| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Ovary |
Molecular Reproduction Research Laboratory,3 Clinical Research Institute of Montreal, Montréal, Québec, Canada H2W 1R7
Department of Physiology,4 McGill University, Montréal, Québec, Canada H3G 1Y6
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
androgen receptor, estradiol receptor, follicle-stimulating hormone receptor, granulosa cells, ovary
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
A major cause of infertility in a significantly increasing number of women of reproductive age is the premature depletion of the follicles from the ovary, leading to a condition called premature ovarian failure. Therefore, the multitude of factors that control the initiation of primordial follicle development and their original pool size ultimately determine reproductive fitness and the age of menopausal transition in women. Developmental factors involved in the recruitment of primordial follicles into the growing pool are not yet fully understood. Among several factors recently suggested to be involved in this process are Müllerian inhibiting substance (MIS), a member of the transforming growth factor (TGF) ß family of signaling molecules implicated in the initiation of primordial follicle growth [5, 6]; bcl-2, an antiapoptotic gene that prevents loss of primordial follicles [7]; and basic fibroblast growth factor (bFGF) [8]. In addition, oocyte-derived entities, such as the transcription factor found in germ cell alpha (FIGalpha) [9], the growth differentiation factor-9 (GDF-9) [10], and the c-kit receptor [11], are also believed to contribute to this process.
The beginning of differentiation of ovarian somatic cells into granulosa cells (GC) that surround the oocyte and form part of the first cohort of primordial follicles heralds folliculogenesis, a process that requires about 3 wk for completion in the mouse. Follicular development in the ovary is critically dependent on the action of FSH, a pituitary glycoprotein hormone that signals through its receptor repertoire on GCs [12, 13]. However, its role in the early stages of differentiation remains to be clarified. Recently, a role for FSH has been suggested in the initiation of growth of the nongrowing pool of follicles in the hamster [14]. Although data in the mouse remain controversial, it is generally believed that this initial event can proceed without gonadotropic support [15]. Indeed, some studies failed to observe the effects of these hormones on small, nongrowing follicles [16], but age selection might be a key factor in determining responsiveness [17]. The appreciable amount of circulating gonadotropin in the perinatal mouse suggests that the hormone is present for a developmental reason [15, 18]. Moreover, the mRNA of the FSH receptor (FSH-R) that selectively mediates the hormone signal has been detected in the ovaries of 1-day-old mice [19], a time at which the rodent ovary contains only primordial follicles. Thus, alternative experimental paradigms might help shed light on such important questions that ultimately determine the beginning and end of the reproductive phase in the female.
In this context, our recently described Follitropin Receptor Knockout (FORKO) mouse [12, 20] generated by deleting all variants of the FSH-R provides additional insights into hormone signaling-related processes in the ovary. This model becomes particularly interesting in light of the diverse early developmental as well as age-dependent abnormalities found in the adult animals, such as infertility/reproductive senescence, hormonal imbalance, obesity, and increased tumor incidences [21]. Although ovarian development proceeds to some extent in the null mouse as seen in adult stages [12, 20], no quantitative analysis of perinatal status had been performed until this study. Thus, our objective was to determine the physiological and molecular aberrations that occur during ovarian development in perinatal FORKO females. Herein, our study reveals that FSH-R signaling is involved very early in the process of formation and recruitment of the primordial follicles into the growing pool.
|
MATERIALS AND METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
The studies described in this article were performed according to the accepted guidelines of the animal care committee of the institution. The FORKO mice were established as described [12, 20] by breeding 129T2/SV EmsJ FSH-R +/- male and females of 3–5 mo. These breeding pairs provided littermate +/+ and -/- females that would allow a direct comparison. All mice were fed a standard laboratory chow (5001/Harland Teklad S-2335 diet 1:1 mixture, Madison, WI) and were maintained under a regimen of 12L:12D. Two-, 10-, and 24-day-old mutants and wild-type mice derived from heterozygous mothers were compared in each experiment. Genotyping by polymerase chain reaction (PCR) was performed on DNA isolated from tails or toes [20]. Mice were killed by decapitation, and ovarian tissue was preserved for histology [22].
Antibodies
Androgen receptor (AR) rabbit polyclonal IgG (N20), MIS goat polyclonal IgG (C-20), GATA-4 goat polyclonal IgG (C-20), and Cyclin D2 rabbit polyclonal IgG (M-20) were purchased from Santa Cruz Biotechnology (Santa Cruz, CA). Estrogen receptor ß (ERß) rabbit polyclonal IgG was the kind donation of Dr. Pierre Chambon (IGBMC, Strasbourg, France). Secondary biotinylated antibodies were included in the ImmunoCruz Staining System (Santa Cruz Biotechnology) that contained the goat anti-rabbit antibody for use with rabbit primary antibody. The ABC staining system (Santa Cruz Biotechnology) contained the donkey anti-goat antibody for use with the goat primary antibodies.
Radioimmunoassay of Gonadotropins and Steroids
Plasma LH levels were determined by RIA as described [21]. All samples were assayed in a single assay with sensitivity of 0.04 ng/ml (standard AFP-5306A). 17ß-Estradiol and testosterone levels were estimated by solid-phase RIAs [20] using commercially available Coat-a-count kits (Diagnostic Products, Los Angeles, CA), and manufacturer's instructions were followed.
Immunohistochemistry
Five-micrometer-thick ovarian sections were deparaffinized and rehydrated using an alcohol gradient. Tissue sections were subjected to a microwave antigen retrieval technique as described [21]. Briefly, after boiling the sections in 10 mM citrate buffer (pH 5.5–6.0) for 30 min, incubation of the cooled sections in 1% Triton-X 100 (EM Science, Gibbstown, NJ) in PBS followed. Sections were then processed for immunostaining using the ImmunoCruz or ABC Staining System as appropriate, following the supplier's instructions. Sections were incubated with the following specific antisera: goat-anti MIS, goat-anti GATA-4, rabbit-anti Cyclin-D2, rabbit-anti AR, and rabbit-anti ERß at a dilution of 1:100 for three nights (except for anti AR, which was incubated overnight) at 4°C. In negative controls, normal serum was substituted for primary antibody in the first reaction. The sections were washed and incubated in biotinylated secondary antibodies for 30 min at room temperature, followed by washing in PBS and incubation in avidin-biotin-horseradish peroxidase for 30 min. Reactions were visualized with 3,3'-diaminobenzidine tetrahydrochloride dihydrate and weakly counterstained with hematoxylin.
Morphology of the Ovaries
Ovaries, uteri, and vaginas were collected from the animals and wet weight was taken. The tissue was fixed in 10% formalin overnight and embedded in paraffin. Five-micrometer-thick sections were stained with hematoxylin and eosin for light microscopic histology. Follicles were classified separately for the neonatal 2-day-old ovaries and for the older (10- and 24-day) ovaries. In the newborn mice, follicle classification was performed as described [23]: oocytes that were not surrounded by somatic cells were classified as naked oocytes; oocytes with a single full layer of flattened GC were termed primordial; those follicles that were between these two stages having an incomplete ring of squamous GC were termed intermediate. In the older animals, follicle classification was done according to previously described terms [24]. A primary follicle has one layer of cuboidal GC; we have included in this group also those follicles that contained a mixed layer of flattened and cuboidal GC because these belong to the growing pool of follicles in rodents [25]; a secondary follicle is comprised of an oocyte surrounded by two layers of GC; a preantral follicle was distinguished by the presence of more than two layers of GC, with the follicle having no antrum; while the antral follicles were the larger ones that had a fluid-filled cavity known as the antrum. Immature mice do not have ovulatory follicles and therefore contain no corpora lutea, differentiated structures that are formed only after ovulation. Follicles were included in the count if the section passed through the nucleolus of the oocyte, and were counted in every second section as described previously [6]. A correction factor of two was applied when calculating the number of naked oocytes due to the very small diameter of these germ cells. Follicle diameter was also measured in the largest cross-section (the section that clearly showed the nucleolus) by taking the average of the two perpendicular diameters of each follicle. Sections were evaluated and photographed using a Carl Zeiss microscope (Jena, Germany) and computer-aided Eclipse image analyzer (Northern Eclipse, Ontario, Canada). At each age, ovaries from four to six different animals were used for counting.
Statistical Analysis
Data are presented as the mean ± SEM and were analyzed by Student t-test or ANOVA with a Fischer least square difference (LSD) post hoc test using P 
0.05 as the level of significance.
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
At 2 days of age, the body weights of female littermates were similar (1.45–1.78 g); however, the FORKO females were smaller by Day 10. At this age, wild-type mice weighed 5.86 ± 0.19 g while the FORKO weighed 4.98 ± 0.18 g, a difference of 10%, which was statistically significant (P < 0.01). This difference was also maintained at 24 days, when FORKO females weighed 9.71 ± 0.22 g and age-matched wild-type animals weighed 10.67 ± 0.22 g (P < 0.01) (Fig. 1A). However, it should be noted that, as reported earlier, metabolic alterations in the adult null female induce obesity by 3–5 mo [20].
|
Estrogen is an important modulator of the structure and function of the reproductive tissues. Hence, the ovaries, uteri, and vaginas were weighed for animals killed at 24 days of age, but not for the younger animals due to difficulties in dissecting and cleaning the tiny tissues. Remarkably, there was a significant atrophy of these steroid hormone-dependent reproductive organs at this early age when the normal females are still sexually immature. Ovarian weight was lower in FORKO females (3.59 ± 0.2 mg) compared with the wild types (6.55 ± 0.65 mg), representing a 45% difference in ovarian weight (P < 0.001) (Fig. 1B). Uterine weight was already 43% lower at 24 days of age in the FORKOs (8.08 ± 0.37 mg in the FORKO vs. 14.12 ± 1.32 mg in the wild-type littermates, Fig. 1B) (P < 0.005). Finally, vaginal weight in FORKO mice was also 47% lower (P < 0.05), indicating that atrophy has already begun at this early age.
Circulating Hormone Levels
Plasma estradiol was undetectable in the FORKO mice at 24 days of age, while its level was 1.6 pg/ml in the wild-type animals (Fig. 1C). The limit of detection of the assay for this hormone was 1.4 pg/ml. Testosterone levels were significantly elevated by 46% in the FORKO females (22.6 ± 4.1 ng/dl) compared with the normal animals (15.5 ± 1.8 ng/dl) (P < 0.05) (Fig. 1D). This hormonal imbalance shows an increasing trend similar to that seen in the adult 3-mo-old mutant animals, where levels are about 10 times higher than normal [20]. The level of LH in the plasma of FORKO females was significantly higher (2.42 ± 0.2 ng/ml) than that of the wild-type females (0.8 ± 0.1 ng/ml) (P < 0.005) (Fig. 1E). This is again indicative of a disrupted negative feedback mechanism(s) at the level of the hypothalamic-pituitary-ovarian axis in the postnatal period.
Morphology of the Ovaries During Perinatal/Postnatal Development
A careful histological assessment was performed to understand the impact of loss of the FSH-R signaling during early postnatal development of the ovary. Ovaries of 2-day-old wild-type animals contained a large number of naked oocytes, intermediate and primordial follicles, with only the occasional appearance of primary follicles (Fig. 2A). The most striking difference between wild-type and FORKO ovaries was the presence of secondary follicles in the FORKO ovary (Fig. 2B, arrow), while wild-type ovaries contained no follicles of this type. This is an indication of advanced follicular development present in the FORKO mice. Another notable difference is the significantly smaller size of the 2-day-old FORKO ovary (Fig. 2B, inset). At 10 days of age, there was thickening of the interstitial tissue, indicating hypertrophy (Fig. 2D, arrow) in addition to other changing dynamics (see below) of the ovaries in the FORKO mice. At 24 days of age, young FORKO females failed to develop antral follicles (Fig. 2F, inset), while their wild-type littermates that are still immature at this age showed proper antral follicle formation (Fig. 2E, inset, arrow). A second feature of the FORKO ovaries at this age was an apparent thickening of the surface epithelial layer (Fig. 2F, arrowhead), which in the wild-type animals appeared thin (Fig. 2E, arrowhead). Besides the overall reduced size of the FORKO ovaries (Fig. 2F, inset), another characteristic aberration was the infiltration of GC into the oocytes (Fig. 2F, arrows). Additional observations pertaining to this oocyte/GC abnormality are included in the accompanying communication [26].
|
Follicular Dynamics in Development
An examination of the different types of follicles reveals changing patterns during development as the animals progress toward the pubertal period around 4 wk of age. At this time, they normally begin to show enhanced responsiveness to hormonal and other factors. In this evaluation of the ovarian structures, we have broadly divided the follicles into two categories, namely the resting pool and the growing pool, as depicted in Figure 3. There was an age-dependent change in the number of follicles in the FORKO ovary relative to the normally developing structures in the wild type. The 2-day-old FORKO ovary contained fewer follicles in the nongrowing pool, with 50% fewer naked oocytes (1877 ± 265 in FORKOs and 3627 ± 240 in +/+ animals, P < 0.005), but the number of intermediate and primordial follicles was not significantly different from those found in the wild-type animals. However, the number of growing follicles (primary and secondary) was higher in the FORKOs. More important, secondary follicles that were completely absent in the wild type at 2 days had already made their appearance in the FORKO ovaries. Thus, there seemed to be a spillover of follicles from the nongrowing pool into the growing pool in the FORKOs as early as 2 days (Fig. 3A). The total number of follicles was significantly smaller (P < 0.05) at this age in the FORKOs (2912 ± 430) compared with the wild type (5269 ± 673). At 10 days of age, the total number of follicles was significantly reduced in the FORKOs (3082 ± 287.6) compared with the wild types (4025 ± 187, P < 0.05). These differences were spread across different developing structures. First, there were lower numbers of primordial follicles (that belong to the nongrowing pool) in the null-mutants (2868 ± 291) compared with the wild-type littermates (3704 ± 178, P = 0.05). The number of growing follicles was also smaller in the FORKOs than the wild types (Fig. 3B). Specifically, there were 20% fewer primary follicles in the FORKOs (145 ± 10.8) compared with the wild-type animals (181 ± 10, P < 0.05); there were 60% fewer preantral follicles in the mutant animals (28 ± 1.8) compared with the wild-type animals (70 ± 5.5, P < 0.001). FORKO ovaries at this age were completely devoid of antral follicles, while age-matched wild-type ovaries contained on average 30 ± 5.5 such mature follicles/ovary. Finally, by 24 days of age, the number of follicles in the resting pool became larger in the FORKOs (1791 ± 122) than in the wild-type animals (1400 ± 122, P < 0.05) indicating that there was an arrest in the recruitment of these follicles into the growing pool (Fig. 3C). The large drop (50%) in the primordial follicle number between Days 10 and 24 in the wild-type animals is part of normal ovarian development [7]. There were fewer growing follicles in the FORKOs, where the number of secondary follicles was 27% of that found in the normal animals (P < 0.001). However, the total number of follicles was not significantly different across the two genotypes (1904 ± 134 in +/+ and 2147 ± 122 in -/-) at this age.
|
Follicle Diameter
The diameter of the follicle provides a direct assessment of the overall influence of growth factors and signaling molecules in coordinating cell proliferation and fluid secretion in the envelope that contains the ovum. Follicular diameter was measured only in 24-day-old ovaries because this was the time when a very large difference was noted in the ovarian weight between the two genotypes. There was a significant difference in the diameter of primary follicles between the two genotypes at 24 days of age (results not shown). The diameter of primary follicles was on average 20%–25% lower in the FORKO females (24 ± 2.5 µm) compared with the diameter of follicles in the wild-type animals (31 ± 2 µm) (P < 0.01). As shown in the accompanying study [26], this is partly attributable to 25% reduction in oocytes in the mutant animals. However, the number of apoptotic GCs was not counted; therefore, a decrease in this number or smaller size of the GCs could also contribute to the smaller size of the follicles.
The diameter of the antral follicles in the wild type was 200 ± 6.5 µm, but this structure is absent in the FORKO ovary. There was no significant difference in the diameter of the other types of follicles (primordial, secondary, and preantral) between the two genotypes.
Immunohistochemistry of Developmental Determinants: MIS Localization
Determining the expression of known markers of differentiation in the ovary is helpful in understanding the severity of developmental aberrations. As MIS expression was previously detected in the postnatal ovary in the GC of growing follicles [5, 6], we assessed it at different ages in our mutants. MIS was found in the GC of preantral follicles in both wild-type and FORKO ovaries but to different extents. Interestingly, at 2 days of age, expression was seen in the oocytes of small follicles in FORKOs, but no staining of the GC was observed at this age (Fig. 4D). There was no expression of MIS in the theca cells or interstitial cells at all ages in both genotypes. MIS appeared to be upregulated in the ovaries of null mutant animals (Fig. 4, E and F vs. B and C) in all ages examined. At 10 days, oocyte MIS expression was still higher in the FORKO ovary but it was minimal (or negative) in the wild type (compare Fig. 4, E vs. B). MIS upregulation in GC at 24 days was obvious (Fig. 4F). On the other hand, FORKO oocytes at this age showed downregulation of MIS expression. MIS expression in the oocytes was observed only in the FORKO mice. Because the negative controls of the same age showed no staining, the atypical expression could be due to either abnormality or defect of the oocytes or due to abnormal localization of the protein in the FORKO mice. As reported elsewhere [21], such early perturbations in MIS expression may be associated with the subsequent development of ovarian tumors in mutants.
|
GATA-4 Localization
The expression of GATA-4 protein, a transcription factor, was previously reported in the nucleus of GC of growing follicles in the human fetal ovary [27] and in the mouse at different stages of development [28]. We detected the protein in the developing ovaries from the earliest ages examined. Nuclear expression was observed in the somatic cells surrounding the germ cells in the 2-day-old +/+ ovaries and in the GC of follicles belonging to the nongrowing pool. The expression of this protein was almost undetectable in the FORKO ovaries (Fig. 4, J–L) at all ages examined, although in the wild-type animals, its expression was obvious at later ages (10 and 24 days).
AR Expression
In the normal ovary, androgens produced by the thecal cells that surround the follicle have effects on the GC within the encapsulated structure in two ways. First, by exerting a direct action on the GCs and second by becoming a substrate for conversion to the female sex hormone 17ß-estradiol, under the catalytic action of aromatase. As androgen functions through the nuclear receptor, we have examined the androgen receptor during development. AR immunostaining occurred in GC nuclei of growing follicles in both genotypes, with no staining in thecal and stromal cells. This pattern of localization was the same as previously seen in immature rodents, where GCs of preantral follicles were highly positive for AR [29]. Weak staining of the oocytes was observed only in the 2-day-old neonates. Positive staining evident in follicles with only one to two granulosa cell layers [30] was also seen in the FORKO immature mice (Fig. 5). There was a marked upregulation of this protein in follicles of 10-day-old FORKO mice (Fig. 5E) compared with the wild-type animals (Fig. 5B) at this age, where immunostaining was either weak or undetectable. The expression of AR at 24 days (Fig. 5F) was weaker than the expression at 10 days in the FORKO ovary; however, it was still more strongly expressed than in the wild-type follicles (Fig. 5C) at this age.
|
ERß Localization
Of the two nuclear estrogen receptors 
and ß that are currently well known, the ERß form is more predominantly expressed in the GC layers of the ovary. 17ß-Estradiol could function in a paracrine or autocrine manner to regulate follicular growth and oocyte characteristics. In agreement with recent reports [29], nuclear ERß staining was observed in GC of follicles at different stages (Fig. 5). Thecal cells, interstitial cells in both wild-type and FORKO ovary were negative for ERß immunoreactivity (Fig. 5, J–L). At 2 and 10 days of age, oocytes in the FORKO primordial as well as primary follicles showed significant ERß staining (Fig. 5, J and K, arrowhead). Staining of comparable structures in the wild type was weak at this age (Fig. 5, G and H, arrowhead). However, oocyte ERß expression was lost by 24 days in the FORKO ovary. At this age and at 10 days of age, GC staining for ERß became stronger in the wild type (Fig. 5, H and I, arrows). There was more expression of ERß in the 2-day-old FORKO ovaries probably because these ovaries contained follicles of a more advanced stage (Fig. 5, J, inset).
Cyclin D2 Expression
The proliferation marker Cyclin D2 that has been linked to GC maturation and function was evaluated to assess the cellular status in immature mice. Immunohistochemical detection of this cell cycle protein was positive in the ovaries of both mutant and wild-type animals at all three ages (Fig. 6). The nuclear staining was very scattered in the 2-day-old ovaries (Fig. 6, A and B). Expression in the FORKO ovary appeared to be downregulated (Fig. 6, D–F) compared with the normal ovaries (Fig. 6, A–C) at all ages examined.
|
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
While FSH, through its interaction with the receptor, is clearly essential for later stages of follicular development [12, 13], its role in the formation of primordial follicles or their initiation to growth was uncertain (reviewed in [3]). This difference could only be discerned by careful age-related comparisons as performed in the present study. Based on these data, we propose that FSH interaction with its receptor is indeed required for the early transition of follicles from the resting pool into the growing pool. Rather than being an all or none phenomenon, this signaling plays an important modulatory role, setting the pace of ovarian growth. Because there are fewer naked oocytes in the FORKO ovary at 2 days of age, it seems that proper formation of the germ cells is disrupted in the absence of FSH-R signaling. Because we compared female pups of different genotypes born to the same mother (with FSH-R +/- genotype), it is reasonable to suppose that they were subject to identical maternal influences during gestation and lactation, except for differences in their own internal endocrine interactions in utero as well as the postnatal life. Using an alternative strategy of selective antibody neutralization of circulating FSH during the final stages of gestation and soon after birth, Roy and Albee [14] also demonstrated a role for the hormone in early follicular development in the hamster. Our interpretation using genetic FSH-R disruption studies noted here are in agreement with these data. Several observations, such as the presence of mRNA transcripts for the FSH-R in the day 1 mouse ovary [19] or in the day 3 rat ovary [32] or in unfertilized mouse oocytes [33], a time at which only resting follicles are present, implies that these target cells are indeed prepared to respond to the endogenous hormone.
In their outward appearance, the FORKO mutants were similar to the wild types except that their body weights were lower at 10 and 24 days of age (Fig. 1A). As reported earlier, the cumulative metabolic effects of estrogen deficiency in the FORKO females leads to obesity by 3 mo [20], a condition that persists until about 12–15 mo, when the animals develop cachexia consequent to ovarian tumors [21].
Although the overall development of the urogenital system appears to be normal in the FORKO females [12, 20], there are significant functional deficits reducing the number of germ cells, as observed in the present study. During normal development, primordial follicles are depleted from the resting pool through initial recruitment by entering the growing pool, attaining the secondary stage by 10–12 days [2]. This process appears to be accelerated in the FORKO ovary (Fig. 3), as there were more growing follicles on day 2. This suggests that the FORKO oocytes might have been triggered to secrete growth factors prematurely to affect the neighboring GCs [31]. However, by Day 24, as the animal approaches the normal stage of puberty (7 wk), there seems to be an arrest in the initial recruitment, with the 10-day ovaries being in the transition period. Therefore, we searched for probable causes and/or alterations in regulators of initial recruitment.
Altered FSH levels are of no consequence in the FORKO mouse as the receptor repertoire is absent. However, higher LH levels (Fig. 1E) could influence the maturation process. Gonadotropin levels in the mouse rise after birth, with FSH level peaking on Day 16 and peak LH level on Day 9 [34]. High LH is known to deplete the primordial follicle pool by enhancing recruitment [35]. LH levels were elevated in FORKO females at 22 days of age. As we could not measure levels prior to that due to technical limitations, we can only speculate that they might also be high even earlier, perhaps right after birth. In transgenic mice with high LH, there are 45% fewer primordial follicles by 5 wk of age [35]. It is of interest to note that, unlike the FORKO ovaries, which are already abnormal by 3 wk, the ovaries of LH transgenic mice remain completely normal at 3 wk of age. This suggests that the high level of LH might not be the only factor affecting the size of the resting pool of follicles in the FORKO mutants.
Accelerated folliculogenesis noted in the neonatal FORKO could also be caused by enhanced ovarian androgen secreted under the influence of high LH (Fig. 1) [36]. In addition, undetectable estrogen in the FORKO is perhaps due to the lack of activity of the P450 aromatase enzyme, which is under FSH regulation in GC of the ovary [37]. Testosterone is an apoptotic factor for follicular cells while estrogen is an antiapoptotic factor [38]. As both are altered in young FORKO females, perhaps the ratio of these two ovarian steroids determines the fate of the cells toward growth or apoptosis. However, a direct correlation to apoptosis in the developmental stages is premature because atresia seen in our mice does not seem to follow apoptosis in the FSH-R +/- animals [22].
To gain additional mechanistic insights, we focused on several known factors that are implicated in early ovarian development. One factor that has recently been suggested to be involved in the recruitment of primordial follicles is Müllerian inhibiting substance (MIS) or Anti-Müllerian Hormone (AMH) [5, 6]. MIS is a dimeric glycoprotein of the growing transforming growth factor-ß (TGFß) family. During fetal life, MIS causes regression of the Müllerian ducts in the male [39]; however, it is also present in the postnatal ovary [40] where its expression is evident in the GC of growing follicles. MIS inhibits initial recruitment, both in vivo [5] and in vitro [6]. Thus, in MIS-KO mice, the pool of primordial follicles decreases faster than in the normal animals [5] because of a shift in recruitment into the growing pool.
Because MIS is a protein under the negative regulation of FSH and estradiol [41], understanding the pattern of its expression in the FORKO ovaries is of interest. Indeed, the upregulated expression of MIS in the follicles of the mutant mice as compared with the wild-type pups might cause stagnation in the entry of the nongrowing follicles into the growing pool. In addition, as mice chronically overexpressing a human MIS transgene have fewer germ cell numbers [42], such an event in the FORKO ovary could also contribute to the lower number of follicles and estradiol production. Additional evidence for the lack of estradiol in immature FORKO females is the significant reduction in growth of the uterus at 3 wk of age in comparison with the wild type (Fig. 1B). The presence of circulating 17ß-estradiol, albeit at a low level in the normal animal, is physiologically important, as confirmed by expression of lactoferrin (an estrogen-regulated gene) in the immature wild type and its absence in the FORKO uterus (unpublished results).
As MIS-producing GC were found to surround only mitotically arrested germ cells [39], the high expression of MIS in the FORKO ovaries could imply the presence of a larger number of arrested oocytes as compared with wild-type females. Alternatively, because later germ cell meiosis coincides with decreased expression of MIS in normal animals [39], an oocyte abnormality could be inferred from the lack of downregulation of this protein in the mutant ovaries. Indeed, high expression of MIS observed as early as 2 days in the mutant oocytes but not in the wild type lend support to this conclusion (Fig. 4). Therefore, communication between the oocyte and the surrounding GC is rendered defective at a very early stage. Accordingly, the potential indications of abnormality in FORKO oocytes are examined in more detail in the accompanying article [26].
The expression of GATA-4, which is a developmentally regulated transcription factor, was also different in the young FORKO ovary (Fig. 4, G–L). GATA-4 is an evolutionarily conserved transcription factor that is abundantly expressed in the developing gonads, with expression pattern closely paralleling that of the MIS hormone [43]. In fact, GATA-4 is involved in the regulation of many ovarian genes through the GATA-binding domain found on the promoter element of genes, such as MIS. In addition, exogenous gonadotropins and estrogens regulate GATA-4 transcript levels (FSH and estradiol positively, LH negatively) in murine ovaries in vivo [44]. The diminished GATA-4 expression in the FORKO mice ovaries at all ages examined is consistent with these observations. It is interesting to note that this pattern of reduced GATA-4 expression is also prevalent in Finnish women with the inactivating mutation of Ala189Val in the FSH-R [45]. Perhaps the lower expression of GATA-4, which is a transcriptional regulator of MIS [43], leads to pronounced expression of the latter in the FORKOs. In conjunction with structural changes (Fig. 2) and diminished GATA-4 expression, the early and enhanced MIS expression could be the precursor of ovarian tumorigenesis that later appears in aging FORKO mice [21].
The action of androgen is at least partially regulated at the receptor level to provide optimum environment for growing follicles [46]. AR is expressed in greatest amounts in ovaries containing immature preantral/early antral follicles and declines throughout the preovulatory follicular development in response to FSH stimulation [46]. Its expression patterns in the FORKO immature females can be related to other models such as the hypophysectomized rat, whose ovaries stained highly positive for AR at 27 days of age [30]. In addition, the inappropriately high level of AR seen in the ERß knockout mice is apparently related to the follicular atresia and the early exhaustion of follicles [29]. Our observation of weak AR expression in the oocytes of primordial follicles at the neonatal age (Fig. 5, A and D) might directly contribute to the recruitment of primordial follicles into the growing pool. As we did not detect the protein on the oocytes of any type of follicle in the older animals, it is also possible that oocyte AR expression is confined to perinatal stages with downregulation in later development.
In the normal developing ovary as well as in the adult, estradiol that is produced by the granulosa cell under FSH-R signaling acts locally in an autocrine/paracrine mode and exerts peripheral actions on numerous targets via the ER [47]. As ERß is expressed predominantly in GCs, we evaluated its expression in the ovary of young FORKO females. The generally weaker expression of ERß in the GCs could be caused by high androgen and LH, both of which are known to downregulate ERß in other animal models [29, 48].
Our findings on Cyclin D2 during ovarian development are important for several reasons. We report here for the first time that this regulator is present in the 2-day-old normal ovary, a stage at which the cells already have a functional FSH-R signaling system. In addition, these data are consistent with our own previous findings in older FORKO mice [12], where we noted a reduction in Cyclin D2 mRNA. Cyclin D2-deficient females are sterile due to the inability of the GC to proliferate normally in response to FSH; in normal GCs, this regulator is specifically induced by FSH [49]. The more severe reduction of Cyclin D2 protein in the FORKO ovary as compared with only slight reduction in its mRNA in the adult FSHß knockout ovary [50] might mean that FSH-R signaling could also be exerting a translational control of protein expression. In addition, we should also note that, unlike in the FORKO females, the normal 17ß-estradiol levels reported in FSHß knockout mice [13] could have maintained Cyclin D2 in these mice. Taken together with the established role of estrogen in Cyclin D2 regulation [51], our data show that impairing FSH-R signaling has more severe developmental implications.
Based on our data in the present study, we visualize the following scenario. At Day 2, FORKO females begin their lives with lower numbers of oocytes. This could be due to increased apoptosis by one of three mechanisms as recently suggested [52], indicating "death by neglect," "death by defect," or "death by self-sacrifice." In addition, it is possible that FORKOs have already elevated LH at birth (because this hormone is overproduced on Day 22). LH transgenic mice have been shown to exhibit accelerated recruitment of primordial follicles. Therefore, pronounced LH interaction with its receptor (which is already present at birth in the normal ovary [53] and presumably in our mutants), could lead to an accelerated recruitment of the primordial follicles in the FORKO mice. In another model, the Bcl-2 transgenic mice, the primordial follicle pool was significantly larger at birth, but by Day 30, the number of primordial follicles did not differ greatly from those in the wild type. The authors suggest that there is an underlying mechanism by which excess numbers of primordial follicles at birth are detected and removed from the ovary by adulthood in order to maintain follicular equilibrium [7]. Similarly, there could be a mechanism by which a smaller nongrowing follicular pool, such as in the FORKO mice, is detected and equilibrated by inhibition of initial recruitment, therefore leading to an increased number of primordial follicles at Day 24. We suggest that the mechanism by which this "catching up" occurs is regulated by FSH-R signaling. Through the loss of inhibition of MIS, initial recruitment is inhibited in the FORKO mice by Day 24.
A diagram outlining the role of important markers of follicular development investigated in this study, and the consequences resulting from the absence of the FSH-R signaling system, is depicted in Figure 7. In the normal animals, because FSH inhibits MIS expression (possibly through the regulation of GATA-4 or its cofactors), its stimulatory effect on aromatase function overrides the inhibitory effects of MIS on this enzyme. In addition, ERß, which mediates estradiol action in the GCs, and FSH are inhibitory to AR expression. In FORKO mice, MIS expression is not inhibited by FSH-R signaling (as GATA-4 is negligible, and perhaps other cofactors, such as FOG-1 or FOG-2 [54] are also differentially expressed); therefore, it is upregulated, inhibiting aromatase action. Absence of the FSH-R also renders aromatase nonfunctional [20], which leads to the absence of estradiol from the circulation of FORKOs. At the same time, elevated serum LH stimulates testosterone production, which accumulates due to lack of aromatization to estradiol. In addition, AR is upregulated in the ovaries of FORKOs because ERß expression is attenuated and FSH-R signaling is absent. A cumulative effect of the imbalance in these regulatory networks changes follicular dynamics in the FORKO ovary, resulting in infertility in the null [20] or reduced fertility in the haplo-insufficient FSH-R female mouse [22].
|
In conclusion, the results of the present study in mutant mice reveal that FSH-R signaling events are critical to proper ovarian development very early in the peri/postnatal period by modulating follicular dynamics. Thus, we propose that this receptor signaling system is involved in quantitative regulation of folliculogenesis and early differentiation of somatic cells. This could explain the presence of streak gonads in certain groups of women who are homozygous to an inactivating mutation of the FSH-R [55] or who have an aberrant FSHß gene that causes premature protein subunit termination [56]. Further investigations examining embryonic development in FSH-R +/- females that exhibit reduced fertility will be of interest in elucidating early alterations in the null mutants.
|
ACKNOWLEDGMENTS |
|---|
|
FOOTNOTES |
|---|
2 Correspondence: M. Ram Sairam, Molecular Reproduction Research Laboratory, Clinical Research Institute of Montréal, 110, avenue des Pins West, Montréal, PQ, Canada H2W 1R7. FAX: 514 987 5585; sairamm{at}ircm.qc.ca
Received: 17 January 2003.
First decision: 18 February 2003.
Accepted: 21 May 2003.
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
This article has been cited by other articles:
|
|
 |
|
  R. S. Viger, S. M. Guittot, M. Anttonen, D. B. Wilson, and M. Heikinheimo Role of the GATA Family of Transcription Factors in Endocrine Development, Function, and Disease Mol. Endocrinol., April 1, 2008; 22(4): 781 - 798. [Abstract] [Full Text] [PDF]
|
|
|
|
