I型干扰素信号在一种新型人源化NOG-FcgammaR⁻/⁻小鼠模型中抗PD-1诱导的肿瘤排斥中的作用

Introductory sentence : To investigate the relationship between anti-PD-1 therapy and immune cells, a model that recapitulates the human tumor microenvironment is required. We previously established a humanized NOG-FcgammaR⁻/⁻ mouse model, in which anti-tumor effects by anti-human PD-1 anibody can be confirmed in vivo. Among several tested cell lines, the head and neck carcinoma line HSC4 showed the highest sensitivity in this model. In the present study, we grafted HSC4 cells into humanized NOG-FcgammaR⁻/⁻ mice and performed single-cell RNA sequencing (scRNA-seq) and spatial transcriptomic analyses to elucidate the molecular and cellular mechanisms of the tumor rejection. Experimental procedures : Human immune cells were reconstituted in NOG and NOG-FcgammaR⁻/⁻ mice by transplantation of human hematopoietic stem cells. One week after grafting HSC4 cells, the mice received intraperitoneal injections of anti-human PD-1 antibody once weekly for two weeks.For scRNA-seq, CD45⁺ cells were isolated from the grafted tumors, whereas FFPE tumor samples were used for spatial transcriptomic analysis following standard protocols. New, unpublished data : In the scRNA-seq data, human immune cells were clustered into naïve-like CD4⁺ T cells, cytotoxic CD8⁺ T cells, regulatory CD4⁺ T cells, CXCL13⁺ helper T cells, and macrophages. Pseudobulk analysis with GSEA identified the interferon-alpha (IFN-alpha) response as a top enriched pathway in nivolumab-treated NOG-FcgammaR⁻/⁻ mice. Feature plots showed that IFN-alpha-related genes were broadly expressed across human immune cell populations. To further explore this response, we analyzed the transcriptional profiles of CXCL13⁺ helper T cells, since their role in tumors remains unclear.In the spatial transcriptomic analysis, after excluding mouse-derived and low-RNA-content spots, five tumor clusters and one immune cluster containing human T cells and macrophages were identified based on marker genes and spatial context. GSEA revealed that tumor cells located near CXCL13⁺ immune cells predominantly exhibited basal-like features, along with upregulation of type I IFN-related pathways. RNAscope analysis confirmed CXCL10, an IFN response marker, in immune cells and tumor cells with basal-like feature but not in tumor cell with squamous differentiation. Conclusions : Our results suggest that tumor regression in this model is mediated by type I IFNs within the tumor microenvironment. This mechanism parallels the effects of nivolumab in human tumors, suggesting that this model provides a valuable platform for studying immune checkpoint inhibitor therapy.