| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Division of Experimental Medicine,3 Department of Obstetrics and Gynecology,4 McGill University, RVH, Montreal, Québec, Canada H3A 1A1
Département de Biologie Médicale,6 Faculté de Médecine, Le Centre de Recherche en Cancérologie de l'Université Laval, L'Hôtel-Dieu de Québec, Québec, Canada G1R 2J6
Department of Physiology,5 McGill University, Montreal, Québec, Canada H3G 1Y6
ABSTRACT
NR5A2 is an orphan nuclear receptor involved in cholesterol metabolism and embryogenesis. The high level of expression of NR5A2 in the ovary and its involvement in the regulation of steroidogenic gene expression also suggest a role for this transcription factor in female reproductive function. In vivo evidence for a role for NR5A2 in fertility, however, is still lacking. In order to address this possibility, we used Nr5a2+/– mice to demonstrate that heterozygosity for a null mutation of Nr5a2 leads to a decreased fertility in females. Our results indicate that although Nr5a2+/– mice display normal follicular development, ovulation, and estrogen production, they exhibit altered luteal function. More specifically, we show that the reduced reproductive ability of Nr5a2+/– females arises from a reduction in circulating progesterone concentrations and can be rescued by exogenous progesterone supplementation. This study therefore provides the first in vivo evidence for a role of NR5A2 in reproductive function and steroidogenesis.
corpus luteum function,, estrogen,, fertility,, granulosa cells,, implantation,, NR5A2,, pregnancy,, progesterone,, steroidogenesis
The orphan nuclear receptor NR5A2, also known as fetoprotein transcription factor (FTF) and liver receptor homolog-1 (LRH-1), plays an important role in embryogenesis and hepatic metabolism. A targeted mutation of the gene coding for NR5A2 in the mouse results in embryonic lethality around the gastrulation period, indicating a requirement for this nuclear receptor early in development [1–5]. Although the early embryonic lethal phenotype of Nr5a2–/– mice precludes the identification of later developmental functions for this nuclear receptor, a significant body of evidence underscores a potential role for this factor during development of endoderm-derived organs. The strong expression pattern of NR5A2 in the embryonic endoderm and its derivatives, the developing liver, intestine, and pancreas, along with its well-established role in the transcriptional cascade leading to hepatopancreatic specification, suggest a role for NR5A2 in endodermal differentiation during organogenesis [3, 6–8]. In adult animals, Nr5a2 expression is predominant in endoderm-derived organs, namely the liver, intestine, and pancreas [6, 9, 10]. Consistent with its enterohepatic expression pattern, NR5A2 has been shown to act as a key regulator of cholesterol homeostasis and to be involved in intestinal crypt cell proliferation in vivo [3, 10, 11]. NR5A2 has been implicated in reverse cholesterol transport, bile acid synthesis and enterohepatic circulation [10], and mice haploinsufficient for or overexpressing NR5A2 exhibit altered cholesterol metabolism [3, 12, 13].
In addition to its role in metabolism and development, recent evidence points to a role for NR5A2 in fertility. Nr5a2 has been shown to be expressed in steroidogenic tissues such as pre-adipocytes [14], the ovary [1, 2, 15–19] and the testis [18–20]. Interestingly, the highest level of NR5A2 expression is detected in the ovary, where it is restricted to the granulosa cells of the developing follicle and to the luteal cells of the corpus luteum [2, 16, 17]. In ovarian follicles, Nr5a2 appears to be regulated by FSH and LH, whereas in the corpus luteum its expression is modulated by prolactin [2, 17]. In addition, NR5A2 has also been shown to control steroid hormone biosynthesis by regulating the expression of genes involved in steroidogenesis, including steroidogenic acute regulatory protein (STAR) [18, 21], cholesterol side-chain cleavage (Cyp11a1) [18, 22, 23], 17
-hydroxylase, 17,20 lyase (Cyp17) [8], 3ß-hydroxysteroid dehydrogenase (3ß-Hsd) [18, 22, 24], 11ß-hydroxylase (Cyp11b1) [18], and P450 aromatase (Cyp19) [14]. Therefore, the elevated gonadal expression of NR5A2 and its ability to regulate the expression of genes involved in steroidogenesis point to a role for NR5A2 in reproductive functions. In vivo evidence for a role for NR5A2 in fertility, however, is lacking.
To evaluate the potential involvement of NR5A2 in reproductive functions in vivo, we used mice heterozygous for a null mutation of Nr5a2 and showed that Nr5a2 haploinsufficiency specifically leads to a reduction in female fertility, thereby establishing, for the first time, a role for NR5A2 in female reproductive function. Furthermore, the results presented herein clearly indicate that the reduction in fertility observed in Nr5a2+/– mice results from impaired progesterone production.
Maintenance, Mating, and Manipulation of Mice
All animal care and experimental procedures were approved by the Animal Care Committee of the Royal Victoria Hospital and were in accordance with the regulations established by the Canadian Council on Animal Care. The generation of Nr5a2+/– mice was described previously in Paré et al. [3]. All mice were housed in filter-topped isolator cages under a 12L:12D cycle (700–1900 hours). Successful fertilization/mating events were assessed by the presence of a vaginal plug. The morning the plug was observed was considered 0.5-day post-coitum (dpc). Hormonal stimulation and superovulation of immature (21- to 24-day-old) Nr5a2+/– and wild-type MF1 females were performed by intraperitoneal injection of 7.5 IU of eCG (Sigma, cat # G4877) alone or followed by intraperitoneal injection of 5 IU of hCG (Sigma cat # CG10) 46 h later.
Immature mice (21 days old) obtained from Nr5a2+/– crosses were genotyped using PCR on tail DNA. The targeted Nr5a2 allele was amplified using oligonucleotides for either the neomycin resistant cassette or the LacZ reporter gene used to disrupt the Nr5a2 gene [3]. LacZ primers sequence: LacZ forward: CAG TGG CGT CTG GCG GAA AAC CTC, LacZ reverse: GGC GGC AGT AAG GCG GTC GG, Neo primer sequences: Neo forward: GGC TAT GAC TGG GCA CAA CAG ACA ATC, Neo reverse: AGC TCT TCA GCA ATA TCA CGG GTA GC. The sequences of the oligonucleotides used to amplify the wild-type Nr5a2 allele were: FTF forward: TAC AGC CTC CAA ATT TTG CC and FTF reverse: TAT CGC CAC ACA CAG GAC AT. The PCR conditions were as follows: denaturation at 94°C for 60 sec, annealing at 58°C for 60 sec, elongation at 72°C for 60 sec for 30 cycles.
Temporal Determination of Pregnancy Loss
Nr5a2+/– females were mated with fertile wild-type or Nr5a2+/– males and were killed at different gestation times from 6.0 to 10.0 dpc, to determine when pregnancy loss occurs. At least 15 Nr5a2+/– females were killed at each gestational time point: 6.0, 7.0, 8.0, 9.0, and 10 dpc. Over 168 Nr5a2+/– females mated with Nr5a2+/– males were killed at 6.5 dpc to obtain embryos for another study [5]. The presence, development, and number of deciduas were evaluated and compared to those seen in wild-type females at the same gestational stage (n > 20). To examine ovulation, Nr5a2+/– females were superovulated and the number of oocytes released was evaluated by counting the number of oocytes recovered from the oviduct.
Ovaries were fixed overnight with 4% paraformaldehyde in PBS at 4°C. Fixed ovaries were then dehydrated through a graded series of ethanol, cleared in xylenes (two times for 15 min each) and embedded in paraffin. Paraffin-embedded ovaries were sectioned (8–10 µm thick) and stained with hematoxylin and eosin.
Whole-cell protein extracts were prepared by homogenization of ovaries in lysis buffer (50 mM Tris-Cl pH 6.8, 100mM dithiothreitol, 2% SDS, 10% glycerol, and 1% protease inhibitor cocktail from Sigma, cat # P8340) followed by a brief sonication and subsequent centrifugation (14 000 x g for 5 min at 4°C) to collect soluble proteins. Prior to gel electrophoresis, protein from whole cell extracts were denatured for 5 min at 90°C in gel loading buffer (50 mM Tris-Cl pH 6.8, 100 mM dithiothreitol, 2% SDS, 10% glycerol, and 0.2% bromophenol blue). Protein extract concentrations were determined as described in Markwell et al. [25].
Approximately 30 µg of ovarian extracts were subjected to SDS-PAGE electrophoresis. Proteins were resolved in 11% acrylamide gels. Resolved proteins were transferred electrophoretically to PVDF membranes (Amersham Biosciences, GE Healthcare) and blocked for at least 1 h at room temperature in Tris-buffered saline containing 0.1% Tween-20 (TBST) and 0.2% blocking reagent (Roche). Blots were incubated overnight at 4°C with a 1:1000 dilution of rabbit anti-STAR antibody (generously provided by Dr. Stocco, Department of Cell Biology and Biochemistry, Texas Tech University Health Sciences Center, Lubbock, TX) or with 1:1000 dilution of goat anti-actin antibody (Santa Cruz Biotechnology, CA, cat # Sc1616) in blocking buffer. Subsequently, blots were washed three times (15 min) in TBST at room temperature, after which they were incubated with a 1:2000 dilution of goat anti-rabbit antibody conjugated to horseradish peroxidase (Sigma, cat # A0545) or with a 1:4000 dilution of donkey anti-goat antibody conjugated to horseradish peroxidase (Santa Cruz Biotechnology, CA, cat # Sc2020) for 1 hr and washed three times (15 min) in TBST. Immunoreactive bands were revealed by chemiluminescence using the ECL plus kit (Amersham Biosciences, GE Healthcare). Blots were exposed to BioMax Light-1 Kodak films (Cat # V1788207, Amersham Biosciences, GE Healthcare). Actin-normalized quantification of the intensity of the chemiluminescence of immunoreactive bands was performed by scanning the blots with a phosphoimager (Storm Scanner 840; Amersham Biosciences, GE Healthcare), and data were analyzed using the ImageQuant V5.2 software (Amersham Biosciences, GE Healthcare).
Silastic Progesterone Implants
Progesterone implants consisted of 6-mm pieces of Silastic medical grade tubing (id, 0.132 in; od, 0.183 in; Dow Corning, Midland, MI) packed with crystalline progesterone (Steraloids). Empty or progesterone-packed implants were inserted subcutaneously on the third day of gestation and mice were killed at 8.5 dpc, 2 days after the time when pregnancy loss occurs in Nr5a2+/– females exhibiting reduced fertility. The frequency of pregnancy was assessed by the presence of well-developed deciduas and embryos.
Measurement of Serum Steroid Concentrations
Mice were anesthetized by inhalation of 1-chloro-2,2,2-trifluoroethyl difluoromethyl ether (Baxter Corporation, Toronto, Ontario), and blood samples were collected by cardiac puncture and serum was obtained by centrifugation using microtainer serum separator tubes (Becton Dickinson) 48 h after eCG treatment, 24 h after eCG/hCG treatment, 24 or 48 h after saline treatment, 5 days post-hCG injection for luteinized ovaries, and after 6.0 or 8.5 (implants) days of natural pregnancy. Serum concentrations of 17ß-estradiol, and progesterone in immature females following gonadotropin stimulation were initially determined in duplicate by ELISA and were performed at the Centre for Bone and Periodontal Research (McGill University, Royal Victoria Hospital, and Montreal) and progesterone levels were confirmed by RIA using the service provided by Dr. Bruce D. Murphy (Centre de Recherche en Reproduction Animale, Faculté de Médecine Vétérinaire, Université de Montréal, St-Hyacinthe, Québec, Canada). Serum levels of progesterone during natural pregnancy following hormonally induced luteinization or progesterone supplementation were measured by RIA in duplicate using the service provided by Dr. Bruce D. Murphy.
All data are presented as the mean ± SEM. Data from the progesterone supplementation experiments in NR5a2+/– animals (see Table 2) were analyzed using the chi-square test. All other statistical analyses of variance were done using two-tailed student t-test, except for Figure 5B, where a one-tailed Student t-test was performed. Based on results presented in previous figures, we assumed that the level of progesterone should also be lower in Nr5a2+/– following luteinization and therefore chose to perform a one-tailed student t-test as a statistical analysis of variance for the results depicted in Figure 5B. A value of P < 0.05 was considered to be significant. Each Western blot experiment was repeated at least twice.
|
|
Nr5a2 Heterozygous Females Exhibit a Reduction in Fertility
The targeted disruption of the Nr5a2 gene in embryonic stem (ES) cells and subsequent generation of Nr5a2+/– mice has been described previously [3]. Mice heterozygous for the Nr5a2 null mutation are viable and appear morphologically normal. In an attempt to examine the potential implication of NR5A2 in reproductive functions in vivo, we evaluated the frequency of pregnancy resulting from different crosses using Nr5a2+/– mice. We reasoned that since Nr5a2 is expressed in both the ovary [1, 2, 15–19] and the testis [18–20], NR5A2 could be involved in either female or male reproductive functions. To determine the frequency of pregnancy in this Nr5a2+/– mating study, successful pregnancy was evaluated by allowing the females to go to term. Table 1 shows that although 100% of wild-type females mated with either wild-type or Nr5a2+/– males led to successful pregnancies, only 56% of Nr5a2+/– females mated with wild-type males had successful pregnancies (P < 0.05). It is important to note, however, that some Nr5a2+/– females that did not support pregnancy after the first breeding event were able to generate offspring on subsequent matings with wild-type males. Conversely, some Nr5a2+/– females that had a successful pregnancy after the first mating event were not always able to sustain a successful pregnancy in subsequent mating events. Collectively, these results clearly indicate that heterozygosity for a null Nr5a2 mutation leads to a reduced reproductive ability in females, providing the first in vivo evidence for the involvement of NR5A2 in female fertility. To determine when the pregnancy loss takes place in Nr5a2+/– females, we killed Nr5a2+/– females at different gestation times, from 6.0 to 10.0 dpc, and found that pregnancy loss occurs prior to 6.0 dpc (data not shown). Furthermore, we have found that out of 168 Nr5a2+/– females killed at 6.5 dpc, only 59.5% exhibited successful pregnancy, as determined by the presence of well-developed deciduas and embryos, and that the remaining 40.5 % were marked by decidual resorption. Interestingly, the frequency at which decidual resorption occurred in Nr5a2+/– females (59.5%) closely matched the frequency of pregnancy loss depicted in Table 1 (56%).
|
Morphologic Phenotype of Nr5a2+/– Ovaries
Having demonstrated that Nr5a2+/– females display a reduction in fertility, we next wanted to determine the underlying cause for this phenotype. The Nr5a2 ovarian expression pattern suggests that NR5A2 might play a role in either folliculogenesis or corpus luteum formation, or both. To assess if defects in folliculogenesis and/or corpus luteum formation were responsible for the decreased reproductive potential of Nr5a2+/– mice, we performed a histologic analysis of Nr5a2+/– ovaries with or without hormonal stimulation. No obvious morphologic defects were observed in Nr5a2+/– ovaries when compared with their wild-type counterparts (Fig. 1). Moreover, no difference was detected between the weight of Nr5a2+/– and wild-type ovaries in either the absence or presence of hormonal treatment (data not shown). The presence of follicles at all stages of development (Fig. 1, C and D) and corpora lutea (Fig. 1, E and F) indicate that folliculogenesis and corpus luteum formation occur properly in Nr5a2+/– mice. This suggests that the Nr5a2+/– fertility phenotype does not result from defects in follicular development or corpus luteum formation, but rather from defects in pregnancy support. To confirm the absence of defects in follicular development and function, we analyzed the ovulation process in Nr5a2+/– females. The number of oocytes released upon gonadotropin stimulation did not differ between wild-type and Nr5a2+/– females (Fig. 2A). Importantly, none of the hormonally treated Nr5a2+/– females failed to ovulate. In agreement with the similar number of oocytes released following superovulation, the average litter size was not significantly different between Nr5a2+/– females that sustained pregnancy and wild-type females (Fig. 2B). The similarity in litter size of wild-type females mated with either wild-type or Nr5a2+/– males is consistent with the absence of fertility problems in Nr5a2+/– males. In addition, we have found that all implantation sites were affected in females exhibiting decidual resorption at 6.5 dpc and that the total number of implantation sites did not differ between Nr5a2+/– females that sustain pregnancy, Nr5a2+/– females showing decidual resorption, and wild-type females at 6.5 dpc (data not shown), which further supports the absence of ovulatory defects in Nr5a2+/– mice. Moreover, this finding also demonstrates that implantation occurs properly in these females.
|
|
Steroid Hormone Production in Nr5a2+/– Females
The establishment and maintenance of pregnancy rely on the integrity of ovarian steroidogenesis. Despite the absence of morphological ovarian abnormalities and ovulatory defects in Nr5a2+/– mice, it is possible that impaired steroidogenesis may be responsible for their reduced reproductive potential. Since NR5A2 has been implicated in the regulation of genes involved in steroidogenesis [14, 21, 23, 24], we examined this possibility by evaluating the serum concentrations of two steroid hormones crucial for female reproductive function, namely estrogen and progesterone, in response to hormonal stimulation. Levels of circulating estrogen in Nr5a2+/– females were similar to those observed in wild-type females in the absence or presence of hormonal treatment (Fig. 3A). Indeed, both wild-type and heterozygous females displayed low levels of circulating estrogen in response to saline injection, whereas estrogen levels were significantly increased upon eCG stimulation, and decreased following hCG treatment. These results demonstrate that a single Nr5a2 allele is sufficient to maintain appropriate levels of ovarian estrogen synthesis.
|
We next evaluated serum levels of progesterone in Nr5a2+/– females in the absence or presence of gonadotropin stimulation. The levels of circulating progesterone were low to undetectable in Nr5a2+/– and wild-type females in the absence of hormonal stimulation (Fig. 3B). Following eCG treatment, progesterone levels increased in both Nr5a2+/– and wild-type females. This increase, however, was significantly lower in Nr5a2+/– mice. When eCG stimulation was followed by hCG treatment, circulating levels of progesterone increased in Nr5a2+/– mice and reached levels similar to those seen in wild-type females. Thus, Nr5a2+/– females display impaired progesterone production in response to gonadotropin stimulation. This finding is supported by assessment of the weight of the uterus. In the absence of gonadotropin stimulation, both Nr5a2+/– and wild-type uterine weights were similar. In response to hormonal treatment, the uterine weights of Nr5a2+/– females were significantly higher, consistent with their lower level of circulating progesterone after eCG stimulation (Fig. 4). Collectively, these results indicate that haploinsufficiency for Nr5a2 leads to a reduction in progesterone production, and more specifically identify a defect in the Nr5a2+/– ovarian progesterone response to eCG.
|
Since progesterone is central for the establishment and maintenance of pregnancy, we further characterized the impairment in progesterone synthesis in Nr5a2+/– females. The level of circulating progesterone was assessed in both wild-type and Nr5a2+/– females on Day 6 of gestation, the period when pregnancy loss was normally observed in Nr5a2+/– females that displayed decreased fertility. The average progesterone level of Nr5a2+/– females, either pregnant or that had undergone embryo loss, was significantly lower than that observed in wild-type females (Fig. 5A). Notably, pregnancy loss occurred specifically in Nr5a2+/– females exhibiting the lowest levels of progesterone. These results demonstrate that progesterone production is decreased in Nr5a2+/– females during the course of natural pregnancy.
The reduction in the level of progesterone during natural pregnancy suggested that NR5A2 might be important for luteal function and corpus luteum maturation. To test this possibility, we determined the serum level of progesterone following hormonal induction of luteinization. The level of circulating progesterone was dramatically reduced in luteinized Nr5a2+/– ovaries compared with their wild-type counterparts (Fig. 5B). Taken together, these results clearly indicate impairment in progesterone production in Nr5a2+/– females and suggest that this reduction in progesterone synthesis might be responsible for their decreased fertility.
Progesterone Supplementation Rescues Pregnancy in Nr5a2+/– Mice
The decrease in progesterone production in Nr5a2+/– females following eCG stimulation, hormonally induced luteinization, and during natural pregnancy strongly suggested that defective progesterone synthesis might underlie the reduced reproductive ability of Nr5a2+/– females. To further examine the potential role of impaired progesterone production in the Nr5a2+/– fertility phenotype, we performed a progesterone supplementation experiment. If impaired progesterone synthesis was responsible for the reduced fertility seen in Nr5a2+/– females, we would expect that administration of progesterone implants at the beginning of pregnancy would rescue the Nr5a2+/– fertility defect. Table 2 shows that insertion of either an empty or progesterone implant did not interfere with normal pregnancy in wild-type females. However, administration of a progesterone implant, but not of an empty implant, restored the frequency of pregnancy to wild-type level in Nr5a2+/– females. Moreover, levels of circulating progesterone following administration of an empty implant were significantly lower in Nr5a2+/– animals (P < 0.05). Conversely, administration of a progesterone implant restored the level of circulating progesterone to wild-type levels in Nr5a2+/– mice. In addition, the number of conceptus was similar between wild-type females and Nr5a2+/– females that sustained pregnancy following the administration of an empty implant. This was also the case for the number of conceptus recovered following the administration of a progesterone implant (Fig. 6). Furthermore, the total number of deciduas did not differ between females that received an empty versus a progesterone implant. Thus, the rescue of the Nr5a2+/– reproductive phenotype by progesterone supplementation indicates that the impairment in progesterone production seen in Nr5a2+/– females is responsible for the reduction in fertility.
|
Ovarian Expression of STAR in Nr5a2+/– Females
Although our results indicate that ovarian progesterone production was impaired in Nr5a2+/– females, the molecular mechanism(s) by which NR5A2 regulates progesterone synthesis in vivo is unclear. Since NR5A2 has been shown to modulate the expression of genes involved in progesterone synthesis, including STAR, which regulates one of the rate-limiting steps of progesterone production [26], we next determined the expression of STAR in Nr5a2+/– ovaries. To this end, Western blot analyses were performed on ovarian extracts obtained from naturally pregnant females and on luteinized ovarian extracts (Fig. 7). Although the level of STAR protein expression was not altered in Nr5a2+/– females on Day 6 of gestation (Fig. 7, A and C), there was a significant reduction (approximately 6-fold) of ovarian STAR protein expression following hormonally induced luteinization in Nr5a2+/– mice (Fig. 7, B and C), which correlates with the decrease in progesterone levels (Fig. 5B). Interestingly, no statistically significant difference was seen in the level of STAR expression in Nr5a2+/– ovaries of females undergoing decidual resorption at 6.0 dpc (data not shown). Collectively, these results suggest that although STAR may participate in Nr5a2-induced progesterone production, it is not the only factor involved, at least during natural pregnancy.
|
In an attempt to elucidate the potential role of NR5A2 in reproductive functions in vivo, we used Nr5a2 haploinsufficient mice to perform different crosses and showed that heterozygosity for Nr5a2 leads to a significant reduction in female fertility with pregnancy loss occuring prior to 6 dpc. This finding, along with the well-established ovarian expression pattern of NR5A2 in granulosa cells and luteal cells, suggests that NR5A2 might be involved in folliculogenesis or corpus luteum formation/function in vivo. To examine this, we first performed a morphological analysis and showed that Nr5a2+/– ovaries do not display obvious defects in response to hormonal stimulation. The presence of corpora lutea in Nr5a2+/– ovaries, the similar number of oocytes released, the similar number of implantation sites, and the similar litter sizes between wild-type and Nr5a2+/– animals all demonstrated that ovulation occurred in Nr5a2+/– females. Furthermore, serum levels of estrogen in heterozygous females were similar to those of wild-type mice in the absence or presence of hormonal stimulation. In addition, the presence of decidual resorption in Nr5a2+/– mice undergoing pregnancy loss also demonstrates that implantation occurs in these females. Collectively, these results indicate that the reduction in fertility observed in Nr5a2+/– females is not a consequence of impaired follicular development or corpus luteum formation, but rather arises from defects in maintaining pregnancy, suggesting that NR5A2 might play a role in luteal function during early pregnancy.
Proper corpus luteum maturation and function are central to the establishment and maintenance of pregnancy. The primary role of the corpus luteum is to produce high levels of progesterone to support pregnancy. We demonstrated that Nr5a2+/– females displayed significantly lower progesterone levels following hormonal stimulation and during natural pregnancy compared with their wild-type littermates. Interestingly, we found that progesterone production in response to hormonal stimulation may be affected in a cell-specific manner in Nr5a2+/– mice. Progesterone synthesis was reduced in Nr5a2+/– female following eCG stimulation but when eCG stimulation was followed by hCG treatment, circulating progesterone reached levels similar to those of wild-type animals. Since eCG exerts predominantly an FSH-like activity in rodents [27], the stimulation of steroidogenesis appears to be limited to granulosa cells, which are the only FSH target cells in the ovary. On the other hand, hCG stimulates steroidogenesis in both granulosa and theca cells. The latter do not express Nr5a2 [2, 16]. This observation suggests that NR5A2 is involved in the regulation of progesterone synthesis in granulosa cells, but not in theca cells during the hormonally regulated periovulation period and that NR5A2-independent regulation of progesterone production in theca cells following hCG stimulation may compensate for the decreased progesterone production in Nr5a2+/– ovaries. One likely factor to regulate progesterone biosynthesis in theca cells is the NR5A2 closely related member NR5A1, which is expressed in these cells [2, 16]. The decreased progesterone production in response to gonadotropin stimulation in Nr5a2+/– mice is also supported by an impaired uterine response. Stimulation by eCG induced an increase in uterine weight, which results from estrogen-dependent water imbibition and increase in cell proliferation. Progesterone is known to counteract and limit the effects of estrogen on the uterus. In Nr5a2+/– females, however, the increase in uterine weight in response to hormonal stimulation was greater than that observed in wild-type females, which is consistent with a reduction in progesterone production.
Our results clearly show that progesterone production is dramatically reduced during luteinization and natural pregnancy in Nr5a2+/– mice compared with their wild-type littermates and suggest that the reduction in ovarian progesterone production observed in Nr5a2+/– mice was responsible for their decreased fertility. To pursue this possibility, we performed a progesterone supplementation experiment and showed that administration of a progesterone implant at the beginning of pregnancy rescues the fertility defects observed in Nr5a2+/– females.
As a first step in elucidating the molecular mechanisms by which NR5A2 regulates ovarian progesterone production, we evaluated the level of ovarian STAR expression in Nr5a2+/– females. STAR regulates one of the rate-limiting steps of progesterone synthesis and its gene was shown to be a NR5A2 target [18, 21]. We showed that, despite the reduction in progesterone production observed in Nr5a2+/– pregnant females, there was no difference in the level of ovarian STAR expression, suggesting that if STAR is involved in mediating NR5A2-induced progesterone synthesis during pregnancy, it is not the only factor. Of particular interest, NR5A2 was shown to regulate the expression of genes coding for the two progesterone-synthesizing enzymes, CYP11A1 and 3ß-HSD, and for the HDL receptor SR-B1, which is required for the cellular uptake of the steroid precursor, cholesterol [1, 18, 22–24]. It is therefore possible that transcriptional activation of a combination of these genes is responsible for the NR5A2-induced progesterone production in vivo. The absence of difference in STAR expression between wild-type and Nr5a2+/– ovaries during natural pregnancy, however, does not completely exclude the involvement of STAR in NR5A2-mediated progesterone production. Indeed, during natural pregnancy, the ovary is composed of a heterogeneous population of cells, in which STAR protein expressed by follicular theca and interstitial cells could mask the reduction in STAR expression from granulosa cells and their derivatives, the granulosa lutein cells of the corpus luteum. To test for this possibility, we hormonally induced luteinization in wild-type and Nr5a2+/– immature females to obtain a more homogenous ovarian cell population, mainly composed of established corpora lutea. Our results show that STAR expression is significantly diminished in Nr5a2+/– luteinized ovaries, indicating that regulation of STAR expression participates in NR5A2-mediated progesterone production during luteinization.
The high ovarian expression levels of Nr5a2 and its closely related member Nr5a1, which shares the same consensus DNA binding site, has led to an extensive debate on their respective roles in the ovary. These two NR5A members, however, exhibit a distinct ovarian expression pattern. Nr5a2 was shown to be expressed in follicular granulosa cells, in newly formed corpora lutea, as well as in mature corpora lutea throughout pregnancy. Conversely, Nr5a1 was shown to be expressed in follicular theca and granulosa cells but absent from corpora lutea of cycling and pregnant females [2, 16, 17]. The distinct ovarian pattern of expression of Nr5a1 and Nr5a2 therefore suggests that they may play different roles in the ovary, in part through functional compartmentalization. Although NR5A2 was originally hypothesized to be involved in ovarian estrogen production [16, 17, 19], a few recent studies suggest a predominant role for NR5A2 in ovarian progesterone production [18, 21–24]. For instance, Saxena et al. [22] provided a direct role for NR5A2 in progesterone biosynthesis, but not estrogen production, during granulosa cell differentiation. In addition, a recent report demonstrated that NR5A1, and not NR5A2, is the endogenous species binding to the Cyp19 promoter in rodent granulosa cells, which suggests that NR5A1 is the most important factor regulating estrogen synthesis in granulosa cells [2]. In agreement with this view, it was shown that a granulosa cell-specific knockout of Nr5a1 leads to sterility accompanying an ovarian phenotype resembling the estrogen receptor (Er) and Cyp19 knockout phenotypes [28]. If NR5A2 was important for estrogen biosynthesis in granulosa cells, we would expect not to see this drastic granulosa cell-specific Nr5a1 knockout phenotype. Together, these findings suggest that NR5A1 and NR5A2 perform distinct functions in the ovary, NR5A1 being important for estrogen synthesis and NR5A2 being predominantly implicated in progesterone production. The results presented herein are in agreement with this idea and clearly identify a role for NR5A2 in murine ovarian progesterone production in vivo.
In conclusion, this study provides the first in vivo evidence for a requirement of NR5A2 in ovarian function. More specifically, we demonstrate that heterozygosity for Nr5a2 leads to a reduction in female reproductive ability and that this decreased fertility results from impaired progesterone production. Therefore, based on our results, we propose that NR5A2 is a regulator of corpus luteum maturation and function and that although a single copy is sufficient to initiate its formation, its complete absence should lead to a more severe phenotype. If this is correct, an ovarian-targeted deletion of Nr5a2 should result in defective corpus luteum formation.
ACKNOWLEDGMENTS
We thank Dr. B. Murphy and Mira Dobias-Goff for performing the progesterone RIAs, Dr. D. M. Stocco for providing us with the anti-peptide STAR antibody, Miren Gratton for performing the ELISA assay for estrogen and progesterone measurements, and Kevin Ebata for helpful discussion.
FOOTNOTES
1Supported by the Canadian Institute of Health Research. D.D. is a Chercheur Boursier du Fonds de la Recherche en Santé du Québec (FRSQ). 
Correspondence: 2Daniel Dufort, Royal Victoria Hospital, 687 Pine Avenue West, Rm F3-24, Montreal, QC, Canada, H3A 1A1. FAX: 514 843 1662; e-mail: daniel.dufort{at}mcgill.ca
Received: 21 November 2006.
First decision: 8 December 2006.
Accepted: 3 April 2007.
REFERENCES
in the developing murine liver intestine and pancreas. Mech Dev 1999; 89:185–188[CrossRef][Medline]1-fetoprotein transcription factor in bile acid biosynthesis. A known nuclear receptor activator that can act as a suppressor of bile acid biosynthesis. J Biol Chem 2004; 279:16813–16821-hydroxysteroid dehydrogenase type II. J Clin Endocrinol Metab 2003; 88:6020–6028This article has been cited by other articles:
|
|
 |
|
  R. Duggavathi, D. H. Volle, C. Mataki, M. C. Antal, N. Messaddeq, J. Auwerx, B. D. Murphy, and K. Schoonjans Liver receptor homolog 1 is essential for ovulation Genes & Dev., July 15, 2008; 22(14): 1871 - 1876. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
