Selective expression of KrasG12D in granulosa cells of the mouse ovary causes defects in follicle development and ovulation

Abstract

Activation of the RAS family of small G-proteins is essential for follicle stimulating hormone-induced signaling events and the regulation of target genes in cultured granulosa cells. To analyze the functions of RAS protein in granulosa cells during ovarian follicular development in vivo, we generated conditional knock-in mouse models in which the granulosa cells express a constitutively active KrasG12D. The KrasG12D mutant mice were subfertile and exhibited signs of premature ovarian failure. The mutant ovaries contained numerous abnormal follicle-like structures that were devoid of mitotic and apoptotic cells and cells expressing granulosa cell-specific marker genes. Follicles that proceeded to the antral stage failed to ovulate and expressed reduced levels of ovulation-related genes. The human chorionic gonadotropin-stimulated phosphorylation of ERK1/2 was markedly reduced in mutant cells. Reduced ERK1/2 phosphorylation was due, in part, to increased expression of MKP3, an ERK1/2-specific phosphatase. By contrast, elevated levels of phospho-AKT were evident in granulosa cells of immature KrasG12D mice, even in the absence of hormone treatments, and were associated with the progressive decline of FOXO1 in the abnormal follicle-like structures. Thus, inappropriate activation of KRAS in granulosa cells blocks the granulosa cell differentiation pathway, leading to the persistence of abnormal non-mitotic, non-apoptotic cells rather than tumorigenic cells. Moreover, those follicles that reach the antral stage exhibit impaired responses to hormones, leading to ovulation failure. Transient but not sustained activation of RAS in granulosa cells is therefore crucial for directing normal follicle development and initiating the ovulation process.

Fig. 1. Conditional knock-in of Kras^G12D in granulosa cells

(A–F) In vivo recombination of the R26R locus in ovaries by the Cyp19-Cre transgene. (A) ROSA26, (B) ROSA26;Cyp19-Cre and (C) ROSA26;Cyp19-Cre, 48 hours after eCG treatment. Images are of ovaries from 23-day-old mice showing β-gal staining (blue). (D) Day10 and (E) Day23 without eCG, and (F) Day23 with eCG treatment. Hematoxylin and Eosin staining of paraffin sections after β-gal staining showing the expression of β-gal in the ovaries of the ROSA26;Cyp19-Cre mouse. F, follicle; GC, granulosa cell. (G) RT-PCR detection of Kras^G12D and total Kras mRNAs in LSL-Kras^G12D;Amhr2-Cre and LSL-L-ras^G12D;Cyp19-Cre mouse ovaries. (H) Immunofluorescence of KRAS in the ovary of a 6-week-old cycling wild-type mouse.

Fig. 2. Expression of KRAS^G12D in granulosa cells causes multiple reproductive defects

(A) Continuous breeding assay showing the cumulative number of progeny per female. The LSL-Kras^G12D;Amhr2-Cre and LSL-Kras^G12D;Cyp19-Cre females (n=6) were subfertile. (B) Superovulation experiments showing that the ovulation rate in response to gonadotropins was reduced in Kras mutant mice (n=10) as compared with the wild type (WT). (C–J) Histology of WT (C,E,G,I) and LSL-Kras^G12D;Amhr2-Cre (D,F,H,J) ovaries at 8 (C–F), 16 (G,H) and 48 (I,J) hours after hCG treatment. Histology of WT (I) and LSL-Kras^G12D;Amhr2-Cre (J) ovaries 48 hours after hCG treatment shows that an oocyte is trapped in the corpus luteum of the Kras mutant ovary (arrow).

Fig. 3. Kras^G12D conditional knock-in mice develop ovarian lesions with altered granulosa cell proliferation, differentiation and apoptosis

(A) Size differences in wild-type (WT) and LSL-Kras^G12D;Cyp19-Cre ovaries at various ages. (B–D) Histology of WT (B) and LSL-Kras^G12D;Cyp19-Cre (C,D) ovaries at 6 months of age. NF, normal follicle; AF, abnormal follicle. (E,F) BrdU incorporation assay in 12-week-old WT (E) and LSL-Kras^G12D;Cyp19-Cre (F) ovaries. Abnormal follicle-like structures are indicated by arrows (as below). (G,H) Immunofluorescent detection of phospho-histone H3 (pHH3, red) in 12-week-old WT (G) and LSL-Kras^G12D;Cyp19-Cre (H) ovaries. (I,J) BrdU incorporation (I) and immunofluorescence for the mitosis marker phospho-histone H3 (J) indicate slightly increased levels of proliferation in Kras^G12D-expressing granulosa cells of antral follicles, as compared with wild type. (K,L) Immunohistochemical detection of PCNA in 12-week-old WT (K) and LSL-Kras^G12D;Cyp19-Cre (L) ovaries. (M,N) Apoptosis assays in 4-week-old WT (M) and LSL-Kras^G12D;Cyp19-Cre (N) ovaries, 2 hours after hCG treatment. (O,P) Immunofluorescent detection of cleaved caspase 3 (CC3) in 12-week-old WT (O) and LSL-Kras^G12D;Cyp19-Cre (P) ovaries.

Fig. 4. Kras^G12D downregulates genes essential for granulosa cell differentiation and ovulation

(A) qRT-PCR of ovulation-related genes from mouse whole ovary mRNAs. Six ovaries from different animals were analyzed. (B) Kras^G12D downregulated the expression of Fshr and prevented the FSH-induced expression of Lhcgr in cultured granulosa cells. Expression of Kras^G12D was induced by infecting the cells with an adenoviral vector encoding Cre recombinase (Ad-Cre). FSH (100 ng/ml) was added to the medium of cells infected, or not, with Ad-Cre. NT, non-treated. Ppib was amplified by RT-PCR in the same samples, as loading control.

Fig. 5. KRAS^G12D activates the RAF1/MEK/ERK1/2 and PI3K/AKT pathways in granulosa cells

(A) RAS activity in wild-type ovaries during ovulation, as measured by a GST pull-down assay using RAF1 RAS-binding domain (RBD) as the bait. NT, non-treated. (B) RAS-GTP levels increased in immature LSL-Kras^G12D;Amhr2-Cre and LSL-KrasG12D;Cyp19-Cre ovaries, as measured by the GST pull-down assay using both RAF1 RBD and p110α RBD. (C) Phosphorylation of MEK1/2, ERK1/2 and AKT in wild-type and LSL-Kras^G12D;Cyp19-Cre ovaries after eCG/hCG treatment. Total ERK1/2 and AKT are shown as loading controls. (D–F) Localization of phospho-ERK1/2 in ovaries. The level of phospho-ERK1/2 was low in immature wild-type mouse ovaries (D), but was increased in the large antral follicles 2 hours after hCG injection (E). By contrast, the phospho-ERK1/2 level remained low in LSL-Kras^G12D;Cyp19-Cre ovaries after the same treatment (F). (G–I) Immunofluorescence of phospho-AKT in wild-type ovaries. (G) Immature ovary; (H) 48 hours after eCG; (I) 2 hours after hCG. (J–L) Immunofluorescence of phospho-AKT in LSL-Kras^G12D;Cyp19-Cre ovaries. (J) Immature ovary before eCG treatment (abnormal follicle-like structures indicated by arrows); (K) 48 hours after eCG treatment; (L) 2 hours after hCG treatment. Three to six ovaries from different animals of each genotype were analyzed in each of these experiments.

Fig. 6. Acute effect of Kras^G12D expression in cultured granulosa cells

(A) Expression of Kras^G12D mRNA in LSL-Kras^G12D granulosa cells after infection with Ad-CMV-Cre. (B) Levels of phospho-ERK1/2 and phospho-AKT post-infection. (C) LSL-Kras^G12D granulosa cells infected with Ad-CMV-Cre and control vectors (Ad-CMV-GFP) for 48 hours were stimulated with FSH, forskolin (Fo) or amphiregulin (AR) for 20 minutes. Each agonist induced ERK1/2 phosphorylation in control granulosa cells, but the responses were reduced in the cells expressing KRAS^G12D.

Fig. 7. Mkp3 is a KRAS^G12D-induced gene involved in normal and abnormal granulosa cell development

(A) Semi-quantitive RT-PCR shows that Mkp3 mRNA levels were elevated in LSL-Kras^G12D;Amhr2-Cre ovaries as compared with those of wild-type mice (n=3 for each genotype). (B) Mkp3 mRNA expression was induced in wild-type granulosa cells by eCG/hCG treatment. (C–H) In situ hybridization for Mkp3 in wild-type and Kras mutant ovaries. Bright-field images show ovarian histology (C,E,G), whereas dark-field images show the signals of Mkp3 antisense probe (D,F,H). Mkp3 mRNA was detected in granulosa/cumulus cells at 4 hours after hCG (D), but not in immature wild-type ovaries (F). By contrast, Mkp3 mRNA was detected in some preantral follicles (H, arrows) in Kras mutant ovaries of the same age. (I,J) Mkp3 siRNA decreased the Mkp3 mRNA levels in unstimulated granulosa cells (I) or those stimulated with AREG (J). (K,L) AREG induced transient phosphorylation of ERK1/2 in control granulosa cells; however, in the granulosa cells treated with Mkp3 siRNA, levels of phospho-ERK1/2 remained elevated for longer (K). (L) Intensity comparison of phospho-ERK2/total ERK2. (M,N) PD98059 blocked AREG-induced Mkp3 expression (M), when the ERK1/2 activation is blocked (N).

Fig. 8. Schematic of the ovarian defects caused by the expression of KRAS^G12D in developing follicles

In ovaries of wild-type mice, the LH/hCG surge transiently activates RAS and its downstream effecters, the ERK1/2 pathway and the PI3K pathway, which impact granulosa cell differentiation and ovulation by regulating the expression of numerous genes. ERK1/2 induces the expression of MKP3, which negatively regulates ERK1/2 activity in ovulating follicles as well as in cells expressing mutant Kras^G12D. PI3K regulates the phosphorylation of AKT and FOXO1. When KRAS^G12D is expressed in the granulosa cells, it interacts with both RAF1 and PI3K, activates ERK1/2 and AKT, respectively, and leads to two major ovarian phenotypes. In small follicles, the granulosa cells fail to differentiate and are devoid of their marker genes such as the FSH receptor. Moreover, these granulosa cells are non-mitotic, non-apoptotic and reside in abnormal follicle-like structures that accumulate in the ovaries of the mutant mice. Those follicles that escape this senescent fate develop to the antral stage but fail to ovulate because of the impaired expression of genes associated with ovulation. In addition, the mutant antral follicles exhibit reduced levels of phospho-ERK1/2 related to abnormally elevated levels of Mkp3. Red lines, RAS-related events in normal ovaries; green lines, KRAS^G12D-related events in mutant ovaries.