CD123-CAR T Cells Manufactured in the Presence of Dasatinib Induce High Grade CRS and/or IEC-HS without Improving Efficacy in Pediatric Patients with Recurrent/Refractory Leukemia

Background

We are conducting a Phase 1 clinical study to evaluate the safety and efficacy of escalating doses of autologous CD123-CAR T cells for pediatric patients with r/r CD123-positive AML or ALL. We previously reported (Naik et al, Blood 2022, 140 (Supplement 1): 4584-5) that infusion of CD123-CAR T cells after lymphodepleting chemotherapy with fludarabine and cyclophosphamide was well tolerated, with transient fevers representing grade 1 cytokine release syndrome (CRS) and without dose limiting toxicities. We observed anti-AML activity as evidenced by responses in two out of six infused patients albeit with limited CAR T-cell expansion. Mechanistic studies revealed that CD123-CAR T cells were predominantly effector memory and expressed markers associated with T cell exhaustion (TIM3, PD1 or CD39). This phenotype was completely reversed by generating CD123-CAR T cells in the presence of dasatinib (CD123-CAR.dasa T cells), a multikinase inhibitor known to transiently inhibit CAR signaling. Based on these findings, we modified the CD123-CAR T cell production for our protocol and now report on an extended comparison of CD123-CAR and CD123-CAR.dasa T cells and on the clinical course of patients who were infused with CD123-CAR.dasa T cells.

Methods and Results

Single cell RNA seq analysis revealed that CD123-CAR.dasa T cells had upregulated IFNa signaling pathways in comparison to CD123-CAR T cells. In addition, they were in a less differentiated state as judged by a methylation-based multipotency index, exhibited preferential oxphos metabolism, and had increased apoptosis after initial antigen exposure as evidenced by caspase 3/7 activity.

Six patients (5 AML, 1 ALL) were infused with CD123-CAR.dasa T cells on dose level (DL) 3 (3x10⁶/kg) or 4 (1x10⁷/kg). All patients developed high-grade fevers within 3 hours of infusion. All 3 patients on DL3 developed grade 2 CRS that responded to a single dose of tocilizumab. On DL4, all 3 patients presented with grade 4-5 CRS with evidence of immune effector cell-associated hemophagocytic lymphohistiocytosis-like syndrome (IEC-HS). All DL4 patients required multiple doses of immune modulatory agents including tocilizumab, steroids and emapalumab. One patient also received ruxolitinib, anakinra and etanercept without benefit, and died of cardiorespiratory failure. Autopsy results were consistent with CRS and no evidence of on target/off cancer toxicity. Patients post CD123-CAR.dasa T-cell infusion had significantly higher levels of circulating chemokines and cytokines, including IFNg, GMCSF, IL10, IL6 and TNFa, in comparison to patients who had received CD123-CAR T cells. Peak expansion of CD123-CAR.dasa T cells, as judged by qPCR analysis, was higher in comparison to CD123-CAR T cells. However, this did not translate into improved anti-leukemia activity as no patients had a response.

Conclusions

CD123-CAR T cells generated in the presence of dasatinib induced high grade CRS and/or IEC-HS without improved anti-leukemia activity. Our findings are in contrast with CAR.dasa T-cell products that target other antigens. While additional mechanistic studies are in progress, our results highlight that dasatinib cannot be considered a universal agent to improve the effector function of CAR T cells for hematological malignancies.