UBC Faculty Research and Publications

No evidence for association of MTHFR 677C>T and 1298A>C variants with placental DNA methylation Del Gobbo, Giulia F; Price, E. Magda; Hanna, Courtney W; Robinson, Wendy P Mar 13, 2018

Your browser doesn't seem to have a PDF viewer, please download the PDF to view this item.

Item Metadata


52383-13148_2018_Article_468.pdf [ 812.83kB ]
JSON: 52383-1.0364280.json
JSON-LD: 52383-1.0364280-ld.json
RDF/XML (Pretty): 52383-1.0364280-rdf.xml
RDF/JSON: 52383-1.0364280-rdf.json
Turtle: 52383-1.0364280-turtle.txt
N-Triples: 52383-1.0364280-rdf-ntriples.txt
Original Record: 52383-1.0364280-source.json
Full Text

Full Text

RESEARCH Open AccessNo evidence for association of MTHFR677C>T and 1298A>C variants withplacental DNA methylationGiulia F. Del Gobbo1,2, E. Magda Price1,2, Courtney W. Hanna3,4 and Wendy P. Robinson1,2,5*AbstractBackground: 5,10-Methylenetetrahydrofolate reductase (MTHFR) is a key enzyme in one-carbon metabolism thatensures the availability of methyl groups for methylation reactions. Two single-nucleotide polymorphisms (SNPs) inthe MTHFR gene, 677C>T and 1298A>C, result in a thermolabile enzyme with reduced function. These variants, inboth the maternal and/or fetal genes, have been associated with pregnancy complications including miscarriage,neural tube defects (NTDs), and preeclampsia (PE), perhaps due to altered capacity for DNA methylation (DNAm). Inthis study, we assessed the association between MTHFR 677TT and 1298CC genotypes and risk of NTDs, PE, ornormotensive intrauterine growth restriction (nIUGR). Additionally, we assessed whether these high-risk genotypesare associated with altered DNAm in the placenta.Results: In 303 placentas screened for this study, we observed no significant association between the occurrenceof NTDs (N = 55), PE (early-onset: N = 28, late-onset: N = 20), or nIUGR (N = 21) and placental (fetal) MTHFR 677TT or1298CC genotypes compared to healthy pregnancies (N = 179), though a trend of increased 677TT genotype in PE/IUGRtogether was observed (OR 2.53, p = 0.048). DNAm was profiled in 10 high-risk 677 (677TT + 1298AA), 10 high-risk 1298(677CC + 1298CC), and 10 reference (677CC + 1298AA) genotype placentas. Linear modeling identified no significantlydifferentially methylated sites between high-risk 677 or 1298 and reference placentas at a false discovery rate < 0.05 andΔβ ≥ 0.05 using the Illumina Infinium HumanMethylation450 BeadChip. Using a differentially methylated region analysisor separating cytosine-guanine dinucleotides (CpGs) by CpG density to reduce multiple comparisons also did not identifydifferential methylation. Additionally, there was no consistent evidence for altered methylation of repetitive DNA betweenhigh-risk and reference placentas.Conclusions: We conclude that large-scale, genome-wide disruption in DNAm does not occur in placentas with thehigh-risk MTHFR 677TT or 1298CC genotypes. Furthermore, there was no evidence for an association of the 1298CCgenotype and only a tendency to higher 677TT in pregnancy complications of PE/IUGR. This may be due to small samplesizes or folate repletion in our Canadian population attenuating effects of the high-risk MTHFR variants. However, givenour results and the conflicting results in the literature, investigations into alternative mechanisms that may explain the linkbetween MTHFR variants and pregnancy complications, or in populations at risk of folate deficiencies, are warranted.Keywords: MTHFR, DNA methylation, Placenta, One-carbon metabolism, 450k array, Neural tube defects, Preeclampsia,IUGR* Correspondence: wrobinson@bcchr.ca1BC Children’s Hospital Research Institute, 950 W 28th Ave, Vancouver, BCV5Z 4H4, Canada2Department of Medical Genetics, University of British Columbia, 4500 Oak St,Vancouver, BC V6H 3N1, CanadaFull list of author information is available at the end of the article© The Author(s). 2018 Open Access This article is distributed under the terms of the Creative Commons Attribution 4.0International License (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, andreproduction in any medium, provided you give appropriate credit to the original author(s) and the source, provide a link tothe Creative Commons license, and indicate if changes were made. The Creative Commons Public Domain Dedication waiver(http://creativecommons.org/publicdomain/zero/1.0/) applies to the data made available in this article, unless otherwise stated.Del Gobbo et al. Clinical Epigenetics  (2018) 10:34 https://doi.org/10.1186/s13148-018-0468-1BackgroundOne-carbon metabolism (OCM) is a fundamental biochem-ical pathway that activates and transfers methyl (CH3)groups for purine synthesis and methylation of DNA, pro-teins, and lipids, making it important for processes such asDNA synthesis, cellular division, and proliferation. Bothfunctional and dietary deficiencies are thought to contributeto altered OCM cycling. Several B vitamins act as substratesor cofactors for OCM, most notably vitamin B9 or folate,the transporter of methyl groups in OCM. Genetic variantsin a central OCM enzyme, 5,10-methylenetetrahydrofolatereductase (MTHFR), have been heavily researched in asso-ciation with human diseases, such as cardiovascular disease,pregnancy complications, and cancers [1–4]. MTHFRcatalyzes the irreversible reduction of 5,10-methylenetetra-hydrofolate (5,10-CH3-THF) to 5-methyltetrahydrofolate(5-CH3-THF). 5-CH3-THF is subsequently used as the sub-strate for the conversion of homocysteine to methionine,catalyzed by the enzyme methionine reductase. Methionineis then used to synthesize S-adenosylmethionine (SAM),the universal methyl donor for methylation reactions, in-cluding DNA methylation (DNAm), catalyzed by DNAmethyltransferases (DNMTs). As such, MTHFR is key todirecting one-carbon units toward DNAm reactions, whichhas motivated the investigation of alterations in DNAm asthe mechanism underlying the association of genetic vari-ants inMTHFR with various pathologies.Two single-nucleotide polymorphisms (SNPs) in theMTHFR gene, 677C>T (rs1801133) and 1298A>C(rs1801131), result in reduced MTHFR function in vitro,particularly in the homozygous recessive state [5–8].These variants are common in the population; globally,the variant allele frequencies are approximately 0.25–0.3(dbSNP [9]), though frequencies vary between differentpopulations. These variants have been associated withmarkers of altered OCM, such as increased levels ofhomocysteine and altered levels of blood folates [10–16],most consistently for the 677 variant. High-risk MTHFRgenotypes (677TT and 1298CC) or variant alleles (677Tand 1298C) have been found in association with a num-ber of reproductive and developmental pathologies.MTHFR 677 and 1298 variants in affected pregnanciesor parents have been associated with miscarriage [17–19]and neural tube defects [20–25]. The 677T allele and677TT genotype in mothers have been associated with pre-eclampsia (PE), a maternal hypertensive disorder in preg-nancy [26–28]. Associations between fetal-placentalMTHFR 677TTgenotypes have been identified [29], thoughthese are not as well studied as the maternal variants.Researchers have hypothesized that increased risk ofpathology might be attributed to aberrant patterns inDNAm, resulting from altered OCM flux caused bythese MTHFR variants [25, 30, 31]. While several studieshave investigated the association of the MTHFR 677C>Tand 1298A>C variants with altered DNAm, results areinconsistent; some have reported associations betweenthe high-risk homozygous MTHFR genotypes and/orfolate levels and altered DNAm [32–37], whereas othersfind no association [38–41]. As gene expression, DNAmpatterns, and metabolic requirements are highly variablebetween tissues, even these conflicting results ascer-tained in adult, non-pregnant blood, may not generalizeto pregnancy complications.The placenta is a directly relevant tissue in which tostudy the interaction between MTHFR variants, alteredDNAm, and pregnancy complications. Due to the demandfor DNA synthesis, cellular division, and proliferation bythe growing fetus and placenta, the requirement for folateduring pregnancy increases by approximately 5–10 timesthe level of non-pregnant women [42]. High-affinity folatereceptors on maternal-facing trophoblast cells allow theplacenta to transport and concentrate folate from the ma-ternal blood up to three times within the placenta [43, 44],ensuring the availability of this crucial nutrient duringdevelopment. Consistent with studies in other tissues, theMTHFR 677T allele is associated with reduced MTHFRenzyme function in the placenta [45]. If OCM flux isimpaired and DNAm patterns are altered in the placentadue to reduced variant MTHFR function, this could haveimplications for placental function and thus increase riskof pregnancy complications. Aberrant DNAm in the formof imprinting is known to have significant impact onplacental development (reviewed in [46, 47]). Genome-wide or imprinted gene-specific alterations in DNAm havebeen noted in placental insufficiency complications ofintrauterine growth restriction (IUGR) [48] and in early-onset PE [49–51]. Additionally, the placenta is a tissuethat may be more likely to exhibit altered DNAm inresponse to reduced MTHFR enzyme function. Theplacenta exhibits a high degree of within- and between-individual variability in DNAm [52, 53], suggesting that itmay be tolerant to changes in DNAm, allowing this organto adapt to environmental conditions [52–54].To date, no studies have investigated the associationbetween DNAm in the placenta and the high-riskMTHFR 677TT and 1298CC genotypes. This may pro-vide insight in to how the MTHFR variants have beenpreviously associated with pregnancy complications, andpotentially help to resolve the currently conflictingliterature investigating the association between thesevariants and DNAm in other tissues. In this study, weevaluated whether fetal high-risk MTHFR genotypeswere more prevalent in pregnancy complications of PE,IUGR, and neural tube defects (NTDs) using 303placental DNA samples. The DNAm patterns of 30placentas were heavily profiled using both site-specificand genome-wide techniques, including the IlluminaInfinium HumanMethylation450 BeadChip array andDel Gobbo et al. Clinical Epigenetics  (2018) 10:34 Page 2 of 16repetitive DNA methylation, to understand the relation-ship between high-risk MTHFR 677TT and 1298CCgenotypes and DNAm in the placenta.MethodsEthics and sample collectionEthics approval for this study was obtained from theUniversity of British Columbia/Children’s Hospital andWomen’s Health Centre of British Columbia ResearchEthics Board (H04-70488, H10-01028). Placentas werecollected from term deliveries at BC Women’s Hospitaland Health Centre and from second trimester stillbirths,elective terminations, and spontaneous abortions throughthe embryo-fetal pathology laboratory. Cases with a pre-natally identified chromosomal abnormality were excluded.A minimum of two distinct sites were sampled fromthe fetal side of each placenta after fetal membranes(amnion and chorion) were removed. Samples werewashed thoroughly with PBS to remove maternal blood.DNA was extracted by a standard salting-out proceduremodified from Miller et al. [55] and quality evaluatedusing a NanoDrop ND-1000 (Thermo Scientific). Onesite from each placenta was selected at random for geno-typing. As DNAm varies significantly within the placenta[52, 53], DNA was combined in equal amounts from atleast two sites to generate a more representative samplein which to evaluate placental DNAm.Case characteristicsA total of 303 placentas were screened for MTHFR 677and 1298 polymorphisms. These included 179 placentasfrom uncomplicated pregnancies, 48 from pregnanciesassociated with PE (28 early-onset PE, 20 late-onset PE),21 from pregnancies associated with intrauterine growthrestriction in the absence of maternal hypertension(normotensive IUGR, nIUGR), and 55 from pregnancieswith a fetal NTD (Table 1). PE was defined according tothe Society of Obstetricians and Gynaecologists ofCanada (SOGC) criteria as pregnancies with (i) gesta-tional hypertension (BP > 140/90 mmHg) and protein-uria (> 300 g/day) arising after 20 weeks of gestation; (ii)pre-existing hypertension with superimposed gestationalhypertension, proteinuria, and/or one or more adversematernal or fetal conditions; or (iii) gestational hyperten-sion without proteinuria, with one or more adversematernal or fetal conditions [56]. PE was subdivided intoearly-onset preeclampsia (EOPE), defined as a diagnosisof PE before 34 weeks of gestation, and late-onset pre-eclampsia (LOPE), a diagnosis of PE after 34 weeks ofgestation [57]. IUGR commonly co-occurs with PE andwas also defined following the SOGC criteria as birthweight < 3rd percentile, accounting for both fetal sexand gestational age (GA), or birth weight < 10th percent-ile with additional clinical findings indicative of poorgrowth such as uterine artery notching, absent orreversed end-diastolic velocity on Doppler ultrasound,or oligohydramnios [58]. nIUGR was defined as unex-plained IUGR without the presence of maternal hyper-tension. NTDs were defined as a fetus diagnosed withspina bifida, anencephaly, or encephalocele on ultra-sound and/or fetal autopsy.MTHFR genotypingPlacental DNA was genotyped for the MTHFR 677 and1298 polymorphisms using pyrosequencing. Primersequences and reaction conditions can be found inAdditional file 1: Table S1. Five microliters of PCR prod-uct was sequenced on a PyroMark Q96 MD Pyrosequen-cer (Qiagen) using standard protocols [59]. A subset ofthe genotyping results from the NTD group (N = 36) hasbeen published elsewhere [60].Population stratificationMinor allele frequencies for the MTHFR 677 and 1298SNPs vary significantly between different populations[61–64], as do the prevalence of NTDs and PE/IUGRpathologies [65]. Both high-risk MTHFR genotypes varyby ethnicity and geography, indicating that selectivepressures have influenced their frequencies [62, 66].Therefore, prior to performing a genetic association ana-lysis, we aimed to assess whether our pregnancy compli-cation groups were matched for ancestry. Maternal self-reported ethnicity was available for only 67% of cases,and no information about the father’s ethnicity wasavailable. We thus used a panel of 57 ancestry-informative marker (AIM) SNPs [67–69] that wereTable 1 Clinical characteristics of casesN Sex; N male (% male) Gestational age (weeks); median (range) Maternal age (years); median (range)Control 179 85 (47) 39.6 (19.4–41.9) 34.4 (23.8–42.7)EOPE 28 17 (61) 32.7 (23.6–38.4)* 36.0 (19.7–42.9)LOPE 20 10 (50) 38.5 (34.9–41.4)* 34.3 (26.1–41.5)nIUGR 21 9 (42) 36.2 (24.0–40.6)* 34.5 (26.1–42.8)NTD 55 28 (51) 21.0 (16.7–23.7)* 30.4 (17.7–40.6)*EOPE early-onset preeclampsia, LOPE late-onset preeclampsia, nIUGR normotensive intrauterine growth restriction, NTD neural tube defect*p < 0.05, calculated in comparison to the control group by Fisher’s exact test for categorical variables and Mann-Whitney test for continuous variablesDel Gobbo et al. Clinical Epigenetics  (2018) 10:34 Page 3 of 16developed to distinguish between African, European,East Asian, and South Asian ancestries to infer the an-cestry of study samples and assess population stratifica-tion along three major axes of variation.Two hundred seventy-seven placental villus DNAsamples were successfully genotyped at 53 AIMs usingthe Sequenom iPlex Gold platform by the Génome Qué-bec Innovation Centre at McGill University, Montréal,Canada, with a call rate > 0.9 for both SNPs and sam-ples. Multidimensional scaling (MDS) with k = 3 dimen-sions was performed in our study samples (N = 277) inaddition to individuals (N = 2157) from African, EastAsian, European, and South Asian populations from the1000 Genomes Project (1kGP) [64] using 50 of the AIMgenotypes that were available in both cohorts. Thismethod allows 1kGP samples to be used as ancestry ref-erence populations for our admixed population and hasbeen used to identify ancestry outliers [70, 71]. The firstthree MDS coordinates were extracted for each sampleand used to describe ancestry along a continuum ratherthan in discrete groups. We believe this better reflectsancestry in admixed populations such as that in Vancou-ver, as well as potentially better representing variationwithin an ancestry group. Further description of thismethod is included in Additional file 2: Methods.MTHFR genotype and DNAmTo assess the association of MTHFR genotype with DNAm,a subset of 30 control/mild pregnancy complication placen-tas were selected for in-depth DNAm profiling, hereafterreferred to as the placental DNAm samples. None of theseplacentas had chromosomal abnormalities, as confirmed bymultiplex ligation-dependent probe amplification in one ormore sites per placenta. The effect of each MTHFR SNPwas assessed independently by comparing 10 placentas withreference genotype at each MTHFR SNP (677CC +1298AA), 10 placentas with the high-risk 677TT genotypein combination with the reference 1298AA (termed “high-risk 677”), and 10 placentas with high-risk 1298CC genotypein combination with the reference 677CC genotype (termed“high-risk 1298”). The reference, high-risk 677, and high-risk 1298 groups were matched by sex, gestational age, birthweight, and maternal reported ethnicity (Table 2). To obtainsufficient numbers in each genotype group, as the high-riskgenotypes are relatively rare in our population and we add-itionally excluded heterozygotes at either locus, some mildpregnancy complication cases (LOPE without IUGR andnIUGR) were included. We previously found no evidencefor altered placental DNAm associated with these pheno-types compared to controls in a separate study [51].Illumina Infinium HumanMethylation450 BeadChipCombined placental DNA from the 30 placental DNAmsamples described in Table 2 was purified using theQiagen DNeasy Blood and Tissue kit, and 750 ng of thisDNA was bisulfite converted using the EZ DNA Methy-lation kit (Zymo Research) following the manufacturer’sprotocols. Samples were processed following theIllumina Infinium HumanMethylation450 BeadChip(450k array) protocol [72] and scanned using the IlluminaHiScan 2000. Raw intensity was read into Illumina GenomeStudio software 2011.1, and background normalization wasapplied. Data processing was performed as described inPrice et al. [73], including sample quality checks, probe fil-tering, data normalization, and batch correction. Thisprocessing pipeline resulted in a final N = 442,355cytosine-guanine dinucleotide (CpG) sites from the450k array for analysis.Repetitive DNA methylationIn addition to the 450k array, genome-wide DNAm wasalso assessed using DNAm at repetitive Alu, LINE-1,and ribosomal DNA (rDNA) sequences [74] by pyrose-quencing in the 30 placental DNAm samples. These se-quences are dispersed throughout the genome, allowingDNAm to be measured at many sites using a single assayper repetitive sequence. DNAm at these three repetitiveDNA sequences has been shown to be altered due to dif-ferent environmental or biological factors [75–79]. Threehundred nanograms of purified combined placental villiDNA was bisulfite converted using the EZ DNAMethylation-Gold Kit (Zymo Research) following themanufacturer’s protocol. Alu and LINE-1 elements wereamplified using primer sets designed to complement theL1H and AluSx consensus sequences [80], and rDNArepeats were amplified using primers designed to targetthe rDNA promoter [81] (Additional file 1: Table S1).PCR products were sequenced on a PyroMark Q96 MDPyrosequencer (Qiagen) using standard protocols [59].The DNAm status of each CpG dinucleotide (Alu: N = 3;LINE-1: N = 4; rDNA: N = 26) was evaluated using thePyroQ CpG software (Biotage). For each assay, the cor-relation of DNAm between CpGs was confirmed, andthen an average DNAm was calculated across the CpGswithin each assay in each sample.Statistical analysesStatistical analyses were conducted in R statistical soft-ware [82]. Deviation from Hardy-Weinberg equilibrium(HWE) at the two MTHFR SNPs in controls wasassessed using an exact test for HWE. Differences in thedistribution of ancestry MDS coordinate values betweencontrol and pregnancy complication groups (EOPE,LOPE, nIUGR, NTD) were assessed using theKolmogorov-Smirnov test. The association between thefrequency of MTHFR 677TT and/or 1298CC genotypesand pregnancy complications was assessed using a one-tailed Fisher’s exact test to test the hypothesis that theDel Gobbo et al. Clinical Epigenetics  (2018) 10:34 Page 4 of 16high-risk genotypes will be increased in pregnancy com-plications compared to controls. For the placentalDNAm samples, 450k array-wide average DNAm andpercent outlier probes per sample were calculated as inPrice et al. [60]. Altered measures of genome-widemethylation, including 450k array-wide average, percentoutlier probes, Alu, LINE-1, and rDNA methylation,were assessed using the Mann-Whitney test. All p valuesfrom statistical tests involving multiple comparisons (an-cestry MDS coordinate values, altered genome-wideDNAm measures) were corrected for multiple testingusing the Bonferroni method.450k array site-specific differential methylation wasalso assessed as in Price et al. [60]. Briefly, a linear modelwith the MTHFR group as the main effect and fetal sexand gestational age included as covariates was fit toevery CpG on the array (N = 442,355). Differentialmethylation results were then extracted for the compari-son of high-risk 677 to reference and for that of high-risk 1298 to reference. These comparisons were used tocalculate group differences in DNAm (delta-beta, Δβ).Significantly differentially methylated CpG sites were con-sidered as those with a false discovery rate (FDR) < 0.05and Δβ ≥ 0.05. Two dimension-reduction techniques wereadditionally used in the 450k array data: a differentiallymethylated region (DMR) analysis, as in Price et al. [60],and an assessment of differential DNAm depending on theCpG density of the surrounding region. 450k probes wereseparated into four groups based on the CpG density: high-density islands, island shores, intermediate-density islands,and non-islands, defined as per Price et al. [83], and the un-adjusted p value distributions from the linear model wereassessed in each CpG density group separately.ResultsAnalysis of ancestry-informative markers identifies nosignificant population stratificationPrior to testing for the association of fetal MTHFR geno-types with NTDs or PE/IUGR groups, we sought to con-firm that these pregnancy complication groups were notconfounded with ancestry as the frequency of theMTHFR 677 and 1298 variants varies between differentancestry groups [61–63]. Self-reported ethnicity wasavailable from mothers for only 67% (203/303) of sam-ples, which were predominately of European (Caucasian)and East Asian ancestries. We thus described ancestryusing coordinates 1, 2, and 3 obtained through MDS ofgenotypes at 50 AIMs (Additional file 3: Table S2) in277 of our placental samples along with 2157 samplesfrom the 1kGP [64] (Additional file 4: Figure S1). Thesethree MDS coordinates were significantly different be-tween the four major continental populations from1kGP (Additional file 4: Figure S1). Furthermore, forthose samples for which we had both maternal reportedethnicities in addition to AIMs (N = 181), the three an-cestry MDS coordinates were highly concordant withmaternal reported ethnicity (Additional file 5: Figure S2).These findings suggest that this method is adequate todescribe major patterns of genetic ancestry. There wasno significant difference in the distribution of ancestryMDS coordinate values 1, 2, and 3 between the NTD,PE, and nIUGR pathology groups in comparison to con-trols (Fig. 1). We thus concluded that our pathologygroups do not show evidence of confounding by ancestry.MTHFR 677TT and 1298CC genotypes are not significantlyassociated with PE or NTDsTo investigate whether the MTHFR 677TT and 1298CCgenotypes were associated with PE, nIUGR, or NTDpathologies, we genotyped placentas at these two locifrom 179 control, 28 EOPE, 20 LOPE, 21 nIUGR, and55 NTD pregnancies. Neither SNP deviated from HWEin controls (Additional file 6: Table S3). In our popula-tion of 303 placentas collected in Vancouver, Canada,the frequencies of the variant MTHFR 677T and 1298Calleles were 0.295 and 0.290, respectively. There was nosignificant increase in the frequency of the high-riskMTHFR 677TT or 1298CC genotypes in EOPE, LOPE,nIUGR, or NTD cases compared to controls (Table 3).There was, however, a tendency for increased MTHFR677TT genotype in placentas from pregnancies compli-cated by placental pathologies (PE or nIUGR). Whenconsidered together (PE or nIUGR; N = 69), the increasein MTHFR 677TT frequency compared to controls wasnominally significant (OR = 2.53, p = 0.048).Table 2 Clinical characteristics of placental DNAm samplesN Sex; N male(%)Gestational age (weeks); median(range)Maternal ethnicity; NCaucasian (%)Birth weight (SD); median(range)Reference (677CC + 1298AA) 10 4 (40) 39.0 (36.1–41.6) 8 (80 %) 0.055 (− 1.13–1.40)High-risk 677 (677TT +1298AA)10 4 (40) 37.9 (34.6–40.3) 6 (60 %) − 0.13 (− 2.97–0.70)High-risk 1298 (677CC +1298CC)10 5 (50) 39.4 (38.6–40.7) 8 (80 %) 0.005 (− 1.61–2.20)p values were calculated in comparison to the reference group by Fisher’s exact test for categorical variables and Mann-Whitney test for continuous variablesSD standard deviationDel Gobbo et al. Clinical Epigenetics  (2018) 10:34 Page 5 of 16High-risk MTHFR 677 and 1298 genotypes are notassociated with altered genome-wide DNAm in theplacentaDue to the central role that MTHFR plays in OCM, thehigh-risk MTHFR genotypes are often hypothesized toaffect the cell’s ability to methylate DNA. We anticipatedthat such effects could potentially be more pronouncedin the placenta due to its high demand for folate in preg-nancy. We selected a subset of 30 placental samples withno, or mild, pathology in which to profile DNAm usingboth genome-wide and site-specific approaches. Theselected samples were of three MTHFR genotype groups:(1) reference (N = 10; MTHFR 677CC + 1298AA), (2)high-risk 677 (N = 10; MTHFR 677TT + 1298AA), and (3)high-risk 1298 (N = 10; MTHFR 677CC + 1298CC). Nocases with high-risk genotypes at both loci were availablein our population to test.First, these 30 placental DNAm samples were run onthe 450k array, from which several measures of DNAmwere obtained. Array-wide DNAm was calculated byaveraging of 442,355 CpG sites in each sample. Thisarray-wide measure of DNAm did not differ significantlybetween either of the high-risk MTHFR groups and thereference group (Table 4). Altered genome-wide DNAmmight not be a characteristic of all carriers of theMTHFR variants; thus, we also calculated the percentageof outlier CpG sites from the 450k array for each sampleto identify individuals exhibiting outlying patterns ofDNAm [84]. Though there was no significant differencein outlier CpGs between the high-risk 677 and referencegroup (Table 4), there was a trend for more outlyingCpG sites in the high-risk 1298 group than in the refer-ence (Table 4, Bonferroni-corrected p = 0.058).Next, the methylation of repetitive DNA sequences wasassessed in the 30 placental DNAm samples. RepetitiveDNAm assays target numerous sites in the genome thatare not well covered by the 450k array, and thus give anadditional measure of genome-wide DNAm. No signifi-cant alterations in the DNAm of Alu, LINE-1, or rDNAsequences were identified between either of the high-riskTable 3 MTHFR 677TT and 1298CC genotypes in pregnancy complicationsN 677TT frequency (N) p value† OR (95% CI) 1298CC frequency (N) p value† OR (95% CI)Control 179 0.056 (10) – – 0.101 (18) – –EOPE 28 0.107 (3) 0.249 2.02 (0.33–8.59) 0 (0) 1.00 0LOPE 20 0.150 (3) 0.129 2.96 (0.48–13.1) 0.150 (3) 0.355 1.57 (0.27–6.27)nIUGR 21 0.143 (3) 0.143 2.80 (0.45–12.3) 0.048 (1) 0.891 0.449 (0.01–3.16)NTD 55 0.091 (5) 0.260 1.69 (0.43–5.72) 0.073 (4) 0.809 0.70 (0.16–2.27)OR odds ratio, CI confidence intervals, EOPE early-onset preeclampsia, LOPE late-onset preeclampsia, nIUGR normotensive intrauterine growth restriction, NTDneural tube defect†p values, calculated by one-tailed Fisher’s exact testFig. 1 Distribution of ancestry derived from multidimensional scaling of AIM genotypes in control and pregnancy complication placentas. In N = 28EOPE, N = 20 LOPE, N = 21 nIUGR, or N = 55 NTD placentas, there were no significant differences in the distribution of ancestry MDS coordinate valuescompared to N = 179 controls at either of the three ancestry MDS coordinates (Kolmogorov-Smirnov tests, Bonferroni-corrected p > 0.05). This suggeststhat there is no population stratification by ancestry in the groups selected for this study. EOPE early-onset preeclampsia, LOPE late-onset preeclampsia,nIUGR normotensive intrauterine growth restriction, NTD neural tube defectDel Gobbo et al. Clinical Epigenetics  (2018) 10:34 Page 6 of 16MTHFR genotype groups and the reference genotypegroup (Table 4). Slightly higher methylation was seen forthe high-risk 677 group in all comparisons, though therange of values was similar. There was, however, a trendfor decreased LINE-1 DNAm in the high-risk 1298 groupcompared to the reference group (nominal p = 0.052), butthis is not meaningful after correction for multiple com-parisons. Overall, we find no conclusive evidence foraltered genome-wide DNAm in association with the high-risk 677 or high-risk 1298 genotypes in the placenta usingthese DNAm measures.High-risk MTHFR 677 and 1298 genotypes are notassociated with altered site-specific DNAm in the placentaThe DNAm status of individual CpG sites targeted bythe 450k array in association with the high-risk MTHFRgenotype groups was next assessed. A linear model wasfit to each CpG site to test for differential methylationby the genotype group, including sex and gestational ageat birth as covariates. None of the 442,355 CpG siteswas differentially methylated at a FDR < 0.05 and methy-lation difference (Δβ) > 0.05 in either of the high-riskMTHFR genotype groups compared to the referencegroup (Fig. 2).Following this finding, two dimension-reduction tech-niques were utilized to explore whether identification ofdifferences between high-risk MTHFR groups and con-trols in the 450k array data was limited due to a smallsample size or large number of test sites. Due to struc-tural or functional differences, some genomic regionsmay be more vulnerable to the effects of a reducedability to methylate DNA potentially caused by the pres-ence of variant MTHFR enzymes. As such, 450k probeswere separated into four groups based on the CpG dens-ity of the surrounding region: high-density islands, islandshores, intermediate-density islands, and non-islands.Additionally, a DMR finding tool was utilized to identifywhether any DMRs existed between high-risk MTHFRgenotype placentas and controls. Unadjusted p value dis-tributions did not show differential methylation at any ofthe four CpG density groups between high-risk MTHFRand reference placentas (Additional file 7: Figure S3),nor were any significant DMRs identified. Given theseresults, we conclude that large-scale alterations inDNAm at CpG sites measured by the 450k array in theplacenta are not commonly associated with high-risk677 or 1298 MTHFR genotypes in our population.DiscussionAltered DNAm has been proposed as a mechanismthrough which MTHFR 677TT and 1298CC genotypeshave been associated with pregnancy complications andother pathologies [25, 30, 31]. In this study, we soughtto investigate alterations in DNAm in association withhigh-risk MTHFR 677TT and 1298CC genotypes in theplacenta, a crucial tissue for the development of thefetus and a healthy pregnancy. Despite deeply profilingDNAm in N = 10 high-risk 677, N = 10 high-risk 1298,and N = 10 reference placentas using a variety of mea-sures, we identified no evidence for altered placentalgenome-wide or site-specific DNAm in association withhigh-risk MTHFR genetic variants.Given the fundamental involvement of OCM in acti-vating and transporting methyl units, if the variantMTHFR alleles influence DNAm, this effect is predictedto be widespread and not gene-specific [85]. By usingthe 450k array, we interrogated DNAm at over 440,000sites in the placental genome, assessing specific CpGsites and also genome-wide trends. This array covers99% of RefSeq genes and is widely dispersed acrossgenomic features and, therefore, can provide an accuratereflection of genome-wide changes associated with spe-cific genomic features or gene regulation. No significantdifferences in the numerous measures of altered 450karray genome-wide or site-specific DNAm were identi-fied, despite additionally utilizing dimension-reductiontechniques to account for a small sample size and largenumber of test sites. DNAm at repetitive DNAsequences, including Alu and LINE-1 repetitive elementsand rDNA repeats, was also assessed, as they are notwell covered by the array, and they allow us to interro-gate numerous locations in the genome in one pyrose-quencing assay. The Alu and LINE-1 repetitive elementshave previously been used as surrogate measures forgenome-wide DNAm [74, 86], and all three repetitivesequences have exhibited alterations in DNAm in certainpathologies and in response to environmental exposures[75–79]. Though small sample size limited our power todetect significant differences in DNAm in this study, weaimed to mitigate this by deeply profiling the 30Table 4 Genome-wide measures of altered DNAm in high-risk MTHFR and reference placentasArray-wide average DNAm (β) Outlier array sites (%) Alu DNAm (%) LINE-1 DNAm (%) rDNAm (%)Reference (N = 10) 0.407 (0.396–0.413) 0.661 (0.262–0.993) 18.9 (17.6–21.7) 52.8 (51.0–53.8) 19.4 (11.2–30.9)High-risk 677 (N = 10) 0.405 (0.397–0.409) 0.851 (0.262–6.357) 20.1 (15.7–21.4) 53.5 (49.8–57.6) 20.0 (8.4–28.9)High-risk 1298 (N = 10) 0.405 (0.395–0.413) 1.14 (0.501–3.37) 19.8 (17.5–21.6) 51.0 (46.5–55.0) 22.9 (10.2–28.9)All results are reported as median (range); p values were calculated by the Mann-Whitney test for the comparison of high-risk 677 or high-risk 1298 to the referencegroup with Bonferroni correction for multiple comparisonsβ beta value, rDNAm ribosomal RNA genesDel Gobbo et al. Clinical Epigenetics  (2018) 10:34 Page 7 of 16placental DNAm samples using a variety of measures ofDNAm to assess whether any differential methylationexists in association with the high-risk MTHFR geno-types. Our study cannot fully exclude that subtle differ-ences in placental DNAm exist in association with high-risk MTHFR genotypes or that a subset of at-risk placen-tas might show changes in DNAm while the groups as awhole did not. Despite this, the results from these nu-merous genome-wide assays reveal that at the very least,large magnitude and/or array-wide differential methyla-tion does not commonly occur in association with high-risk MTHFR genotype in the placenta.Our study is the first that has investigated the associa-tions between MTHFR 677 and 1298 variants andaltered DNAm in the placenta, and only the second thathas done so using a genome-wide DNAm microarrayplatform. Numerous studies have investigated alteredDNAm in association with MTHFR 677 and/or 1298variants using different measures of genome-wideDNAm and/or targeted gene DNAm, summarized inTable 5. Results from these various studies, mainly inthe blood, are conflicting. Certain studies have foundassociations between MTHFR 677 or 1298 polymor-phisms and altered DNAm; however, many do not findsignificant associations with altered DNAm or only findaltered DNAm in the presence of low levels of OCMnutrients (Table 5). Some of these inconsistencies maybe explained by the use of different measures of alteredDNAm (i.e., genome-wide, candidate site-specific, repeti-tive element DNAm) between studies, lack of multiple-test correction, use of different tissues, or inconsistenciesin accounting for confounding variables. Nonetheless,the effect of the MTHFR 677 and 1298 variants onDNAm is clearly complex.Several studies reviewed in Table 5 suggest that alteredDNAm in association with MTHFR 677 and 1298variants might only be present under limited folateconditions [33, 35]. The presence of folate stabilizes thevariant MTHFR 677 enzyme [87], and adequate folateattenuates the effects of high-risk MTHFR 677TT geno-type on increased homocysteine [88, 89]. Due to theretrospective nature of the study, we were unable toassess folate concentrations in the placenta or maternalblood and did not have complete information on mater-nal folic acid supplementation. Though folate status wasunknown for the cases in this study, we assume thatmost of the women in our Canadian cohort were folate-replete due to folic acid fortification in cereal and grain,increased literacy around healthy pregnancies, and highuptake of gestational monitoring. Additionally, in a studyof 368 pregnant women in Toronto, Canada, with simi-lar demographics as our population in Vancouver,Plumptre et al. found that none of these women werefolate deficient during pregnancy, even considering that7% of women did not take folic acid supplements [90]. Itis possible that in the presence of adequate folate levels,the activity of the variant MTHFR 677 or 1298 enzymesin the placentas of our study was not diminished enoughto result in a compromised OCM and altered DNAm.Despite this potential limitation, investigating alterationsin placental DNAm in association with MTHFR variantsin a folate-replete population under the hypothesis thatthis may increase risk of pregnancy complications is stillwarranted. Fortification of grain products with folic acida bFig. 2 450k array-wide differential DNAm volcano plots in high-risk MTHFR 677 and high-risk 1298 placentas. Differential methylation was determined using alinear model with the MTHFR group as the main effect and fetal sex and gestational age included as covariates. The magnitude of difference (Δβ) betweenrisk groups and reference group is depicted on the x-axis, and the significance of the comparison (−log10(adjusted p value)) is on the y-axis, for every CpGtested on the 450k array (N=442,355). a Differential methylation results between high-risk 677 and reference placentas. b Differential methylation resultsbetween high-risk 1298 and reference placentas. Neither comparison identified any CpG sites differentially methylated between the high-risk MTHFR placentasand reference placentas. 450k array Illumina Infinium HumanMethylation450 BeadChip, FDR false discovery rateDel Gobbo et al. Clinical Epigenetics  (2018) 10:34 Page 8 of 16Table 5 Literature assessing associations between MTHFR 677 or 1298 variants and altered DNAm in healthy tissuesStudy Type of DNAm assessed: specific assay Tissue Study size† ResultsStudies finding associations with DNAmStern et al. (2000) [32] Genome-wide: radiolabeled methylgroup incorporation assayBlood 677CC: N = 9677TT: N = 10677TT associated withapproximately 40% higher[3H]-methyl acceptancecapacity than 677CC (p = 0.04),reflecting globalhypomethylationCastro et al. (2004) [34] Genome-wide: cytosine extension assay Blood 677CC/1298AA: N = 17677CT/1298AA: N = 22677TT/1298AA: N = 9677CT/1298AC: N = 22677CC/1298AC: N = 20677CC/1298CC: N = 7677TT associated with higher [3H]-dCTP relative incorporationcompared to 677CC (p < 0.05).677TT/1298AA and 677CC/1298CC associated with higherrelative incorporation than677CC/1298AA (p < 0.05)McKay et al. (2012) [31] Genome-wide: LUMACandidate sites (N = 3): pyrosequencingUmbilicalcord blood677: N = 1601298: N = 132mother-infant pairsMaternal 677T allele associatedwith altered mean DNAm inIGF2 in infant cord blood(p = 0.017); maternal 1298Callele associatedwith altered DNAm at 1 CpG inZNT5 (p = 0.012) in infantcord blood.No associations withgenome-wide DNAmvan Mil et al. (2014) [100] Candidate sites (N = 11): MassArrayEpiTYPERUmbilical cordblood677CC or 677CT:N = 413 677TT: N = 50Maternal 677TT genotypeassociated with lower DNAmin infant blood at candidateCpG sites in NR3C1, DRD4,5-HTT, IGF2DMR, H19,KCNQ1OT1, and MTHFR genes(p = 0.03)Weiner et al. (2014) [36] Genome-wide: Methyl FlashMethylated DNA Quantification KitBlood 677CC: N = 40677TT: N = 40677TT associated withsignificantly lower mean DNAmethylation compared to677CC (p = 0.0034)Llanos et al. (2015) [37] LINE-1: pyrosequencing Female breasttissue1298AA: N = 731298AC or 1298CC: N = 451298C allele associated withlower LINE-1 methylation(OR 0.96; 95% CI 0.93–0.98)Song et al. (2016) [101] Genome-wide: Illumina 450k array Female breasttissueStudy population: N = 81 677T and 1298C alleles associatedwith differential methylation at 5and 3 CpGs, respectively(unadjusted p value < 5.0 × 10−5).No sites reached significanceat an adjusted p value < 0.05Studies finding no association with DNAmNarayanan et al. (2004) [38] Genome-wide: radiolabeled methylgroup incorporation assayBlood 677CC: N = 90677CT: N = 84677TT: N = 251298AA: N = 931298AC: N = 771298CC: N = 29No altered DNAm in associationwith 677T or 1298C allelesJung et al. (2011) [102] Genome-wide: LC/MS Blood (Folic acid supplemented/placebo)677CC: N = 36/40677CT: N = 36/34677TT: N = 33/37No altered DNAm between the3-year folic acid-supplemented(0.8 mg/day) group and placebogroup and no difference inDNAm when stratified by theMTHFR 677 genotypeGomes et al. (2012) [39] Genome-wide: IMDQ kit Blood 677CC: N = 72677CT: N = 39677TT: N = 12No altered DNAm betweenMTHFR 677 genotype groupsOno et al. (2012) [103] Genome-wide: LUMA Blood 677CC: N = 112677CT or 677TT: N = 272No altered DNAm in associationwith MTHFR 677 or 1298 variantsDel Gobbo et al. Clinical Epigenetics  (2018) 10:34 Page 9 of 16Table 5 Literature assessing associations between MTHFR 677 or 1298 variants and altered DNAm in healthy tissues (Continued)Study Type of DNAm assessed: specific assay Tissue Study size† Results1298AA: N = 2541298AC or 1298CC: N = 130No interaction betweengenome-wide DNAm, folateintake, and MTHFR 677 or1298 variantsHanks et al. (2013) [104] Genome-wide: LC/MSCandidate sites (N = 7): pyrosequencingColon 677CC: N = 185677CT: N = 119677TT: N = 32No difference in DNAmbetween MTHFR 677 genotypegroups, even when accountingfor folate biomarkersNo significant difference inDNAm at ESR1, MYOD1, IGF2,N33, MLH1, MGMT, and APCgenes by the genotype groupde Arruda et al. (2013) [40] Genome-wide: IMDQ kit Oralepithelialcells677CC: N = 17677CT: N = 19677TT: N = 8No difference in DNAm betweenMTHFR 677 genotype groupsDeroo et al. (2014) [105] LINE-1: pyrosequencing Blood Study population: N = 646women without breastcancerN = 294 with breast cancer677 or 1298 genotypes notassociated with altered LINE-1DNAm in women withoutbreast cancerLouie et al. (2016) [106] Candidate sites (N = 3): bisulfitesequencingSperm 677CC: N = 21677CT: N = 19677TT: N = 4677 genotype not associatedwith altered DNAm at MEST,H19, or IG-GTL2 imprinteddifferentially methylated regionsWang et al. (2016) [41] Meta-analysis 11 studies 677: N = 11471298: N = 1053No altered DNAm associatedwith 677T and 1298C allelesStudies finding association with DNAm only with interaction with altered OCM nutrient statusFriso et al. (2002) [33] Genome-wide: LC/MS Blood 677CC: N = 187677TT: N = 105677TT associated withapproximately half the meanlevel of mCytosine than in the677CC group (p < 0.0001).This effect was driven by TTindividuals with low-folate statusShellnut et al. (2004) [107] Genome-wide: radiolabeled methylgroup incorporation assay and LC/MSBlood 677CC: N = 22677TT: N = 19No significant difference inDNAm between 677TT and677CC groups.In response to 7-week folaterepletion following 7-week folatedepletion, significantly increasedmean % change and raw changein DNAm in 677TT individuals(p= 0.04 and 0.03, respectively)Friso et al. (2005) [35] Genome-wide: LC/MS Blood 677CC/1298AA: N = 19677TT/1298AA: N = 72677CC/1298CC: N = 42In the presence of low folate,1298AA associated with lowergenome-wide DNAm comparedto 1298AC or 1298CC genotypes(p=0.0001 and p=0.021,respectively), and 677TT/1298AAassociated with lower DNAmcompared to 677CC/1298AA(p<0.05) and 677CC/1298CC(p<0.0001).In 677TT/1298AA individuals,DNAm significantly reduced inlow-folate vs high-folateindividuals (p < 0.0001)Axume et al. (2007) [108] Genome-wide: cytosine extension assay Blood 677CC: N = 14677CT: N = 12677TT: N = 17677TT associated with lowerDNAm compared to 677CC(p < 0.05) after 7-week folaterestriction followed by 7-weekfolate repletion treatmentLa Merrill et al. (2012) [109] Genome-wide: LUMA 677CC: N = 31677CT or 677TT: N = 164677T or 1298C alleles notassociated with alteredDel Gobbo et al. Clinical Epigenetics  (2018) 10:34 Page 10 of 16has not entirely reduced the incidence of NTDs infolate-replete populations [91, 92]; in Canada, NTDs arethe most common congenital abnormality [93]. Add-itionally, pathologies such as PE and IUGR are alsopresent at a high frequency in folate-replete populations,and associations between PE and MTHFR have beenobserved in such populations [27], indicating a mechan-ism for association with pathology beyond low-folate/folic acid status.In our study population, we found no significant associ-ation of NTDs, EOPE, LOPE, or nIUGR with high-risk677TT or 1298CC placental genotypes, although there wasa tendency to increased MTHFR 677TT in pregnancies af-fected by PE or IUGR as a whole (OR 2.53, p = 0.048). Thistrend is consistent with the literature noting an increasedrisk of PE in association with the 677T allele in both thematernal blood and the placenta [27, 29]. In a recent meta-analysis of 52 different studies, with a combined total of7398 PE cases and 11,230 controls, Wu et al. identified asignificantly increased risk of PE in association with theMTHFR 677T allele [28]. However, in 1103 cases and 988controls, no association between the MTHFR 1298C alleleand PE [28] was found. As for NTDs, our data was not sug-gestive of any association with fetal MTHFR 677TT or1298CC genotype. Sample size limited our power to detectsignificant differences between study groups; however, fewstudies have investigated associations between placental/fetal MTHFR variants and PE/IUGR and NTD pathologiesin the Canadian population post-folic acid fortification, andthe main focus of the current research was to assess alteredplacental DNAm in association with high-risk MTHFRgenotypes. Larger studies in NTDs have identifiedincreased risk in association with the 677 variant [25], butthere is inconsistent evidence for an association with the1298CC genotype [24, 94].Population stratification, the presence of systematicdifferences in allele frequencies between cases and con-trols, typically due to differences in ancestry, can be alimitation of genetic or epigenetic association studies.Specifically, false positive or negative results can be aconsequence of failing to match study groups on thisvariable. To address this, we utilized a novel approach toassess population stratification in our study groups usingthree continuous variables of ancestry based on a MDSanalysis of a panel of AIMs. This is similar to studiesusing MDS of genome-wide genotypes (i.e., from a SNParray or DNA sequencing) in study samples combinedwith ancestry reference populations to identify ancestryoutliers, select homogeneous groups, or infer ancestry[70, 71, 95], and to studies that include principal compo-nents or MDS coordinates highly associated with ances-try in statistical models to correct for ancestry [96, 97].Other potential confounding factors for our study, suchas maternal smoking status, diet, and medications takenduring pregnancy, were not well documented in all casesincluded in this study and thus could have resulted inheterogeneity between study groups that we were unableto account for in statistical modeling.Currently, the evidence supporting the relationshipbetween MTHFR 677 or 1298 variant and pathologyor altered DNAm is not conclusive enough for physi-cians to support implementing MTHFR genetic testingas a clinical practice [98]. Despite this, MTHFR geno-typing is available from 50 certified labs in the USA[98], and testing is widely promoted in the naturo-pathic field, where patients are told that a “faultygenotype” may explain a list of symptoms and dis-eases including “anxiousness, adrenal fatigue, brainfog, cervical dysplasia, increased risk of many cancers,low thyroid, leaky gut, high blood pressure, heartTable 5 Literature assessing associations between MTHFR 677 or 1298 variants and altered DNAm in healthy tissues (Continued)Study Type of DNAm assessed: specific assay Tissue Study size† ResultsBlood(pregnantwomen)1298AA: N = 1581298AC or 1298CC: N = 37genome-wide DNAm, butvitamin B6 deficiency andpresence of 677T alleleassociated with hypomethylation(p= 0.02)Aarabi et al. (2015) [110] Genome-wide: RRBSCandidate sites (N = 6): pyrosequencingSperm 677CC: N = 13677CT or 677TT: N = 17After 6 months of high-dosefolic acid supplementation,there was significant reductionin methylation in intergenicregions in 677CC men, whereas677CT or 677TT men hadsignificantly reducedmethylation in promoters, exons,introns, and intergenic regions(p< 0.05)LUMA luminometric methylation assay, LINE-1 LINE-1 repetitive elements, LC/MS liquid-chromatography tandem mass spectrometry, IMDQ imprint methylatedDNA quantification†Sample size given for each MTHFR SNP was assessed in publication. If the sample size of specific genotypes is not present, it was not reported in publication. Thecombined MTHFR 677/1298 genotypes are specified when available; otherwise, these SNPs were assessed independently from one another in the samesample populationDel Gobbo et al. Clinical Epigenetics  (2018) 10:34 Page 11 of 16attacks, stroke, Alzheimer’s disease, diabetes, and miscar-riages” (https://sciencebasedmedicine.org/dubious-mthfr-genetic-mutation-testing/). These patients are advised totake supplements containing “methyl folate” and “methylB12” to increase methylation and decrease their risk of dis-ease development (https://www.jillcarnahan.com/2013/05/12/mthfr-gene-mutation-whats-the-big-deal-about-methy-lation/). Our findings, coupled with variable results fromother studies, suggest that these variants may not be ofsuch strong concern in terms of DNAm, particularly inhealthy individuals meeting folate requirements; however,studies with larger sample sizes are required to validate this.At the very least, the negative results from our study sug-gest that if these variants have an effect on placental andthereby newborn health in Canada, it may not be throughaltered DNA methylation.ConclusionsDNA methylation (DNAm) alterations have been pro-posed to be the link between MTHFR 677C>T and1298A>C variants and increased risk of pregnancy compli-cations. In this novel study of DNAm in human placentasof high-risk MTHFR 677TT and 1298CC individuals, wedid not find evidence of altered DNAm associated withthese genotypes in numerous measures of genome-wideand CpG site-specific methylation. We conclude thatwidespread changes in DNAm do not occur in the placen-tas of MTHFR 677 and 1298 variant carriers in our folate-replete population. Further studies with larger samplesizes and/or in populations that are folate deficient maysupport or refute our results. The results from this studysuggest that factors other than alterations in DNAm maycontribute to the previously reported association betweenhigh-riskMTHFR genotypes and pathology.Additional filesAdditional file 1: Table S1. PCR and pyrosequencing conditions.(DOCX 22 kb)Additional file 2: Methods. (DOCX 20 kb)Additional file 3: Table S2. Global minor allele frequencies of 50 AIMSNPs used to assess ancestry. (DOCX 25 kb)Additional file 4: Figure S1. Distribution of ancestry coordinatesderived from MDS of 50 AIM SNP genotypes. (DOCX 166 kb)Additional file 5: Figure S2. Distribution of N = 277 study samplesalong 3 MDS ancestry coordinates. (DOCX 138 kb)Additional file 6: Table S3. MTHFR 677 and 1298 genotype counts andHardy-Weinberg equilibrium. (DOCX 20 kb)Additional file 7: Figure S3. Distribution of unadjusted p-values byCpG density between high-risk 677 or high-risk 1298 placentas comparedto reference placentas. (DOCX 230 kb)Abbreviations1kGP: 1000 Genomes Project; 450k array: Illumina Infinium HumanMethylation450BeadChip; AIMs: Ancestry-informative markers; CpG: Cytosine-guaninedinucleotide; DMR: Differentially methylated region; DNAm: DNA methylation;EOPE: Early-onset preeclampsia; FDR: False discovery rate; HWE: Hardy-Weinbergequilibrium; IUGR: Intrauterine growth restriction; LOPE: Late-onset preeclampsia;MDS: Multidimensional scaling; MTHFR: 5,10-Methylenetetrahydrofolate reductase;nIUGR: Normotensive intrauterine growth restriction; NTD: Neural tube defect;OCM: One-carbon metabolism; PE: Preeclampsia; rDNA: Ribosomal DNA;SNP: Single-nucleotide polymorphism; Δβ: Delta-betaAcknowledgementsWe thank all the participants of this study for kindly donating the samples.We also thank Dr. Deborah McFadden, Dr. Peter von Dadelszen, Dr. MargotVan Allen, Dr. Hayley Bos, Kristal Louie, Luana Avila, and Ruby Jiang for thesample acquisition and processing; Dr. Elodie Portales-Casamar for her advicein the ancestry analysis; Dr. Maria Peñaherrera for the array processing andadvice in the manuscript preparation; Samantha Wilson and Chaini Konwarfor their advice in the data analysis and manuscript preparation; and theKobor Lab for the generous use of their facilities.FundingThis work was supported by Canadian Institutes of Health Research grantsFRN106430 and FRN49520 (to WPR). WPR receives salary support through aninvestigatorship award from the BC Children’s Hospital Research Institute (BCCHRI).Availability of data and materialsThe 450k array dataset generated and analyzed in this current study hasbeen deposited in NCBI’s Gene Expression Omnibus [99] and is accessiblethrough GEO Series accession number GSE108567 (https://www.ncbi.nlm.nih.gov/geo/query/acc.cgi?acc=GSE108567). The remainingdata are available from the corresponding author upon reasonable request.Authors’ contributionsGFDG contributed to the study design, ran the pyrosequencing assays,performed the statistical analysis, interpreted the results, and wrote the draftof the manuscript. EMP contributed to the study design, ran the assays,performed the statistical analysis, interpreted the results, and contributed tothe draft of the manuscript. CWH contributed to the study design and ranthe assays and 450k arrays. WPR conceived of and supported the study andcontributed to the statistical analyses and interpretation. All authors read,critically revised, and approved the final manuscript.Ethics approval and consent to participateEthics approval for this study was obtained from the University of BritishColumbia/Children’s Hospital and Women’s Health Centre of British ColumbiaResearch Ethics Board (H04-70488, H10-01028). For all control, PE, nIUGR, andsome NTD cases (N= 261), mothers provided informed written consent toparticipate prior to delivery or termination of pregnancy. For certain NTD casesobtained retrospectively from pathological autopsy specimens (N= 42),biospecimens were de-identified and unlinked to clinical data. For all cases, onlynon-identifiable information is presented in this publication.Consent for publicationNot applicable.Competing interestsThe authors declare that they have no competing interests.Publisher’s NoteSpringer Nature remains neutral with regard to jurisdictional claims inpublished maps and institutional affiliations.Author details1BC Children’s Hospital Research Institute, 950 W 28th Ave, Vancouver, BCV5Z 4H4, Canada. 2Department of Medical Genetics, University of BritishColumbia, 4500 Oak St, Vancouver, BC V6H 3N1, Canada. 3EpigeneticsProgramme, Babraham Institute, Cambridge CB22 3AT, UK. 4Centre forTrophoblast Research, University of Cambridge, Cambridge CB2 3EG, UK.5Child and Family Research Institute, Room 2082, 950 W 28th Avenue,Vancouver, BC V5Z 4H4, Canada.Del Gobbo et al. Clinical Epigenetics  (2018) 10:34 Page 12 of 16Received: 21 December 2017 Accepted: 1 March 2018References1. Botto LD, Yang Q. 5,10-methylenetetrahydrofolate reductase genevariants and congenital anomalies: a HuGE review. Am J Epidemiol.2000;151(9):862–77.2. Kim YI. Methylenetetrahydrofolate reductase polymorphisms, folate, andcancer risk: a paradigm of gene-nutrient interactions in carcinogenesis. NutrRev. 2000;58(7):205–9.3. Klerk M, Verhoef P, Clarke R, Blom HJ, Kok FJ, Schouten EG, MTHFR StudiesCollaboration Group. MTHFR 677C–>T polymorphism and risk of coronaryheart disease: a meta-analysis. JAMA. 2002;288(16):2023–31.4. Kumar A, Kumar P, Prasad M, Sagar R, Yadav AK, Pandit AK, Jali VP, Pathak A.Association of C677T polymorphism in the methylenetetrahydrofolatereductase gene (MTHFR gene) with ischemic stroke: a meta-analysis. NeurolRes. 2015;37(7):568–77.5. Kang SS, Wong PW, Susmano A, Sora J, Norusis M, Ruggie N. Thermolabilemethylenetetrahydrofolate reductase: an inherited risk factor for coronaryartery disease. Am J Hum Genet. 1991;48(3):536–45.6. van der Put NM, Gabreels F, Stevens EM, Smeitink JA, Trijbels FJ, Eskes TK,van den Heuvel LP, Blom HJ. A second common mutation in themethylenetetrahydrofolate reductase gene: an additional risk factor forneural-tube defects? Am J Hum Genet. 1998;62(5):1044–51.7. Weisberg I, Tran P, Christensen B, Sibani S, Rozen R. A second geneticpolymorphism in methylenetetrahydrofolate reductase (MTHFR) associatedwith decreased enzyme activity. Mol Genet Metab. 1998;64(3):169–72.8. Weisberg IS, Jacques PF, Selhub J, Bostom AG, Chen Z, Curtis Ellison R,Eckfeldt JH, Rozen R. The 1298A–>C polymorphism inmethylenetetrahydrofolate reductase (MTHFR): in vitro expression andassociation with homocysteine. Atherosclerosis. 2001;156(2):409–15.9. Sherry ST, Ward MH, Kholodov M, Baker J, Phan L, Smigielski EM, Sirotkin K.dbSNP: the NCBI database of genetic variation. Nucleic Acids Res. 2001;29(1):308–11.10. Engbersen AM, Franken DG, Boers GH, Stevens EM, Trijbels FJ, Blom HJ.Thermolabile 5,10-methylenetetrahydrofolate reductase as a cause of mildhyperhomocysteinemia. Am J Hum Genet. 1995;56(1):142–50.11. Bagley PJ, Selhub J. A common mutation in themethylenetetrahydrofolate reductase gene is associated with anaccumulation of formylated tetrahydrofolates in red blood cells. ProcNatl Acad Sci U S A. 1998;95(22):13217–20.12. Friedman G, Goldschmidt N, Friedlander Y, Ben-Yehuda A, Selhub J, BabaeyS, Mendel M, Kidron M, Bar-On H. A common mutation A1298C in humanmethylenetetrahydrofolate reductase gene: association with plasma totalhomocysteine and folate concentrations. J Nutr. 1999;129(9):1656–61.13. Dekou V, Whincup P, Papacosta O, Ebrahim S, Lennon L, Ueland PM,Refsum H, Humphries SE, Gudnason V. The effect of the C677T and A1298Cpolymorphisms in the methylenetetrahydrofolate reductase gene onhomocysteine levels in elderly men and women from the British RegionalHeart Study. Atherosclerosis. 2001;154(3):659–66.14. Shelnutt KP, Kauwell GP, Chapman CM, Gregory JF 3rd, Maneval DR, BrowdyAA, Theriaque DW, Bailey LB. Folate status response to controlled folateintake is affected by the methylenetetrahydrofolate reductase 677C–>Tpolymorphism in young women. J Nutr. 2003;133(12):4107–11.15. Tanaka T, Scheet P, Giusti B, Bandinelli S, Piras MG, Usala G, Lai S, Mulas A,Corsi AM, Vestrini A, Sofi F, Gori AM, Abbate R, Guralnik J, Singleton A,Abecasis GR, Schlessinger D, Uda M, Ferrucci L. Genome-wide associationstudy of vitamin B6, vitamin B12, folate, and homocysteine bloodconcentrations. Am J Hum Genet. 2009;84(4):477–82.16. Clarke R, Bennett DA, Parish S, Verhoef P, Dotsch-Klerk M, Lathrop M, Xu P,Nordestgaard BG, Holm H, Hopewell JC, Saleheen D, Tanaka T, Anand SS,Chambers JC, Kleber ME, Ouwehand WH, Yamada Y, Elbers C, Peters B,Stewart AF, Reilly MM, Thorand B, Yusuf S, Engert JC, Assimes TL, Kooner J,Danesh J, Watkins H, Samani NJ, Collins R, Peto R, MTHFR StudiesCollaborative Group. Homocysteine and coronary heart disease: meta-analysis of MTHFR case-control studies, avoiding publication bias. PLoS Med.2012;9(2):e1001177.17. Kim SY, Park SY, Choi JW, Kim DJ, Lee SY, Lim JH, Han JY, Ryu HM, Kim MH.Association between MTHFR 1298A>C polymorphism and spontaneousabortion with fetal chromosomal aneuploidy. Am J Reprod Immunol. 2011;66(4):252–8.18. Behjati R, Modarressi MH, Jeddi-Tehrani M, Dokoohaki P, Ghasemi J, ZarnaniAH, Aarabi M, Memariani T, Ghaffari M, Akhondi MA. Thrombophilicmutations in Iranian patients with infertility and recurrent spontaneousabortion. Ann Hematol. 2006;85(4):268–71.19. Zetterberg H, Regland B, Palmer M, Ricksten A, Palmqvist L, Rymo L,Arvanitis DA, Spandidos DA, Blennow K. Increased frequency of combinedmethylenetetrahydrofolate reductase C677T and A1298C mutated alleles inspontaneously aborted embryos. Eur J Hum Genet. 2002;10(2):113–8.20. De Marco P, Calevo MG, Moroni A, Arata L, Merello E, Finnell RH, ZhuH, Andreussi L, Cama A, Capra V. Study of MTHFR and MSpolymorphisms as risk factors for NTD in the Italian population. J HumGenet. 2002;47(6):319–24.21. Parle-McDermott A, Mills JL, Kirke PN, O’Leary VB, Swanson DA, Pangilinan F,Conley M, Molloy AM, Cox C, Scott JM, Brody LC. Analysis of the MTHFR1298A–>C and 677C–>T polymorphisms as risk factors for neural tubedefects. J Hum Genet. 2003;48(4):190–3.22. Rampersaud E, Melvin EC, Siegel D, Mehltretter L, Dickerson ME, George TM,Enterline D, Nye JS, Speer MC, NTD Collaborative Group. Updatedinvestigations of the role of methylenetetrahydrofolate reductase in humanneural tube defects. Clin Genet. 2003;63(3):210–4.23. Kirke PN, Mills JL, Molloy AM, Brody LC, O’Leary VB, Daly L, Murray S, ConleyM, Mayne PD, Smith O, Scott JM. Impact of the MTHFR C677Tpolymorphism on risk of neural tube defects: case-control study. BMJ. 2004;328(7455):1535–6.24. Yadav U, Kumar P, Yadav SK, Mishra OP, Rai V. Polymorphisms in folatemetabolism genes as maternal risk factor for neural tube defects: anupdated meta-analysis. Metab Brain Dis. 2015;30(1):7–24.25. Yang Y, Chen J, Wang B, Ding C, Liu H. Association between MTHFR C677Tpolymorphism and neural tube defect risks: a comprehensive evaluation inthree groups of NTD patients, mothers, and fathers. Birth Defects Res A ClinMol Teratol. 2015;103(6):488–500.26. Sohda S, Arinami T, Hamada H, Yamada N, Hamaguchi H, Kubo T.Methylenetetrahydrofolate reductase polymorphism and pre-eclampsia.J Med Genet. 1997;34(6):525–6.27. Chedraui P, Salazar-Pousada D, Villao A, Escobar GS, Ramirez C, Hidalgo L,Perez-Lopez FR, Genazzani A, Simoncini T. Polymorphisms of themethylenetetrahydrofolate reductase gene (C677T and A1298C) innulliparous women complicated with preeclampsia. Gynecol Endocrinol.2014;30(5):392–6.28. Wu X, Yang K, Tang X, Sa Y, Zhou R, Liu J, Luo Y, Tang W. Folatemetabolism gene polymorphisms MTHFR C677T and A1298C and risk forpreeclampsia: a meta-analysis. J Assist Reprod Genet. 2015;32(5):797–805.29. Chedraui P, Andrade ME, Salazar-Pousada D, Escobar GS, Hidalgo L, RamirezC, Spaanderman ME, Kramer BW, Gavilanes AW. Polymorphisms of themethylenetetrahydrofolate reductase gene (C677T and A1298C) in theplacenta of pregnancies complicated with preeclampsia. GynecolEndocrinol. 2015;31(7):569–72.30. Blom HJ, Shaw GM, den Heijer M, Finnell RH. Neural tube defects and folate:case far from closed. Nat Rev Neurosci. 2006;7(9):724–31.31. McKay JA, Groom A, Potter C, Coneyworth LJ, Ford D, Mathers JC, Relton CL.Genetic and non-genetic influences during pregnancy on infant global andsite specific DNA methylation: role for folate gene variants and vitamin B(12).PLoS One. 2012;7(3):e33290. https://doi.org/10.1371/journal.pone.0033290.32. Stern LL, Mason JB, Selhub J, Choi SW. Genomic DNA hypomethylation, acharacteristic of most cancers, is present in peripheral leukocytes ofindividuals who are homozygous for the C677T polymorphism in themethylenetetrahydrofolate reductase gene. Cancer Epidemiol Biomark Prev.2000;9(8):849–53.33. Friso S, Choi SW, Girelli D, Mason JB, Dolnikowski GG, Bagley PJ, Olivieri O,Jacques PF, Rosenberg IH, Corrocher R, Selhub J. A common mutation inthe 5,10-methylenetetrahydrofolate reductase gene affects genomic DNAmethylation through an interaction with folate status. Proc Natl Acad Sci US A. 2002;99(8):5606–11.34. Castro R, Rivera I, Ravasco P, Camilo M, Jakobs C, Blom H, de Almeida IT.5,10-methylenetetrahydrofolate reductase (MTHFR) 677C–>T and 1298A–>Cmutations are associated with DNA hypomethylation. J Med Genet. 2004;41(6):454–8.35. Friso S, Girelli D, Trabetti E, Olivieri O, Guarini P, Pignatti PF, Corrocher R,Choi SW. The MTHFR 1298A>C polymorphism and genomic DNAmethylation in human lymphocytes. Cancer Epidemiol Biomark Prev. 2005;14(4):938–43.Del Gobbo et al. Clinical Epigenetics  (2018) 10:34 Page 13 of 1636. Weiner AS, Boyarskikh UA, Voronina EN, Mishukova OV, Filipenko ML.Methylenetetrahydrofolate reductase C677T and methionine synthaseA2756G polymorphisms influence on leukocyte genomic DNA methylationlevel. Gene. 2014;533(1):168–72.37. Llanos AA, Marian C, Brasky TM, Dumitrescu RG, Liu Z, Mason JB,Makambi KH, Spear SL, Kallakury BV, Freudenheim JL, Shields PG.Associations between genetic variation in one-carbon metabolism andLINE-1 DNA methylation in histologically normal breast tissues.Epigenetics. 2015;10(8):727–35.38. Narayanan S, McConnell J, Little J, Sharp L, Piyathilake CJ, Powers H, BastenG, Duthie SJ. Associations between two common variants C677T andA1298C in the methylenetetrahydrofolate reductase gene and measures offolate metabolism and DNA stability (strand breaks, misincorporated uracil,and DNA methylation status) in human lymphocytes in vivo. CancerEpidemiol Biomark Prev. 2004;13(9):1436–43.39. Gomes MVM, Toffoli LV, Arruda DW, Soldera LM, Pelosi GG, Neves-Souza RD, Freitas ER, Castro DT, Marquez AS. Age-related changes inthe global DNA methylation profile of leukocytes are linked tonutrition but are not associated with the MTHFR C677T genotype orto functional capacities. PLoS One. 2012;7(12):e52570. https://doi.org/10.1371/journal.pone.0052570.40. de Arruda ITS, Persuhn DC, de Oliveira NFP. The MTHFR C677Tpolymorphism and global DNA methylation in oral epithelial cells. GenetMol Biol. 2013;36(4):490–3.41. Wang L, Shangguan S, Chang S, Yu X, Wang Z, Lu X, Wu L, Zhang T.Determining the association between methylenetetrahydrofolatereductase (MTHFR) gene polymorphisms and genomic DNAmethylation level: a meta-analysis. Birth Defects Res A Clin MolTeratol. 2016;106(8):667–74.42. Antony AC. In utero physiology: role of folic acid in nutrient delivery andfetal development. Am J Clin Nutr. 2007;85(2):598S–603S.43. Solanky N, Requena Jimenez A, D’Souza SW, Sibley CP, Glazier JD.Expression of folate transporters in human placenta and implications forhomocysteine metabolism. Placenta. 2010;31(2):134–43.44. Laanpere M, Altmae S, Stavreus-Evers A, Nilsson TK, Yngve A, Salumets A.Folate-mediated one-carbon metabolism and its effect on female fertilityand pregnancy viability. Nutr Rev. 2010;68(2):99–113.45. Daly SF, Molloy AM, Mills JL, Lee YJ, Conley M, Kirke PN, Weir DG, Scott JM. Theinfluence of 5,10 methylenetetrahydrofolate reductase genotypes on enzymeactivity in placental tissue. Br J Obstet Gynaecol. 1999;106(11):1214–8.46. Frost JM, Moore GE. The importance of imprinting in the humanplacenta. PLoS Genet. 2010;6(7):e1001015. https://doi.org/10.1371/journal.pgen.1001015.47. Varmuza S, Miri K. What does genetics tell us about imprinting and theplacenta connection? Cell Mol Life Sci. 2015;72(1):51–72.48. Bourque DK, Avila L, Penaherrera M, von Dadelszen P, Robinson WP.Decreased placental methylation at the H19/IGF2 imprinting control regionis associated with normotensive intrauterine growth restriction but notpreeclampsia. Placenta. 2010;31(3):197–202.49. Yuen RK, Penaherrera MS, von Dadelszen P, McFadden DE, Robinson WP.DNA methylation profiling of human placentas reveals promoterhypomethylation of multiple genes in early-onset preeclampsia. Eur J HumGenet. 2010;18(9):1006–12.50. Gao WL, Li D, Xiao ZX, Liao QP, Yang HX, Li YX, Ji L, Wang YL. Detection ofglobal DNA methylation and paternally imprinted H19 gene methylation inpreeclamptic placentas. Hypertens Res. 2011;34(5):655–61.51. Wilson SL, Leavey K, Cox B, Robinson WP. Mining DNA methylationalterations towards a classification of placental pathologies. Hum Mol Genet.2018;27(1):135–46.52. Avila L, Yuen RK, Diego-Alvarez D, Penaherrera MS, Jiang R, Robinson WP.Evaluating DNA methylation and gene expression variability in the humanterm placenta. Placenta. 2010;31(12):1070–7.53. Penaherrera MS, Jiang R, Avila L, Yuen RK, Brown CJ, Robinson WP. Patternsof placental development evaluated by X chromosome inactivationprofiling provide a basis to evaluate the origin of epigenetic variation. HumReprod. 2012;27(6):1745–53.54. Novakovic B, Yuen RK, Gordon L, Penaherrera MS, Sharkey A, Moffett A,Craig JM, Robinson WP, Saffery R. Evidence for widespread changes inpromoter methylation profile in human placenta in response to increasinggestational age and environmental/stochastic factors. BMC Genomics. 2011;12:529. https://doi.org/10.1186/1471-2164-12-529.55. Miller SA, Dykes DD, Polesky HF. A simple salting out procedure forextracting DNA from human nucleated cells. Nucleic Acids Res. 1988;16(3):1215.56. Magee LA, Helewa M, Moutquin JM, von Dadelszen P, HypertensionGuideline Committee, Strategic Training Initiative in Research in theReproductive Health Sciences (STIRRHS) Scholars. Diagnosis, evaluation, andmanagement of the hypertensive disorders of pregnancy. J ObstetGynaecol Can. 2008;30(3 Suppl):S1–48.57. von Dadelszen P, Magee LA, Roberts JM. Subclassification of preeclampsia.Hypertens Pregnancy. 2003;22(2):143–8.58. Kramer MS, Platt RW, Wen SW, Joseph KS, Allen A, Abrahamowicz M,Blondel B, Breart G, Fetal/Infant Health Study Group of the CanadianPerinatal Surveillance System. A new and improved population-basedCanadian reference for birth weight for gestational age. Pediatrics. 2001;108(2):E35.59. Royo JL, Hidalgo M, Ruiz A. Pyrosequencing protocol using a universalbiotinylated primer for mutation detection and SNP genotyping. Nat Protoc.2007;2(7):1734–9.60. Price EM, Penaherrera MS, Portales-Casamar E, Pavlidis P, Van Allen MI,McFadden DE, Robinson WP. Profiling placental and fetal DNA methylationin human neural tube defects. Epigenetics Chromatin. 2016;9:6. https://doi.org/10.1186/s13072-016-0054-8. eCollection 201661. Pepe G, Camacho Vanegas O, Giusti B, Brunelli T, Marcucci R, AttanasioM, Rickards O, De Stefano GF, Prisco D, Gensini GF, Abbate R.Heterogeneity in world distribution of the thermolabile C677T mutationin 5,10-methylenetetrahydrofolate reductase. Am J Hum Genet. 1998;63(3):917–20.62. Wilcken B, Bamforth F, Li Z, Zhu H, Ritvanen A, Redlund M, Stoll C, AlembikY, Dott B, Czeizel A, Gelman-Kohan Z, Scarano G, Bianca S, Ettore G, TenconiR, Bellato S, Scala I, Mutchinick O, Lopez M, de Walle H, Hofstra R,Joutchenko L, Kavteladze L, Bermejo E, Martinez-Frias M, Gallagher M,Erickson J, Vollset S, Mastroiacovo P, Andria G, Botto L. Geographical andethnic variation of the 677C>T allele of 5,10 methylenetetrahydrofolatereductase (MTHFR): findings from over 7000 newborns from 16 areas worldwide. J Med Genet. 2003;40(8):619–25.63. Esfahani ST, Cogger EA, Caudill MA. Heterogeneity in the prevalence ofmethylenetetrahydrofolate reductase gene polymorphisms in women ofdifferent ethnic groups. J Am Diet Assoc. 2003;103(2):200–7.64. 1000 Genomes Project Consortium, Auton A, Brooks LD, Durbin RM,Garrison EP, Kang HM, Korbel JO, Marchini JL, McCarthy S, McVean GA,Abecasis GR. A global reference for human genetic variation. Nature. 2015;526(7571):68–74.65. Bryant AS, Worjoloh A, Caughey AB, Washington AE. Racial/ethnic disparitiesin obstetric outcomes and care: prevalence and determinants. Am J ObstetGynecol. 2010;202(4):335–43.66. Wang X, Fu J, Li Q, Zeng D. Geographical and ethnic distributions of theMTHFR C677T, A1298C and MTRR A66G gene polymorphisms in Chinesepopulations: a meta-analysis. PLoS One. 2016;11(4):e0152414. https://doi.org/10.1371/journal.pone.0152414.67. Phillips C, Salas A, Sanchez JJ, Fondevila M, Gomez-Tato A, Alvarez-Dios J,Calaza M, de Cal MC, Ballard D, Lareu MV, Carracedo A, SNPforID Consortium.Inferring ancestral origin using a single multiplex assay of ancestry-informativemarker SNPs. Forensic Sci Int Genet. 2007;1(3–4):273–80.68. Fondevila M, Phillips C, Santos C, Freire Aradas A, Vallone PM, Butler JM,Lareu MV, Carracedo A. Revision of the SNPforID 34-plex forensic ancestrytest: assay enhancements, standard reference sample genotypes andextended population studies. Forensic Sci Int Genet. 2013;7(1):63–74.69. Phillips C, Freire Aradas A, Kriegel AK, Fondevila M, Bulbul O, Santos C,Serrulla Rech F, Perez Carceles MD, Carracedo A, Schneider PM, Lareu MV.Eurasiaplex: a forensic SNP assay for differentiating European and SouthAsian ancestries. Forensic Sci Int Genet. 2013;7(3):359–66.70. Tardif JC, Rheaume E, Lemieux Perreault LP, Gregoire JC, Feroz Zada Y,Asselin G, Provost S, Barhdadi A, Rhainds D, L’Allier PL, Ibrahim R, UpmanyuR, Niesor EJ, Benghozi R, Suchankova G, Laghrissi-Thode F, Guertin MC,Olsson AG, Mongrain I, Schwartz GG, Dube MP. Pharmacogenomicdeterminants of the cardiovascular effects of dalcetrapib. Circ CardiovascGenet. 2015;8(2):372–82.71. Michailidou K, Hall P, Gonzalez-Neira A, Ghoussaini M, Dennis J, Milne RL,Schmidt MK, Chang-Claude J, Bojesen SE, Bolla MK, Wang Q, Dicks E, Lee A,Turnbull C, Rahman N, Breast and Ovarian Cancer SusceptibilityCollaboration, Fletcher O, Peto J, Gibson L, Dos Santos Silva I, Nevanlinna H,Del Gobbo et al. Clinical Epigenetics  (2018) 10:34 Page 14 of 16Muranen TA, Aittomaki K, Blomqvist C, Czene K, Irwanto A, Liu J, Waisfisz Q,Meijers-Heijboer H, Adank M, Hereditary Breast and Ovarian CancerResearch Group Netherlands (HEBON), van der Luijt RB, Hein R, Dahmen N,Beckman L, Meindl A, Schmutzler RK, Muller-Myhsok B, Lichtner P, HopperJL, Southey MC, Makalic E, Schmidt DF, Uitterlinden AG, Hofman A, HunterDJ, Chanock SJ, Vincent D, Bacot F, Tessier DC, Canisius S, Wessels LF,Haiman CA, Shah M, Luben R, Brown J, Luccarini C, Schoof N, Humphreys K,Li J, Nordestgaard BG, Nielsen SF, Flyger H, Couch FJ, Wang X, Vachon C,Stevens KN, Lambrechts D, Moisse M, Paridaens R, Christiaens MR, RudolphA, Nickels S, Flesch-Janys D, Johnson N, Aitken Z, Aaltonen K, Heikkinen T,Broeks A, Veer LJ, van der Schoot CE, Guenel P, Truong T, Laurent-Puig P,Menegaux F, Marme F, Schneeweiss A, Sohn C, Burwinkel B, Zamora MP,Perez JI, Pita G, Alonso MR, Cox A, Brock IW, Cross SS, Reed MW, Sawyer EJ,Tomlinson I, Kerin MJ, Miller N, Henderson BE, Schumacher F, Le MarchandL, Andrulis IL, Knight JA, Glendon G, Mulligan AM, kConFab Investigators,Australian Ovarian Cancer Study Group, Lindblom A, Margolin S, HooningMJ, Hollestelle A, van den Ouweland AM, Jager A, Bui QM, Stone J, Dite GS,Apicella C, Tsimiklis H, Giles GG, Severi G, Baglietto L, Fasching PA, HaeberleL, Ekici AB, Beckmann MW, Brenner H, Muller H, Arndt V, Stegmaier C,Swerdlow A, Ashworth A, Orr N, Jones M, Figueroa J, Lissowska J, Brinton L,Goldberg MS, Labreche F, Dumont M, Winqvist R, Pylkas K, Jukkola-VuorinenA, Grip M, Brauch H, Hamann U, Bruning T, GENICA (Gene EnvironmentInteraction and Breast Cancer in Germany) Network, Radice P, Peterlongo P,Manoukian S, Bonanni B, Devilee P, Tollenaar RA, Seynaeve C, van AsperenCJ, Jakubowska A, Lubinski J, Jaworska K, Durda K, Mannermaa A, Kataja V,Kosma VM, Hartikainen JM, Bogdanova NV, Antonenkova NN, Dork T,Kristensen VN, Anton-Culver H, Slager S, Toland AE, Edge S, Fostira F, KangD, Yoo KY, Noh DY, Matsuo K, Ito H, Iwata H, Sueta A, Wu AH, Tseng CC,Van Den Berg D, Stram DO, Shu XO, Lu W, Gao YT, Cai H, Teo SH, Yip CH,Phuah SY, Cornes BK, Hartman M, Miao H, Lim WY, Sng JH, Muir K,Lophatananon A, Stewart-Brown S, Siriwanarangsan P, Shen CY, Hsiung CN,Wu PE, Ding SL, Sangrajrang S, Gaborieau V, Brennan P, McKay J, Blot WJ,Signorello LB, Cai Q, Zheng W, Deming-Halverson S, Shrubsole M, Long J,Simard J, Garcia-Closas M, Pharoah PD, Chenevix-Trench G, Dunning AM,Benitez J, Easton DF. Large-scale genotyping identifies 41 new lociassociated with breast cancer risk. Nat Genet. 2013;45(4):353-61, 361e1-2.72. Sandoval J, Heyn H, Moran S, Serra-Musach J, Pujana MA, Bibikova M,Esteller M. Validation of a DNA methylation microarray for 450,000 CpG sitesin the human genome. Epigenetics. 2011;6(6):692–702.73. Price EM, Robinson WP. Adjusting for batch effects in DNA methylationmicroarray data, a lesson learned. Front Genet. 2018;9:83.74. Price EM, Cotton AM, Penaherrera MS, McFadden DE, Kobor MS,Robinson W. Different measures of “genome-wide” DNA methylationexhibit unique properties in placental and somatic tissues. Epigenetics.2012;7(6):652–63.75. Powell MA, Mutch DG, Rader JS, Herzog TJ, Huang TH, Goodfellow PJ.Ribosomal DNA methylation in patients with endometrial carcinoma: anindependent prognostic marker. Cancer. 2002;94(11):2941–52.76. Rusiecki JA, Baccarelli A, Bollati V, Tarantini L, Moore LE, Bonefeld-Jorgensen EC.Global DNA hypomethylation is associated with high serum-persistent organicpollutants in Greenlandic Inuit. Environ Health Perspect. 2008;116(11):1547–52.77. Baccarelli A, Wright R, Bollati V, Litonjua A, Zanobetti A, Tarantini L, SparrowD, Vokonas P, Schwartz J. Ischemic heart disease and stroke in relation toblood DNA methylation. Epidemiology. 2010;21(6):819–28.78. Zhu ZZ, Hou L, Bollati V, Tarantini L, Marinelli B, Cantone L, Yang AS,Vokonas P, Lissowska J, Fustinoni S, Pesatori AC, Bonzini M, Apostoli P, CostaG, Bertazzi PA, Chow WH, Schwartz J, Baccarelli A. Predictors of globalmethylation levels in blood DNA of healthy subjects: a combined analysis.Int J Epidemiol. 2012;41(1):126–39.79. Teschler S, Gotthardt J, Dammann G, Dammann RH. Aberrant DNAmethylation of rDNA and PRIMA1 in borderline personality disorder. Int JMol Sci. 2016;17(1) https://doi.org/10.3390/ijms17010067.80. Bollati V, Baccarelli A, Hou L, Bonzini M, Fustinoni S, Cavallo D, Byun HM,Jiang J, Marinelli B, Pesatori AC, Bertazzi PA, Yang AS. Changes in DNAmethylation patterns in subjects exposed to low-dose benzene. Cancer Res.2007;67(3):876–80.81. Qiao Y, Mondal K, Trapani V, Wen J, Carpenter G, Wildin R, Price EM,Gibbons RJ, Eichmeyer J, Jiang R, DuPont B, Martell S, Lewis SM, RobinsonWP, O’Driscoll M, Wolf FI, Zwick ME, Rajcan-Separovic E. Variant ATRXsyndrome with dysfunction of ATRX and MAGT1 genes. Hum Mutat. 2014;35(1):58–62.82. R Core Team: R: a language and environment for statistical computing. 201683. Price ME, Cotton AM, Lam LL, Farre P, Emberly E, Brown CJ, RobinsonWP, Kobor MS. Additional annotation enhances potential forbiologically-relevant analysis of the Illumina InfiniumHumanMethylation450 BeadChip array. Epigenetics Chromatin. 2013;6(1):4. https://doi.org/10.1186/1756-8935-6-4.84. Hanna CW, McFadden DE, Robinson WP. DNA methylation profiling ofplacental villi from karyotypically normal miscarriage and recurrentmiscarriage. Am J Pathol. 2013;182(6):2276–84.85. O’Neill RJ, Vrana PB, Rosenfeld CS. Maternal methyl supplemented diets andeffects on offspring health. Front Genet. 2014;5:289.86. Yang AS, Estecio MR, Doshi K, Kondo Y, Tajara EH, Issa JP. A simple methodfor estimating global DNA methylation using bisulfite PCR of repetitive DNAelements. Nucleic Acids Res. 2004;32(3):e38.87. Guenther BD, Sheppard CA, Tran P, Rozen R, Matthews RG, Ludwig ML. Thestructure and properties of methylenetetrahydrofolate reductase fromEscherichia coli suggest how folate ameliorates humanhyperhomocysteinemia. Nat Struct Biol. 1999;6(4):359–65.88. de Bree A, Verschuren WM, Bjorke-Monsen AL, van der Put NM, Heil SG,Trijbels FJ, Blom HJ. Effect of the methylenetetrahydrofolate reductase677C–>T mutation on the relations among folate intake and plasma folateand homocysteine concentrations in a general population sample. Am JClin Nutr. 2003;77(3):687–93.89. Hustad S, Midttun Ã, Schneede J, Vollset SE, Grotmol T, Ueland PM. Themethylenetetrahydrofolate reductase 677C–>T polymorphism as amodulator of a B vitamin network with major effects on homocysteinemetabolism. Am J Hum Genet. 2007;80(5):846–55.90. Plumptre L, Masih SP, Ly A, Aufreiter S, Sohn KJ, Croxford R, Lausman AY,Berger H, O’Connor DL, Kim YI. High concentrations of folate andunmetabolized folic acid in a cohort of pregnant Canadian women andumbilical cord blood. Am J Clin Nutr. 2015;102(4):848–57.91. De Wals P, Tairou F, Van Allen MI, Uh SH, Lowry RB, Sibbald B, Evans JA, Vanden Hof MC, Zimmer P, Crowley M, Fernandez B, Lee NS, Niyonsenga T.Reduction in neural-tube defects after folic acid fortification in Canada. NEngl J Med. 2007;357(2):135–42.92. Williams J, Mai CT, Mulinare J, Isenburg J, Flood TJ, Ethen M, Frohnert B,Kirby RS, Centers for Disease Control and Prevention. Updated estimates ofneural tube defects prevented by mandatory folic acid fortification—UnitedStates, 1995-2011. MMWR Morb Mortal Wkly Rep. 2015;64(1):1–5.93. Irvine B, Luo W, Leon JA. Congenital anomalies in Canada 2013: a perinatalhealth surveillance report by the Public Health Agency of Canada’sCanadian Perinatal Surveillance System. Health Promot Chronic Dis PrevCan. 2015;35(1):21–2.94. Wang XW, Luo YL, Wang W, Zhang Y, Chen Q, Cheng YL. Associationbetween MTHFR A1298C polymorphism and neural tube defectsusceptibility: a metaanalysis. Am J Obstet Gynecol. 2012;206(3):251.e1–7.95. Oskoui M, Gazzellone MJ, Thiruvahindrapuram B, Zarrei M, Andersen J, WeiJ, Wang Z, Wintle RF, Marshall CR, Cohn RD, Weksberg R, Stavropoulos DJ,Fehlings D, Shevell MI, Scherer SW. Clinically relevant copy numbervariations detected in cerebral palsy. Nat Commun. 2015;6 https://doi.org/10.1038/ncomms8949.96. Price AL, Patterson NJ, Plenge RM, Weinblatt ME, Shadick NA, Reich D.Principal components analysis corrects for stratification in genome-wideassociation studies. Nat Genet. 2006;38(8):904–9.97. Hibar DP, Stein JL, Renteria ME, Arias-Vasquez A, Desrivieres S, Jahanshad N,Toro R, Wittfeld K, Abramovic L, Andersson M, Aribisala BS, Armstrong NJ,Bernard M, Bohlken MM, Boks MP, Bralten J, Brown AA, Chakravarty MM,Chen Q, Ching CR, Cuellar-Partida G, den Braber A, Giddaluru S, GoldmanAL, Grimm O, Guadalupe T, Hass J, Woldehawariat G, Holmes AJ, HoogmanM, Janowitz D, Jia T, Kim S, Klein M, Kraemer B, Lee PH, Olde Loohuis LM,Luciano M, Macare C, Mather KA, Mattheisen M, Milaneschi Y, Nho K,Papmeyer M, Ramasamy A, Risacher SL, Roiz-Santianez R, Rose EJ, Salami A,Samann PG, Schmaal L, Schork AJ, Shin J, Strike LT, Teumer A, vanDonkelaar MM, van Eijk KR, Walters RK, Westlye LT, Whelan CD, Winkler AM,Zwiers MP, Alhusaini S, Athanasiu L, Ehrlich S, Hakobjan MM, Hartberg CB,Haukvik UK, Heister AJ, Hoehn D, Kasperaviciute D, Liewald DC, Lopez LM,Makkinje RR, Matarin M, Naber MA, McKay DR, Needham M, Nugent AC,Putz B, Royle NA, Shen L, Sprooten E, Trabzuni D, van der Marel SS, vanHulzen KJ, Walton E, Wolf C, Almasy L, Ames D, Arepalli S, Assareh AA,Bastin ME, Brodaty H, Bulayeva KB, Carless MA, Cichon S, Corvin A, CurranJE, Czisch M, de Zubicaray GI, Dillman A, Duggirala R, Dyer TD, Erk S, FedkoDel Gobbo et al. Clinical Epigenetics  (2018) 10:34 Page 15 of 16IO, Ferrucci L, Foroud TM, Fox PT, Fukunaga M, Gibbs JR, Goring HH, GreenRC, Guelfi S, Hansell NK, Hartman CA, Hegenscheid K, Heinz A, HernandezDG, Heslenfeld DJ, Hoekstra PJ, Holsboer F, Homuth G, Hottenga JJ, IkedaM, Jack CR Jr, Jenkinson M, Johnson R, Kanai R, Keil M, Kent JW, Jr KP, KwokJB, Lawrie SM, Liu X, Longo DL, McMahon KL, Meisenzahl E, Melle I, MohnkeS, Montgomery GW, Mostert JC, Muhleisen TW, Nalls MA, Nichols TE, NilssonLG, Nothen MM, Ohi K, Olvera RL, Perez-Iglesias R, Pike GB, Potkin SG,Reinvang I, Reppermund S, Rietschel M, Romanczuk-Seiferth N, Rosen GD,Rujescu D, Schnell K, Schofield PR, Smith C, Steen VM, Sussmann JE,Thalamuthu A, Toga AW, Traynor BJ, Troncoso J, Turner JA, ValdesHernandez MC, van’t Ent D, van der Brug M, van der Wee NJ, van Tol MJ,Veltman DJ, Wassink TH, Westman E, Zielke RH, Zonderman AB, AshbrookDG, Hager R, Lu L, FJ MM, Morris DW, Williams RW, Brunner HG, Buckner RL,Buitelaar JK, Cahn W, Calhoun VD, Cavalleri GL, Crespo-Facorro B, Dale AM,Davies GE, Delanty N, Depondt C, Djurovic S, Drevets WC, Espeseth T,Gollub RL, Ho BC, Hoffmann W, Hosten N, Kahn RS, Le Hellard S, Meyer-Lindenberg A, Muller-Myhsok B, Nauck M, Nyberg L, Pandolfo M, PenninxBW, Roffman JL, Sisodiya SM, Smoller JW, van Bokhoven H, van Haren NE,Volzke H, Walter H, Weiner MW, Wen W, White T, Agartz I, Andreassen OA,Blangero J, Boomsma DI, Brouwer RM, Cannon DM, Cookson MR, de GeusEJ, Deary IJ, Donohoe G, Fernandez G, Fisher SE, Francks C, Glahn DC, GrabeHJ, Gruber O, Hardy J, Hashimoto R, Hulshoff Pol HE, Jonsson EG,Kloszewska I, Lovestone S, Mattay VS, Mecocci P, McDonald C, McIntosh AM,Ophoff RA, Paus T, Pausova Z, Ryten M, Sachdev PS, Saykin AJ, Simmons A,Singleton A, Soininen H, Wardlaw JM, Weale ME, Weinberger DR, AdamsHH, Launer LJ, Seiler S, Schmidt R, Chauhan G, Satizabal CL, Becker JT, YanekL, van der Lee SJ, Ebling M, Fischl B, Longstreth WT, Jr GD, Schmidt H,Nyquist P, Vinke LN, van Duijn CM, Xue L, Mazoyer B, Bis JC, Gudnason V,Seshadri S, Ikram MA, Alzheimer’s Disease Neuroimaging Initiative, CHARGEConsortium, EPIGEN, IMAGEN, SYS, Martin NG, Wright MJ, Schumann G,Franke B, Thompson PM, Medland SE. Common genetic variants influencehuman subcortical brain structures. Nature. 2015;520(7546):224–9.98. Hickey SE, Curry CJ, Toriello HV. ACMG practice guideline: lack of evidencefor MTHFR polymorphism testing. Genet Med. 2013;15(2):153–6.99. Edgar R, Domrachev M, Lash AE. Gene Expression Omnibus: NCBI geneexpression and hybridization array data repository. Nucleic Acids Res. 2002;30(1):207–10.100. van Mil NH, Bouwland-Both MI, Stolk L, Verbiest MM, Hofman A, Jaddoe VW,Verhulst FC, Eilers PH, Uitterlinden AG, Steegers EA, Tiemeier H, Steegers-Theunissen RP. Determinants of maternal pregnancy one-carbonmetabolism and newborn human DNA methylation profiles. Reproduction.2014;148(6):581–92.101. Song MA, Brasky TM, Marian C, Weng DY, Taslim C, Llanos AA, DumitrescuRG, Liu Z, Mason JB, Spear SL, Kallakury BV, Freudenheim JL, Shields PG.Genetic variation in one-carbon metabolism in relation to genome-wideDNA methylation in breast tissue from heathy women. Carcinogenesis.2016;37(5):471–80.102. Jung AY, Smulders Y, Verhoef P, Kok FJ, Blom H, Kok RM, Kampman E,Durga J. No effect of folic acid supplementation on global DNA methylationin men and women with moderately elevated homocysteine. PLoS One.2011;6(9):e24976.103. Ono H, Iwasaki M, Kuchiba A, Kasuga Y, Yokoyama S, Onuma H, NishimuraH, Kusama R, Ohnami S, Sakamoto H, Yoshida T, Tsugane S. Association ofdietary and genetic factors related to one-carbon metabolism with globalmethylation level of leukocyte DNA. Cancer Sci. 2012;103(12):2159–64.104. Hanks J, Ayed I, Kukreja N, Rogers C, Harris J, Gheorghiu A, Liu CL, Emery P,Pufulete M. The association between MTHFR 677C>T genotype and folatestatus and genomic and gene-specific DNA methylation in the colon ofindividuals without colorectal neoplasia. Am J Clin Nutr. 2013;98(6):1564–74.105. Deroo LA, Bolick SC, Xu Z, Umbach DM, Shore D, Weinberg CR, Sandler DP,Taylor JA. Global DNA methylation and one-carbon metabolism genepolymorphisms and the risk of breast cancer in the Sister Study.Carcinogenesis. 2014;35(2):333–8.106. Louie K, Minor A, Ng R, Poon K, Chow V, Ma S. Evaluation of DNA methylationat imprinted DMRs in the spermatozoa of oligozoospermic men in associationwith MTHFR C677T genotype. Andrology. 2016;4(5):825–31.107. Shelnutt KP, Kauwell GP, Gregory JF 3rd, Maneval DR, Quinlivan EP, TheriaqueDW, Henderson GN, Bailey LB. Methylenetetrahydrofolate reductase 677C–>Tpolymorphism affects DNA methylation in response to controlled folate intakein young women. J Nutr Biochem. 2004;15(9):554–60.108. Axume J, Smith SS, Pogribny IP, Moriarty DJ, Caudill MA. The MTHFR 677TTgenotype and folate intake interact to lower global leukocyte DNA methylationin young Mexican American women. Nutr Res. 2007;27(1):1365–17.109. La Merrill M, Torres-Sanchez L, Ruiz-Ramos R, Lopez-Carrillo L, Cebrian ME,Chen J. The association between first trimester micronutrient intake, MTHFRgenotypes, and global DNA methylation in pregnant women. J MaternFetal Neonatal Med. 2012;25(2):133–7.110. Aarabi M, San Gabriel MC, Chan D, Behan NA, Caron M, Pastinen T, BourqueG, MacFarlane AJ, Zini A, Trasler J. High-dose folic acid supplementationalters the human sperm methylome and is influenced by the MTHFR C677Tpolymorphism. Hum Mol Genet. 2015;24(22):6301–13.•  We accept pre-submission inquiries •  Our selector tool helps you to find the most relevant journal•  We provide round the clock customer support •  Convenient online submission•  Thorough peer review•  Inclusion in PubMed and all major indexing services •  Maximum visibility for your researchSubmit your manuscript atwww.biomedcentral.com/submitSubmit your next manuscript to BioMed Central and we will help you at every step:Del Gobbo et al. Clinical Epigenetics  (2018) 10:34 Page 16 of 16


Citation Scheme:


Citations by CSL (citeproc-js)

Usage Statistics



Customize your widget with the following options, then copy and paste the code below into the HTML of your page to embed this item in your website.
                            <div id="ubcOpenCollectionsWidgetDisplay">
                            <script id="ubcOpenCollectionsWidget"
                            async >
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:


Related Items