Donor-Derived CD19 CAR Cytokine Induced Killer (CIK) Cells Engineered with Sleeping Beauty Transposon for Relapsed B-Cell Acute Lymphoblastic Leukemia (B-ALL)

Background Immunotherapy using patient-derived CAR T cells has achieved complete remission and durable response in highly refractory populations. However, logistical complexity and high costs of manufacturing autologous viral products limit CAR T cell availability. Allogeneic Cytokine Induced Killer (CIK) cells, a T-cell population characterized by the enrichment of CD3+CD56+ cells, have demonstrated a high profile of safety in acute lymphoblastic leukemia (ALL) patients (Introna M et al. Biol Blood Marrow Transplant. 2017). CIK cells could be easily engineered by the non-viral Sleeping Beauty (SB) transposon for the clinical application (Magnani CF et al, Hum Gene Ther. 2018, Biondi A et al. J Autoimmun. 2017).

Methods CIK cells were generated from 50 ml of donor-derived peripheral blood (PB) by electroporation with the GMP-grade CD19.CAR/pTMNDU3 and pCMV-SB11 plasmids according to the method enclosed in the filed patent EP20140192371. After lymphodepletion with Fludarabine (30 mg/m²/day) x 4 days and Cyclophosphamide (300 mg/m²/day) x 2 days, CARCIK-CD19 were infused in pediatric and adult B-cell ALL (B-ALL) patients relapsed after allogeneic hematopoietic stem cell transplantation (HSCT). The clinical trial follows a four-dose escalation scheme (1x10⁶, 3x10⁶, 7.5x10⁶ and 15x10⁶ transduced CAR+ T cells/kg) using the novel Bayesian Optimal Interval Design (BOIN). During the cell manufacturing period, bridging anti leukemic therapy from patient registration to the beginning of the lymphodepletion, was allowed. The primary endpoint was to define the Maximum Tolerated Dose (MTD) and a safety assessment. Key secondary endpoints included the assessment of complete hematologic response (CR), defined as < 5% bone marrow (BM) blasts, circulating blasts < 1%, no clinical evidence of extramedullary disease, as well as the characterization of CARCIK-CD19 persistence in PB and BM (NCT03389035).

Results We manufactured eighteen batches by seeding a median of 126.8x10⁶ allogeneicPBMCs. At the end of expansion, the mean harvesting was 6.46x10⁹ nucleated cells (range 1.39 - 16.00x10⁹). Manufactured cells were mostly CD3+ lymphocytes (mean 98.90% ±SE 0.30%). Of these, 43.57%±3.73% were CAR+, 47.07%±2.74% were CD56+, 80.44%±2.53% were CD8+. The quality requirements for batch release were met in 17 productions. As of the data cut-off date (July 19, 2019), 4 pediatric and 7 adult patients were infused with a single dose of CARCIK-CD19 (n=2 HLA identical sibling, n=4 MUD, n=5 haploidentical donor). The leukemic burden in the BM post lymphodepletion/pre-infusion ranged from 0% to 96%. CARCIK-CD19 were characterized by a high profile of safety in all treated patients. Toxicities reported were a grade I cytokine release syndrome and an infusion-related DMSO-associated seizure, with absence of dose-limiting toxicities, neurotoxicity and graft-versus-host disease (GvHD) in any of the treated patients. Four out of 5 patients, receiving the highest doses, achieved CR and CRi at day 28. The 3 patients in CR were also MRD- (by flow cytometry and RT-PCR) while the CRi was MRD+ and relapsed at day+49. Robust expansion was achieved in the majority of the patients as defined by detectable CAR T-cell detection (vector copy number VCN, range 4645-977992 transgene copies/ug) and flow (range 0.5-30%) in PB. The median time to peak engraftment was 14 days. The magnitude of expansion was correlated with the CD19+ burden in the BM at the time of the infusion (P value = 0.0006, R square 0.7469). CD8+ T cells represented the predominant CARCIK-CD19 T-cell subset (78.88%±5.33% d14 n=6) along with CD3+CD56+ CIK cells and CD4+ T cells to a lesser extent. The majority of CAR T cells had a central and effector memory phenotype. CAR T cells were measurable by VCN up to 6 months with a mean persistence of 70.5 ± 14.85 days (follow up ranging from 28 days to 1 year). No major difference was observed by integration analyses of the patients' PB and the cell products. The vector integration sites reflected the classical random distribution of SB without any tendency for gene dense regions.

Conclusions Our ongoing phase I/II trial demonstrates that SB-engineered CARCIK-CD19 cells are able to expand and persist in pediatric and adult B-ALL patients relapsed after HSCT, with important implications for a non-viral technology. These encouraging results prompted us to expand our study.