CD22-Directed CAR T Cell Single Cell Multiomic Features Associated with Immune Effector Cell-Associated Hemophagocytic Lymphohistiocytosis-like Syndrome (IEC-HS)

Introduction

In a Phase 1 trial we studied CD22-directed CAR T cell therapy (CAR22) for the treatment of patients with relapsed/refractory large B cell lymphoma (LBCL) who had progressed after CD19-directed CAR T cell therapy (CAR19). In this study of 38 patients, CAR22 demonstrated an overall objective response rate (ORR) of 68% and a complete response (CR) rate of 53% (Frank MJ, et al., Lancet 2024). Despite overall favorable toxicity, high CAR22 expansion was linked to the development IEC-HS (n=5) versus no IEC-HS (n=33) (median 2378 vs. 189 CAR22-positive cells/μL; p=0.0016). IEC-HS is associated with delayed recovery, a severe immunocompromised state, and increased risk of non-relapse morbidity and mortality. CAR22-associated IEC-HS typically occurs 1-2 weeks after cytokine release syndrome (CRS) resolves, suggesting it has unique pathophysiology. Here, we report on multiomic single-cell analyses identifying features in both CAR22 products and expanding in vivo CAR22 cells associated with IEC-HS.

Methods

Single-cell RNA transcriptome (scRNA-seq) and αßTCR sequence were assessed via single cell sequencing on the 10x Genomics platform. We tracked individual T cell clonotypes from pre-manufacture apheresis material, final CAR22 products, at day 14 (peak expansion), and at day 28 post-infusion. We also performed bulk assay for transposase-accessible chromatin using sequencing (ATAC-seq) analyses on CAR22 products.

Results

Our computational pipeline was applied to a cohort of 19 patients, representative of the study population in terms of age, sex, CAR22 cell dose, adverse events, and clinical outcomes. A total of 70 samples, including 781,747 cells and 280,449 unique TCR clonotypes, were analyzed by scRNA-seq. Pairwise differential accessibility analyses of bulk ATAC-seq of CAR+ cells from products associated with the development of IEC-HS (n=4) versus not (n=15), identified enrichment of open chromatin regions in genes associated with NF-κB signaling (RELA, NFKB1, RELB); NFAT family members crucial for T cell activation, differentiation, and cytokine production (NFATC1, NFATC2, NFATC3); and the RFX family involved in regulating MHC class II expression (RFX3, RFX4, RFX8). Covarying neighborhood analysis (CNA) identified a region of clonally expanded CD8 T cells from product to peak expansion, 14 days after infusion, associated with IEC-HS. Differential gene expression analyses of these expanded cells revealed significant upregulation of interferon-stimulated genes (ISGs) IFITM1, IFI6, and ISG15. Transcription factor activity analysis showed a marked upregulation of STAT2 activity. When comparing all CD8 T cells at peak expansion between those who developed IEC-HS and those who did not, gene set enrichment analyses (GSEA) revealed both interferon alpha and interferon gamma as the top two pathways upregulated in IEC-HS.

Conclusions

Our results identify CAR22 product attributes associated with developing IEC-HS. Specifically, we identify open chromatin regions in genes associated with NF-κB and NFAT signaling pathways, along with the upregulation of interferon-stimulated genes (ISGs) and STAT2 activity at peak expansion, suggesting a link between chromatin accessibility and interferon signaling. This analysis provides insights into the transcriptional and epigenetic mechanisms underlying the syndrome. The distinct chromatin landscape of these cells warrants further investigation into the mechanisms and causes. Our results are consistent with current concepts implicating IFNγ in HLH pathophysiology and evidence for activity of interferon-gamma antagonists. Future work is needed to confirm whether these features could provide predictor biomarkers for increased risk of IEC-HS in patients. Overall, these findings offer potential avenues for early detection and further targeted intervention, with the goal to improve the safety of CAR T cell therapies.