Allogeneic Veto-CAR CD8 T Cells Enabling Prolonged Graft Survival, Anti-Tumor Specificity, and Low Risk of Graft Versus Host Disease

Abstract

¥ Equal contribution Studies in mice have demonstrated that CD8 T cells exhibit marked veto activity against host T cells with anti-graft specificity, enhancing engraftment after non-myeloablative T cell depleted mismatched bone marrow transplantation. To reduce the risk of GVHD associated with allogeneic CD8 veto T cells, these studies used naïve CD8 T cells stimulated against 3rd party stimulators under cytokine deprivation and subsequent expansion in the presence of IL-15. Recently, we showed that mouse CD8 veto T cells can be generated by stimulating CD8 memory T cells from ovalbumin-immunized mice under cytokine deprivation, using ovalbumin as a third-party antigen (Or-Geva et al., Leukemia, 2019). Based on these insights, human anti-viral CD8 central memory veto cells were successfully generated from CMV and EBV positive donors. Donor PBMC collected by leukapheresis were initially depleted of CD45RA+, CD4+ and CD56+ cells, and co-cultured with dendritic cells from the same donor, pulsed with a mixture of CMV, EBV, Adeno and BKV viral peptides. Within 12 days of culture, starting with 3 days of cytokine deprivation, more than 3x10 9 highly homogenous CD8 T cells, predominantly expressing a central memory phenotype could be routinely harvested. These cells exhibited marked veto activity in-vitro with more than a 3 log-depletion of alloreactivity as measured by limiting dilution analysis. Based on these results, a first in human phase 1-2 clinical trial is currently in progress, testing the safety of such veto cells in the context of non-myeloablative haploidentical T cell-depleted hematopoietic stem cell transplantation (HSCT). Preliminary results indicate that infusion of 5-10x10 6 veto cells / kg supports engraftment with a low risk of GVHD (Champlin et al. Blood, 2022). Most attempts to use allogenic chimeric antigen receptor (CAR) T cells involve gene editing, knocking down the TCR to avoid GVHD, and eliminating MHC expression to avoid graft rejection. Our veto cell preparation could potentially offer an attractive platform for attaining allogeneic CAR T cells capable of deleting host anti-donor T cell clones by virtue of their veto activity, with a very low risk for GVHD and without the need of gene editing. Using our current protocol for production of anti-viral central memory veto CD8 T cells, we first evaluated transduction efficacy at different time points during the veto production using a retroviral CAR directed against the Her-2 antigen (N29). The scFv N29 was attached to a costimulatory determinant of CD28 and a T cell activating domain followed by a GFP reporter gene. Our results suggested that optimal CAR transduction was attained on day 5 of culture. The veto cell product harvested on day 12 exhibited 92% CD8 T cells, of which 72% expressed the transduced N29 CAR. Notably, the transduced cells continued to exhibit veto activity and functional assays showed killing of target cells in a dose dependent manner. Next, we performed similar experiments transducing veto Tcm with a clinical grade gammaretroviral vector encoding a second-generation 4-1BB co-stimulated CD19-specific CAR. A typical experiment, shown in Fig.1, demonstrates a transduction level of 66.7 % (Fig. 1B,) and confirms the central memory phenotype of the transduced cells (Fig. 1C). Notably, upon infusion of 3x10 6 veto-CD19CAR into NSG mice implanted 3 days earlier with 0.25x10 6 Nalm-6-Luc+ tumor cells (CD19+), a marked eradication of tumor cells (Fig.1D) and survival benefit (Fig.1E) were found, while infusion of 3x10 6 non-transduced veto cells were ineffective. Based on these results, we plan to initiate clinical trials using haploidentical VETO-CD19CAR for the treatment of B-cell malignancies, in combination with hematopoietic transplantation and as off the shelf cellular therapy. .