UBC Faculty Research and Publications

Placental DNA methylation at term reflects maternal serum levels of INHA and FN1, but not PAPPA, early… Wilson, Samantha L; Blair, John D; Hogg, Kirsten; Langlois, Sylvie; von Dadelszen, Peter; Robinson, Wendy P Dec 11, 2015

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

Item Metadata

Download

Media
52383-12881_2015_Article_257.pdf [ 1.33MB ]
Metadata
JSON: 52383-1.0361954.json
JSON-LD: 52383-1.0361954-ld.json
RDF/XML (Pretty): 52383-1.0361954-rdf.xml
RDF/JSON: 52383-1.0361954-rdf.json
Turtle: 52383-1.0361954-turtle.txt
N-Triples: 52383-1.0361954-rdf-ntriples.txt
Original Record: 52383-1.0361954-source.json
Full Text
52383-1.0361954-fulltext.txt
Citation
52383-1.0361954.ris

Full Text

RESEARCH ARTICLE Open AccessPlacental DNA methylation at term reflectsmaternal serum levels of INHA and FN1,but not PAPPA, early in pregnancySamantha L. Wilson1,2, John D. Blair1,2,3, Kirsten Hogg1,4, Sylvie Langlois1,2, Peter von Dadelszen1,5and Wendy P. Robinson1,2*AbstractBackground: Early detection of pregnancies at risk of complications, such as intrauterine growth restriction (IUGR)and preeclampsia (PE), is critical for improved monitoring and preventative treatment to optimize health outcomes.We predict that levels of placental-derived proteins circulating in maternal blood reflect placental gene expression,which is associated with placental DNA methylation (DNAm) profiles. As such, placental DNAm profiling may be usefulto distinguish pregnancies at risk of developing complications and correlation between DNAm and protein levels inmaternal blood may give further evidence for a protein’s use as a biomarker. However, few studies investigate all clinicalparameters that may influence DNAm and/or protein expression, which can significantly affect the relationship betweenthese measures.Results: Candidate genes were chosen based on i) reported alterations of protein levels in maternal bloodand ii) observed changes in placental DNAm (Δβ > 0.05 and False Discovery Rate (FDR) <0.05) in pregnanciescomplicated by PE/IUGR. Fibronectin (FN1) enhancer DNAm and placental gene expression were inverselycorrelated (r = −0.88 p < 0.01). The same trend was observed between promoter DNAm and gene expression forINHBA and PAPPA, though not significant. INHBA and FN1 DNAm was associated with gestational–age corrected birthweight, while INHA levels were associated with fetal: placental weight ratio and FN1 level was associated with maternalbody mass index (BMI).DNAm at the INHBA promoter in the term placenta was negatively correlated with second trimester maternal serumlevels (r = −0.50 p = 0.01) and DNAm at the FN1 enhancer was negatively associated with third trimester maternalserum levels (r = −0.38, p = 0.009). However, a similar correlation was not found for PAPPA.Conclusions: These results show that establishing a correlation between altered DNAm in the term placenta andaltered maternal serum levels of the corresponding protein, is affected by a number of factors. Nonetheless, thecorrelation between placental DNAm of INHBA/FN1 and maternal serum INHA/FN1 levels indicate that DNAm maybe a useful tool to identify novel biomarkers for adverse pregnancy outcomes in some cases.Keywords: Preeclampsia, Intrauterine growth restriction, DNA methylation, Maternal serum screening, Placenta* Correspondence: wrobinson@cfri.ca1Child & Family Research Institute, 950 W 28th Ave, Vancouver, BC V5Z 4H4,Canada2Department of Medical Genetics, University of British Columbia, C201-4500Oak St, Vancouver, BC V6H3N1, CanadaFull list of author information is available at the end of the article© 2015 Wilson et al. 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.Wilson et al. BMC Medical Genetics  (2015) 16:111 DOI 10.1186/s12881-015-0257-zBackgroundPlacental insufficiency is the inability of the placenta toprovide an adequate supply of nutrients to the growingfetus. This can lead to a number of pregnancy complica-tions including intrauterine growth restriction (IUGR) [1]and preeclampsia (PE), a maternal hypertensive disorder,which manifests as maternal hypertension and proteinuriaafter 20 weeks (wks) gestation [2]. Early diagnosis of PEand IUGR before clinical signs of disease can improvemanagement and outcomes of affected pregnancies.Placental-derived proteins may be released into the mater-nal circulation where they can be quantified and used toassess placental function during pregnancy [3–6]. Suchprotein markers have been investigated for the predictionof PE and/or IUGR with varying success [7–9]. Nicolaideset al. (2013) reported a detection rate of 95 % for early-onset PE (EOPE, diagnosis <34 wks) using decreased levelsof maternal serum markers, pregnancy associated plasmaprotein A (PAPPA) and placental growth factor (PlGF), incombination with maternal factors [7]. However, thesemeasures might not be generalizable, as the etiology andconfounding environmental factors vary between popula-tions [8]. Moreover, the ability to predict women at risk oflate-onset PE (LOPE, diagnosis >34 wks) and IUGR islimited using these markers.Differential gene expression between placentas from PEand/or IUGR pregnancies [9–12] may be utilized to iden-tify additional biomarkers to distinguish women at highrisk of these complications early in gestation. DNA methy-lation (DNAm) is associated with gene expression, but ismore robust to variation in technical conditions and lesssubject to short-term biological change [13]. We previ-ously reported numerous changes in DNAm in placentasfrom pregnancies complicated by EOPE [14]. Alterationsof placental DNAm were noted in genes for which theexpression of the encoded protein is altered in maternalblood in PE and/or IUGR pregnancies (e.g.: PAPPA, sENG,PAPPA2) [14]. Furthermore we found that sites of alteredDNAm in PE frequently reflected changes in gene expres-sion. While proteins produced in the placenta can bereleased into maternal circulation, their levels in maternalserum may be affected by many additional factors includ-ing size of the placenta, the cell type expressing the protein,and how such proteins are transported and metabolized.The purpose of the present study was to delineate the rela-tionship between changes we observed in DNAm at termand maternal protein levels in early pregnancy. Weselected three genes for which there was evidence for bothaltered maternal protein levels and altered DNAm in PE;we then evaluated 1) the relationship between placentalDNAm and gene expression; 2) the role of variables thatmight confound measurement of DNAm, mRNA or pro-tein levels including gestational age, fetal sex, placental effi-ciency (fetal: placental weight ratio), fetal birth weight,placental breadth: width ratio and maternal body massindex (BMI); and 3) whether placental DNAm at termreflected protein levels in maternal blood during gestationafter correcting for these variables.Results and discussionCandidate site selection and characteristicsTo isolate loci for which altered DNAm might reflectmaternal serum levels early in pregnancy, we chosecandidate genes that not only had sites showing alteredDNAm in EOPE, but also encode for proteins previouslyreported to show altered maternal serum protein levelsin pregnancies that subsequently developed PE and/orIUGR. Previous studies have shown upregulation of bothPAPPA and INBHA in the placentas of pregnanciescomplicated by PE and IUGR [12, 14–16]. In addition,several studies have reported DNAm alterations inplacentas from pregnancies complicated by PE and/orIUGR [14, 17, 18]. FN1 [19] was selected due to thelarge magnitude of change in DNAm between EOPEand control placentas (Δβ = −0.24, FDR < 0.05) (seemethods). INHBA (Δβ = −0.16, FDR < 0.05) and PAPPA(Δβ = −0.074, FDR < 0.05) were selected because theyadditionally encode for proteins for which first (PAPPA)or second trimester (INHA) maternal serum measureswere available from clinical prenatal serum screeningtesting. We also focused on DNAm alterations in generegulatory elements. The CpGs of interest for INHBAand PAPPA were 76 base pairs (bp) and 163 bp up-stream of the transcriptional start sites, respectively. Inrelation to FN1, the CpG site was ~100 kb upstream ofthe transcriptional start site, within an enhancer region.Although these sites were selected based on a significantassociation with EOPE, we also wanted to know if thesechanges were conserved in other clinical groups (Fig. 1). Inaddition to hypomethylation of these sites in EOPE, theLOPE + IUGR group was hypomethylated for the INHBA(promoter) (Δβ = −0.18, p < 0.001) (Fig. 1a) and the FN1upstream enhancer (Δβ = −0.25, p < 0.01) (Fig. 1c). Whilereduced methylation at the PAPPA promoter was onlyfound in EOPE (Fig. 1b). As differences in DNAm wereonly found in the EOPE and LOPE + IUGR groups, as po-tential biomarkers, these candidate genes would presum-ably only be useful in identifying this subset of pregnancies[20]. Markers useful to detect LOPE or normotensiveIUGR may be more challenging to identify due to theirweak association with placental pathology.Is DNAm at candidate sites inversely correlated with geneexpression?To confirm that the DNAm change resulted in a changein gene expression, we assessed the relationship betweenplacental DNAm (measured by Illumina 450 k array)and gene expression at these three candidate sites. FN1Wilson et al. BMC Medical Genetics  (2015) 16:111 Page 2 of 10showed an inverse correlation between DNAm of an up-stream enhancer and gene expression at term (r = −0.88,p < 0.0001). INHBA and PAPPA, showed a non-significanttrend with increasing DNAm being associated with de-creased gene expression in the placenta (Fig. 2). Thisphenomenon may be due to alterations in cellcomposition between pathogenic and healthy placen-tas related to the pathology of PE/IUGR [14]. For allcandidate genes, there was an observable divide be-tween the controls and EOPE cases, where cases haddecreased DNAm corresponding to increased geneexpression in the placenta.What clinical factors are associated with DNAm atcandidate sites?To better understand what factors might affect the meas-urement of DNAm and therefore the relationship withprotein expression levels in maternal blood, we also evalu-ated several potential confounding factors including gesta-tional age at delivery [21], fetal sex [22, 23], fetal birthweight [24], placental dimensions and maternal BMI. Bi-sulfite pyrosequencing was used to extend our assess-ment of DNAm at the candidate sites into a largercohort of controls for which clinical serum measure-ments (INHBA N = 36, PAPPA N = 33) or serum sam-ples for assaying FN1 (N = 76) were available.Birth-weight standard deviation (SD) was associatedwith DNAm at the INHBA promoter (p = 0.05) and theupstream enhancer of FN1 (p = 0.02). Gestational agewas only associated with FN1 DNAm (p = 0.03). None ofthe clinical factors assessed was associated with DNAmat the PAPPA site (Table 1). The observation that birthweight (SD) was associated with INHBA DNAm, withoutan association with gestational age, emphasizes the im-portance of including both gestational age and birthweight when considering the relationship betweenDNAm and other variables.What clinical factors are associated with proteinconcentration in maternal blood?We also assessed the same clinical parameters for associ-ation to protein concentration in maternal blood (Table 2).Gestational age at blood draw was only assessed as acovariate for FN1 as clinical values for INHA and PAPPAwere given in multiples of the median (MoM), which wasalready corrected for GA at blood draw. Placental effi-ciency (fetal: placental weight-ratio, at birth) was associ-ated with increased second trimester INHA levels inmaternal blood. FN1 level was not associated with mater-nal BMI in the controls for which we had this information(N = 37), though it was significant when evaluating allclinical groups together (EOPE, LOPE, IUGR, Controls)(N = 75). It was therefore included in subsequent analyses.None of the assessed factors were associated with PAPPAmaternal blood levels during pregnancy (Table 3).What is the relationship between DNAm and maternalserum levels?DNAm in the promoter of INHBA correlated with sec-ond trimester protein levels in maternal blood (r = −0.50,Fig. 1 DNAm distribution at INHBA, PAPPA, and FN1 across all clinicalgroups. The DNAm distribution (β values ± SD) at each site across clinicalgroups for a INHBA, b PAPPA, and c FN1. EOPE = early-onset PE (N= 20),LOPE = late-onset PE (N= 11), IUGR = Intrauterine growth restriction(N = 12), Control (N = 37). *p< 0.05Wilson et al. BMC Medical Genetics  (2015) 16:111 Page 3 of 10ABCDEFFig. 2 Correlation between placental DNAm and gene expression at term in control samples and between placental DNAm and maternal blood proteinlevels during gestation in control samples. The correlation between DNAm at a regulatory element and gene expression (log2) in eight early-onset PEand eight control placentae in a INHBA b PAPPA and c FN1 † All gene expression graphs were produced from data published in Blair et al. (2013). Therelationship between d INHBA (N= 36) promoter DNAm in the term placenta and second trimester INHA levels in maternal blood, plotted as residualscorrected for fetal birth weight (SD) and fetal: placental ratio, e PAPPA (N= 34) promoter DNAm in the term placenta and first trimester PAPPA levels inmaternal blood, and f FN1(N= 76) enhancer DNAm in the term placenta and second/third trimester FN1 levels in maternal blood, plotted as residualscorrected for fetal birth weight (SD), gestational age, and maternal body mass index (BMI). MoM=multiple of the medianWilson et al. BMC Medical Genetics  (2015) 16:111 Page 4 of 10p = 0.01) while modeling for both fetal birth weight (SD)and fetal: placental weight ratio (Fig. 2d). Additionally,DNAm in an upstream enhancer of FN1 correlated withthird trimester protein levels in maternal blood (r = −0.38,p = 0.009) while adjusting for birth weight (SD), gesta-tional age, and maternal BMI (Fig. 2f). This supported ourprediction that DNAm changes observed in the placentacould explain some of the previous reports of alteredINHA and FN1 levels in maternal blood in PE. It isremarkable that these serum measurements from thesecond and third trimesters of pregnancy reflected DNAmat term. This implies that this DNAm change may be anearly alteration in PE. In contrast, a similar result was notobserved for PAPPA/PAPPA (Fig. 2e).We had predicted that protein levels in maternal bloodwould reflect placental DNAm and gene expression.While this may be true in some instances (e.g. INHA,FN1), in other cases establishing a relationship may bechallenging (e.g. PAPPA). Establishing such a relation-ship may be complicated by several factors. Protein leveldepends not only on the level of gene expression, butalso on the total number of cells expressing that protein,the number of mRNA transcripts being translated intoprotein in those cells, and the rate and mode of release ofthe protein into maternal blood. These factors may beinfluenced by the underlying pathology (i.e. more proteinmay be released with increased apoptosis) and placentalsize; which, in turn may be associated with fetal weightand/or fetal: placental weight ratio. Other factors such asexpression of the same protein from maternal tissues, andthe metabolism of proteins by the placenta, reducing theamount of protein being secreted into the maternal circu-lation may have a substantial influence of the total proteinconcentration in maternal blood (Fig. 3). PAPPA has beenfound to be expressed from other maternal sources (e.g.ovary, some epithelial and endometrial cells, and breast)Table 1 Univariate linear analysis results (DNAm vs. Clinical parameters) in controls. Reported in correlation coefficient (r) valuesGene N= FetalSexGA atDeliveryBirth Weight(SD)Fetal: PlacentalWeightPlacental Maternal BMILength: Breadth (Number of samples BMI was available)INHBA 36 0.53 0.055 0.29* 0.08 0.01 0.35 (N = 18)PAPPA 34 0.18 0.17 0.24 0.00 0.26 0.25 (N = 21)FN1 76 0.12 0.22* 0.23* 0.10 0.23 0.30 (N = 75, all samples)0.12 (N = 37,control only)GA gestational age*p < 0.05Table 2 Samples used for pyrosequencing and to assess maternal FN1 protein levelsControl EOPE LOPE + IUGR LOPE IUGRINHA N= 36 - - - -Mean GA at blood draw (weeks ± SD) 14–20wks - - - -Mean GA at delivery (weeks ± SD) 39.3 (±1.3) - - - -Mean BW (grams ± SD) 3480.3 (±483.4) - - - -Mean MA (years ± SD) 33.5 (±4.4) - - - -Sex (Female/N, %) 18/36, 50 % - - - -PAPPA N= 33 - - - -Mean GA at blood draw(weeks ± SD) 11–13wks - - - -Mean GA at delivery (weeks ± SD) 39.6 (±1.4) - - - -Mean BW (grams ± SD) 3428.9 (±355.9) - - - -Mean MA (years ± SD) 34.2 (±4.6) - - - -Sex (Female/N, %) 18/34, 53 % - - - -FN1 N= 76 13 6 10 9Mean GA at blood draw(weeks ± SD) 31.6 (±6.1) 32.3 (±3.2) 35.9 (±1.3) 37.4 (±2.4) 33.5 (±4.5)Mean GA at delivery (weeks ± SD) 39.1 (±2.9) 33.1 (±3.2) 36.1 (±1.1) 38.4 (±1.9) 35.2 (±4.5)Mean BW (grams ± SD) 3465.3 (±398.94) 1663 (±710) 1921 (±402) 3187 (±683) 1932 (±746)Mean MA (years ± SD) 33.5 (±3.6) 33.4 (±6.4) 32.4 (±5.3) 35.5 (5.5) 33.5 (±3.5)Sex (Female/N, %) a36/74, 49 % 6/13,46 % 3/6, 50 % 6/10, 60 % 6/9,66 %aSex not available on 2 samplesWilson et al. BMC Medical Genetics  (2015) 16:111 Page 5 of 10besides the placenta, and it is possible be that thesesources mask any relationship between placental derivedprotein and DNAm in the placenta [25–28]. It is alsoimportant in the case of PAPPA to note that maternal pro-tein levels were measured in the first trimester andadditional variation may arise over gestation affectingcorrelation with placental DNAm at term.Are there any differences in protein levels between caseand control placentas?To confirm a previous report of altered maternal FN1 inassociation with PE and/or IUGR [19], FN1 levels weremeasured in maternal blood samples from pregnancieswhich subsequently developed EOPE, LOPE + IUGR,LOPE without IUGR, or normotensive IUGR, in additionTable 3 Univariate linear analysis results (Protein Levels vs. Clinical parameters) in controls. Reported in correlation coefficient (r)valuesProtein N= Fetal Sex GA at Delivery GA at Blood Drawa Birth Weight (SD) Fetal Weight:Placental WeightPlacentalLength:BreadthMaternal BMI (Number ofsamples BMI was available)INHA 36 0.20 0.00 NA 0.12 0.44* 0.30 0.34 (N = 18)PAPPA 34 0.11 0.20 NA 0.26 0.08 0.00 0.05 (N = 21)FN1 76 0.05 0.10 0.16 0.11 0.13 0.063 0.25* (N = 75, all samples)0.10 (N = 37, control onlyGA gestational age*p < 0.05aOnly measured for FN1 as INHA and PAPPA levels were obtained from maternal serum screening program and already corrected for gestational age atblood drawFig. 3 Processes that may influence the relationship between DNAm, gene expression and protein expression. Outlines reasons why we may notsee a correlation between placental DNAm and gene expression or between placental gene expression and circulating levels of placental-specificproteins in maternal bloodWilson et al. BMC Medical Genetics  (2015) 16:111 Page 6 of 10to our control cohort (Table 2). Similar to the alterationsin DNAm, changes in FN1 levels were found to be signifi-cantly different from controls only in the EOPE group(Mann U Whitney test), although there was a trend ofincreased FN1 levels between LOPE + IUGR and controls(p = 0.08) (Fig. 4). Our results were in concordance toAuer et al. (2010) who also reported increased levels ofmaternal FN1 in pregnancies complicated by EOPE andLOPE + IUGR. We did not confirm their observation of adecrease of FN1 in pregnancies complicated by IUGR;however, we may have been under-powered to observethis small difference. Furthermore, although we observe adifference in EOPE and LOPE + IUGR compared to con-trols, the range of FN1 levels completely overlap betweenthe groups, hindering FN1 to be an adequate biomarkerused alone.ConclusionThis study provides a link between changes in placentalDNAm at term and protein biomarkers present in themother’s circulation earlier in pregnancy. It emphasizesthe many confounding factors that may influence thisrelationship, explaining why this linkage may not be ob-served for all loci. We chose three genomic sites withsignificantly altered DNAm in term placenta associatedwith PE and that were associated with genes for whichthe protein product is altered in PE/IGUR. Despite this,for only two of the three loci (INHA and FN1) did wefind a correlation between placental DNAm and secondand third trimester maternal serum protein expressionin control samples. Nonetheless, this does suggest thatother DNAm marks may be associated with early differ-ences in gene expression. Furthermore, with the adventof techniques to quantify placental nucleic acids in ma-ternal serum [29], DNAm changes may be more directlylinked to measurable miRNA and RNA in maternal blood.Factors such as placental surface area and mechanisms forrelease into maternal blood, will also affect serum levels ofplacental nucleic acids [30]. Future studies measuring pro-tein levels directly in placental tissue, correlating with ma-ternal levels and investigating the factors affecting rate ofrelease are needed to help translate findings measured inthe term placenta into maternal biomarkers of pregnancyoutcomes in early gestation.MethodsSample informationEthics approval was obtained from both the Universityof British Columbia and BC Women’s and Children’sHospital ethics committees in Vancouver, BC, Canada(H04-70488). Placental samples were obtained with con-sent via recruitment through the Medical Genetics andObstetrics and Gynecology departments. Case informa-tion such as: maternal age, maternal BMI, mode of deliv-ery, gestational age at delivery, fetal sex, birth weight,gestational age at blood draw, results on any moleculartesting, and placental dimensions were recorded.Preeclampsia (PE) was defined according to Society ofObstetricians and Gynecologists of Canada (SOGC) cri-teria as one of i) hypertension (BP > 140/90 mm Hg) andproteinuria (>300 g/day) arising after 20 weeks gestation[2]; ii) HELLP syndrome without hypertension or protein-uria [31]; or iii) eclamptic seizure without previous hyper-tension or proteinuria [32]. EOPE was defined by adiagnosis of PE prior to 34 weeks gestation, and LOPEwas defined as a diagnosis after 34 weeks gestation [33].Intrauterine growth restriction (IUGR) was also definedfollowing SOGC criteria [34] as birth weight < 3rd per-centile accounting for fetal sex and gestational age, orbirth weight < 10th percentile with additional clinical find-ings indicative of poor growth such as: absent or reversedend diastolic velocity on Doppler ultrasound, or oligohy-dramnios. Criteria for exclusion were chronic/pre-existingmaternal hypertension, gestational diabetes, multi-fetalpregnancies, and fetal chromosomal abnormalities. Con-trols were selected based on absence of any criteria listedabove and a placenta with no observable pathology.Whole chorionic villi were sampled from four sites, eachfrom distinct cotelydons of the placenta [13]. Samplingfrom infarcts or other abnormal regions of the placentawas avoided. DNA was extracted from each sampled siteand pooled together in equal proportions. DNA wasassessed for quality on the Nanodrop 1000 spectropho-tometer (ThermoScientific, Wilmington, DE, USA). Threehundred nanograms of each DNA sample was bisulfiteconverted for subsequent analyses. Additionally, RNAextracted from the placental villi with RNeasy kit (Qiagen,Heiden, Germany) and was stored in RNAlater at−80 °C.Fig. 4 FN1 protein levels in maternal blood during gestation across allclinical groups. FN1 levels (Median with interquartile range) in maternalblood are increased in EOPE compared with controls, with a increasingtrend in LOPE + IUGR compared to controls. EOPE = early-onset PE(N = 20), LOPE = late-onset PE (N = 11), IUGR = Intrauterine growthrestriction (N = 12), Control (N = 37). **p < 0.05, *p < 0.1Wilson et al. BMC Medical Genetics  (2015) 16:111 Page 7 of 10RNA quality was assessed on a Bioanalyzer 2100 (Agilent,Santa Clara, USA).While we used a total of 171 placentas for our studies,not all placentas were used in all studies as we were lim-ited by samples run on the 450 K array (N = 66); samplesrun on the Illumina expression array (N = 16), maternalserum screening results (first trimester N = 34, secondtrimester N = 36), or maternal serum samples for FN1testing (N = 114). Additional file 1: Table S1 outlines alist of all samples and which analyses they were used in.Gene expression analysisGene expression was measured with the HT-12v4 Ex-pression BeadChip (Illumina, Inc.) as per Blair et al.(2013) protocol, comparing eight EOPE and eight con-trols [14] (Additional file 2: Table S2).DNA methylation analysisIllumina infinium HumanMethylation450 BeadChip arrayTo compare the DNAm differences between clinicalgroups for each of our candidate genes twenty EOPE, 11LOPE, 8 LOPE + IUGR, 10 IUGR, and 37 control caseswere run on the Illumina Infinium HumanMethylation450BeadChip (450 k) array, which interrogates >480,000 CpGsites in >20,000 genes [35]. Some of these samples werepreviously analyzed in the study reported by Blair et al.(2013). To compare the association between DNAm andprotein levels in maternal blood, 122 placental DNA sam-ples (750 ng) bisulfite converted using the EZ DNA Methy-lation kit (Zymo Research, Irvine, USA). Hybridization ofsamples to the array was completed as per the manufac-turer’s protocol. The microarray chips were scanned by theHiScan 2000 or iScan (Illumina). Data was normalized andanalyzed as per Blair et al. (2010) methods [14].Bisulfite pyrosequencingCandidate CpGs determined from the 450 k array data inBlair et al. (2013) were followed up with bisulfite pyrose-quencing in control cohorts for each candidate gene(Table 4). To compare the association between DNAmand protein levels in maternal blood, 122 placental DNAsamples (750 ng) were bisulfite converted using the EZDNA methylation-Gold kit (Zymo Research Corp, Irvine,CA, USA) as per manufacturer’s protocol. Bisulfite con-verted DNA was PCR amplified prior to pyrosequencing.PCR reactions consisted of 20 ng of bisulfite convertedDNA, 1x PCR buffer (with MgCl2) (Qiagen Ltd.), 0.18UDNA polymerase (HotStarTaq, Qiagen Ltd.),0.2 mM dNTP(Invitrogen, Carlsbed, CA),0.4uM forward and reverseprimers (Integrated DNA Technologies, Coralville,IA) forINHABA,PAPPA, and FN1. PCR conditions were 95 °C(15 min), [95 °C (30s), 55 °C (30s), 72 °C (30s)]x40 cycles,72 °C (10 min). Pyrosequencing assays for the candidategenes were designed in PSQ Assay Design software(Biotage, Upsalsa, Sweden) and run on a Qiagen PyromarkQ96 MD (Qiagen) (Additional file 3: Table S3).Candidate DNAm selectionCpG sites chosen to investigate in the present studywere selected on i) a significant change in placentalDNAm, defined as a false discovery rate (FDR) < 0.05 aΔ β > 0.05 (i.e. at least 5 percentage points difference inDNAm), a cut-off that enriches for changes in DNAmthat would likely have biological impact [36], in placen-tas associated with PE and ii) genes encoding for pro-teins reported to show altered levels in maternal bloodin pregnancies complicated by PE and/or IUGR. Inaddition to meeting these criteria, INHBA and PAPPAwere chosen as we had maternal serum measures avail-able on INHA and PAPPA from the maternal serum-screening program. We chose FN1 since the differencein DNAm between EOPE and controls was Δβ = 0.24.We also took into account where the DNAm alterationwas in the genome, taking interest in alterations in generegulatory elements (Table 4).Maternal blood protein measurementsMeasurements of Pregnancy associated plasma protein A(PAPPA) and Inhibin alpha (INHA) were obtained fromclinical maternal serum screening data for 36 and 33women, respectively, and are measured in multiples of themedian (MoM). Additionally, blood was drawn in EDTAtubes during the second trimester for a subset of 158women (Table 2). Plasma was obtained via centrifugationat 3000 rpm for 10 min 4 °C. Plasma Fibronectin (FN1)was measured using a FN1 ELISA kit (eBioscience, SanDiego, CA, USA). FN1 measurements were run in dupli-cate and absorbance was measured at 450 nm. A 5 param-eter asymmetrical logistic curve was generated from thestandard data points which ranged from 0.31-20.0 ng/mL.Samples were diluted as per manufacturer’s protocol; sam-ples which FN1 concentration was over the standard curvewere diluted to 1 in 80,000, and 4 samples which remainedwere further diluted to 1 in 100,000.Table 4 Candidate CpG sites chosen for follow-upGene Site Genomic Region Distance to TSS (bp) EOPE (Change in Beta value from control group)INHBA cg11079619 Active Promoter 76 0.434 (−0.162)PAPPA cg08189448 Active Promoter −163 0.326 (−0.074)FN1 cg12436772 Intergenic/Upstream enhancer −101593 0.465 (−0.240)Wilson et al. BMC Medical Genetics  (2015) 16:111 Page 8 of 10Statistical analysisDNAm at the two CpGs in the PAPPA pyrosequencingassay were correlated (r = 0.85, p < 0.001, Spearman’s cor-relation) and the measurements for these two sites werethus averaged (Additional file 4: Figure S1).Potential covariates which may be associated with eitherDNAm or protein concentration in maternal blood wereassessed for each candidate site. Univariate linear regres-sion analyses were performed, investigating gestational ageat delivery, fetal sex, fetal birth weight (SD), fetal: placentalweight ratio, placental length: breadth ratio, maternal BMI,and when appropriate, gestational age at blood draw. Asabsolute fetal birth weight is confounded by gestational ageat delivery, fetal birth weight was measured as a standarddeviation relative to the mean for that gestational age.PAPPA and INHA protein levels were expressed in MoMto correct for gestational age a blood draw.Correlations were performed when testing any associ-ation between placental gene expression at term andplacental DNAm at term. Spearman’s correlations wereperformed between protein concentration and DNAm insites where there were no covariate factors. For sites withcovariate factors, which needed to be modeled for, partialcorrelations were performed. Non-parametric t-tests wereperformed to determine if DNAm in the EOPE, LOPE +IUGR, LOPE, and IUGR placentas were significantlydifferent from controls. Statistics were calculated usingSPSS v19.0 statistical package.Additional filesAdditional file 1: Table S1. All samples used and the measurescompleted on each one. (PDF 353 kb)Additional file 2: Table S2. Information on samples used to assessDNAm and gene expression in the placenta. (PDF 174 kb)Additional file 3: Table S3. Primer sequences for bisulfitepyrosequencing. Specific locations are based on UCSC hg/18assembly. (PDF 296 kb)Additional file 4: Figure S1. Spearman’s correlation between CpG 1and CpG 2 within the PAPPA assay. DNA methylation was averaged overthe two CpGs (PDF 249 kb)Abbreviations450 k array: Illumina HumanMethylation450 bead chip array.; DNAm: DNAmethylation; EOPE: Early-onset preeclampsia; FN1: Fibronectin; INHBA: Inhibinbeta-alpha; IUGR: Intrauterine growth restriction; LOPE: Late-onsetpreeclampsia; PAPPA: Pregnancy associated plasma protein A;PE: Preeclampsia; SD: Standard deviation; wks: Weeks.Competing interestsThe authors declare no competing interests.Authors’ contributionsSLW participated in study design, acquisition and analysis of data, and draftingthe manuscript. WPR participated in study design and data interpretation. PVDand SL were involved in patient recruitment/sample collection JDB collectedgene expression and 450 k array data. KH aided in design of PAPPA primers. Allauthors contributed to and approved the final manuscript.AcknowledgmentsWe would like to thank Kristal Louie and Johanna Schuetz for recruitingpatients for this study; as well as Ruby Jiang for her assistance in placentadissection and DNA extraction. Thanks to Dr. Maria Peñaherrera and MagdaPrice for reviewing and providing valuable feedback on the manuscript. Thanksalso to Dr. Michael Kobor for use of the pyrosequencing machine. SLW isfunded through the University of British Columbia Four Year DoctoralFellowship, and PvD and WPR receive salary support through investigatorshipawards from the CFRI. Work related to this study was funded through CanadianInstitute of Health Research (CIHR) (#49520) to WPR, SL and PvD.Author details1Child & Family Research Institute, 950 W 28th Ave, Vancouver, BC V5Z 4H4,Canada. 2Department of Medical Genetics, University of British Columbia,C201-4500 Oak St, Vancouver, BC V6H3N1, Canada. 3Department ofMolecular & Cell Biology, University of California Berkeley, Berkeley, CA, USA.4Hudson Institute of Medical Research, Centre for Genetic Diseases, 27-31Wright Street, Melbourne, Australia. 5Department of Obstetrics andGynaecology, University of British Columbia, 4500 Oak St, Vancouver, BC V6H3 V5, Canada.Received: 31 March 2015 Accepted: 27 November 2015References1. Alberry M, Soothill P. Management of fetal growth restriction. Arch Dis ChildFetal Neonatal Ed. 2007;92(1):F62–7.2. 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):S1–48.3. Huang T, Hoffman B, Meschino W, Okun N. Prediction of adverse pregnancyoutcomes by combinations of first and second trimester biochemistrymarkers used in the routine prenatal screening of down syndrome. PrenatDiagn. 2010;30(5):471–7.4. Smith GC, Stenhouse EJ, Crossley JA, Aitken DA, Cameron AD, Connor JM.Early pregnancy levels of pregnancy-associated plasma protein a and therisk of intrauterine growth restriction, premature birth, preeclampsia, andstillbirth. J Clin Endocrinol Metab. 2002;87(4):1762–7.5. Ong CY, Liao AW, Spencer K, Munim S, Nicolaides KH. First trimestermaternal serum free β human chorionic gonadotrophin and pregnancyassociated plasma protein A as predictors of pregnancy complications.BJOG. 2000;107(10):1265–70.6. Bersinger NA, Smárason AK, Muttukrishna S, Groome NP, Redman CW. Womenwith preeclampsia have increased serum levels of pregnancy-associatedplasma protein A (PAPP-A), inhibin A, activin A and soluble E-selectin.Hypertens Pregnancy. 2003;22(1):45–55.7. Poon LC, Syngelaki A, Akolekar R, Lai J, Nicolaides KH. Combined screeningfor preeclampsia and small for gestational age at 11–13 weeks. Fetal DiagnTher. 2013;33(1):16–27.8. Audibert F, Boucoiran I, An N, Aleksandrov N, Delvin E, Bujold E, et al. Screeningfor preeclampsia using first-trimester serum markers and uterine artery Dopplerin nulliparous women. Am J Obstet Gynecol. 2010;203(4):383. e1-383. e8.9. Sitras V, Paulssen R, Leirvik J, Vartun A, Acharya G. Placental gene expressionprofile in intrauterine growth restriction due to placental insufficiency.Reprod Sci. 2009;16(7):701–11.10. Sitras V, Paulssen R, Grønaas H, Leirvik J, Hanssen T, Vårtun Å, et al.Differential placental gene expression in severe preeclampsia. Placenta.2009;30(5):424–33.11. Smets EM, Visser A, Go AT, van Vugt JM, Oudejans C. Novel biomarkers inpreeclampsia. Clin Chim Acta. 2006;364(1):22–32.12. Tsai S, Hardison NE, James AH, Motsinger-Reif AA, Bischoff SR, Thames BH, et al.Transcriptional profiling of human placentas from pregnancies complicated bypreeclampsia reveals disregulation of sialic acid acetylesterase and immunesignalling pathways. Placenta. 2011;32(2):175–82.13. Avila L, Yuen R, Diego-Alvarez D, Peñaherrera M, Jiang R, Robinson W.Evaluating DNA methylation and gene expression variability in the humanterm placenta. Placenta. 2010;31(12):1070–7.14. Blair JD, Yuen RK, Lim BK, McFadden DE, Von Dadelszen P, Robinson WP.Widespread DNA hypomethylation at gene enhancer regions in placentasWilson et al. BMC Medical Genetics  (2015) 16:111 Page 9 of 10associated with early-onset preeclampsia. Mol Hum Reprod.2013;19(10):697–708.15. Nishizawa H, Ota S, Suzuki M, Kato T, Sekiya T, Kurahashi H, et al. Comparativegene expression profiling of placentas from patients with severe pre-eclampsiaand unexplained fetal growth restriction. Reprod Biol Endocrinol. 2011;9:107.16. Gurusinghe S, Wallace EM, Lim R. The relationship between Activin A andanti-angiogenic factors in the development of pre-eclampsia. PregnancyHypertens: An Int J Women’s Cardiovas Health. 2014;4(1):3–6.17. Hogg K, Blair JD, von Dadelszen P, Robinson WP. Hypomethylation of theLEP gene in placenta and elevated maternal leptin concentration in earlyonset pre-eclampsia. Mol Cell Endocrinol. 2013;367(1):64–73.18. Yuen RK, Peñaherrera 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):1006–12.19. Auer J, Camoin L, Guillonneau F, Rigourd V, Chelbi ST, Leduc M, et al. Serumprofile in preeclampsia and intra-uterine growth restriction revealed byiTRAQ technology. J Proteome. 2010;73(5):1004–17.20. Akolekar R, Syngelaki A, Sarquis R, Zvanca M, Nicolaides KH. Prediction of early,intermediate and late pre‐eclampsia from maternal factors, biophysical andbiochemical markers at 11–13 weeks. Prenat Diagn. 2011;31(1):66–74.21. Novakovic B, Yuen RK, Gordon L, Penaherrera MS, Sharkey A, Moffett A,et al. Evidence for widespread changes in promoter methylation profile inhuman placenta in response to increasing gestational age andenvironmental/stochastic factors. BMC Genomics. 2011;12(1):529.22. Tobi EW, Lumey LH, Talens RP, Kremer D, Putter H, Stein AD, et al. DNAmethylation differences after exposure to prenatal famine are common andtiming- and sex-specific. Hum Mol Genet. 2009;18(21):4046–53.23. El-Maarri O, Becker T, Junen J, Manzoor SS, Diaz-Lacava A, Schwaab R, et al.Gender specific differences in levels of DNA methylation at selected locifrom human total blood: a tendency toward higher methylation levels inmales. Hum Genet. 2007;122(5):505–14.24. Engel SM, Joubert BR, Wu MC, Olshan AF, Haberg SE, Ueland PM, et al.Neonatal genome-wide methylation patterns in relation to birth weight in theNorwegian Mother and Child Cohort. Am J Epidemiol. 2014;179(7):834–42.25. Overgaard MT, Oxvig C, Christiansen M, Lawrence JB, Conover CA, Gleich GJ, etal. Messenger ribonucleic acid levels of pregnancy-associated plasma protein-Aand the proform of eosinophil major basic protein: expression in humanreproductive and nonreproductive tissues. Biol Reprod. 1999;61(4):1083–9.26. Lawrence JB, Oxvig C, Overgaard MT, Sottrup-Jensen L, Gleich GJ, Hays LG.Yates JR,3rd, Conover CA: The insulin-like growth factor (IGF)-dependent IGFbinding protein-4 protease secreted by human fibroblasts is pregnancy-associated plasma protein-A. Proc Natl Acad Sci U S A. 1999;96(6):3149–53.27. Mosesson MW, Amrani DL. The structure and biologic activities of plasmafibronectin. Blood. 1980;56(2):145–58.28. CLEMMENSEN I. Fibronectin and its role in connective tissue diseases. Eur JClin Invest. 1981;11(3):145–6.29. Manokhina I, Wilson SL, Robinson WP: Non-invasive nucleic-acid basedapproaches to monitor placental health and predict pregnancy-relatedcomplications. Am J Obs Gynecol 2015 (In Press).30. Metzenbauer M, Hafner E, Hoefinger D, Schuchter K, Stangl G, Ogris E, et al.Three-dimensional ultrasound measurement of the placental volume inearly pregnancy: method and correlation with biochemical placentaparameters. Placenta. 2001;22(6):602–5.31. Audibert F, Friedman SA, Frangieh AY, Sibai BM. Clinical utility of strictdiagnostic criteria for the HELLP (hemolysis, elevated liver enzymes, and lowplatelets) syndrome. Am J Obstet Gynecol. 1996;175(2):460–4.32. Douglas KA, Redman CW. Eclampsia in the United Kingdom. BMJ.1994;309(6966):1395–400.33. Von Dadelszen P, Magee LA, Roberts JM. Subclassification of preeclampsia.Hypertens Pregnancy. 2003;22(2):143–8.34. Kramer MS, Platt RW, Wen SW, Joseph KS, Allen A, Abrahamowicz M, et al.Fetal/infant health study group of the Canadian perinatal surveillancesystem: a new and improved population-based Canadian reference for birthweight for gestational age. Pediatrics. 2001;108(2):E35.35. Bibikova M, Barnes B, Tsan C, Ho V, Klotzle B, Le JM, et al. High density DNAmethylation array with single CpG site resolution. Genomics. 2011;98(4):288–95.36. Hogg K, Price EM, Robinson WP. Improved reporting of DNA methylationdata derived from studies of the human placenta. Epigenetics.2014;9(3):333–7.•  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:Wilson et al. BMC Medical Genetics  (2015) 16:111 Page 10 of 10

Cite

Citation Scheme:

        

Citations by CSL (citeproc-js)

Usage Statistics

Share

Embed

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"
                            src="{[{embed.src}]}"
                            data-item="{[{embed.item}]}"
                            data-collection="{[{embed.collection}]}"
                            data-metadata="{[{embed.showMetadata}]}"
                            data-width="{[{embed.width}]}"
                            async >
                            </script>
                            </div>
                        
                    
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:
http://iiif.library.ubc.ca/presentation/dsp.52383.1-0361954/manifest

Comment

Related Items