Genome-wide association analysis of psoriatic arthritis

Document Type

Conference Proceeding

Publication Date


Publication Title

Br J Dermatol


Psoriasis vulgaris (PsV) manifests at least 36 genetic susceptibility loci in the white population. About 20% of patients with PsV develop psoriatic arthritis (PsA). Nearly all patients with PsA have cutaneous psoriasis (PsC); thus the question arises, do specific genetic loci drive the development of PsA? Understanding PsA-predisposing loci could lead to prevention of and better treatment of PsA. To date, all known PsA loci (RUNX3, TRAF3IP2, REL, TNIP1, FBXL19, IL12B and HLA-C) have also been implicated in PsV. To identify genetic determinants potentially distinguishing PsA from PsC, we performed a genome-wide association study using 1430 cases of PsA and 1417 controls of European ancestry. We confirmed four previously known PsA susceptibility loci (TNIP1, IL12B, HLA-C and TRAF3IP2), and identified a new one near TYK2, a known PsV locus, at genome-wide levels of significance (P < 0.001). To increase statistical power, we performed a meta-analysis including five additional psoriasis datasets with high-density genotyping for which PsA phenotype data were available. The combined dataset included 9293 cases of PsV, 3061 cases of PsA cases, 3110 cases PsC (defined as skin lesions of psoriasis with absence of PsA for at least 10 years) and 13 670 controls. All datasets were imputed for the entire genome using the European cohort of the 1000 Genomes Project, and association analysis was performed using a logistic regression model. Analysis of all cases of PsV vs. controls confirmed genome-wide significant association at 24 of the 36 previously known loci and identified two new ones: CDKAL1 (a methythiotransferase family protein of unknown function) and CAMK2G (subunit of a serine/threonine protein kinase). The PsA vs. control analysis identified five new loci attaining genome-wide significance (IFNLR1, IL23R, IFIH1, NFKBIA and TYK2), along with five previously known PsA loci (REL, TNIP1, IL12B, HLA-C and TRAF3IP2). All 10 identified PsA loci are also known PsV loci. PsA vs. PsC analysis showed the HLA-C region to be the only significantly associated locus (P < 0.001), with human leucocyte antigen-Cw6- associated markers showing higher risk for PsC than for PsA. Conditional analysis, performed by sequential addition of the best associated signals as covariates, identified multiple independent PsA association signals at IFIH1, IL12B and HLA-C. While 10 of 12 known PsA loci associate equally with PsA and PsC, one signal in the IL23R region and three independent signals at HLA-C identified variants more strongly associated with PsA than PsC. In summary, PsA may have aetiological elements distinct from PsC, with attendant therapeutic implications.





First Page


Last Page


This document is currently not available here.