Genetic mapping with multiple levels of phenotypic information reveals new determinants of lymphocyte glucocorticoid sensitivity

Clinical response to glucocorticoids (GCs), steroid hormones widely used as pharmaceuticals, varies extensively, with many patients (~30%) showing a weak response to treatment. Although little is known about the molecular basis of this variation, regulatory polymorphisms are likely to play a key role as GCs act largely through activation of a transcription factor, the GC receptor. In an effort to characterize the molecular basis of variation in GC sensitivity, we measured in vitro lymphocyte GC sensitivity (LGS) and transcriptome-wide response to GCs in peripheral blood mononuclear cells (PBMCs) from African-American healthy donors. We found that variation in LGS was correlated with transcriptional response at 27 genes (FDR<0.1). Furthermore, a genome-wide association scan revealed a quantitative trait locus (QTL) for LGS (rs11129354, P=4x10-8) that was also associated with transcriptional response at multiple genes, including many (14 of 27) where transcriptional response was correlated with LGS. Using allelic imbalance assays, we show that this QTL is a GC-dependent cis-regulatory polymorphism for RBMS3, which encodes an RNA-binding protein known as a tumor suppressor. We found that siRNA-mediated knockdown of RBMS3 expression increases cellular proliferation in PBMCs, consistent with the role of the gene as a negative regulator of proliferation. We propose that differences in LGS reflect variation in transcriptional response, which are influenced by a GC-dependent regulatory polymorphism that acts in cis relative to RBMS3 and in trans to affect the transcriptional response of multiple distant genes. Total RNA was obtained from paired aliquots of peripheral blood mononuclear cells treated with dexamethasone and phytohemagglutinin, vehicle (EtOH) and phytohemagglutinin, or blank (no treatment) for 6 hours.