T cell-targeted gene delivery: Rapid CAR T cell generation without full activation

Lentiviral vectors (LVs) are potent tools to genetically modify hematopoietic stem cells (HSCs), T cells or B cells. Genetic engineering of T cells to express a chimeric antigen receptor (CAR) was termed the breakthrough therapy of the year 2013 by the journal Science and ever since numerous approaches using CAR T cells for different targets are being investigated. CAR T cell therapy celebrated its greatest successes so far in 2018, when two CAR T cell products received marketing authorization by the food and drug administration (FDA) and the European commission (EC). Despite its promising results, CAR T cell therapy still needs to overcome various hurdles, including a complicated and time-consuming manufacturing process. Additionally, a protocol which generates high amounts of less differentiated T cells, that were shown to be beneficial in regard of CAR T cell persistence, would be of advantage to avoid cumbersome selection of specific CAR T cell subsets. In this thesis, a receptor-targeted LV specific for CD62L was established to ease and shorten the CAR T cell manufacturing process and obtain higher amounts of preferred CAR T cell phenotypes than are derived after transduction with a vesicular stomatitis virus glycoprotein (VSV-G) pseudotyped LV. These beneficial T cell subsets are expressing CD62L and this circumstance was crucial for producing a CD62L receptor-targeted LV (62L-LV). First steps for establishment of 62L-LV included choosing the appropriate LV receptor-targeting proteins. Therefore, surface expression of a CD62L specific single chain variable fragment (scFv) fused to different envelope proteins was analyzed. Afterwards, vector particles were generated that encoded the green fluorescent protein (gfp). As a first proof-of-concept these vector particles were tested for their ability to selectivity deliver the transgene to CD62L+ cell lines. Besides, an αCD19-CAR was analyzed that later was planned to be packaged into 62L-LV. As a start to characterize this CAR, it was packaged into the already established receptor-targeted LVs: CD4-LV and CD8-LV. The CAR gene delivery potency of respective LVs to human T cells was analyzed in presence and absence of a transduction enhancer. The arising transduced T cells were proven for their ability to kill target cells. Moreover it was uncovered that the CAR and the reporter protein can be transferred as protein from LV packaging cells to T cells. Notably, this was discovered only because of the usage of T cell-targeted LVs. After initial testing of CD62L-LV-gfp and the αCD19-CAR both were combined and the resulting vector was analyzed for correct vector particle assembly via Western blot. Robustness of LV production was analyzed by quantifying particle numbers and gene delivery approaches. In addition, specificity of gene transfer was proven on cell lines and primary cells. Furthermore, the influence of shed CD62L on vector particle binding was investigated. Additionally, transduction efficiency of T cells and phenotypes of transduced T cells were evaluated. Transduction rate using a transduction enhancer was reaching levels similar to those obtained with VSV-LV. Interestingly, T cells transduced with 62-LV were less differentiated in vitro even after full activation. Even more so, the ability to deliver transgenes to minimally activated T cells was investigated in vitro and in vivo. While minimally activated T cells were functional in vitro after both transduction with VSV-LV and 62L-LV, an initial in vivo study suggested that receptor-targeted LV outperformed VSV-LV in an antitumoral mouse model. Fully and minimally activated transduced T cells were injected into mice as early as 24 hours after incubation with vector and were challenged with antigen by application of CD19 positive tumor cells three days later. Tumor growth was followed regularly by in vivo bioluminescence imaging. Indeed, the full activated CAR T cells were antitumorally active, however, 62L-LV incubated T cells outperformed T cells that were incubated with VSV-LV in the setting with minimally activated T cells. Thus, only 62L-LV incubated minimally activated T cells were antitumorally functional in the antitumor mouse model. Distribution of transgene expressing T cells in various organs was inspected, as well as their phenotypes and exhaustion profiles. In mice that had less tumor load compared to a control group, CAR T cells populated various organs. Interestingly, CAR T cells from mice receiving minimally activated T cells incubated with 62L-LV had a higher amount of less differentiated cells in comparison to fully activated T cells. In conclusion, a novel receptor-targeted LV was successfully generated which is specific for CD62L. Usage of 62L-LV allowed generation of potent CAR T cells within less than two days and did not require full activating stimuli. 62L-LV can be a potent tool in the CAR T cell therapy field to generate CAR T cells with beneficial phenotypes very rapidly