Abstract:
Neoadjuvant immunotherapy with anti-programmed death-1 (neo-antiPD1) has revolutionized perioperative treatment for improving the overall survival (OS) of patients with early-stage lung cancer. Despite its transformative impact, neo-antiPD1 provides limited benefit to nearly 50% of patients, who do not achieve a major pathologic response (non-MPR). Currently, there is an urgent need for methods to predict MPR before surgery and strategies to overcome resistance in non-MPR patients.
In this study, we performed an integrated analysis of publicly available single-cell sequencing data from 55 patients, leading to the identification of a novel subset of ADGRE5+ CD8+ T cells that expand in the tumor microenvironment (TME) of lung cancer patients achieving MPR. These ADGRE5+ CD8+ T cells exhibit stem-like properties compared to ADGRE5- CD8+ T cells. Using single-cell sequencing and CYTOF profiling of murine TME, we confirmed the presence of stem-like ADGRE5+ CD8+ T cells in mice that responded to neo-antiPD1 treatment. We further validated the significant expansion of ADGRE5+ CD8+ T cells in MPR patients from our own neo-antiPD1 cohort of 16 patients.
To explore its broader application, we constructed a pan-cancer immunotherapy cohort using RNA-seq data from 782 patients. Leveraging machine learning, we developed an ADGRE5-centered prognostic model (Tsurv), which could accurately distinguish between MPR and non-MPR patients in both TME and peripheral blood mononuclear cell (PBMC) sequencing data (AUC: 0.88). Mechanistically, using a 3D ex vivo system developed by us previously, we revealed that anti-PD1 therapy upregulated ADGRE5 expression in a STAT5-IL32-dependent manner. Moreover, ADGRE5+ CD8+ T cells demonstrated greater stemness, proliferative capacity, and cytotoxicity compared to ADGRE5- CD8+ T cells. Notably, adoptive transfer (ADT) of ADGRE5+ CD8+ T cells effectively overcame neo-antiPD1 resistance in murine models.
In summary, our research highlights a novel subset of ADGRE5+ CD8+ T cells with potential utility in non-MPR identification and as a promising candidate for ADT therapy to overcome neo-antiPD1 resistance. Additionally, the robust performance of the Tsurv model in PBMC RNA-seq data suggests its potential application as a liquid biopsy tool for real-time monitoring of neo-antiPD1 response.
