TAMI-49. JAK STAT INHIBITION REVERSES MYELOID CELL INDUCED ANTI TUMOR IMMUNITY IN T CELLS
Abstract BACKGROUND Many central questions about the immunosuppressive microenvironment in glioblastoma (GBM) remain unanswered, particularly the interaction with lymphoid and myeloid populations. Here, we combined single-cell (scRNA) and spatial transcriptomics (stRNA) to comprehensively characterize the immune interaction with GBM. MATERIAL AND METHODS We performed scRNA-Seq of 50k CD45+ cells (8 patients) and inferred transcriptional programs and fate decisions in T cells. A novel algorithm (Nearest functionally connected neighbor) was used to predict interacting cells, further validated using spatial transcriptomics and immunofluorescence. Our findings were validated in our human neocortical glioblastoma model with autografted T cells. RESULTS Integration of st/scRNA-seq revealed a transcriptional shift of T cells towards exhaustion/hypoxia induced dysfunction. Pseudo-time analysis revealed increased Interleukin 10 (IL10) response during the Tcell transformation from the effector to the exhausted state. Using NFCN we identified a subset of HMOX1+ myeloid cells (STAT/HMOX axis), responsible for this IL10 release. Computational findings were validated using our human neocortical glioblastoma model with autografted T cells, where IL10R-inhibition/myeloid cell depletion prevented T cell exhaustion/dysfunction (p < 0.01) . In order to target the STAT3/HMOX1 axis we used a JAK/STAT inhibitor in our model which showed a drastic reduction of IL10 release (p< 0.02) and concordant activation of T cells. Clinically, one patient treated with a JAK/STAT-inhibitor in a neoadjuvant setting, 4 weeks prior to the recurrent GBM surgery, led to a significant increase (p< 0.001) in effector T cell population. CONCLUSION Our findings suggest that targeting the myeloid compartment of GBM provides an opportunity to convert a “cold” into “hot” immune environment which might be helpful to improve all T cell based therapies in the future.