Identification of human tumor-unique signaling networks between antigen-presenting cells and T cells using multi-omic single cell analysis

Tissue residence and prolonged exposure to inflammation profoundly affect lymphocyte and myeloid cell function. Immune cells derived solid tumors are thought to undergo additional adaptation, and the identification of such tumor-driven unique immune cell alterations might allow for more efficient and precise tumor therapies. We used a multi-omic single-cell analysis approach to define the immune landscape of human oral squamous cell carcinomas (OSCC) and inflamed non-malignant oral tissues to identify tumor-unique immune cell interactions. The immune infiltrate in non-malignant, inflamed tissues showed substantial phenotypic congruence with immune phenotypes that are typically tumor-associated. Yet, computational and machine learning analyses of single-cell data allowed identification of putative tumor-unique subsets and interactions of regulatory T cells (Tregs) and antigen-presenting cells (APCs). Subsequent experimental validation confirmed tumor-unique chemokine and cytokine signaling axes between APCs and Tregs. We identified ICOS and IL-1R1 as uniquely co-expressed biomarkers on tumor Tregs that allow for specific therapeutic targeting. Immune cells were isolated from inflamed oral mucosa, oral squamous cell carcinomas, or peripheral blood. These cells were then process for transcriptional profiling using both targeted and whole transcriptome approaches (Rhapsody and 10x, respectivly)