DOP71 TREM1, OSM and a co-expressed transcriptional module are core components of the molecular resistome to anti-cytokine therapy in Ulcerative Colitis
Saifuddin, A.(1)*;Madgwick, M.(2);Cozzetto, D.(1);Pavlidis, P.(3);Verstockt, B.(4,5);Hart, A.(1,6);Korcsmaros, T.(1);Powell, N.(1);
(1)Imperial College London, Department of Metabolism- Digestion and Reproduction, London, United Kingdom;(2)Earlham Institute, Organisms and Ecosystems, Norwich, United Kingdom;(3)King's College Hospital NHS Trust, Department of Gastroenterology, London, United Kingdom;(4)University Hospitals Leuven, Department of Gastroenterology and Hepatology, Leuven, Belgium;(5)KU Leuven, Department of Chronic Diseases and Metabolism, Leuven, United Kingdom;(6)St Mark's Academic Institute, Department of Gastroenterology, London, United Kingdom;
Background
Molecular biomarkers are emerging as promising tools to predict response to advanced therapy in IBD. For instance, mucosal expression of Triggering Receptor Expressed on Myeloid cells-1 (TREM1) and oncostatin M (OSM) are associated with anti-TNF non-response in Ulcerative Colitis (UC). The specificity of such transcriptional features is poorly understood. We investigated whether mucosal expression of TREM1 and OSM influences ustekinumab response in UC.
Methods
UNIFI is a landmark, phase III, registration trial that evaluated the efficacy of ustekinumab as induction and maintenance therapy in patients with moderate to severe UC. In a subset of patients (n=358), colonic biopsies were sampled at baseline and gene expression profiled by RNA microarray (Affymetrix HT HG-U133+ PM). Normalised expression levels of TREM1 and OSM for responders and non-responders to induction (week 8 mucosal healing) were compared using Mann-Whitney U test on GraphPad Prism 9.2.0. Co-expressed transcripts were determined with Morpheus matrix visualisation software (Broad Institute, USA). Gene set variation analysis was used to calculate gene set enrichment scores (ES) within the expression data. Multivariable models were developed to construct receiver operating characteristic curve (ROC) analyses. Deeper biological insights of TREM1- and OSM-expressing cells were explored using single cell RNA sequencing and Gene Ontology over-representation analysis (R 4.2.2).
Results
Mucosal expression of TREM1 and OSM in inflamed colonic biopsies correlated highly (r = 0.81, p<0.0001) and TREM1 was the third-highest correlate with OSM across the entire genome. Pathway analysis showed that a 20-gene module (GS1) of transcripts co-expressed with TREM1 and OSM was enriched for immune cell adhesion, migration and differentiation (Fig 1). Expression of TREM1 (median: 6.27 vs 5.55, p<0.001), OSM (5.00 vs 4.72, p=0.002) and GS1 (ES: 0.00221 vs -0.496, p<0.0001) were all significantly increased in ustekinumab non-responders compared to responders (Fig 2). Although key clinical parameters correlated weakly with TREM1 and OSM expression, they improved the performance of multivariate models to predict non-response (Fig 3). TREM1 and OSM were predominantly expressed by the same population of inflammatory monocytes accumulating in inflamed colon tissue (Fig 4).
Conclusion
Colonic expression of TREM1 and OSM is associated with resistance to both anti-TNF therapy and ustekinumab, supporting the notion that they constitute a core molecular resistome in UC patients treated with anti-cytokine agents. Their predominant co-expression in infiltrating, inflammatory monocytes indicates that treatment resistance is likely underpinned by biological pathways regulated by this cell population.