Abstract 574
Relapsed and refractory pediatric acute lymphocytic leukemia (ALL) remains a difficult therapeutic challenge and accounts for a sizable number of cancer-related deaths in children. Chimeric antigen receptors (CAR) are genetically engineered molecules that combine antibody specificity for a target antigen with the potent cytotoxic potential of activated T cells. CAR based T cell immunotherapy is currently under study in several clinical trials and encouraging early response data is beginning to emerge. However, studies in pediatric malignancies are still lacking. We have studied the efficacy of a second-generation CAR containing the T-cell receptor zeta signaling subunit, the signaling domain of CD28, and a single chain variable fragment directed against CD19. CD19 is present on nearly 100% of pediatric ALL blasts and normal B cells but not on hematopoietic stem cells. Anti-CD19-CAR T cells are generated using retroviral transduction of T cells activated with K562 based artificial antigen presenting cells expressing the high-affinity Fc-receptor (CD64) loaded with anti-CD3 and the co-stimulatory molecule, 41BB-ligand. Transduction efficiency, as measured by flow cytometry, averaged 50–80%. Unselected anti-CD19-CAR T-cells specifically and robustly killed four CD19+ ALL but not CD19– target cell lines (45-60% lysis at E:T ratios as low as 2.5:1 in 4h
51Cr release assay) and produced significant levels of IFNg, TNFa, and IL-2 when encountering CD19+ ALL cell lines but not CD19– cells. In xenograft models, 3 × 10^6 unselected anti-CD19-CAR T-cells injected via tail vein eliminated engrafted ALL (NALM6 cell line stably expressing both GFP and firefly luciferase, NALM6-GL) within 48 hours in 5/5 immunodeficient NOG mice (Figure 1) whereas all animals that received 3 × 10^6 activated but untransfected T-cells required sacrifice within 18 days. Nearly two months after T cell infusion, CD19-CAR T cells bearing an effector memory phenotype could still be detected in peripheral blood. When lower T cell doses were used, antitumor effects were less potent and were associated with a lack of persistence of anti-CD19-CAR T cells in vivo.
CD4+ T cells are not classically cytotoxic and normally require Class II presentation of target peptides for recognition. However, since CAR based recognition is MHC independent, we hypothesized that both CD4+ and CD8+ anti-CD19-CAR T-cells may mediate cytotoxicity. To test this, CD4+ and CD8+ T-cells from the same donor were negatively selected using immunomagnetic beads, then activated and transduced as before. A 4-hour 51Cr release assay demonstrated cytotoxicity from CD8+ CAR T-cells similar to unselected cells but no appreciable cytotoxicity from CD4+ CAR T cells. However, in NOG xenografts, 1 × 10^6 CD4+ CD19-CAR T cells produced complete responses in 4/5 mice by Day 10 as compared to 5/5 complete responses in animals receiving 1 × 10^6 CD8+ CD19-CAR T cells. While antitumor effects mediated by CD8+ T cells were rapid and nearly complete within 3 days, antitumor effects of CD4+ T cells were slower, peaking at approximately day 14. Together, these results demonstrate notable differences in efficacy of anti-CD19-CAR based therapy with small changes in cell dose, unexpected activity of CD4+ anti-CD19-CAR based T cells in xenografts which is not reflected by activity in short term killing assays, and evidence that cure of mice is associated with persistence of anti-CD19-CAR modified T-cells in this model. These results inform clinical development of this emerging therapy by emphasizing the importance of cell dose, cell persistence and the novel observation that CD4+ anti-CAR T cells may be important effectors in mediating antitumor effects of this therapy.
Disclosures:
No relevant conflicts of interest to declare.
