OP14 Interpreting genome-wide association studies of Inflammatory Bowel Disease through the lens of single-cell sequencing
Krzak, M.(1);Taylor, L.(1);Alegbe, T.(1);Ghouraba, M.(1);Strickland, M.(1);Przybilla, M.(1);Noell, G.(1);Iyer, V.(1);Cheam, A.X.(2);Matte, J.(1);Parkes, M.(3);McIntyre, R.(1);Raine, T.(3);Anderson, C.(1);
(1)Sanger Institute, Genomics of inflammation and immunity, Hinxton, United Kingdom;(2)Cambridge University, UK Dementia Research Institute, Cambridge, United Kingdom;(3)Addenbrooke’s Hospital, Department of Gastroenterology, Cambridge, United Kingdom;
Background
Inflammatory bowel disease (IBD) is a complex disease characterised by chronic inflammation of the digestive tract. Genome-wide association studies (GWAS) have identified 241 risk loci significantly associated with the two common forms of IBD, Crohn’s disease and ulcerative colitis. The vast majority of these risk loci reside in non-coding regions of the genome, and we only know which gene is dysregulated to increase risk of disease for a minority. This knowledge gap makes it difficult to draw insights into disease pathology and identify new candidate drug targets.
Methods
To improve biological insights from IBD GWAS, we generated single cell RNA-sequencing data from ileal biopsies ascertained from 25 CD patients with active ileal inflammation and 26 non-IBD controls. We identified 49 different cell types among the ~140K sequenced cells (Fig.1), including all major immune, enterocyte, secretory and mesenchymal populations. Our optimized single-cell dissociation protocol preserves the top of villus epithelial cells, which are inherently prone to anoikis, enabling generation of high-quality transcriptomes for the first time. 
Fig. 1
Single-cell atlas of terminal ileum biopsies from Crohn’s disease and non-IBD individuals.
Results
We identified 797 unique genes differentially expressed between CD patients and controls, with notable expression differences in stem cell, secretory, and enterocyte populations. Genes involved in antigen presentation and interferon-gamma signaling were enriched among those most frequently dysregulated cell types. In an attempt to identify which of these expression differences are likely causal of disease, rather than simply a consequence of it, we integrated results from the latest IBD GWAS to assess the extent to which genes captured disease heritability, and in which cell-types. Genes specifically expressed in Tregs, monocytes and IL10RA-negative monocyte-derived macrophages captured a significant fraction of disease heritability, strongly implicating these cell types in disease pathogenesis. We investigated which genes were driving these enrichment signals and identified candidate effector genes at many IBD risk loci. Reassuringly, many confirmed IBD effector genes known to have a role in the normal functioning of these cell types were found, including NOD2, IL18RAP, IL23R, NCF4, and IL2RA.
Conclusion
Single-cell analysis combined with IBD genetics has generated strong evidence for a causal role of novel disease mechanisms that have therapeutic potential. Further experiments are underway to validate this finding. At ECCO we will present an updated version of this analysis, including genetic mapping of gene-regulation across cell types to identify IBD effector genes and causal variants.