Introduction Chimeric antigen receptor T-cell therapy targeting CD22 (CAR22) shows promise for patients with relapsed / refractory B-cell acute lymphoblastic leukemia (ALL) and large B cell lymphoma (LBCL) (Fry et al. Nat Med 2018; Baird et al. Blood 2020). The overall response rate for CAR22 in adults who received prior CD19 CAR therapy is 100% and 68% in ALL and LBCL respectively (Tandem 2023) (NCT04088890). However, 33% of pediatric ALL patients who received CAR22 developed immune effector cell-associated hemophagocytic lymphohistiocytosis-like syndrome (IEC-HS) (Shah et al. JCO 2020). IEC-HS is characterized by hyperferritinemia, cytopenias, coagulopathy and end organ damage. Pediatric ALL data suggest many cytokines including IFN-γ, IL-1β, IL-18 and TNF-α are upregulated in IEC-HS (Ishii et al. JCI 2020, Lichtenstein et al. Blood 2021). Here, we report the clinical features and cytokines associated with IEC-HS after CAR22 infusion in predominantly adult subjects with ALL and LBCL. Methods Plasma was collected weekly from Day 0 to 28 for 54 patients (ALL: n = 16, LBCL: n = 38; adult: n = 51, pediatrics: n = 3) who were treated with CAR22 manufactured using an automated, closed-system Miltenyi CliniMACS Prodigy device. Three dose levels were tested (ALL: 3 x 10 5 cells/kg, LBCL dose level 1 (DL1): 1 x 10 6 cells/kg, LBCL dose level 2 (DL2): 3 x 10 6 cells/kg). Clinical characteristics associated with IEC-HS were retrospectively collected and graded according to consensus criteria (Hines et al. TCT 2023). Eighty cytokines were quantified in batch via multiplexed fluorescence immunoassay (Luminex) to generate mean fluorescent indices (MFIs). Unsupervised hierarchical clustering using Euclidean distance and ward.D2 linkage were employed to analyze cytokine dynamics. Statistical analyses were performed using R (v 4.2.2). Results 8 of 54 (15%) patients developed IEC-HS, all were ≥18 years old. 5 patients had LBCL, 3 had ALL. 5 cases were grade 2, 3 were grade 4. Median onset was D+12 (range 8 - 21). All patients had prior CRS with median onset at D+1 (range 0 - 9). The 3 ALL patients had ongoing CRS and spectral overlap with IEC-HS at time of diagnosis whereas LBCL patients' CRS resolved prior to IEC-HS onset. IEC-HS patients had significant increases in inflammatory markers (ferritin, CRP, LDH, triglycerides), coagulation markers (INR, PTT), cytopenias (hemoglobin, platelets) and transaminases (Table 1). Patients most commonly received high-dose glucocorticoids, tocilizumab and anakinra. Additional agents for grade 4 IEC-HS included anti-thymoglobulin, dasatanib and emapalumab. 5 of 8 patients developed serious infections requiring IV antimicrobials including bacteremia, invasive fungal infections and viral reactivation. 5 patients with IEC-HS died after achieving complete response: 3 due to IEC-HS and concomitant infection, 1 from t-AML and 1 lost to follow up. IEC-HS occurred more frequently at higher doses in LBCL where multiple doses were tested (DL2: 33% vs DL1: 7%, p = 0.08). Peak serum CAR22 concentration was markedly elevated in patients who developed IEC-HS compared to those who did not (median 648 vs 42 cells/µL, p<0.001). Unsupervised hierarchical clustering performed on serial cytokine measurements identified a signature of IFN-γ-related cytokines including CXCL9, CXCL10, TNFα, FasL, IL6 and IL18 as being closely related in IEC-HS (Figure 1). Most cytokines in the signature peaked at D+14, consistent with onset of clinical IEC-HS. Peak levels of nearly all cytokines in the interferon-γ cluster were significantly elevated in IEC-HS patients compared to patients who did not develop IEC-HS, and this cluster contained most cytokines with significant increases in peak levels (Figure 1). Conclusions IEC-HS is a potentially life-threatening toxicity associated with CAR22 across age and underlying malignancy. The high rates of non-relapse mortality and severe infections in this study highlight the importance of understanding the immune dysregulation underlying IEC-HS. Higher CAR22 dose and expansion were associated with IEC-HS, indicating optimal dosing as a potential preventive strategy. The observed IFN-γ cytokine signature is consistent with myeloid cell activation and Th1 T-cell skewing that likely contributes to IEC-HS pathogenesis. These cytokines may be amenable to therapeutic intervention and further study of approaches targeting JAK/STAT and IFN-γ are warranted.
Read full abstract