Vectors for Algernon: Receptor-Targeted Lentiviral and Adeno-Associated Vectors in Syngeneic Mouse Models of In Vivo Gene Therapy

Driven by recent leaps in receptor targeting technology, advanced vector platforms are currently reshaping gene therapeutic strategy, promising to enable a new generation of accessible, safe and unprecedentedly effective products. Their transformative potential has been impressively demonstrated by recent preclinical reports describing the in vivo generation of chimeric antigen receptor (CAR) T cells in short-term mouse models. Beyond proof-of-principle, however, important questions remain, many of them motivated by an insufficient understanding of the host immune response to vector administration, which will likely critically impact the products’ real-world safety and efficacy. Toward a better understanding of the host response, syngeneic mouse models - capable of recapitulating the response of a complete, complex mammalian immune system to vector administration - can helpfully complement the existing body of work on humanized models and enable the thorough preclinical examination of vector-host interplay which is warranted by the tumultuous history of clinical gene therapy research. Such models require surrogate reagents, which have so far been unavailable for certain classes of receptor-targeted vectors, especially viral vectors. This thesis describes the generation and characterization of such mouse-compatible viral vectors, as well as their use in the syngeneic mouse models of in vivo gene therapy they enable. Lentiviral vectors targeted to CD8+ and CD4+ murine lymphocytes (mCD8- & mCD4-LVs) were generated by insertion of anti-mCD8α MSE10 designed ankyrin repeat protein (DARPin) and anti-mCD4 GK1.5 single chain variable fragment, respectively, into the blinded measles virus pseudotype pioneered by the host lab- oratory. Crucially, in spite of the well-documented, multicausal inability of LVs’ infamous parent virus HIV-1 to productively infect murine cells, the introduction of mouse receptor-targeted binders was sufficient to confer mouse-compatibility to the particles, which displayed transducing titers on primary mouse splenocytes similar to those observed for hCD8- and hCD4-LVs on human T cell lines. Additionally, binder insertion rendered mCD4- and mCD8-LV highly selective for cells expressing their cognate receptor: Five days after vector addition, >98% of GFP+ lymphocytes extracted from whole mouse blood treated with mCD8-LV were found to be CD8+. The subtype-specific presence of tagged viral glycoproteins on T cells only two hours after addition of mCD4- and mCD8-LV to whole mouse blood observed on closer examination suggest that the vectors’ selectivity is achieved at the stage of cell binding. Interestingly, receptor incompatibility is a principal barrier not only for the efficient transduction of murine cells with lentiviral, but also with adeno-associated vectors (AAVs). When the mCD8α-specific MSE10 DARPin was inserted into the GH2/GH3 loop of VP1 of the AAV2 capsid, the resulting DARPin-targeted (DART) mCD8-AAV displayed near-absolute selectivity for CD8+ primary murine lympho- cytes and transducing titers six- to sevenfold higher than those of regular AAV2. When mCD8- and mCD4-LV were used for GFP transfer in BALB/c mice, protein and genome level transfer signals were close to the lower limit of detection, but were strongest in the tissues with the highest T cell content. Notably, pronounced remodeling of the lymphoid compartment, i.e. decreases in the relative frequency and increases in size and granularity of CD3+ cells in spleen and blood, were observed, indicating an immune response to vector administration. Host response was minimal when a mix of phagocytosis-shielded CD47hi mCD8- and mCD4-LVs was systemically injected into BALB/c mice to generate αmCD19-CAR T cells directly in vivo, as no signs of immune activation were observed and no CAR-related changes in peripheral blood were found within 43 days post injection. On final analysis, qPCR identified splenic vector integration only in treated mice, and flow cytometric analysis yielded a shift of the CD3+/CD19+ composition in spleens of vector-treated animals. Data highlighting the considerable influence of receptor targeting on vector biodistribution was obtained when mCD8-AAV was tested in Ai9 mice. In this reporter-tolerant system, infusion of mCD8-AAV resulted in >87% specific trans- duction of mCD8+ cells and overall transduction rates in blood, spleen and bone marrow approximately 4-100 fold higher than for unmodified AAV2. Additionally, targeting of AAVs by MSE10 was found to reduce liver transduction, assessed by measuring whole organ surface fluorescence, twentyfold. The observed utility of receptor targeting in the context of two molecularly distinct vector platforms well-illustrates its crucial role in the maturation of gene therapy, enabling pivotal vector platforms as well as their detailed preclinical examination. Indeed, the syngeneic mouse models enabled here through the generation of mouse-compatible viral vectors well-summarized current concerns in the field, as they confirmed both the challenge in vivo gene therapy faces from restrictive host responses – stressing the urgent need for the evaluation and implementation of immune-modulatory strategies to enable productive in vivo transduction in immunocompetent systems – and the key role of receptor targeting technology in profoundly improving genetic treatment, e.g. by decreasing liver burden for a vector class whose liver toxicity upon systemic administration is an emerging issue