13. Identifying distal interactions between RUNX1 and JIA associated SNP by Chromosome Conformation Capture

Background: 17 genetic loci have now been identified to confer susceptibility to juvenile idiopathic arthritis (JIA); several of these loci harbour genes involved in the Interleukin-2 (IL2) pathway suggesting that this may be an important signalling cascade involved in JIA. It is hypothesised that these regions may act as regulatory elements that modulate gene expression. Indeed capture Hi-C data, which identifies physical DNA interactions, has shown that interactions occur between a JIA susceptibility SNP rs9979383 located in a non-coding region, and the Runt-Related Transcription Factor (RUNX1) gene a crucial transcription factor involved in the regulation of IL-2. Here we use bioinformatics analysis to prioritise the likely functional variant at this locus and chromatin conformation capture (3C) to validate suggested genotype specific interactions. Aims: To design a bioinformatics pipeline to prioritise the most likely functional candidate SNPs and to design and perform functional experiments to define the mechanisms by which these JIA associated variants contribute to disease pathogenesis. Methods: To prioritise the most likely functional candidate SNPs several bioinformatics databases were curated along with In-house generated Capture Hi-C data to assess interactions between associated SNPs and nearby IL2 pathway genes. Fragments near JIA associated SNPs showed looping at several points around the haematopoiesis master regulator gene; RUNX1, a key IL2 pathway gene. Chromosome Conformation Capture (3C) experiments were implemented to validate interactions in the selected regions. To test for a genotype specific effect nine B-lymphocyte cell lines, three of each genotype, were selected for this experiment. Results: he highest prioritised SNP in the RUNX1 gene region was identified as rs9799383 based on transcription factors, Hi-C data and histone marks. Preliminary 3C data shows some genotype specific interactions more frequently in a single fragment which, however when all cell lines are grouped they show a trend towards increased frequency at multiple locations across the RUNX1 gene region. Conclusion: The bioinformatics approach proved to be highly informative and aided the design of 3C experiments. Base on initial experiments cell lines grouped by genotype show mostly insignificant differences when compared to controls. However, one interaction in particular appears to show evidence of a 400kb looping interaction between the JIA associated SNP fragment and a fragment containing evidence of PU.1 binding. Further experiments have been designed to examine whether a similar genotype specific effect can be observed in PU.1 binding in this fragment and could be indicative of a distal regulatory effect that may influence gene expression of RUNX1.