A novel co-culture system to validate results of a phase I trial testing new treatment for Barrett’s esophagus.

Abstract

e15099 Background: The role of the mesenchyme in the development of Barrett’s esophagus (BE) is largely unknown. In a Phase I trial “Tamoxifen to Treat Barrett's Metaplasia” (NCT02089386), six BE patients were treated with 20mg of oral tamoxifen daily for 12 weeks. Pre and post treatment biopsies were analyzed and showed consistent and unique gene expression changes that may be produced by the stroma. Patient-derived organoids (PDOs) are three-dimensional cultures that have emerged as tools to model BE and esophageal adenocarcinoma (EAC). Co-culturing of human esophageal myofibroblasts (HEMF) and BE PDOs is a novel model to investigate the epithelial-mesenchymal interactions in the BE development and EAC progression. Methods: Established PDOs of BE were previously grown from human endoscopic biopsy samples and characterized by whole-exome sequencing and transcriptomic analysis. Previously established HEMFs were grown for seven days until fully confluent. BE PDO lines were then added to HEMF-seeded transwells, and the co-cultures were maintained via the PDO culture media thereafter. Immunofluorescence staining was performed for E-cadherin, CDX2, and a-SMA. The co-culture system and HEMFs alone were treated with and without active metabolites of tamoxifen, 4-hydroxytamoxifen and endoxifen, for seven days. Pre and post treatment co-cultured cells and HEMFs were isolated for transcriptomic analyses. Results: Two co-cultures, HEMFs grown with a TP53 -mutated dysplastic BE PDO line and a non-dysplastic BE PDO line, were established. Immunofluorescence staining demonstrated the presence and viability of the HEMF mesenchymal cells marked by a-SMA and BE epithelial cells marked by E-cadherin and CDX2. Co-cultures remained confluent and viable while being treated with tamoxifen metabolites for seven days. Bulk RNA transcriptomic sequencing and gene set enrichment analysis (GSEA) on the co-cultures was compared to that of HEMF and BE PDO cells alone. We found similar gene expression changes in GSEA Hallmark pathways between HEMFs treated with tamoxifen metabolites and in our in vivo BE samples collected post tamoxifen treatment. The top five overlapping Hallmark pathways are epithelial-mesenchymal transition, coagulation, angiogenesis, adipogenesis, and apoptosis. Conclusions: In the case of this Phase I trial studying the effects of tamoxifen on BE, we found that it is likely the patient mesenchymal cells that respond to tamoxifen because the HEMF cells in co-culture showed similar gene expression changes. In addition, the co-culture approach we describe here can be used in the future to augment trials with other therapeutic agents in a system that allow for mechanistic dissection of cellular and molecular responses that cannot be done in vivo in patients.