TRAIL Expression in Tumor Cells Predicts and Drives Poor Response to Neoadjuvant Immunochemotherapy in Esophageal Squamous Cell Carcinoma

Abstract Background: Neoadjuvant PD-1 blockade combined with chemotherapy has shown promising antitumor efficacy in the therapy of esophageal squamous cell carcinoma (ESCC). However, effective biomarkers and potential mechanisms of resistance to this treatment are still unclear. Methods: We prospectively collected pre- and post-neoadjuvant treatment ESCC tissues from 40 patients who received PD-1blockade (camrelizumab, 200mg) and chemotherapy (albumin-paclitaxel, 260mg/m2 and carboplatin, area under the curve = 5). Based on their pathological responses to neoadjuvant immunochemotherapy (NICT), patients were classified into “responder” and “non-responder” groups. All samples were subjected to bulk RNA-sequencing and whole-exome sequencing. Of note, samples from 14 patients in both groups were subjected to single cell RNA sequencing and TCR sequencing. We characterized and validated the changes in the immunological landscape of “responder” and “non-responder” groups using multi-omics sequencing data, multiplex immunofluorescence, in vitro and in vivo experiments. External validation cohorts of ESCC treated with NICT were analyzed to validate the predictive biomarkers for NICT. Results: We found that TNF-related apoptosis-inducing ligand (TRAIL) signaling was significantly higher in pre-treatment non-responders than in responders, and was inversely correlated with T cell proliferation and TCR expansion. TRAIL was mainly secreted by tumor cells and bound to its specific receptor DR5 in T cells, thereby inhibiting the activation of T cells. Functional co-culture experiments and T cell cytotoxicity assays revealed that TRAIL+ tumor cells mediated T cell suppression and caused CD8+ T cell dysfunction. Furthermore, TRAIL inhibition could improve the efficacy of anti-PD-1 in a xenografted mouse model. In bulk RNA-seq (n=40), TRAIL expression was associated with worse disease-free survival and overall survival, however, these results were not observed in the ESCC cohort treated with upfront surgery from the TCGA database, suggesting that the predictive value of TRAIL is specific to NICT. In an external validation cohort of ESCC patients treated with NICT (n=81), multivariable Cox regression analysis showed that high TRAIL expression was an independent risk factor for unfavorable outcomes. Importantly, TRAIL expression exhibited a higher accuracy in predicting NICT response compared with PD-L1 expression in ESCC (AUC of ROC curves: 0.886 vs 0.607, respectively P<0.001). Meanwhile, the serum concentrations of TRAIL (measured by ELISA) in pre-NICT ESCC patients were also significantly higher in non-responders (n=58). Conclusions: TRAIL expression in tumor cells is negatively correlated with the response of ESCC to NICT. TRAIL expression may serve as a promising biomarker to predict NICT response and guide NICT selection in ESCC patients. Citation Format: Weixiong Yang, Fang Wang, Jianping Guo, Junchao Cai, Zekang Wang, Yao Liu, Zengli Fang, Wenfang Chen, Lixia Xu, Shuishen Zhang, Bo Zeng, Zhenguo Liu, Sui Peng, Sai-Ching Jim Yeung, Chao Cheng. TRAIL expression in tumor cells predicts and drives poor response to neoadjuvant immunochemotherapy in esophageal squamous cell carcinoma [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2024; Part 1 (Regular Abstracts); 2024 Apr 5-10; San Diego, CA. Philadelphia (PA): AACR; Cancer Res 2024;84(6_Suppl):Abstract nr 5173.