TY - JOUR
AU - Kim, Kyung Mo
AB - Abstract Background and Aims The genetic contribution to the prognosis of ulcerative colitis [UC] is poorly understood, and most currently known susceptibility loci are not associated with prognosis. To identify genetic variants influencing the prognosis of UC, we performed an Immunochip-based study using an extreme phenotype approach. Methods Based on the finding that the only association, Pdiscovery-meta <1 × 10-4, was located in the human leukocyte antigen [HLA], we focused our analyses on the HLA region. We performed the analysis using HLA imputation data from three independent discovery cohorts of 607 UC patients [243 poor-prognosis and 364 good-prognosis], followed by replication in 274 UC patients [145 poor-prognosis and 129 good-prognosis]. Results We found that rs9268877, located between HLA-DRA and HLA-DRB, was associated with poor-prognosis of UC at genome-wide significance (odds ratio [ORdiscovery] = 1.82; ORreplication = 1.55; ORcombined-meta = 1.72, pcombined-meta = 1.04 × 10-8), with effect size [OR] increasing incrementally according to worsening of prognosis in each of the three independent discovery cohorts and the replication cohort. However, rs9268877 showed no association with UC susceptibility [ORcombined-meta = 1.07, pcombined-meta = 0.135]; rs9268877 influenced 30-year clinical outcomes, and the presence of the rs9268877 risk allele had a sensitivity of 80.0% and specificity of 38.1% for colectomy. Conclusions Our results provide new insights into prognosis-associated genetic variation in UC, which appears to be distinct from the genetic contribution to disease susceptibility. These findings could be useful in identifying poor-prognosis patients who might benefit from early aggressive therapy. Ulcerative colitis, genetics, biomarker 1. Introduction Genome-wide association studies [GWAS] have identified and confirmed a substantial number of susceptibility loci for ulcerative colitis [UC].1,2 However, the effects of these loci tend to be relatively weak, leaving their clinical utility open to debate. The identification of genetic variants associated with UC progression would have significant clinical utility, allowing clinicians to identify patients at high risk of progressive disease, for whom a more aggressive therapeutic strategy would be appropriate. However, most currently known UC susceptibility variants exhibit weak or inconsistent associations with the clinical course of the disease,3–5 suggesting that the genetic variants associated with disease progression might differ from those associated with susceptibility to UC. Accurate phenotyping and proper strategies for using phenotypic data are critical in identifying prognosis-associated variants. Previous studies aimed at identifying the genetic predictors of UC prognosis were limited by the classification of patients into only two groups, i.e., colectomy versus no colectomy.3,6,7 This classification is simple and easy to apply, but potentially ignores prognostic heterogeneity among patients who do not undergo colectomy, thus preventing robust conclusions. To identify genetic predictors that impact on the prognosis of UC, we adopted an extreme phenotype approach8–10 in the current study. First, we identified patients at opposite ends of the prognostic spectrum and performed a single-centre Korean UC population association analysis in three UC cohorts genotyped using Immunochip v1.0, Immunochip v2.0, and OmniExpress chip, respectively, followed by a meta-analysis of the results. This study, by integrating fine clinical phenotyping data with human leukocyte antigen [HLA] imputation data, provides evidence that a particular HLA variant independently confers poor-prognosis in Korean UC patients. 2. Materials and Methods 2.1. Study population Among 2147 UC patients who were seen at the Asan Medical Center, Seoul, Korea, 1961 patients were included in the study. The remaining 186 patients who required neither immunosuppressive agents (corticosteroids, thiopurines, or anti-tumour necrosis factor [TNF] agents] nor colectomy over follow-up of less than 5 years, were excluded from the present study because their prognosis was considered undetermined. Of the 1961 patients included in the present study [Supplementary Table S1, available as Supplementary data at ECCO-JCC online], 1336 patients with any single nucleotide polymorphism [SNP] array data were used in the discovery phase [671 patients in discovery cohort 1, 475 in discovery cohort 2, and 190 in discovery cohort 3], and the remaining 625 patients were used in the replication phase. Demographic and clinical characteristics of the 1961 subjects are presented in Supplementary Table S1. All patients were of Korean descent. Diagnosis of UC was made on the basis of conventional clinical, radiological, endoscopic, and histopathological criteria.11 Patient demographic and clinical information was retrieved from the Asan IBD [inflammatory bowel disease] registry, which has been prospectively maintained since 1997 and previously described in detail.12 For each subject, any changes in clinical information during the follow-up period, including changes in disease extent according to the Montreal Classification,13 commencement of new medications, and information on surgical procedures, are updated at each clinic visit or at least yearly by IBD clinical specialists. 2.2. Patient classification All cases were treated using a step-up approach, in which therapies with the least toxicity are used first, and subsequent therapies may be administered due to toxicity or lack of response [Figure 1a].14 To reflect the prognostic spectrum, we first classified all 1961 patients into five subgroups based on the highest treatment level reached: 5-aminosalicylates, corticosteroids, thiopurines, anti-TNF agents, or colectomy [Figure 1b]. We next selected subgroups of patients at opposite ends of the prognostic spectrum for association analyses. Good-prognosis UC was defined as mild to moderate disease that required neither immunosuppressive agents [corticosteroids, thiopurines, or anti-TNF agents] nor colectomy despite ≥5 years of disease duration. Poor-prognosis UC was defined as disease that required anti-TNF agents and/or colectomy regardless of disease duration [Figure 1c]. 2.3. Treatment strategy for UC in Korea Our treatment strategies for UC have been detailed previously12,15,16 and are based on a step-up approach [Figure 1a]. Topical and/or oral 5-aminosalicylates are the first-line therapy to induce and maintain remission in mild to moderate UC. Systemic corticosteroid therapy is used in patients with moderate to severe UC, as well as in those who do not respond to 5-aminosalicylates, and is tapered and discontinued over 2 to 3 months. Thiopurines and, in the event of their failure, anti-TNF agents, are used in steroid-dependent or steroid-refractory patients. In Korea, first-line biologic therapy is seldom reimbursed, and patients need to satisfy strict criteria for reimbursement before they become eligible for anti-TNF agent coverage.12,15–17 For patients to receive government reimbursement, they must exhibit an inadequate response to conventional treatment with corticosteroids and/or thiopurines, and moderate or severe disease activity. Figure 1. View largeDownload slide Definition of phenotyping. [a] All ulcerative colitis patients [n = 1961] were treated using a ‘step-up’ treatment strategy. [b] Patients were categorised according to the highest treatment level. [c] Good-prognosis ulcerative colitis was defined as mild to moderate active disease that required neither immunosuppressive agents [corticosteroids, thiopurines, or anti-TNF agents] nor colectomy during a disease duration of ≥5 years. Poor-prognosis ulcerative colitis was defined as disease activity that required anti-TNF agents and/or colectomy. Figure 1. View largeDownload slide Definition of phenotyping. [a] All ulcerative colitis patients [n = 1961] were treated using a ‘step-up’ treatment strategy. [b] Patients were categorised according to the highest treatment level. [c] Good-prognosis ulcerative colitis was defined as mild to moderate active disease that required neither immunosuppressive agents [corticosteroids, thiopurines, or anti-TNF agents] nor colectomy during a disease duration of ≥5 years. Poor-prognosis ulcerative colitis was defined as disease activity that required anti-TNF agents and/or colectomy. 2.4. Genotyping and quality controls Samples in discovery cohort 1 were genotyped as part of an earlier Immunochip study,18 in discovery cohort 2 using the Immunochip v2.0 array, and in discovery cohort 3 as part of an earlier GWAS.19 Genotyping of the replication cohort was conducted using TaqMan genotyping technology [7900HT Fast Real-Time PCR System, Applied Biosystems]. The top SNP imputed in the Immunochip v1.0 data was also genotyped to confirm imputation. Samples and SNPs that failed quality-control criteria in the original studies were excluded.18,19 Genotyping of discovery cohort 2, containing 753 UC cases and 497 unrelated healthy controls, was carried out using the Immunochip v2.0 at the Cedars-Sinai Medical Center, Los Angeles, CA, USA. All SNPs on the X, Y, and mitochondrial chromosomes were excluded. SNPs with call rate <98%, Hardy–Weinberg equilibrium p <1.0 × 10-5 in controls, or minor allele frequency [MAF] <0.01, were excluded. Eight samples [three UC cases and five controls] were removed due to close genetic relatedness [PI_HAT >0.25, IBS >0.8]. Subsequently, principal-component analysis [PCA] was performed to detect population outliers and stratification. Three samples [two UC cases and one control] were removed following PCA. After SNP and sample quality-control analyses of Immunochip v2.0 data, 166 843 SNPs in 748 cases and 491 controls in discovery cohort 2 [average call rate of 99.99%] remained for further analyses. All analyses were performed in PLINK 1.07 and R 3.4.1. 2.5. Statistical analyses An analysis using extreme phenotypes was performed using genotype data from three independent discovery cohorts of 607 UC patients [243 with poor-prognosis and 364 with good-prognosis], followed by replication in 274 UC patients [145 poor-prognosis and 129 good-prognosis cases] [Supplementary Figure S1, available as Supplementary data at ECCO-JCC online]. Because meta-analysis of our Immunochip v1.0, Immunochip v2.0, and OmniExpress datasets across 607 UC patients [243 with poor-prognosis and 364 with good-prognosis] showed association pdiscovery-meta<1 × 10-4 at the HLA locus only [Supplementary Figure S2, available as Supplementary data at ECCO-JCC online], we focused our subsequent analysis on the HLA region. 2.5.1. HLA imputation Genotype data from 29M-34M at chromosome 6 were extracted from the post–quality-control datasets and used to impute SNPs, classical HLA alleles, and amino acids using SNP2HLA.20 The reference panel for imputation, which consists of 413 individuals of Korean origin, has been shown to be accurate for imputing HLA in Koreans;21 the panel included 10 159 binary markers defined for HLA-A, B, C, DRB1, DRB3, DRB4, DRB5, DQB1, and DPB1. Cases and controls were imputed together for 10 159 binary markers separately for three cohorts typed with three different genotyping platforms [Immunochip v1.0, Immunochip v2.0, and GWAS chip]. After imputation, markers with INFO score ≤0.8, MAF <1%, or missing rate >10% of genotypes, in any of the three cohorts were excluded from further analyses. The final set included 4053 SNPs, 87 HLA alleles at four-digit resolution, 58 HLA alleles at two-digit resolution, and 2475 single amino acid variants. 2.5.2. Association testing Associations for binary markers were tested using probabilistic genotypic dosages that took imputation uncertainty into account. First, an analysis between patients with good-prognosis [n = 493] and poor-prognosis [n = 388] was performed using logistic regression. Second, a stratified analysis based on the highest treatment level reached [n = 1468] was performed with reference to good-prognosis cases [n = 493]. In addition, a case-control analysis was performed to estimate the susceptibility risk compared with unrelated healthy controls [n = 5415; 1185 for Immunochip v1.0, 491 for Immunochip v2.0, and 3739 for GWAS chip]. Conditional logistic regression was performed to identify additional associations. The data were combined using a fixed-effect meta-analysis model [inverse-variance weighted average] in discovery analyses with the three discovery cohorts [Meta discovery], as well as in joint analysis with the three discovery cohorts and the replication cohort [Meta combined]. The genome-wide significance threshold was defined as p <5 × 10-8. 2.5.3. Adjusting for clinical heterogeneity Because extensive colitis is a well-established risk factor for colectomy,12,22 discovery analysis was performed including a binary model for disease extent at diagnosis (extensive versus non-extensive disease [left-sided colitis or proctitis]) as covariates. To test whether the prognosis of UC was affected by differences in follow-up duration, disease duration was also included as a covariate in the regression. 2.5.4. Adjusting for genetic heterogeneity To evaluate associations with prognosis, polygenic risk scores [PRS] were computed from the 24 known UC susceptibility loci in Koreans [Supplementary Table S2, available as Supplementary data at ECCO-JCC online].18,19,23 PRS were calculated by summing the number of risk alleles multiplied by their corresponding β coefficients. Statistical significance was assessed separately for the three discovery collections and combined using a t-test between poor-prognosis and good-prognosis cases, as implemented in Prizm [GraphPad, San Diego, CA, USA]. We then compared the frequencies of patients with good-prognosis and patients and with poor-prognosis between the first and fourth quartile of PRS. To assess the effect of the PRS on prognosis, effect size was calculated by linear regression of the PRS against the prognosis while adjusting for the effect of other phenotypes. In addition, logistic regression including PRS as covariate was run to estimate the influence of PRS on prognosis of UC for the discovery collection. 2.5.5. Clinical outcomes according to genotype The influence of the identified variants on long-term outcomes and prognosis, including the cumulative probabilities of colectomy and medication use [anti-TNF agents, thiopurines, and corticosteroids], were assessed in all 1961 patients with UC. The cumulative probabilities were analysed using the Kaplan–Meier method, with differences determined by the log-rank test. To distinguish direct from indirect associations, all regression models were also adjusted for other phenotypes [age of onset, sex, smoking status at UC diagnosis, and disease extent at UC diagnosis]; p <0.05 was considered statistically significant. Statistical evaluations were performed using the packages Prizm and SPSS 21.0 for Windows [IBM SPSS, Ver. 21.0; IBM Co., Armonk, NY, USA]. 2.6. Candidate genes and functional effects Genes of interest in the associated region were identified using Ensembl, UCSC Genome Bioinformatics, and GeneCards. ENCODE project data24 and HaploReg[v4]25 were used to predict the functional effects of the top associated variants. GTEx26 was used to investigate expression of quantitative trait loci within genomic regions of interest. 3. Results 3.1. Patient classification based on a step-up treatment strategy Among 1961 patients, the highest level of treatment was 5-aminosalicylates in 493 patients, corticosteroids in 710 patients, thiopurines in 370 patients, anti-TNF agents in 268 patients, and colectomy in 120 patients [Figure 1b]. We categorized patients at opposite ends of the prognostic spectrum for comparison [Figure 1c]: 493 patients met the definition of good-prognosis [cohort 1, n = 190; cohort 2, n = 115; cohort 3, n = 59; replication, n = 129], and another 388 patients met the definition of poor-prognosis [cohort 1, n = 160; cohort 2, n = 58; cohort 3, n = 25; replication, n = 145]. The number of follow-up visits per year was 3.1 ± 1.4 for patients with good-prognosis [discovery cohorts: 3.2 ± 1.4, replication cohort: 3.0 ± 1.6] and 6.6 ± 5.0 for patients with poor-prognosis [discovery cohorts: 6.4 ± 3.9, replication cohort: 6.9 ± 6.5]. 3.2. Variants associated with UC prognosis To identify genetic predictors of UC prognosis, we first performed association analyses in three UC cohorts genotyped using Immunochip v1.0, Immunochip v2.0, and OmniExpress chip, respectively, followed by a meta-analysis of the results. Based on the finding that the only association pdiscovery-meta <1 × 10-4 was located in the HLA [Supplementary Figure S2, available as Supplementary data at ECCO-JCC online], we focused our subsequent association analyses between binary markers and good versus poor UC prognosis on the HLA region. Discovery meta-analyses of the three cohorts revealed the strongest association between rs9268877 and patients with poor-prognosis who required anti-TNF agents or colectomy [Figure 2 and Table 1]. Conditioning on rs9268877 did not reveal additional independent signals [pconditioned-meta ≥0.006 [Figure 2b; and Supplementary Table S3, available as Supplementary data at ECCO-JCC online]. Linkage disequilibrium between rs9268877 and classic HLA alleles was low [r2 <0.4]. The most strongly associated classical allele was HLA-DRB4*0101 [Supplementary Table S4, available as Supplementary data at ECCO-JCC online]. However, reciprocal conditional analysis on HLA-DRB4*0101 and rs9268877 [r2 = 0.15] showed that the effect of association with HLA-DRB4*0101 was substantially attenuated when accounting for rs9268877, indicating that HLA-DRB4*0101 is not independently associated with prognosis [Supplementary Table S5, available as Supplementary data at ECCO-JCC online]; rs9268877 was selected for replication in an additional 145 poor-prognosis and 129 good-prognosis patients. Combining the results from the three discovery collections and the replication collection using a fixed-effect meta-analysis model, rs9268877 exhibited the strongest association with poor-prognosis UC (poor-prognosis versus good-prognosis: odds ratio [ORcombined-meta] = 1.72; 95% confidence interval [CI] = 1.43–2.07; pcombined-meta = 1.04 × 10-8; Table 1 and Supplementary Table S6, available as Supplementary data at ECCO-JCC online). Consistent associations were observed even after adjustments were made for clinical phenotypes, including disease extent and PRS [Supplementary Table S7, available as Supplementary data at ECCO-JCC online]. Prognosis-associated rs9268877 was not associated with susceptibility to UC in Koreans and East Asians2 [Supplementary Table S8, available as Supplementary data at ECCO-JCC online]. Table 1. Association of rs9268877 with prognosis of ulcerative colitis in the Korean population SNP Position [hg19] Alleles Study No. of subjects RAF OR 95% CI pa I2 pQ Poor prognosis Good prognosis Poor prognosis Good prognosis rs9268877 32431147 A/G Discovery 243 364 0.51 0.35 1.82 1.44–2.29 3.88 × 10-7 0.12 0.32 Replication 145 129 0.50 0.37 1.55 1.14–2.12 5.72 × 10-3 Combined 388 493 0.51 0.35 1.72 1.43–2.07 1.04 × 10-8 0.00 0.41 SNP Position [hg19] Alleles Study No. of subjects RAF OR 95% CI pa I2 pQ Poor prognosis Good prognosis Poor prognosis Good prognosis rs9268877 32431147 A/G Discovery 243 364 0.51 0.35 1.82 1.44–2.29 3.88 × 10-7 0.12 0.32 Replication 145 129 0.50 0.37 1.55 1.14–2.12 5.72 × 10-3 Combined 388 493 0.51 0.35 1.72 1.43–2.07 1.04 × 10-8 0.00 0.41 The odds ratio is presented with respect to the risk allele and the risk of poor-prognosis ulcerative colitis. CI, confidence interval; hg19, human genome version 19; I2, I2 statistics; OR, odds ratio; pQ, p-value for Cochrane Q; RAF, risk allele frequency; SNP, single-nucleotide polymorphism. ap-Value was calculated using the logistic regression model. The combined p-value was calculated using a fixed-effect model of meta-analysis. View Large Table 1. Association of rs9268877 with prognosis of ulcerative colitis in the Korean population SNP Position [hg19] Alleles Study No. of subjects RAF OR 95% CI pa I2 pQ Poor prognosis Good prognosis Poor prognosis Good prognosis rs9268877 32431147 A/G Discovery 243 364 0.51 0.35 1.82 1.44–2.29 3.88 × 10-7 0.12 0.32 Replication 145 129 0.50 0.37 1.55 1.14–2.12 5.72 × 10-3 Combined 388 493 0.51 0.35 1.72 1.43–2.07 1.04 × 10-8 0.00 0.41 SNP Position [hg19] Alleles Study No. of subjects RAF OR 95% CI pa I2 pQ Poor prognosis Good prognosis Poor prognosis Good prognosis rs9268877 32431147 A/G Discovery 243 364 0.51 0.35 1.82 1.44–2.29 3.88 × 10-7 0.12 0.32 Replication 145 129 0.50 0.37 1.55 1.14–2.12 5.72 × 10-3 Combined 388 493 0.51 0.35 1.72 1.43–2.07 1.04 × 10-8 0.00 0.41 The odds ratio is presented with respect to the risk allele and the risk of poor-prognosis ulcerative colitis. CI, confidence interval; hg19, human genome version 19; I2, I2 statistics; OR, odds ratio; pQ, p-value for Cochrane Q; RAF, risk allele frequency; SNP, single-nucleotide polymorphism. ap-Value was calculated using the logistic regression model. The combined p-value was calculated using a fixed-effect model of meta-analysis. View Large Figure 2. View largeDownload slide Result of association analysis for prognosis in Korean patients with ulcerative colitis. The horizontal dashed line indicates variants with pdiscovery-meta = 1.0 × 10-4. [a] Poor-prognosis (patients received anti-tumour necrosis factor [TNF] agents or colectomy) versus good-prognosis [patients received 5-aminosalicylates only despite ≥5 years of disease duration]; the most significant association was observed at rs9268877. [b] Conditioning on rs9268877. Figure 2. View largeDownload slide Result of association analysis for prognosis in Korean patients with ulcerative colitis. The horizontal dashed line indicates variants with pdiscovery-meta = 1.0 × 10-4. [a] Poor-prognosis (patients received anti-tumour necrosis factor [TNF] agents or colectomy) versus good-prognosis [patients received 5-aminosalicylates only despite ≥5 years of disease duration]; the most significant association was observed at rs9268877. [b] Conditioning on rs9268877. Based on the GTEx v6p dataset, in colon tissue and whole blood rs9268877 acts in cis on the expression of class II genes, including HLA-DQA1, HLA-DQA2, HLA-DQB1, HLA-DQB1-AS1, and HLA-DRB1 [Supplementary Table S9, available as Supplementary data at ECCO-JCC online].26 In blood, rs9268877 also acts in trans on the expression of plausible candidate genes, including the T-cell receptor beta variable 18 [TRBV18] and acyloxyacyl hydrolase [AOAH].27–29 Notably in this regard, TCRβ mutant mice develop spontaneous colitis,30 and AOAH modulates microbiota-related signals that influence Th17 response and mucosal T-cell immunity.31 3.3. Clinical outcomes according to rs9268877 To examine the association of rs9268877 with use of medication beyond extreme cases, we performed the association analyses in 1961 patients including 881 extreme cases and 1080 non-extreme cases. When the patients were grouped based on the highest treatment level reached [Figure 1b], the effect size increased according to treatment level [Supplementary Table S10, available as Supplementary data at ECCO-JCC online], with the largest effect in patients who underwent colectomy [OR = 2.22; 95% CI = 1.67–2.95; pcombined-meta-stratified = 4.26 × 10-8], suggesting that the associations might be involved in the biology that determines UC prognosis. We next performed analyses integrating clinical parameters to evaluate long-term probability of treatment. In multivariate Cox analysis, the risk of colectomy was 3-fold higher in risk allele homozygotes [AA] than in non-carriers [GG] after adjusting for clinical variables [p = 1.16 × 10-4; Table 2]. A Kaplan–Meier analysis of 1961 UC patients indicated that rs9268877 was associated with a higher risk of colectomy and need for medications [Figure 3]. The cumulative probability of colectomy was higher in homozygotes [AA] than in non-carriers [GG]: 10.3% versus 3.0% at 10 years, 19.6% versus 4.9% at 20 years, and 26.3% versus 6.4% at 30 years [p = 3.30 × 10-7; Figure 3a]. The presence of the rs9268877 ‘A’ allele had sensitivity of 80.0% [96/120] and specificity of 38.1% [701/1841] for colectomy [Supplementary Figure S3, available as Supplementary data at ECCO-JCC online]. Table 2. Multivariate Cox analysis of risk factors for colectomy and medication use in 1961 ulcerative colitis patients Colectomy Multivariate Cox analysis Anti-TNF agents Multivariate Cox analysis HR 95% CI p HR 95% CI p Age at diagnosisa 1.02 1.01–1.04 0.010 Age at diagnosisa 1.00 0.99–1.01 0.552 Sex Women Reference Sex Women Reference Men 1.88 1.08–3.30 0.026 Men 1.38 1.00–1.91 0.049 Smoking status at Dx Never Reference Smoking status at Dx Never Reference Ex-smoker 1.09 0.59–2.01 0.784 Ex-smoker 1.34 0.93–1.94 0.122 Current smoker 0.57 0.29–1.12 0.102 Current smoker 0.75 0.51–1.11 0.146 Extent at Dx Non-extensive colitis Reference Extent at Dx Non-extensive colitis Reference Extensive colitis 2.84 1.85–4.35 2.00 × 10-6 Extensive colitis 1.96 1.51–2.55 3.69 × 10-7 rs9268877 Non-carrier [GG] Reference rs9268877 Non-carrier [GG] Reference Heterozygote [AG] 1.84 1.07–3.16 0.029 Heterozygote [AG] 1.44 1.07–1.95 0.016 Homozygote [AA] 3.09 1.74–5.49 1.16 × 10-4 Homozygote [AA] 2.00 1.44–2.79 4.20 × 10-5 Thiopurines Corticosteroids Age at diagnosisa 0.99 0.99–1.00 0.050 Age at diagnosisa 0.99 0.99–1.00 6.96 × 10-4 Sex Women Reference Sex Women Reference Men 1.36 1.09–1.69 0.006 Men 1.04 0.90–1.20 0.618 Smoking status at Dx Never Reference Smoking status at Dx Never Reference Ex-smoker 1.20 0.93–1.56 0.168 Ex-smoker 1.12 0.93–1.34 0.226 Current smoker 0.82 0.63–1.06 0.127 Current smoker 0.97 0.81–1.15 0.703 Extent at Dx Non-extensive colitis Reference Extent at Dx Non-extensive colitis Reference Extensive colitis 2.14 1.78–2.56 1.79 × 10-16 Extensive colitis 2.48 2.18–2.83 1.06 × 10-43 rs9268877 Non-carrier [GG] Reference rs9268877 Non-carrier [GG] Reference Heterozygote [AG] 1.21 0.99–1.47 0.064 Heterozygote [AG] 1.26 1.11–1.44 5.01 × 10-4 Homozygote [AA] 1.48 1.17–1.86 0.001 Homozygote [AA] 1.22 1.04–1.44 0.017 Colectomy Multivariate Cox analysis Anti-TNF agents Multivariate Cox analysis HR 95% CI p HR 95% CI p Age at diagnosisa 1.02 1.01–1.04 0.010 Age at diagnosisa 1.00 0.99–1.01 0.552 Sex Women Reference Sex Women Reference Men 1.88 1.08–3.30 0.026 Men 1.38 1.00–1.91 0.049 Smoking status at Dx Never Reference Smoking status at Dx Never Reference Ex-smoker 1.09 0.59–2.01 0.784 Ex-smoker 1.34 0.93–1.94 0.122 Current smoker 0.57 0.29–1.12 0.102 Current smoker 0.75 0.51–1.11 0.146 Extent at Dx Non-extensive colitis Reference Extent at Dx Non-extensive colitis Reference Extensive colitis 2.84 1.85–4.35 2.00 × 10-6 Extensive colitis 1.96 1.51–2.55 3.69 × 10-7 rs9268877 Non-carrier [GG] Reference rs9268877 Non-carrier [GG] Reference Heterozygote [AG] 1.84 1.07–3.16 0.029 Heterozygote [AG] 1.44 1.07–1.95 0.016 Homozygote [AA] 3.09 1.74–5.49 1.16 × 10-4 Homozygote [AA] 2.00 1.44–2.79 4.20 × 10-5 Thiopurines Corticosteroids Age at diagnosisa 0.99 0.99–1.00 0.050 Age at diagnosisa 0.99 0.99–1.00 6.96 × 10-4 Sex Women Reference Sex Women Reference Men 1.36 1.09–1.69 0.006 Men 1.04 0.90–1.20 0.618 Smoking status at Dx Never Reference Smoking status at Dx Never Reference Ex-smoker 1.20 0.93–1.56 0.168 Ex-smoker 1.12 0.93–1.34 0.226 Current smoker 0.82 0.63–1.06 0.127 Current smoker 0.97 0.81–1.15 0.703 Extent at Dx Non-extensive colitis Reference Extent at Dx Non-extensive colitis Reference Extensive colitis 2.14 1.78–2.56 1.79 × 10-16 Extensive colitis 2.48 2.18–2.83 1.06 × 10-43 rs9268877 Non-carrier [GG] Reference rs9268877 Non-carrier [GG] Reference Heterozygote [AG] 1.21 0.99–1.47 0.064 Heterozygote [AG] 1.26 1.11–1.44 5.01 × 10-4 Homozygote [AA] 1.48 1.17–1.86 0.001 Homozygote [AA] 1.22 1.04–1.44 0.017 CI, confidence interval; Dx, diagnosis; HR, hazard ratio; TNF, tumour necrosis factor.. aPer year of increase in age. View Large Table 2. Multivariate Cox analysis of risk factors for colectomy and medication use in 1961 ulcerative colitis patients Colectomy Multivariate Cox analysis Anti-TNF agents Multivariate Cox analysis HR 95% CI p HR 95% CI p Age at diagnosisa 1.02 1.01–1.04 0.010 Age at diagnosisa 1.00 0.99–1.01 0.552 Sex Women Reference Sex Women Reference Men 1.88 1.08–3.30 0.026 Men 1.38 1.00–1.91 0.049 Smoking status at Dx Never Reference Smoking status at Dx Never Reference Ex-smoker 1.09 0.59–2.01 0.784 Ex-smoker 1.34 0.93–1.94 0.122 Current smoker 0.57 0.29–1.12 0.102 Current smoker 0.75 0.51–1.11 0.146 Extent at Dx Non-extensive colitis Reference Extent at Dx Non-extensive colitis Reference Extensive colitis 2.84 1.85–4.35 2.00 × 10-6 Extensive colitis 1.96 1.51–2.55 3.69 × 10-7 rs9268877 Non-carrier [GG] Reference rs9268877 Non-carrier [GG] Reference Heterozygote [AG] 1.84 1.07–3.16 0.029 Heterozygote [AG] 1.44 1.07–1.95 0.016 Homozygote [AA] 3.09 1.74–5.49 1.16 × 10-4 Homozygote [AA] 2.00 1.44–2.79 4.20 × 10-5 Thiopurines Corticosteroids Age at diagnosisa 0.99 0.99–1.00 0.050 Age at diagnosisa 0.99 0.99–1.00 6.96 × 10-4 Sex Women Reference Sex Women Reference Men 1.36 1.09–1.69 0.006 Men 1.04 0.90–1.20 0.618 Smoking status at Dx Never Reference Smoking status at Dx Never Reference Ex-smoker 1.20 0.93–1.56 0.168 Ex-smoker 1.12 0.93–1.34 0.226 Current smoker 0.82 0.63–1.06 0.127 Current smoker 0.97 0.81–1.15 0.703 Extent at Dx Non-extensive colitis Reference Extent at Dx Non-extensive colitis Reference Extensive colitis 2.14 1.78–2.56 1.79 × 10-16 Extensive colitis 2.48 2.18–2.83 1.06 × 10-43 rs9268877 Non-carrier [GG] Reference rs9268877 Non-carrier [GG] Reference Heterozygote [AG] 1.21 0.99–1.47 0.064 Heterozygote [AG] 1.26 1.11–1.44 5.01 × 10-4 Homozygote [AA] 1.48 1.17–1.86 0.001 Homozygote [AA] 1.22 1.04–1.44 0.017 Colectomy Multivariate Cox analysis Anti-TNF agents Multivariate Cox analysis HR 95% CI p HR 95% CI p Age at diagnosisa 1.02 1.01–1.04 0.010 Age at diagnosisa 1.00 0.99–1.01 0.552 Sex Women Reference Sex Women Reference Men 1.88 1.08–3.30 0.026 Men 1.38 1.00–1.91 0.049 Smoking status at Dx Never Reference Smoking status at Dx Never Reference Ex-smoker 1.09 0.59–2.01 0.784 Ex-smoker 1.34 0.93–1.94 0.122 Current smoker 0.57 0.29–1.12 0.102 Current smoker 0.75 0.51–1.11 0.146 Extent at Dx Non-extensive colitis Reference Extent at Dx Non-extensive colitis Reference Extensive colitis 2.84 1.85–4.35 2.00 × 10-6 Extensive colitis 1.96 1.51–2.55 3.69 × 10-7 rs9268877 Non-carrier [GG] Reference rs9268877 Non-carrier [GG] Reference Heterozygote [AG] 1.84 1.07–3.16 0.029 Heterozygote [AG] 1.44 1.07–1.95 0.016 Homozygote [AA] 3.09 1.74–5.49 1.16 × 10-4 Homozygote [AA] 2.00 1.44–2.79 4.20 × 10-5 Thiopurines Corticosteroids Age at diagnosisa 0.99 0.99–1.00 0.050 Age at diagnosisa 0.99 0.99–1.00 6.96 × 10-4 Sex Women Reference Sex Women Reference Men 1.36 1.09–1.69 0.006 Men 1.04 0.90–1.20 0.618 Smoking status at Dx Never Reference Smoking status at Dx Never Reference Ex-smoker 1.20 0.93–1.56 0.168 Ex-smoker 1.12 0.93–1.34 0.226 Current smoker 0.82 0.63–1.06 0.127 Current smoker 0.97 0.81–1.15 0.703 Extent at Dx Non-extensive colitis Reference Extent at Dx Non-extensive colitis Reference Extensive colitis 2.14 1.78–2.56 1.79 × 10-16 Extensive colitis 2.48 2.18–2.83 1.06 × 10-43 rs9268877 Non-carrier [GG] Reference rs9268877 Non-carrier [GG] Reference Heterozygote [AG] 1.21 0.99–1.47 0.064 Heterozygote [AG] 1.26 1.11–1.44 5.01 × 10-4 Homozygote [AA] 1.48 1.17–1.86 0.001 Homozygote [AA] 1.22 1.04–1.44 0.017 CI, confidence interval; Dx, diagnosis; HR, hazard ratio; TNF, tumour necrosis factor.. aPer year of increase in age. View Large Figure 3. View largeDownload slide Long-term clinical outcomes according to rs9268877 in 1961 patients with ulcerative colitis [three discovery collections and a replication collection]. The cumulative probability of colectomy and medication use in Korean patients with ulcerative colitis: [a] colectomy, [b] use of anti- tumour necrosis factor [TNF] agents, [c] use of thiopurines, and [d] use of corticosteroids. Figure 3. View largeDownload slide Long-term clinical outcomes according to rs9268877 in 1961 patients with ulcerative colitis [three discovery collections and a replication collection]. The cumulative probability of colectomy and medication use in Korean patients with ulcerative colitis: [a] colectomy, [b] use of anti- tumour necrosis factor [TNF] agents, [c] use of thiopurines, and [d] use of corticosteroids. 3.4. UC risk score analysis To determine whether susceptibility variants collectively influence prognosis, we calculated PRS for each patient using 24 UC risk loci confirmed in Korean populations [Supplementary Table S2].18,19,23 We found no significant differences in the PRS between the good- and poor-prognosis subgroups [Figure 4; and Supplementary Figure S4, available as Supplementary data at ECCO-JCC online]. However, there was a statistically significant difference in the percentage of good-prognosis and poor-prognosis patients between the lowest quartile group [Q1: bottom 25%] and the top quartile group [Q4: top 25%] of PRS [Q1: 65.4% versus 34.6% and Q4: 53.0% versus 47.0%, p = 0.04, Figure 4b]. When the HLA susceptibility SNP rs9271366 was excluded from the calculation of PRS, there was no statistically significant difference in patients with extremes of the distribution [Q1: 65.9% versus 34.1% and Q4: 59.3% versus 40.7%, p = 0.26, Figure 4c]. We then analysed the effect sizes of PRS for clinical prognosis based on 24 susceptibility loci or 23 loci excluding rs9271366 in the HLA region. Composite PRS from all 24 independent susceptibility signals was associated with prognosis in multiple linear regression analysis. After excluding the HLA susceptibility SNP rs9271366, we did not observe significant associations between prognosis and PRS [Supplementary Table S11, available as Supplementary data at ECCO-JCC online]. Figure 4. View largeDownload slide [a] Box-and-whisker plots of weighted polygenic risk scores for ulcerative colitis prognosis subgroups. Each box represents the mean and 95% confidence interval. Weights were based on β coefficients for the 24 ulcerative colitis susceptibility single nucleotide polymorphism [SNPs] identified in our previous studies18,19,23 of 538 patients [215 poor-prognosis, 323 good-prognosis] and unaffected controls [n = 4936]. [b] Polygenic risk score [PRS] was divided into four parts [quartiles 1, 2, 3, and 4]. The differences in frequencies of patients with good-prognosis and those with poor-prognosis using the extremes of the distribution of PRS [quartile 1 versus quartile 4] were analysed by chi-square test. [c] Removing rs9271366 in human leukocyte antigen [HLA] region from the calculation of PRS [23 loci]. The differences in frequencies of patients with good-prognosis and those with poor-prognosis using the extremes of the distribution of PRS [quartile 1 versus quartile 4] were analysed by chi-square test. Figure 4. View largeDownload slide [a] Box-and-whisker plots of weighted polygenic risk scores for ulcerative colitis prognosis subgroups. Each box represents the mean and 95% confidence interval. Weights were based on β coefficients for the 24 ulcerative colitis susceptibility single nucleotide polymorphism [SNPs] identified in our previous studies18,19,23 of 538 patients [215 poor-prognosis, 323 good-prognosis] and unaffected controls [n = 4936]. [b] Polygenic risk score [PRS] was divided into four parts [quartiles 1, 2, 3, and 4]. The differences in frequencies of patients with good-prognosis and those with poor-prognosis using the extremes of the distribution of PRS [quartile 1 versus quartile 4] were analysed by chi-square test. [c] Removing rs9271366 in human leukocyte antigen [HLA] region from the calculation of PRS [23 loci]. The differences in frequencies of patients with good-prognosis and those with poor-prognosis using the extremes of the distribution of PRS [quartile 1 versus quartile 4] were analysed by chi-square test. 4. Discussion By integrating fine phenotyping, including longitudinal data, with HLA imputation data, our results demonstrate that rs9268877 is associated at genome-wide significance with poor-prognosis of UC, but not with susceptibility. To reflect clinical practice, patients were classified according to the highest treatment received, based on a step-up approach. Our patients were enrolled from a registry containing detailed clinical information on medication use and colectomy, enabling this classification.12,15,32 Furthermore, because the Korean national health insurance system requires patients to meet strict criteria for reimbursement of anti-TNF agents, first-line biologic therapy is rare, supporting our classification of patients in need of anti-TNF agents as poor-prognosis.12,15–17 Regarding the disease duration criteria for good-prognosis, previous studies applied criteria of 4 or 10 years.4,8,9 We adopted 5 years, based on a report that two-thirds of the UC patients in remission for the first 5 years after UC diagnosis remained in remission at 10 years.33 Based on these clinical phenotype definitions, we found that the OR of rs9268877 progressively increased according to treatment level, and confirmed that rs9268877 affects 30-year clinical outcomes. Although the genetic associations for UC overlap more extensively among different ethnic groups than those for Crohn’s disease [CD],2,19,23,34 the clinical course of UC is generally more favourable in Asian patients than in patients of European ancestry, whereas the course of CD is comparable between Asians and Europeans.12,15,16,22,32 The HLA region shows strong evidence of an association with UC in both populations,2,35,36 but the specific HLA alleles associated with UC and their influences on UC prognosis differ between Asians and Europeans.2 HLA-DRB1*010336 was associated with pancolitis and the need for colectomy in European studies,37 but Asian UC risk allele HLA-DRB1*150235 was not associated with prognosis [Supplementary Table S4]; rs4151651 on Complement Factor B [CFB] in HLA class III, a genome-wide significant risk SNP for colectomy in Europeans,37 was monomorphic in East Asians [the 1000 Genomes Phase 3]. Another SNP associated with poor-prognosis in Europeans,6 rs17207986 on the Tenascin XB [TNXB] gene, is located ~350 kb away from rs9268877, but was not replicated in the present study. By contrast, rs9268877, identified as a prognosis-associated SNP in this study, was previously identified as a UC susceptibility SNP for Europeans.36,38,39 Recent studies in CD have shown that the genetic contribution to prognosis is independent of the contribution to disease susceptibility.8,9,40 Previous studies focused on susceptibility genes to identify prognosis-associated variants in UC, but failed to demonstrate consistent associations.3,4 Our study in Korean UC patients also suggest that the genetic burden of susceptibility variants might not be associated with disease course. Our analysis, based on the extremes of the distribution of genetic burden [the first versus the fourth quartile of PRS], showed no significant differences in percentage of patients with good- and poor-prognosis after excluding the susceptibility HLA SNP. The significant correlation between the PRS and clinical prognosis also disappeared after excluding the susceptibility SNP in HLA region. Therefore, we cannot exclude the possible role of susceptibility variants in the HLA region in UC prognosis. Considering the complex linkage disequilibrium patterns of the HLA region and our modest cohort sizes, further research should be carried out to test if susceptibility and prognosis in UC are determined by the same or different HLA genes. However, our results, together with those of previous studies,3,4,8,9,40 suggest that the genetic architecture associated with UC prognosis in non-HLA region is distinct from that of disease susceptibility. Therefore, a great deal remains to be understood about the genetic contribution to prognosis. We speculate that a hypothesis-free approach using a GWAS chip might be a more informative strategy for identifying prognosis-associated variants than the Immunochip array, which was designed for dense genotyping of 186 loci associated with 12 immune-mediated diseases. This study must be interpreted against the background of its potential limitations. First, our cohort was recruited at a single tertiary referral centre that often handles more severe and refractory cases. Although a single-centre study has an advantage over multicentre studies as a result of more consistent assessment of clinical phenotypes and more reliable longitudinal data, our results need to be replicated in other populations. Second, we only performed in silico analysis of rs9268877, warranting further studies on the biological mechanism underlying the influence of rs9268877 on the prognosis of UC. Third, despite our efforts to address the issue of clinical and genetic heterogeneity, further studies are required to incorporate additional factors, including environmental factors, that might influence the disease course. Fourth, the performance of the rs9268877 ‘A’ allele in sensitivity and specificity might be overestimated, given that it was based on a cohort that included the discovery samples. Finally, our results are based on a step-up approach; consequently, in the era of rapid expansion of the therapeutic armamentarium, emerging therapeutic options might influence the association between rs9268877 and prognosis. In conclusion, we identified an intergenic variant between HLA-DRA and HLA-DRB that is associated with UC prognosis, but not with susceptibility, and confirmed its clinical impact on 30-year outcomes. This study thus provides new insights into the prognosis of UC, which might be useful for establishing therapeutic strategies and personalised treatment. Supplementary Data Supplementary data are available at ECCO-JCC online. Funding This work was supported by a Korean Health Technology R&D Project grant from the Korea Health Industry Development Institute to S-KY [A120176], funded by the Ministry of Health & Welfare, as well as a Mid-Career Researcher Program grant from the National Research Foundation of Korea to KS [2017R1A2A1A05001119], funded by the Ministry of Science, Information & Communication Technology, and Future Planning of the Republic of Korea. DPBM and TH are supported by NIH/NIDDK grants P01 DK046763 and U01 DK062413, and the Leona M and Harry B Helmsley Charitable Trust. JJL is supported by the Agency for Science, Technology and Research [A*STAR], Singapore. Conflict of Interest S-KY received a research grant from Janssen Korea Ltd.; however, this grant was not related to the topic of the study. The remaining authors have no conflicts of interest to declare. Author Contributions KS and SKY obtained financial support and conceived of the study. HSL and SKY designed the study, assessed clinical phenotyping, and performed data analyses. KS supervised genotyping, data analysis, and interpretation. SKY supervised all sample collection, data analysis, and interpretation. SKY, BDY, SHP, HSL, KMK, SHO, CSY, and YSY recruited subjects and participated in diagnostic evaluation. JL provided critical comments on the study design. TH and DPM supervised the Immunochip genotyping. HSL, MH, SJ, BMK, CHL, and BH performed data analyses. MH, SJ, and JWM prepared DNA samples and performed genotyping. HSL, SKY, and KS drafted the manuscript. HSL, SKY, KS, and DPM revised the manuscript. Conference presentations: American Society of Human Genetics 2017 Annual Meeting [Orlando, FL, USA]. Acknowledgments We would like to thank all participating patients and healthy donors who provided the DNA and clinical information necessary for this study. This work was supported by the PLSI supercomputing resources of the Korea Institute of Science and Technology Information. Web resources: URLs for data presented herein are as follows: HLA Korean reference panel, https://sites.google.com/site/scbaehanyang/hla_panel R, http://www.r-project.org The 1000 Genome Project, http://www.1000genomes.org UCSC Genome Browser, http://genome.ucsu.edu Ensembl, http://asia.ensembl.org GeneCards, http://www.genecards.org HaploReg, http://archive.broadinstitute.org/mammals/haploreg/haploreg.php Genotype-Tissue Expression project, http://www.gtexportal.org References 1. McGovern DP , Gardet A , Törkvist L , et al. ; NIDDK IBD Genetics Consortium . Genome-wide association identifies multiple ulcerative colitis susceptibility loci . Nat Genet 2010 ; 42 : 332 – 7 . Google Scholar CrossRef Search ADS PubMed 2. Liu JZ , van Sommeren S , Huang H , et al. ; International Multiple Sclerosis Genetics Consortium; International IBD Genetics Consortium . Association analyses identify 38 susceptibility loci for inflammatory bowel disease and highlight shared genetic risk across populations . Nat Genet 2015 ; 47 : 979 – 86 . Google Scholar CrossRef Search ADS PubMed 3. Ananthakrishnan AN , Huang H , Nguyen DD , Sauk J , Yajnik V , Xavier RJ . Differential effect of genetic burden on disease phenotypes in Crohn’s disease and ulcerative colitis: analysis of a North American cohort . Am J Gastroenterol 2014 ; 109 : 395 – 400 . Google Scholar CrossRef Search ADS PubMed 4. Kopylov U , Boucher G , Waterman M , et al. Genetic predictors of benign course of ulcerative colitis - a North American Inflammatory Bowel Disease Genetics Consortium Study . Inflamm Bowel Dis 2016 ; 22 : 2311 – 6 . Google Scholar CrossRef Search ADS PubMed 5. Yang SK . Personalizing IBD therapy: the Asian perspective . Dig Dis 2016 ; 34 : 165 – 74 . Google Scholar CrossRef Search ADS PubMed 6. Haritunians T , Taylor KD , Targan SR , et al. Genetic predictors of medically refractory ulcerative colitis . Inflamm Bowel Dis 2010 ; 16 : 1830 – 40 . Google Scholar CrossRef Search ADS PubMed 7. Asano K , Matsushita T , Umeno J , et al. A genome-wide association study identifies three new susceptibility loci for ulcerative colitis in the Japanese population . Nat Genet 2009 ; 41 : 1325 – 9 . Google Scholar CrossRef Search ADS PubMed 8. Lee JC , Espéli M , Anderson CA , et al. ; UK IBD Genetics Consortium . Human SNP links differential outcomes in inflammatory and infectious disease to a FOXO3-regulated pathway . Cell 2013 ; 155 : 57 – 69 . Google Scholar CrossRef Search ADS PubMed 9. Lee JC , Biasci D , Roberts R , et al. ; UK IBD Genetics Consortium . Genome-wide association study identifies distinct genetic contributions to prognosis and susceptibility in Crohn’s disease . Nat Genet 2017 ; 49 : 262 – 8 . Google Scholar CrossRef Search ADS PubMed 10. Yang SK , Hong M , Baek J , et al. A common missense variant in NUDT15 confers susceptibility to thiopurine-induced leukopenia . Nat Genet 2014 ; 46 : 1017 – 20 . Google Scholar CrossRef Search ADS PubMed 11. Lennard-Jones JE . Classification of inflammatory bowel disease . Scand J Gastroenterol Suppl 1989 ; 170 : 2 – 6 ; discussion 16–9. Google Scholar CrossRef Search ADS PubMed 12. Lee HS , Park SH , Yang SK , et al. Long-term prognosis of ulcerative colitis and its temporal change between 1977 and 2013: a hospital-based cohort study from Korea . J Crohns Colitis 2015 ; 9 : 147 – 55 . Google Scholar CrossRef Search ADS PubMed 13. Silverberg MS , Satsangi J , Ahmad T , et al. Toward an integrated clinical, molecular and serological classification of inflammatory bowel disease: report of a Working Party of the 2005 Montreal World Congress of Gastroenterology . Can J Gastroenterol 2005 ; 19 : 5A – 36A . Google Scholar CrossRef Search ADS PubMed 14. Panaccione R , Rutgeerts P , Sandborn WJ , Feagan B , Schreiber S , Ghosh S . Review article: treatment algorithms to maximize remission and minimize corticosteroid dependence in patients with inflammatory bowel disease . Aliment Pharmacol Ther 2008 ; 28 : 674 – 88 . Google Scholar CrossRef Search ADS PubMed 15. Park SH , Kim YM , Yang SK , et al. Clinical features and natural history of ulcerative colitis in Korea . Inflamm Bowel Dis 2007 ; 13 : 278 – 83 . Google Scholar CrossRef Search ADS PubMed 16. Lee HS , Yang SK , Soh JS , et al. Short- and long-term outcomes of acute severe ulcerative colitis in Korea: the 1999-2005 cohort . Inflamm Bowel Dis 2015 ; 21 : 1825 – 31 . Google Scholar CrossRef Search ADS PubMed 17. Wei SC . Differences in the public medical insurance systems for inflammatory bowel disease treatment in Asian countries . Intest Res 2016 ; 14 : 218 – 23 . Google Scholar CrossRef Search ADS PubMed 18. Ye BD , Choi H , Hong M , et al. Identification of ten additional susceptibility loci for ulcerative colitis through immunochip analysis in Koreans . Inflamm Bowel Dis 2016 ; 22 : 13 – 9 . Google Scholar CrossRef Search ADS PubMed 19. Yang SK , Hong M , Oh H , et al. Identification of loci at 1q21 and 16q23 that affect susceptibility to inflammatory bowel disease in Koreans . Gastroenterology 2016 ; 151 : 1096 – 9.e4 . Google Scholar CrossRef Search ADS PubMed 20. Jia X , Han B , Onengut-Gumuscu S , et al. Imputing amino acid polymorphisms in human leukocyte antigens . PLoS One 2013 ; 8 : e64683 . Google Scholar CrossRef Search ADS PubMed 21. Kim K , Bang SY , Lee HS , Bae SC . Construction and application of a Korean reference panel for imputing classical alleles and amino acids of human leukocyte antigen genes . PLoS One 2014 ; 9 : e112546 . Google Scholar CrossRef Search ADS PubMed 22. Bernstein CN , Ng SC , Lakatos PL , Moum B , Loftus EV Jr ; Epidemiology and Natural History Task Force of the International Organization of the Study of Inflammatory Bowel Disease . A review of mortality and surgery in ulcerative colitis: milestones of the seriousness of the disease . Inflamm Bowel Dis 2013 ; 19 : 2001 – 10 . Google Scholar PubMed 23. Yang SK , Hong M , Zhao W , et al. Genome-wide association study of ulcerative colitis in Koreans suggests extensive overlapping of genetic susceptibility with Caucasians . Inflamm Bowel Dis 2013 ; 19 : 954 – 66 . Google Scholar CrossRef Search ADS PubMed 24. Bernstein BE , Epstein CS , Shoresh N , et al. A user’s guide to the encyclopedia of DNA elements [ENCODE] . PLoS Biol 2011 ; 9 : e1001046 . Google Scholar CrossRef Search ADS PubMed 25. Ward LD , Kellis M . HaploReg: a resource for exploring chromatin states, conservation, and regulatory motif alterations within sets of genetically linked variants . Nucleic Acids Res 2012 ; 40 : D930 – 4 . Google Scholar CrossRef Search ADS PubMed 26. The GTEx Consortium . The genotype-tissue expression [GTEx] pilot analysis: multitissue gene regulation in humans . Science 2015 ; 348 : 648 – 60 . CrossRef Search ADS PubMed 27. Fehrmann RS , Jansen RC , Veldink JH , et al. Trans-eQTLs reveal that independent genetic variants associated with a complex phenotype converge on intermediate genes, with a major role for the HLA . PLoS Genet 2011 ; 7 : e1002197 . Google Scholar CrossRef Search ADS PubMed 28. Fairfax BP , Humburg P , Makino S , et al. Innate immune activity conditions the effect of regulatory variants upon monocyte gene expression . Science 2014 ; 343 : 1246949 . Google Scholar CrossRef Search ADS PubMed 29. Zeller T , Wild P , Szymczak S , et al. Genetics and beyond–the transcriptome of human monocytes and disease susceptibility . PLoS One 2010 ; 5 : e10693 . Google Scholar CrossRef Search ADS PubMed 30. Mombaerts P , Mizoguchi E , Grusby MJ , Glimcher LH , Bhan AK , Tonegawa S . Spontaneous development of inflammatory bowel disease in T cell receptor mutant mice . Cell 1993 ; 75 : 274 – 82 . Google Scholar CrossRef Search ADS PubMed 31. Janelsins BM , Lu M , Datta SK . Altered inactivation of commensal LPS due to acyloxyacyl hydrolase deficiency in colonic dendritic cells impairs mucosal Th17 immunity . Proc Natl Acad Sci U S A 2014 ; 111 : 373 – 8 . Google Scholar CrossRef Search ADS PubMed 32. Park SH , Yang SK , Park SK , et al. Long-term prognosis of Crohn’s disease and its temporal change between 1981 and 2012: a hospital-based cohort study from Korea . Inflamm Bowel Dis 2014 ; 20 : 488 – 94 . Google Scholar CrossRef Search ADS PubMed 33. Solberg IC , Lygren I , Jahnsen J , et al. ; IBSEN Study Group . Clinical course during the first 10 years of ulcerative colitis: results from a population-based inception cohort [IBSEN Study] . Scand J Gastroenterol 2009 ; 44 : 431 – 40 . Google Scholar CrossRef Search ADS PubMed 34. Yang SK , Hong M , Zhao W , et al. Genome-wide association study of Crohn’s disease in Koreans revealed three new susceptibility loci and common attributes of genetic susceptibility across ethnic populations . Gut 2014 ; 63 : 80 – 7 . Google Scholar CrossRef Search ADS PubMed 35. Okada Y , Yamazaki K , Umeno J , et al. HLA-Cw*1202-B*5201-DRB1*1502 haplotype increases risk for ulcerative colitis but reduces risk for Crohn’s disease . Gastroenterology 2011 ; 141 : 864 – 71.e1–5 . Google Scholar CrossRef Search ADS PubMed 36. Goyette P , Boucher G , Mallon D , et al. ; International Inflammatory Bowel Disease Genetics Consortium; Australia and New Zealand IBDGC; Belgium IBD Genetics Consortium; Italian Group for IBD Genetic Consortium; NIDDK Inflammatory Bowel Disease Genetics Consortium; United Kingdom IBDGC; Wellcome Trust Case Control Consortium; Quebec IBD Genetics Consortium . High-density mapping of the MHC identifies a shared role for HLA-DRB1*01:03 in inflammatory bowel diseases and heterozygous advantage in ulcerative colitis . Nat Genet 2015 ; 47 : 172 – 9 . Google Scholar CrossRef Search ADS PubMed 37. Cleynen I , Boucher G , Jostins L , et al. Inherited determinants of Crohn’s disease and ulcerative colitis phenotypes: a genetic association study . Lancet 2016 ; 387 : 156 – 67 . Google Scholar CrossRef Search ADS PubMed 38. Franke A , Balschun T , Karlsen TH , et al. ; IBSEN study group . Sequence variants in IL10, ARPC2 and multiple other loci contribute to ulcerative colitis susceptibility . Nat Genet 2008 ; 40 : 1319 – 23 . Google Scholar CrossRef Search ADS PubMed 39. Barrett JC, Lee JC, Lees CW, et al.; UK IBD Genetics Consortium. Genome-wide association study of ulcerative colitis identifies three new susceptibility loci, including the HNF4A region. Nat Genet 2009;41:1330–4. 40. Kugathasan S , Denson LA , Walters TD , et al. Prediction of complicated disease course for children newly diagnosed with Crohn’s disease: a multicentre inception cohort study . Lancet 2017 ; 389 : 1710 – 8 . Google Scholar CrossRef Search ADS PubMed Copyright © 2018 European Crohn’s and Colitis Organisation (ECCO). Published by Oxford University Press. All rights reserved. For permissions, please email: journals.permissions@oup.com This article is published and distributed under the terms of the Oxford University Press, Standard Journals Publication Model (https://academic.oup.com/journals/pages/open_access/funder_policies/chorus/standard_publication_model)
TI - An Intergenic Variant rs9268877 Between HLA-DRA and HLA-DRB Contributes to the Clinical Course and Long-term Outcome of Ulcerative Colitis
JO - Journal of Crohn's and Colitis
DO - 10.1093/ecco-jcc/jjy080
DA - 2018-09-01
UR - https://www.deepdyve.com/lp/oxford-university-press/an-intergenic-variant-rs9268877-between-hla-dra-and-hla-drb-mjPhUlSDhs
SP - 1113
EP - 1121
VL - 12
IS - 9
DP - DeepDyve
ER -