{"paper_id":"823b985d-30be-4273-82b6-431320d1e12c","body_text":"Twin Research and Human Genetics\nVolume 18 Number 5 pp. 518–525 C⃝ The Author(s) 2015 doi:10.1017/thg.2015.61\nIndependent Replication and Meta-Analysis for\nEndometriosis Risk Loci\nYadav Sapkota,1, ∗ Amelie Fassbender,2,3, ∗ Lisa Bowdler,1 Jenny N. Fung,1 Dani ¨elle Peterse,2,3\nDorien O,2,3 Grant W. Montgomery,1, ∗ Dale R. Nyholt,1,4, ∗ and Thomas M. D’Hooghe2,3,5, ∗\n1QIMR Berghofer Medical Research Institute, Brisbane, Queensland, Australia\n2Department of Development and Regeneration, Organ Systems, KU Leuven, Leuven, Belgium\n3Department of Obstetrics and Gynaecology, Leuven University Fertility Centre, University Hospital Leuven, Leuven,\nBelgium\n4Institute of Health and Biomedical Innovation, Queensland University of Technology, Brisbane, Queensland, Australia\n5Division of Reproductive Biology, Institute of Primate Research, Karen, Nairobi, Kenya\nEndometriosis is a complex disease that affects 6–10% of women in their reproductive years and 20–50% of\nwomen with infertility. Genome-wide and candidate-gene association studies for endometriosis have identi-\nfied 10 independent risk loci, and of these, nine (rs7521902, rs13394619, rs4141819, rs6542095, rs1519761,\nrs7739264, rs12700667, rs1537377, and rs10859871) are polymorphic in European populations. Here we\ninvestigate the replication of nine SNP loci in 998 laparoscopically and histologically confirmed endometrio-\nsis cases and 783 disease-free controls from Belgium. SNPs rs7521902, rs13394619, and rs6542095 show\nnominally significant (\np < .05) associations with endometriosis, while the directions of effect for seven SNPs\nare consistent with the original reports. Association of rs6542095 at the IL1A locus with ‘All’ ( p = .066)\nand ‘Grade_B’ (p = .01) endometriosis is noteworthy because this is the first successful replication in an\nindependent population. Meta-analysis with the published results yields genome-wide significant evidence\nfor rs7521902, rs13394619, rs6542095, rs12700667, rs7739264, and rs1537377. Notably, three coding\nvariants in\nGREB1 (near rs13394619) and CDKN2B-AS1 (near rs1537377) also showed nominally significant\nassociations with endometriosis. Overall, this study provides important replication in a uniquely charac-\nterized independent population, and indicates that the majority of the original genome-wide association\nfindings are not due to chance alone.\n■ Keywords: endometriosis, genome-wide association study, replication, meta-analysis, IL1A\nEndometriosis is the most common cause of pelvic pain that\na f f e c t s6 – 1 0 %o fw o m e ni nt h e i rr e p r o d u c t i v ey e a r s( T r e l o a r\net al., 1999) and 20–50% of women with infertility (Gao\net al., 2006). It is primarily characterized by the presence of\nendometrium-like tissue in cavities other than the uterus.\nBesides severe pelvic pain, women with endometriosis may\nalso suffer from heavy or irregular menstrual bleeding, pain\nduring intercourse and exercise, infertility, lower abdominal\nand back pain, diarrhea, and/or constipation and chronic\nfatigue. Endometriosis has a complex etiology that results\nfrom an interplay of genetic and non-genetic risk factors and\nhas an estimated heritability of approximately 51% (Treloar\net al., 1999).\nWe and others have conducted genome-wide associa-\ntion (GWA) studies involving individuals of European and\nJapanese ancestries to identify genetic risk factors for en-\ndometriosis (Albertsen et al., 2013; Nyholt et al., 2012;\nPainter et al., 2011; Uno et al., 2010). Results from the four\nGWA studies strongly associated nine independent single\nnucleotide polymorphism (SNP) loci with endometriosis,\nproviding genome-wide significant (p < 5× 10\n−8)e v i d e n c e\nin at least one study. These risk loci include the following:\nrs10965235 in the CDKN2BAS gene at 9p21.3, rs12700667\nat 7p15.2, rs7521902 near WNT4 at 1p36.12, rs13391619\nin the GREB1 at 2p25.1, rs10859871 near VEZT at 12q22,\nrs4141819 at 2p14, rs7739264 nearID4 at 6p22.3, rs1537377\nRECEIVED 1 July 2015; ACCEPTED 9 July 2015.\n∗Equal contributions.\nADDRESS FOR CORRESPONDENCE : Y adav Sapkota, Department of\nGenetics and Computational Biology, QIMR Berghofer Medical\nResearch Institute, 300 Herston Road, Herston 4006 Queensland,\nAustralia. E-mail: Y adav.Sapkota@qimrberghofer.edu.au\n518\nhttps://doi.org/10.1017/thg.2015.61 Published online by Cambridge University Press\n\nReplication Study for Endometriosis Risk Loci\nnear CDKN2B-AS1 at 9p21.3, and rs1519761 at 2q23.3.\nThe GREB1 signal was implicated in a meta-analysis of the\nEuropean (QIMRHCS +OX) and Japanese (BBJ) GWA\ndata after combining published results for rs1339416 from\nAdachi et al. ( 2010) — a small GWA study comprising\n696 endometriosis cases and 825 controls of Japanese de-\nscent. A recent follow-up study by ourselves further im-\nplicated GREB1 locus in endometriosis risk, with stronger\nassociation signals within the region (Fung et al., 2015).\nThree risk loci (rs4141819 at 2p14, rs7739264 near ID4 at\n6p22.3, and rs1537377 near CDKN2B-AS1 at 9p21.3) were\nimplicated when analysis was conducted after excluding\nendometriosis cases with minimal or mild (revised Amer-\nican Fertility Society [rAFS] stage 1 or 2 disease; Ameri-\ncan Society for Reproductive Medicine,1997)e n d o m e t r i o -\nsis (Nyholt et al., 2012). More recently, we conducted a\ncandidate-gene association study (Sapkota, et al., 2015b)\nto investigate potential role of the interleukin 1A ( IL1A)\nvariants reported by two small Japanese GWA studies for\nendometriosis (Adachi et al., 2010; Hata et al., 2013). The\nstudy results provided genome-wide significant evidence\nfor the association of rs6542095 in IL1A with endometrio-\nsis. Taken together, results suggest 10 independent SNP\nloci for endometriosis at genome-wide significant level; of\nthese, nine are polymorphic in populations of European\nancestry.\nThe three replication studies (Pagliardini et al., 2013,\n2015; Sundqvist et al., 2013) performed to date have only\nreplicated association of rs7521902 (p= 5.6× 10\n−3; Pagliar-\ndini et al., 2013) and rs10859871 (p = 6.9 × 10−5; Pagliar-\ndini et al.,2015) with endometriosis. Inherently hidden fine\nstratification among different populations of similar ances-\ntry, which is often difficult to tease apart, power, and possi-\nble variations in disease definition and/or classification may\nhave contributed to these inconsistent results. Moreover,\nthese replication studies have investigated only a fraction of\nthe implicated SNP loci, and hence additional replication\nstudies are required to examine association of all risk loci\nwith endometriosis.\nHere we extended our previous study (Sundqvist et al.,\n2013) to evaluate all nine implicated SNP loci for en-\ndometriosis that are polymorphic in populations of Euro-\npean ancestry, utilizing GWA data for surgically confirmed\n998 endometriosis cases and 783 disease-free controls from\nBelgium. Importantly, severity of disease in endometrio-\nsis cases has been prospectively graded using the rAFS\nclassification system, and hence is less likely to be mis-\nclassified as compared with the retrospective disease stag-\ning based on clinical records. We also performed meta-\nanalysis for the nine risk loci after combining results from\nthe current study with the relevant published results. Fi-\nnally, we conducted GWA analysis of the observed data in\nt h eB e l g i a nc o h o r tt os e ew h e t h e rt h e r ew e r ea n yn o v e l\nloci associated with endometriosis risk identified in this\nsample.\nMaterials and Methods\nStudy Participants\nThe cases ( n = 1,077) and controls ( n = 900) who were\nrecruited in this study at the Leuven University Hospital,\nBelgium during 1993–2012 had undergone laparoscopy for\nsub-fertility with or without pain. Presence of endometrio-\nsis in cases was confirmed laparoscopically and histologi-\ncally based on electronic medical file records. The disease\nseverity in women with endometriosis was prospectively\ngraded according to the rAFS classification system. En-\ndometriosis cases had either minimal (stage I, n = 380),\nmild (stage II, n = 229), moderate (stage III, n = 174),\nsevere (stage IV ,n = 284), or unknown ( n = 10) disease.\nAbsence of endometriosis in controls was confirmed la-\nparoscopically. Both cases and controls were Caucasian in\norigin. All the study participants provided written informed\nconsent, and the study was approved by the Commission of\nMedical Ethics of the Leuven University Hospital, Belgium\nand QIMR Berghofer Human Ethics Research Committee,\nAustralia.\nDNA Extraction and Genotyping\nDNA was purified from EDTA-stabilized whole blood col-\nlected for routine molecular diagnostic tests at the Cen-\ntre for Human Genetics of University Hospitals, Leuven,\nBelgium. Following the manufacturer’s protocol, DNA was\npurified using Chemagic DNA blood special kit (Chema-\ngen MSM I, PerkinElmer Chemagen T echnologies GmbH,\nBaesweiler, Germany) based on the specific binding of DNA\nto paramagnetic beads, and Auto Pure LS Puregene chem-\nistry (Qiagen, V enlo, The Netherlands) based on salting-\nout extraction and a manual salting-out procedure (home-\nb r e w ) .T h ec h o i c eo ft h ee x t r a c t i o nm e t h o dw a sb a s e do n\nthe available amount of blood and on the type of required\nmolecular diagnostic test. DNA concentration was mea-\nsured using Victor (PerkinElmer, Massachussetts, USA).\nWhole genome genotyping of the DNA samples was\nperformed using the Illumina HumanCoreExome 12v1.1\narray at the Molecular Epidemiology Laboratory, QIMR\nBerghofer Medical Research Institute, Brisbane, Australia,\nfollowing the manufacturer’s standard protocol. For a qual-\nity control (QC) check, DNA concentrations of majority of\nthe samples were re-measured on the BioTech Powerwave\nat the QIMR Molecular Epidemiology Laboratory before\ngenotyping. The Illumina HumanCoreExome genotyping\narrays are the newer generation of Illumina GWA arrays,\nwhich comprised /223c250,000 common tag SNPs (‘core’) and\n/223c250,000 predominantly rare coding variants (‘exome’).\nThe exome variants in the Illumina HumanCoreExome\nwere carefully selected based on exome sequencing data\nin /223c12,000 individuals.\nGenotype Calling and Quality Control\nGenotype data were called using a custom cluster file gen-\nerated using /223c2,000 good quality ( <1% missing rate)\nTWIN RESEARCH AND HUMAN GENETICS 519\nhttps://doi.org/10.1017/thg.2015.61 Published online by Cambridge University Press\n\nYadav Sapkota et al.\nsamples and the GenCall algorithm within Illumina\nGenome Studio. Data were then further processed by zCall\n(Goldstein et al.,2012) ,ar a r ev a r i a n tc a l l e r ,i na na t t e m p tt o\nrecall missing genotypes. Following manufacturer’s guide-\nlines and the protocols developed for the Exome chip data,\nquality control measures were applied to the Belgian GWA\ndata. Briefly, samples with >1% missing rates, outlying\nheterozygosity, non-European ancestries (based on 1,000\nGenomes Project’s European populations), cryptic related-\nness (pi-hat > 0.2), and gender discordances were excluded.\nSimilarly, markers with poor separation of three genotype\nclusters, excess heterozygosity, outlying mean theta and in-\ntensity values for heterozygote genotypes, >1% missing\nrates, the Hardy–Weinberg Equilibrium (HWE) p < 10\n−6\nin controls, and minor allele frequency (MAF) < .05% in\neither cases or controls were dropped.\nAssociation Analysis\nSince rs1096523 in the CDKN2BAS gene at 9p21.3 is\nmonomorphic in populations of European ancestry, we\nconsidered the remaining nine SNP loci (rs7521902,\nrs13394619, rs4141819, rs6542095, rs1519761, rs7739264,\nrs12700667, rs1537377, and rs10859871) for further anal-\nysis. T o see whether there were any novel association sig-\nnals for endometriosis in the Belgian cohort, GWA analy-\nsis of the observed genotypes was performed using –assoc\n(for ‘core’ SNPs) and –fisher (for ‘exome’ variants) com-\nmands in Plink for data, including all endometriosis cases\n(‘All’) and controls. Considering the relatively greater ge-\nnetic loading of moderate-to-severe (rAFS stage III/IV or\n‘Grade_B’) endometriosis compared with mild or minimal\n(rAFS stage I/II or ‘Grade_A’) disease (Nyholt et al., 2012;\nPainter et al., 2011; Sapkota et al., 2015a), additional anal-\nysis for ‘Grade_B’ endometriosis cases versus controls was\nalso performed. Strengths of association of SNPs with en-\ndometriosis are reported in terms of odds ratio (ORs) and\nconfidence intervals (CIs).\nImputation and Meta-Analysis\nOf the nine SNPs, only six (rs7521902, rs13394619,\nrs4141819, rs6542095, rs7739264, and rs12700667) are\nassayed on Illumina HumanCoreExome 12v1.1 geno-\ntyping platform. Therefore, we imputed genotypes in\nchromosomes containing the nine SNP loci in the Belgian\nGWA data, using a reference panel of 1,000 Genomes\nProject (March 2012 release). Imputation was carried out\nusing SHAPEIT (Delaneau et al., 2012) and minimac\nprograms (Li et al., 2009, 2010) and following the two-step\napproach outlined in the online Minimac: 1,000 Genomes\nImputation Cookbook (http://genome.sph.umich.edu/\nwiki/Minimac:_1000_Genomes_Imputation_Cookbook).\nQuality of the imputed genotypes was assessed byr\n2 metric,\nwhich estimates the squared correlation between true and\nimputed genotypes. Poorly imputed SNPs indicated by r2\n< .3 were excluded from downstream analyses. Association\nTABLE 1\nSummary of the Datasets Used in the Current Study\nEndometriosis\nGWA study cases (‘Grade_B’) Controls Ethnicity\nBelgian 998 (423) 783 Caucasian\nQIMRHCS 2,262 (905) 2,924 European\nOX 919 (452) 5,151 European\nBBJ 1,423 1,318 Japanese\nAlbertsen et al. (2013) 1,514 12,660 Caucasian\n— discovery\nAlbertsen et al. (2013) 505 1,811 Caucasian\n— replication\nAdachi et al. (2010) 696 825 Japanese\nNote: QIMRHCS = Queensland Institute of Medical Research and Hunter\nCommunity Study; OX = Oxford (UK); BBJ = BioBank of Japan.\nanalyses of imputed genotype dosage scores of the nine\nSNP loci were conducted using Plink for ‘All’ and ‘Grade_B’\nendometriosis cases separately.\nAfter combining results of imputed dosage scores for the\nnine SNP loci from Belgian data with the published results\nof Nyholt et al. (2012), Adachi et al. (2010), Albertsen et al.\n(2013), and Sapkota et al. ( 2015b), we performed meta-\nanalysis for ‘All’ endometriosis cases and controls. A brief\nsummary of the datasets used in this study is provided in\nTable 1.R e s u l t sf r o mN y h o l te ta l .(2012) included summary\nstatistics of rs7521902, rs13394619, rs4141819, rs7739264,\nrs12700667, rs1537377, and rs10859871 obtained from the\nEuropean (QIMRHCS+OX) and Japanese (BBJ) GWA data\n(Table 1). Similarly, we included results of rs13394619 and\nrs6542095 from Adachi et al. ( 2010), obtained from the\ncombined analysis of Affymetrix 500K and 6.0 arrays in 696\ncases and 825 controls of Japanese ancestry. Furthermore,\nresults from Albertsen et al. ( 2013) included summary\nstatistics of rs1519761 in their discovery and replication\nstages. Finally, we obtained summary results of rs6542095\nin the QIMRHCS, OX, and BBJ imputed data from Sapkota\net al. (2015b).\nInitial meta-analysis was conducted using a fixed-effect\n(inverse variance-weighted) model implemented in the\nGWAMA program (Magi & Morris, 2010). Heterogeneity\nof allelic associations was examined using the Cochran’s Q\nstatistic p\nhet < .1 (Cochran, 1954), as well as the I2 index\n(Ioannidis et al., 2007), which indicates the proportion of\nvariance attributable to between-study heterogeneity. Meta-\nanalysis of SNPs associated in fixed-effect model with an\nevidence of heterogeneity ( p < .1) was carried out using\nthe Han–Eskin random-effects model (RE2; Han & Eskin,\n2011) implemented in the METASOFT program. In con-\ntrast with the conventional random-effects model, the RE2\nmodel increases power under heterogeneity. Furthermore,\nadditional meta-analysis for the nine SNP loci was also\nperformed by restricting to ‘Grade_B’ endometriosis cases\n(wherever available) versus controls.\nResults\nFollowing the QC steps, a total of 998 endometriosis\ncases and 783 disease-free controls with 316,467 markers\n520 TWIN RESEARCH AND HUMAN GENETICS\nhttps://doi.org/10.1017/thg.2015.61 Published online by Cambridge University Press\n\nReplication Study for Endometriosis Risk Loci\nTABLE 2\nSummary Results of the Nine Known SNP Loci for Endometriosis in the Current Study\n‘All’ ‘Grade_B’\nChr SNP Position RA OA RAF cases RAFcontrols OR (95% CI) p RAFcases RAFcontrols OR (95% CI) p\n1 rs7521902 22490724 A C 0.243 0.221 1.13 (0.97–1.32) .125 0.271 0.221 1.30 (1.08–1.58) .007\n2 rs13394619 11727507 G A 0.538 0.503 1.14 (1.00–1.30) .045 0.533 0.503 1.13 (0.95–1.33) .164\n2 rs4141819 67864675 C T 0.307 0.328 0.91 (0.79–1.05) .183 0.311 0.328 0.92 (0.77–1.11) .384\n2 rs6542095 113529183 C T 0.324 0.295 1.14 (0.99–1.32) .066 0.346 0.295 1.26 (1.06–1.51) .010\n2 rs1519761 151633204 G A 0.410 0.414 0.98 (0.86–1.13) .814 0.408 0.414 0.97 (0.82–1.16) .755\n6 rs7739264 19785588 T C 0.520 0.515 1.02 (0.89–1.16) .806 0.545 0.515 1.12 (0.95–1.32) .172\n7 rs12700667 25901639 A G 0.758 0.747 1.06 (0.91–1.24) .445 0.772 0.747 1.15 (0.94–1.40) .166\n9 rs1537377 22169700 C T 0.391 0.413 0.91 (0.79–1.04) .178 0.415 0.413 1.01 (0.85–1.20) .935\n12 rs10859871 95711876 C A 0.322 0.327 0.98 (0.84–1.13) .747 0.329 0.327 1.01 (0.84–1.21) .915\nNote: Chr = chromosome; Position = chromosomal position (bp) based on Human Build 37 (GRCh37/hg19); RA = risk allele from original study; OA = other\nallele; OR = odds ratio; CI = confidence interval.\nremained in the Belgian GWA data for downstream anal-\nysis. Of these, 246,071 were ‘core’ SNPs whereas 70,396\nwere ‘exome’ variants. The GWA analysis of observed geno-\ntypes of the 316,467 markers in the Belgian GWA study\nalone did not produce any genome-wide significant hits\nin either ‘All’ or ‘Grade_B’ analysis, with few suggestive\n(p < 1 × 10\n−5) associations (data not shown). All nine\nSNPs were accurately imputed with r2 > .95. We also com-\npared imputed genotypes (dosage scores) of six SNP loci\n(rs7521902, rs13394619, rs4141819, rs6542095, rs7739264,\nand rs12700667) for endometriosis with the observed true\ngenotypes available in the Belgian data. Genotype concor-\ndances (as measured by the Pearson’s correlation coeffi-\ncient) between two sets of genotypes for the six SNPs were\n>0.99 (p < 2.2 × 10\n−16).\nAssociation analysis of the dosage scores of the nine im-\nplicated SNP loci in the Belgian data provided further in-\nsights into the associations of these SNPs with endometrio-\nsis. Risk alleles and their frequencies of all nine SNPs were\nsimilar to the ones reported in the original studies (Table 2;\nAlbertsen et al.,2013; Nyholt et al.,2012; Painter et al.,2011;\nSapkota et al., 2015b) and their associations were stronger\nwith ‘Grade_B’ than ‘All’ endometriosis. Furthermore, ef-\nfect directions of seven out of nine tested SNPs in either ‘All’\nor ‘Grade_B’ endometriosis were in line with the published\nresults. Three SNPs showed statistically significant associa-\ntion with endometriosis in either ‘All’ or ‘Grade_B’ disease\nat a nominal p < .05. SNP rs7521902 showed borderline\nmarginal association (OR = 1.13; p = .12) with ‘All’ en-\ndometriosis. As expected, its association was stronger and\nstatistically significant (OR= 1.30; p= .007) with ‘Grade_B’\ncases. A statistically significant association (OR= 1.14; p =\n.045) for rs13394619 was also observed for ‘All’ endometrio-\nsis; however, the signal was slightly weaker (OR = 1.13; p\n= .164) in ‘Grade_B’ cases. A borderline association (OR\n= 1.14; p = .06) with ‘All’ endometriosis was observed for\nrs6542095, which was stronger and significant (OR= 1.26;\np = .01) in ‘Grade_B’ cases.\nMeta-analysis, including imputed data from Belgian co-\nhort and the published results, provide insights into SNP\nloci associated with endometriosis. Six SNP loci showed\nassociations with either ‘All’ or ‘Grade_B’ endometriosis at\ngenome-wide significance level ( p < 5 × 10\n−8)a n dw i t h\nsimilar directions of effect across all studies included in the\nanalysis (Table 3). Of these, three SNPs were associated with\nboth ‘All’ and ‘Grade_B’ endometriosis, with a genome-\nwide significant evidence in the fixed-effect meta-analysis.\nThese included: SNP rs7521902 near WNT4 (‘All’ , OR=\n1.17; 95% CI = 1.11–1.23; p = 3.63 × 10\n−8;‘ G r a d e _ B ’ ,O R\n= 1.25; 95% CI= 1.17–1.34; p = 1.72 × 10−10), rs13394619\nin GREB1 (‘All’ , OR= 1.15; 95% CI = 1.10–1.20; p = 9.13\n× 10−9;‘ G r a d e _ B ’ ,O R= 1.17; 95% CI = 1.11–1.24; p =\n3.02 × 10−8), and rs12700667 at 7p15.2 (‘All’ , OR= 1.19;\n95% CI = 1.13–1.26; p = 7.10 × 10−10;‘ G r a d e _ B ’ ,O R=\n1.29; 95% CI = 1.20–1.39; p = 1.47 × 10−11). The IL1A\nSNP (rs6542095) was genome-wide significantly associated\nwith only ‘Grade_B’ (OR = 1.22; 95% CI = 1.14–1.30; p =\n1.00 × 10\n−9) endometriosis in fixed-effect meta-analysis,\nbut after appropriate modeling for between-study hetero-\ngeneity in ‘All’ endometriosis ( p\nhet = .007) in the RE2\nmodel, the association reached genome-wide significance\n(p = 3.35 × 10\n−8). Statistical significance of association\nof rs6542095 with ‘Grade_B’ endometriosis also became\nstronger ( p = 4.90 × 10\n−10)i nt h eR E 2m o d e la f t e ra c -\ncounting for between-study heterogeneity ( phet = .01). A\nstrong association between rs7739264 near ID4 and ‘All’\nendometriosis (p = 1.93 × 10−7) was observed, and the sig-\nnal was further enriched in ‘Grade_B’ endometriosis (OR\n= 1.20; 95% CI = 1.12–1.27; p = 1.98 × 10\n−8), achieving\na genome-wide significance. Similarly, near genome-wide\nsignificant evidence for association between rs1537377 near\nCDKN2B-AS1 and ‘Grade_B’ endometriosis (OR = 1.19;\n95% CI= 1.12–1.27; p = 9.27 × 10\n−8)w a so b s e r v e d ,w h i c h\nwas genome-wide significant (p = 4.80 × 10−8)a f t e rm o d -\neling for borderline between-study heterogeneity ( phet =\n.1). However, the effect of rs1537377 in ‘All’ endometriosis\nwas in opposite direction of the published results (Table 2).\nWhile the remaining three SNPs (rs4141819, rs1519761,\nand rs10859871) did not produce genome-wide signifi-\ncant evidence for association with either ‘All’ or ‘Grade_B’\nTWIN RESEARCH AND HUMAN GENETICS 521\nhttps://doi.org/10.1017/thg.2015.61 Published online by Cambridge University Press\n\nYadav Sapkota et al.\nTABLE 3\nMeta-Analysis for the Nine Known Endometriosis SNP Loci After Combining Summary Statistics From Current Study With the\nPublished Results\nPosition No. of\nChr SNP (bp) RA OA Cases OR (95% CI) pp het I2 studies Direction RE2 p\n1 rs7521902 22490724 A C ‘All’ 1.17 (1.11–1.23) 3.63 × 10−8 0.533 0.00 4 ++++ NC\n‘Grade_B’ 1.25 (1.17–1.34) 1.72 × 10−10 0.762 0.00 4 ++++ NC\n2 rs13394619 11727507 G A ‘All’ 1.15 (1.10–1.20) 9.13 × 10−9 0.183 0.36 5 +++++ NC\n‘Grade_B’ 1.17 (1.11–1.24) 3.02 × 10−8 0.243 0.27 5 +++++ NC\n2 rs4141819 67864675 C T ‘All’ 1.12 (1.07–1.19) 2.38 × 10−5 0.017 0.71 4 +++-1 . 8 8 × 10−5\n‘Grade_B’ 1.17 (1.09–1.25) 1.01 × 10−5 0.004 0.77 4 +++-3 . 6 3 × 10−7\n2 rs6542095 113529183 C T ‘All’ 1.15 (1.09–1.21) 1.05 × 10−7 0.007 0.72 5 +++++ 3.35 × 10−8\n‘Grade_B’ 1.22 (1.14–1.30) 1.00 × 10−9 0.014 0.68 5 +++++ 4.90 × 10−10\n2 rs1519761 151633204 G A ‘All’ 1.15 (1.08–1.22) 5.62 × 10−6 0.025 0.73 3 ++-1 . 0 7 × 10−5\n‘Grade_B’ 1.17 (1.09–1.24) 1.94 × 10−6 0.067 0.63 3 ++-7 . 9 9 × 10−6\n6 rs7739264 19785588 T C ‘All’ 1.14 (1.08–1.20) 1.93 × 10−7 0.359 0.07 4 ++++ NC\n‘Grade_B’ 1.20 (1.12–1.27) 1.98 × 10−8 0.756 0.00 4 ++++ NC\n7 rs12700667 25901639 A G ‘All’ 1.19 (1.13–1.26) 7.10 × 10−10 0.405 0.00 4 ++++ NC\n‘Grade_B’ 1.29 (1.20–1.39) 1.47 × 10−11 0.338 0.11 4 ++++ NC\n9 rs1537377 22169700 C T ‘All’ 1.10 (1.05–1.16) 9.88 × 10−5 0.041 0.64 4 +++-1 . 0 0 × 10−4\n‘Grade_B’ 1.19 (1.12–1.27) 9.27 × 10−8 0.101 0.52 4 ++++ 4.80 × 10−8\n12 rs10859871 95711876 C A ‘All’ 1.16 (1.09–1.22) 4.29 × 10−7 0.137 0.46 4 +++-N C\n‘Grade_B’ 1.17 (1.10–1.25) 2.46 × 10−6 0.385 0.02 4 ++++ NC\nNote: Chr = chromosome; Position = chromosomal position (bp) based on Human Build 37 (GRCh37/hg19); RA = risk allele; OA = other allele; OR = odds\nratio; CI = confidence interval; Phet = Cochran’s Q between-study heterogeneity test p value; I2, percentage of variance attributable to between-study\nheterogeneity; RE2 = Han Eskin’s random effects model meta-analysis; NC = not calculated.\nendometriosis in fixed-effect meta-analysis, they still\nshowed strong associations with the disease ( p < 2.38 ×\n10−5). Nonetheless, the effects of rs4141819 and rs1519761\nin both ‘All’ and ‘Grade_B’ endometriosis, and that of\nrs10859871 in ‘All’ endometriosis were in opposite direc-\nt i o n so ft h ep u b l i s h e dr e s u l t s(Table 2). SNP rs4141819\nshowed between-study heterogeneity ( p\nhet < .01) in both\n‘All’ and ‘Grade_B’ endometriosis, and after accounting\nfor this heterogeneity in the RE2 model, association of\nrs4141819 with ‘Grade_B’ disease became stronger with\nnear genome-wide significant evidence (p = 3.63 × 10\n−7).\nSignificant between-study heterogeneity for rs1519761 was\nalso observed in both ‘All’ and ‘Grade_B’ endometriosis,\nbut its association with the disease (‘All’ ,p = 5.62 × 10\n−6;\n‘Grade_B’ ,p = 1.94 × 10−6) slightly diluted in the RE2\nmodel (‘All’ , p = 1.07 × 10−5;‘ G r a d e _ B ’ ,p = 7.99 ×\n10−6). A near genome-wide significant association between\nrs10859871 near VEZT and ‘All’ endometriosis (OR= 1.16;\n95% CI = 1.09–1.22; p = 4.29 × 10−7) was observed, with\nslightly larger effect size (OR = 1.17; 95% CI = 1.10–1.25)\nin ‘Grade_B’ disease, although statistical significance of the\nsignal was weaker (p = 2.46 × 10\n−6).\nDiscussion\nEndometriosis is a complex disease and studies have shown\nthat genetic risk factors substantially contribute to the risk\nof endometriosis. Genetic studies, especially the GWA stud-\nies for endometriosis, have identified 10 SNP loci, of which\nnine are polymorphic in the populations of European ori-\ngin (Albertsen et al., 2013; Nyholt et al.,2012; Painter et al.,\n2011; Sapkota et al., 2015b). While much larger and well-\npowered GWA studies are needed to identify additional\ngenetic risk factors involved in the risk of endometriosis,\nreplication studies are crucial to provide credibility that the\ninitial genotype–phenotype associations are valid. Repeated\nobservation of such associations in independent popula-\ntions of similar ethnicity adds evidence that the associations\nare not due to chance alone. The previous three replication\nstudies for endometriosis have investigated only a handful\nof the 10 implicated SNP loci to date. Here we report the\nmost comprehensive replication study performed to date,\nin which we examine all nine implicated SNP risk loci for\nendometriosis that are polymorphic in populations of Eu-\nropean ancestry, by utilizing GWA data in uniquely char-\nacterized 998 endometriosis cases and 783 controls from\nBelgium.\nThe risk alleles and their frequencies for all the nine\nSNPs in the Belgian replication cohort were comparable\nwith the original studies (Table 2; Albertsen et al.,2013;N y -\nholt et al., 2012; Painter et al., 2011; Sapkota et al., 2015b).\nMoreover, direction of effects for seven of the nine SNPs\nfor either ‘All’ or ‘Grade_B’ endometriosis was also con-\nsistent with the published results. Among these, we could\nsuccessfully replicate associations of three SNPs (rs7521902,\nrs13394619, and rs6542095) with either ‘All’ or ‘Grade_B’\nendometriosis at nominalp < .05, which is more often than\nby chance alone (p = .008; one-sided binomial test). Signif-\nicant association of rs6542095 at the IL1A locus with ‘All’\n(p = .066) and ‘Grade_B’ (p = .01) endometriosis is note-\nworthy as this is the first successful replication in an inde-\npendent population, providing further supporting evidence\nfor a potential link between inflammation and endometrio-\nsis pathogenesis. More importantly, all the SNPs showed\nlarger effect sizes with ‘Grade_B’ than ‘All’ endometriosis —\nan observation consistent with the previous reports\n522 TWIN RESEARCH AND HUMAN GENETICS\nhttps://doi.org/10.1017/thg.2015.61 Published online by Cambridge University Press\n\nReplication Study for Endometriosis Risk Loci\nTABLE 4\nAssociation Statistics for the ‘Best’ SNPs at the Six Genome-Wide Significant Loci for Endometriosis Reported in Nyholt et al. (2012)\n‘All’ ‘Grade_B’\nChr SNP Position RA OA RAF cases RAFcontrols OR (95% CI) p RAFcases RAFcontrols OR (95% CI) p\n1 rs56318008 22470407 T C 0.178 0.151 1.23 (1.02–1.48) .031 0.190 0.151 1.33 (1.06–1.66) .014\n2 rs77294520 11660955 C G 0.166 0.147 1.18 (0.97–1.44) .098 0.175 0.147 1.28 (1.00–1.63) .051\n2 rs2861694 67845739 G A 0.692 0.670 1.11 (0.96–1.28) .161 0.688 0.670 1.09 (0.91–1.30) .368\n6 rs6901079 19776659 C T 0.249 0.245 1.02 (0.88–1.19) .762 0.260 0.245 1.09 (0.90–1.32) .402\n9 rs7041895 22162794 A C 0.599 0.583 1.07 (0.93–1.23) .327 0.576 0.583 0.97 (0.81–1.15) .716\n12 rs11107968 95690444 A G 0.680 0.676 1.02 (0.89–1.18) .779 0.672 0.676 0.98 (0.82–1.17) .846\nNote: Chr = chromosome; Position = chromosomal position (bp) based on Human Build 37 (GRCh37/hg19); RA = risk allele; OA = other allele; OR = odds\nratio; CI = confidence interval.\nsupporting greater genetic loading in moderate-to-severe\ndisease (Nyholt et al., 2012; Painter et al., 2011;S a p k o t a\net al., 2015a).\nOur meta-analysis, including results from the cur-\nrent replication study and the published results, pro-\nduced genome-wide significant evidence for six (rs7521902\nnear WNT4, rs13394619 in GREB1, rs6542095 in IL1A,\nrs7739264 near ID4, rs12700667 at 7p15.2, and rs1537377\nnear CDKN2B-AS1) of the nine implicated SNPs in ei-\nther ‘All’ or ‘Grade_B’ endometriosis, after accounting for\nbetween-study heterogeneity using the RE2 model, wher-\never appropriate (Table 3). With the exception of rs1519761\nat 2q23.3 reported by Albertsen et al. (2013), the other two\nSNP loci (rs4141819 at 2p14 and rs10859871 near VEZT)\nalso showed near genome-wide significance for ‘Grade_B’\nendometriosis in the RE2 model (p = 3.63 × 10\n−7)a n df o r\n‘All’ endometriosis in the fixed-effect model ( p = 4.29 ×\n10−7). The association signal for rs1519761 was the weak-\nest (‘All’ ,p = 5.62 × 10−6;‘ G r a d e _ B ’ ,p = 1.94 × 10−6)\namong the nine risk loci, and the signal was slightly diluted\nafter accounting for observed between-study heterogene-\nity in the RE2 model (‘All’ , p = 1.07 × 10\n−5;‘ G r a d e _ B ’ ,p\n= 7.99 × 10−6). Association signal at this locus was also\nnot replicated in a recent meta-analysis for endometriosis\n(Rahmioglu et al., 2014), suggesting that further investiga-\ntion is required to confirm a role for this locus in the risk of\nendometriosis.\nIn our multi-ethnic GWA meta-analysis that strongly\nassociated seven risk loci with endometriosis, we found\nstronger associations at six loci (rs56318008 at 1p36.12,\nrs77294520 at 2p25.1, rs2861694 at 2p14, rs6901079 at\n6p22.3, rs7041895 at 9p21.3, and rs11107968 at 12q22)\nwhen we imputed genotypes in the region 2,500 kb up-\nstream and downstream of the most significant geno-\ntyped SNP using the full reference panel from the 1,000\nGenomes Project Interim Phase 1 Haplotypes (2010–2011\ndata freeze). For the risk loci at 7p15.2, the genotyped\nSNP rs12700667 was the best signal. For the remaining\nsix loci with stronger association signals (‘best’ SNPs) post-\nimputation than the genotyped SNP , we assessed for their\nreplication in the Belgian cohort ( Table 4). All six SNPs\nwere accurately imputed in the current study withr\n2 > .85.\nAssociation results were consistent with that of the orig-\ninal genotyped SNPs, as shown in Table 2, in particular\nfor SNPs rs56318008 at 1p36 and rs77294520 at 2p25.1,\nwhich showed nominally significant associations with ‘All’\n(p < .098) and ‘Grade_B’ ( p < .051), providing further\nsupporting evidence for implication of these risk loci in\nendometriosis.\nAs a first step to help identify causal variants at nine\nSNP loci, we interrogated the ExomeChip data for puta-\ntively functional coding variants within genes harboring or\nclosest to GWA SNPs. For the GREB1 locus, we found an-\nother coding variant, rs10929757, showing nominally sig-\nnificant association ( p = .015) with endometriosis. The\neffect size of rs10929757 was similar (OR = 1.18) to the\nGWA SNP rs13394619, although they are poorly correlated\n(r\n2 = .25). Similarly, we also observed nominally significant\nassociation (p < .018) for two coding variants (rs2383207\nand rs4977574) in CDKN2B-AS1 —t h ec l o s e s tg e n et ot h e\nGWA SNP rs1537377 at 9p21.3. In spite of lack of corre-\nlation ( r\n2 = .011 and .005, respectively) with rs1537377,\nthe effect sizes for both variants were similar (ORs = 1.17\nand 1.19 respectively). While these data may suggest inde-\npendent association signals at GREB1 and 9p21.3 loci, the\nthree coding variants need to be further investigated in a\nlarger sample size for a more conclusive interpretation. We\ndid not observe evidence of association for other coding\nvariants at nominal p < .05, even though the effect sizes\nf o rs o m ew e r ec o m p a r a b l ew i t hG W AS N P sa te a c hr i s k\nloci (data not shown). We did not detect any rare coding\nvariants at GWA loci despite the ExomeChip data. This may\nbe due, in part, to reduced power in the Belgian sample to\ndetect such rare variants. Assuming a disease prevalence of\n8%, our sample size only had 45% power to detect alle-\nles of frequency .20 contributing to genotype relative risk\nof 1.15 (Purcell et al., 2003). As such, larger ExomeChip\nstudies may be required to adequately investigate potential\nrole of coding/rare variants in the risk of endometriosis\nand other complex traits. We cannot rule out the possibil-\nity of other types of rare functional variants at these loci,\nwhich are not adequately captured by either ExomeChip\nor current imputation methods, contributing to increased\nrisk of endometriosis. These issues may be addressed by\nTWIN RESEARCH AND HUMAN GENETICS 523\nhttps://doi.org/10.1017/thg.2015.61 Published online by Cambridge University Press\n\nYadav Sapkota et al.\nthe future studies utilizing larger sample sizes, coupled\nwith re-sequencing and further fine-mapping required to\nidentify causal variants within the implicated GWA loci.\nFurthermore, controls used in this study were clinic-based\nendometriosis-free individuals who presented with symp-\ntoms of sub-fertility. As such, they may have different allele\nfrequencies as compared with ‘population-based’ controls\nused in most GWA studies, and hence may partly explain the\nopposite direction of effect sizes observed for some SNPs.\nHowever, this needs to be investigated further using a larger\nsample size with population-based controls, and therefore\ncaution should be used interpreting these results.\nOverall, results from the current replication study pro-\nvide further supporting evidence for associations of the\nimplicated SNP loci with endometriosis. Meta-analysis\nfor these loci after including additional published results\nproduced genome-wide significant evidence for six loci,\nwith similar magnitudes and directions of effect across\nstudies, and hence provided further evidence against any\npossibility of inflated genetic effects due to the ‘winner’s\ncurse’ bias in the original study. More importantly, all the\nnine SNPs showed larger effect sizes with stage III/IV en-\ndometriosis than all cases, corroborating our previous ob-\nservation for greater genetic loading in moderate-to-severe\nendometriosis.\nAcknowledgments\nWe would like to thank all the participants in the en-\ndometriosis studies who were included in this analysis. We\nalso thank many hospital directors and staff, gynecologists,\ngeneral practitioners, and pathology services who provided\nassistance with confirmation of diagnoses. Dale R. Nyholt\nwas supported by an NHMRC Fellowship (613674) and\nARC Future Fellowship (FT0991022) schemes, and Grant\nW. Montgomery was supported by the NHMRC Fellow-\nships Scheme (339446, 619667).\nReferences\nA d a c h i ,S . ,T a j i m a ,A . ,Q u a n ,J . ,H a i n o ,K . ,Y o s h i h a r a ,K . ,\nMasuzaki, H., . . . Tanaka, K. (2010). Meta-analysis of\ngenome-wide association scans for genetic susceptibility\nto endometriosis in Japanese population.Journal of Human\nGenetics, 55, 816–821.\nAlbertsen, H. M., Chettier, R., Farrington, P ., & Ward, K.\n(2013). Genome-wide association study link novel loci to\nendometriosis. PLoS One, 8, e58257.\nAmerican Society for Reproductive Medicine. (1997). Revised\nAmerican Society for Reproductive Medicine classification\nof endometriosis: 1996. Fertility and Sterility, 67 , 817–821.\nCochran, W. G. (1954). The combination of estimates from\ndifferent experiments. Biometrics, 10 , 101–129.\nDelaneau, O., Marchini, J., & Zagury, J. F. (2012). A linear com-\nplexity phasing method for thousands of genomes. Nature\nMethods, 9, 179–181.\nFung, J. N., Holdsworth-Carson, S. J., Sapkota, Y., Zhao, Z. Z.,\nJones, L., Girling, J.E., . . . Montgomery, G. W. (2015).\nFunctional evaluation of genetic variants associated with\nendometriosis near GREB1. Human Reproduction, 30 ,\n1263–1275.\nGao, X., Outley, J., Botteman, M., Spalding, J., Simon, J. A., &\nPashos, C. L. (2006). Economic burden of endometriosis.\nFertility and Sterility, 86 , 1561–1572.\nGoldstein, J. I., Crenshaw, A., Carey, J., Grant, G. B., Maguire,\nJ., Fromer, M., . . . Neale, B. M. (2012). zCall: A rare variant\ncaller for array-based genotyping: Genetics and population\nanalysis. Bioinformatics, 28, 2543–2545.\nHan, B., & Eskin, E. (2011). Random-effects model aimed at\ndiscovering associations in meta-analysis of genome-wide\nassociation studies. American Journal of Human Genetics,\n88, 586–598.\nHata, Y., Nakaoka, H., Y oshihara, K., Adachi, S., Haino, K.,\nY amaguchi, M., . . . Tanaka, K. (2013). A non-synonymous\nvariant of IL1A is associated with endometriosis in\nJapanese population. Journal of Human Genetics, 58 , 517–\n520.\nIoannidis, J. P ., Patsopoulos, N. A., & Evangelou, E. (2007).\nHeterogeneity in meta-analyses of genome-wide associa-\ntion investigations. PLoS One, 2, e841.\nLi, Y., Willer, C. J., Ding, J., Scheet, P ., & Abecasis, G. R. (2010).\nMaCH: Using sequence and genotype data to estimate hap-\nlotypes and unobserved genotypes. Genetic Epidemiology,\n34, 816–834.\nLi, Y., Willer, C., Sanna, S., & Abecasis, G. (2009). Genotype\nimputation. Annual Review of Genomics and Human Genet-\nics, 10, 387–406.\nMagi, R., & Morris, A. P . (2010). GWAMA: Software for\ngenome-wide association meta-analysis.BMC Bioinformat-\nics, 11, 288.\nNyholt, D. R., Low, S. K., Anderson, C. A., Painter, J. N., Uno,\nS., Morris, A. P ., . . . Montgomery, G. W. (2012). Genome-\nwide association meta-analysis identifies new endometrio-\nsis risk loci. Nature Genetics, 44, 1355–1359.\nPagliardini, L., Gentilini, D., Sanchez, A. M., Candiani, M.,\nVigano, P ., & Di Blasio, A. M. (2015). Replication and meta-\nanalysis of previous genome-wide association studies con-\nfirm vezatin as the locus with the strongest evidence for\nassociation with endometriosis. Human Reproduction, 30 ,\n987–993.\nPagliardini, L., Gentilini, D., Vigano, P ., Panina-Bordignon, P .,\nBusacca, M., Candiani, M., . . . Di Blasio, A. M. (2013). An\nItalian association study and meta-analysis with previous\nGWAS confirm WNT4, CDKN2BAS and FN1 as the first\nidentified susceptibility loci for endometriosis. Journal of\nMedical Genetics, 50, 43–46.\nPainter, J. N., Anderson, C. A., Nyholt, D. R., Macgregor, S.,\nLin, J., Lee, S. H., . . . Zondervan, K. T. (2011). Genome-\nwide association study identifies a locus at 7p15.2 associated\nwith endometriosis. Nature Genetics, 43, 51–54.\nPurcell, S., Cherny, S. S., & Sham, P . C. (2003). Genetic power\ncalculator: Design of linkage and association genetic map-\nping studies of complex traits. Bioinformatics, 19 , 149–\n150.\n524 TWIN RESEARCH AND HUMAN GENETICS\nhttps://doi.org/10.1017/thg.2015.61 Published online by Cambridge University Press\n\nReplication Study for Endometriosis Risk Loci\nRahmioglu, N., Nyholt, D. R., Morris, A. P ., Missmer, S. A.,\nMontgomery, G. W., & Zondervan, K. T. (2014). Genetic\nvariants underlying risk of endometriosis: Insights from\nmeta-analysis of eight genome-wide association and repli-\ncation datasets. Human Reproduction Update, 20, 702–716.\nSapkota, Y., Attia, J., Gordon, S. D., Henders, A. K., Holliday,\nE. G., Rahmioglu, N., . . . Nyholt, D. R. (2015a). Genetic\nburden associated with varying degrees of disease severity\nin endometriosis. Molecular Human Reproduction, 21, 594–\n602.\nS a p k o t a ,Y . ,L o w ,S .K . ,A t t i a ,J . ,G o r d o n ,S .D . ,H e n d e r s ,A .K . ,\nHolliday, E. G., . . . Nyholt, D. R. (2015b). Association be-\ntween endometriosis and the interleukin 1A (IL1A) locus.\nHuman Reproduction, 30, 239–248.\nSundqvist, J., Xu, H., Vodolazkaia, A., Fassbender, A.,\nKyama, C., Bokor, A., . . . Falconer, H. (2013). Replica-\ntion of endometriosis-associated single-nucleotide poly-\nmorphisms from genome-wide association studies in a\nCaucasian population. Human Reproduction, 28, 835–839.\nTreloar, S. A., O’Connor, D. T., O’Connor, V . M., & Martin,\nN. G. (1999). Genetic influences on endometriosis in an\nAustralian twin sample. Fertility and Sterility, 71 , 701–\n710.\nUno, S., Zembutsu, H., Hirasawa, A., Takahashi, A., Kubo,\nM., Akahane, T., . . . Nakamura, Y. (2010). A genome-\nwide association study identifies genetic variants in\nthe CDKN2BAS locus associated with endometriosis in\nJapanese. Nature Genetics, 42, 707–710.TWIN RESEARCH AND HUMAN GENETICS 525\nhttps://doi.org/10.1017/thg.2015.61 Published online by Cambridge University Press","source_license":"CC0","license_restricted":false}