Introduction

Various culture systems have been described for the maturation of female gametes in vitro.1 However, the development of primordial follicles to maturity, in which oocytes acquire in vitro competence to undergo maturation, fertilization and embryonic development, remains a considerable challenge. Recent evidence suggests that mechanisms regulating oocyte growth underlie the establishment of maternal primary imprinting during mouse oogenesis.2 This lead to monoallelic expression/repression of imprinted genes involved in embryonic growth, placentation and behavioral development.3 During oogenesis, expression and/or methylation analysis of imprinted genes provide evidence for the erasure of the parental imprint in primordial germs cells.4 New imprints are then later initiated during the development of nongrowing oocytes to fully grown (fg) oocytes as indicated by a marked change of maternal germline-specific imprints.2,5 Thus, maternal primary imprinting may be particularly susceptible to methylation changes that occur at imprinted loci during oogenesis. Since previous studies in mice have suggested that culture conditions could affect the epigenetic regulation of genomic imprinting,6,7 the aim of our study was to determine the consequence of in vitro folliculogenesis on the progress of maternal primary imprinting during germinal vesicle-stage oocyte growth.

Materials and methods

Preantral follicles containing early growing germinal vesicle-stage oocytes (eg oocytes) were isolated from ovaries of 11- and 12-day old F1 hybrid female mice (C57bl6 × CBA). Early preantral follicles were selected according to the following criteria: intact follicle structure with one or two layers of granulosa cells, visible round and central oocyte, adhering thecal cells, follicle diameter between 100 and 130 μm. Follicles were placed individually into 10 μl drops of culture medium, under pre-equilibrated mineral oil (Sigma-Aldrich) at 37°C in an atmosphere of 5% CO2 in air, using a system similar to the one previously described.8 Culture medium consisted of α minimal essential medium (Life Technologies) supplemented with 5% heat inactivated FCS, 10 μg/ml transferrin (Boehringer Mannheim, France), 5 μg/ml insulin (Boehringer Mannheim) and 100 mIU/ml r-FSH (Serono, France). Healthy oocyte–cumulus complexes consisting of compact granulosa cells with regular borders surrounding the entire mass were used in these experiments. On day 11 of culture, fg oocytes were freed of their surrounding cumulus cells by repeated pipetting in 1 mg/ml hyaluronidase (Sigma-Aldrich) and washed in serum-free modified HTF medium (Irvine Scientific, USA). As control, we used antral follicle grown in vivo collected on the surface of the ovary of adult mice.

Bisulfite treatment of 5–8 fg oocytes imbedded in agarose beads was performed as previously reported.9 Special care was taken to avoid contamination from granulosa cells. DNA methylation analysis of specific differentially methylated regions (DMRs) of H19, Mest/Peg1 and Igf2R was performed on independent pools of 5–8 fg oocytes by denaturing high-performance liquid chromatography (DHPLC) analysis as recently reported.10

Results and discussion

We have analysed the methylation status of DMRs of imprinted genes using a newly developed DHPLC-based procedure.10 Analysis of control fg oocytes from different PCR products indicated that H19 alleles remained essentially all unmethylated (Figure 1c), whereas Mest/Peg1 alleles (Figure 2c) and Igf2R alleles (Figure 3c) were almost all methylated when compared to curves representing 100 and 0% methylated/unmethylated control alleles. These methylation profiles were characteristic of the maternal imprint and provided evidence that the maternal methylation imprint was already established in in vivo fg oocytes as previously described.2,5,11

Figure 1
figure 1

Methylation analysis of imprinted H19, (a) Genomic structures of the DMR studied here. We amplified 13 CpGs of the H19 imprinting control region (accession no. U19619.1; nt −2975, −3204), Arrow, transcription start site of the gene; black box, exons; small black squares, individual CpGs within the areas amplified. (b) DHPLC chromatograms of PCR products of control samples. Methylation profiles were studied by newly developed DHPLC-based method, as recently described.10 Single peaks consisting of homoduplexes characterized homogeneous samples with methylated or unmethylated alleles, whereas multiple peaks consisting of homoduplex and heteroduplex formations characterized heterogeneous samples: (*), heteroduplexes; granulosa cells; control methylated alleles; control unmethylated alleles. (c) DHPLC chromatograms of PCR products of in vivo grown fg oocytes. n=9 chromatograms are superposed for H19, n= 8 for Mest/Peg1 and n=7 for Igf2R. (d) DHPLC chromatograms of PCR products of in vitro grown fg oocytes n=7 chromatograms are superposed: (→), unmethylated alleles; (↓), methylated alleles; (*), heteroduplexes.

Full size image
Figure 2
figure 2

Methylation analysis of imprinted Mest/Peg1 gene. (a) Genomic structure of the DMR studied. We amplified 22 CpGs of the Mest/Peg1 promoter region (accession no. AF017994; nt −201, +116 relative to transcriptional start). See Figure 1(a) for details. (b) DHPLC chromotograms of PCR products of control samples. See Figure 1(b) for details. (c) DHPLC chromotograms of PCR products of in vivo grown fg oocytes n=8 chromotograms are superposed. (d) DHPLC chromotograms of PLR products of in vitro grown fg oocytes. n=7 chromotograms are superposed. See Figure 1(d) for details.

Full size image
Figure 3
figure 3

Methylation analysis of imprinted Igf2R gene. (a) Genomic structure of the DMR studied. We amplified 13CpGs of the Igf2R within intronic DMR2 (accession no. L06446; nt 1857–2101). See Figure 1(a) for details. (b) DHPLC chromotograms of PCR products of control samples. See Figure 1(b) for details. (c) DHPLC chromotograms of PCR products of in vivo grown fg oocytes. n=7 chromotograms are superposed. (d) DHPLC chromotograms of PCR products of in vitro grown fg oocytes. n=8 chromotograms are superposed. For details see Figure 1(d).

Full size image

After in vitro folliculogenesis, DHPLC profiles of PCR products from in vitro fg oocytes showed almost unmethylated alleles in six out of seven samples at the H19 locus (Figure 1d) and methylated alleles in six out of seven samples at the Mest/Peg1 locus (Figure 2d). In one (one out of seven) PCR product, a gain of allelic methylation and a loss or absence of allelic methylation were unexpectedly observed at the H19 and Mest/Peg1 loci, respectively. Such abnormal monoallelic profiles for H19 and Mest/Peg1 were different from multiallelic patterns that have been observed in granulosa cells (Figure 1, 2 and 3b). In contrast, at the Igf2R-R2 locus, DHPLC chromatograms showed profiles corresponding to almost methylated alleles in only one out of eight PCR products, whereas multiple peaks were found in seven out of eight samples (Figure 3d). These multiple peaks, consisting of homoduplex and heteroduplex formations were characteristics of heterogeneous samples with methylated alleles and unmethylated alleles, as confirmed by genomic bisulphite sequencing (not shown). However, since PCR amplifications were performed on pools of five to eight in vitro fg oocytes, we could not establish whether this heterogeneity affected individual fg oocytes (methylated and unmethylated alleles in the same fg oocyte) or the entire population of fg oocytes (fg oocytes with methylated alleles, fg oocytes with unmethylated alleles). Nevertheless, these results indicated, for the first time, the disruption (loss, absence or delay) of maternal methylation imprint establishment during in vitro folliculogenesis. The mechanisms that could affect maternal primary imprinting establishment as a consequence of in vitro culture remain elusive. We hypothesize that these dysregulations could be stochastic events at some imprinted loci as illustrated by improper de novo methylation at the H19 locus and improper demethylation at the Mest/Peg1 locus. Alternatively, these imprinting errors could be enhanced by in vitro cultivation at other specific imprinted loci, as exemplified by results at the Igf2R-R2 locus. The consequence of imprinting methylation errors for incomplete oocyte competence requires further investigation. However, it is tempting to speculate that such epigenetic abnormalities could be involved in subsequent abnormal development.12,13,14 Recently, complete oogenesis supporting development to full term after nuclear transfer and in vitro fertilization was successfully and efficiently accomplished by in vitro folliculogenesis using a 28-day in vitro culture system.15 This complete nuclear reprogramming was consistent with establishment of normal methylation imprinting at the Igf2R locus and illustrates the critical role of culture procedures on subsequent epigenotype and phenotype. Although, in vitro folliculogenesis will enable important progress in different fields of biology and medicine, these results highlight the need for improving in vitro culture conditions and developing biomarkers to identify complete nuclear competence.