Endothelial-Leukocyte Adhesion in CAR T Cell Treatment Associated Neurotoxicity

PROJECT SUMMARY This project studies the mechanism of neurologic toxicity in chimeric antigen receptor (CAR) T cell therapy. CAR T cells are genetically modified, patient derived T cells that use the CAR to recognize and destroy malignant target cells. Although CAR T therapy has shown impressive results against leukemia and lymphoma,

approximately 30-40% of patients experience neurologic side effects in the first month after receiving CD19- targeted CAR T cells. This includes cognitive disturbances, seizures, and in rare cases fatal cerebral edema. Systemic cytokine release syndrome after CAR T cell infusion is a well-established risk factor for neurotoxicity,

but the connection between systemic inflammation and brain dysfunction is poorly understood. To study the mechanisms of neurotoxicity, we have developed an immunocompetent mouse model. After treatment with high dose CD19-directed murine CAR T cells, mice develop motor and balance difficulties, as well as brain microhemorrhages. Surprisingly, we found that >10% of cortical capillaries are obstructed by white

blood cells during neurotoxicity. This was accompanied by capillary remodeling and decreased vessel coverage by pericytes. Based on these findings, we now propose the following experiments: Aim 1: What molecular mechanisms cause white blood cells to plug capillaries during neurotoxicity? We will use in vivo two-photon imaging in mice to determine which cell types cause the capillary plugging – the

mouse’s own or the transferred CAR T cells? We will then measure how CAR T cell treatment changes the expression of adhesion molecules in brain capillary endothelial cells and in leukocytes, and test whether blockade of these adhesion interactions can prevent capillary plugging and neurotoxicity. Aim 2: Is neuroendothelial-leukocyte adhesion increase in human microvessels during neurotoxicity? In

parallel to our work in mice, we will use a 3D in vitro model of human brain capillaries to measure how soluble factors in patient plasma affect adhesion molecule expression in the endothelium. We will then test whether white blood cells from CAR T cell patients with neurotoxicity have increased predilection for plugging synthetic

microvessels that mimic capillaries, and whether we can prevent this plugging by blocking adhesion molecules. Aim 3: Can the strength of cell-cell adhesion signaling separate CAR T cell efficacy from toxicity? We will use quantitative multiplex immunoprecipitation to probe protein-protein interaction networks in CAR T cells that

cause high or low neurotoxicity in mice to understand what activation states are conducive to neurotoxicity. We will then test whether knock down of adhesion molecule expression can direct CAR T cells away from a toxicity phenotype by impairing their ability to signal to other cells, and to adhere to the brain microvasculature.

This work is innovative because it combines advanced imaging, in vitro modeling techniques, and protein network analysis in a unique collaboration between neuroscience, vascular biology, and oncology. The work is significant because it addresses key safety issues in emerging cancer immunotherapy modalities.