CAR T Cells Targeted with a High Affinity Scfv Against the HCMV Glycoprotein Gb As Adoptive T Cell Therapy after Hematopoietic Stem Cell Transplantation

Introduction: Reactivation of human cytomegalovirus (HCMV) in immune compromised patients after hematopoietic stem cell transplantation (HSCT) is associated with high morbidity and mortality, particularly after cord blood transplantation (CBT). Adoptive transfer of T cells expanded in vitro is currently used as therapy for drug-refractory HCMV disease. A major limitation of this approach is the requirement of HLA-restricted HCMV-specific memory T cells. An alternative approach exploring HLA-independent T cell recognition was sought. Because the HCMV envelope glycoprotein B (gB) is highly expressed during lytic infection and in latently infected cells, we hypothesized that T cells can be redirected to recognize and kill HCMV-specific cells by means of a gB-specific chimeric antigen receptor (CAR). We have synthesized and tested a gB-specific CAR derived from the SM5-1 monoclonal antibody which binds with high affinity (K_D 5.7x10¹¹) to a conserved antigenic and non-glycosylated domain of gB.

Methods: We generated two codon-optimized SM5-derived scFvs (VH->VL and VL->VH) and fused with an existing CAR backbone containing an IgG Fc spacer and intracellular signaling domains. CARs containing either CD28.zeta or 4-1BB.zeta were synthesized and expressed in T cells following a standard retroviral transduction protocol yielding 80-90% transduction rate. Expression of the CARs on T cells was confirmed by flow cytometry using goat anti-human immunoglobulin reactive against the IgG Fc region. 293T cells co-expressing gB and dTomato were used for in vitro cytotoxicity assays.

Results: T cells expressing gB-CAR/CD28.zeta were cytotoxic against gB⁺ target cells producing 90% killing of 293T/gB-dTom cells compared with control CD19 CAR/CD28.zeta cells at an effector-to-target ratio 3:1 for 48 h (parental 293T cells were not killed). The cytolytic activity correlated with expansion of CAR T cells and concomitant loss of gB-dTom expression in the remaining viable 293T cells. Sequential co-culture of these gB-CAR T cells with freshly seeded 293T/gB-dTom resulted into further elimination of target cells. We are currently evaluating the effects of different gB-CAR T cell designs in the killing of HCMV-infected cell lines and primary cells using HCMV laboratory strains expressing the GFP and Gaussia Luciferase reporter genes. Pilot experiments indicated that gB-CAR/CD28.zeta cells with the scFv in the VL->VH orientation resulted into more clustering and killing of HepG2 cells infected with HCMV-GFP after 24h of co-culture than a control CD19 CAR/CD28.zeta. Humanized mice transplanted with cord blood CD34⁺ stem cells and challenged with these HCMV laboratory strains will be used to evaluate the in vivo effectivity of cord blood-derived donor-matched gB-CAR-T cells to eliminate acute and latent HCMV infections.

Conclusion: These studies explore a novel approach in preventing HCMV reactivation in immunosuppressed patients by redirecting T cells expressing a high-affinity gB-CAR to eliminate HCMV-infected cells in a TCR/MHC-independent manner.