Single-cell transcriptome analysis of the heterogeneous effects of differential expression of tumor PD-L1 on responding TCR-T cells

Rationale: TCR-T cell therapy plays a critical role in the treatment of malignant cancers. However, it is unclear how TCR-T cells are affected by PD-L1 molecule in the tumor environment. We performed an in-depth evaluation on how differential expressions of tumor PD-L1 can affect the functionality of T cells. Methods: We used MART-1-specific TCR-T cells (TCR-T(MART-1)), stimulated with MART-1(27-35) peptide-loaded MEL-526 tumor cells, expressing different proportions of PD-L1, to perform cellular assays and high-throughput single-cell RNA sequencing. Results: Different clusters of activated or cytotoxic TCR-T(MART-1) responded divergently when stimulated with tumor cells expressing different percentages of PD-L1 expression. Compared to control T cells, TCR-T(MART-1) were more sensitive to exhaustion, and secreted not only pro-inflammatory cytokines but also anti-inflammatory cytokines with increasing proportions of PD-L1(+) tumor cells. The gene profiles of chemokines were modified by increased expression of tumor PD-L1, which concurrently downregulated pro-inflammatory and anti-inflammatory transcription factors. Furthermore, increased expression of tumor PD-L1 showed distinct effects on different inhibitory checkpoint molecules (ICMs). In addition, there was a limited correlation between the enrichment of cell death signaling in tumor cells and T cells and increased tumor PD-L1 expression. Conclusion: Overall, though the effector functionality of TCR-T cells was suppressed by increased expression percentages of tumor PD-L1 in vitro, scRNA-seq profiles revealed that both the anti-inflammatory and pro-inflammatory responses were triggered by a higher expression of tumor PD-L1. This suggests that the sole blockade of tumor PD-L1 might inhibit not only the anti-inflammatory response but also the pro-inflammatory response in the complicated tumor microenvironment. Thus, the outcome of PD-L1 intervention may depend on the final balance among the highly dynamic and heterogeneous immune regulatory circuits.