Imaging Activated Innate Immune Myeloid Cells in Immune Effector Cell-Associated Neurotoxicity Syndrome (ICANS) Following CAR T-Cell Therapy

Chimeric antigen receptor (CAR) T-cell therapies induce durable responses in many patients with hematological malignancies. However, these therapies are associated with unique and significant toxicities, including cytokine release syndrome (CRS) and immune effector cell-associated neurotoxicity syndrome (ICANS), with grade 3-4 ICANS occurring in up to 30% of patients. The pathophysiology of ICANS is poorly understood with evidence suggesting elevation of cytokines, notably IFNϒ and GM-CSF, resulting in innate myeloid cell hyperactivation. Blood biomarkers for ICANS show group differences but are largely non predictive. Furthermore, there are no characteristic radiographic features in patients with ICANS. Developing a non-invasive strategy for earlier detection and monitoring treatment responses is paramount and provides an opportunity for earlier intervention using approved and emerging therapies. Recently, we reported the synthesis and validation of 4-[18F]fluoro-1-naphthol ([18F]4FN), a novel redox tuned radiotracer for selective imaging of high redox potential reactive oxygen/nitrogen species (RONS) by positron emission tomography (PET), thus interrogating the activation state of innate immunity [Nature Biotechnology, 2022]. Leveraging non-invasive imaging of RONS bursts we aim to unravel underlying biological mechanisms and test novel therapeutic interventions in models of ICANS. First, we sought to identify key effector cell types and their biochemical co-correlates that could be used to stratify outcomes or interventions. To study the role of activated innate immune cells in a co-culture assay, we polarized THP-1 acute promyelocytic leukemia cells to an activated macrophage-like state, as well as autologous CAR T-cells manufactured from healthy donors and incubated with L-012, a bioluminescent reporter of high energy RONS. The kinetics and dynamic activation of the innate immune system was evaluated in a humanized B-cell acute lymphoblastic leukemia mouse model of ICANS by bioluminescence imaging. Media and peripheral blood were collected for cytokine analysis from in cellulo and in vivo experiments. Representative sections of mouse brains were evaluated by immunohistochemistry (IHC). In vitro experiments demonstrated that polarized M1 macrophage-like THP-1 cells that were activated with PMA and incubated with L-012, resulted in 3 orders of magnitude enhancement in RONS-generated bioluminescent signal. Similarly, CAR T-cell coculture assays with autologous peripheral blood mononuclear cells (PBMCs) showed a 30-fold increase in bioluminescence, which was significantly reduced upon monocyte depletion. Cytokine analysis confirmed elevation in monocyte-derived cytokines and a correlation with L-012 bioluminescence (e.g. Spearman's correlation 0.97, L-012 vs IL-1). Importantly peak bioluminescence preceded all peak cytokine response. Similarly, humanized ICANS mouse models demonstrated significantly higher central nervous system (CNS) inflammatory hotspots following CAR T-cell administration at day +1 and up to day +3 post CAR T-cell therapy. CNS enhancement was also observed in non-tumor bearing humanized mice treated with CAR T-cells. Very low luminescence was detected in non-humanized mice and was comparable to control groups. Cytokine analysis performed on peripheral blood at day +3 post CAR T-cell therapy also revealed elevations in GM-CSF, IL-1B, IL-6. IHC demonstrated infiltration of innate immune myeloid cells in the CNS compared to control groups. Finally, we validated the use of [18F]4FN in a cohort of patients enrolled in our phase I first-in-human study. First-in-human pharmacokinetic and dosimetry analysis for [18F]4FN confirmed a whole body effective dose 16 +/- 1.9 µSv/MBq, and no serious attributed adverse events were observed. Importantly [18F]4FN lacked retention in patients without CNS inflammation, a PET radiotracer property optimal for imaging inflammatory conditions within the CNS. In conclusion, RONS activity using L-012 was dependent on monocytes in co-culture assays and CNS inflammatory hotspots were observed in humanized mice following CAR T-cell therapy. These preliminary findings provide a rationale for investigating [18F]4FN PET as a possible translational biomarker of ICANS, and a validated platform where modulators of RONS respiratory bursts can be evaluated as potent therapeutics for ICANS.