817 Identification of Biomarkers and Mechanistic Insight for Upadacitinib in Crohn’s Disease: Serum Inflammatory Mediator Analysis From the Phase 2b CELEST Study

Abstract

INTRODUCTION: We report pharmacodynamic profiling of serum inflammatory mediators in CELEST (patients) pts with the aim of understanding UPA mechanism of action in CD and linking biomarkers to intestinal inflammation, abdominal pain (AP), and stool frequency (SF). METHODS: The CELEST Ph2 study evaluated the efficacy/ safety of multiple dosing regimens of UPA vs PBO in adults with moderately to severely active CD and inadequate response or intolerance to immunomodulators and/or TNF inhibitors.1,2 Serum samples from baseline (BL) visit and end of induction period (Wk16) were analyzed by OLINK® inflammation panel (92 proteins) and Singulex immunoassay for interleukin-17A(IL-17A), IL-17F, IL-22, and IL-23p19. Protein level changes (BL–Wk16) were analyzed by an ANCOVA model where BL protein level was adjusted as a covariate and treatment effect was considered significant when a false discovery rate < 0.10 was observed after multiplicity adjustment on F-test P-values. Spearman rank correlation coefficients(rs) determined relationships between BL levels or BL–Wk16 level changes of serum biomarkers (also including serum C-reactive protein and fecal calprotectin) and improvements in Simple Endoscopic Score for CD (SES-CD), AP, and SF. RESULTS: Paired BL and Wk16 serum samples were available from 104 pts (UPA 3 mg twice daily [BID], n = 18; 6 mg BID, n = 19; 12 mg BID, n = 13; 24 mg BID, n = 18; 24 mg once daily, n = 18; and PBO, n = 18). UPA treatment significantly reduced expression of serum pro-inflammatory mediators (Table). that are associated with immune cell migration, type I/II interferon (IFN) responses, T cell activation, T helper cell-1 (Th1) responses, and CD8+ T cell responses. UPA treatment did not significantly modulate expression of IL-23/Th17-related serum cytokines in CD. Correlations were observed for BL–Wk16 changes in serum proteins and SES-CD, AP, or SF. Importantly, oncostatin-M reductions correlated with SES-CD (rs = 0.35, P = 0.02) and SF improvement (rs = 0.33, P = 0.04), suggesting similar drivers of these aspects of disease. AP improvement correlated with reductions in CSF-1 (rs = 0.37, P = 0.01), TNFRSF9 (rs = 0.34, P = 0.03), IL-12B (rs = 0.31, P = 0.07), CXCL9 (rs = 0.31, P = 0.07), and CD8A (rs = 0.31, P = 0.09), which are all associated with the Th1/CD8/IFN-γ pathway. CONCLUSION: This analysis of inflammatory mediators in serum of CD pts from CELEST identifies potential pharmacodynamic biomarkers and mechanistic insights for the JAK-1 inhibitor, UPA.