Complex regulatory interactions at GDF5 shape joint morphology and osteoarthritis disease risk

Our ability to pinpoint causal variants using GWAS is dependent on understanding the dynamic epigenomic and epistatic context of each associated locus. Being the best studied skeletal locus, GDF5 associates with many diseases and has a complex cis-regulatory architecture. We interrogate GDF5 regulatory interactions and model disease variants in vitro and in vivo. For all regulatory regions we see that local epigenetic activation/repression impacts patterns of joint-specific expression and disease risk. By modeling the most cited risk variant in mice we found that it had no impact on expression, joint morphology, or disease. Yet, we identified significant epistatic expression interactions between this risk variant and others lying within regulatory regions subject to repression or activation. These findings are important lessons on how regulatory interactions and local epistasis work in the etiology of disease risk, and that assessment of individual variants of high GWAS significance need not alone be considered causal.