03 TOWARDS DISEASE INTERCEPTION: CONTRIBUTION OF POLYGENIC RISK SCORES TOWARDS AGE-DEPENDENT RISK OF IBD IN A HIGH-RISK POPULATION

Abstract

As the focus shifts to precision medicine, there is increasing prioritization of intervening before disease inception and early intervention as these will allow for disease prevention and/or prompt treatment with therapies with improved side-effect profiles and enhanced efficacy. Given this, we sought to define factors that could predict disease onset within inflammatory bowel disease (IBD) using an Ashkenazi Jewish (AJ) population who carry a 4-fold increased risk of IBD. We began with an examination of genetic factors, using polygenic risk scores (PRS). Multiplex AJ families with IBD, defined as having at least three first degree relatives affected, were enrolled. Genotyping was performed using a global screening array (GSA) custom chip; imputation was performed using 1000 Genomes. PRS were calculated using imputed IBD association statistics from Liu, J. Z. Nat Genet (2015) and GSA data from the affected cohort. PLINK was used to perform IBD association p-value and linkage disequilibrium-based pruning and thresholding as described in Khera Nat Genet (2018) using 24 models for the best separation between affected, unaffected, and controls. Major NOD2 and AJ-predominant LRRK2 (N2081D) uncommon risk alleles were analyzed. PRS were compared amongst unaffected multiplex family members by age. PRS of affected were compared to age at diagnosis using the Spearman rank correlation. A total of 88 multiplex families (N=620) and 6 control Jewish families (N=39) were enrolled with median [IQR] family size of 6.5 [4-8] individuals, including 167 affected. Median [interquartile range] age of diagnosis was 20 [13-24] years. GSA chip was performed on a preliminary 282 (77 affected, 191 unaffected, 14 controls). The most significant PRS model at separating cases, controls, and unaffected individuals used p-value < 0.0005 and R2 value of 0.4 (Figure 1). We observed no unaffected individuals above the age of 25 with a score in the top 10% (Figure 2). PRS did not correlate with age at diagnosis (rho = 0.26, p=0.1). In ~half of the families, co-segregation of NOD2 and/or LRRK2 risk alleles was observed. While uncommon risk alleles contribute substantially to familial risk, co-segregation of NOD2 and LRRK2 was observed in only ~half of the multiplex AJ families. A substantive fraction of familial risk is driven by the composite of common variation from the over 200 other IBD loci that are also present in non-AJ IBD. We observed no unaffected individuals above the age of 25 with a PRS in the top 10%, suggesting that children carrying a PRS in the top 10% in this high-risk cohort should be carefully monitored for disease inception. Further refinement of disease risk prediction will likely be achieved by analysis of less common AJ-predominant alleles in AJ case-control cohorts and by the addition of serologies and proteomics.