Expression of mRNAs for DNA-methyltransferases and histone deacetylases in granulosa cells and follicular fluid of women undergoing in vitro fertilization − a pilot study

INTRODUCTION

Epigenetic reprogramming during follicular maturation, oogenesis and pre-implantation embryo development has been shown to contribute to the reproductive outcome.^[1,2] Epigenetic modifications are achieved by four major mechanisms: (a) DNA methylation, (b) histone acetylation, (c) micro RNA expression and (d) nucleosome positioning.^[3,4] Currently, the impact of DNA-methylation and histone acetylation of maturing oocytes and early embryos on the reproductive performance is the subject of intensive research.^{[5,6,7,8,9,10,11]} The pattern of DNA- methylation is mediated by the enzymes DNA methyltransferases that add methyl groups to DNA. Two types of DNA methylation have been documented; maintaining methylation is catalysed by DNMT1, while de novo methylation is performed by DNMT3a and DNMT3b.^[12,13] The acetylation status of histone proteins has also been claimed to play a critical role in transcriptional regulation and structural organization of chromatin, and it has been proposed to be a potential biomarker for oocyte quality.^[10,14] Histone acetylation occurs via the histone acetyltransferase enzyme (HAT) and deacetylation via the enzymes histone deacetylases (HDACs).^[15] NAD-dependent histone deacetylases comprise the Sirtuin family that confers cellular protection by regulating redox state, stress signalling, cell cycle and genome stability.^[16,17] In this regard, it is relevant to mention that our group recently reported that FF Sirtuin 6 and serum Sirtuin 1 and 6 were positively related to the number of mature oocytes and clinical pregnancy, respectively, when correction was made for confounders in women undergoing in vitro fertilization (IVF).^[18]

In addition to histone acetylation, methylation of arginine residues of histone- and non-histone proteins are also thought to be an important regulator of cellular functions, in particular, the structure and function of DNA. Therefore, it has also been suggested to contribute to post-translational modifications.^[19,20,21] With this notion in line, significant negative association was found between FF l-arginine methylation products and the number of mature oocytes and viable embryos. Specifically, elevated levels of FF l-arginine, symmetric and asymmetric dimethylarginines and monomethyl arginine appeared to have an adverse influence on the reproductive performances in IVF patients.^[22]

On the basis of these observations, the present study was designed to further explore the impact of gene expression profiles of the enzymes DNMT1, DNMT3a and DNMT3b, as well as HDAC5 and HDAC6 on the success of IVF program. The mRNA expression of these enzymes was determined in GC and GC-free F obtained from IVF patients at the oocytes retrieval. Attempts were also made to find clinical correlates of the gene expression of these enzymes involved in mediating DNA methylation and histone deacetylation.

MATERIALS AND METHODS

RESULTS

Table 2 shows the mRNA expression of DNMTs and HDACs in the GC and FF in all patients and separately, the pregnancy-positive and pregnancy-negative groups. Each transcript could be detected in GC and FF samples without significant differences between samples from either source.

Patients who underwent successful IVF treatment and progressed to clinical pregnancy (nine patients) were compared with those who failed to become clinically pregnant (22 patients); no consistent changes could be detected between the two groups in their DNMT1, DNMT3a, DNMT3b or in HDAC5and HDAC6 expression in GC. However, DNMT3a expression level in FF proved to be significantly higher in the pregnant than in the non-pregnant group (P < 0.016). The mRNA expression of other enzymes studied in FF appeared to be similar irrespective of the IVF outcome.

To assess the possible contribution of mRNA expressions measured in GC to their respective levels in FF, we examined the relationship between the corresponding parameters of GC and FF. Except for DNMT3a transcript (r = 0.478, P < 0.033), we failed to document associations between enzyme transcripts obtained simultaneously from GC and FF, suggesting that the mRNAs of DNMT1, DNMT3b, HDAC5 and HDAC6 originate mostly from sources other than the GC.

Interestingly, there were significant positive relationships of DNMT1 to DNMT3b, DNMT3b to HDAC5 and HDAC5 to HDAC6, respectively, in GC. Furthermore, FF DNMT1 was significantly related to FF DNMT3a and to FF DNMT3b. FF DNMT3b was also related to FF HDAC5 and HDAC6, whereas FF HDAC5 was related to FF HDAC 6 [Table 3].

The effects of mRNA expressions on outcome measures in our patients were also evaluated. The number of oocytes, matured oocytes and viable embryos, as well as serum hCG levels on day 12 and clinical pregnancy were used as indices of outcome. Out of the gene expressions studied, only the GC HDAC6 had significant impact on the number of oocytes (r = 0.404, P < 0.030), matured oocytes (r = 0.383, P < 0.040) and viable embryos (r = 0.413, P < 0.026). An attempt was also made to explore the influence of some common clinical/laboratory variables on the pattern of gene expression. No association was found between the age, BMI, number of previous IVF procedures, dose and duration of FSH stimulation, serum LH, progesterone andestradiol levels and the mRNA expression of any methyltransferases or deacetylases investigated. Because these clinical and laboratory parameters failed to significantly affect the gene expression patterns of DNMTs and HDACs, they were not considered as confounders to be adjusted for.

DISCUSSION

The present study showed that in women who underwent IVF, the mRNA expressions for DNMT1, DNMT3a, DNMT3 band for HDAC5 and HDAC6 were present in both GC and FF and except for DNMT3a, there were no significant differences between their expression levels in the two compartments. This suggests that the mRNAs of enzymes studied may also be derived from sources other than GC. Concerning the association of gene expression with IVF outcome, FF DNMT3a was significantly higher in the pregnant than in the non-pregnant group and the GC HDAC6 proved to be related to the number of oocytes, matured oocytes and viable embryos.

GCs are the major somatic cell compartments of the ovarian follicles and play a crucial role in achieving developmental competence of the maturing oocytes. Accordingly, stage-specific regulation of growing mouse oocytes by GC has been demonstrated^[25] and the beneficial effects of GC co-culture on in vitro oocyte maturation in murine and human models have been established.^[26,27] Experimental and clinical studies have been published to reveal the importance of GC in basic biological processes related to oocyte quality and fertilization potential. In this regard, investigations on telomere length and telomerase activity,^[28,29,30] identification of microRNAs^[31,32] and apoptosis markers^[33,34] are to be considered.

In a recent mRNA-seq and genome-wide DNA methylation study, human ovarian GCs have been used to explore the role of genome and epigenome in the age-related decline in ovarian functions and female fertility. It was assumed that epigenetic alterations in these cells may reflect the interaction between the genome and environment. In support of this notion, significant, non-random changes in transcriptome and DNA methylome features were demonstrated in human ovarian GCs as women age and their ovarian functions deteriorate.^[34] In this regard, it is to be noted that ovarian aging not only involves decreased quality and quantity of oocytes but also those of the surrounding GCs.^[34]

In a most recent study, Yang et al. evaluated the genome-wide DNA methylation profile of human preimplantation embryos. In trophoectoderm biopsy samples from blastocytes, they demonstrated negative correlations of genome-wide methylation levels to embryo quality and to maternal age confirming that increased level of DNA methylation may compromise embryo competence.^[35]

From clinical point of view, it is to be stressed that there are marked differences in DNA methylation profiles of human oocytes and preimplantation embryos between in vitro and in vivo conceived children, which can be attributed to the assisted reproductive technologies (culture media, ovarian stimulation) rather than infertility.^{[8,36,37,38,39]}

It is to be noted that variations of DNA methylation in samples from the placental and umbilical cord tissues, saliva and cord blood of new birth weight neonates have been claimed to mediate perinatal programming of non-communicable diseases later in life. Genome-wide DNA methylation studies have revealed association between differentially methylated DNA and neurodevelopmental impairments and compromised immune functions of the neonates.^{[40,41,42,43]} The histone acetylation is a prominent player of post-translational modifications. It is controlled by two opposing enzymes: histone acetyl-transferases and histone deacetylases. In addition to histone, these latter enzymes (HDACs) catalyse deacetylation of other non-histone proteins. It has been shown that global histone deacetylation and activity of HDACs are essential for oocyte growth and survival.^[44] Histone hyperacetylation induced by HDAC inhibitor trichostatin A during meiosis resulted in chromosome instability in pre-ovulatory and in vitro matured mammalian oocytes.^[45] Importantly, GC from women with polycystic ovarian syndrome had widespread lysine acetylation of proteins, and enhanced acetylation was associated with markedly reduced two pronuclear rates and the number of viable embryos during assisted reproduction.^[46] With this observation in line, in vitro maturation of oocytes from polycystic ovaries reduced the expression of HDAC1 in MII oocytes and two-cell embryos.^[47] To further support the involvement of histone acetylation in fertilization, meiosis-specific deacetylation^[48] and increased acetylation levels during post-ovulatory aging^[49] have been reported in mouse oocytes.

In our study, the gene expression profiles of DNMTs and HDACs were evaluated in the easily accessible GCs. It was assumed that due to the close interactions of GCs with oocytes,^[50] we could get information about the quality and developmental potential of oocytes. Both cellular compartments release biologically active substances into the FF, and FF is thought to mediate their cross-talk; therefore, mRNA expression levels in FF were also analysed. Except for DNMT3a, the gene expressions of other enzymes studied proved to be independent of their levels in GC which can be attributed to the possible contribution of transcripts derived from other follicular cells. Concerning the clinical relevance of our observations, it should be emphasized that only the HDAC6 expression in GC was associated with the number of retrieved oocytes, matured oocytes and viable embryos and the FF levels of DNMT3a were significantly elevated in the pregnant versus non-pregnant group. These findings may be regarded as indicating limited predictive values of the transcripts measured, so one needs to be cautious when interpreting these laboratory data.

There have been reports on the interrelations of the multiple post-translational modifications during follicular development and oocyte growth.^[1,51,52,53] The significant associations of mRNA expressions for DNMT and for HDAC we found in GC and FF are in agreement with these observations; however, further studies are warranted to confirm or exclude their causal relationships.

In conclusion, mRNAs for DNMTs and for HDACs are expressed not only in GC but also in FF. The FF transcript may originate partly from GC, partly from other cell types of follicles. In our clinical settings, mRNAs for DNMT3a and HDAC6 have the potential to assess IVF outcome.

SUMMARY

FF of IVF women contains mRNAs for DNA-methyltransferases and histone deacetylases. The FF transcripts for DNMT3a are higher in pregnant than in non-pregnant patients, also transcripts for GC HDAC6 are associated with the number of oocyte and viable embryos, therefore expression level of FF DNMT3a and GC HDAC6 has predictive potential for IVF outcome.

COMMENTARY