OP28 Gene expression (GE) values in a phase 2 trial of mirikizumab in ulcerative colitis (UC) correlate better with histopathology (HP) than endoscopy (EN) and Mayo scores. (15th January 2020)

• Record Type: Journal Article
• Authors: Steere, B; GOTTLIEB, K; Schmitz, J; Higgs, R; Jia, B; Milch, C; Tuttle, J; Krishnan, V; Belin, R; Reinisch, W
• Is Part Of: Journal of Crohn's and colitis. Volume 14(2020)Supplement 1
• Page Start: S025
• Page End: S026
• Publication Date: 2020-01-15
• Subjects: Inflammatory bowel diseases -- Periodicals; 616.344005
• DOI: 10.1093/ecco-jcc/jjz203.027
• Languages: English
• ISSNs: 1873-9946

Abstract

Abstract: Background: HP is considered integral to the definition of mucosal healing of UC. The regulated transcription of genes determines cell identity and function, and, consequentially, HP and EN appearance. Here we correlate GE in a UC clinical trial with EN and HP indices, and the modified Mayo score (MMS). Methods: 249 patients. with moderately-to-severely active UC were randomised to placebo or different doses of mirikizumab, a p19-directed IL-23 antibody (NCT02589665). Gene expression in baseline biopsy samples from the most affected area in the colon was evaluated using Affymetrix HTA2.0 exon-format microarray with a typical Bioconductor workflow (DOI: 10.1186/gb-2004-5-10-r801). Genes were represented by their largest groups of highly correlated exon groups. Expression values were correlated with Robart's Histopathology Index (RHI), Geboes score (GS), Ulcerative Colitis Endoscopic Index of Severity (UCEIS), and modified Mayo score (MMS) at baseline. A threshold of false-discovery rate (FDR)-adjusted p -value ≤0.05 was applied to the significance of the r -values. Subsets of genes of interest were subjected to pathway enrichment using MetaCore (Clarivate Analytics). Simulation analysis (SIM) was performed to estimate any bias in correlation estimates caused by the number of levels used for different clinical scores. Results: The plot (Figure 1) shows more genes with expression values that correlate significantly ( q < 0.05) with RHI (4647), than UCEIS (1232), or MMS (869) (refer to legend). There are 498 genes that intersect all scores, representing predominantly the pathways/networks in Table 1. The RHI covers 86 % (1064/1232) of the genes that correlate with UCEIS. The genes unique to the UCEIS (168) do not significantly map to a specific pathway/network. Highest absolute r-values were RHI 0.56, GS 0.45, UCEIS 0.4 and MMS 0.3. The corresponding lowest q values expressed in −Log10 are 15, 9, 4, and 2.5. SIM shows only a minimal bias of <0.03 units (Pearson's and Spearman's r ). Conclusion: In our analysis, RHI tracked baseline gene expression values best, followed by another HP index (GS), followed by EN (UCEIS) and finally MMS. Judging by the orders-of-magnitude greater significance of correlation with gene expression values, the histology indices are particularly information-rich. Regulated gene expression, fundamental to a phenotypic appearance in HP, seems to best mirrored by the RHI, followed by GS. The genes that track with all indices predominantly modulate cell adhesion and leukocyte trafficking. The genes that are unique to the UCEIS, compared with the RHI, do not significantly belong to any recognisable pathway or network. Further evaluation of what current EN scores contribute relative to HP disease activity assessment in UC may be useful.