Objective: To identify ankylosing spondylitis (AS)-associated copy number variations (CNVs) in Korean subjects and their synergistic roles in the development of AS.
Methods: A genome-wide association study (GWAS) was performed in 309 patients with AS and 309 control subjects, using a copy number variant (CNV) microarray. AS-associated CNV regions were replicated in 2 independent sets (625 patients and 891 control subjects) by quantitative polymerase chain reaction (PCR) and deletion-typing PCR.
Results: In the CNV GWAS, 227 CNV regions were shown to be significantly associated with the risk of AS. Of the candidate CNV regions, 9 were successfully replicated in the first replication analysis: 1q32.2 (HHAT), 1p34.2 (BMP8A), 2q31.2 (PRKRA), 6p21.32 (HLA-DPB1), 11q22.1 (CNTN5), 13q13.1 (EEF1DP3), 14q24.2 (RGS6), 16p13.3, and 22q11.1 (IL17RA). The 5 deletion-type CNV regions, in 1q32.2, 2q31.2, 6p21.32, 13q13.1, and 16p13.3, were associated with an increased risk of AS, and the other 4 CNV regions were protective. In the second replication analysis, 4 CNV regions in 1q32.2, 2q31.2, 6p21.32, and 16p13.3 were replicated. Among patients with CNV regions in ≥4 risk-increasing loci, the risk was 18.0 times higher than that in patients without any deletions (odds ratio [OR] 17.98, P = 2.3 × 10(-7) ). Among patients with CNV regions in ≥2 protective loci, the risk was 5.2 times lower than that in those without any deletions (OR 0.19, P = 4.0 × 10(-10) ). The additive effects of simultaneous events were shown to be dependent on the frequency of CNV regions. Through deletion-typing PCR and sequencing, the exact sizes and breakpoint sequences were defined in 4 CNV regions. The mechanism of all 3 deletions was shown to be microhomology-based nonhomologous end joining.
Conclusion: The results of this study can help to identify pathogenic mechanisms of AS and can easily be applied in the development of algorithms estimating the risk of AS.
Copyright © 2014 by the American College of Rheumatology.