help button home button Endocrine Society Molecular Endocrinology
HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS
Molecular Endocrinology, doi:10.1210/me.2003-0301
This Article
Right arrow Abstract Freely available
Right arrow Full Text (PDF)
Right arrow All Versions of this Article:
18/4/953    most recent
Author Manuscript (PDF)
Right arrow Purchase Article
Right arrow View Shopping Cart
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Right arrow Citation Map
Services
Right arrow Email this article to a friend
Right arrow Similar articles in this journal
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Download to citation manager
Right arrow Request Copyright Permission
Citing Articles
Right arrow Citing Articles via HighWire
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Jorgez, C. J.
Right arrow Articles by Matzuk, M. M.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Jorgez, C. J.
Right arrow Articles by Matzuk, M. M.
Molecular Endocrinology 18 (4): 953-967
Copyright © 2004 by The Endocrine Society

Granulosa Cell-Specific Inactivation of Follistatin Causes Female Fertility Defects

Carolina J. Jorgez, Michal Klysik, Soazik P. Jamin, Richard R. Behringer and Martin M. Matzuk

Program in Developmental Biology (C.J.J., R.R.B., M.M.M.), Departments of Pathology (M.K., M.M.M.), Molecular and Human Genetics (M.M.M.), and Molecular and Cellular Biology (M.M.M.), Baylor College of Medicine, and Department of Molecular Genetics (S.P.J., R.R.B.) University of Texas M.D. Anderson Cancer Center, Houston, Texas 77030

Address all correspondence and requests for reprints to: Martin M. Matzuk, M.D., Ph.D., The Stuart A. Wallace Chair and Professor, Department of Pathology, Baylor College of Medicine, One Baylor Plaza, Houston, Texas 77030. E-mail: mmatzuk{at}bcm.tmc.edu.


   ABSTRACT
TOP
 ABSTRACT
INTRODUCTION
 RESULTS
 DISCUSSION
 MATERIALS AND METHODS
 REFERENCES
 
Follistatin plays an important role in female physiology by regulating FSH levels through blocking activin actions. Failure to regulate FSH has been implicated as a potential cause of premature ovarian failure. Premature ovarian failure is characterized by amenorrhea, infertility, and elevated gonadotropin levels in women under the age of 40. Because follistatin is essential for postnatal viability, we designed a cre/loxP conditional knockout system to render the follistatin gene null specifically in the granulosa cells of the postnatal ovary using Amhr2cre transgenic mice. The follistatin conditional knockout females develop fertility defects, including reduced litter number and litter sizes and, in the most severe case, infertility. Reduced numbers of ovarian follicles, ovulation and fertilization defects, elevated levels of serum FSH and LH, and reduced levels of testosterone were observed in these mice. These findings demonstrate that compromising granulosa cell follistatin function leads to findings similar to those characterized in premature ovarian failure. Follistatin conditional knockouts may therefore be a useful model with which to further study this human syndrome. These studies are the first report of a granulosa cell-specific deletion of a gene in the postnatal ovary and have important implications for future endeavors to generate ovary-specific knockout mouse models.


   INTRODUCTION
TOP
 ABSTRACT
 INTRODUCTION
RESULTS
 DISCUSSION
 MATERIALS AND METHODS
 REFERENCES
 
FOLLISTATIN, A CYSTEINE-RICH monomeric glycoprotein encoded by a single gene, was originally identified from ovarian follicular fluid as a potent inhibitor of pituitary FSH (1). The murine follistatin gene contains six exons encoding the full-length 315-amino acid protein. An alternative splicing event produces a shorter 288-amino acid protein (2, 3). Follistatin is an important regulator of cellular differentiation and function due to its ability to bind and neutralize activins and bone morphogenetic proteins (BMPs) (4, 5).

Activins and inhibins are dimeric protein members of the TGF-ß superfamily that share ßA and ßB subunits [activin A (ßA:ßA), activin B (ßB:ßB), activin AB (ßA:ßB), inhibin A ({alpha}:ßA), and inhibin B ({alpha}:ßB)]. These proteins are named for their respective roles in stimulating and inhibiting the secretion of FSH from the pituitary (6). Follistatin interacts with activin at physiological pH; this affinity is similar to or exceeds that of activin binding to its receptor. The bound complex consists of one follistatin molecule bound to each of the two activin ß-subunits. Follistatin has low affinity for inhibins, which have only one ß-subunit (7, 8).

Follistatin also regulates members of the BMP subfamily by binding to them with lower affinities than for activin (5); however, this activity may be sufficient to allow follistatin to act as a regulator in the numerous systems expressing these factors. In Xenopus laevis embryos, follistatin is able to directly bind, inhibit, and neutralize BMP2, BMP4, and BMP7. The phenotype caused by overexpression of follistatin in early X. laevis embryos is similar to those achieved by overexpression of dominant negative BMP receptors (9, 10). Likewise, mouse BMP11 is able to induce dorsal mesoderm and neural tissue formation in X. laevis assays, and both effects are inhibited by follistatin (11). Follistatin also attenuates BMP15 stimulation of rat granulosa cell proliferation (12). These studies suggest that both ovarian activin and BMPs are potential targets for inhibition of ovarian-derived follistatin. Thus, knockout of follistatin in the ovary may affect both activin and BMP systems in the ovary because BMP4 and BMP7 are expressed in thecal cells; activin ßA and ßB and ALK3, ALK6, and BMPR2 are expressed in granulosa cells; and growth differentiation factor 9, BMP15, and BMP6 are expressed in oocytes (13, 14, 15).

To study the in vivo roles of follistatin (Fst), knockout mice were generated (16). Unfortunately, these mice die within hours of birth due to multiple defects including growth retardation, craniofacial defects, and decreased mass of the diaphragm and intercostal muscles, making it difficult to study the postnatal roles of follistatin in reproduction (16). Conversely, mice overexpressing follistatin are viable and develop to adulthood, but have extensive reproductive defects, including decreased testis size, an arrest in spermatogenesis in males, and a block in folliculogenesis in females (17). These findings, together with the abundance of follistatin mRNA in the ovary (18, 19), suggest that follistatin produces important effects within the gonads by regulating the local functions of activins and/or other TGF-ß family members.

Due to the lethality of the Fst knockout mice and to further study the postnatal functions of follistatin, we chose to produce mice in which the follistatin gene could be manipulated with the cre/loxP system. Mice carrying a follistatin allele flanked by loxP sites (FstFlox) were generated. Anti-Müllerian hormone receptor type II (Amhr2) knock-in mice were chosen for granulosa cell-specific expression of Cre recombinase. The Amhr2cre mice were hypothesized to be a key reagent to study the postnatal roles of granulosa cell genes in the ovary where Amhr2 and follistatin are highly expressed. These studies open new avenues of investigation using conditional mice to study the roles of essential granulosa cells expressed genes involved in the female reproductive system.


   RESULTS
TOP
 ABSTRACT
 INTRODUCTION
 RESULTS
DISCUSSION
 MATERIALS AND METHODS
 REFERENCES
 
Generation of a Follistatin Conditional Allele in Embryonic Stem (ES) Cells and Follistatin Conditional Mutant Mice
In previous studies, we used a PgkHPRT selectable marker cassette to replace 5.1 kb of the mouse Fst gene, thereby deleting essentially all six exons. This allele was called Fsttm1Zuk and herein is called Fsttm1 (16). In the present study, a floxed follistatin allele was created (Fsttm2Zuk; herein called FstFlox) for subsequent conditional deletion of the follistatin gene in vivo. The FstFlox targeting vector contains loxP sites inserted into intron 1 and 3' of exon 6. A PgkNeo cassette was positioned between the follistatin polyadenylation site and the 3'-loxP site (Fig. 1AGo). This construct was electroporated into the original HPRT-positive Fsttm1/+ ES cell line (FS3-C2) that is heterozygous for a null mutation at the Fst locus (16). We used 6-thioguanine for negative selection of the PgkHPRT cassette expressed in the ES cells and G418 for positive selection of the PgkNeo cassette in the targeting construct. Of the 36 ES cell clones that survived the double selection, 33 of them (92%) were correctly targeted at the Fst locus and contained two loxP sites. Two of these cell lines were used to produce chimeric male mice that were fertile and transmitted the FstFlox allele to F1 progeny. Heterozygous FstFlox mice were crossed to produce FstFlox homozygous mice. FstFlox/FstFlox mice are viable and fertile and were obtained at the expected Mendelian frequency. Intercrossing of heterozygotes (FstFlox/+) yielded 299 progeny including 81 wild-type mice (27%), 150 heterozygous (FstFlox/+) mice (50%), and 68 homozygous (FstFlox/ FstFlox) mice (23%) out of 39 litters analyzed with an average of 7.7 ± 0.37 pups per litter.



View larger version (39K):
[in this window]
[in a new window]
  		Fig. 1. Follistatin Conditional Gene Targeting Strategy

A, The follistatin conditional-targeting vector to delete exons 2–6 is shown. The targeting vector was generated by inserting loxP sites into intron 1 and 3' of the untranslated region of the Fst gene. A PgkNeo cassette was used for positive selection and 6-thioguanine for negative selection against the PgkHPRT cassette located at the Fst locus in the ES cells (FS3-C2) that contain the Fsttm1 allele. The four possible alleles are shown: Wt, Fsttm1, FstFlox, and Fst{Delta}. B, Southern blot analysis of tail DNA derived from six offspring from the cross of an EIIacre+ female to a heterozygous FstFlox/+ male. Lanes 1 and 6, Wt;EIIacre- mouse. Lanes 2, 3, and 5, FstFlox/+; EIIacre+ mice in which Cre-mediated recombination has now produced the genotype Fst{Delta}/+. Lane 4, FstFlox/+; EIIacre- mouse. Genomic DNA was digested using SacI/EcoRI and analyzed using the 5'-probe that can differentiate the four possible alleles. The 5'-probe detects a 13.7-kb wild-type band, a 10.5-kb FstFlox band, a 5.5-kb Fsttm1 band, as well as a 5.6-kb Fst{Delta} band produced by Cre recombination after crossing the FstFlox mice with EIIacre mice. Genotype analysis of the EIIacre transgene was determined by PCR as described in Materials and Methods.

 
To verify the presence of the loxP sites and their ability to recombine in vivo, we crossed the FstFlox/+ mice to EIIacre transgenic mice that express Cre recombinase in multiple tissues including germ cells (20). Southern blot analysis using a 5'-probe demonstrates the presence of the various Fst alleles, including the recombined Fst{Delta} allele (Fig. 1BGo). Thus, Cre recombinase can efficiently recombine the loxP sites flanking the Fst exons to produce a null allele.

Cre Activity in the Ovaries of Amhr2cre Knock-in Mice
Amhr2 together with its ligand anti-Müllerian hormone (Amh) are members of the TGF-ß family of signaling proteins expressed in the female gonads. AMH induces the regression of Müllerian ducts, precursors of the oviducts, uterus, and upper vagina by binding to AMHR2 expressed in the mesenchyme surrounding the ductal epithelium during embryogenesis. Amhr2 is expressed in Sertoli and Leydig cells in testes and granulosa cells in ovaries (21, 22, 23, 24, 25, 26, 27).

As a driver of Cre expression, the Amhr2cre knock-in mice were selected. Amhr2cre knock-in mice were generated with an insertion of a cre-neo cassette into the fifth exon of Amhr2 to genetically modify Amhr2-expressing tissues (28). The Amhr2cre mice were initially used to selectively disrupt the expression of Bmpr1a in the mesenchymal cells surrounding the Müllerian ducts, thereby identifying BMPR1A as a type I receptor for AMH-induced regression of the Müllerian ducts (28).

To monitor the Cre activity in the postnatal ovaries, the Amhr2cre mice were crossed to Gt(ROSA)26Sortm1Sor (R26R) mice that express ß-galactosidase activity only in cells that express Cre (29). After staining with 5-bromo-4-chloro-3-indolyl-ß-D-galactosidase, we detected ß-galactosidase activity in the ovaries of Amhr2cre R26R double heterozygous females (Fig. 2AGo). ß-Galactosidase (ß-gal) activity was present in the ovary as early as embryonic d 17.5 (Fig. 2BGo). Throughout the postnatal ovary, ß-gal activity was present in granulosa cells of all secondary and small antral follicles, but was at lower or undetectable levels in granulosa cells of primordial and primary follicles (Fig. 2Go, C and D). These observations are consistent with the endogenous Amhr2 expression pattern. However, low Cre activity was also found in some theca cells and oocytes (Fig. 2DGo). In addition to the ovary, we checked for Cre activity in the endometrium and the muscular layer of the uterus. The endometrium was negative for Cre; on the other hand, the muscular layer of the uterus was Cre positive as expected because Amhr2 is expressed in the mesenchyme of the Müllerian duct that gives rise to the uterine musculature and not the endometrium (24). These results indicate that the Amhr2cre knock-in line of mice is a good cre deleter strain to ablate follistatin expression in the granulosa cells of the ovary.



View larger version (127K):
[in this window]
[in a new window]
  		Fig. 2. Cre Activity in the Ovaries of Amhr2cre Mice Using the R26R Reporter

A, Whole-mount X-gal staining of a 6-wk-old Amhr2cre; R26R ovary. B, Transverse sections of an embryonic d 17.5 Amhr2cre; R26R ovary. ß-gal activity is observed in almost all of the ovarian somatic cells. C and D, Transverse sections of a 6-wk-old Amhr2cre; R26R ovary. C, ß-gal activity is observed mainly in granulosa cells of preantral and small antral follicles and is absent in granulosa cells of primordial and primary follicles; however, Cre activity was also found in some thecal cells. The apparent ß-gal activity in some oocytes may be a staining artifact because the intensity of the staining is not consistent with that in the adjacent granulosa cells (x50). D, Higher magnification (x200) of a follicle from the ovary shown in panel C.

 
Follistatin Expression in Conditional Mutant Mice
Follistatin mRNA is expressed at highest levels in the mammalian ovary (18). By in situ hybridization studies, follistatin expression in the adult ovary is primarily confined to the granulosa cells of preantral and antral follicles and to a lesser degree in the granulosa cells of primordial follicles (13, 18). Because follistatin null mice die at birth (16), the postnatal functions of follistatin in female reproductive physiology remained unclear.

To produce follistatin deficiency in the ovaries, Amhr2cre+; Fsttm1/+ mice were generated, and these mice were subsequently mated to FstFlox/FstFlox mice. All four expected genotypes were recovered from the above cross at close to the expected Mendelian frequency.

To determine the efficiency of the Cre recombinase of the Amhr2cre knock-in mice in ablating follistatin expression in the ovary, we performed Southern blot analysis of granulosa cells from immature females hormonally treated with 5 IU of pregnant mare’s serum gonadotropin (PMSG) for 48 h and Northern blot analysis of ovaries from adult untreated and hormonally stimulated mice.

Granulosa cells derived from six females (two Amhr2cre-; Fsttm1/FstFlox control and four Amhr2cre+; Fsttm1/FstFlox experimental mice) were isolated from the ovaries after PMSG stimulation, and DNA from these cells was analyzed by Southern blot (Fig. 3AGo). Whereas the Amhr2cre negative granulosa cells contained the FstFlox allele, the granulosa cells from all four Amhr2cre positive mice had the Fst{Delta} (deleted) allele at equal intensity to the Fsttm1 allele, demonstrating efficient recombination of the loxP sites of the FstFlox allele and conversion to the Fst{Delta} allele. Thus, these mice are follistatin null in their granulosa cells.



View larger version (29K):
[in this window]
[in a new window]
  		Fig. 3. Efficiency of Amhr2cre at Abolishing Follistatin Expression in the Ovary

A, Southern blot analysis of granulosa cell DNA derived from six offspring. Genomic DNA was isolated from granulosa cells of superovulated 19- to 21-d-old mice and digested using SacI/EcoRI and analyzed using the 5'-probe. The probe detects a 13.7-kb wild-type band, a 10.5-kb FstFlox band, a 5.5-kb Fsttm1 band, and a 5.6-kb Fst{Delta} band produced by Cre recombination after crossing the FstFlox mice with Amhr2cre mice. Lanes 2 and 4, Amhr2cre-; Fsttm1/FstFlox. Lanes 1, 3, 5, and 6, Genomic analysis of Amhr2cre+; Fsttm1/FstFlox mice shows the recombination event induced by the Cre recombinase. B, Northern blot analysis of ovarian RNA derived from three different Amhr2cre+ mice at 3 months of age. Ovarian RNA samples were: WT control female, WT female superovulated with 5 IU of PMSG for 48 h, two Amhr2cre+; Fsttm1/FstFlox females, and one Amhr2cre+; Fsttm1/FstFlox female superovulated with 5 IU of PMSG for 48 h. Gapd was used as a control for RNA loading.

 
To test for the presence of the follistatin mRNA throughout the entire ovary (i.e. ensuring that follistatin was not expressed from other cell types), Northern blot analysis was conducted using a follistatin cDNA probe. As shown in Fig. 3BGo, Northern blot analysis of wild-type untreated or PMSG-treated ovaries using a follistatin cDNA probe identified a follistatin mRNA band of approximately 2.6 kb. However, no follistatin mRNA signal was detected in the mRNA derived from Amhr2cre+; Fsttm1/FstFlox ovaries, either untreated or PMSG treated. Thus, inactivation of the follistatin gene in granulosa cells abolishes follistatin mRNA expression in the entire adult ovary in the presence or absence of PMSG treatment.

Fertility Analysis of Follistatin Conditional Mice
To study the fertility of mice with follistatin-deficient gonads, male or female Amhr2cre-; Fsttm1/FstFlox (control mice) and Amhr2cre+; Fsttm1/FstFlox (experimental mice) were mated to wild-type female or male mice for a 6-month period. There was no significant difference in the fertility between the seven experimental males (9.62 ± 0.59 pups per litter and 0.96 ± 0.03 litter/month) and six control males (8.74 ± 0.82 pups per litter and 0.92 ± 0.04 litter/month) (Table 1Go). On the other hand, experimental females exhibited fertility defects; one of the 11 experimental female mice was infertile, whereas the remaining 10 experimental females displayed significantly reduced fertility. The number of litters per month for the experimental females (0.42 ± 0.07) was less than half of that of the control females (0.97 ± 0.01). The number of pups per litter for the experimental group (3.75 ± 0.35) was less than half of the values of control females (7.88 ± 0.52) (Table 1Go). Thus, ovarian follistatin plays an important role in female reproduction, and the absence of follistatin in the granulosa cells of the ovary reduces fertility.


View this table:
[in this window]
[in a new window]
  		Table 1. Six-Month Mating Data for Fsttm1/FstFlox Mice

 
Morphological and Histological Analysis
To determine the causes of subfertility in the Amhr2cre +; Fsttm1/FstFlox mice, ovaries were analyzed at multiple time points. There were no obvious differences in ovarian size between the experimental mice (Amhr2cre+; Fsttm1/FstFlox) and the control mice (Amhr2cre-; Fsttm1/FstFlox); however, the majority of the ovaries of 8-month old experimental mice looked hemorrhagic (Fig. 4IGo).



View larger version (138K):
[in this window]
[in a new window]
  		Fig. 4. Histological Analysis of the Ovaries of Amhr2cre-; Fsttm1/Fstflox (Control) Mice and Amhr2cre+; Fsttm1Fstflox (Experimental) Mice

A–C, Amhr2cre-; Fsttm1/FstFlox mice. D–H, Amhr2cre+; Fsttm1/FstFlox mice. Panels A and D show ovaries from 3-month-old female mice (low power). A, Amhr2cre-; Fsttm1/FstFlox mouse ovary with several corpora lutea, primary, secondary, and antral follicles. D, Amhr2cre+; Fsttm1/FstFlox mice ovary with lower number of antral follicles, presence of PAS-positive material in the interstitium, and decrease in the number of corpora lutea. Panels B and E show ovaries of 6-month-old female mice. B, Amhr2cre-; Fsttm1/FstFlox mouse ovary (low power) with several corpora lutea, primary, secondary, and antral follicles. E, Amhr2cre+; Fsttm1/FstFlox mouse ovary (low power) with one antral follicle and few primary and secondary ones, high abundance of PAS-positive material in the interstitium, no corpora lutea. Panels C, F, G, and H show ovaries of 8-month-old female mice. C, Amhr2cre-; Fsttm1/FstFlox mouse ovary (low power) with several corpora lutea, primary, secondary, and antral follicles. F, Amhr2cre+; Fsttm1/FstFlox mouse ovary (low power) with no follicles in any stage, absence of corpora lutea, high abundance of PAS-positive material in the interstitium, no corpora lutea, presence of a cyst (indicated by asterisk) and presence of a Sertoli cell tubule-like structure (indicated by arrow). G, Amhr2cre+; Fsttm1/FstFlox mouse ovary (low power) with one antral follicle and no primary or secondary ones, high abundance of PAS-positive material in the interstitium with an area of high abundance (indicated by arrowhead), no corpora lutea, and presence of Sertoli cell tubule-like structure (indicated by arrow). H, High magnification of panel G showing a Sertoli cell tubule-like structure. I, Comparison between the ovaries of Amhr2cre-; Fsttm1/FstFlox mouse (left) and Amhr2cre+; Fsttm1/FstFlox mouse (right) at 8 months of age. Ovaries from Amhr2cre-;Fsttm1/FstFlox mouse shows normal appearance. In contrast, ovaries from Amhr2cre+; Fsttm1/FstFlox mouse are hemorrhagic in appearance.

 
Histologically, differences were found in adult mice indicating a degeneration of ovarian function with age. At 5 and 21 d of age, there was no significant difference in the ovarian histology between experimental and control mice (data not shown). We began to find differences between the ovaries of experimental and control mice after 3 months of age. At 3 months of age, the ovaries of all experimental mice that were analyzed showed follicles at all stages of folliculogenesis (Fig. 4DGo); however, the experimental ovaries exhibited a decrease in the number of antral follicles and corpora lutea and an increase in accumulation of periodic acid Schiff (PAS)-positive material in the interstitium that was likely due to zona pellucida (ZP) remnants (Fig. 4DGo; see Ref.30) when compared with the ovaries of control mice (Fig. 4AGo). By 6 months of age, a marked contrast in control vs. experimental ovaries was observed. All four experimental mice analyzed had a reduction in the number of follicles (specifically the antral follicles), an abundance of ZP remnants, and a decrease in the number of corpora lutea (Fig. 4EGo) when compared with control mice of the same age with follicles at all the stages of folliculogenesis, few ZP remnants, and abundant corpora lutea (Fig. 4BGo). One of the 6-month-old experimental ovaries showed Sertoli cell tubule-like structures (data not shown).

Comparative histology of 8-month-old mice demonstrated remarkable differences. Ten of the 11 experimental mice analyzed had few follicles (<five primary or later follicles per cross-section of the ovary; Fig. 4Go, F and G). This is a significant decrease when compared with control mice of the same age that had more than 10 follicles at different stages of folliculogenesis, abundant corpora lutea, and few ZP remnants (Fig. 4CGo). The experimental ovaries showed an increase in ZP remnants when compared with control and younger experimental mice, indicating an active oocyte death process (Fig. 4Go, F and G). Another important finding was that seven of the 11 experimental mice had Sertoli cell tubule-like structures (Fig. 4Go, F–H), reminiscent of the ovarian tumors from inhibin-deficient mice (31, 32), mice overexpressing follistatin (17) and FSH (33), mice lacking both estrogen receptor {alpha} and ß (34), and mice overexpressing AMH (35). Two of the experimental mice had single ovarian cysts (Fig. 4FGo), previously reported in mice lacking growth differentiation factor 9 (36). Therefore, these findings suggest that follistatin plays roles in oocyte survival and folliculogenesis due to the accelerated loss of oocytes and follicles with age in the experimental mice.

Consistent with the normal fertility of the mutant males, there was no significant difference in testes size, weight, or histology between the control and experimental mice (data not shown). These results indicate that either follistatin does not play an essential role in the testis or the Amhr2cre knock-in mouse model is not an appropriate cre deleter strain to abolish follistatin expression in the Sertoli cells of the testis where follistatin is expressed (37).

Ovarian and Oocyte Physiology
To assess whether follistatin plays a critical role in successful fertilization of the released ovum, 3-wk-old females were superovulated and then mated to stud males. Approximately 20 h after the human chorionic gonadotropin (hCG) injection, oocytes and one-cell embryos were isolated. Eighty-one percent of the oocytes from the control mice (Amhr2cre-; Fsttm1/ FstFlox) and 70% of the oocytes from the experimental mice (Amhr2cre+; Fsttm1/FstFlox) developed to two-cell embryos after 24 h culture (Table 2Go). No significant differences were found either in the number of oocytes released or the fertilization rates between the two groups, indicating that the number of oocytes ovulated and fertilized in the presence or absence of follistatin, at least under pharmacological conditions, were the same in these immature mice.


View this table:
[in this window]
[in a new window]
  		Table 2. In Vivo Fertilization of 21-d-Old Fsttm1/FstFlox Mice

 
Because histology defects are observed in experimental mice as they aged, the in vivo superovulation experiments were performed in 6-month-old females. The experiments were performed under the same conditions as the ones for the immature mice. From the 185 oocytes released from six control mice (30.8 ± 4.8), 70% of them developed to two-cell embryos after 24 h culture. However, from the 122 oocytes released from seven experimental mice (17.4 ± 3.8), only 54% of them developed to two-cell embryos after 24 h culture (Table 3Go). A significant difference was found in the number of oocytes released as well as the number that developed to the two cell stage in the two groups. These findings indicate that adult mice lacking follistatin in the granulosa cells do not respond as well as the control mice to exogenous gonadotropins because they 1) release fewer oocytes and 2) these oocytes are less competent to be fertilized.


View this table:
[in this window]
[in a new window]
  		Table 3. In Vivo Fertilization of 6-Month-Old Fsttm1/FstFlox Mice

 
To further study follicle development in these mice, we injected 5 IU of PMSG and 47 h later 5 IU of hCG to 32-d-old control and experimental mice. The ovaries were collected 5 h after hCG injection, and the tissues were fixed in formalin. The histology of these ovaries showed that both control and experimental follicles became periovulatory after gonadotropin stimulation (Fig. 5Go, A and B) even though it seemed that control mice have a better response to gonadotropin stimulation. One additional important finding in this experiment was the presence of abnormal follicle development (i.e. a follicle with three oocytes) in the ovary of one of the experimental mice (Fig. 5Go, B and C).



View larger version (67K):
[in this window]
[in a new window]
  		Fig. 5. Defects with Gonadotropin Stimulation in Amhr2cre+; Fsttm1/Fstflox Mice

Histology of ovaries of 32-d-old mice after stimulation with 5 IU of PMSG and 47 h later 5 IU of hCG. The ovaries were collected 5 h after hCG injection, and the tissues were fixed in formalin and stained with PAS and hematoxylin. A, Ovary of a control (Amhr2cre-; Fsttm1/FstFlox) mouse in which the majority of follicles became periovulatory. B and C, Ovary of an experimental (Amhr2cre+; Fsttm1/FstFlox) mouse with follicles in different stage from primary to periovulatory. Only a fraction of the follicles became periovulatory. The presence of an abnormal follicle with three oocytes is indicated by an arrow. C, High magnification of the abnormal follicle from panel B. Three oocytes are present surrounded by granulosa and theca cells.

 
Determination of Hormonal Levels
To further analyze the reproductive deficits of the Amhr2cre+; Fsttm1/FstFlox females, we measured the serum levels of the gonadotropins FSH and LH, the estrogen estradiol, and the androgen testosterone by RIA. The hormonal data of experimental (Amhr2cre+; Fsttm1/FstFlox) and control (Amhr2cre-; Fsttm1/FstFlox) mice are shown in Table 4Go. There was a significant difference in the values of three of the four measured hormones, specifically in the oldest mice.


View this table:
[in this window]
[in a new window]
  		Table 4. Hormonal Data for Fsttm1/FstFlox Mutant Mice

 
In the case of FSH, there was a difference between experimental and control mice both in immature (21 d old) and mature (3 and 8 month-old) mice. At 3 months of age, we found the biggest difference in the levels of circulating FSH; the experimental mice had an average of 13.28 ± 0.99 ng/ml when compared with control mice of the same age with an average of 8.04 ± 0.69 ng/ml. The lowest FSH level of the experimental mice was comparable to the highest levels seen in control mice (data not shown). The experimental mice at 8 months of age had an increase in the values of FSH when compared with the control (12.68 ± 2.74 ng/ml vs. 7.81 ± 1.37 ng/ml). These data are consistent with the histological findings. For example, the experimental mouse that exhibited the presence of cysts in the ovary (Fig. 4GGo) had one of the highest serum levels of FSH (i.e. 23.63 ng/ml). The increase in the FSH values supports the histological findings that show a pattern resembling that of premature ovarian failure (POF). The levels of the gonadotropin LH were also evaluated. A significant difference between the LH levels of control and experimental mice in immature and adult animals was also discovered with the differences in the oldest mice being the most substantial. At 8 months of age, the experimental mice exhibited a value of 0.53 ± 0.07 ng/ml whereas the control mice had a level of 0.23 ± 0.07 ng/ml. These results of the gonadotropins (FSH and LH) indicate a dysfunction in the ovary and in the hypothalamic pituitary-ovarian feedback system, a finding that might be expected from the ovarian histology (Fig. 4Go).

Estradiol is an estrogen produced in the granulosa cells of the ovary. Levels of estradiol were measured in experimental and control mice at 3 months, 6 months, and 8 months of age. Surprisingly, no significant differences between experimental and control mice at any age were found. The estradiol levels start to decrease in the experimental mice (15.82 ± 1.41 pg/ml) at 6 months of age when compared with the control mice (18.17 ± 3.99 pg/ml) of the same age. In the 8-month-old mice, the estradiol levels were reduced to 6.61 ± 1.11 pg/ml but we also noticed a decrease in the levels in the control mice (7.93 ± 0.99).

Half of the circulating levels of testosterone are produced in the ovary. Deficiencies in testosterone have been reported in women with premature ovarian failure (38). Consistent with studies in those women, we found a significant difference in the values of testosterone when we compared experimental (0.18 ± 0.02 pg/ml) and control (0.40 ± 0.11 pg/ml) mice at 6 months of age (Table 4Go). Also, at 8 months of age, a significant difference in testosterone levels were found in experimental mice (0.19 ± 0.01 pg/ml) vs. control mice (0.35 ± 0.07 pg/ml). No significant differences were found in the 3-month-old mice.

Study of the Ovarian Expression of TGF-ß Superfamily Members in the Amhr2cre+; Fsttm1/FstFlox Mice
Discrepancies between the phenotypes of transgenic mice deficient in either activin or activin type II receptor and the phenotype of transgenic mice overexpressing follistatin (17, 39) suggest that activins are not the only growth factors that may be affected by the absence of follistatin. Possible candidates include members of the TGF-ß family such as inhibins that share the same ß-subunits with activins and BMPs (e.g. oocyte-derived factor BMP15) that are inhibited by follistatin (12).

To determine whether mRNAs of any TGF-ß superfamily members are regulated by follistatin, the levels of activin ßA and ßB, inhibin {alpha}, and Bmp15 mRNAs were measured by Northern blot analysis of Amhr2cre+; Fsttm1/Fstflox (experimental mice) and Amhr2cre-; Fsttm1/Fstflox (control mice) ovarian RNA. Initially individual sets of ovaries from 3-month-old females untreated or treated with 5 IU of PMSG for 48 h before ovary collection were used to assess whether the absence of follistatin in the gonads has an effect on these ovarian-expressed TGF-ß members. Regarding the levels of {alpha}-subunit mRNA, no differences were found. The {alpha}subunit was present in both experimental and control mice, both untreated and treated (Fig. 6CGo). However, some individual differences were found when the blots were probed with cDNA against activin ßA and ßB, and Bmp15 (data not shown). These results suggest that follistatin could have an effect on the expression of these genes. To corroborate these findings, Northern blot analysis using a pools of ovaries from four mice at 6 months of age was performed. The results from the Northern blot of pool of ovaries using cDNA probe against activin ßA and ßB and BMP15 revealed no significant differences between the ovaries of Amhr2cre-; Fsttm1/FstFlox (control) mice, and ovaries from Amhr2cre+; Fsttm1/FstFlox (experimental) mice (Fig. 6Go, A and B). These result indicate that follistatin does not seem to have a role in the transcriptional regulation of any of these ovarian TGF-ß members.



View larger version (46K):
[in this window]
[in a new window]
  		Fig. 6. Northern Blot Analyses of TGF-ß Family Ligand mRNA in the Ovary

A, Lanes represent pool of ovaries of four mice at 6 months of age in the following order: ovaries from wild-type mice, ovaries from Amhr2cre-; Fsttm1/FstFlox (control) mice, and ovaries from Amhr2cre+; Fsttm1/FstFlox (experimental) mice. The Northern blot was probed with cDNAs against activin ßA, Bmp15, and Gapd. Gapd was used as a control for RNA loading. B, Lanes represent pool of ovaries of four mice at 6 months of age in the following order: ovaries from wild-type mice, ovaries from Amhr2cre-; Fsttm1/FstFlox (control) mice, and ovaries from Amhr2cre+; Fsttm1/FstFlox (experimental) mice. The Northern blot was probed with cDNA against activin ßB and Gapd. Gapd was used as a control for RNA loading. C, Lanes represent ovaries from individual mice at 3 months of age in the following order: one wild-type mouse untreated, one wild-type mouse treated with 5 IU of PMSG for 48 h, two different Amhr2cre+; Fsttm1/FstFlox mice untreated, one Amhr2cre+; Fsttm1/FstFlox mouse treated with 5 IU of PMSG for 48 h, and two different Amhr2cre-; Fsttm1/FstFlox mice untreated. The Northern blot was probed with cDNA against inhibin {alpha}-subunit and Gapd. Gapd was used as a control for RNA loading.

 

   DISCUSSION
TOP
 ABSTRACT
 INTRODUCTION
 RESULTS
 DISCUSSION
MATERIALS AND METHODS
 REFERENCES
 
To study the in vivo roles of follistatin in the mammalian ovary, a cre-loxP system was used. A follistatin conditional allele was generated by insertion of loxP sites flanking exons 2–6 of the Fst gene (Fig. 1AGo). This system allowed for removal of all of the exons required for the synthesis of the functional follistatin protein when these mice were crossed to cre-expressing mice. As a cre deleter strain, the Amhr2cre knock-in mice were selected due to the high levels of Cre activity in granulosa cells of postnatal ovaries of these mice (Fig. 2Go).

No expression of ovarian follistatin was detected in experimental mice tested at 3 months of age (Fig. 3AGo). Also, Southern blot analysis of the granulosa cells of 21-d-old experimental ovaries showed the Fst{Delta} allele and not the FstFlox allele (Fig. 3BGo), demonstrating that recombination occurs in essentially 100% of the granulosa cells. These results indicate that using a mouse to deliver cre under the control of the Amhr2 gene represents an excellent system for studying follistatin function in the granulosa cells of the postnatal ovary. These results also open new avenues of research for using conditional mutant mice to study the roles of other (ubiquitous) genes expressed in the granulosa cell of the ovary. Even though we obtained almost 100% recombination in the granulosa cells studied, the concern is always present that Cre activity will be insufficient to cause loxP recombination in all cells early in development, resulting in mosaicism. However, transplanting follistatin null ovaries into the bursa of wild-type mice creates a similar phenotype to the Fst conditional mutant mice. These mice are infertile, and minimal follicle formation in the ovaries is observed (Matzak, M., unpublished data). One advantage of using the conditional knockout mice as a model rather than ovarian transplant recipients is the opportunity to have an unlimited number of mice to use in the studies. Another advantage is that FstFlox mice will allow study of the in vivo roles of follistatin not only in the reproductive system but also in other systems. Areas of interest include the development of the skin, hair, and whiskers (16, 17, 40, 41, 42, 43), as well as kidney tubule morphogenesis, where Fst is highly expressed and its expression is reduced by injury (44), and bone and palate development, where it plays a role in the transition from cartilage to bone (16, 45).

The preliminary results of ovarian transplantation experiments in which the females were infertile and the fact that our experimental mice were also infertile or subfertile indicate the importance of follistatin in communication between different cell types of the follicle for its formation and maturation. The reason that only one of the experimental mice was completely infertile may be due to delayed postnatal deletion of the FstFlox allele in the subfertile mice. Mosaicism, in which gene inactivation is induced in only a fraction of the desired cells, has been reported in the other cre/loxP systems (46). Although a range of phenotypes was found with use of Amhr2cre knock-in mice (28), mosaicism appears minimal based on our Southern and Northern blot analyses (Fig. 3Go).

Our results show that mice with the deletion of the follistatin gene in ovarian granulosa cells exhibited a reduced number of litters per month and reduced litter size, and, in the most severe case, infertility. No defects in ovulation or fertilization were found with superovulation of immature mice (Table 2Go). However, 6-month-old experimental mice did show a significant reduction in ovulation and fertilization of oocytes, suggesting reductions in the number and quality of available oocytes (Table 3Go). Although five of the seven experimental mice responded to stimulation with exogenous gonadotropins to varying degrees, two mice did not respond.

Histological examination of experimental mice beyond 6 months of age reveals a drastic reduction in the number of the follicles (Fig. 4Go, E–G), corroborating our hypothesis that there is a reduction in the pool of available oocytes as a consequence of the loss of follistatin. Furthermore, some of the 8-month-old females did not have any ovarian follicles (Fig. 4FGo). These results indicate premature ovarian failure in mice lacking follistatin in the granulosa cells. One interesting histological finding in the experimental mice was the presence of Sertoli cell tubule-like structures (Fig. 4Go, F–H) noted in seven of the eleven 8-month-old and one of the four 6-month-old, experimental mice. Sertoli cell tubule-like structures have been reported in female mice lacking inhibin {alpha} (31, 32), both estrogen receptors {alpha} and ß (34), and in those overexpressing AMH, follistatin, and FSH (17, 33, 35). All of the above mentioned genes are expressed in granulosa cells, and their perturbation can cause changes in granulosa cell fate. Furthermore, experimental mice experience an increase in oocyte loss (demonstrated by an abundance of ZP remnants) (Fig. 4Go, D and G). When a follicle loses its oocyte, the follicle is no longer regulated by oocyte factors, and that subsequently results in an alteration in cell fate. Follicle development is a very sensitive process influenced by a variety of factors and disturbances of any of them, either by removal or excess, can cause a change in fate pattern. The Sertoli cell tubule-like structures are a consequence of an imbalance in the follicle equilibrium.

Ovarian failure in women may be diagnosed by high concentrations of the gonadotropins, FSH and LH, and a low concentration of estradiol. A rise in FSH is the most sensitive and the best early marker for ovarian failure (47). FSH levels in our experimental mice lacking ovarian follistatin were significantly increased in both immature and mature mice. Also LH levels were elevated in the experimental mice (Table 4Go). Interestingly, there was no significant difference in the level of estradiol between experimental and control mice in our study. Women with premature ovarian failure (POF) produce estrogen intermittently and may ovulate despite the presence of high gonadotropin levels (48). Ovaries of POF patients sporadically go through a temporary phase of low activity, which can return to normal later in life. For that reason, at least two hormone level measurements, taken some weeks apart, are necessary before a POF diagnosis is made (47, 49). This could be a possible explanation for the lack of variation in the estradiol levels of the experimental mice. Unfortunately, we can not take a second sample to confirm whether or not there is a significant reduction in the levels of estradiol in the experimental mice.

We also evaluated testosterone levels in experimental mice. Women with POF have the potential for loss of ovarian androgens due to the atrophy of the ovarian cortex (38, 50, 51, 52). In agreement with the human POF clinical findings, we found a significant decrease in testosterone in experimental mice beyond 6 months of age (Table 4Go). Thus, the hormonal profile of mice lacking ovarian follistatin resembles the hormonal profile of women with POF.

The quality and number of oocytes in the experimental mice are compromised as seen by a reduction in the ovulation and fertility rate in the 6-month-old experimental mice when in vivo fertilization experiments were performed (Table 3Go). Two of the seven experimental mice did not respond to the ovulation induction. Women with POF, who underwent attempts to induce ovulation using different regimens, have reduced ovulation rates (53, 54, 55). This result indicates that defects in granulosa cell function during follicle development can compromise not only ovulation but also oocyte competence.

Follistatin has been associated with polycystic ovary syndrome (PCOS) in women through genetic linkage studies. PCOS is an endocrine disorder characterized by reduced fertility, hyperandrogenism, and chronic anovulation (56). After the original study linking follistatin to PCOS, three different studies (including one from the same group) did not show any relation between mutations in the human follistatin gene and PCOS (57, 58, 59). We did not find any indication of hyperandrogenism or multiple ovarian cysts in our ovarian follistatin deletion mice. In contrast, the values of testosterone in 6- and 8-month-old female mice showed a significant reduction when compared with controls. Thus, loss of ovarian follistatin in mice fails to model several features of PCOS.

Our findings of early deterioration of ovarian function, increased gonadotropins, decreased testosterone, decreased number of follicles, and failure of follicles to be fertilized and ovulated in response to exogenous gonadotropins suggest that a mutation compromising follistatin in the granulosa cell of the ovary could be a cause of POF. POF is a condition causing amenorrhea, infertility, and elevated gonadotropin concentrations in women under the age of 40 yr. In some women, intermittent ovarian function has been reported and pregnancy can occur in 5–10% of patients subsequent to diagnosis. In most patients with POF and normal female chromosome constitution, no cause can be identified. The hypotheses of the causes of POF are: 1) failure to attain the appropriate primordial follicle pool, and 2) accelerated loss of oocytes and follicles. Our animal model provides us with an insight into a new candidate gene possibly involved in POF. Follistatin should be included in future POF genetic studies. If follistatin is found to be associated with POF in women, our animal model will be useful in elucidating the mechanism of this disease that has significant psychosocial sequelae and major health implications.

One model to explain our findings is that mouse follistatin is required in the antral follicle to neutralize atretic effects of activin and allow the follicle to progress to ovulation. This model is based on in vitro evidence that follistatin is able to bind and neutralize activins (8). Activin ßA is able to induce apoptosis in different cell lines including B9 (60) and immortalized ovarian surface epithelium (IOSE-29) (61). Furthermore, when recombinant activin is injected directly into the ovary of immature female rats, in the presence or absence of systemic PMSG, follicles showed morphological signs of atresia in which granulosa and theca cell layers were atrophied and oocytes were highly fragmented (62). In our mice, absence of follistatin may facilitate activin-induced follicle atresia and oocyte loss. Due to excess of activin and continuous loss of oocytes, the oocyte pool is subsequently reduced in 6-month-old experimental mice, leading to a loss of ovarian function. We have shown that activin subunits are highly expressed in the granulosa cells of antral and atretic follicles, and follistatin is highly expressed in the granulosa cells of preantral and antral follicles using in situ hybridization (13). These expression patterns in wild-type mice support the model that follistatin is required in the antral follicle to neutralize activin-induced atresia. We are currently producing and studying activin conditional knockout mice, which will help us to understand better the interrelationships of activins and follistatin in ovarian physiology.


   MATERIALS AND METHODS
TOP
 ABSTRACT
 INTRODUCTION
 RESULTS
 DISCUSSION
 MATERIALS AND METHODS
REFERENCES
 
ES Cell Technology and Southern Blot Analysis
More than 20 kb of DNA, encompassing the six-exon mouse follistatin gene sequence, was isolated from a 129S6/SvEv genomic library using a follistatin cDNA. This genomic sequence was used to generate a conditional targeting vector. The targeting vector contained 3.7 kb of intron 1 sequence, the exon 1 of the follistatin gene, a loxP site, exons 2–6 of the follistatin gene, a positive selectable marker [the phosphoglycerate kinase promoter, neomycin (PgkNeo) expression cassette], which was inserted between the apparent polyadenylation site and the 3'-loxP site, a loxP site, and 2.9 kb of genomic DNA downstream of exon 6. The linearized vector was electroporated into the ES cell line (FS3-C2) containing a selectable marker, the phosphoglycerate kinase promoter with the hypoxanthine-guanine phosphoribosyl transferase minigene (PgkHPRT) expression cassette, replacing 5.1 kb of the mouse follistatin (Fst) gene, that include essentially all six follistatin exons (16). Clones were selected in 6-thioguanine for negative selection of the PgkHPRT cassette expressed in the ES cells and G418 for positive selection of the PgkNeo cassette in the construct. DNA from surviving clones was analyzed by Southern blot analysis, and targeted ES cell clones were expanded and injected into blastocysts as previously described (31). Ninety two percent of the ES cell was correctly targeted at the Fst locus (data not shown), and two of these ES cell clones, FstFlox-D1 and FstFlox-D12, were injected into blastocysts and produced male chimeras that successfully transmitted the mutant FstFlox allele to F1 progeny. F1 heterozygous mice were intercrossed to produce FstFlox homozygous mice. Chimeras were mated to C57Bl6/J females to produce 129S6/SvEv/C57BL6/J hybrid mice. Southern blot analysis was used for genotyping analysis of all FstFlox mutant offspring as shown in Fig. 1BGo. Southern blot analysis of granulosa cells was performed with mural granulosa cells isolated from large antral follicles of 3-wk-old control and experimental mice treated with PMSG for 48 h before granulosa cell collection as described in Ref.63 and shown in Fig. 2AGo.

Genotyping of the Amhr2cre Allele and EIIacre and ROSA 26 Reporter (R26R) Transgene
Genotyping analysis of the Amhr2cre allele and EIIacre and R26R transgene was determined using PCR. The sequences of the primers for the Amhr2cre genotyping were 5'-cgcattgtctgagtaggtgt-3' and 5'-gaaacgcagctcggccagc-3'. The sequences of the primers for the EIIacre genotyping were 5'-ccgggctgccacgaccaa-3' and 5'-ggcgcggcaacaccattttt-3'. The sequences of the primers for the R26R genotyping were 5'-gcgttacccaacttaatcg-3' and 5'-tgtgagcgagtaacaacc-3'.

RNA Isolation and Northern Blot Analysis
Total ovarian RNA was isolated from individual mice by acid guanidium thiocyanate-phenol-chloroform extraction using the RNA STAT-60 reagent (Leedo Medical Laboratories, Houston, TX). Each RNA sample (12 µg) was used for electrophoresis and transferred to nylon membranes as described previously (13). Radioactive complementary cDNA probes were synthesized from the templates listed in Table 5Go using [{alpha}32P]dATP and the Strip-EZ kit (Ambion, Inc., Austin, TX). Autoradiography and phosphor imaging allowed for visualization and quantification of probe hybridization, respectively. Phosphor imaging plates were scanned and analyzed using ImageQuant software (Molecular Dynamics, Inc., Sunnyvale, CA). A background level for each blot was determined and subtracted. Blots were stripped and reprobed for glyceraldehyde 3-phosphate dehydrogenase (Gapd), and phosphor imaging of the Gapd signal allowed us to correct each lane for RNA loading.


View this table:
[in this window]
[in a new window]
  		Table 5. Northern Blot Probes

 
Morphological and Histological Analysis
Mice of each sex and genotype were killed and tissues collected immediately into 10% buffered formalin for ovaries or Bouin’s fixative for testes. After overnight fixation, tissues were embedded in paraffin blocks, sectioned, and stained with hematoxylin and eosin or PAS and hematoxylin. Embedding and staining procedures were performed by standard protocols at the Baylor College of Medicine Pathology Core Services laboratory.

Serum Analysis
Mice were anesthetized by isoflurane inhalation (Abbott Laboratories, North Chicago, IL), and blood was recovered by closed cardiac puncture. Serum was separated by centrifugation in Microtainer tubes (Becton Dickinson, Franklin Lakes, NJ) and stored at -20 C before analysis. FSH, LH, estradiol, and testosterone measurements were made by the University of Virginia Ligand Core Facility (Specialized Cooperative Centers Program in Reproduction Research NICHD/NIH U54 HD28934). Estradiol and testosterone were measured using commercially prepared kits from Diagnostic Systems Laboratories [ULTRA ESTRADIOL (3rd Gen.) DA RIA (DSL-39100), sensitivity, 1.5 pg/ml; and FREE TESTOSTERONE RIA (DSL-4900) sensitivity, 0.25 pg/ml].

Mouse LH Sandwich Assay (MLHS).
LH was measured in serum by a modified supersensitive two-site sandwich immunoassay (1) using monoclonal antibodies MAB1 (no. 581B7) against bovine LH and TMA (no. 5303: Medix, Kauniainen, Finland) against the human LH-ß-subunit. The tracer antibody (no. 518B7 kindly provided by Dr. Janet Roser, Department of Animal Science, University of California, Davis) was iodinated by the chloramine T method and purified on Sephadex G-50 columns. The capture antibody (no. 5303) was biotinylated and immobilized on avidin-coated polystyrene beads (7 mm; Nichols Institute, San Juan Capistrano, CA). Mouse LH reference preparation provided by Dr. A. F. Parlow and the National Hormone and Peptide Program was used as standard. The assay has a sensitivity of 0.07 ng/ml.

Mouse FSH RIA.
Mouse FSH measurements were determined by RIA using reagents provided by Dr. A.F. Parlow and the National Hormone and Peptide Program and procedures validated earlier. Mouse FSH reference preparation was used for assay standards, and mouse FSH antiserum (guinea pig) AFP-1760191, diluted to a final concentration of 1:200,000, was used as a primary antibody. Secondary antibody was purchased from Antibodies, Inc. (catalog no. 51-534) and was diluted to a final concentration of 1:60. The assay has a sensitivity of 4.5 ng/ml and less than 0.5% cross-reactivity with other pituitary hormones.

Superovulation and Isolation of Oocytes/Embryos
Experimental and control female mice (19–21 d-old and 6 months old) were injected ip with PMSG (5 IU/mouse) and given hCG ip (5 IU/mouse) 48 h later. Mice were bred to C57/129 hybrid stud males. The following morning eggs and/or embryos were recovered in M2 medium, counted, and cultured in vitro for 24 h in M16 medium.

Histochemical Analysis
X-galactosidase (X-gal) staining on ovaries was performed by fixing the tissue in 2% paraformaldehyde in 1x PBS, pH 7.2, overnight, rinsing three times for 30 min at 4 C in rinse buffer (2 mM MgCl2; 0.1% sodium deoxycholate; 0.2% Nonidet P-40 in PBS, pH 7.2) and staining overnight at 37 C in X-gal staining solution (5 mM potassium ferricyanide; 5 mM potassium ferrocyanide; X-gal, 1 mg/ml in rinse buffer). After X-gal staining, tissues were washed in three changes of 1x PBS. Tissues were embedded in paraffin, and 5-µm tissue sections were stained with nuclear fast red and analyzed by light microscopy.


   ACKNOWLEDGMENTS
 
We thank Drs. Kathleen H. Burns and Stephanie A. Pangas for critical reading of the manuscript, Ms. Shirley Baker for her assistance in manuscript formatting, Mr. Julio Agno for assistance with genotyping and Southern blot analysis, and Ms. Cathy Guo for help in the construction of the targeting vector. We thank Dr. Heiner Westphal for the gift of the EIIacre transgenic mice and Dr. Philippe Soriano for the gift of the R26R mice. We acknowledge Ms. Valerie Long and Aleisha Schoenfelder at the University of Virginia Ligand Core Facility (Specialized Cooperative Centers Program in Reproduction Research) for assistance with the hormonal determination.


   FOOTNOTES
 
This work was supported in part by National Institutes of Health Grants HD32067 (to M.M.M.) and HD30284 (to R.R.B.). Hormonal determination was supported by National Institute of Child Health and Human Development/National Institutes of Health through cooperative agreement [U54 HD28934] as part of the Specialized Cooperative Centers Program in Reproduction Research. S.P.J. is the recipient of a postdoctoral fellowship from the Lalor foundation.

Abbreviations: AMH, Anti-Müllerian hormone; Amhr2, anti-Müllerian hormone receptor type II; BMP, bone morphogenetic protein; ES, embryonic stem; Fst, follistatin; ß-gal, ß-galactosidase; hCG, human chorionic gonadotropin; PAS, periodic acid Schiff; PCOS, polycystic ovary syndrome; PMSG, pregnant mare’s serum gonadotropin; POF, premature ovarian failure; X-gal, X-galactosidase; ZP, zona pellucida.

Received for publication July 31, 2003. Accepted for publication December 23, 2003.


   REFERENCES
TOP
 ABSTRACT
 INTRODUCTION
 RESULTS
 DISCUSSION
 MATERIALS AND METHODS
 REFERENCES
 

  1. Ueno N, Ling N, Ying SY, Esch F, Shimasaki S, Guillemin R 1987 Isolation and partial characterization of follistatin: a single-chain Mr 35,000 monomeric protein that inhibits the release of follicle-stimulating hormone. Proc Natl Acad Sci USA 84:8282–8286[Abstract/Free Full Text]
  2. Esch FS, Shimasaki S, Mercado M, Cooksey K, Ling N, Ying S, Ueno N, Guillemin R 1987 Structural characterization of follistatin: a novel follicle-stimulating hormone release-inhibiting polypeptide from the gonad. Mol Endocrinol 1:849–855[Abstract]
  3. Shimasaki S, Koga M, Esch F, Cooksey K, Mercado M, Koba A, Ueno N, Ying SY, Ling N, Guillemin R 1988 Primary structure of the human follistatin precursor and its genomic organization. Proc Natl Acad Sci USA 85:4218–4222[Abstract/Free Full Text]
  4. Nakamura T, Takio K, Eto Y, Shibai H, Titani K, Sugino H 1990 Activin-binding protein from rat ovary is follistatin. Science 247:836–838[Abstract/Free Full Text]
  5. Welt C, Sidis Y, Keutmann H, Schneyer A 2002 Activins, inhibins, and follistatins: from endocrinology to signaling. A paradigm for the new millennium. Exp Biol Med (Maywood) 227:724–752[Abstract/Free Full Text]
  6. Vale W, Rivier J, Vaughan J, McClintock R, Corrigan A, Woo W, Karr D, Spiess J 1986 Purification and characterization of an FSH releasing protein from porcine ovarian follicular fluid. Nature 321:776–779[CrossRef][Medline]
  7. Schneyer AL, Rzucidlo DA, Sluss PM, Crowley Jr WF 1994 Characterization of unique binding kinetics of follistatin and activin or inhibin in serum. Endocrinology 135:667–674[Abstract]
  8. Shimonaka M, Inouye S, Shimasaki S, Ling N 1991 Follistatin binds to both activin and inhibin through the common subunit. Endocrinology 128:3313–3315[Abstract]
  9. Fainsod A, DeiBler K, Yelin R, Marom K, Epstein M, Pillemer G, Steinbeisser H, Blum M 1997 The dorsalizing and neural inducing gene follistatin is an antagonist of BMP-4. Mech Dev 63:39–50[CrossRef][Medline]
  10. Iemura S, Yamamoto TS, Takagi C, Uchiyama H, Natsume T, Shimasaki S, Sugino H, Ueno N 1998 Direct binding of follistatin to a complex of bone-morphogenetic protein and its receptor inhibits ventral and epidermal cell fates in early Xenopus embryo. Proc Natl Acad Sci USA 95:9337–9342[Abstract/Free Full Text]
  11. Gamer LW, Wolfman NM, Celeste AJ, Hattersley G, Hewick R, Rosen V 1999 A novel BMP expressed in developing mouse limb, spinal cord, and tail bud is a potent mesoderm inducer in Xenopus embryos. Dev Biol 208:222–232[CrossRef][Medline]
  12. Otsuka F, Moore RK, Iemura S, Ueno N, Shimasaki S 2001 Follistatin inhibits the function of the oocyte-derived factor BMP-15. Biochem Biophys Res Commun 289:961–966[CrossRef][Medline]
  13. Elvin JA, Yan C, Wang P, Nishimori K, Matzuk MM 1999 Molecular characterization of the follicle defects in the growth differentiation factor-9-deficient ovary. Mol Endocrinol 13:1018–1034[Abstract/Free Full Text]
  14. Shimasaki S, Zachow RJ, Li D, Kim H, Iemura S-I, Ueno N, Sampath K, Chang RJ, Erickson GF 1999 A functional bone morphogenetic protein system in the ovary. Proc Natl Acad Sci USA 96:7282–7287[Abstract/Free Full Text]
  15. Dube JL, Wang P, Elvin J, Lyons KM, Celeste AJ, Matzuk MM 1998 The bone morphogenetic protein 15 gene is X-linked and expressed in oocytes. Mol Endocrinol 12:1809–1817[Abstract/Free Full Text]
  16. Matzuk MM, Lu H, Vogel H, Sellheyer K, Roop DR, Bradley A 1995 Multiple defects and perinatal death in mice deficient in follistatin. Nature 372:360–363
  17. Guo Q, Kumar TR, Woodruff T, Hadsell LA, DeMayo FJ, Matzuk MM 1998 Overexpression of mouse follistatin causes reproductive defects in transgenic mice. Mol Endocrinol 12:96–106[Abstract/Free Full Text]
  18. Shimasaki S, Koga M, Buscaglia ML, Simmons DM, Bicsak TA, Ling N 1989 Follistatin gene expression in the ovary and extragonadal tissues. Mol Endocrinol 3:651–659[Abstract]
  19. Michel U, Albiston A, Findlay JK 1990 Rat follistatin: gonadal and extragonadal expression and evidence for alternative splicing. Biochem Biophys Res Commun 173:401–407[CrossRef][Medline]
  20. Lakso M, Pichel JG, Gorman JR, Sauer B, Okamoto Y, Lee E, Alt FW, Westphal H 1996 Efficient in vivo manipulation of mouse genomic sequences at the zygote stage. Proc Natl Acad Sci USA 93:5860–5865[Abstract/Free Full Text]
  21. di Clemente N, Wilson C, Faure E, Boussin L, Carmillo P, Tizard R, Picard JY, Vigier B, Josso N, Cate R 1994 Cloning, expression, and alternative splicing of the receptor for anti-Mullerian hormone. Mol Endocrinol 8:1006–1020[Abstract]
  22. Baarends WM, van Helmond MJ, Post M, van der Schoot PJ, Hoogerbrugge JW, de Winter JP, Uilenbroeck JT, Karels B, Wilming LG, Meijers JH, Themmen APN, Grootegoed JA 1994 A novel member of the transmembrane serine/threonine kinase receptor family is specifically expressed in the gonads and in mesenchymal cells adjacent to the Mullerian duct. Development 120:189–197[Abstract]
  23. Baarends WM, Uilenbroek JT, Kramer P, Hoogerbrugge JW, van Leeuwen EC, Themmen AP, Grootegoed JA 1995 Anti-mullerian hormone and anti-mullerian hormone type II receptor messenger ribonucleic acid expression in rat ovaries during postnatal development, the estrous cycle, and gonadotropin-induced follicle growth. Endocrinology 136:4951–4962[Abstract]
  24. Teixeira J, He WW, Shah PC, Morikawa N, Lee MM, Catlin EA, Hudson PL, Wing J, MacLaughlin DT, Donahoe PK 1996 Developmental expression of a candidate Mullerian inhibiting substance type II receptor. Endocrinology 137:160–165[Abstract]
  25. Racine C, Rey R, Forest MG, Louis F, Ferre A, Huhtaniemi I, Josso N, di Clemente N 1998 Receptors for anti-mullerian hormone on Leydig cells are responsible for its effects on steroidogenesis and cell differentiation. Proc Natl Acad Sci USA 95:594–599[Abstract/Free Full Text]
  26. Mishina Y 2001 The in vivo function of Mullerian-inhibiting substance during mammalian sexual development. In: Matzuk MM, Brown CW, Kumar TR, eds. Contemporary endocrinology: transgenics in endocrinology. Totowa, NJ: Humana Press Inc; 41–59
  27. Teixeira J, Maheswaran S, Donahoe PK 2001 Mullerian inhibiting substance: an instructive developmental hormone with diagnostic and possible therapeutic applications. Endocr Rev 22:657–674[Abstract/Free Full Text]
  28. Jamin SP, Arango NA, Mishina Y, Hanks MC, Behringer RR 2002 Requirement of Bmpr1a for Mullerian duct regression during male sexual development. Nat Genet 32:408–410[CrossRef][Medline]
  29. Soriano P 1999 Generalized lacZ expression with the ROSA26 Cre reporter strain. Nat Genet 21:70–71[CrossRef][Medline]
  30. Yan C, Wang P, DeMayo J, DeMayo F, Elvin J, Carino C, Prasad S, Skinner S, Dunbar B, Dube J, Celeste A, Matzuk M 2001 Synergistic roles of bone morphogenetic protein 15 and growth differentiation factor 9 in ovarian function. Mol Endocrinol 15:854–866[Abstract/Free Full Text]
  31. Matzuk MM, Finegold MJ, Su J-GJ, Hsueh AJW, Bradley A 1992 {alpha}-Inhibin is a tumor-suppressor gene with gonadal specificity in mice. Nature 360:313–319[CrossRef][Medline]
  32. Matzuk MM, Kumar TR, Shou W, Coerver KA, Lau AL, Behringer RR, Finegold MJ 1996 Transgenic models to study the roles of inhibins and activins in reproduction, oncogenesis, and development. Recent Prog Horm Res 51:123–157
  33. Kumar TR, Palapattu G, Wang P, Woodruff TK, Boime I, Byrne MC, Matzuk MM 1999 Transgenic models to study gonadotropin function: the role of follicle-stimulating hormone in gonadal growth and tumorigenesis. Mol Endocrinol 13:851–865[Abstract/Free Full Text]
  34. Couse JF, Hewitt SC, Bunch DO, Sar M, Walker VR, Davis BJ, Korach KS 1999 Postnatal sex reversal of the ovaries in mice lacking estrogen receptors {alpha} and ß. Science 286:2328–2331[Abstract/Free Full Text]
  35. Behringer RR, Cate RL, Froelick GJ, Palmiter RD, Brinster RL 1990 Abnormal sexual development in transgenic mice chronically expressing Müllerian-inhibiting substance. Nature 345:167–170[CrossRef][Medline]
  36. Dong J, Albertini DF, Nishimori K, Kumar TR, Lu N, Matzuk MM 1996 Growth differentiation factor-9 is required during early ovarian folliculogenesis. Nature 383:531–535[CrossRef][Medline]
  37. Meinhardt A, O’Bryan MK, McFarlane JR, Loveland KL, Mallidis C, Foulds LM, Phillips DJ, de Kretser DM 1998 Localization of follistatin in the rat testis. J Reprod Fertil 112:233–241[Abstract]
  38. Hartmann BW, Kirchengast S, Albrecht A, Laml T, Soregi G, Huber JC 1997 Androgen serum levels in women with premature ovarian failure compared to fertile and menopausal controls. Gynecol Obstet Invest 44:127–131[Medline]
  39. Matzuk MM, Kumar TR, Bradley A 1995 Different phenotypes for mice deficient in either activins or activin receptor type II. Nature 374:356–360[CrossRef][Medline]
  40. Kawakami S, Fujii Y, Winters SJ 2001 Follistatin production by skin fibroblasts and its regulation by dexamethasone. Mol Cell Endocrinol 172:157–167[CrossRef][Medline]
  41. Nakamura K, Matzuk MM, Gerstmayer B, Bosio A, Lauster R, Miyachi Y, Werner S, Paus R 2003 Control of pelage hair follicle development and cycling by complex interactions between follistatin and activin. FASEB J 17:497–499[Abstract/Free Full Text]
  42. Jhaveri S, Erzurumlu RS, Chiaia N, Kumar TR, Matzuk MM 1998 Defective whisker follicles and altered brainstem patterns in activin and follistatin knockout mice. Mol Cell Neurosci 12:206–219[CrossRef][Medline]
  43. Roberts VJ, Bentley CA, Guo Q, Matzuk MM, Woodruff TK 1996 Tissue-specific binding of radiolabeled activin A by activin receptors and follistatin in postimplantation rat and mouse embryos. Endocrinology 137:4201–4209[Abstract]
  44. Kojima I, Maeshima A, Zhang YQ 2001 Role of the activin-follistatin system in the morphogenesis and regeneration of the renal tubules. Mol Cell Endocrinol 180:1