions of progesterone are mediated via specific interactions with its corresponding receptors. Progesterone receptors are expressed in two isoforms, PR-A and PR-B, in the granulosa cells of mature antral follicles and CL (Conneely et al., 2001). Mice lacking both PR isoforms (PRKO) are infertile because they fail to ovulate, even though follicles grow to the antral stage (Lydon et al., 1995). Similar to PRKO ovaries, ovaries from mice deficient in only the PR-A isoform (PRAKO) contain numerous mature, arrested anovulatory follicles (Conneely et al., 2001). In addition, PRAKO mice have a severely impaired ability to respond to hormone treatments during ovulation induction (Conneely et al., 2001). Taken together, these data indicate that PR-B alone cannot support normal ovulation and that PR-A is primarily responsible for follicle rupture. These