Epigenetics to predict response to treatment for Crohn’s Disease

Review by Dr. Jonathan O'Donnell (Australia)

Study Summary 

Previous work has shown that responses to biologics in Crohn’s disease (CD) are variable, but predictive biomarkers are lacking. DNA methylation is an epigenetic process where a methyl group is added to a DNA molecule, typically at the C5 position of a cytosine base. This chemical modification does not change the underlying DNA sequence but can alter gene activity by preventing transcription factors from binding, effectively "turning off" the gene. Given the reported stability of DNA methylation across previously reported IBD cohorts, as well as known interactions with host microbiome and metabolome, it is a theoretically promising biomarker.

Adults with active CD initiating adalimumab, vedolizumab, or ustekinumab in Amsterdam (discovery, n=183) and Oxford (validation, n=90) were studied. Whole-blood DNA methylation was analyzed with machine learning to identify response-associated CpGs (cytosine followed by guanine sites). Models predicted vedolizumab and ustekinumab response with areas under the curve values ( AUCs) ~0.75 in validation, outperforming clinical decision support tools (CDST). Findings suggest epigenetic predictors are stable and may improve precision medicine in CD.

Commentary 

Using epigenetics for predictive markers or for new treatments is an exciting area of research in IBD precision medicine, given the potential for post-translational modification of histone proteins with histone deacetylase (HDAC) inhibitors. This is the first large epigenome-wide association study (EWAS)-based attempt to predict biologic response in CD using blood methylation. Strengths include external validation, rigorous machine learning, and integration with transcriptomics. Weaknesses were heterogeneity of primary outcome, poor replication for adalimumab and comparing to clinical decision support tools. Generalizability is uncertain: European ancestry predominates, and epigenetic profiles vary by ethnicity, environment, and disease phenotype. Still, the data support methylation as a feasible biomarker class. Next steps include validation in larger, more diverse cohorts, integration with multi-omics and mucosal data, and prospective testing in clinical algorithms before routine adoption.

Citation

https://www.thelancet.com/journals/langas/article/PIIS2468-1253(25)00102-5/fulltext

Joustra VW, et al. Development and validation of peripheral blood DNA methylation signatures to predict response to biological therapy in adults with Crohn’s disease (EPIC-ICD): an epigenome-wide association study. Lancet Gastroenterol Hepatol 2025;10:818-30.

View past News You Can Use commentaries