Tertiary lymphoid structure (TLS) representative cells elicit response to anti-PD-1 + therapy in a cold tumor

The composition and activation phenotype of immune cells within the tumor microenvironment (TME) can significantly influence therapeutic response to PD-1 based immunotherapy. Using the Farcast TruTumor histoculture platform, we investigated potential factors contributing to the unexpected anti-PD-1 responsiveness in an immune cold TME. Freshly resected Head and Neck Squamous Cell Carcinoma (HNSCC) samples (n=5) and matched blood were collected from consented patients. Explants were generated from the tumors and treated in culture with anti PD-1 (Nivolumab132 µg/ml) for 72 h. The response was evaluated using histopathology and flow cytometry readouts. Among the five samples analysed, one responder (S2) exhibited elevated cleaved caspase-3 expression (2.5-fold increase) in the tumor relative to the control. Interestingly, this sample (immune content <2%), exhibited a cold tumor immune microenvironment compared with the four non-responder (NR) samples (mean immune content = 4.5%). To investigate factors underlying the observed response, we evaluated the baseline immune profile. Sample S2 exhibited a more heterogenous Lo-SSC compartment compared to NR samples. This was characterized by a high proportion of CD23⁺ follicular B cells (3.9-fold increase), CD11c⁺ dendritic cells (1.5-fold increase), and DC-LAMP + expressing mature dendritic cells (2.9-fold increase) compared to NR samples. These sub-populations are indicative of immune cell subsets typically present in or associated with TLS. Although the responder sample contained a relatively low overall proportion of CD8⁺T cells, it exhibited 15 fold lower proportions of both early (CD8⁺PD-1⁺) and late (CD8⁺CD39⁺) exhausted T-cells, while displaying a 12-fold increase in stem-like (CD8⁺TCF1⁺) T cells. Upon ex vivo culture of tumor explants on TruTumor platform, the responder sample showed a distinct increase of 1.8-fold in pro-inflammatory M1 macrophages in the treated relative to the untreated control, whereas no appreciable change was observed in the non-responder samples. Interestingly, anti PD-1 treatment led to the emergence of an NKT-cell population (CD3⁺CD56⁺) and an increase in NK cells expressing Granzyme B (CD3⁻CD56⁺Granzyme B⁺). In summary, our data reveals that response to anti-PD-1 is not solely dependent on the abundance of CD8⁺ T cells but also involves substantial contributions from TLS representative immune sub-populations within the TME. Specifically, the presence of follicular B cells, stem-like CD8⁺ T cells, CD11c⁺ cells, and DC-LAMP + dendritic cells may contribute to treatment responsiveness, even in the context of an otherwise immune-poor phenotype. The TruTumor histoculture platform provides a translational model to study tumor response to immunotherapy with insights into how tumor responsiveness, transcends beyond mere presence of CD8 + T cells in the TME.