P108 LONGITUDINAL MQTL STUDY IN BLOOD FROM PEDIATRIC CROHN’S DISEASE PATIENTS

Abstract

Crohn’s disease (CD) is a relapsing-remitting inflammatory disorder with flares of inflammation. We have recently shown that CD associated DNA methylation signatures in blood are a transient consequence of inflammation. Here, we investigated how genetic associations of DNA methylation change during the course of the disease, as a result of varying clinical features in pediatric CD patients. We used a subset of 164 pediatric CD patients from the RISK inception cohort. DNA was extracted from whole blood samples at two time points – at diagnosis and 1-3 years after diagnosis. Genome-wide genotyping and DNA methylation profiling was performed using the Illumina’s Multi-Ethnic Genotyping array and HumanMethylationEPIC ‘850K’ array, respectively. The effect size for each variant-CpG association (within ±500 kb; cis-methylation quantitative trait loci (mQTLs)) was estimated using linear mixed effects model implemented in GEMMA with age, gender, cell-type proportions and genetic relationship matrix as covariates. A total of ~613,000 genetic variants were tested against ~625,000 CpG sites at two time points, separately. At baseline, we identified a total of 62,970 mQTLs comprising 35,942 unique SNPs and 13,053 CpGs (Bonferroni P < 3.25 x 10-10). Among these mQTLs were SNPs previously found to be associated with CD. Of the 142 CD variants from large, meta-analyses of genome-wide association studies, 21 SNPs showed an association with a total of 39 CpGs. Notably, rs1819333 and rs1250546 depicted the strongest association with CpG sites annotated to known CD susceptibility genes, RNASET2 and RSP6KA2, and ZMIZ1, respectively. On the other hand, when we examined the plasma CRP levels, patients had elevated levels of CRP at diagnosis which were significantly lower during the time of the follow-up (P = 8.4 x 10-9). However, when we compared the mQTL effects from follow-up to baseline, we noted an extremely strong positive correlation suggesting the non-inflammatory nature of the identified CD-mQTLs (R2=0.99; P < 2.2e-16) [Figure 1]. Longitudinal stability of the relationship between genetic variation and DNA methylation despite effective treatment of the inflammation symptoms strongly suggests that these mQTL are not being driven by inflammation. Whether or not the associations are causally related to IBD remains an open question. Our findings also support the utility of using DNA methylation to refine the functional consequences of genetic variation in CD. Figure 1: Effects of all baseline significant mQTLs compared with follow-up showing perfect correlation suggessting non-inflammatory nature.