P011 Signalling and transcriptional network propagation uncovers novel ulcerative colitis pathogenetic pathways from single-nucleotide polymorphisms
D. Modos*1, J. Brooks2,3,4,5, P. Sudhakar2,4,6, B. Verstockt6,7, B. Alexander-Dann1, A. Zoufir1, D. Fazekas4,8, S. Vermeire6,7, T. Korcsmaros2,4, A. Bender1
1University of Cambridge, Chemistry, Cambridge, UK, 2The Quadram Institute Bioscience, Gut Microbes and Health Programme, Norwich, UK, 3Norfolk and Norwich University Hospitals, Norwich Medical School, Norwich, UK, 4Earlham Institute, Norwich, UK, 5University Hospitals, Department of Gastroenterology Norfolk and Norwich, Norwich, UK, 6KU Leuven, Department of Chronic Diseases, Leuven, The Netherlands, 7University Hospitals Leuven, Department of Gastroenterology and Hepatology, Leuven, The Netherlands, 8Eötvös Loránd University, Department of Genetics, Budapest, Hungary
ulcerative colitis (UC) is a complex disease with poorly understood pathogenesis. In recent years, enormous genome-wide association studies have identified 242 single-nucleotide polymorphisms (SNPs) which cause UC susceptibility. However, their exact functions and effects remain unknown. To help discover novel pathogenic pathways in UC, we developed network biology approaches to study these SNPs in the context of their signalling and regulatory landscapes.
We used immunochip profiles of 941 UC patients and focussed on UC-associated SNPs which altered miRNA target sites or transcription factor (TF) binding sites. We identified the SNP affected proteins, and mapped them to a comprehensively curated signalling database, OmniPath (http://omnipathdb.org/), to uncover their known interactions. We run a simulation using an approach called random walks to link the effect of the SNP affected proteins to TFs. We calculated how many signals reached each TF from the SNP affected proteins in each patient. Afterwards, we connected the TFs to their target genes, using a manually curated TF-target gene dataset we developed in-house (TFlink) and the Gene Transcription Regulation Database. Following a randomised control, we kept those genes that were significantly affected in more than 50% in the analysed patients.
We found 24 genes with putative links to UC. The 24 genes linked the immune-related kinase LYN and STAT4 to the immune-based pathogenesis of UC. UC SNPs affected CSKA1, CSKA2, and PKCA kinases. These kinases regulate major parts of cellular signalling networks, indicating their key role in pathogenic rewiring. Furthermore, we identified TFs involved in myofibroblast development including MYOD1 and MEF2A and MEF2D. We also identified EPCAM and ACTN4A which are involved in the focal adhesion complex, which is regulated indirectly by LYN. The involvement of these genes suggests a defected wound healing mechanism in the colon as a key player in UC pathogenesis.
Our findings suggest that the SNPs in UC can affect, via their signalling interactions, a wide variety of cellular functions with known pathogenic relevance. The functions of the affected genes indicate the focal adhesion complex and the myofibroblast development to be involved in UC pathogenesis. The described effects suggest novel pathogenetic pathways involved in UC which may be used to illuminate potential novel therapeutic intervention points.