Document resource
Background Patients with advanced head and neck squamous cell carcinoma (HNSCC) have poor prognosis in part due to insufficient anti-tumor immune responses. Unfortunately, immune checkpoint inhibitors (ICIs) have poor or no efficacy for the majority of advanced HNSCC patients. Radiation therapy (RT) as an adjuvant treatment is a promising combination partner for ICI due to its inflammatory properties and has been shown to enhance response to ICI in the clinic. Thus far, the mechanisms by which RT improves anti-tumor immunity and mechanisms leading to treatment resistance are only partially understood.Methods To investigate which pre-treatment tumor characteristics contribute to a successful immune response to RT, we investigated The Cancer Genome Atlas (TCGA) HNSC database for gene expression biomarkers of response. Given the relevance of natural killer (NK) cells in that analysis, we employed two immune-competent mouse models of HNSCC to study NK cell responses with and without RT treatment: one which responds to RT with increased survival (responder), and one which does not (non-responder). We conducted in vitro experiments where cell lines from both models were treated +/- 10 Gray RT and rested for 48 hours prior to either single cell RNA sequencing (scRNAseq) or co-culture with syngeneic healthy mouse splenocytes followed by flow cytometry.Results TCGA analysis revealed that pre-treatment expression of a signature gene for NK cell maturation positively correlated with overall survival in RT-treated patients but not untreated patients, suggesting that the presence of mature NK cells is crucial to response to RT. In our preclinical mouse models of HNSCC, scRNAseq revealed that RT increased the expression of ligands for the NK cell activating receptor NKG2D in the responder cells. This was correlated with Irf3 transcription factor activity, which is downstream of cGAS-STING signaling. These changes were not observed in the non-responder model, suggesting a cancer cell-intrinsic mechanism of resistance to RT in that model. Co-culture experiments revealed that RT induced NK cell activation against responder cells as determined by CD107A surface expression on NK cells.Conclusions These studies support the prognostic value of mature NK cell tumor infiltration in HNSCC patients to be treated with RT, providing a potential clinical biomarker to predict response to RT/ICI combination therapy. These findings also show that RT impacts cancer cell immunogenicity, which may underly the prognostic value of NK cells. These findings will inform the design of new NK cell-focused combination therapies including RT.