CELL INTELLIGENCE · 肿瘤细胞治疗研究

Many patients develop poor clinical response to immune checkpoint inhibitors (ICIs), especially PD-1/PD-L1 blockade. However, transcriptomic features of chemokine receptors associated with poor response remain incompletely characterized. We analyzed publicly available single-cell RNA sequencing datasets from non-small-cell lung cancer (NSCLC) and melanoma cohorts, with additional exploratory analyses in hepatocellular carcinoma (HCC) and colorectal cancer datasets. Chemokine receptor expression on CD8+ T cells from clinical responsive and non-responsive samples to anti-PD-1 therapy was systematically profiled. Differential gene expression, cell-state scoring, pseudotime trajectory inference, and ligand-receptor interaction analysis were performed to characterize associated transcriptional states and predicted cellular interactions. CCR5 transcript expression in tumor-infiltrating CD8+ T cells was associated with lower responsiveness to PD-1/PD-L1 blockade therapy. CCR5+ CD8+ T cells exhibited transcriptional features associated with increased exhaustion, reduced stemness, and advanced differentiation. Pseudotime inference suggested progressively increased CCR5 expression along the inferred differentiation trajectory. Ligand-receptor interaction analysis further identified predicted interactions between CCR5+ CD8+ T cells and tumor-associated myeloid cells, with elevated expression of CCL3 and CCL4 observed in myeloid populations from non-responsive tumors. Together, these findings identify transcriptomic associations between CCR5+ CD8+ T cell states and poor clinical response to PD-1 blockade therapy. These observations support the CCL3/4-CCR5 axis as a candidate pathway for future spatial, functional, and experimental validation.