Methylation pattern at the KvDMR in a child with Beckwith–Wiedemann syndrome conceived by ICSI

To the Editor: Epigenetic modifications are reversible and heritable changes in DNA packing and chromatin structure that can affect the access of transcription factors and gene expression without involving the DNA sequence. Genomic imprinting is an epigenetic mechanism in which gene expression depends on the parental origin (paternal or maternal). Approximately 60 imprinted genes have been identified in mammals and are particularly implicated in the regulation of fetal growth and development, function of the placenta, and postnatal behaviors [Lucifero et al., 2004]. During embryo/fetal development paternally expressed genes tend to favor the extensive use of maternal resources and to enhance fetal growth, whereas the majority of maternally expressed genes suppress fetal growth. Experiments with animal knockout models demonstrated the role of the paternally expressed imprinted gene Igf2 in the regulation of nutrient supply by the placenta and the role of the maternally expressed Cdkn1c in the nutrient requirements of the fetus [Constancia et al., 2002; Reik et al., 2003]. The precise imprinting mechanism is still unclear and the differential methylation of imprinting control regions (ICRs) mapped in or near imprinted genes is thought to be a key mechanism involved in the regulation of gene expression dependent on parental origin. Several human disorders involving birth defects have been shown to involve epigenetic alterations of genomic imprinting and recent data have suggested an association between these congenital malformation syndromes and assisted reproductive technology, including in vitro fertilization (IVF) and intracytoplasmic sperm injection (ICSI) [Cox et al., 2002; Orstavik et al., 2002; DeBaun et al., 2003; Gicquel et al., 2003; Maher et al., 2003; Halliday et al., 2004]. Beckwith–Wiedemann syndrome (BWS—OMIM #130650) is a model of imprinting congenital disorder that involves overgrowth and neoplasias. The classical clinical features of BWS are macrosomia, macroglossia, abdominal wall defects, and predisposition to embryonal cancer. BWS has been shown to involve alteration of imprinting of a cluster of imprinted genes on 11p15.5. Approximately 20% of BWS cases result from paternal uniparental disomy (UPD), whereas paternally derived duplications and maternally derived translocations of 11p15.5 account for approximately 2% of cases. Mutations in the maternally derived allele of the CDKN1C gene (also known as p57kip2) are found in approximately 10% of sporadic cases. Aberrant methylation at the paternal ICR H19DMR have been found in approximately 2–7% of cases and the loss of methylation at the maternal allele of KvDMR, an intronic CpG island within KCNQ1 gene, is the most frequent alteration found in approximately 50% of cases. In 10–15% of patients with BWS, the etiology is unknown [Weksberg et al., 2005]. We report on a different methylation pattern at the KvDMR in embryonic and extra-embryonic tissues of a child with BWS who was conceived by ICSI. The study was approved by the National Ethic Committee (CONEP) and informed consent was obtained from participating family. When ICSI was performed the maternal age was 41 and there were no characteristics of a BWS phenotype in the mother and/or father. Unfortunately, the details of the assisted reproductive procedure such as medium, culture conditions, the cause of infertility (paternal or maternal) was not accessed because of privacy issues (ICSI was performed in a private service). Prenatal ultrasonographic examination at 17 weeks demonstrated polyhydramnios, overgrowth, macroglossia, omphalocele, and renal enlargement (Fig. 1). Although these alterations suggested a diagnosis of BWS, the differential diagnosis of Pallister–Killian syndrome (PKS or mosaic tetrasomy 12p—OMIM #601803) could not be excluded by ultrasonography, and despite the different etiology, both BWS and PKS were recently reported in cases of children conceived by assisted reproductive techniques [Gicquel et al., 2003; Maher et al., 2003; Chiurazzi et al., 2004]. Cytogenetic analysis by GTG-banding and by fluorescent in situ hybridization (FISH) using probes for 12p13 was carried out on cells obtained from the amniotic fluid (AF). A normal karyotype (46, XY) and the presence of only two copies of the 12p13 region (FISH), that excluded PKS, were detected (data not shown). Clinical analysis of the newborn confirmed the suspicion of BWS (Fig. 1). Prenatal ultrasonographic findings showing: (A) omphalocele, (B and C) macroglossia, and (D) renal enlargement. Prenatal 3D ultrasonographic features: (E) omphalocele. Postnatal aspect of the patient: (F) face with macroglossia and ear indentation (arrow). Umbilical cord blood (UCB) and placental tissue were obtained immediately after delivery. Placental tissue was rigorously washed in a 0.9% NaCl solution to exclude contamination with maternal blood. DNA obtained from the placenta was genotyped by the informative RFLP IGF2/ApaI to confirm the fetal origin of the sample. Paternal UPD of 11p15.5 was analyzed by RFLPs IGF2/ApaI [Tadokoro et al., 1991] and H19/RSaI [Zhang and Tycko, 1992] of a fetal DNA sample from AF (19 weeks) and from parental peripheral blood. Paternal UPD was considered unlikely by the identification of a maternally inherited allele after comparison between fetal and parental genotypes, although, it is difficult to exclude the presence of low level of mosaicism. The methylation pattern at KvDMR (GenBank U90095) was analyzed by methylation-specific PCR (MS-PCR) [Herman et al., 1996] of DNA samples obtained from AF, UCB, and placenta, chemically modified by sodium bisulfite [Clark et al., 1994]. We used the Methprimer software [Li and Dahiya, 2002] for prediction of the CpG island and for primer designs (Left M primer 5′-TTTTTTCGGTTAATGATAGGATACG-3′, Right M primer 5′-TCTACCTAAAAACTACGACAACGCT-3′ and Left U primer 5′-TTTTTTGGTTAATGATAGGATATGG-3′, Right U primer 5′-TCTACCTAAAAACTACAACAACACT-3′). We observed hypomethylation at KvDMR (consistent with the BWS phenotype) in DNA samples from AF and from UCB. Nevertheless a normal monoallelic methylation pattern at KvDMR was observed in DNA samples obtained from the placenta (Fig. 2). Molecular results of a BWS patient conceived by ICSI: (A) UPD analysis by RFLP H19/RsaI genotype: F, father; P, patient; Mo, mother. B: MS-PCR from the BWS patient: M, methylated allele; U, unmethylated allele; AF, amniotic fluid; UCB, umbilical cord blood; PL, placental tissue; (C) genotyping of patient's placenta (PL) and maternal peripheral blood (Mo) samples by RFLP IGF2/ApaI; Mar, molecular marker. Molecular results for control samples: (D) MS-PCR for naturally conceived infant (negative control); UCC, umbilical cord blood control; PLC, placenta control; PBC, peripheral blood control; (E) MS-PCR for BWS positive control; PBB, peripheral blood from a naturally conceived BWS patient. In order to confirm the different methylation pattern and quantify the percentage of methylated and unmethylated DNA we developed a real-time PCR approach associated to enzymatic restriction with the methylation-sensitive HpaII. The amount of methylated DNA template was quantified by an ABI PRISM 7500 Sequence Detection System (Applied Biosystems, Foster City, CA) in a total reaction volume of 25 µl containing the SYBR Green PCR Master Mix (Applied Biosystems) and 10 pmol of each primer (sense: 5′-GTGCCTCTCAGCGTGGTCC-3′; antisense: 5′-AACCACGATGACTGACGCAC-3′). The quantitative methylation analysis was obtained by comparison of the amount of amplifiable template remaining after HpaII digestion and undigested DNA. Both digested and undigested DNA were amplified by real-time PCR and the cycle at which the level of fluorescence reached an arbitrary threshold level of fluorescence (the ct value) was obtained for each sample using ABI software. Each reaction was carried out in triplicate and an average ct was obtained. According to the real-time PCR principle of exponential sample amplification (i.e., a doubling of product at every cycle) the percentage of methylated DNA (amplifiable template remaining after HpaII digestion) was quantified relatively to the undigested DNA (100%). The percentage of methylated sample was calculated by (1/2)n, where n = number of cycles, which was obtained by subtraction of the average ct of digested DNA by the average ct of undigested DNA. Normal differential methylation in placenta and hypomethylation in AF and UCB were confirmed by the real-time PCR approach since it was observed that 50.3% of placenta amplifiable template remained after HpaII digestion whereas only 2.1% and 2.7% of AF and UCB, respectively, were observed (Table I). We used the software Prism Graphpad for graphic representation of real-time PCR results as average and standard deviations (Fig. 3). Graphic representation of results obtained by Real-Time PCR approach: (A) average threshold cycles of digested (with HpaII) and undigested DNA samples from placenta (PL), amniotic fluid (AF), and umbilical cord blood (UCB); (B) percentage of amplifiable DNA remaining after HpaII digestion. The evidence of this BWS case with loss of methylation at KvDMR (normally methylated in the maternal allele) may support the hypothesis of an epigenetic vulnerability of maternal gametes to the methods used for fertilization by ICSI. Literature data suggest that the origin of epigenetic disorders following assisted reproduction could be a consequence of the hormonal hyperstimulation of the ovary of women who undergo fertilization, and that gonadotrophins might cause the premature release of immature oocytes that have not completed the establishment of their imprints [Shi and Haaf, 2002, Borghol et al., 2006]. In a recent study involving 19 BWS patients conceived by IVF and ICSI, Chang et al. [2005] did not find associations between predisposition to BWS and the in vitro media or timing of embryo transfer. Nevertheless, these authors speculated an association with the use of ovarian stimulation medication, since it was a common feature for the analyzed group. The unexpected result of different maternal methylation pattern at KvDMR in placenta (normal) and AF/UCB (hypomethylated) could be explained by three hypotheses: The first is an epigenetic error of acquisition or maintenance of epigenetic marks occurring at an early stage of embryo development, after the epigenetic cell fate (embryonic or extra-embryonic), and resulting in an abnormal methylation pattern at maternal KvDMR only in cells from embryonic origin (AF and UCB) and not in cells from extra-embryonic origin (placenta). Studies have suggested that the maternal genome appears to be geared towards embryo development while the paternal genome is directed towards the development of extra-embryonic tissues. Since KvDMR is methylated in the maternal allele and unmethylated in the paternal allele, errors in the methylation acquisition or maintenance in cells from embryonic origin result in alterations only in tissues from the embryo itself, without affecting normal development and function of extra-embryonic tissues (placenta) [Surani and Barton, 1983]. The second possible hypothesis is that there was a vanishing twin phenomenon in early pregnancy followed by placenta fusion. Weksberg et al. [2002] showed that the incidence of monozygotic twins among patients with BWS is dramatically increased over that of the general population. The authors also suggested that KvDMR is less efficiently methylated by the human ortholog of Dnmt10 (an oocyte-specific form of the major maintenance DNA methyltransferase) than are other DMRs, and that BWS arises when maintenance methylation at KvDMR fails to occur [Weksberg et al., 2002]. Although, in the present study, singleton pregnancy and normal karyotype (46,XY) were first demonstrated by ultrasonography and villous biopsy, respectively (data not shown), at the 13th week of gestation, an early loss of a second unaffected fetus and placenta fusion could not be excluded. A third hypothesis to consider is an epigenetic adaptation (reversion) of the placenta. Although in a previous study in mid-gestation mice Mann et al. [2004] showed that the placenta displays a loss of imprinting at H19 and a decrease in methylation at the paternal H19DMR suggesting that mechanisms that safeguard imprinting in mice may be more robust in the embryo than in the placenta, some differences in methylation pattern were evidenced in studies involving human embryos [Fuke et al., 2004; Fulka et al., 2004]. A recent study of DNA methylation patterns in human embryos conceived by IVF or ICSI revealed, by the anti 5-methylcytosine (5-Mec) staining method, a completely different pattern of labeling between trophectodermal cell (intensively labeled) and inner cell mass (ICM; weaker intensity of labeling) contrary to the typical pattern in mouse (heavily labeled ICM and weakly labeled trophectoderm) [Fulka et al., 2004]. Additionally, an increase in 5metC content was reported in normal placenta with advance of gestational age by high performance liquid chromatography (HPLC) method [Fuke et al., 2004]. Since both methodologies, 5-Mec staining and HLPC, do not provide information in terms of specific loci, the global gain of methylation reported in these studies may be involved in human placental function. Furthermore, these opposite findings could also suggest that a different scenario might exist between human and mice. This is the first report, in humans, in which embryonic and extra-embryonic tissues from the same patient have been analyzed for methylation. Based on the present finding of a different methylation pattern at KvDMR in cells from embryonic and extra-embryonic origin, we argue that care must be taken in prenatal sampling for genetic testing of BWS since chorionic villus may reveal a normal methylation pattern whereas the hypomethylated pattern consistent with BWS could be present in AF and UCB samples. Future studies are required to investigate this further. In the past 10 years, the complexity of the machinery of epigenetic programming before and after fertilization and during embryo development has been demonstrated, and the influence of (assisted reproductive technologies) ART on this complex process is still an enigma. Despite ideological, religious, and financial interests, a consensus is growing that ART does not replicate the in vivo environment. The consequence of this procedure for the embryo (child) may manifest during development, later in adult age or not at all. For this reason, clinical monitoring of children conceived by IVF is desired. The authors would like to thank the directors, doctors, and nurses of Maternidade do Complexo Aeroporto (MATER) and from the Hospital of the School of Medicine of Ribeirão Preto, University of São Paulo, Brazil, for their help in obtaining control samples and Dr. Jeremy Squire for suggestions. The authors also acknowledge FAEPA and CNPq (process no. 141560/2003-7) for financial support.