Single-cell dynamics of breakthrough toxicities following anakinra prophylaxis for axicabtagene ciloleucel treatment in lymphoma

Chimeric antigen receptor T cells have transformed the treatment of multiple hematologic malignancies, but are limited in application due to treatment-related toxicities. Cytokine release syndrome (CRS) and neurotoxicity (NT) are the primary toxicities associated with these therapies and are thought to be mediated by broad immune activation due to robust T-cell expansion. Current treatment of these toxicities utilize anti-IL-6 directed therapy and/or systemic glucocorticoids to mitigate the consequences of this complex inflammatory cascade. We sought to use anakinra, an IL-1R antagonist, as a prophylactic strategy to prevent clinically meaningful CRS/NT, defined as grade 2 or higher toxicity necessitating clinical intervention. Although our study, in line with others, demonstrated that IL-1R antagonism is insufficient to prevent such toxicities entirely, we were able to gain insight into the molecular immune signaling associated with breakthrough CRS & NT in the presence of anakinra prophylaxis using our scRNA dataset of patients treated with and without IL-1R antagonism. We demonstrate that IL-4 and IL-10 anti-inflammatory pathways in infused CAR-T products of both anakinra and non-anakinra cohorts were negatively associated with grade 2+ toxicities. We also found that expression of IFNg pathways and ligand-receptor activities, as well as cytokine levels of IFNg and CXCL10 in CD14+ monocytes, were significantly enriched in patients with breakthrough toxicity in the anakinra cohort. This correlated with an increase in IFNg in the peripheral blood of patients with breakthrough toxicities, among other cytokines. These data identify IFNg as a potential key mechanism in the development of CAR-T cell-associated toxcities and suggest that this pathway is also targetable and not inhibited by anakinra alone. Cryopreserved infusion product was provided by Kite Pharma. PBMCs from day 7 (± 1 day) after CAR-T cell infusion or cryopreservd infusion product were thawed and resuspended in staining buffer for flow cytometry (PBS consisting of 2% FBS). Human Fc block (5 mL) was added per sample (BD Bioscience catalog no. 564219). After 5 minutes, the following antibodies (in BD brilliant stain buffer, BD Bioscience catalog no. 566349) were added for 20 minutes at 4°C: CD45 (BV786 BD Biosciences catalog no. BD563716) at a dilution of 1 in 167, CD3 (APC BD Biosciences catalog no. 340661) at a dilution of 1 in 83.5, CD4 (V450 BD Bioscience catalog no. 651850) at a dilution of 1 in 83.5, CD8 (V500c BD Bioscience catalog no. 647458) at a dilution of 1 in 83.5, CD14 (FITC BD Bioscience catalog no. 347493) at a dilution of 1 in 83.5 and Kip1 (PE conjugated, Yescarta product provided by Kite) at a dilution of 1 in 150. After washing twice, cells were resuspended in FACS buffer (2% FBS in PBS) followed by the addition of 7-AAD (Thermo Fisher Scientific). Using a BD FACSAria sorting system, live CAR-T cells and non-CAR-T cells were acquired separately. The gating strategy was based on healthy non-CAR transduced donor PBMC and was prepared separately for each group as previously performed by Haradhvala et al. (2022). Before loading on a 10x Genomics Chromium platform, samples were resuspended in 0.04% ultrapure bovine serum albumin (BSA) in PBS at 1,000 cells per μl density. Day 7 samples and infusion product were thawed similarly and combined into groups of up to five samples per 10x channel loaded at a density of 3,000 cells per μl as previously performed by Haradhvala et al. (2022). *************************************************************** Submitter states that missing raw data are being made available for controlled access in dbGaP. ***************************************************************