SCRI-CAR19x22v2 T Cell Product Demonstrates Bispecific Activity in B-ALL

Introduction: CAR T cells in B-ALL have recently focused on the dual targeting of CD19 and CD22 to enhance long term remissions and prevent antigen negative recurrence that is frequently encountered with single antigen targeting. However, a barrier to this approach has been the retention of dual specificity killing and ongoing persistence. PLAT-05 is a multisite phase 1 trial (NCT03330691) that was undertaken to evaluate the safety and feasibility of SCRI-CAR19x22v1, a dual transduced patient-derived product with lentiviral vectors encoding for either a CD19- or CD22-specific CAR, both with 4-1BB co-stimulation. Early results of the first 27 subjects infused demonstrated feasibility and a favorable safety profile with encouraging CR rates. Products were fractionated evenly between CD19 CAR, CD19+CD22 CAR and CD22 CAR. However, engraftment was predominated by the single CD19 CAR population, leading to unsuccessful eradication of CD19-CD22+ leukemia. This finding led to re-engineering the CD22 CAR construct for enhanced CD22 targeting, and re-initiation of dose finding with the new product, SCRI-CAR19v2. Methods: After enrollment, subjects undergo apheresis followed by a combined CD4/CD8 positive immunomagnetic selection and seeded at a prescribed ratio for co-culture in a closed-system G-Rex bioreactor. Following anti-CD3xCD28 bead stimulation, T cells are transduced with two lentiviral vectors that encode for either a CD19- or CD22-specific CAR. After flu/cy lymphodepletion, CAR T cells are infused at one of three dose levels: 0.5, 1 or 3 X 10 6 CAR T cells/kg. Toxicity is graded according to CTCAEv5 except for CRS and ICANS which are graded per ASTCT criteria. Leukemic response and CAR T cell persistence are evaluated by flow cytometry. Results: 14 subjects enrolled onto PLAT-05 for the SCRI-CAR19x22v2 dose escalation and products were successfully manufactured in all subjects with an average of 8.9 days in culture (range 7-12 days). In contrast to v1 products, the CAR composition of v2 products was skewed in favor of CD22 CAR expression, with median expression of each population as follows: 42% CD22 only, 33% CD19 and CD22, 3.2% CD19 only. Twelve subjects were infused (0.5x10 6/kg n=3, 1x10 6/kg n=3, 3x10 6/kg n=6), 11 of whom had prior exposure to CD19 or CD22 targeted therapies with diverse expression of CD19 and CD22 on the leukemic blasts. No dose limiting toxicities occurred in the 11 fully evaluable subjects (1 subject is pending) and the recommend