Open AccessContribution of a haplotype in the HLA region to anti–cyclic citrullinated peptide antibody positivity in rheumatoid arthritis, independently of HLA–DRB1
Author(s) -
Okada Yukinori,
Yamada Ryo,
Suzuki Akari,
Kochi Yuta,
Shimane Kenichi,
Myouzen Keiko,
Kubo Michiaki,
Nakamura Yusuke,
Yamamoto Kazuhiko
Publication year - 2009
Publication title -
arthritis & rheumatism
Language(s) - English
Resource type - Journals
eISSN - 1529-0131
pISSN - 0004-3591
DOI - 10.1002/art.24939
Subject(s) - haplotype , single nucleotide polymorphism , allele , human leukocyte antigen , genetics , logistic regression , biology , immunology , genotype , medicine , gene , antigen
Objective To examine the risk of anti–cyclic citrullinated peptide (anti‐CCP) antibody positivity in rheumatoid arthritis (RA) patients carrying certain haplotypes in the HLA region. Methods A total of 1,389 Japanese patients with RA were genotyped for 30 single‐nucleotide polymorphisms (SNPs) in the HLA region using commercial oligonucleotide arrays (from Perlegen or Affymetrix) as well as for HLA–DRB1 alleles using a sequence‐specific polymerase chain reaction method. Stepwise logistic regression was used to select from among the 30 SNPs the ones that represented a risk of anti‐CCP antibody positivity. Haplotypes of the selected SNPs were inferred using an expectation‐maximization algorithm. Associations of individual SNPs were evaluated with the Cochran‐Armitage test for trend. DRB1 alleles and haplotypes were evaluated with the chi‐square test. Heterogeneities of risks among the shared epitope (SE) and non‐SE HLA–DRB1 alleles were examined using the exact test. Haplotype associations that were independent of individual HLA–DRB1 alleles were evaluated using the likelihood ratio test. Results Significant associations were found for 9 SNPs (smallest P value being 2.4 × 10 −8 ) and in 4 HLA–DRB1 alleles (smallest P value being 2.0 × 10 −10 in DRB1*0405). Stepwise logistic regression selected 4 SNPs (rs9262638, rs7775228, rs4713580, and rs9277359). Among the 16 inferred haplotypes of these 4 SNPs, 6 indicated significant associations (smallest P value being 1.9 × 10 −11 ). Risks among SE and non‐SE alleles were significantly heterogeneous ( P = 0.0095 and P = 9.8 × 10 −9 , respectively), indicating the importance of stratification with individual DRB1 alleles rather than SE alleles. Conditional analysis of the risk associated with individual DRB1 alleles identified a risk haplotype that was independent of DRB1 (odds ratio 2.00 [95% confidence interval 1.44–2.79], P = 2.6 × 10 −5 ). Conclusion Heterogeneous risks of anti‐CCP antibody positivity were confirmed among SE and non‐SE alleles in our patient population. A risk haplotype in the HLA region that is independent of HLA–DRB1 was confirmed.