Exploring potential biomarkers and therapeutic targets in inflammatory bowel disease: insights from a mega-analysis approach

Abstract

Background: Understanding the molecular pathogenesis of inflammatory bowel disease (IBD) has led to the discovery of new therapeutic targets that are more specific and effective. Our aim was to explore the molecular pathways and genes involved in IBD pathogenesis and to identify new therapeutic targets and novel biomarkers that can aid in the diagnosis of the disease.

Methods: To obtain the largest possible number of samples and analyze them comprehensively, we used a mega-analysis approach. This involved reprocessing raw data from multiple studies and analyzing them using bioinformatic and machine learning techniques.

Results: We analyzed a total of 697 intestinal biopsies of Ulcerative Colitis (n = 386), Crohn's disease (n = 183) and non-IBD controls (n = 128). A machine learning analysis detected 34 genes whose collective expression effectively distinguishes inflamed biopsies of IBD patients from non-IBD control samples. Most of these genes were upregulated in IBD. Notably, among these genes, three novel lncRNAs have emerged as potential contributors to IBD development: ENSG00000285744, ENSG00000287626, and MIR4435-2HG. Furthermore, by examining the expression of 29 genes, among the 34, in blood samples from IBD patients, we detected a significant upregulation of 12 genes (p-value < 0.01), underscoring their potential utility as non-invasive diagnostic biomarkers. Finally, by utilizing the CMap library, we discovered potential compounds that should be explored in future studies for their therapeutic efficacy in IBD treatment.

Conclusion: Our findings contribute to the understanding of IBD pathogenesis, suggest novel biomarkers for IBD diagnosis and offer new prospects for therapeutic intervention.

Keywords: Crohn’s disease; biomarkers; inflammatory bowel disease; machine learning; mega-analysis; ulcerative colitis.