Phase I Study of AUTO3, a Bicistronic Chimeric Antigen Receptor (CAR) T-Cell Therapy Targeting CD19 and CD22, in Pediatric Patients with Relapsed/Refractory B-Cell Acute Lymphoblastic Leukemia (r/r B-ALL): Amelia Study

Introduction

CAR T-cell therapies directed against CD19 or CD22 have shown remarkable activity in r/r B-ALL but relapse due to target antigen down-regulation/loss has been the major cause of treatment failure. To address this, we developed AUTO3, a CAR T-cell therapy designed to target CD19 and CD22 simultaneously. Preliminary results of this study showed an acceptable safety profile and encouraging efficacy in pediatric r/r B-ALL (all 6 patients treated in active doses ≥3 x 10⁶ CAR T-cells/ Kg achieved complete remission (CR) with negative minimal residual disease (MRD) (Amrolia et al, Blood 2018 132:279). Here we present the updated results of CAR naïve patients treated at the active doses.

Methods & Patients

We constructed a bicistronic retroviral vector encoding both an anti-CD19 CAR and an anti-CD22 CAR. This second-generation CAR incorporated an OX40 co-stimulatory domain for the CD19 CAR and a 41BB for the CD22 CAR. The cell product was manufactured on a semi-automated/closed process.

Patients (aged 1‒24 years) with high risk relapsed (IBFM criteria) or refractory B-ALL, adequate performance score/organ function, an absolute lymphocyte count ≥0.5 x 10⁹/L are eligible. Patients with CNS Grade 3 disease, active graft versus host disease are excluded. Patients receive lymphodepletion with 30 mg/m²/day fludarabine x 4 days and 500 mg/m²/day cyclophosphamide x 2 days prior to AUTO3 infusion.

Three dose levels were explored (1 x 10⁶, 3 x 10⁶, and 5 x 10⁶ cells/kg), CAR T cells are infused as a single (for <25% blasts) or split (for ≥25% blasts) dose based on leukemia burden. Bridging therapy is allowed during the manufacturing period. The primary endpoint is the frequency of dose-limiting toxicities (DLTs) and key secondary endpoints include morphological/MRD negative CR rate, disease-free survival, overall survival, as well as AUTO3 levels and persistence in blood and bone marrow.

Results

As of the data cut-off date (June 17, 2019), 10 patients received AUTO3 at 3 x 10⁶ cells/Kg (n= 5, of whom 1 received split dose) or 5 x 10⁶ CAR T-cells/Kg (n= 5, all of them received single infusion). The median transduction efficiency was 15.5% (range 8.6‒39.3%). Median age was 8.5 years (range 5‒16 years) and 5 (50%) patients had prior haemopoietic stem cell transplant (HSCT). One patient (10%) had prior anti-CD19 CAR-T cells. The disease burden at Day ‒7 ranged from 0 to 38% (median 7.5%) blasts.

Among the 10 treated patients, 2 have not completed the 30 days post-infusion DLT observation period as of the cut-off date. No deaths or DLTs were observed. MTD has not yet been reached. The most common grade (Gr) ≥3 adverse events were neutropenia (60%), anaemia (50%), pyrexia (40%), febrile neutropenia (40%) and thrombocytopenia (30%). Eight patients (80%) had Gr 1 cytokine release syndrome (CRS), one (10%) had Gr 2 CRS; no ≥Gr 3 CRS was observed. Only one patient was treated with tocilizumab and none required admission to ICU due to CRS. One patient (10%) experienced Gr 1 neurotoxicity; no ≥ Gr 2 neurotoxicity was reported.

Among the 9 CAR naïve patients, 7 (4 in the 3 x 10⁶ cells/Kg dose cohort, 3 in the 5 x 10⁶ cells/Kg dose cohort) had a minimum of 8 weeks' follow up and were evaluable for efficacy analysis. All 7 patients achieved CR/CRi (100%) following AUTO3 infusion as well as molecular negative remission (100%). After a median follow-up of 8 months (range 2-12), emergence of MRD by PCR occurred in four patients, lack of persistence of circulating CAR T-cells was observed in 3 of the 4 patients. Three relapses were reported including one with CD19 negative/CD22 low expression at 1 year after treatment. One patient in ongoing molecular remission proceeded to HSCT. All the remaining 4 patients in ongoing CR/CRi maintain B-cell aplasia. The median CAR T-cell expansion (expressed as vector copy number per microgram of DNA) at peak was 102K (range 56-128). The median persistence of CAR-T cells in blood was 180 days (range 21-330). Updated data with longer follow up and additional patient data will be presented.

Conclusion

This interim data analysis demonstrates that AUTO3 at ≥3 x 10⁶ cells dose achieved 100% molecular remission rate with a favourable safety profile, no ≥ Gr 3 CRS or ≥ Gr 2 neurotoxicity was reported. The most common cause of relapse was antigen positive relapse due to lack of CAR T cell persistence. Evaluation of patients with a modified manufacturing process is planned.