Do assisted reproductive technologies and in vitro embryo culture influence the epigenetic control of imprinted genes and transposable elements in children?

STUDY QUESTION: Do assisted reproductive technologies (ART) and in vitro embryo culture influence the epigenetic control of imprinted genes (IGs) and transposable elements (TEs) in children? SUMMARY ANSWER: Significant differences in the DNA methylation of IGs or transposon families were reported between ART and naturally conceived children, but there was no difference between culture media. WHAT IS KNOWN ALREADY: There is concern that ART may play a role in increasing the incidence of adverse health outcomes in children, probably through epigenetic mechanisms. It is crucial to assess epigenetic control, especially following non-optimal in vitro culture conditions and to compare epigenetic analyses from ART-conceived and naturally conceived children. STUDY DESIGN, SIZE, DURATION: This follow-up study was based on an earlier randomized study comparing in vitro fertilization outcomes following the use of two distinct culture media. We compared the epigenetic profiles of children from the initial randomized study according to the mode of conception [i.e. ART singletons compared with those of a cohort of naturally conceived singleton children (CTL)], the type of embryo culture medium used [global medium (LifeGlobal) and single step medium (Irvine Scientific)] and the mode of in vitro fertilization (i.e. IVF versus ICSI). PARTICIPANTS/MATERIALS, SETTING, METHODS: A total of 57 buccal smears were collected from 7- to 8-year-old children. The DNA methylation profiles of four differentially methylated regions (DMRs) of IGs (H19/IGF2: IG-DMR, KCNQ1OT1: TSS-DMR, SNURF: TSS-DMR, and PEG3: TSS-DMR) and two TEs (AluYa5 and LINE-1) were first assessed by pyrosequencing. We further explored IGs and TEs' methylation changes through methylation array (Human MethylationEPIC BeadChip referred as EPIC array, Illumina). MAIN RESULTS AND THE ROLE OF CHANCE: Changes in the IGs' DNA methylation levels were found in ART children compared to controls. DNA methylation levels of H19/IGF2 DMR were significantly lower in ART children than in CTL children [52% versus 58%, P = 0.003, false discovery rate (FDR) P = 0.018] while a significantly higher methylation rate was observed for the PEG3 DMR (51% versus 48%, P = 0.007, FDR P = 0.021). However, no differences were found between the culture media. After observing these targeted modifications, analyses were performed at wider scale. Again, no differences were detected according to the culture media, but imprinted-related DMRs overlapping promoter region near the genes major for the development (MEG3, BLCAP, and DLX5) were detected between the ART and CTL children. LIMITATIONS, REASONS FOR CAUTION: The sample size could seem relatively small, but the high consistency of our results was ensured by the homogeneity of the cohort from the initial randomized study, the standardized laboratory techniques and the robust statistical analyses accounting for multiple testing. WIDER IMPLICATIONS OF THE FINDINGS: Although this study did not report DNA methylation differences depending on the culture medium, it sheds light on epigenetic changes that could be observed in some children conceived by ART as compared to CTL children. The clinical relevance of such differences remains largely unknown, and it is still unclear whether such changes are due to some specific ART procedures and/or to parental infertility. STUDY FUNDING/COMPETING INTEREST(S): This work was supported by funding from the Agence Nationale pour la Recherche ('CARE'-ANR JCJC 2017). The authors have no conflicts of interest. TRIAL REGISTRATION NUMBER: Not concerned.

The epigenetic control of transposable elements and imprinted genes in newborns is affected by the mode of conception: ART versus spontaneous conception without underlying infertility.

STUDY QUESTION: Do assisted reproductive technologies alter DNA methylation and/or transcription of transposable elements and imprinted genes in cord blood and placenta? SUMMARY ANSWER: After ART, DNA methylation and/or transcription changes of some transposable elements and imprinted genes were found in placenta samples while transcription modifications for some transposable elements were also discovered in cord blood. WHAT IS KNOWN ALREADY: Recent studies have confirmed the increased risk of placenta-related adverse pregnancy outcomes and the excess of imprinted disorders with abnormal methylation patterns after ART, which raises the issue of a potential ART-induced epigenetic risk. STUDY DESIGN, SIZE, DURATION: A total of 51 IVF/ICSI (15 conventional and 36 ICSI) singleton pregnancies were prospectively included from January 2013 to April 2015 and compared to 48 spontaneously conceived singleton pregnancies. PARTICIPANTS/MATERIALS, SETTING, METHODS: The DNA methylation and transcription of three imprinted loci (H19/IGF2, KCNQ1OT1 and SNURF DMRs) and four transposon families (LINE-1, ERVFRD, AluYa5 and ERVW) in cord blood and placenta obtained at birth were assessed by pyrosequencing and quantitative RT-PCR, respectively. All data were adjusted for gestational age at delivery, sex of the newborn, parity and maternal age. MAIN RESULTS AND THE ROLE OF CHANCE: DNA methylation levels of H19/IGF2, KCNQ1OT1, LINE-1Hs and ERVFRD-1 were significantly lower in IVF/ICSI placentas than in control placentas, while there was no difference for cord blood. Moreover, the expression of ERVFRD-1 and LINE-1 ORF2 in cord blood and ERVFRD-1 in placenta was lower in the IVF/ICSI group than in controls. The expression of ERVFRD-1 in placenta correlated positively with birth weight and placenta weight, but only in the control group, thus pointing to the potential deregulation of syncytin function after ART. LARGE SCALE DATA: N/A. LIMITATIONS, REASONS FOR CAUTION: The control group of fertile couples having conceived within 1 year prevented us from deciphering the distinct roles of ART and infertility. WIDER IMPLICATIONS OF THE FINDINGS: These novel findings of ERVFRD (syncytin-2) expression correlating with birth weight and placenta weight suggest that more research on syncytins and pregnancy-associated diseases could lead to them being used as biomarkers or even as therapeutic targets. The epigenetic modifications in placenta for sequences involved in foetal development raise the question of their potential effects on pregnancy and future life. These results should encourage us to analyse the exact causes and consequences of epigenetic changes and strive to minimize these variations in the interests of epigenetic safety after ART. STUDY FUNDING/COMPETING INTEREST(S): The study was funded by a grant from Besançon and Dijon University Hospitals. The authors have no conflicts of interest to declare.

Transposable elements in the mammalian germline: a comfortable niche or a deadly trap?

Retrotransposable elements comprise around 50% of the mammalian genome. Their activity represents a constant threat to the host and has prompted the development of adaptive control mechanisms to protect genome architecture and function. To ensure their propagation, retrotransposons have to mobilize in cells destined for the next generation. Accordingly, these elements are particularly well suited to transcriptional networks associated with pluripotent and germinal states in mammals. The relaxation of epigenetic control that occurs in the early developing germline constitutes a dangerous window in which retrotransposons can escape from host restraint and massively expand. What could be observed as risky behavior may turn out to be an insidious strategy developed by germ cells to sense retrotransposons and hold them back in check. Herein, we review recent insights that have provided a detailed picture of the defense mechanisms that concur toward retrotransposon silencing in mammalian genomes, and in particular in the germline. In this lineage, retrotransposons are hit at multiple stages of their life cycle, through transcriptional repression, RNA degradation and translational control. An organized cross-talk between PIWI-interacting small RNAs (piRNAs) and various nuclear and cytoplasmic accessories provides this potent and multi-layered response to retrotransposon unleashing in early germ cells.

Sex-specific promoters regulate Dnmt3L expression in mouse germ cells.

BACKGROUND: Dnmt3L, a member of the DNA methyltransferase 3 family, lacks enzymatic activity but is required for de-novo methylation of imprinted genes in oocytes and for transposon repression in male germ cells. METHODS: We used northern blots, RT-PCR, 5' rapid amplification of complementary DNA (cDNA) ends (RACE), RNase H mapping, real-time/quantitative RT-PCR and in situ hybridization to identify and characterize Dnmt3L transcripts produced during germ cell development. RESULTS: Mouse Dnmt3L uses three sex-specific promoters, not the single promoter previously thought. A promoter active in prospermatogonia drives transcription of an mRNA encoding the full-length protein in perinatal testis, where de-novo methylation occurs. Late pachytene spermatocytes activate a second promoter in intron 9 of the Dnmt3L gene. After this stage, the predominant transcripts are three truncated mRNAs, which appear to be non-coding. We could also detect similar adult testis transcripts in humans. In the mouse ovary, an oocyte-specific promoter located in an intron of the neighbouring autoimmune regulator (Aire) gene produces a transcript with the full open reading frame (ORF). This is the only Dnmt3L transcript found in growing oocytes and is absent in the oocytes of Dnmt3L-/- females. CONCLUSIONS: Sex-specific promoters control Dnmt3L expression in the mouse germ line, mirroring the situation at the Dnmt1 and Dnmt3A loci.

Origins of extreme sexual dimorphism in genomic imprinting.

Roughly equal numbers of imprinted genes are subject to repression from alleles of maternal and of paternal origin. This masks the strong sexual dimorphism that underlies major aspects of imprinted gene regulation. First, imprints are established very early in the male germ line and persist for the reproductive life of the organism, while maternal genomic imprints are established shortly prior to ovulation and are erased soon thereafter in the primordial germ cells of the next generation. Second, many CpG island-associated promoters are subject to maternal methylation but no known promoters are subject to paternal-specific germline methylation. The few known paternal methylation marks are kilobases distant from the affected genes and have a low CpG density. Third, Dnmt3L is required for imprint establishment but not transposon methylation in female germ cells, while Dnmt3L is required for transposon methylation and has only a minor role in de novo methylation at imprinted loci in male germ cells. Fourth, maternally expressed genes are commonly repressed on the paternal allele by paternally expressed imprinted genes produced in cis and encoding nontranslated RNAs. It is here suggested that rapid loss of highly mutable methylated CpG sites has led to the depletion of methylation target sites in paternally repressed imprinted genes, and that an imprinting mechanism based on RNAs or local inhibitory influences of ongoing transcription of regulatory loci has evolved to counter the erosion of paternally methylated regulatory regions. This mutability model is based on the fact that paternally methylated sequences are maintained in the methylated state for a much longer time than are maternally methylated sequences, and are therefore lost at a correspondingly faster rate. The difference in timing of imprint establishment is likely to underlie the increasing sexual dimorphism of other aspects of imprinted gene expression.

DNMT3B mutations and DNA methylation defect define two types of ICF syndrome.

ICF syndrome is a rare autosomal recessive disease characterized by variable immunodeficiency, centromeric instability, and facial abnormalities. Mutations in the catalytic domain of DNMT3B, a gene encoding a de novo DNA methyltransferase, have been recognized in a subset of patients. ICF syndrome is a genetic disease directly related to a genomic methylation defect that mainly affects classical satellites 2 and 3, both components of constitutive heterochromatin. The variable incidence of DNMT3B mutations and the differential methylation defect of alpha satellites allow the identification of two types of patients, both showing an undermethylation of classical satellite DNA. This classification illustrates the specificity of the methylation process and raises questions about the genetic heterogeneity of the ICF syndrome.

Dnmt3L and the establishment of maternal genomic imprints.

Complementary sets of genes are epigenetically silenced in male and female gametes in a process termed genomic imprinting. The Dnmt3L gene is expressed during gametogenesis at stages where genomic imprints are established. Targeted disruption of Dnmt3L caused azoospermia in homozygous males, and heterozygous progeny of homozygous females died before midgestation. Bisulfite genomic sequencing of DNA from oocytes and embryos showed that removal of Dnmt3L prevented methylation of sequences that are normally maternally methylated. The defect was specific to imprinted regions, and global genome methylation levels were not affected. Lack of maternal methylation imprints in heterozygous embryos derived from homozygous mutant oocytes caused biallelic expression of genes that are normally expressed only from the allele of paternal origin. The key catalytic motifs characteristic of DNA cytosine methyltransferases have been lost from Dnmt3L, and the protein is more likely to act as a regulator of imprint establishment than as a DNA methyltransferase.

Delayed and incomplete reprogramming of chromosome methylation patterns in bovine cloned embryos.

Full-term development has now been achieved in several mammalian species by transfer of somatic nuclei into enucleated oocytes [1, 2]. Although a high proportion of such reconstructed embryos can evolve until the blastocyst stage, only a few percent develop into live offspring, which often exhibit developmental abnormalities [3, 4]. Regulatory epigenetic markers such as DNA methylation are imposed on embryonic cells as normal development proceeds, creating differentiated cell states. Cloned embryos require the erasure of their somatic epigenetic markers so as to regain a totipotent state [5]. Here we report on differences in the dynamics of chromosome methylation between cloned and normal bovine embryos before implantation. We show that cloned embryos fail to reproduce distinguishable parental-chromosome methylation patterns after fusion and maintain their somatic pattern during subsequent stages, mainly by a highly reduced efficiency of the passive demethylation process. Surprisingly, chromosomes appear constantly undermethylated on euchromatin in morulae and blastocysts, while centromeric heterochromatin remains more methylated than that of normal embryos. We propose that the abnormal time-dependent methylation events spanning the preimplantation development of clones may significantly interfere with the epigenetic reprogramming, contributing to the high incidence of physiological anomalies occurring later during pregnancy or after clone birth.

Direct analysis of chromosome methylation.

DNA methylation is a possible candidate for a genomic imprinting marker in mammals. This epigenetic modification of DNA satisfies several essential criteria for the identification of the parental origin of individual alleles and larger portions of the genome: DNA methylation is stably propagated in somatic cells during cell division, it is reversible, it may inactivate the target sequence, and male and female gametes have different methylation patterns (reviewed in ref. 1).

Whole-genome methylation scan in ICF syndrome: hypomethylation of non-satellite DNA repeats D4Z4 and NBL2.

The ICF (immunodeficiency, centromeric instability and facial abnormalities) syndrome is a rare recessive disease characterized by immunodeficiency, extraordinary instability of certain heterochromatin regions and mutations in the gene encoding DNA methyltransferase 3B. In this syndrome, chromosomes 1 and 16 are demethylated in their centromere-adjacent (juxtacentromeric) heterochromatin, the same regions that are highly unstable in mitogen-treated ICF lymphocytes and B cell lines. We investigated the methylation abnormalities in CpG islands of B cell lines from four ICF patients and their unaffected parents. Genomic DNA digested with a CpG methylation-sensitive restriction enzyme was subjected to two-dimensional gel electrophoresis. Most of the restriction fragments were identical in the digests from the patients and controls, indicating that the methylation abnormality in ICF is restricted to a small portion of the genome. However, ICF DNA digests prominently displayed multicopy fragments absent in controls. We cloned and sequenced several of the affected DNA fragments and found that the non-satellite repeats D4Z4 and NBL2 were strongly hypomethylated in all four patients, as compared with their unaffected parents. The high degree of methylation of D4Z4 that we observed in normal cells may be related to the postulated role of this DNA repeat in position effect variegation in facio- scapulohumeral muscular dystrophy and might also pertain to abnormal gene expression in ICF. In addition, our finding of consistent hypomethylation and overexpression of NBL2 repeats in ICF samples suggests derangement of methylation-regulated expression of this sequence in the ICF syndrome.

Abnormal methylation does not prevent X inactivation in ICF patients.

DNA undermethylation is a characteristic feature of ICF syndrome and has been implicated in the formation of the juxtacentromeric chromosomal abnormalities of this rare syndrome. We have previously shown that in female ICF patients the inactive X chromosome (Xi) is also undermethylated. This result was unexpected since female ICF patients are not more severely affected than male patients. Here we show that CpG island methylation is abnormal in some ICF patients but in other ICF patients, the difference in methylation pattern between Xi and Xa (active X) is maintained. The consequences of Xi undermethylation on gene expression were investigated by enzyme assays. They showed that significant gene expression did not correlate with CpG island methylation status. The widespread Xi undermethylation does not affect overall Xi replication timing and does not prevent Barr body formation suggesting that a normal methylation pattern is not required for normal chromatin organization of Xi. Molecular investigation of some X-chromosome intron regions showed that the methylation changes in ICF female patients extend to non CpG islands sequences. Our results suggest that the genetic alteration of DNA methylation in ICF syndrome has little consequence on X chromosome gene expression and chromatin organization.

Chromosome instability and immunodeficiency syndrome caused by mutations in a DNA methyltransferase gene.

The recessive autosomal disorder known as ICF syndrome (for immunodeficiency, centromere instability and facial anomalies; Mendelian Inheritance in Man number 242860) is characterized by variable reductions in serum immunoglobulin levels which cause most ICF patients to succumb to infectious diseases before adulthood. Mild facial anomalies include hypertelorism, low-set ears, epicanthal folds and macroglossia. The cytogenetic abnormalities in lymphocytes are exuberant: juxtacentromeric heterochromatin is greatly elongated and thread-like in metaphase chromosomes, which is associated with the formation of complex multiradiate chromosomes. The same juxtacentromeric regions are subject to persistent interphase self-associations and are extruded into nuclear blebs or micronuclei. Abnormalities are largely confined to tracts of classical satellites 2 and 3 at juxtacentromeric regions of chromosomes 1, 9 and 16. Classical satellite DNA is normally heavily methylated at cytosine residues, but in ICF syndrome it is almost completely unmethylated in all tissues. ICF syndrome is the only genetic disorder known to involve constitutive abnormalities of genomic methylation patterns. Here we show that five unrelated ICF patients have mutations in both alleles of the gene that encodes DNA methyltransferase 3B (refs 5, 6). Cytosine methylation is essential for the organization and stabilization of a specific type of heterochromatin, and this methylation appears to be carried out by an enzyme specialized for the purpose.

Chromosome methylation patterns during mammalian preimplantation development.

DNA methylation patterns were evaluated during preimplantation mouse development by analyzing the binding of monoclonal antibody to 5-methylcytosine (5-MeC) on metaphase chromosomes. Specific chromosome patterns were observed in each cell stage. A banding pattern predominated in chromosomes at the one-cell stage. Banding was replaced at the two-cell stage by an asymmetrical labeling of the sister chromatids. Then, the proportion of asymmetrical chromosomes decreased by one-half at each cell division until the blastocyst stage, and chromosomes became progressively symmetrical and weakly labeled. Our results indicate that chromosome demethylation is associated with each DNA replication and suggest that a passive mechanism predominates during early development.

alpha-satellite DNA methylation in normal individuals and in ICF patients: heterogeneous methylation of constitutive heterochromatin in adult and fetal tissues.

The methylation profile of ten alpha-satellites was investigated in normal individuals and in ICF (Immunodeficiency, Centromeric instability, Facial abnormalities) patients. Two out of three ICF patients showed modified methylation of these sequences, reproducing a placental profile. CENP-B boxes, the binding sites of centromeric protein B, were always skewed toward nonmethylation. Unexpected results were observed in normal individuals: in somatic adult tissues the methylation pattern of alpha-satellite DNA varied between chromosomes, and in fetal tissues these satellites were homogeneously undermethylated. Detailed methylation analysis of CENP-B boxes revealed that unmethylated alpha-satellite units coexist with thoroughly methylated regions. These observations showed that the two major components of constitutive heterochromatin are differently methylated in normal somatic and fetal tissues, since classical satellites are consistently methylated. The definite changes in the methylation profile of heterochromatin in somatic chromosomes and the asynchronous timing of methylation of classical and alpha-satellites during development may reflect specific roles of highly repeated sequences in genomic organization.

Undermethylation of Alu sequences in ICF syndrome: molecular and in situ analysis.

The methylation status of young Alu sequences was investigated in four ICF patients. In fibroblast and leukocyte DNAs, Alu repeats were either undermethylated (HhaI and HpaII digestion) or demethylated (BstUI digestion), in contrast with the methylated status of Alus in control subjects. The methylation profile exhibited in ICF patients reproduces the normal profile of placental or sperm DNA. High-sensitivity immunocytochemical detection of HhaI and HpaII restriction sites on metaphase chromosomes provided further evidence of this undermethylation. The DNA methylation defect in ICF patients, first detected in satellite DNAs (constitutive heterochromatin) and CpG islands of genes on the inactive X chromosome (facultative heterochromatin), thus includes Alu sequences that are widely distributed throughout the human genome.