Background: Both genetic & environmental risk factors contribute to Crohn's disease (CD) susceptibility, & genetic biomarkers may allow for the characterization of disease activity & severity as well as response to therapeutic agents. Here, we evaluated the association of genetic polymorphisms with baseline disease characteristics & the response to ustekinumab (UST), an anti-IL12/23p40 monoclonal antibody, in the phase 3 UNITI program.
Methods: The UNITI studies assessed the safety & efficacy of UST induction & maintenance therapy in patients (pts) with moderate-severe CD who had previously failed TNF-antagonist therapy (UNITI-1) or who had previously failed conventional therapy & were largely TNF antagonist-naïve (UNITI-2). Pts that responded in the induction studies were re-randomized in IM-UNITI maintenance study. 902 pts were genotyped genome-wide on the Illumina Infinium Omni5Exome platform (UNITI-1 n=479, UNITI-2 n=423). We evaluated genetic associations with baseline disease, induction wk8 response, & maintenance wk44 response phenotypes using linear or logistic regression models. Analyses were conducted separately within each study & meta-analyzed. We performed both targeted & genome-wide analyses, where the targeted analyses evaluated associations with 185 candidate SNPs in the IL12/23 pathway or previously associated with IBD risk. Significance thresholds were set at 2.7×10–4 for candidate gene analyses & 5×10–8 for the GWAS to account for multiple testing.
Results: We did not identify any statistically significant associations with IL12/23 candidate SNPs, suggesting that pts respond equally well to UST regardless of genotype or that we did not have the power to detect these associations given our limited sample size. GWAS analyses identified two associations with response phenotypes that met statistical significance. The first was an association at a locus upstream of the TWSG1 gene on chromosome 18;these SNPs were associated with change in CDAI & remission at wk8, where pts carrying the minor allele had better response to UST. Interestingly, SNPs at this locus have been associated with TWSG1 expression within the GTEx dataset (http://www.gtexportal.org), suggesting a possible functional link. The second association was between change in CDAI at wk8 & an intergenic locus on chromosome 8 located between the SFRP1 & GOLGA7 genes, where pts carrying two copies of the minor allele had the greatest decrease in CDAI after treatment.
Conclusions: These results suggest that SNPs in the IL12/23 pathway & SNPs associated with IBD disease risk may not influence responses to UST. Additionally, the genome-wide analyses nominate new candidate genes that may be influencing UST response, although these results must be replicated in an independent cohort.