Dual Oxidase 2 (DUOX2) as a Proteomic Biomarker for Predicting Treatment Response to Chemoradiation Therapy for Locally Advanced Rectal Cancer: Using High-Throughput Proteomic Analysis and Machine Learning Algorithm

Abstract

High-throughput mass-spectrometry-based quantitative proteomic analysis was performed using formalin-fixed, paraffin-embedded (FFPE) biopsy samples obtained before treatment from 13 patients with locally advanced rectal cancer (LARC), who were treated with concurrent chemoradiation therapy (CCRT) followed by surgery. Patients were divided into complete responder (CR) and non-complete responder (nCR) groups. Immunohistochemical (IHC) staining of 79 independent FFPE tissue samples was performed to validate the predictive ability of proteomic biomarker candidates. A total of 3637 proteins were identified, and the expression of 498 proteins was confirmed at significantly different levels (differentially expressed proteins—DEPs) between two groups. In Gene Ontology enrichment analyses, DEPs enriched in biological processes in the CR group included proteins linked to cytoskeletal organization, immune response processes, and vesicle-associated protein transport processes, whereas DEPs in the nCR group were associated with biosynthesis, transcription, and translation processes. Dual oxidase 2 (DUOX2) was selected as the most predictive biomarker in machine learning algorithm analysis. Further IHC validation ultimately confirmed DUOX2 as a potential biomarker for predicting the response of nCR to CCRT. In conclusion, this study suggests that the treatment response to RT may be affected by the pre-treatment tumor microenvironment. DUOX2 is a potential biomarker for the early prediction of nCR after CCRT.