Impaired fertility and FSH synthesis in gonadotrope-specific Foxl2 knockout mice

Abstract

Impairments in pituitary FSH synthesis or action cause infertility. However, causes of FSH dysregulation are poorly described, in part because of our incomplete understanding of mechanisms controlling FSH synthesis. Previously, we discovered a critical role for forkhead protein L2 (FOXL2) in activin-stimulated FSH β-subunit (Fshb) transcription in immortalized cells in vitro. Here, we tested the hypothesis that FOXL2 is required for FSH synthesis in vivo. Using a Cre/lox approach, we selectively ablated Foxl2 in murine anterior pituitary gonadotrope cells. Conditional knockout (cKO) mice developed overtly normally but were subfertile in adulthood. Testis size and spermatogenesis were significantly impaired in cKO males. cKO females exhibited reduced ovarian weight and ovulated fewer oocytes in natural estrous cycles compared with controls. In contrast, ovaries of juvenile cKO females showed normal responses to exogenous gonadotropin stimulation. Both male and female cKO mice were FSH deficient, secondary to diminished pituitary Fshb mRNA production. Basal and activin-stimulated Fshb expression was similarly impaired in Foxl2 depleted primary pituitary cultures. Collectively, these data definitively establish FOXL2 as the first identified gonadotrope-restricted transcription factor required for selective FSH synthesis in vivo.

Figure 1.

In vivo recombination of the Foxl2 locus is restricted gonadotropes and testes of cKO mice. (A) PCR detection of floxed and recombined Foxl2 alleles in gDNA isolated from pituitary, testis, ovary, and tail of control and cKO animals. (B) RT-qPCR detection of Foxl2 mRNA levels in whole pituitaries of control and cKO mice (n = 10/group). All data here and in subsequent figures are means (±SEM), and an asterisk indicates statistical significance with P < .05 by t test (unless stated otherwise). (C) Pituitary FOXL2 protein expression in control and cKO mice as assessed by immunoblot (IB). Pooled protein lysates (100 μg) from 3 pituitaries per genotype per lane. Duplicates are shown for each genotype. LβT2 cell lysates were included as a positive control. ACTB, β-Actin. (D) Immunofluorescence for FOXL2 (green) and TSHβ (red) in pituitary sections from adult control (left) and cKO (right) mice. Arrows indicate examples of double-labeled cells. Scale bar, 100 μm. (E) PCR detection of floxed, recombined, and wild-type (WT) Foxl2 alleles in gDNA isolated from FACS-purified pituitary cells of control (Foxl2^+/+;Gnrhr^GRIC/+;Rosa26^EYFP/+) and cKO (Foxl2 ^flox/flox;Gnrhr ^GRIC/+;Rosa26^EYFP/+) mice. YFP−, Nongonadotropes; YFP+, gonadotropes. (F) Foxl2 mRNA levels from the sorted cells in E. Data are from 5 independent experiments (n = 3 using males and n = 2 using females), and pooled results were analyzed by 2-way ANOVA (Tukey post hoc).

Figure 2.

Foxl2 cKO mice are subfertile. (A) Cumulative numbers of litters over 6 months in control and cKO males and females (n = 7/genotype). At 6 months, an asterisk indicates a statistically significant difference (P < .05) between control and cKO groups within the same sex. (B) Average litter size per couple. (C) Litter frequency. In C, data from 1 cKO male were omitted, because he did not sire any pups.

Figure 3.

Testicular weights, SC numbers, and sperm count are reduced in cKO males. (A) Paired testicular weights (in grams) normalized to body weight (in grams) for adult control (n = 14) and cKO (n = 34) males. (B) HE-stained sections of testes from representative control and cKO males showing variable disruption of spermatogenesis. 1, Control testis with normal spermatogenesis with open lumens and all stages of spermatogenesis; 2, cKO testis with abnormal retention of spermatids as denoted by the asterisk; 3, cKO testis with pyknotic cells (arrow) and sloughed germ cells filling the lumen; 4, cKO testis with spermatogenesis arrested at meiosis; few to no haploid round spermatids observed. (C) SC counts from left testis of control and cKO males (n = 6/group). (D) Posthomogenization sperm counts from right testis and cauda epididymis of control (n = 11) and cKO (n = 10).

Figure 4.

Ovarian weights and ovulation rates are decreased in cKO females. (A) HE-stained ovarian sections from representative 8-week-old control and cKO females. Red arrows indicate CL. (B) Paired ovarian weight (in grams) normalized to body weight (in grams) in metestrous/diestrous control (n = 9) and cKO (n = 12) females aged 8–13 weeks. (C) COC and CL counts after mating in control (n = 10 for COC, n = 5 for CL) and cKO (n = 6 for COC, n = 4 for CL) females. (D) COC counts after exogenous gonadotropin stimulation in 24- to 28-day-old control and cKO females (n = 8/genotype). ns, nonsignificant.

Figure 5.

Pituitary FSH synthesis is impaired in cKO mice. (A) Immunofluorescence for FSH (green) and LH (red) on pituitary sections from adult control and cKO mice (×400 magnification; scale bar, 50 μm). Mean (±SEM) mRNA level as assessed by RT-qPCR for the indicated genes from pituitaries of control and cKO (B) males and (C) females; n = 7–10/genotype. ns, nonsignificant.

Figure 6.

Decreased Fshb transcription and FSH synthesis in primary pituitary cultures after ex vivo Foxl2 recombination. (A) Recombination of floxed Foxl2 in male primary pituitary cells. Ad-GFP, Adenovirus expressing GFP; Ad-Cre, adenovirus expressing Cre recombinase; floxed and recombined Foxl2 alleles are indicated. Mean (±SEM) relative (B) Foxl2 mRNA, (C) Fshb mRNA, and (D) secreted FSH levels in treated cultures from 5 independent experiments. Data were analyzed by 2-way ANOVA (Tukey post hoc). Bars with different symbols were statistically different; those sharing symbols did not differ.