Molecular and epigenetic predictors and mechanisms of treatment response to topical steroids in eosinophilic esophagitis ABSTRACT Eosinophilic esophagitis (EoE) is a chronic immune-mediated disease defined by abnormal infiltration of eosinophils into the esophageal mucosa, leading to dysphagia, progressive esophageal stenosis, and food impaction. The incidence and prevalence are rising dramatically, and EoE has rapidly become a major cause of upper gastrointestinal morbidity. Despite this increasing importance, much is unknown about the ideal approach to treatment, and we cannot predict which patients will respond to specific treatments. Practitioners must frustratingly use trial and error, and patients are subjected to expensive therapies and invasive and costly follow-up procedures without knowing their individual likelihood of response. While swallowed/topical corticosteroids are currently a mainstay of therapy for EoE, non-response is common and can be seen in up to 50% of EoE patients. Identifying which patients are less likely to have a histologic response would allow for personalization of therapy, but no predictors have been validated for clinical use. We have begun to address this major knowledge gap, and now have promising data for two novel techniques. The first is differential correlation of gene expression, where the correlation between gene pairs is assessed between responders and non-responders. We identified a 22 gene module where there was significantly higher differential correlation in non-responders at baseline prior to treatment, and also identified potential associated miRNA pathway regulators. The second is epigenetic methylation, where we identified differential methylation at 18 CpG sites when comparing treatment responders and non-responders at baseline. Because differential correlation and methylation both imply differential regulation of biologic pathways, they can provide highly needed information about molecular mechanisms of non-response, and also may yield new treatment targets. The overall goal of the proposed study is to validate these two novel methods for prediction of treatment response in EoE and use these results to investigate molecular mechanisms of non-response The specific aims are to 1) validate a pre- treatment differential correlation gene expression module for prediction of topical steroid non-response in patients with EoE; and 2) to validate a pre-treatment epigenetic methylation profile for prediction of topical steroid non-response in patients with EoE. To achieve these aims, we will perform a secondary analysis of a rich clinical and biosample set from an NIH-funded randomized clinical trial conducted by the PI (R01 DK101856) that compared budesonide to fluticasone for initial treatment of EoE. This innovative, hypothesis- driven, and rigorously designed study will be conducted by an existing multidisciplinary team with recognized expertise in EoE, clinical trials, epidemiology, genetics, statist...