Targeted Disruption of Mapk14 (p38MAPKα) in Granulosa Cells and Cumulus Cells Causes Cell-Specific Changes in Gene Expression Profiles that Rescue COC Expansion and Maintain Fertility

Abstract

MAPK14 (p38MAPKα) is critical for FSH and prostaglandin E (PGE)2 signaling cascades in granulosa cells (GCs) and cumulus cell-oocyte complexes (COCs) in culture, indicating that this kinase might impact follicular development and COC expansion in vivo. Because Mapk14 knockout mice are embryonic lethal, we generated GC specific Mapk14 knockout mice (Mapk14gc^−/−) by mating Mapk14^fl/fl and Cyp19-Cre mice. Unexpectedly, the Mapk14gc^−/− female mice were fertile. Analyses of gene expression patterns showed that amphiregulin (Areg) and epiregulin (Ereg), two key regulators of ovulation and COC expansion, were up-regulated in the GCs but down-regulated in cumulus cells of the mutant mice in vivo. COCs from the mutant mice expanded and expressed matrix-related genes, if cultured with AREG, but not when cultured with forskolin or PGE2, the latter being a key factor regulating MAPK14 activity in cumulus cells. Conversely, when GCs from the Mapk14gc^−/− mice were cultured with forskolin, they produced more Areg and Ereg mRNA than did wild-type GCs. These results indicate that disruption of Mapk14 selectively alters the expression of Areg and other genes in each cell type. Greater AREG and EREG produced by the GCs appears to by-pass and compensate for the critical need for MAPK14 signaling and induction of Areg/Ereg (and hence matrix genes) by PGE2 in cumulus cells of the mutant mice. In conclusion, although MAPK14 is not overtly essential for preovulatory follicle development or events associated with ovulation and luteinization in vivo, it does impact gene expression profiles.

Disruption of Mapk14 in granulosa and cumulus cells causes cell specific changes in the expression of Areg/Ereg and other genes that regulate COC expansion.

The pituitary gonadotropic hormones, FSH and LH, are essential for controlling ovarian follicle development, ovulation, and luteinization (1). One primary signaling cascade downstream of FSH and LH is the activation of adenylyl cyclase and protein kinase A (PKA) (2). However, recent studies indicate that multiple additional signaling cascades are activated by the gonadotropins, including the MAPKs (3,4,5). There are three major subgroups of MAPKs, the ERKs (ERK/MAPKs), the stress-activated protein kinases (SAPKs)/Jun kinases (SAPK-1/Jun kinases), and the p38MAPKs (MAPKs11-14) (6). Several groups have documented the important roles of MAPK3/1 (ERK1/2) in follicle development, especially in regulating events associated with ovulation and luteinization (5,7,8,9). Using gene targeting approaches, we have conditionally disrupted the Erk2 gene in the Erk1^−/− background using our Cyp19-Cre mice (10) that express Cre recombinase driven by the human aromatase PII promoter, resulting in the selective expression in granulosa cells (GCs). Analyses of ovarian function in the resulting Erk1^−/−;Erk2^fl/fl;Cyp19-Cre (Erk1/2^gc−/−) mouse model document that these two kinases are essential for the global physiological consequences by which LH induces ovulation and luteinization (11).

Based on studies in cultured cells, the p38MAPKs also appear critical for mediating FSH and prostaglandin signaling cascades in GCs and cumulus cell-oocyte complexes (COCs). Using SB203580 (SB20), a highly selective p38α and p38β MAPK-specific inhibitor, we and others have shown that it blocks FSH-induced ERK1/2 phosphorylation (5) and small heat shock protein 27 phosphorylation in GCs of immature rats (12). Many other studies have shown that SB20 blocks COC expansion and the expression of genes required for this process that are induced by FSH, prostaglandin E (PGE)2, or IL-6 in both cultured mouse COCs (7,13,14,15) and porcine COCs (16,17). The genes that are induced in COCs (see animals section in Materials and Methods and reference therein) and that are required for expansion include: the epidermal growth factor (EGF)-like factors (Areg, Btc, and Ereg) (9), matrix forming and stabilizing factors [hyaluronan synthase 2 (Has2), prostanglandin synthase 2 (Ptgs2), tumor necrosis factor α-induced protein 6 (Tnfaip6), and pentraxin 3 (Ptx3)[ (18,19,20,21) and inflammatory cytokines (Il6) (15). Importantly, amphiregulin (AREG) and PGE2 appear to comprise a regulatory feedback loop that is essential for COC expansion and depends not only on the activation of ERK1/2 and induction of Ptgs2 but also p38MAPK and its regulation of specific functions. These results have led to the hypothesis that the p38MAPKs play potentially important physiological roles during follicular development and COC expansion in vivo.

The p38MAPK super family is comprised of four members, p38MAPKα (MAPK14), p38MAPKβ (MAPK11), p38MAPKγ (MAPK12), and p38MAPKδ (MAPK13). In most mammalian tissues, MAPK14 is the most abundant isoform (6). Mice in which the Mapk11 (22,23), Mapk12 (24), and Mapk13 (24,25) genes have been disrupted are viable and exhibit normal fertility. However, different strains of MAPK14 null mice (26,27,28,29,30) are all embryonic lethal, which has prevented analyses of the in vivo roles of MAPK14 in the adult ovary. To overcome this limitation, specific p38MAPK inhibitors have been used as indicated above. However, inhibitor studies in culture do not always mimic the events that occur in vivo. Therefore, Mapk14^fl/fl mice, in which the first exon of Mapk14 gene encoding the start codon was flanked by the loxP sites, have been generated and used by others to target the disruption of the Mapk14 gene in selected cell types using different mouse strains expressing cell specific Cre recombinase (28,31). In general, these studies have determined that in some cells, MAPK14 exerts negative regulatory effects on the c-Jun N-terminal kinase (JNK) pathway to block cell proliferation and cell fate decisions (28,31,32). In other cell contexts, MAPK14 is critical for mediating cytokine-induced inflammatory reactions and senescence (23). Loss of MAP3K4 (MEKK4), a kinase upstream from MAPK14, alters mouse sex determination, suggesting that MAPK14 may impact this process as well (33). To determine the role of MAPK14 in the ovary, we generated GC specific Mapk14 knockout mice (Mapk14gc^−/−) by mating Mapk14^fl/fl mice (28) to mice expressing Cre recombinase driven by Cyp19a1 (GC specific) promoter (10). Using this mouse model, we demonstrate for the first time that disruption of the Mapk14 gene alone selectively alters gene expression profiles in cumulus cells vs. GCs of preovulatory follicles, providing evidence that GCs and cumulus cells each represent a special niche in the preovulatory follicle. For example, expression of the EGF-like factors Areg and Ereg in GCs is increased specifically, whereas their expression in cumulus cells is dramatically impaired and associated with defective COC expansion in culture. Therefore, the enhanced activity in GCs may provide one compensatory mechanism that maintains COC expansion in vivo and, thus, fertility in the Mapk14gc^−/− mice.

Results

Localization and activation of MAPK14 in ovarian cells

Although MAPK14 inhibitors have been used to analyze the function of this kinase in ovarian cells, the cell-specific localization of total and activated (phosphorylated) MAPK14 in the mouse ovary in vivo has not been determined previously. Therefore, immunohistochemical and Western blot analyses were done to determine which cells express this kinase and its phosphorylated, activated form in vivo. As shown in Fig. 1, immunopositive MAPK14 and phospho-MAPK14 were detected in GCs, cumulus cells, and oocytes within follicles from the preantral to preovulatory stages of development. MAPK14 was not detected in the murine corpora lutea (Fig. 1), in marked contrast to previous studies in the rat (34), indicating a species difference in expression patterns of the kinase in luteal cells. Western blot analyses confirmed that levels of phospho-MAPK14 were increased by equine chorionic gonadotropin (eCG) during follicular development but were down regulated rapidly by human CG (hCG) treatment during luteinization. Levels of total MAPK14 were relatively constant in the whole ovarian lysates due to the presence of MAPK14 in GCs of growing follicles.

Figure 1.

Expression and activation of MAPK14 in the mouse ovary. A, Immunohistochemistry shows that total MAPK14 and phospho-MAPK14 (p-MAPK14) are localized selectively to oocytes, GCs, and cumulus cells present in small and preovulatory follicles [d27 sections (Aa)] but are not expressed in the corpus luteam (CL) [hCG 24 h (Ab); immature (Bc); hCG 48 h (Bd)]. Scale bars, 100 μm. B, Western blot analyses of lysates prepared from ovaries of immature (IM) mice before and after treatment with eCG and hCG confirm that the levels of phospho-MAPK14 decrease after hCG treatment, whereas total MAPK14 remains relatively constant, representing GCs present in the whole-ovarian lysates. AKT was used as an internal standard.

MAPK14 inhibitors block COC expansion

MAPK14 has been implicated in regulating ovarian cell activities, including cumulus cell-oocyte (COC) expansion and steroidogenesis based on numerous studies using specific MAPK14 inhibitors to block hormone-induced changes in GC and cumulus cell functions (5,7,13,14,15). LH-mediated COC expansion is mediated by the EGF-like factors [AREG, betacellulin, and epiregulin (EREG)] that activate ERK1/2, as well as by cumulus cell production of PGE2 that induces Areg in a feedback loop (35). In addition, FSH can substitute for LH and PGE2 to initiate COC expansion in culture (7,13,14). To analyze the ability of the MAPK14 inhibitor SB20 to block COC expansion and regulate the expression of selected genes induced in culture by FSH, AREG or PGE2 in the same experiment, COCs were isolated and cultured using previous protocols. Specifically, we examined the ability of the MAPK14 inhibitor SB20 to block COC expansion induced in culture by FSH, AREG or PGE2. As shown in Fig. 2A, SB20 potently inhibited COC expansion induced by each agonist in a dose-dependent manner (Fig. 2A), confirming previous results (7,13,14,15). Using another MAPK14 inhibitor, PD169316, we observed similar results (data not shown). To determine the molecular basis by which MAPK14 regulates COC expansion, real-time RT-PCR analyses were done for selected genes known to be important for this process (36,37). FSH, AREG, and PGE2 each induced marked increases in Areg, Ptgs2, Has2, Ptx3, Tnfaip6, and Il6 mRNA (Fig. 2B). The presence of SB20 significantly reduced the expression of all of these genes in response to each agonist with the exception of AREG induction of Ptgs2, which was not reduced by the MAPK14 inhibitor (Fig. 2B) but can be blocked by inhibition of ERK1/2 (11). Collectively, these data suggest that inhibition of MAPK14 impairs COC expansion by regulating PGE2-induced expression of specific matrix-related genes.

Figure 2.

The MAPK14 inhibitor SB20 blocks COC expansion and represses the induction of genes critical for this process. A, SB20 blocks FSH-, AREG-, and PGE2-induced COC expansion in a dose-dependent manner. COCs (15 per well) were isolated from immature mice primed with eCG for 48 h and cultured for 20 h in 50 μl medium. SB20 was added 1 h before agonists. A representative of two experiments is shown. The insertion of small pictures on the first row represents the unexpanded COCs in control vehicle-treated group. B, SB20 repressed the induction of genes important for COC expansion. COCs (100 per well) were isolated from immature mice primed with eCG for 48 h and cultured 4 h in 100 μl medium. Real-time RT-PCR data for each gene were first normalized to L19 and then calculated as fold change relative to the level induced by FSH for each gene. Data represent the means ± sem for three individual experiments. An asterisk indicates statistically significant differences between the group treated with and without SB20 (P < 0.05).

Generation of Mapk14gc^−/− mice

Although the in vitro involvement of MAPK14 in COC expansion by various agonists appears clear, the in vivo importance of MAPK14 in this process, as well as in follicular development, has not been determined, because germ line disruption of the Mapk14 gene causes embryonic lethality (26,27,28,29,30). To overcome this severe limitation, we generated GC-specific Mapk14gc^−/− by mating Mapk14^fl/fl mice (28) to mice expressing Cre recombinase driven by Cyp19a1 (GC specific) promoter (Fig. 3A) (10). The successful disruption of Mapk14 expression in GCs is demonstrated by the marked decrease of Mapk14 mRNA (based on real-time RT-PCR analyses) (Fig. 3B) and protein (Western blot analyses) (Fig. 3C) in samples prepared from GCs of immature control mice and Mapk14gc^−/− 24 and 48 h after treatment with eCG. This regimen stimulates preovulatory follicle development and greater expression of the Cyp19-Cre, hence increased recombination (10). Immunohistochemical analyses further document the selective reduction of MAPK14 in GCs and cumulus cells of growing and antral follicles but not in oocytes, theca cells, or the stromal compartment (Fig. 3D, a–d).

Figure 3.

Generation of Mapk14gc^−/− mice. A, The Mapk14gc^−/− mice were generated by mating Mapk14^fl/fl mice to mice expressing Cre recombinase driven by Cyp19a1 (GC specific) promoter. B, Real-time RT-PCR data show the relative levels of Mapk14 mRNA in GCs isolated from ovaries of control and Mapk14gc^−/− mice at 0, 24, and 48 h after eCG. For each sample, a pool of GCs from three mice was used, and the experiments were repeated twice. C, Total MAPK14 protein was partially reduced in GCs isolated from immature (d 23) Mapk14gc^−/− mice compared with control. Upon eCG treatment, total MAPK14 protein was dramatically reduced at both 24 and 48 h. AKT was used as a loading control. A representative Western blot analysis of three individual experiments is shown. D, Immunohistochemical images show the distribution of total MAPK14 in control (a and c) and Mapk14gc^−/− mice (b and d). In control mice, the MAPK14 signal is strong in GCs, cumulus cells, and oocytes [d 27 (a); hCG 8 h (c)]. In Mapk14gc^−/− mice, MAPK14 is only detected in oocytes [d 27 (b); hCG 8 h (d)]. Scale bars, 100 μm (a and b) and 200 μm (c and d). E and F, Mapk14gc^−/− mice are fertile and give birth to the same number of pups as wild-type mice (n = 4 for each). Mapk14gc^−/− mice ovulate a comparable number of oocytes as wild-type mice after superovulatory regimen of eCG and hCG 16 h (16 h after hCG; n = 6 for each).

Despite the depletion of MAPK14 in the GCs and cumulus cells, the Mapk14gc^−/− mice exhibit normal fertility. The Mapk14gc^−/− females gave birth to a similar number of pups as wild-type females over a 6-month breeding period (Fig. 3E). Under superovulation conditions using eCG and hCG, the Mapk14gc^−/− females released similar numbers of COCs (and oocytes) into the oviduct in vivo as wild-type females (Fig. 3F), and there were no differences on COC expansion in vivo (data not shown). These results were unexpected and indicated that there might be compensatory mechanisms operating in ovulating follicles of the mutant mice.

COCs from Mapk14gc^−/− mice have defective responses to FSH, forskolin, and PGE2 but not to AREG

Although the Mapk14gc^−/− mice did not exhibit overt ovarian defects, altered responses of COCs to specific hormones and growth factors may occur but not be readily detected in the in vivo context. Therefore, COCs from control and Mapk14gc^−/− mice were isolated and cultured with AREG, FSH, forskolin, or PGE2 to induce expansion. As expected, each agonist induced the expansion of the COCs isolated from control mice. However, only AREG induced expansion in the Mapk14gc^−/− COCs. FSH, forskolin, and PGE2 were totally ineffective in this culture context, indicating that MAPK14 is critical for mediating the effects of endogenous PGE2 that are essential in the feedback loop for inducing Areg and Ereg (Fig. 4A, C, and D) and specific matrix-related genes. Specifically, after 4 h of culture, when genes encoding the critical factors that comprise the hyaluronon-rich matrix are expressed at a maximal level, AREG induced similar levels of Has2 in control and mutant COCs, but the levels of Areg and Ptgs2 were reduced by 50% in Mapk14gc^−/− COCs (Fig. 4B). In contrast, the responses to forskolin and PGE2 in the Mapk14gc^−/− COCs were severely impaired with more than 80% loss of Areg and Ptgs2 expression compared with wild-type COCs (Fig. 4B). PGE2 was also less effective in inducing Has2 in Mapk14gc^−/− COCs compared with wild type. These data indicate that the forskolin and PGE2 signaling cascades are significantly altered in Mapk14gc^−/− COCs compared with wild-type COCs, because PGE2 cannot increase Areg that is required for activation of the ERK1/2 pathway and the induction of Ptgs2 (Fig. 4, C and D).

Figure 4.

COCs isolated from Mapk14gc^−/− mice exhibit impaired responses to FSH, forskolin, and PGE2 but not AREG. A, COCs were isolated from Mapk14gc^−/− mice primed with eCG for 48 h and cultured with AREG, FSH, forskolin, or PGE2. Although AREG induced COC expansion, the other agonists did not. B, COCs isolated from Mapk14gc^−/− mice at 48 h after eCG exhibit impaired gene expression profiles in response to treatment with forskolin or PGE2 for 4 h compared with WT COCs. Data shown are mean ± sem from two experiments. An asterisk indicates statistically significant differences between the two groups at the same time point (P < 0.05). C, Schematic representation of the AREG-PGE2 regulatory feedback loop in WT COCs. In cumulus cells, AREG is required to induce expression of Ptgs2, leading to increased PGE2 production. PGE2 in turn activates MAPK14, which enhances cumulus cell expression of Areg. D, As shown in the schematic representation, when MAPK14 is absent, PGE2 cannot induce Areg and Ereg. Therefore, ERK1/2 phosphorylation is reduced, and expression of matrix-related genes is impaired. However, exogenous AREG can by-pass the lack of MAPK14, leading to expansion and the induction of genes required for this process. E, AREG increases ERK1/2 phosphorylation in control and mutant COCs 90 min after treatment, whereas PGE2 increases pJNK in the mutant but not control COCs.

Because one potential explanation for the impaired PGE2 signaling might be reduced expression of PGE2 receptors, we analyzed the expression levels of Ptger2 and Ptger4. AREG induced similar levels of Ptger2 in control and Mapk14gc^−/− COCs, but the levels of Ptger4 were reduced 50% in the Mapk14gc^−/− COCs (Fig. 4B). The induction of Ptger2 and Ptger4 mRNAs in Mapk14gc^−/− COCs by PGE2 or forskolin were repressed at 4 h after treatments compared with wild-type COCs. These results further support a critical role of MAPK14 in mediating PGE2 actions via PTGER2/4 in cumulus cells (Figs. 4, C and D).

Furthermore, we document that phosphorylation of the stress kinase JNK is sustained in the Mapk14-depleted COCs compared with controls when cultured in the presence of PGE2 for 90 min (Fig. 4E), whereas no differences were observed at 20 min (data not shown). Both controls and the Mapk14-depleted COCs exhibited increased phosphorylation of ERK1/2 in response to AREG (Fig. 4E).

These dramatic changes in gene expression profiles and JNK phosphorylation in the Mapk14gc^−/− COCs indicated that there might be altered expression of these same or different genes in vivo. To analyze this, COCs were isolated from control and Mapk14gc^−/− mice at 4 h after hCG in vivo, a time when the major matrix-related factors are induced. As shown in Fig. 5A, the expression levels of Has2 were comparable between control and Mapk14gc^−/− COCs. However, all other genes, Areg, Ereg, Ptgs2, Tnfaip6, Ptx3, and Il6, were expressed at significantly lower (50%) levels in Mapk14gc^−/− COCs, indicating that the AREG-Ptgs2-PGE2 regulatory loop is compromised in the Mapk14gc^−/− COCs in vivo as well as in culture.

Figure 5.

Gene expression profiles are altered in COCs and GCs of the Mapk14gc^−/− mice in vivo. A, COCs isolated from Mapk14gc^−/− mice at 4 h after hCG exhibit reduced expression of Areg, Ereg, Ptgs2, Tnfaip6, and Ptx3 but normal level of Has2. B, GCs isolated from the same Mapk14gc^−/− mice at 4 h after hCG exhibit selectively increased expression of Areg and Ereg and reduced expression of Il6. C, Lysates prepared from GCs isolated from Mapk14gc^−/− mice at eCG 48 h exhibit increased levels of pCREB and pJNK. D, Schematic representation of ovarian responses of WT and Mapk14gc^−/− to FSH and LH. In wild-type GCs, FSH induces expression of genes associated with GC differentiation (Cyp19a1 and Lhcgr), whereas LH induces genes associated with ovulation (Areg) and luteinization (Cyp11a1 and Star). AREG then activates ERK1/2 in cumulus cells, leading to the induction of Ptgs2 and genes associated with expansion. PGE2, in turn, can induce Areg, providing a regulatory loop in preovulatory follicles that is disrupted in the Mapk14gc^−/− cells. However, because the mutant GCs produce increased amounts of AREG, this by-passes the need for PGE2.

Depletion of Mapk14 alters signaling pathways and gene expression profiles in GCs in vivo and in culture

Knowing that the COCs from Mapk14gc^−/− mice exhibited reduced expression of specific genes in vivo, we determined whether the GCs from Mapk14gc^−/− mice might also exhibit altered gene expression profiles relating to ovulation or luteinization. Therefore, GCs from Mapk14gc^−/− and control mice were isolated at 4 h after hCG, and the expression levels of Areg, Ereg, Ptgs2, Tnfaip6, Ptx3, Il6, and Has2 were compared. In marked contrast to the cumulus cells, we found that the levels of Areg and Ereg were significantly elevated in the Mapk14gc^−/− mutant cells compared with wild type (Fig. 5B). Because LH/hCG-mediated induction of Areg and Ereg is initiated, at least in part, by activation of the PKA pathway (9,35) and occurs in Erk1/2-depleted GCs that fail to initiate ovulation or luteinization (11), we hypothesized that PKA might exhibit higher activity in Mapk14gc^−/− GCs than in control cells. Therefore, we prepared lysates from GCs of control and Mapk14gc^−/− mice at 48 h after eCG in vivo and show by Western blot analyses that there is a 1.5- to 2-fold increase in phospho-cAMP response element-binding protein (pCREB) in the mutant cells (Fig. 5C). The enhanced activity of the PKA pathway is supported further by the observation that forskolin, a direct activator of adenylyl cyclase, induced significantly higher levels of Areg and Ereg in Mapk14gc^−/− GCs compared with controls (Fig. 6A). As in the Mapk14-depleted COCs (shown above), levels of phsopho-JNK were also elevated in the Mapk14-depleted GCs. Thus, in normal cells, MAPK14 appears to exert negative regulatory effects on both the PKA as well as the JNK signaling pathways (Fig. 5D).

Figure 6.

GCs from Mapk14gc^−/− mice exhibit altered responses to FSH and forskolin. A, When control and mutant GCs were cultured, forskolin induced higher expression of Areg and Ereg in Mapk14gc^−/− GCs 1.5 h after treatment. B, When GCs were isolated from control and Mapk14gc^−/− mice cells and cultured with either FSH or PGE2 for 24 h, several distinct gene expression profiles were observed. Basal levels of some genes (Fshr, Cyp19a1, and Pgr) were similar but exhibited enhanced responses to FSH or PGE2. Other GC marker genes (Inhba, Inhbb, Foxo1, and Bmp2) exhibited elevated basal levels of expression. Lastly, some genes associated with differentiation (Lhcgr), ovulation and inflammation (Il6), and luteinization (Cyp11a1 and Star) showed lower basal levels of expression. C, In control GCs, basal levels of Lhcgr were repressed by the MAPK14 inhibitor SB20, whereas basal levels of Inhbb were increased. All real-time RT-PCR data are presented as the mean ± sem from at least two experiments. An asterisk indicates significant differences between two groups (P < 0.05).

Additionally, when GCs from control and Mapk14gc^−/− mice were isolated and cultured in the presence of FSH or PGE2, the expression profiles of several genes known to control follicular development and luteinization were markedly altered and fell into three categories. Several genes (Fshr, Cyp19a1, and Pgr) exhibited similar basal levels in the control and Mapk14gc^−/− GCs, but their induction by FSH and PGE2 was enhanced in the mutant cells compared with controls. For other GC marker genes (Inhba, Inhbb, Foxo1, and Bmp2), the basal expression levels were elevated selectively in the Mapk14gc^−/− GCs compared with controls. In contrast, genes associated with luteinization (Lhcgr, Cyp11a1, and Star) and inflammation (Il6) were dramatically suppressed and not induced by hormones in Mapk14gc^−/− GCs (Fig. 6B), documenting that the Mapk14gc^−/− GCs exhibit more pronounced changes in gene expression profiles when placed in culture compared with in vivo (Fig. 6B). The suppressed levels of Lhcgr and Cyp11a1 mRNA in Mapk14gc^−/− GCs may be due to the elevated expression of activins (Inhba and Inhbb), because the inhibin α (Inha) knockout mice are characterized by elevated expression and unopposed action of activins and markedly reduced levels of Lhcgr and steroidogenic genes (38). To test this hypothesis, GCs from control mice were treated with the MAPK14 inhibitor SB20 for 24 h and total RNA was prepared. Real-time RT-PCR analyses show that inhibition of MAPK14 in this context was associated with increased expression of Inhbb and decreased expression of Lhcgr (Fig. 6C), in agreement with the observations in Mapk14gc^−/− GCs. However, the in vivo levels of Lhcgr mRNA in GCs were not affected in the Mapk14gc^−/− (data not shown), indicating that there are additional pathways that support Lhcgr expression in the in vivo context.

Discussion

The embryonic lethality of germ line disruption of Mapk14 provided an initial dramatic example of the critical importance of this protein kinase during the establishment of the trophoblast invasion and the initiation of angiogenesis in the placenta of mice (27). This lethality also precluded studies of MAPK14 in adult tissues forcing investigators to analyze the functions of MAPK14 in cells, including follicular cells, in culture with presumed MAPK14 specific inhibitors (5,7,12,13,16,17). Because these inhibitors had pronounced inhibitory effects on COC expansion and altered hormone-induced steroidogenesis in GCs, MAPK14 was presumed to play a major role in events associated with ovulation, oocyte maturation, and luteinization in vivo.

The availability of the Mapk14^fl/fl mice now provides a model in which cell specific targeted disruption of this gene can be engineered in vivo and thereby used to determine the physiological relevance of MAPK14 in adult tissues. The studies described here document for the first time that the disruption of Mapk14 in GCs and cumulus cells of growing follicles alters gene expression profiles in GCs and cumulus cells but does not impair fertility in the Mapk14gc^−/− mice in vivo, most likely because expression levels of Areg and Ereg are increased in a compensatory manner in GCs. Strikingly, COC expansion and ovulation were normal in the Mapk14gc^−/− mice. However, when COCs from the Mapk14gc^−/− mice were isolated and cultured in the presence of FSH, forskolin, or PGE2, COC expansion was blocked and the expression of genes associated with expansion was impaired severely. Of critical relevance was the lack of induction of the EGF-like factors Areg and Ereg in the Mapk14gc^−/− COCs. These EGF-like growth regulatory factors are known targets of LH/cAMP/PKA and act down-stream of LH via RAS (rat sarcoma virus oncogene)/MEK1 (mitogen-activated protein kinase kinase 1)/ERK1/2 to mediate the actions of LH in preovulatory follicles (9,11,35). Thus, when COCs from the Mapk14gc^−/− mice were cultured with AREG, COC expansion was normal and the need for MAPK14 was by-passed (Fig. 4, C and D).

These results strongly suggest that in vivo the EGF-like factors coming from GCs might override the defect in EGF-like factor production in cumulus cells. Indeed, when the expression of Areg and Ereg was analyzed in GCs isolated from control and Mapk14gc^−/− mice, expression of both EGF-like factors was elevated in the mutant cells compared with controls in vivo. Moreover, the expression of Areg and Ereg mRNA was enhanced in the Mapk14gc^−/− GCs compared with controls when the cells were cultured in the presence of forskolin, indicating that MAPK14 can exert negative regulatory effects on the PKA pathway. Furthermore, when COCs from the Mapk14gc^−/− mice were cocultuerd on a GC monolayer and treated with FSH or forskolin, COC expansion was normal (data not shown), providing additional evidence that factors from GCs can restore normal expansion in the Mapk14gc^−/− mutant COCs. Induction of Pgr, a gene critical for ovulation and a target of ERK1/2, was also greater in the mutant GCs cultured with FSH compared with controls, indicating a positive effect of AREG and EREG on these cells. These results provide an explanation for why COC expansion and ovulation occur normally in vivo. That is, elevated levels of AREG and EREG (and perhaps other factors) produced by the Mapk14gc^−/− GCs can by-pass the defects in the MAPK14 pathway regulation of Areg and Ereg expression in mutant COCs to maintain ERK1/2 phosphorylation and the induction of matrix genes at a level sufficient for expansion (Fig. 5D). These physiological compensatory events are supported by the evidence that elevated EGF can overcome the inhibitory effects of SB20 in cultured COCs in mice (7) and pig (17) and our data showing that high but not low concentrations of SB20 can block AREG induced COC expansion (Fig. 2A). Moreover, the results of these studies indicate unequivocally that the loss of Mapk14 in GCs has a dramatically different effect on the expression of Areg and Ereg than does its loss in cumulus cells. The molecular mechanisms that mediate these differences are not entirely clear but appear to be associated with the greater role of MAPK14 in cumulus cells and the increased activity of PKA in the Mapk14gc^−/− GCs. Therefore, MAPK14 may act as a negative regulatory factor in PKA activity in GCs but not cumulus cells.

The enhanced expression of other GC marker genes, such as activins (Inhba and Inhbb), Foxo1, and Bmp2, as well as Cyp19a1, in the Mapk14gc^−/− GCs indicates further that MAPK14 has the potential to exert negative regulatory effects on PKA during normal follicular development in vivo, depending on context. Conversely, the expression of many genes associated with GC differentiation (Lhcgr) and luteinization (Cyp11a1 and Star) was markedly suppressed in the Mapk14gc^−/− GCs, indicating that MAPK14 can promote differentiation, perhaps by reducing activins. This hypothesis is supported by the evidence that elevated levels of activins present in the inhibin α KO mice are associated with suppressed expression of Lhcgr and other genes associated with GC differentiation (38). Alternatively, the increased levels of JNK in the Mapk14gc^−/− GCs may contribute to the observed changes in gene expression profiles, although little is known at this time about the effects of JNK in the ovary. Collectively, these results indicate that in normal mice, MAPK14 has the potential to restrict the expression of genes in GCs of growing follicles and favors the expression of genes associated with luteinization.

The increased expression of Cyp19a1 and reduced expression of Cyp11a1 and Star in the Mapk14gc^−/− GCs in culture contrast with previous studies, in which the MAPK14 inhibitor SB20 suppressed Cyp19a1 and estradiol production and increased Star and progesterone biosynthesis (39,40). The underlying molecular differences remain unclear but could be related to the stage of GC differentiation (DES-treated rats vs. eCG 24-h primed mice) or the use of mice vs. rats.

The molecular basis for why differences in gene expression profiles in COCs and GCs in the Mapk14gc^−/− mice were more potently revealed in culture remains to be determined. One possibility is that other MAPK isoforms exert redundant functions (6). Although we cannot rule this out, the expression levels of Mapk11, Mapk12, and Mapk13 are extremely low in ovarian cells, and they are not increased in a compensatory manner in the Mapk14gc^−/− mouse ovaries (data not shown). Moreover, MAPK14 alone has been shown to be both necessary and sufficient for responding to inflammatory signals in vivo (23), and this is highly relevant to the events that control ovulation (15). Alternatively, because MAPK14 is activated in other cells and tissues during stress situations, it is possible that COCs and GCs in culture are under some kind of stress and that, in the absence of Mapk14, specific responses dependent on MAPK14 are uncovered in this context more easily than in vivo. Evidence to support this hypothesis is based on studies of other adult Mapk14 conditional knockout mice, in which the phenotypes in vivo appear to be mild but are dramatically enhanced when specific cell types are isolated and challenged in culture (28,41,42). For example, the effects of targeted disruption of Mapk14 in cardiomyocytes and hematopoetic stem cells is most potently observed when the cells are isolated and exposed to some kind of oxidative or metabolic stress. In this regard, a key role for MAPK14 has been associated with radiation damage and DNA repair, as well as the induction and release of potent cytokines from the damaged senescent cells. Strikingly, one gene that is most potently suppressed in the Mapk14 null GCs and cumulus cells in vivo and in culture is that encoding the cytokine IL-6 that is induced during the ovulation process. In addition, the activation of JNK is increased in the Mapk14gc^−/− GCs and cumulus cells and may contribute to the observed context specific changes in gene expression profiles in each cell type in the Mapk14 depleted cells.

In summary, the results of these studies show for the first time that disruption of Mapk14 gene alone in murine GCs and cumulus cells selectively alters the expression of specific genes in each cell type. Although COCs in the Mapk14gc^−/− mice fail to expand in the presence of PGE2 in culture, they can respond to AREG, indicating that one major role of PGE2 and MAPK14 in cumulus cells is the induction of AREG and EREG. That COC expansion and ovulation occur in the Mapk14gc^−/− mice in vivo may be due in part to the increase of Areg and Ereg expression in GCs of the Mapk14gc^−/− mice. Greater AREG and EREG from the GCs appears to by-passes the need for PGE2 in cumulus cells (Fig. 5D). Thus, the cell-specific effects in GCs appear to exert compensatory mechanisms allowing the Mapk14gc^−/− mice to remain fertile. Although MAPK14 is not overtly essential for preovulatory follicle development or events associated with ovulation and luteinization in vivo, it does impact gene expression profiles in both GCs and cumulus cells.

Materials and Methods

Materials

Pregnant mare serum gonadotropin (PMSG/eCG), SB20, and PD169316 were purchased from Calbiochem (La Jolla, CA). Pregnyl (hCG) was from Organon Special Chemicals (West Orange, NJ). Recombinant human FSH was the generous gift of Schering-Plough (Oss, The Netherlands). AREG was obtained from R&D Systems, Inc. (Minneapolis, MN). PGE2 was purchased from Cayman Chemical Co. (Ann Arbor, MI), and routine chemicals and reagents were obtained from Fisher Scientific (Pittsburgh, PA) or Sigma (St. Louis, MO). Antibodies for Thymoma viral proto-oncogene 1, serine-threonine kinase (AKT) (no. 9272), phospho-p44/42 MAPK (pERK1/2) (no. 9101), phospho-P38 MAPKs (no. 9211), total P38 MAPKs (no. 9212), MAPK14 (no. 9228), pJNK/SAPK (no. 9251), and pCREB (no. 9198) were from Cell Signaling Technology, Inc. (Danvers, MA).

Animals

Immature female C57BL/6 mice were obtained from Harlan, Inc. (Indianapolis, IN). On d 23 of age, female mice were injected ip with 4 IU of eCG to stimulate follicular growth, followed 48 h later with 5 IU hCG to stimulate ovulation and luteinization (43). Animals were housed under a 16-h light, 8-h dark cycle schedule in the Center for Comparative Medicine at Baylor College of Medicine and provided food and water ad libitum. Animals were treated in accordance with the National Institutes of Health Guide for the Care and Use of Laboratory Animals, as approved by the Animal Care and Use Committee at Baylor College of Medicine.

Cumulus oocyte complex isolation and culture

The procedures used for in vivo COCs isolation and in vitro culture are as described previously (15). Briefly, COCs and GCs (GCs) from preovulatory follicles of mice primed with eCG for 48 h were released into the culture medium by needle puncture of the ovary. The COCs were collected separately from the GCs by pipette, pooled, and treated as described (15). For the COC expansion assay, 15 COCs were cultured in 50 μl of COC medium. For gene expression and Western blot analyses, 100 COCs were cultured in 100 μl of COC medium.

To culture GCs, primary GCs were isolated by needle puncture from antral follicles of immature mice primed with eCG for 24 h. The cells were washed twice and plated in 24-well plates in DMEM/F12 medium (Invitrogen, Carlsbad, CA) with 5% FBS. After 24 h, the attached cells were washed twice and changed into fresh DMEM/F12 medium without serum for 2 h. The cells were then treated with specific reagents for the time interval indicated in the text.

RNA isolation and real-time PCR

Total RNA was isolated using the RNeasy mini kit (QIAGEN, Valencia, CA) according to the manufacturer’s instructions. Total RNA was reverse transcribed using 500 ng poly-dT (Amersham Pharmacia Biotech, Newark, NJ) and 200 U of M-MLV reverse transcriptase (Promega Corp., Madison, WI). Real-time PCR was performed using the Rotor-Gene 6000 thermocycler (QIAGEN) as described previously (15). The primers were designed using software Primer3 (44) and listed in Supplemental Table 1, published on The Endocrine Society’s Journals Online web site at http://mend.endojournals.org. The relative levels of gene expression were calculated using Rotor-Gene 6.0 software and normalized to levels of ribosomal protein L19 mRNA.

Western blot analyses

COCs, GCs, and whole ovaries were lysed with RIPA buffer [20 mm Tris (pH 7.5), 150 mm NaCl, 1% Nonidet P-40, 0.5% sodium deoxycholate, 1 mm EDTA, and 0.1% sodium dodecyl sulfate] containing complete protease inhibitors (Roche, Indianapolis, IN). Western blot analyses were performed using cell lysates equivalent to 100 COCs or 15 μg of protein for GCs and whole ovaries. All the primary antibodies are used at 1:1000 dilutions.

Immunohistochemistry

Immunohistochemical analyses were done using Tyramide Signal Amplification kit (PerkinElmer, Waltham, MA) according to the manufacture’s recommendations with the antibodies specified in the text and figure legends. The primary antibodies were used at 1:100 dilutions. Digital images were captured using a Zeiss AxioPlan2 Microscope with ×5, ×10, and ×20 objectives (Carl Zeiss, Thornwood, NY) in the Integrated Microscopy Core.

Statistics

Differences among groups were analyzed by Student’s t test. P ≤ 0.05 was considered significant. Data shown were mean ± sem.