MOLECULAR AND RESTING-STATE BRAIN CONNECTIVITY LINKS TO STRESS REACTIVITY IN ULCERATIVE COLITIS

Abstract BACKGROUND Perceived stress has previously been associated with increased flare risk in ulcerative colitis (UC) patients, and we have demonstrated that a psychometrically-based Stress Reactivity (SR) phenotype is associated with increased risk of flares and altered gut microbiome composition in ulcerative colitis (UC) patients. Here, we further delineate the biological basis of increased SR by investigating associations between SR with blood transcriptomic measures of sympathetic nervous system (SNS) gene regulation and resting state brain connectivity. METHODS Blood transcriptome profiling, brain imaging, and psychosocial symptom assessments were obtained in 92 (53 High SR, 39 Low SR) biopsy-confirmed UC patients. Pro-inflammatory, Type I interferon, and Conserved Transcriptional Response to Adversity (CTRA) gene expression composites were analyzed, along with genome-wide analysis of differential gene expression and transcription factor activation patterns. Resting state connectivity analyses was focused on central autonomic network (CAN) and its connection with the emotional arousal, salience, sensorimotor, default mode networks, and with brainstem regions. RESULTS In blood cell gene regulation, high SR was associated with increased activation of the SNS response pathway CREB, the HPA-axis glucocorticoid receptor (GR) pathway, and multiple immune activation pathways (STAT, IRF2/3, PPAR-gamma, PU1, and MAF) while the canonical pro-inflammatory NF-kB pathway was down-regulated. UC SR groups showed differences in the intrinsic connectivity of the CAN, with a large effect size decrease in High compared to the Low SR group in the Left(L) parabrachial complex (PBC) to L anterior insula (aINS) connectivity (Cohen’s d=-.92,p=6.48e-.05, q = 0.02). SR group differences were also observed for connectivity measures of several other brain regions Cohens D =|0.59-0.43|). Integrative multivariate analyses revealed an intrinsic CAN connectivity signature comprised by 9 features including pairwise connectivity from emotional arousal to salience regions and default mode to PBC and salience regions. Higher scores on this signature were associated with higher inflammatory composite scores, b = 0.80(SE=.12), B=.65, t = 6.50 p = 1.77e-08, and increased CREB activity, p< .01. CONCLUSIONS To our knowledge, this is the first demonstration of a high SR phenotype with distinct brain and transcriptomic features that predict flare risk. The High SR group shows a profile of increased SNS and HPA-axis gene regulation, with alterations in the intrinsic connectivity of the CAN. Furthermore, across the SR groups, higher scores on a multivariate CAN signature associated with parasympathetic and sympathetic activity are linked to higher inflammatory composite scores, a consequence of greater SNS modulation of gene expression.