P1.106. Proteomic Signatures of Chemoresistance in Oesophageal Adenocarcinoma Cell Models

Abstract

Abstract Topic Esophageal Cancer: Molecular Biology/Pathology Background Oesophageal adenocarcinoma is a rapidly rising cancer with poor prognosis. Despite advances in surgery, chemotherapy, and targeted therapy, five-year survival remains below 20%. Treatment failure is largely driven by acquired chemoresistance, enabling tumour cells to survive under therapeutic pressure. Understanding resistance mechanisms is critical to identify vulnerabilities and improve outcomes. Methods Matched parental and acquired chemoresistant oesophageal adenocarcinoma cell models were profiled by quantitative proteomics. Cell lysates were processed by filter-aided sample preparation with tryptic digestion and analysed by data-independent acquisition LC-MS/MS. Peptides were identified and quantified using established DIA analysis pipelines with standard quality control, alignment and normalisation; thousands of proteins were consistently quantified across datasets. Protein abundance profiles were compared between resistant and parental counterparts and across resistant models generated against distinct chemotherapeutic agents to delineate shared versus agent-specific adaptations. Differential abundance testing used pre-specified thresholds for statistical significance (p < 0.05) and effect size (fold-change cut-offs), implemented in Perseus and complemented by custom R scripts for data integration, overlap analysis, and visualisation. Prioritisation focused on robustly upregulated proteins supported by reproducibility across models. Functional annotation, pathway enrichment, and protein interaction analyses were applied to identify convergent resistance programmes and nominate candidates for follow-up. Results Thousands of proteins were consistently quantified across datasets, enabling robust comparison of chemoresistant versus parental counterparts and cross-model integration. Differential abundance analysis used p < 0.05 with an initial fold-change cut-off greater than 2 to define resistance-associated changes. Resistant and parental phenotypes segregated clearly, showing remodelling of metabolic, stress-response, and structural pathways. Cross-comparison of six resistant models identified substantial heterogeneity, with many alterations unique to individual models, but also recurrently enriched proteins shared across multiple datasets, including subsets shared between models exposed to the same agent or drug class. In dose-stratified oxaliplatin-resistant models (5 μM and 10 μM), low and high exposure states displayed highly concordant proteomic responses, with conserved alterations that became more pronounced at higher exposure. More stringent filtering (fold change >4) of the shared oxaliplatin response yielded nine consistently upregulated candidates. Three of these candidates also recurred in a carboplatin-resistant model, supporting a potential platinum-class resistance signature. Conclusion Quantitative proteomic profiling of acquired chemoresistant oesophageal adenocarcinoma cell models identifies broad adaptive changes with both heterogeneous and recurrent components. Cross-model integration highlights shared adaptations, including patterns consistent with drug class–specific responses, and suggests a platinum-class component supported by overlap between oxaliplatin- and carboplatin-resistant models. The prioritised candidates provide a focused starting point for orthogonal validation in independent models and translational settings, and for mechanistic studies to define actionable vulnerabilities relevant to therapy resistance.