Physiological effects of major genes affecting ovulation rate in sheep

Genetic mutations with major effects on ovulation rate in sheep were recently identified in two genes of the transforming growth factor (TGFβ) superfamily and a TGFβ receptor, namely bone morphogenetic protein 15 (BMP15), otherwise known as the growth differentiation factor 9b (GDF9b), GDF9 and activin-like kinase 6 (ALK6) otherwise known as the BMP receptor type IB (BMPRIB). Animals homozygous for the BMP15 or GDF9 mutations are anovulatory whereas animals heterozygous for BMP15 or GDF9 or heterozygous or homozygous for ALK6 have higher than normal ovulation rates. Immunisation of ewes against BMP15 or GDF9 shows that both are essential for normal follicular development and control of ovulation rate. Common features of fertile animals with the BMP15, ALK6 (and possibly GDF9) mutations are changes in oocyte development during early preantral follicular growth, earlier maturation of granulosa cells and ovulation of mature follicles at smaller diameters. In summary, these findings have led to a new paradigm in reproductive biology, namely that the oocyte plays a key role in regulating the ovulation rate.


INTRODUCTION
Sheep (Ovis aries) are a highly diverse species with more than 900 different breeds that vary substantially in their physiological characteristics including ovulation rate and fecundity [30]. In some instances, the differences in ovulation rate have been attributed to the action of a single or a closelylinked group of genes [4]. Indeed mutations with major effects on ovulation rate have been identified in two genes of the transforming growth factor beta (TGFβ) superfamily and a TGFβ receptor, namely bone morphogenetic protein (BMP15), growth differentiation factor 9b (GDF9) and BMP receptor-IB (ALK6) [10,11,23,28,33]. Further mutations in some of these genes or in different genes are likely to be present in other prolific breeds.
The aforementioned gene discoveries have led to a new paradigm in reproductive biology, namely that the oocyte plays a key role in regulating the ovulation rate [10][11][12]. Thus, sheep are proving to be remarkably informative in offering new insights into why some mammals have a predisposition to produce two or three offspring rather than one [18]. The purpose of this review was to summarise some of the physiological effects of the BMP15, GDF9 and ALK6 mutations affecting ovulation rate in sheep.

THE BMP15 AND GDF9 MUTATIONS
Five separate point mutations in the BMP15 gene and one point mutation in the GDF9 gene have been identified as affecting the ovulation rate [2,10,11]. Animals homozygous for each of these mutations are anovulatory and thus sterile whereas heterozygous animals have mean ovulation rates between 0.8 and 2.4 above that of the respective non-carrier flocks [5,11]. Two of the five BMP15 mutations have premature stop codons, one of these (FecX G in Belclare and Cambridge sheep) is at amino acid position −29 in the proregion of exon 2 before the mature region thus no mature protein is produced and the other is the Hanna mutation (FecX H ) at amino acid 23 of the mature protein rendering it inactive. Another two mutations are nonconservative amino acid substitutions within the mature protein at amino acid positions 31 (Inverdale; FecX I ) and 99 (Belclare; FecX B ). The other identified mutation has been reported in Lacaune ewes and is reported elsewhere in this journal [2]. This mutation was found as a co-dominant mutation in an animal with an autosomal gene affecting ovulation rate. For all mutations, the ovarian phenotypes are indistinguishable in homozygous carriers, in that few, if any, follicles develop beyond the primary (i.e. type 2) stage of development (Fig. 1). The GDF9 mutation (Belclare; FecG H ) corresponds to a non-conservative amino acid substitution at position 77 of the mature protein region. The ovarian phenotype in animals homozygous for this mutation is different than that for the BMP15 mutations in that ovarian follicles continue to develop to the antral (i.e. type 5) stages although most, if not all, are abnormal with respect to oocyte morphology and the arrangement and appearance of the granulosa and cumulus cell-types (Fig. 1).  4 (preantral) follicles in a wild-type sheep (a) and abnormal ovarian follicles with similar diameter oocytes (-80 µm) in an Inverdale ewe with a homozygous non-conservative aspartic acid substitution for valine at amino acid 31 in BMP15 (b); in a Belclare ewe homozygous for a stop codon at amino residue −29 (i.e. in the proregion of BMP15) (c) and in a Belclare ewe with a homozygous isoleucine substitution for serine at amino acid residue 77 in GDF9 (d). The photomicrographs (e) and (f) show immunostaining for GDF9 and BMP15 in oocytes of type 1 (i.e. primordial) follicles and a type 2 (primary) follicle, respectively, whereas (g) and (h) show the corresponding photomicrographs where the signal has been blocked with an ovine GDF9 peptide or an E. coli derived ovine BMP15 mature protein respectively. The follicles shown in photomicrographs (e-h) are from normal follicles in wild-type sheep. Scale bars a-d = 70 µm; e-h = 25 µm.
The primary cause of infertility in ewes homozygous for the BMP15 or GDF9 mutations is the apparent lack of bioactive BMP15 or GDF9. These findings are in contrast to those reported for mice [7,34]. GDF9 null mice have been found to be sterile and follicular growth is blocked at the primary stage, whereas in BMP15 null mice, no obvious effects on follicular growth and embryo development have been noted, although some defects in fertilisation have been observed. In the heterozygous mice, no differences in ovulation rate have been reported. To explain this paradox, Liao et al. [15] suggest that this might be due to altered intracellular processing and secretion of BMP15 and GDF9 because of the nature of the mutations in sheep and of the way in which the mouse BMP15 gene is mutated by deleting the entire second exon. These authors produced a recombinant human BMP15 with an amino acid substitution at position 31 which mimics the mutation in Inverdale sheep. They found that a mutant dimerised form of BMP15 can be secreted from a cell line that does not contain a GDF9 expression system. However, when the mutant form of BMP15 is co-expressed with GDF9, the proteolytic processing and secretion of BMP15 is inhibited and the secretion of GDF9 is impaired. It was proposed that the mutations in the BMP15 gene in sheep may actually affect the level of GDF9 secretion and that abnormal concentrations of GDF9 are the cause of the Inverdale phenotype. However, when sheep are immunised against peptides specific for BMP15 or GDF9, the evidence is that secreted forms of BMP15 as well as GDF9 are essential for both normal follicular development and ovulation rate [12,18].

The BMP15 and GDF9 immunisation experiments
The purpose of the immunisation experiments in sheep was to test whether extracellular concentrations of GDF9 or BMP15 are important for the regulation of follicular development and ovulation rate. To test this, the ewes were immunised at monthly intervals for 7 months with keyhole limpet haemacyanin (KLH) or KLH conjugated with an E. coli derived mature BMP15 protein or with a 15mer GDF9 or BMP15 peptide respectively in Freunds adjuvant. The major effects on ovarian phenotype are a significant reduction in ovarian volume, no normal follicular development beyond the type 2 stage, few, if any type 5 (i.e. antral) follicles and significantly larger oocytes in the type 2-4 follicles in both the BMP15 and GDF9 groups [12]. This phenotype is similar to that observed in sheep with the BMP15 and GDF9 mutations. Passive immunisation of ewes with BMP15 or GDF9 specific antibodies during the follicular phase of the oestrous cycle showed that both factors are also essential during the preovulatory follicular maturation. In follow-up studies, where ewes were given a primary and booster injection of KLH or bovine serum albumin (BSA) conjugated to the BMP15 or GDF9 peptide, in a waterbased adjuvant, significant increases in mean ovulation rates of around 0.4 to 1.2 were achieved relative to the KLH or BSA immunised controls [13] (McNatty, Juengel, unpublished data). These findings suggest that a partial neutralisation of BMP15 or GDF9 in biological fluid leads to an ovulation response similar to those observed in animals with the heterozygous mutations of either BMP15 or GDF9 [10,11]. While these studies show that both GDF9 and BMP15 are essential for follicular growth and ovulation rate, they do not rule out the possibility that some of their effects are due to the presence of BMP15/GDF9 heterodimers.

Intraovarian sites of BMP15 and GDF9 gene expression and possible mechanisms of action
Within the sheep ovary, BMP15 and GDF9 are expressed exclusively by the oocyte (Fig. 2). Both GDF9 mRNA and protein are present in normal sheep oocytes in non-growing type 1 follicles (i.e. primordial follicles) as well as in growing follicles [3,12] (Fig. 1). In contrast, BMP15 mRNA and protein are both present from the type 2 (i.e. primary) stage (Fig. 2). With respect to receptors for the TGFβ/BMP ligands, BMPR-II, activin receptor-like kinase 6 (ALK6), ALK3, and ALK5 mRNA are present in oocytes at all stages of follicular growth [1,29,33] (Fig. 2; Juengel, unpublished data). In granulosa cells, BMPR-II and ALK3 are present at all stages including type 1 follicles, whereas ALK6 is present from the type 2 stage and ALK5 from the type 4 stage of growth. In theca interna cells, BMPR-II, ALK3, ALK5 but not ALK6 are present from the type 4 stage of growth [1,32] (Fig. 2). However, a low level of ALK6 protein has been detected in theca by immunohistochemistry suggesting that this receptor might be present [29]. In rats, BMP15 binds to ALK6 with very little affinity for ALK3 and BMPR-II inhibits BMP15 induced cell functions [22] suggesting that cells expressing both ALK6 and BMPR-II are probable targets for the BMP15 ligand. If this holds true for sheep, then it is likely that granulosa cells and the oocyte are affected by oocyte-derived BMP15 from the type 2 stage of growth [33]. In rat granulosa cells, GDF9 is known to bind to BMPR-II [32] and to stimulate inhibin A and B production via the Smad2 pathway [26] inferring that the type 1 receptor might be an activin or TGFβ receptor rather than ALK3 or ALK6 [24]. Indeed a recent report shows that ALK5 (TGFβRI) is the type 1 receptor for GDF9 [19]. This finding suggests that GDF9 could have autocrine effects on the oocyte from the type 1 stage and paracrine effects on both the granulosa and theca cells from the type 4 (range  preantral) stage of growth. It is worth noting that GDF9 mRNA and protein are present before follicular formation in sheep as well as in non-growing type 1 follicles in several species including sheep, cows and possum [3,8,12]. Thus, either the form of the protein is biologically inactive, the appropriate receptors are not present on the adjacent granulosa cells in type 1 follicles or that GDF9 is a relatively inactive growth factor when other companion factors are absent.
The roles of GDF9 and BMP15 as assessed from in vitro studies appear complex with some possible species differences. In rodents, both BMP15 and GDF9 stimulate proliferation of cultured granulosa cells from antral follicles, whereas in sheep only recombinant ovine BMP15 has been shown to induce modest proliferation [25,31] (McNatty, unpublished data). An explanation for the higher ovulation rate observed in heterozygous Inverdale or Hanna ewes has been offered by Shimasaki and his colleagues [25] based on their in vitro findings that BMP15 may reduce the sensitivity of granulosa cells to FSH by inhibiting expression of the FSH receptor. Thus, in animals such as the heterozygous Inverdale or Hanna with a lower than normal concentration of BMP15, a higher FSH-induced granulosa cell responsiveness would result. In granulosa cells from heterozygous Inverdale ewes, a higher FSH-induced cAMP responsiveness has been observed relative to wild-type animals in the absence of any differences in the plasma concentrations of FSH, LH or ovarian steroids [27]. Since GDF9 has also been shown in rats to inhibit FSH-induced cAMP synthesis [31], the overall effect of the GDF9 mutation in sheep may be to enhance the sensitivity of the ovarian follicle to FSH and thereby increase the ovulation rate in a manner similar to that noted for BMP15.

The ALK6 (BMPR-IB) mutation
The ALK6 mutation was first identified in Booroola ewes (FecB) as a point mutation at nucleotide position 830 leading to an arginine replacing a glutamine amino acid (i.e. Q249R) in a highly conserved region of the intra cellular kinase domain [23,28,33]. The dominant phenotype is precocious ovarian follicular development [20] and exceptional prolificacy [4]. Some significant secondary effects have been observed in Booroola sheep, most likely because ALK6 mRNA is present in a wide range of tissues as well as the ovaries, including the brain, pituitary, kidneys, skeletal muscle, uterus and prostate and testes [20,33]. ALK6 mRNA is also likely to be expressed in several tissues during fetal development such as the heart and mesonephros [17]. The evidence for effects of the Q249R ALK6 mutation in various tissues during fetal development as well as on pituitary function and ovarian follicular formation have been summarised elsewhere [20]. As reported by Davis [4], the Booroola (Q249R ALK-6) mutation has now been identified in a number of different sheep breeds throughout the world.
The ALK6 knockout mouse does not display the same phenotype as the Booroola ewe [36]. ALK6 knockout mice appear to have normal ovarian follicular development and ovulation rates although there is a reduction in fertility perhaps caused by the failure of normal cumulus expansion. Unlike Booroola sheep, ALK6 null mice exhibit defects in chondrogenesis leading to defects in skeletal development [35]. In two unrelated human families, point mutations have been identified in the ALK6 gene [14]. In one family, the mutation resulted in a lysine residue substituting for an isoleucine (I200K) in a region of ALK6 involved in phosphorylation of the receptor. In the other family, a tryptophan is substituted for an arginine (R486W) in a highly conserved region of the C-terminal of the ALK6 kinase domain. These mutations cause malformations in limb development but there are no reports of effects on fertility. Experiments to block the actions of putative ligands interacting with the ALK6 receptor in sheep or mutating the ALK6 receptor in mice to generate the Booroola mutant have not been reported. Thus, comparisons of the ALK6 mutation in sheep with the ALK6 knockout mice to explore mechanistic interpretations of species differences in ovulation rate are premature.

Intraovarian effects of ALK6 expression and possible mechanisms of action
ALK6 together with BMPR-II mRNA have been identified in ovine oocytes of type 1 follicles and expression levels for both remain high throughout the follicular types 1-4 and early type 5 stages of growth [37] (Fig. 2). Thereafter, the levels of ALK6 but not BMPR-II mRNA in oocytes decline in large antral follicles [33]. Both ALK6 and BMPR-II mRNA are also present in granulosa cells from the type 2 stage of growth but little or no ALK6 is detectable in theca interna cells (Fig. 2). It seems likely that there is a functional interaction between BMP15 and ALK6 since the effect on the ovulation rate is multiplicative in Booroola Inverdale crossed sheep that are heterozygous for the ALK6 and BMP15 mutations [6]. Moreover, it has been shown that the BMP but not the TGFβ1 or activin A signalling pathway is altered in granulosa cells from carriers of the Q249R ALK6 mutation [9,21]. Given that BMP6 and BMP15 were expressed in ovine oocytes from the type 1 and 2 stages respectively ( Fig. 2; Juengel, unpublished data), it is reasonable to expect that one or  both of these intrafollicular ligands would interact with the BMPR-II/mutant ALK6 complex [10,11] to affect some change in oocytes or granulosa cells in early growing follicles and this appears to be the case. In animals that are homozygous carriers and non-carriers for the Booroola mutation, oocytes reach their mature diameter of ∼130 µm at the types 4 and 5 stages of development respectively [33]. This divergence between the genotypes in diameters of the oocytes is already evident at the type 3 follicular stage. Changes in granulosa cell numbers with respect to follicular diameters (e.g. 75 to ≥3000 µm) have also been examined after fitting a broken stick regression model (i.e. 2 straight lines with an inflection point) for each genotype (see Tab. I). The results from analyses of the granulosa cell populations from the type 2 to type 5 stages of development show that the Booroola mutation influences granulosa cell development both before as well as after antrum formation [16] (Tab. I).
In preantral follicles, the granulosa cells accumulate in greater numbers with respect to follicular diameter in BB compared to ++ ewes until follicles reach a 639 µm diameter. Thereafter in ++ ewes, the granulosa cells continue to accumulate as beforehand until the follicles reach a 1224 µm diameter. Once BB and ++ follicles reach the aforementioned respective diameters, the accumulation of granulosa cells with respect to follicular diameter changes but it continues to remain higher in the BB genotype (Tab. I). However follicles in BB ewes reached an ovulatory size at −3000 µm diameter and in ++ ewes between 5000 and 6000 µm diameter with the mean numbers of granulosa cells per preovulatory follicle being 1.9 million and ∼5 million, respectively. These differences in oocyte and granulosa cell development in sheep with the Q249R ALK6 mutation are associated with a significantly higher FSH-induced cAMP responsiveness and earlier onset of follicular steroidogenesis including earlier onset of aromatase activity in granulosa cells, together with an earlier onset of LH responsiveness in the granulosa cells (for a review, see [20]). All of these events in the ALK6 mutated ewes lead to maturation and ovulation of multiple ovarian follicles at smaller follicular diameters without any changes in the ovarian secretion rates of steroid or inhibin during either the follicular or luteal phases of the oestrous cycle [20]. These results when taken together with those from Fabre et al. (2003) [9] suggest that some intraovarian factors have antidifferentiative roles (e.g. anti-luteinising effects) on ovine granulosa cells and that one important consequence of the mutated ALK6 receptor is a decreased ability of some BMP to inhibit differentiation. It remains to be determined how specific BMP act on the BMPR-II complexes and how the downstream signalling pathways of the mutated ALK6 receptor influences cell function in Booroola sheep.

CONCLUSION
A common feature of the BMP15 and ALK6 mutations is that changes in oocyte are evident during early preantral (i.e. type 3) development followed by precocious follicular maturation with each preovulatory follicle containing a reduced population of granulosa cells. Although not proven, the same characteristics are likely to be true for sheep with the GDF9 mutation. The immunisation studies show that both GDF9 and BMP15 are essential for follicular development in sheep but it remains to be determined whether GDF9 and BMP15 act as heterodimers and whether these growth factors either separately or together exert their effects through the ALK6 receptor. In summary, these data suggest new possibilities for developing novel therapeutic reagents for regulating fertility in mammals.