Suppression of Ovarian Follicle Activation in Mice by the Transcription Factor Foxo3a

See allHide authors and affiliations

Science  11 Jul 2003:
Vol. 301, Issue 5630, pp. 215-218
DOI: 10.1126/science.1086336


Foxo transcription factors have been implicated in diverse biological processes, including metabolism, cellular stress responses, and aging. Here, we show that Foxo3a–/– female mice exhibit a distinctive ovarian phenotype of global follicular activation leading to oocyte death, early depletion of functional ovarian follicles, and secondary infertility. Foxo3a thus functions at the earliest stages of follicular growth as a suppressor of follicular activation. In addition to providing a molecular entry point for studying the regulation of follicular growth, these results raise the possibility that accelerated follicular initiation plays a role in premature ovarian failure, a common cause of infertility and premature aging in women.

The FOXO subfamily of forkhead transcription factors consists of Foxo3a (FKHRL1), Foxo1 (FKHR), and Foxo4 (AFX), all downstream effectors of the PTEN/PI3K/AKT pathway (1). In Caenorhabditis elegans, systematic genetic analyses have revealed the existence of a conserved insulinlike signaling pathway involved in development, longevity, and fertility (25). Conservation of this pathway has fueled speculation that Foxo factors regulate related biological processes in mammals (6), prompting us to generate mice bearing a null mutation in the Foxo3a locus (fig. S1 and supporting online text). Despite a broad pattern of expression (fig. S1, D and E) (79), Foxo3a–/– animals were outwardly normal and did not show a prominent cancer-prone condition, abnormal weight gain (Fig. 1A), or statistically significant differences in mortality up to 48 weeks of age. Foxo3a–/– animals did exhibit some physiologic abnormalities, consistent with the view that Foxo3a serves diverse physiologic roles. Peripheral blood smears from Foxo3a–/– animals revealed hematologic abnormalities (apparent as early as 3 weeks of age), including a mild compensated anemia with associated reticulocytosis. Foxo3a–/– animals also exhibited a decreased rate of glucose uptake in glucose tolerance tests after an overnight fast (10).

Fig. 1.

Phenotypic characterization of Foxo3a–/– mice. (A) Normal weight gain in Foxo3a–/– females. Males also exhibited normal weight gain. (B) Continuous breeding assay starting at 6 weeks of age, showing cumulative number of progeny per female.

Foxo3a–/– and control females bore first litters of similar sizes (supporting online text), indicating normal sexual maturation. However, in contrast to males, Foxo3a–/– females exhibited a marked age-dependent decline in reproductive fitness and were sterile by 15 weeks of age (Fig. 1B). Gross and histologic evaluation revealed no abnormalities of Müllerian structures or the pituitary in Foxo3a–/– females. Ovaries from Foxo3a–/+ and Foxo3a–/– females showed no size differences through postnatal day (PD) 3. However, by PD8 to 14, ovaries from Foxo3a–/– mice were consistently enlarged (Fig. 2A).

Fig. 2.

Gross and histologic analysis of Foxo3a–/– and control ovaries. (A) Size differences in Foxo3a–/+ and Foxo3a–/– ovaries; scale bar, 1 mm. (B) Histology of Foxo3a–/+ and Foxo3a–/– ovaries at different ages. Scale bar = 100 μ for all panels, except that at PD3, scale bar = 20 μ. All sections were stained with hematoxylin and eosin except for those at 8.5 weeks and the inset sections at 4.5 weeks, which were stained with periodic acid–Schiff. For each time point, insets are at the same magnification.

In mammals, follicular growth is irreversible, and follicles recruited from the resting (primordial) follicle pool to the growing pool undergo apoptotic death (atresia) if not selected for further growth at subsequent stages of maturation (11). Follicular activation is characterized by oocyte growth and the transition of squamous to cuboidal granulosa cells (GCs), followed by GC proliferation (12). The activation of individual follicles involves unknown triggering mechanisms intrinsic to the ovary (11), although circulating factors likely modulate the overall rate of initiation (13). Follicular activation is independent of pituitary gonadotropins (14).

PD3 Foxo3a–/+ and Foxo3a–/– ovaries had similar numbers of oocytes (Fig. 3A) and patterns of apoptosis (10). By PD14, however, Foxo3a–/– ovaries contained markedly elevated numbers of early-growing follicles characterized by an increased oocyte diameter and flattened GCs; mitotic activity began in more advanced primary follicles with cuboidal GCs (Fig. 2B). This pervasive activation of follicular growth in Foxo3a–/– females resulted in the progression of increased numbers of follicles to more advanced stages of follicular development. Compared with controls, PD14 Foxo3a–/– ovaries showed 1.9- and 2.1-fold increases in the number of primary and secondary follicles (Fig. 3C, P = 0.012 and 0.000) and in the number of atretic secondary follicles (Fig. 3C, P = 0.016). At 4.5 weeks of age, around the onset of sexual maturity, Foxo3a–/– ovaries still contained large numbers of early-growing follicles and greater numbers of primary and secondary follicles relative to controls (3.5 and 3.9 times as many, respectively) (Fig. 3D).

Fig. 3.

Histomorphometric analyses of mouse ovaries. (A) Relative oocyte numbers at PD3. Numbers represent total counts of every fifth section from serially sectioned ovaries (n = 4 animals per genotype). (B) Relative primordial/primary follicle counts in aged females (48 weeks). Numbers represent total counts of every 10th section from serially sectioned ovaries (n = 3 animals per genotype). (C to E) Relative follicle counts at PD14, 4.5 weeks, and 8.5 weeks of age. Numbers represent the average counts of each follicle type per section (n = 3 animals per genotype). (F) Primordial oocyte diameters in Foxo3a+/+ and Foxo3a–/– females at PD3, PD8, PD14, and 4.5 weeks of age. (G) Estimated volumes of newborn Foxo3a+/+ (n = 6), Foxo3a–/+ (n = 10), and Foxo3a–/– (n = 4) mouse ovaries grown in culture. (H) Primordial oocyte diameters in Foxo3a+/+, Foxo3a–/+, and Foxo3a–/– ovaries from (G) after 8 days in culture. For all panels, data are represented as mean values with error bars representing the SEM. P values are shown above the bars and were calculated by the exact t test.

At 8.5 weeks of age, Foxo3a–/– ovaries exhibited a 14-fold increase in zona pellucida remnants (ZPRs), which are remnants of oocytes that have undergone atresia subsequent to synthesis of the zona pellucida (15, 16) (Figs. 2B and 3E). This increase likely reflects widespread follicular initiation followed by secondary follicular atresia. All oocytes in remaining primordial follicles in Foxo3a–/– ovaries appeared large and misshapen (Fig. 2B, inset), and many had fragmented nuclei. Thus, the widespread follicular activation in Foxo3a–/– females eventually led to follicular and oocyte atresia in follicles that did not progress. The smaller litter sizes in aging Foxo3a–/– females (Fig. 1B) are consistent with this widespread oocyte degeneration by 8.5 weeks of age. At 18 weeks, Foxo3a–/– ovaries contained many ZPRs and occasional old corpora lutea but were devoid of viable follicles, in contrast to controls (Fig. 2B). In aged Foxo3a–/+ and Foxo3a+/+ females (48 weeks), counts of primordial plus primary follicles were similar (i.e., there was no evidence of Foxo3a haploinsufficiency) (Fig. 3B).

Normal primordial follicles (those without enlarged oocytes) were absent in serially sectioned Foxo3a–/– ovaries by PD14, indicating total depletion of the reserve follicle pool by this age. The complete absence of Foxo3a function thus resulted in global activation of primordial follicles before the onset of sexual maturity, leading secondarily to the premature depletion of ovarian follicles. We conclude that Foxo3a is a critical regulator and suppressor of ovarian follicle activation.

Oocyte enlargement in primordial follicles was evident by PD8 (Fig. 3F) and occurred in the great majority of oocytes by PD8, consistent with global activation (fig. S2). Activated oocytes in Foxo3a–/– females grew until 4.5 weeks of age despite a lack of GC growth and maturation (Fig. 3F). To investigate whether Foxo3a functions within the ovary, we explanted newborn ovaries (17). Enlargement was evident in Foxo3a–/– ovaries after 8 days of in vitro culture (Fig. 3G) and was associated with increased oocyte diameter (Fig. 3H and fig. S3) similar in magnitude to that observed in Foxo3a–/– ovaries in vivo (Fig. 3F). The recapitulation of these phenotypes in culture is consistent with the hypothesis that Foxo3a functions within the ovary itself to suppress follicular activation and that extrinsic circulating factors do not play a major role in this Foxo3a-mediated effect. However, it remains formally possible that Foxo3a operates in an ovarian nonautonomous manner.

Serum follicle-stimulating hormone (FSH) and luteinizing hormone (LH) levels were measured at 20 weeks of age. Compared with Foxo3a+/+ controls, Foxo3a–/– females showed significant elevations in both FSH and LH levels (Fig. 4A, P = 0.000 and 0.001). Therefore, Foxo3a–/– females exhibit classic signs of hypergonadotropic hypogonadism secondary to premature ovarian failure (POF), indicating a normal pituitary response to follicular depletion (18). We also obtained serial vaginal smears for cytologic evaluation. Whereas all heterozygous and wild-type control females cycled regularly during 4-week observation periods, Foxo3a–/– females cycled sporadically from 6 to 15 weeks but were subsequently acyclic (10), consistent with depletion of the follicular reserve and consequent ovarian failure.

Fig. 4.

Hormonal studies of adult female mice. (A) Serum FSH and LH measurements of Foxo3a+/+ (n = 8), Foxo3a–/+ (n = 4), and Foxo3a–/– (n = 12) females at 20 weeks of age. (B) Average number of released oocytes per female after superovulation (n = 4 to 9 animals per genotype) at 3.0 and 4.5 weeks of age. For all panels, data are represented as mean values ± SEM. P values are shown above the bars and were calculated by the exact t test.

We induced ovulation by administration of gonadotropins at 3.0 and 4.5 weeks of age. At 3 weeks, the numbers of eggs released after superovulation were similar among +/+, –/+, and–/–females (Fig. 4B), and >90% of these released eggs were competent to initiate embryonic development in culture (10), demonstrating that Foxo3a is not required for gonadotropin responsiveness, as expected given that Foxo3a–/– females bore litters before follicular depletion, which occurs at around 15 weeks of age. At 4.5 weeks, more eggs were released by Foxo3a–/– females than by controls (Fig. 4B, P = 0.039), which is probably a reflection of the greater numbers of advanced follicles observed in Foxo3a–/– females at this age. Eggs from superovulated Foxo3a–/– females were somewhat larger than those from controls, likely secondary to the abnormal, prolonged interval between oocyte activation and subsequent follicular maturation (supporting online text). Taken together, these findings suggest that a lack of Foxo3a function results in an intrinsic ovarian defect specifically involving follicular growth activation, without directly impairing other aspects of follicular maturation or reproduction.

We confirmed that Foxo3a is expressed in the mouse ovary before the earliest manifestation of the Foxo3a–/– phenotype (supporting online text). FOXO3a expression is also readily detectable in human newborn and adult ovaries, as has been shown for adult human ovaries (7) and in a wide range of other tissues (fig. S4).

POF (secondary infertility with persistently elevated gonadotropin levels before the age of 40) is a common cause of infertility and premature aging in women, with an estimated 1% incidence (18). Some cases of POF clearly have a genetic basis (20). Interestingly, haploinsufficiency for another distantly related (21) forkhead transcription factor, FOXL2, results in eyelid anomalies and POF. In contrast to Foxo3a, Foxl2 has a limited pattern of expression in the mouse; it is expressed in the ovaries and in the mesenchyme of developing eyelids (22). Primordial follicles in affected women are growth-arrested, indicating that mutations of these forkhead transcription factors result in POF through distinct mechanisms (23). Oocyte apoptosis is likely a relevant mechanism in the genesis of POF in patients who have undergone chemo- or radiotherapy (24). However, the vast majority of cases of POF are idiopathic (18). Studies of POF have been hindered by the inaccessibility of the ovary and by the occurrence of follicle depletion before the onset of symptoms. Mouse models of POF such as the Foxo3a–/– knockout mouse described here provide a molecular entry point to delineate further the regulation of ovarian follicle activation and a system to determine how misregulation of this process can lead to POF. Our findings that increased follicle activation can result in POF in the mouse suggest that a similar mechanism (perhaps influenced by genetic factors) could contribute to follicle depletion in women with POF. Further study of the Foxo3a pathway might lead to the development of improved contraceptives that delay follicular activation, preserving the follicular reserve pool until fertility is desired.

Supporting Online Material

Materials and Methods

SOM Text

Figs. S1 to S4


References and Notes

View Abstract

Navigate This Article