Abstract Introduction: Immune effector cell-associated hemophagocytic-like syndrome (IEC-HS) is an increasingly recognized, life-threatening, hyperinflammatory complication of chimeric antigen receptor (CAR) T-cell therapy across multiple CAR constructs. The pathogenesis and optimal management of IEC-HS remains poorly understood. We previously reported the potential pathogenic role of persistent, expanded CAR and non-CAR activated T-cells in IEC-HS. To further delineate the mechanisms of IEC-HS, we conducted a comprehensive proteomic and transcriptional evaluation of patients receiving CD22 CAR T-cell therapy for relapsed/refractory B-cell acute lymphoblastic leukemia (NCT02315612). Methods: We performed an in-depth plasma proteomic evaluation using the SomaScan® platform in 10-patients that developed IEC-HS with samples collected during their initial cytokine release syndrome (CRS) and at IEC-HS onset compared to 10 control-patients that developed CRS only with samples collected at a matched timepoint corresponding to IEC-HS (median day +13). Additionally, to explore cellular transcriptional changes and tissue-level immune dysregulation, we performed single cell RNA (scRNA) sequencing on peripheral blood samples at IEC-HS onset and spatial transcriptomics using the 10x Genomics Visium platform on bone marrow biopsy samples pre- and post-CAR T-cell therapy (Day +28) from 5-patients: 3 that developed IEC-HS and 2 who experienced CRS only. Unsupervised clustering, differential abundance, and pathway analyses were performed to compare the groups. Results: Twenty-five patients that received CD22 CAR T-cells with a median age of 19-years (interquartile range: 15-23) were analyzed. Fifteen (60%) were male and 12 (48%) were non-White. All patients developed CRS (median CRS max grade: 2, range: 1-4) and 96% had a complete remission after treatment. Hierarchical clustering of the 1000-most variable proteins demonstrated the IEC-HS group at IEC-HS onset clustered together and independently of samples collected at CRS onset from the same patients. These samples from CRS onset in IEC-HS patients more closely resembled the CRS only control population. Pathway analyses comparing IEC-HS onset to the matched timepoint of CRS only patients identified a significant increase in cell proliferation pathways (Myc and E2F Targets, normalized enrichment score NES=1.6, p0.05), type II interferon responses (NES=1.7, p0.05), and glycolysis (NES=1.3) (p0.05). A similar analysis comparing the paired samples from the IEC-HS group at CRS vs IEC-HS onset identified upregulation of the same pathways during IEC-HS (p0.05). Due to the potential role of inflammasome activation in IEC-HS, associated proteins were specifically interrogated; significant increases in IL18, the IL18 receptor, and the IL1 receptor in IEC-HS compared to the other groups (Kruskal-Wallis, p0.001) were seen. scRNA sequencing identified a unique cluster of rapidly proliferating, hyperactivated T-cells producing interferon-γ (IFNγ) in IEC-HS patients supporting that activated T-cells are driving the proliferation and Type-II interferon signaling proteomic signatures. Spatial transcriptomics of the bone marrow revealed a pronounced myeloid expansion at Day +28 post-infusion compared to the pre-infusion sample in IEC-HS patients. This corresponded to marked increases in Visium spots consisting of granulocyte-macrophage progenitors (GMPs), CD14+ macrophages, and progenitor red blood cells. Gene ontology analysis of clusters corresponding to these monocyte/myeloid progenitors populations demonstrated increased G-protein receptor activation/signaling pathways (adjusted p-value 0.001) and exhibited increased transcripts associated with innate immune activation including S100A12, ferritin light chain, and IL18, consistent with the hyperinflammatory milieu typical of HLH. Conclusions: IEC-HS represents a novel hyperinflammatory toxicity distinct from CRS and characterized by a strong proliferative and glycolytic proteomic signature with increased IFNγ signaling driven by pathologic T-cell activation and expansion similar to findings observed in HLH. However, prominent innate immune activation is also observed potentially due to local myeloid cell development and activation within the bone marrow. Targeting the IFNγ-IL18 axis in these patients may provide a therapeutic benefit while preserving CAR-T efficacy.
Rocco et al. (Mon,) studied this question.