The HLA Alleles B*0801 and DRB1*0301 Are Associated with Dermatitis Herpetiformis in a Chinese Population
Yonghu Sun,Yan Lin,Baoqi Yang,Chuan Wang,Xi’an Fu,Fangfang Bao,Gongqi Yu,Qilin Zhang,Zhenzhen Wang,Guangjun Zhou,Xuechao Chen,Mei Wu,Guangjin Wang,Hong Liu,Jing Liu,Furen Zhang
DOI: https://doi.org/10.1016/j.jid.2015.10.057
2016-01-01
Abstract:Dermatitis herpetiformis (DH) is a rare autoimmune blistering disease linked to gluten sensitivity; DH follows a chronically relapsing and remitting course, characterized by pruritic polymorphic lesions and granular deposition of IgA in the dermal papillary tips (Bolotin and Petronic-Rosic, 2011Bolotin D. Petronic-Rosic V. Dermatitis herpetiformis: part I. Epidemiology, pathogenesis, and clinical presentation.J Am Acad Dermatol. 2011; 64: 1017-1024Abstract Full Text Full Text PDF PubMed Scopus (120) Google Scholar). Currently, DH is considered as the specific cutaneous manifestation of celiac disease (CD) because of the presence of circulating IgA antitissue, epidermal transglutaminase antibodies, and the same typical histological features of small bowel villous atrophy (Reunala, 1998Reunala T. Dermatitis herpetiformis: coeliac disease of the skin.Ann Med. 1998; 30: 416-418Crossref PubMed Scopus (74) Google Scholar). The incidence and prevalence of DH varies across ethnic groups and is more common in Caucasians (Salmi et al., 2011Salmi T.T. Hervonen K. Kautiainen H. Collin P. Reunala T. Prevalence and incidence of dermatitis herpetiformis: a 40-year prospective study from Finland.Br J Dermatol. 2011; 165: 354-359Crossref PubMed Scopus (107) Google Scholar) but rare in Asians (Zhang et al., 2012Zhang F. Yang B. Lin Y. Chen S. Zhou G. Wang G. et al.Dermatitis herpetiformis in China: a report of 22 cases.J Eur Acad Dermatol Venereol. 2012; 26: 903-907Crossref PubMed Scopus (23) Google Scholar). For the original Chinese population, only a few sporadic cases have been reported in the population of mainland China (Zhang et al., 2012Zhang F. Yang B. Lin Y. Chen S. Zhou G. Wang G. et al.Dermatitis herpetiformis in China: a report of 22 cases.J Eur Acad Dermatol Venereol. 2012; 26: 903-907Crossref PubMed Scopus (23) Google Scholar), Hong Kong, and Singapore (Ratnam, 1988Ratnam K.V. IgA dermatosis in an adult Chinese population. A 10-year study of linear IgA and dermatitis herpetiformis in Singapore.Int J Dermatol. 1988; 27: 21-24Crossref PubMed Scopus (18) Google Scholar). The heritability of DH has not been well understood, but a genetic predisposition has been suggested by the reported associations of both HLA class I and class II molecules in various populations. Case-control studies in early stage patients with DH showed that DH is strongly associated with HLA-B8, DR3, and DQw2 haplotypes (Hall et al., 1989Hall R.P. Sanders M.E. Duquesnoy R.J. Katz S.I. Shaw S. Alterations in HLA-DP and HLA-DQ antigen frequency in patients with dermatitis herpetiformis.J Invest Dermatol. 1989; 93: 501-505Abstract Full Text PDF PubMed Scopus (41) Google Scholar, Hall et al., 1996Hall M.A. Lanchbury J.S. Ciclitira P.J. HLA class II region genes and susceptibility to dermatitis herpetiformis: DPB1 and TAP2 associations are secondary to those of the DQ subregion.Eur J Immunogenet. 1996; 23: 285-296Crossref PubMed Scopus (11) Google Scholar, Katz, 1977Katz S.I. HLA-B8 and dermatitis herpetiformis in patients with IgA deposits in skin.Arch Dermatol. 1977; 113: 155-156Crossref PubMed Scopus (39) Google Scholar). As in CD in Caucasians, more than 90% of patients share the HLA-DQ2 haplotype, with the majority of the remainder expressing HLA-DQ8 (Kaukinen et al., 2002Kaukinen K. Partanen J. Mäki M. Collin P. HLA-DQ typing in the diagnosis of celiac disease.Am J Gastroenterol. 2002; 97: 695-699Crossref PubMed Google Scholar). NOD DQ8+ murine models have subsequently confirmed these associations (Marietta et al., 2004Marietta E. Black K. Camilleri M. Krause P. Rogers R.S. David C. et al.A new model for dermatitis herpetiformis that uses HLA-DQ8 transgenic NOD mice.J Clin Invest. 2004; 114: 1090-1097Crossref PubMed Scopus (124) Google Scholar). By contrast, in Asians, literature reviews of previous studies (Ohata et al., 2012Ohata C. Ishii N. Hamada T. Shimomura Y. Niizeki H. Dainichi T. et al.Distinct characteristics in Japanese dermatitis herpetiformis: a review of all 91 Japanese patients over the last 35 years.Clin Dev Immunol. 2012; 2012: 1-9Crossref Scopus (43) Google Scholar) have revealed distinct characteristics of DH in Asian from Caucasian, which may be due to the absence of the HLA-B8/DR3/DQ2 haplotype in Japan (Shibahara et al., 2002Shibahara M. Nanko H. Shimizu M. Kanda N. Kubo M. Ikeda M. et al.Dermatitis herpetiformis in Japan: an update.Dermatology. 2002; 204: 37-42Crossref PubMed Scopus (37) Google Scholar). To our knowledge, there has been no genetic investigation of Chinese patients with DH published, nor any relative HLA association reported. To establish genetic risk factors for DH in the Chinese population, we conducted a comprehensive genetic study consisting of a two-stage genome-wide variant screen and a next-generation sequence-based HLA typing of DH in the Chinese population. The discovery analysis included 24 patients with DH and 486 controls who were all recruited from the Shandong Provincial Institute of Dermatology and Venereology in China from 2006 (Supplementary Table S1 online). An additional 12 patients with DH and 244 controls were recruited in the same way for validation analysis. The clinical diagnostic criteria and the serological test of patients with DH were described in the Supplementary Methods online. The controls for both the discovery and validation analyses were clinically confirmed to lack any autoimmune diseases. The study was approved by the institutional review board at the Shandong Provincial Institute of Dermatology and Venereology. All the subjects gave written informed consent. In our patients, men seemed to be more frequently affected than women with a male-to-female ratio of 2.27:1. The mean age of DH onset was in the fourth decade of the patients and the mean duration of symptoms was approximately 44.9 months. Twenty-five percent of patients experienced gastrointestinal symptoms, mostly comprising chronic diarrhea, steatorrhea, weight loss, and nonspecific abdominal discomfort. Among these patients, only two patients received a duodenal biopsy and both specimens showed mild-to-moderate villous atrophy on histological examination (Supplementary Figure S1 online). For the serological data (Supplementary Table S1), 11 of 25 patients (44%) were found to be anti-endomysial antibody positive, with titers ranging from 1:10 to 1:320; and 10 of 17 patients (58.8%) were found to be anti-tissue transglutaminase antibody positive (cutoff value > 20 RU/ml), of whom 8 patients also showed anti-endomysial antibody positivity. For each participant, genomic DNA was extracted for genotyping analysis. We genotyped all 24 cases and 486 controls of the discovery analysis using Illumina Omni Zhonghua chips with 900,015 single nucleotide polymorphisms (SNPs). After standard sample and SNP QC filtering (Zhang et al., 2011Zhang F. Liu H. Chen S. Low H. Sun L. Cui Y. et al.Identification of two new loci at IL23R and RAB32 that influence susceptibility to leprosy.Nat Genet. 2011; 43: 1247-1251Crossref PubMed Scopus (136) Google Scholar), a total of 24 cases and 480 controls with 806,342 SNPs were included for genome-wide association analyses. Rs2187668, located at the intronic site of HLA-DQA1, was identified as the most significant association. Furthermore, a group of SNPs within the major histocompatibility complex region also showed significant associations (P < 1.0 × 10−5). To clarify the extensive associations observed within the major histocompatibility complex region, we imputed the genotyping SNPs to classical HLA alleles and corresponding amino acid polymorphisms using the reference panel published previously (Okada et al., 2014Okada Y. Kim K. Han B. Pillai N.E. Ong R.T. Saw W.Y. et al.Risk for ACPA-positive rheumatoid arthritis is driven by shared HLA amino acid polymorphisms in Asian and European populations.Hum Mol Genet. 2014; 23: 6916-6926Crossref PubMed Scopus (111) Google Scholar), which contains genotype data for 6,173 SNPs and classical alleles of HLA-A, B, C, DRB1, DQA1, DQB1, DPA1, and DPB1 at 4-digit resolution in 530 unrelated individuals of Asian descent. Two strong associations were found for HLA-B*0801 and HLA-DRB1*0301, whereas conditioning for either could not eliminate the effect of the other one. In addition, no further associations (P < 1.0 × 10−5) were discovered after conditioning for the effects of HLA-DRB1*0301 and HLA-B*0801. To validate the results of the imputation analysis, we sequenced HLA-DRB1 and HLA-B in all discovery samples as well as an additional 12 DH cases and 244 controls by using long-range PCR with previously published primers (Bentley et al., 2009Bentley G. Higuchi R. Hoglund B. Goodridge D. Sayer D. Trachtenberg E.A. et al.High-resolution, high-throughput HLA genotyping by next-generation sequencing.Tissue Antigens. 2009; 74: 393-403Crossref PubMed Scopus (157) Google Scholar). Individual libraries of each amplicon were prepared by enzymatic fragmentation, end repair, adenylation, and ligation of indexed adaptors. The 12 indexed libraries were pooled before paired-end 2 × 250 bp sequencing on an illumina Miseq platform. Sequence reads were aligned and analyzed to the HLA region using Omixon target software (www.omixon.com) with default settings. A mean coverage of ×114 was identified for exon 2 of HLA-DRB1 and exon 2+3 of HLA-B per sample across all the samples. A direct comparison between the results of imputation and sequencing analyses revealed a high concordance of 87.31% at HLA-DRB1 and 92.57% at HLA-B for the 4-digit resolution. Analysis of the validation sample and the joint analysis of sequenced samples confirmed the strong association of HLA-B*0801 (P = 6.87 × 10−15, OR = 28.67) and HLA-DRB1*0301 (P = 2.1 × 10−13, OR = 9.7) with DH (Table 1). In the patients with DH, there was a significant increase in the HLA-B*0801 allele frequency (Freq = 18.1%) and in HLA-DRB1*0301 allele frequency (25%) compared with the healthy population. Conditioning on HLA-B*0801 could not eliminate the effect of HLA-DRB1*0301 (PconB*0801 < 0.05), and vice versa.Table 1Results of association analyses for significant classical HLA allelesClassical HLA alleleDiscovery setValidation setJoint analysesF_AF_UPORF_AF_UPORF_AF_UPORHLA-B*08010.1670.0077.33E-0827.230.2080.0081.77E-0531.320.1810.0086.87E-1528.67HLA-DRB1*03010.2500.0353.92E-079.080.2500.0291.48E-0411.140.2500.0332.10E-139.70F_A represents the frequency of present alleles in cases, whereas F_U represents the frequency of present alleles in controls. P and OR, P value and corresponding odds ratio per present allele, respectively. The P values and odds ratios for the joint analysis were adjusted by the study variable. Open table in a new tab F_A represents the frequency of present alleles in cases, whereas F_U represents the frequency of present alleles in controls. P and OR, P value and corresponding odds ratio per present allele, respectively. The P values and odds ratios for the joint analysis were adjusted by the study variable. As 25% of the patients with DH suffered from gastrointestinal symptoms, we performed subgroup analysis of the patients with and without gastrointestinal symptoms on the two identified risk alleles. Interestingly, we found that HLA-DRB1*0301 seemed to be a protective factor for patients with DH against gastrointestinal symptoms with nominal significance (P = 0.03, OR = 0.128, Supplementary Table S2 online). Finally, we evaluated the effectiveness of these two risk variants (B*0801 and DRB1*0301) for risk prediction using all the sequenced samples (Figure 1). B*0801 was present in 36.1% of the patients with DH (13 of 36), but only 1.5% of the healthy controls (11 of 720). Meanwhile, the presence of DRB1*0301 has a sensitivity of 50% and specificity of 93.3% (with an area under the curve value of 0.717) for risk prediction. Inclusion of these two risk variants (presence of either) could improve the performance of risk predictions but in a moderate way, increasing the sensitivity to 61.1% and the area under the curve value to 0.776 and decreasing the specificity to 93.1%. Based on the estimated prevalence of 1/100,000, the negative predictive value and positive predictive value were almost the same when employing the two risk variants (negative predictive value = 99.9%, positive predictive value = 0.009%) compared with DRB1*14 alone (negative predictive value = 99.9%, positive predictive value = 0.008%) or HLA-B*0801 alone (negative predictive value = 99.9%, positive predictive value = 0.024%) for predicting DH. Human HLA genes are extremely polymorphic, with specific alleles determined by complex gene sequence polymorphisms with distinct affinity for antigenic peptides for an effective immune response. In our study, we confirmed two independent classical HLA alleles associated with DH, of which HLA-DRB1*0301 was associated with the avoidance of gastrointestinal symptoms in patients with DH. Although DH was currently considered as the specific cutaneous manifestation of CD, only a very small number of patients with DH were considered to have CD because only 25% of patients with DH had gastrointestinal symptoms in our data and the rarity of CD incidence in China. Epidemiologically, HLA-B*0801 and HLA-DRB1*0301 are rare in Chinese populations (Qin et al., 2011Qin P.Q. Su F. Yan W.X. Xing Z. Meng P. Chengya W. et al.Distribution of human leucocyte antigen-A, -B and -DR alleles and haplotypes at high resolution in the population from Jiangsu province of China.Int J Immunogenet. 2011; 38: 475-481Crossref PubMed Scopus (32) Google Scholar) (B*0801 = 0.88%, DRB1*0301 = 4.9%), whereas the frequency of these alleles is much higher in the Caucasians (B*0801 = 11.44%, DRB1*0301 = 12.16%). Previously reported risk alleles in the Caucasians, such as HLA-DQ2 and DQ8, showed highly genetic heterogeneity between Caucasian and Chinese populations (Supplementary Table S3 online). In our data, the most significant signals encoding the DQ2 group are DQB1*0201 (29.2% vs. 11%, P = 7.64 × 10−4) and DQA1*0501 (27.1% vs. 14.1%, P = 0.02), whereas genes encoding DQ8 showed no significance (DQA1*0301 [25% vs. 25.8%, P = 1], DQB1*0302 [6.25% vs. 7.6%, P = 1]). This might be explained by the population heterogeneity between Caucasian and Chinese populations. Although many studies have emphasized that genetic factors other than HLA play an important role in the pathogenesis of DH, we have not discovered any significant associations outside of the major histocompatibility complex region. This finding strongly supports that the pathogenic mechanism of DH is highly driven by an HLA-mediated autoimmune response. The authors state no conflicts of interest. We thank the study participants and research staffs (Paul I.W. de Bakker at University Medical Center Utrecht, Yukinori Okada at Tokyo Medical and Dental University, Soumya Raychaudhuri at Brigham and Women’s Hospital) for their contributions on the HLA imputation analyses. We also thank the Shandong Computer Science Center (National Supercomputer Center in Jinan) for their contribution to the data processing. This work was funded by grants from the Shandong Provincial Advanced Taishan Scholar Construction Project, the Science and Technology Plan of Shandong Academy of Medical Science (2014-09, 201201), and the Health and Family Planning Commission of Shandong Province (2013WS0368, 2014WS0064). Download .pdf (2.46 MB) Help with pdf files Supplementary Data