Introduction: Chimeric antigen receptor (CAR) T cell therapy targeting CD19 has recently demonstrated high success but also shown limitations regarding their toxicity and development of CD19negative variants. Here we reported results from a phase I study designed to determine the safety of the CD19 CAR-T and CD22 CAR-T cocktail and the feasibility of making enough quantities to treat patients with CD19+CD22+ relapsed/refractory B cell acute lymphoblastic leukemia (B-ALL).
Patients and Methods: From July 2017 to July 2018, a total of 15 patients with CD19+CD22+ relapsed/refractory B-ALL were treated, including 5 children and 10 adults (Table 1). All patients received fludarabine 30mg/m2/d´3d and cyclophosphamide 250mg/m2/d´2d before infusion of CAR-T cells, followed by a cocktail CAR-T cell infusion with a median number of 2 (0.9-5)´105 CD19 CAR+ T cells/kg and a median number of 0.5 (0.4-12)´105 CD22 CAR+ T cells/kg. The lentiviral backbone containing constructs of CD19 CAR and CD22 CAR are shown in Figure 1. CD19 CAR includes a truncated EGFR sequence which can be used to identify and select CAR+ cells. CD22 CAR includes a single chain variable fragment (ScFv) sequence derived from a monoclonal antibody against human PD-L1 which attempts to reduce the exhaustion of CAR-T cells by blocking the PD-1/PD-L1. Real-time quantitative PCR using primers with specificity for the ScFv of CD19 CAR and ScFv of CD22 CAR can detect the in vivo CAR-T persistence for either CAR. Sequential transduction was performed 2 days after activation of sorted T cells stimulated with CD3 and CD28 antibodies. Percentages of CD19 CAR+ and CD22 CAR+ T cells were determined by flow cytometry through staining with an antibody against EGFR and a fusion protein of CD22-Fc, respectively, and expression of anti-PDL1 ScFv in CD22 CAR-T cells were demonstrated by flow cytometry through intracellular staining with a PD-L1-Fc fusion protein. The primary end points of this study were to evaluate feasibility and toxicity, and the secondary end points included disease response and persistent CAR-T engraftment of infused CAR-T cell.
Results: The median observation period was 133 days (24-392 days). The median percentage of pre-treatment bone marrow CD19+CD22+ blasts was 21.5%(0.11-74.1%). On day 20-30 after CAR-T infusion, 15/15 (100%) cases achieved complete remission (CR) or incomplete CR(CRi), 14/15 (93.3%) cases had negative minimal residual disease (MRD). Patient P098 had residual (0.58%) CD19+CD22+ BM blasts at day 30 post-infusion and thereafter achieved negative MRD after re-infusion with CD19 CAR-T cells. 11/17 patients were bridged into allo-HSCT and have remained in remission state with a median follow up of 133 (97-214) days. 2/5 patients without bridging allo-HSCT relapsed on day 240 and day 105 post-infusion, respectively. Notably, both patients (100%) relapsed with CD19+CD22+ leukemia cells. Despite achievement of a very high CR rate, a very low treatment-related toxicity was observed in this trial. Only 1 patients experienced grade 3 cytokine release syndrome (CRS) and another patient (6.7%) developed grade 3 central nervous system (CNS) toxicity; all other patients were CRS grade<2 and CNS grade 0. On days -1, 1, 4, 7, 10, 14, 21 and 28 after infusion, peripheral blood (PB) was drawn and the level of infused CD19 and CD22 CAR-T cells were analyzed by either qPCR or flow cytometry assay. Results demonstrated obvious in vivo proliferation of both CD19 and CD22 CAR-T cells. The median peak level was 3.5 (0.47-79.1)´104 copy number/mg PB genomic DNA for CD19 CAR-T, and 0.9 (0.08-80.8)´104 copy number/mg PB genomic DNA for CD22 CAR-T. The median day to reach the peak value was day 10 for both CARs, ranging mostly from day 7 to day 14.
Conclusion: This study demonstrates technical feasibility, high efficacy and low toxicities of CD19 and CD22 CAR-T cocktail in treating patients with CD19+CD22+ relapsed/refractory B-ALL. Both patients relapsed with CD19+ leukemia suggests this cocktail treatment may reduce the risk of CD19 negative relapse. Low toxicities may relate with small number of infused CAR-T, but involvement of anti-PDL1 ScFv which is co-expressed with CD22 CAR construct cannot be excluded. Therefore, related mechanisms are currently being investigated in the lab.
No relevant conflicts of interest to declare.