DOP86 Mirikizumab-induced upregulation of colonic transcripts correlates with improvements in stool frequency in a phase 2 study of patients with moderately to severely active Ulcerative Colitis
Steere, B.(1);Powell, N.(2);Higgs, R.(1);Wang, Y.C.(1);Milch, C.(1);Sandborn, W.J.(3);Sands, B.E.(4);Reinisch, W.(5);D'Haens, G.(6);Krishnan, G.(1);
(1)Eli Lilly and Company, Biomedicines, Indianapolis, United States;(2)Imperial College London, Gastroenterology, London, United Kingdom;(3)University of California San Diego, Gastroenterology, San Diego, United States;(4)Icahn School of Medicine at Mount Sinai, Gastroenterology, New York, United States;(5)Medical University of Vienna, Internal Medicine, Vienna, Austria;(6)Amsterdam UMC, Gastroenterology, Amsterdam, The Netherlands;
Background
We have previously shown that treatment with mirikizumab (miri), a p19-directed IL-23 antibody, significantly downregulates inflammatory genes associated with disease activity and upregulates genes expressing epithelial transporter proteins in colonic tissue in patients with ulcerative colitis (UC). Here we explored the correlation between the expression of colonic mucosa genes and stool frequency (SF), a symptom reflective of disease activity, during the 12-week induction period of a Phase 2 study of patients with moderately to severely active UC (NCT02589665).
Methods
Patients were randomised 1:1:1:1 to receive intravenous placebo (PBO), miri 50mg or 200mg with possibility of exposure-based dose increases, or fixed miri 600mg every 4 weeks for 12 weeks. SF was reported daily by patients and transformed on a 4-level ordinal scale [0-3] representing increased SF above their normal or healthy baseline (BL). Patient colonic biopsies (PBO N=58, miri 50mg N=52, 200mg N=51, 600mg N=54) were collected at BL and Week (W)12, and gene expression measured using an Affymetrix HTA2.0 microarray workflow. BL and W12 gene expression or SF values were pooled and associations identified based on non-parametric Kendall’s tau. Pathway analysis (Hallmark and Reactome) of correlated genes was performed using over-representation analysis. p values of enrichment were determined by hypergeometric distribution test and adjusted for multi-testing with Benjamini-Hochberg procedure. Differential gene expression after miri treatment was determined by paired t-test comparing expression levels at BL and at W12 using data from the 200mg treatment group.
Results
A total of 267 genes were correlated with SF (|tau| >0.3 and qval <0.001). Of these, 212 were positively correlated (high expression associated with high SF) and 55 were negatively correlated (high expression associated with low SF). The 212 transcripts that were positively correlated with SF were uniformly and consistently downregulated with miri treatment, while the 55 transcripts that negatively correlated with SF, were consistently upregulated with miri treatment (Table 1). Biological pathways significantly associated with the miri-responsive transcripts that correlated with SF included inflammatory response, extracellular matrix dysregulation, neutrophil degranulation and cytokine signaling pathways, especially TNF and IL6 pathways (Table 2).
Conclusion
This is the first study to identify colon-based transcripts that correlate with a clinical disease activity measure, stool frequency, and it demonstrates that treatment with miri may upregulate genes associated with normalization of SF and down regulate genes associated with inflammation in colonic tissue samples of patients with UC.
