EGF-Like Growth Factors As Mediators of LH Action in the Ovulatory Follicle

See allHide authors and affiliations

Science  30 Jan 2004:
Vol. 303, Issue 5658, pp. 682-684
DOI: 10.1126/science.1092463


Before ovulation in mammals, a cascade of events resembling an inflammatory and/or tissue remodeling process is triggered by luteinizing hormone (LH) in the ovarian follicle. Many LH effects, however, are thought to be indirect because of the restricted expression of its receptor. Here, we demonstrate that LH stimulation induces the transient and sequential expression of the epidermal growth factor (EGF) family members amphiregulin, epiregulin, and beta-cellulin. Incubation of follicles with these growth factors recapitulates the morphological and biochemical events triggered by LH, including cumulus expansion and oocyte maturation. Thus, these EGF-related growth factors are paracrine mediators that propagate the LH signal throughout the follicle.

In mammals, the midcycle LH surge triggers a series of changes in the ovarian follicle critical for the ovulation of a fertilizable egg. LH stimulation induces resumption of meiosis in oocytes, a reprogramming of the granulosa cells of the follicle wall (mural granulosa cells) accompanied by the expression of new sets of mRNAs and proteins, and changes in secretory properties of the cells surrounding the oocyte (cumulus cells) (1, 2). Countless in vivo models have linked a failure of these LH-activated processes to different degrees of female infertility (3). One of the most perplexing aspects of LH action on the follicle, however, is the restricted expression of the LH receptor (LHR) to mural granulosa cells (4). In spite of the profound LH effects on their function, cumulus cells and oocytes express few or no LHRs and are insensitive to direct LH stimulation. Several models have been proposed to reconcile the actions of LH in the absence of receptor expression. Mural granulosa cells, cumulus cells, and oocytes are interconnected via gap junctions that may allow the flow of intracellular mediators from the periphery to the core of the follicle (5). Alternatively, factors released by mural granulosa cells may convey the LH stimulus to cumulus cells and the oocyte (2). In view of their function as short-range mediators in tissue remodeling and cell growth and as potential intermediates in G protein–coupled receptor signaling (6), we explored the possibility that the epidermal growth factor (EGF)–related proteins fulfill the role of autocrine and/or paracrine signals propagating the LH stimulus.

A signaling network consisting of seven EGF-related ligands and four tyrosine kinase receptors (ErbBs) mediates cell-cell interactions in development and oncogenesis (7, 8). Amphiregulin (AR), epiregulin (EPI), and beta-cellulin (BTC) are three of these growth factors, with an EGF-like motif organization synthesized as integral membrane precursors. These growth factors bind homo- and heterodimers of the EGF receptor (EGFR/ErbB1) and other members of the family, including ErbB3 and ErbB4 (7, 8). The ensuing signals are required for the organogenesis of mammary gland, heart, and pancreas and for the autonomous growth of tumors (7). Here, we provide evidence of a hitherto unappreciated role for this EGF network in a physiological process central to ovulation and fertility.

An ovulatory dose of the LH analog human chorionic gonadotropin (hCG) brought about the rapid and transient expression of AR, EPI, and BTC mRNAs within 1 to 3 hours after injection (Fig. 1A). AR and EPI mRNAs were the first to appear, followed by BTC. Whereas AR and BTC expression was transient, EPI mRNA remained elevated up to 12 hours after the gonadotropin stimulation. When monitored by in situ hybridization, the expression of these genes was restricted to mural granulosa cells of preovulatory follicles, and, with the exception of BTC, a specific signal was not detected in cumulus cells surrounding the oocyte (Fig. 1B). A major reduction in expression of these mRNAs was observed in PDE4D–/– ovaries, an in vivo model where ovulation but not luteinization is impaired (9, 10) (figs. S1 and S2). Expression of EPI on the surface of granulosa cells after gonadotropin stimulation was detected by flow cytometry or by metabolic labeling and immunoprecipitation (fig. S3). Taken together with the detection of EGF-like activity in follicular fluid (11), these data strongly indicate that hCG stimulation induces the expression of EGF-like growth factors.

Fig. 1.

In vivo regulation of EPI, AR, and BTC expression by LH-hCG. (A) Northern blot analysis of AR, EPI, and BTC mRNA expression in wild-type mouse ovaries at different times (h, hours) after hCG stimulation in vivo. Total RNA was prepared and hybridized to probes corresponding to EPI, BTC, and AR. Staining of 28S RNA was used as a control for RNA loading. (B) In situ hybridization of AR, EPI, and BTC mRNA expression in wild-type ovaries before or 3 hours after hCG injection. In the bright and dark field micrographs, asterisks mark cumulus-oocyte complexes and the antrum of preovulatory follicles. Arrows indicate mural granulosa cells.

In view of the above findings, we tested whether these factors reproduce some of the LH effects in a follicle culture model that recapitulates most of the events occurring in vivo, including oocyte meiotic maturation and cumulus expansion (12). In this organ culture, LH promotes resumption of oocyte meiosis measured as germinal vesicle breakdown (GVBD) (Fig. 2A). Although BTC was only partially effective, exposure of the follicles to AR and EPI induced meiotic maturation to the same extent as LH (Fig. 2A), albeit with an important difference (Fig. 2B). Whereas LH-induced GVBD was completed only after 4 hours, EPI and AR induction was maximal in 2 to 3 hours (Fig. 2B). This finding is consistent with the hypothesis that AR and EPI function distal to LH, and the more-than-1-hour delay is attributed to the time required for LH to induce expression of the growth factors. It is noteworthy that BTC mRNA is expressed with a time course incompatible with induction of oocyte maturation because it accumulates only after 3 hours of hCG stimulation, when oocyte maturation is well underway both in vivo and in vitro (Fig. 2B and fig. S3).

Fig. 2.

EGF-like growth factors stimulate oocyte maturation in vitro. (A) Induction of meiotic maturation by AR, EPI, and BTC. Preovulatory follicles were incubated for 4 hours in the presence of LH (1 μg/ml) or increasing concentrations of the three peptides. Resumption of meiosis was assessed by scoring GVBD in oocytes dissected free of granulosa cells. Each point is the mean ± SEM of three experiments using 30 to 40 follicles for each concentration. Con., control. (B) Time course of AR, EPI, and LH stimulation of meiotic resumption. Follicles were incubated with AR, EPI (100 nM), or LH (1 μg/ml). At the indicated times, oocytes were retrieved from the follicles and scored for GVBD. Each point is the mean ± SEMof three experiments using 40 to 60 follicles per time point.

LH stimulation of intact follicles causes cumulus expansion within 10 hours of stimulation (13, 14), a phenomenon reproduced by each of the three growth factors (Fig. 3A), again with a faster time course (15). This remodeling of the extracellular matrix is preceded by the expression of Ptgs2, Has2, and Tnfaip6 genes, shown to be essential for this process. These mRNAs were all induced by the EGF-like growth factors (Fig. 3B). In a simplified model of cumulusoocyte complexes (COCs) where mural granulosa cells are removed, LH no longer affected cumulus expansion, but these growth factors were even more effective than in intact follicles. COCs isolated from preovulatory follicles displayed a compact structure with cells tightly packed around the oocyte (Fig. 3C). AR, EPI, and BTC all produced marked cumulus expansion at concentrations of 0.1 to 1 nM (shown in Fig. 3C). Moreover, they induced GVBD when spontaneous maturation was prevented with the phosphodiesterase inhibitor hypoxanthine (Fig. 3D). Conversely, denuded oocytes arrested with hypoxanthine did not resume meiosis when incubated with these factors (Fig. 3E).

Fig. 3.

EGF-like growth factors stimulate expansion and oocyte maturation in COCs. (A) Cumulus expansion is induced by BTC in follicle cultures treated as described in (C), as well as by EPI and AR (not shown). (B) LH, EPI, AR, and BTC induction of genes involved in cumulus expansion. Follicles were cultured for 3 hours with LH (1 μg/ml) or AR, BTC, or EPI (100 nM). At the end of the culture, mRNA was extracted and analyzed by reverse transcription polymerase chain reaction (RT-PCR) with primers specific for Ptgs2, Has2, Tnfaip6, and GAPDH (Materials and Methods). Cont., control. (C) Expansion of COCs exposed for 8 to 12 hours in minimum essential medium (MEM), or MEM supplemented with 5% fetal bovine serum (FBS), follicle-stimulating hormone (FSH), AR, EPI, or BTC. FSH was used because LH has no effect on COC. (D and E). AR, EPI, and BTC induction of oocyte maturation in COCs. COCs were isolated and cultured for 18 hours in the presence of 4 mM hypoxanthine with either vehicle or FSH (1 μg/ml) and, EPI, AR, or BTC (100 nM). As a control, a group of oocytes was denuded in the presence of hypoxanthine and incubated with the same concentrations of growth factors (E). At the end of the incubation, oocytes were dissected free of cumulus cells and GVBD was scored. DO, denuded oocytes.

Thus, these growth factors are among the most potent stimuli for cumulus expansion and oocyte maturation. In vitro EGF stimulation of cumulus expansion has been reported (13, 16, 17); however, the physiological importance of this stimulation was not appreciated until now because the endogenous EGFR ligands were unknown. Inhibition of EGFR tyrosine kinase with AG1478 completely blocked the effects of the growth factors, confirming that their action is mediated by EGFR (fig. S4).

Given the rapid induction of AR, EPI, and BTC, LH-hCG should cause a delayed activation of the EGFR present in follicles (fig. S5). Indeed, in vivo treatment with hCG caused EGFR phosphorylation that reached a maximum 3 to 4 hours after injection (Fig. 4, A and B). In follicle cultures, the EGFR kinase inhibitors AG1478 and compound 56 (C56) but not the inactive analog AG43 completely blocked the LH-induced oocyte maturation and cumulus expansion with EC50 (median effective concentrations) of 300 to 500 nM (Fig. 4, C and D), concentrations identical to those required to block EPI-induced maturation (fig. S4). These compounds are not toxic for the oocyte, because spontaneous maturation, which occurs when the oocyte is removed from the follicle, is not affected (Fig. 4C). Moreover, these inhibitors did not prevent the induction of AR mRNA, suggesting that the initial LH signaling is not affected (Fig. 4E).

Fig. 4.

EGF receptor phosphorylation is required for LH action. (A and B) hCG stimulation of EGFR phosphorylation. Ovaries were collected at different times after hCG injection and homogenized in radioimmunoprecipitation assay buffer, and supernatants were immunoprecipitated with an antibody against EGFR. Immunoprecipitated samples were separated by SDS–polyacrylamide gel electrophoresis and transferred and detected with either a phosphotyrosine antibody (PY-EGFR) or an EGFR-specific antibody. The ratio between the intensity of the bands is reported as the mean ± SEM of the four to five experiments performed. (C and D) Follicles (POF) were incubated with LH and increasing concentrations of the EGFR tyrosine kinase–selective inhibitors C56 (◯) and AG1478 (⚫) or the inactive analog AG43 (◼). The effect of the inhibitors on spontaneous maturation was determined with the use of COCs. Incubation was terminated after 4 hours for GVBD (C) or 12 hours for cumulus expansion (D). (E) Induction of the AR mRNA was determined by semiquantitative RT-PCR in follicles incubated for 1 hour with vehicle, LH (1μg/ml), or LH in the presence of AG1478 (3 μM) or C56 (1 μM).

Collectively, these experiments demonstrate that EGF-like growth factors induced by LH are sufficient to switch on cumulus expansion and oocyte maturation and that their signaling through EGFR is necessary for LH action. On the basis of these findings, we propose that EGF-like growth factors are paracrine mediators of LH signals during ovulation. These findings explain the observation that EGF mimics some of the LH in vitro effects. EGF induction of cumulus expansion and oocyte maturation have been observed in rat (17), porcine (18), and human follicles (19). In a clinical setting, EGF improves the quality of oocytes derived from in vitro follicle cultures (19). Finally, EPI mRNA up-regulation was detected in subtractive hybridization of LH-induced genes in rat and cow (20, 21). Therefore, we surmise that these EGF-like growth factors are, in several species, the endogenous ligands that activate the EGFR in the follicle after the LH surge. In concert with cyclic adenosine monophosphate signaling and other paracrine regulators (i.e., prostaglandins) (2, 5), this EGF network orchestrates the follicular changes that lead to ovulation.

Mice with inactivation of AR and BTC genes are available; however, no reproductive phenotype was observed (22, 23). Given the extreme redundancy of the EGF network, these findings are not inconsistent with our model. Our data show that AR, EPI, and BTC expression and their effects are overlapping, suggesting a redundant mechanism. Whereas inactivation of the EGFR produces a profound disruption of embryonic development, only double or triple knockouts for EPI, AR, and BTC would yield detectable phenotypes in vivo. This strategy would be necessary to investigate their effects on the follicle, but triple knockouts will be difficult to produce in view of the clustering of the three genes on the same chromosome (24). Thus, alternative strategies will be needed for genetic testing of our model.

Our findings have considerable implications in both physiological and clinical settings. They provide a framework with which to explain the complex effects of the LH surge in the follicle and the concerted changes taking place in somatic and germ cell compartments. The discovery of this function for the EGF-related growth factors offers a rational basis for improving in vitro maturation conditions in assisted reproduction. Finally, the presence of this network in the follicle opens new opportunities for the manipulation of follicular functions for contraception or the treatment of infertility.

Supporting Online Material

Materials and Methods

Figs. S1 to S5


References and Notes

View Abstract

Stay Connected to Science

Navigate This Article