Conditional ablation of Insr and Igf1r using Esr2-Cre leads to abnormal ovarian follicle development and infertility in mice

Abstract

Conditional ablation of Igf1r in early folliculogenesis has demonstrated the necessity of insulin signaling to progress to the antral stage, whereas ablation of both Insr and Igfr1 in the periovulatory window allows the formation of the antrum but reduces the efficiency of ovulation and subsequent luteinization. For this study, we examined the independent and shared actions in single and double knockouts (DKO) for Insr and Ifg1r using Esr2-Cre. As this recombinase is active during neonatal ovarian development and the initial wave of folliculogenesis, we hypothesized that abnormalities in ovary formation and establishment of the initial follicle pool would occur, which could alter female reproductive lifespan. We found that ablation of both receptors led to a delay in puberty, altered mating frequency, and ultimately infertility for Igf1r^d/d and DKO females. Quantitation of germ cell cyst breakdown, and formation of primordial and primary follicles were normal in the neonatal window and at puberty, suggesting insulin signaling was not essential for establishment of ovarian reserve. However, the loss of IGF1R signaling impaired transition from primary to secondary follicles, which was worsened when IGF ligand cross-reactivity from INSR signaling was lost in DKO mice. DKO mice also exhibited abnormal follicle activation in the absence of hormone stimulation, but no subsequent proliferation of granulosa cells or antrum formation occurred. In adult mice, loss of either receptor disrupted estrous cyclicity, with DKO mice rarely leaving metestrus indicating abnormal regulation of the HPG axis contributing to subfertility and infertility observed in single and double receptor knockouts.

Summary Sentence

Granulosa-specific loss of IGF1R and INSR does not significantly disrupt establishment of the follicle reserve, but female mice are infertile due to impaired growth of follicles to the antral stage.

Graphical Abstract

Introduction

Insulin receptors are critical activators of several developmental and physiological processes including cellular differentiation, migration, proliferation, and survival [1, 2]. The primary insulin receptor (INSR) and insulin-like growth factor one receptor (IGF1R) belong to class II receptor tyrosine kinase superfamily. In mammals, they exhibit high structural homology, as well as amino acid identities ranging from 45–65% in the ligand binding domain and 65–85% in the tyrosine kinase domain [3, 4]. Both receptors form heterotetramers comprised of two extracellular alpha subunits and two transmembrane beta subunits, bound by disulfide bridges [5]. Their best-known ligands: INS, IGF1 and IGF2 have a potential to bind each receptor with different affinity (Reviewed in [6, 7]). Thus, when studied mechanistically, the potential activity of both receptors should be considered in order to avoid any functional redundancy or compensation.

Unfortunately, mice null for Insr and Igf1r, as well as combinations of their ligands leads to abnormal development, severe hyperglycemia or hyperketonemia, early lethality and reduced body size which has precluded using those models to investigate their roles in reproductive processes [8–10]. This has led many groups to employ conditional CRE/LoxP knockout models in different cell-type and temporal–spatial windows in the gonads and female reproductive tract. Receptor expression in germ cells is low, and their ablation in sperm (using Ngn3-Cre) or oocytes (using Gdf9-Cre) produces gametes that combine efficiently in vitro and produce offspring [11]. However, Sry is not properly induced. Thus, XY pups become females, and XX ovarian development is delayed, but ultimately elicits no severe defects in adult fertility [12]. We have recently used Amhr2-Cre to ablate Insr and Igf1r in granulosa cells of secondary follicles and found that the Cre recombinase did not elicit sufficient follicular ablation of INSR and IGF1R to alter the expression of ovulation-promoting factors or steroidogenesis enzymes to hinder the normal progression of folliculogenesis and ovulation [13]. However, when we employed a later acting and higher efficiency Cre driver under the control of progesterone receptor, we found that mice lacking both INSR and IGF1R exhibited reduced ovulation with about 50% of oocytes remaining trapped in the corpus luteum which eventually were lost due to follicular atresia [14]. Luteinization was also impaired and DKO mice exhibited reduced production of progesterone. Pgr-Cre mediated Igf1r^d/d females were subfertile and DKO mice were infertile. However, reduced litter sizes were predominantly due to failures in implantation and decidualization as uterine ablation within the epithelium and stroma was also elicited by Pgr-Cre.

The subfertility of Igf1r^d/d female mice using the Esr2-Cre driver has previously been characterized and it was found that when paired with Cyp19-Cre, female mice were completely infertile. In their analysis, the fertility block was due to failure of the follicles to progress past the secondary stage and form an antrum [15]. While they see no fertility problems in Insr granulosa cell knockout mice, deletion of Igf1r leads to infertility due to disruption of follicular progression, and absence of antrum formation [15]. Secondary follicles also exhibited an elevated amount of granulosa cell apoptosis. Further characterization revealed significantly reduced estrogen production (~10% relative to WT mice) and no release of oocytes into the oviduct even after the administration of superovulatory gonadotropins [15]. The goal of the prior report was to examine the synergy between FSH and IGF signaling in granulosa cell function that had previously been investigated using primarily only in vitro models. We hypothesized that in the absence of residual cross-reacting IGF or INS stimulation of the non-cognate receptor in single knockouts, that ovarian development and establishment of the ovarian reserve might be impacted. This represents a novel window compared to our prior Pgr-Cre model were ablation of both receptors after antral follicle formation resulted in impaired ovulation [14], and the Amhr2-Cre deletion model where receptor proteins present at the time of gene deletion were able to support normal ovulation by carrying the follicle past the antral stage [13]. In the present study, we sought to understand the involvement of INSR and IGF1R in early follicular development in the neonatal window. In addition, this is the first study to examine dual receptor ablation using the Esr2-Cre in order to determine novel features of insulin signaling that may have been previously masked by having at least one functional receptor in granulosa cells. For simplicity’s sake, unless specifically stated, we have used insulin signaling to refer to insulin-related receptor signals that could be transmitted by INSR, IGF1R, or a combination of both receptors in response to their primary ligand or lower affinitiy ligand binding to the non-cognate receptor (e.g. INS-IGF1R, or the converse).

Materials and methods

Experimental animals and generation of conditional insulin receptor KO mice

All animal handling was done according to NIH guidelines and in compliance with the Southern Illinois University Carbondale Institutional Animal Care and Use Committee (protocols 16–043, 19–007) and Washington State University (protocols 6757, 6767). Mice used for the study were maintained on a C57BL6 genetic background. All animals were housed under a 12 h light, 12 h dark schedule, and fed Purina Labdiet 5008 mouse chow. Genomic DNA was collected from tail snips for genotyping by PCR using the primer sequences shown in Supplementary Table S1. We generated mice carrying double floxed alleles for Insr and/or Igf1r from two previously generated lines where LoxP sites were introduced to flank exon 4 of Insr (Insr^fl/fl, JAX6955, [16]) and exon 3 of Igf1r (Igf1r^fl/fl, JAX12251, [17]). For conditional ablation of Insr or Igf1r in granulosa cells of growing follicles, we bred female Insr^fl/fl, Igf1r^fl/fl or double floxed Insr^fl/fl/ Igf1r^fl/fl mice to male Cre^+/− mice expressing the recombinase under the control of the Esr2 promoter (Esr2-iCre, JAX28717, [18]) with heterozygocity for either Insr and Igf1r so that conditional null and control littermates could be examined where desired. The characterization of intermediate floxed/null mice was performed in a prior study [15] and found to exhibit normal ovarian function.

Superovulation

For the superovulation studies, female mice aged PND21–28, selected by size of at least 15 g for maximal response, were treated with 5 IU equine chorionic gonadotropin (eCG; Biovendor RP178272, Asheville, NC). This treatment was followed 48 hours later by treatment with 4 IU human chorionic gonadotropin (hCG; Sigma C0434, St. Louis, MO). Both eCG and hCG were dissolved in 0.85% saline solution and injected intraperitoneally (IP) in a total volume of 0.1 mL. Mice were euthanized by CO₂ asphyxiation 0–72 hours after hCG injection and ovaries collected for histological or molecular analyses.

Vaginal opening and estrous cycle assessment

Animals were checked daily for signs of vaginal opening starting at postnatal day 20 (PND20). Once the vaginal opening was detected, vaginal smears were collected daily for 22–24 days beginning 1 week after vaginal opening. Smears were collected at the same time each morning (9 a.m.) by gently flushing the vagina three times with sterile PBS and the eluate transferred to glass slides and stained with 0.2% methylene blue to differentiate the proportions of nucleated epithelial cells, cornified squamous epithelial cells, and leukocytes by light microscopy as previously described [19]. Diestrus consisted predominantly of leukocytes, proestrus showed both leukocytes and nucleated epithelial cells in approximately equal numbers, estrus was characterized by cornified squamous epithelial cells and metestrus consisted of equal numbers of leukocytes and cornified epithelial cells with translucent nuclei.

Histology and immunohistochemistry

Ovaries were extirpated and fixed in 4% paraformaldehyde in PBS and then embedded in paraffin. Sections (5 μm) were mounted on glass slides, deparaffinized with xylene, and rehydrated in 100%, 95%, 70% ethanol, and rinsed in distilled water (dH₂O). For histology, slides were dipped in freshly filtered hematoxylin for 30 seconds to 1 minute. Hematoxylin is a basic dye that has an affinity for nucleic acids in the nucleus. Excess hematoxylin was washed off in dH₂O and the slides dipped in diluted ammonia hydroxide. Slides were rinsed again in dH₂O and the slides were dehydrated with xylene (x3) and mounted with Permount mounting medium (Thermo Fisher). Immunohistochemistry and immunofluorescence were performed as described previously [14, 20]. Antigen retrieval was performed using 10 mM citrate buffer at 110°C for 10 min. Supplementary Table S2 shows the primary antibodies and conditions used for each experiment.

Follicle analysis and quantitation

Embedded ovaries were serially sectioned at 5 μm thickness and every fifth serial section was mounted on a glass slide and stained with H&E for histological evaluation of follicle numbers. Primordial follicles contained an oocyte surrounded by a single layer of squamous granulosa cells; primary follicles contained an oocyte surrounded by a single layer of cuboidal granulosa cells; secondary follicles contained an oocyte surrounded by at least two layers of cuboidal granulosa cells and theca cells; antral follicles contained an oocyte surrounded by multiple layers of cuboidal granulosa cells with a fluid filled antral space and theca cells; and CLs were much larger structures containing luteal cells. All primordial and primary follicles were counted in each section regardless of nuclear material in the oocyte, whereas secondary and antral follicles where the oocyte’s nucleolus was visible were counted to avoid the risk of double counting the larger follicle types that can span multiple sections. Follicles that were comprised of 10% apoptotic bodies were scored as atretic. Follicles transitioning between stages were counted as follicles within the more mature stage of the two stages. The total numbers of all follicles, total numbers of each type of follicle, percentages of each follicle type, and CL numbers per number of sections counted were recorded. For germ cell nest analysis, if two or more germ cells were adjacent and clearly defined within a single unit of pre-granulosa cells, they were considered a nest. The average follicle counts for two observers blinded with respect to corresponding genotype were averaged for each mouse.

Quantitative RT-PCR

Total RNA was isolated from testes using TRIzol^® reagent (Invitrogen), and then RT was performed using High-Capacity cDNA Reverse Transcription Kit (Applied Biosystems). Real-time PCR was performed using PowerUp SYBR Green Master Mix (Applied Biosystems) in a BioRad Cfx96 Real-time PCR System. Real-time PCR was performed using the following protocol: 2 min at 95°C, 40–45 cycles of denaturation (15 sec at 95°C) and annealing/extension (1 min at 60°C), and a final step of melting curve analysis. As an internal control, Rpl19 was used. The relative levels of mRNA were calculated using the 2^−ΔΔCt method. The primers for gene amplification have been previously reported [20–23] or are listed in Supplementary Table S2.

Statistical analysis

Data were analyzed using Prism software (Ver. 5.0, GraphPad). All experimental data are presented as mean with standard error of the mean (SEM). For the analyses of grouped data, one-way ANOVA, paired with Tukey’s or Dunnett’s post hoc test was employed. Tukey’s multiple range test was used to analyze the differences between individual means of multiple grouped data. Dunnett’s tests was used when comparisons were made between a control group and more than one experimental group to identify differences between individual means. All data met necessary criteria for ANOVA analysis including equal variance as determined by Bartlett’s test. P value less than 0.05 was considered to be statistically significant.

Results

Conditional ablation of Insr and Igf1r in primordial granulosa cells

To examine the role of insulin signaling, which we apply generally to INS-INSR, IGF1-IGF1R, or instances where cross-reactivity occurs, in the ovary without the severe defects reported for mice with global deletion of INSR or IGF1R [8–10], we generated transgenic mice that carry granulosa cell-specific deletions of either receptor insulin receptor (Insr^d/d) or insulin-like growth factor 1 receptor (Igf1r^d/d), or double ablation of both receptors (DKO) by cross breeding floxed Insr and Igf1r mice with Esr2-Cre mice [18]. The DKO eliminates the potential for confounding redundancy from ligand cross-talk with non-cognate receptors. Analysis of mRNA expression in whole ovaries exhibited a significant reduction (45–55%) of Insr or Igf1r in the expected conditional knockout lines in post-natal day 8 (PND8) ovaries compared to wild type (WT) control mice (Fig. 1). Few granulosa cells exhibited recepter protein (less than 10%) after conditional ablation with Esr2-Cre, but some oocytes retained receptor expression, particularly INSR, in DKO mice. Thus, residual mRNA expression is due to retained transcription in non-granulosa cells such as the thecal layer or stroma where the Cre is not active. The Esr2-Cre exhibits low-level activity in the oocyte, either via unexpected promoter activity or the transfer of Cre-encoding mRNA from granulosa cells to the oocyte [18], which may account for absence of receptors in the oocyte of some (less than 5%) follicles.

Figure 1.

Confirmation of Insr and Igf1r conditional knockout in granulosa cells. The relative mRNA expression of Insr and Igf1r was determined in total tissue RNA from PND8 ovaries. Compared to WT controls, mRNA levels of both Insr and Igf1r were reduced in DKO uteri and specifically reduced in the appropriate single knockout line as determined by quantitative real-time RT-PCR (qPCR, n = 12). INSR and IGF1R protein were nearly eliminated specifically in granulosa cells of ovarian follicles at PND8 and in ovaries at vaginal opening. Representative images, n = 4 per genotype. Bar heights indicate the Mean ± SEM for each genotype. Letters denote means which are significantly different (P < 0.05), one way ANOVA with Tukey multiple comparison post-test.

Loss of ovarian INSR and IGF1R alters follicular development and impairs fertility

In mice, one external indicator of the onset of puberty is the opening of the vagina. Opening is induced by an increase in estradiol (E2), mainly produced by large secondary follicles. In WT mice, vaginal opening occurred around PND28 and neither Insr^d/d or Igf1r^d/d female mice exhibited precocious nor delayed puberty and could be mated in the same window as WT mice. In contrast, DKO mice exhibited an average delay of 3 days compared to WT controls (Fig. 2A). Female fertility was assessed by mating WT, single knockout and double knockout females with wildtype males of proven fertility. Adult females were paired with males on the night of estrous and examined for vaginal plugs the morning after as an indicator of successful mating. For most WT and Insr^d/d females, plugs were observed within 1–2 days after the start of mating. The average time to plug for Igf1r^d/d mice was 5 ± 0.6 days and 12.6 ± 3.9 days for DKO females, which were highly variable but significantly different than WT females (Fig. 2A). One DKO female did not exhibit a plug for the duration of the 2-month breeding trial and was excluded from further fertility analyses. After observing a plug, the females were separated from males, and pregnancy outcomes determined. Mean litter sizes for WT dams was 8.12 ± 0.35 pups and 6.13 ± 1.36 pups for Insr^d/d dams, which were not significantly different even with two dams that failed to produce pups included in the comparison. However, no pups were produced from Igf1r^d/d or DKO breeding pairs (Fig. 2A).

Figure 2.

Subfertility in Esr2-Cre mediated Insr and Igf1r conditional knockout mice. (A) Onset of puberty was assessed by the timing of vaginal opening. DKO mice had a significant delay (~5 days) in vaginal opening compared to the WT and single knockout females. Eight-week-old females (n = 5) were paired with males of proven fertility and the number of days necessary to observe a copulatory plug is shown. Subsequent first litter sizes from each pairing are shown (n = 8) for each genotype. Exclusion of the two Insr^d/d dams that failed to produce litters does not alter the statistical outcome. Bar heights indicate the Mean ± SEM for each genotype. Letters denote means which are significantly different (P < 0.05), one way ANOVA with Tukey multiple comparison post-test. (B) Representative images (n = 12) showing histological architecture of the ovary at puberty. The ovaries from DKO mice were smaller in diameter and characterized by cross-sections with multiple follicles with activated oocytes with flat granulosa cells and no obvious growing follicles. Subsequently, mice were superovulated (n = 6) to determine whether exogenous gonadotropin stimulation would induce follicular growth. At 48 h post eCG stimulation, follicular development should be actively progressing but only a few secondary follicles were detected. (C) Gross inspection of uterine horns after hormone treatment did not reveal the typical endometrial thickening response and ovaries were noticeably smaller than WT mice even 24 h after hCG injection. No corpora lutea were observed in any DKO mice.

To understand the involvement of INSR and IGF1R in follicular development, ovaries were collected at the onset of puberty. Histological evaluation of DKO ovaries at vaginal opening revealed much smaller ovarian diameters with no mature follicles present (Fig. 2B). Oocytes were larger and irregularly shaped with little or no proliferation of granulosa cells, suggesting that normal ovulation is not possible. To determine whether hyperstimulation with exogenous gonadotropins could improve follicle growth and ovulation success, mice were subjected to standard superovulatory doses of eCG and hCG to mimic FSH and LH. Subsequent histological examination of eCG-primed ovaries showed an arrest in follicular development (Fig. 2B). While ovarian cross-sections were slightly larger than those at vaginal opening, only a few secondary follicles were detected, all of which were at the early stages of development. The rest of the follicles were at either primordial or primary stages, indicating that insulin signaling in granulosa cells is important in follicular development, and that DKO females do not respond to exogenous eCG stimulus. Gross anatomy of the uterine horns showed no characteristic thickening of the endometrium from fluid retention and proliferation of the stroma in DKO mice (Fig. 2C). No corpora lutea were observed at 24 h post-hCG (Fig. 2C) or at any later time points in subsequent analyses.

In our previous analyses of Insr and Igf1r conditional knockout mice using Amhr2-Cre and Pgr-Cre, we observed no significant differences in the numbers of follicles at each developmental stage [13, 14]. However, the severity of the fertility phenotype, abnormal appearance of DKO, and the prior characterization of Igf1r^d/d mice by the Stocco lab [15], suggested that this was not likely to be the case with Esr2-Cre mediated Insr and Igf1r deletions. Thus, we serially sectioned ovaries to assess the distribution of follicle subtypes present at vaginal opening (Fig. 3). At vaginal opening, littermate WT ovaries were comprised of primordial (~70.5%), primary (~13.6%), secondary (~13.1%) and few antral follicles (~2.7%). Total primordial follicle counts varied between individuals in all genotypes tested, but no obvious trends or significance in mean follicles were observed. There tended to be fewer primary follicles in DKO mice, but the difference relative to WT mice was not significant (p > 0.07). In contrast, there was a significant reduction in secondary follicles in Igf1r^d/d and DKO mice. When present in DKO mice, secondary follicles contained oocytes surrounded by only 2–3 layers of granulosa cells indicating greater impairment of follicle growth (Fig. 3A). All three mutants had a significant reduction in antral follicles, with none being observed in any DKO ovary. DKO mice did possess a high percentage, 12.8%, of follicles with abnormal morphology characterized by potentially activated but not maturing oocytes where the single layer of cuboidal granulosa cells were detached from the oocyte (Fig. 3B).

Figure 3.

Analysis of follicular development at puberty in Insr and Igf1r conditional knockout mice. (A) Representative histological sections of ovaries from mice at vaginal opening. The scale bar = 100 μm. Primary, secondary, and antral follicles (An) are clearly present in WT and single receptor knockout mice. However, few growing follicles are observed in DKO mice and several abnormal follicles with activated oocytes and no granulosa cell proliferation are present (representative follicles are marked with *). (B) Total follicle counts from serially sectioned ovaries (n = 6) revealed no significant differences in primordial or primary follicles, but secondary and antral follicles were reduced in mice lacking one or both receptors. Bars indicate the Mean ± SEM for each genotype. Data is presented on a log scale to better delineate differences between groups were low numbers of each follicle type were observed. Letters denote means which are significantly different (P < 0.05), one way ANOVA with Tukey multiple comparison post-test.

Assessment of follicle pool establishment in Insr and Igf1r conditional knockout mice

In mice, the breakdown of germ cell cysts occurs shortly after birth, significant numbers of germ cells are lost due to apoptosis, and the formation of the primordial follicle pool is completed by postnatal day 4. Once finalized, the size of this pool is considered to define the reproductive longevity of adult mice (reviewed in [24–26]). To determine if loss of insulin signaling alters these processes, we determined the numbers of follicles present at PND1, PND4, and PND8. At PND1, a larger number of isolated germ cells were present in ovaries from DKO mice, but no significant differences in the number of subsequent primordial or primary follicles were observed in any group (Fig. 4A). Similarly, germ cell nests within ovaries of DKO mice had 3.59 ± 0.26 oocytes per nest, which was significantly more than the single knockouts or WT ovaries which averaged 2.6 to 2.8 cells/nest, suggesting a modest delay in nest breakdown (Fig. 4B). At PND4, the number of isolated germ cells that were not organized into follicles was reduced. However, no differences between genotypes were observed until the secondary follicle stage of development (Fig. 4C). DKO mice exhibited a 76% reduction in secondary follicles, 28.57 ± 7.21 compared to WT mice 119.2 ± 20.1.

Figure 4.

Establishment of the neonatal follicle reserve appears normal in mice lacking INSR and IGF1R in granulosa cells. Ovaries were serially sectioned and total follicle counts determined and analyzed as in Fig. 3. Unincorporated oocytes were significantly higher in ovaries of DKO mice at PND1 (n = 6–8), but no differences in formed primordial or primary follicles were observed. (B) Germ cell nests from DKO mice had an additional oocyte relative to WT mice or single knockouts (n = 10), suggesting selection for atresia may be altered. (C) Analysis of follicle distribution at PND4 revealed a significant decrease in secondary follicles (n = 6–8).

If formation of the initial follicle pool is proceeding in the established timeline, if FOXL2 signaling in granulosa cells and NOBOX is active in oocytes, all germ-cell nests should be broken into primordial follicles by PND8 [27, 28], and they should be steadily progressing into primary and secondary follicles [29]. There were no obvious differences in gross morphology or size of the ovaries between groups (Fig. 5A). Loss of both Insr and Igf1r signaling did not significantly alter primordial counts between groups, suggesting that primary follicle formation and/or survival occurred at the same rate as WT control follicles (Fig. 5B). However, the transition from primary to secondary follicles was reduced. While Insr and Igf1r single mutants showed a 2-fold decrease in secondary follicles (t-test significance P < 0.04 single knockout vs WT), this did not reach statistical significance by one-way ANOVA (P > 0.05) when all groups were compared together. However, PND8 DKO ovaries exhibited the same decrease in relative numbers of secondary follicles ~7.1% (308 ± 41) vs ~1.1% (55 ± 21.6), observed at PND4. The morphology of follicles present in DKO mice was largely normal, with oocytes remaining in close contact with surrounding cuboidal granulosa layers, in contrast to those present in DKO mice at vaginal opening (Fig. 3A). Taken together, these results suggest insulin signaling is not required for germ cell nest breakdown and organization of the follicle pool, but rather the potential recruitment and growth of follicles that are resident in the ovary.

Figure 5.

The primary follicle pool is unaltered in mice lacking Insr and Igf1r. (A) Representative histological sections of ovaries from conditional knockout mice at PND8. Primordial and primary follicles (1°) are clearly present in WT and single receptor knockout mice, but few secondary (2°) follicles were observed. The scale bar = 100 μm. (B) Ovaries were serially sectioned and total follicle counts determined and analyzed as in Fig. 3 (n = 5–10). The number secondary follicles in single receptor knockout mice appeared lower but was not significantly different. However, secondary follicles were significantly reduced to nearly absent in DKO mice. (C) Relative mRNA expression of insulin receptors, ligands, glucose-dependent transporter, and Foxl2 a driver of granulosa proliferation and development as determined by qPCR in WT mice. The data are presented as the means ± SEM (n = 6). Letters denote means which are significantly different (P < 0.05), one way ANOVA with Tukey multiple comparison post-test.

While Insr levels are static during the neonatal window (Fig. 5C), follicles enhance their ability to respond to insulin signaling as evidenced by the gradual increase in Igf1r between PND1 and PND8 in wild-type ovaries (Fig. 5C). Expression of the IGF1R ligand, Igf1 transiently increases at PND4 followed by significant increase in the local production of Ins1 in granulosa cells and the primary murine INSR ligand, Ins2 at PND8. While we did not examine the facilitated transport of glucose in growing follicles via Slc2a4 (aka, Glut4), its expression was enhanced at PND8. The increase in insulin signaling elements were correlated with a significant increase in the granulosa cell differentiation factor, Foxl2 (Fig. 5C).

Loss of INSR and IGF1R signaling alters the expression of genes involved in granulosa cell proliferation and follicular activation at PND8

The phenotype observed in dual insulin receptor mutant mice mimics that previously described for female Foxl2-null mice [30, 31]. Chiefly, inappropriate activation of oocytes with no subsequent proliferation of their surrounding granulosa cells. At PND8, qPCR analysis of total ovarian mRNA revealed a significant decrease in Foxl2 expression in Insr^d/d mice, that was further decreased in Igf1r^d/d and DKO mice (Fig. 6A). The localization of FOXL2 in granulosa cells of growing follicles was correlated to follicle size (i.e. the number of granulosa cell layers) which were relatively greater in secondary follicles from WT and Insr^d/d mice than those in Igf1r^d/d and DKO mice (Fig. 6B). Expression of the cell proliferation marker MKI67 was nearly absent in follicles of DKO mice (Fig. 6B). In addition, Amh which encodes a regulator of the primordial follicle pool by suppressing follicle activation was similarly decreased in all receptor mutants and nearly undetectable in DKO mice (Fig. 6A). The localization of AMH was prominent in the secondary follicles of WT mice and could be detected in Igf1r^d/d secondary follicles (Fig. 6B). However, it was greatly reduced in similar follicles from Igf1r^d/d mice and nearly absent in DKO mice (Fig. 6B). Gdf9 is an oocyte specific marker which should increase in expression level during the proper activation of the follicle, but in each case mRNA levels were significantly decreased after Insr ablation and further decreased with Igf1r loss (Fig. 6A).

Figure 6.

Loss of Insr and Igf1r impairs follicle development and growth. (A) Relative mRNA expression of markers of granulosa cell development and follicular activation were determined as in Fig. 5. (n = 12). (B) Immunofluorescent localization of oocyte factor FOXO3, granulosa marker FOXL2, the proliferation marker MKI67, and the regulator of follicle selection AMH. Protein expression was well-correlated to mRNA levels and were greatly reduced in Igf1r^d/d and DKO mice. (*) Abnormal follicles where FOXO3 disappeared from the oocyte in accordance with normal follicle dynamics but no subsequent proliferation of the granulosa cell layer was observed.

To determine whether oocyte development is affected by loss of insulin signaling, the protein expression pattern for a marker of oocyte development, FOXO3 was analyzed [32]. In oocytes of primordial and primary follicles, FOXO3 is abundantly expressed and localized to the nucleus. After activation and transition into the secondary follicle stage, FOXO3 is translocated into the cytoplasm and degraded. In WT and single receptor mutant mice, this pattern is preserved (Fig. 6B). However, DKO mice exhibited normal FOXO3 expression in primordial follicles, but FOXO3 was absent in putatively abnormal activated follicles with a single layer of granulosa cells (*).

Adult mice lacking granulsoa cell INSR and IGF1R exhibit irregular of estrous cyclicity

To determine whether abnormal follicle growth and progression resulted in abnormal communication with the HPG axis, we assess estrous cyclicity in mice for ~3 weeks beginning one week after vaginal opening. Wild-type mice exhibited the expected 2–4 day progression of estrous cyclicity (Fig. 7). In contrast, Insr^d/d females were inconsistent in their overall cyclicity and exhibited a normal progression only a few times during the observation window with significant pauses in estrus and metestrus. Both Igf1r^d/d and DKO mice were primarily trapped in metestrus with only one Igf1r^d/d female exhibiting a potentially successful ovulatory cycle which corresponded to overall fertility outcomes observed in these lines (Fig. 2).

Figure 7.

Estrous cyclicity is disrupted in mice lacking ovarian insulin receptor signaling. One week after vaginal opening, estrous cycles were monitored for 24 days (n = 6). Representative plots for three WT and three knockout females for each receptor are shown. Wild-type mice exhibit expected 2–4 day repeating cycles. However, all three mutants exhibit extended pauses in metestrus (M) or estrus (E) with DKO mice never successfully completing a cycle. Proestrus (P), diestrus (D).

Discussion

In this study, we conditionally deleted Insr and Igf1r using Esr2-Cre, which led to impaired ovulation and infertility in Igf1r^d/d and DKO in a 2-month breeding trial. We chose to ablate both receptors simultaneously to eliminate the potential cross-reactivity of insulin ligands (e.g. INS low affinity binding to IGF1R) that might mask fertility defects after single ablation of their cognate receptors individually. As expected, in contrast to complete insulin receptor knockout lines where body growth and development are impaired, our conditional DKO mice exhibited no such complications and exhibited similar growth curves as control animals. However, abnormal ovarian follicle progression resulted in a delay in puberty and improperly timed or stalled ovarian cycles. While all, but one, females in the breeding trial exhibited vaginal plugs, indicating they were receptive to copulation, the mean time for the consistent presence of vaginal plugs in pairings of Igf1r^d/d or DKO female mice and WT males was greater and highly varied in DKO mice.

Prior to this report, studies examining the essential synergy of ovarian insulin signaling and fertility have concentrated on IGF modulation of FSH receptor signaling [15, 33]. To examine the role of IGF signaling in the ovary without the deleterious effects of global IGF or IGF1R knockout [34], Baumgarten et al performed conditional ablation of Igf1r [15]. To achieve complete ablation of IGF1R in granulosa cells, they employed a combination of Esr2-Cre and Cyp19-Cre expressing lines and the same Igf1r^fl/fl strain and Esr2-Cre used in this study. Mice lacking granulosa IGF1R had smaller ovaries and were sterile. Unexpectedly, FSH receptor levels in these mice were not significantly different than WT control females. However, they did observe a complete arrest of follicle growth at the preantral stage in adult mice. Granulosa cells that were present failed to differentiate, exhibited premature apoptosis, and ultimately gave rise to follicles with markedly reduced estrogen production which could not be alleviated by exogenous gonadotropin stimulation.

In our prior report, we used Pgr-Cre, which is highly active relatively later than both Esr2-Cre or Cyp19-Cre, to ablate Insr and Igf1r and demonstrated that folliculogenesis could continue past the antral follicle block observed in the Stocco report and confirmed in this study. However, Pgr-Cre mediated single knockout mice were subfertile and DKO mice infertile, suggesting distinct roles for insulin receptor signaling in early follicle, periovulatory window, and follicle function after luteinization [14]. DKO mice exhibited a 50% decline in ovulation with oocytes often found trapped in partially luteinized follicles, but the mice were completely infertile due to Pgr-Cre driven ablation of insulin receptors in the uterus leading to complete failure in implantation [35]. As with our prior report, the majority of the fertility defects in our conditional knockout mice come from loss of IGF1R signaling. While two Insr^d/d dams failed to produce a litter during the breeding trial, those that did generated 8.13 ± 0.75 live pups per litter which matched WT (8.13 ± 0.99) litter sizes. This agrees with findings from INSR^gcko mice that exhibited greater variability in average litter size and oocytes ovulated, but no significant differences in diminished mean values compared to WT mice in a 6-month breeding trial [15] and our Pgr-Cre driven Insr^d/d ablation model that produced ~1.5 pup/litter fewer per litter most likely due to uterine, not ovarian, abnormality [14]. Ligands from the IGF family can bind INSR [8, 36], but are apparently not required to do so when IGF1R is intact (our work [13, 14], and others’ [15, 37]). Supplemental INS protein is commonly added to granulosa cell culture media and has been demonstrated to elicit maximal hormone production in cultured ovarian cells [38–40]. These studies and others examining similar mechanisms have often relied on the non-physiological effects of insulin-related hormone, so the ratio of relevant INSR vs IGF1R action is difficult to assess in light of potential cross-reactivity. We recently showed that Ins2, the gene which encodes is the primary insulin hormone in mice [41], is highly transcribed locally in the granulosa cell layer of growing follicles [42]. The regulation Ins2 expression in the ovary is at least partially controlled by two luteinizing hormone and progesterone-regulated homeobox factors, RHOX5 and RHOX8, but ovarian hormones might directly stimulate Ins2 as well. We had previously reported the regulation of Ins2 expression in Sertoli cells, the male counterpart of granulosa nurse cells in the testis [43]. However, it is clear that INSR-mediated growth and development cannot compensate for the loss of IGF1R. At vaginal opening, the two receptors appear to cooperate to drive maximal growth of follicles to the secondary and antral stages with Insr^d/d lying intermediary between WT mice and Igf1r^d/d mice, with DKO mice essentially lacking antral follicles. While IGF1R appears to be more involved in the development of large follicles [15, 44, 45], it is unclear which receptor plays a more important role in early follicular development.

One unique goal of this proposal was to examine the impact of differential insulin receptor signaling in the neonatal window. Prior studies had demonstrated sex reversal when insulin signaling is lost in embryonic Sertoli cells, but that did not provide insights into the role of granulosa-specific actions of insulin receptors in germ cell nest formation and breakdown because the hormonal axes that controlled that aspect of development would have been normal in female mice [12]. We observed no abnormal male development in DKO mice following ablation with Esr2-Cre, and female differentiation of the ovary appeared normal. While it is possible that germ cell selection may be altered in ovaries of DKO mice at birth, the establishment of the primordial and primary follicle pools was normal. However, by PND4 the defect in follicle growth is evident in DKO mice with significantly fewer secondary follicles being observed and by PND8, the follicle distributions observed at puberty are established.

Starting from the transition from primordial follicles to preantral follicles, the main driver of proliferation and development of granulosa cells is FOXL2 [28, 46]. Interestingly, we saw static levels of Insr during the PND1–8 window, but a gradual increase in insulin ligands and Igf1r expression which correlated with the induction of Foxl2 in WT animals. At PND8, Foxl2 mRNA levels were reduced in all three knockout models with Insr deletion being more mild than loss of Igf1r or both receptors. This could explain why there is a similarity in ovarian morphology between Foxl2 mutant mice [28, 46] and our current model. In Foxl2^−/− mice, ovarian follicles are arrested at the primordial/primary stages, and their granulosa cells do not proliferate. The loss of Foxl2 results in primary premature ovarian failure (POF) by the age of 15 weeks [28, 46]. Our DKO model exhibit a similar ovarian defect with large, activated oocytes and no subsequent proliferation of granulosa cells. The majority of other models exhibiting similar defects stem from the loss of oocyte-specific gene expression. Thus, taken together, this suggests that INSR and IGF1R potentially act upstream of Foxl2 and regulate its expression, at least during early follicular development.

In addition to expression, another way insulin signaling could regulate FOXL2 is by phosphorylating residues S33, a principal phosphorylation site, as well as S211, Y186, Y215, S238, S263, S323 and/or S326 [47–49]. All residues have a similar PxSP sequence – a target of MAPK and related kinases [47, 50], one of the two canonical insulin signaling pathways. Of these, pS33 and pY186 seem to be positively correlated with granulosa cell proliferation, especially in granulosa cell tumors (GCTs) [48, 49]; ~60% of cells in GCTs were positive for pS33, compared to less than 20% in other tumors. FOXL2 is also involved in sex determination, unfortunately in our model, the knockout of Insr and Igf1r occurs in neonates [18], thus a different Cre would be necessary to unravel their potential role in female sex development.

One of the main regulators of the primordial pool is AMH. AMH suppresses the activation of primordial follicles, its marked reduction in mRNA expression in all knockout models supports the hypothesis that there is an enhanced activation of oocytes prematurely. Thus, its loss when insulin receptors are ablated may relieve the repression signal allowing the activation of primary follicles before the desired hormonal conditions for successful ovulation are met. In support of this, multiple large abnormal oocytes with few granulosa cells are observed in ovarian sections of DKO animals at puberty. Whether these follicles could recover if subsequent granulosa proliferation was restored is unknown. Unfortunately, we only examined AMH expression in younger adult dams included herein. Thus, we do not know if older adult DKO mice would exhibit POF due to accelerated unregulated activation of their normal follicle reserves as we did not quantify this since the mice were completely infertile by 2 months of age. However, perturbation of insulin signaling could potentially be employed to generate novel mouse models of estropause and ovarian aging, which we plan to pursue in future studies.

In summary, employing an Esr2-Cre conditional knockout model, we successfully ablated ovary-specific insulin signaling beginning in the neonatal window. We hoped to find significant alteration in the establishment of the ovarian reserve. However, if insulin signaling is critical during this process, the activity of Esr2-Cre was insufficient to significantly alter the expected developmental program. We demonstrated that insulin actions are necessary for successful the differentiation of ovulation competent oocytes, granulosa cell support of ovulation, and ultimately fertility. Our findings, in agreement with prior studies from our lab and others, indicate that IGF1R plays a more prominent role in governing overall female fertility than INSR. However, as the DKO has a unique phenotype from that resulting from single ablation of either receptor in granulosa cells, there must be novel co-regulated factors and pathways that have yet to be elucidated. Insulin receptor signaling is highly conserved in mammalian species, thus the elucidation of these pathways may ultimately provide insights into ovulatory dysfunction in women with metabolic syndromes where insulin receptor action is impaired.

Supplementary Material

Supplementary material

Supplementary data are available at BIOLRE online.

Data availability

The data underlying this article are available and will be shared on reasonable request to the corresponding author.