Using Gene Therapy to Solve Challenges with CAR-T Cell Immunotherapy: Lead Selection and Preclinical Development of an Adeno-Associated Virus with Reduced Immunogenicity Exhibiting Efficient and Long-Term Expression of an Anti-CD19 T-Cell Engager

We previously described adeno-associated virus (AAV) gene therapy to achieve off-the-shelf, long-term in vivo T cell engagement for cancer (PMID 35857488). A single intravenous dose of AAV expressing a secreted anti-CD19/anti-CD3 bispecific diabody, termed GP101, resulted in a persistent long-term steady-state levels and tumor elimination in a flank model of Raji Burkitt lymphoma. Like CAR-T, AAV gene therapy is a single infusion with long-lasting effects, but potential advantages of AAV over CAR-T include it is available off-the-shelf, provides a gentle onset of action, doesn’t require lymphodepletion chemotherapy, engages all T cells including those freshly produced from the bone marrow continuously, and utilizes highly efficient signaling through the native T cell receptor. Thus, we sought to develop a clinical lead construct. To reduce transgene-mediated immunogenicity, we deleted the unnecessary C-terminus 6xHis tag and used alternative codons to eliminate CpGs (toll-like receptor agonists), which we showed did not compromise expression in primary human hepatocytes. Based on FDA feedback, we also modified the woodchuck hepatitis virus post-transcriptional regulatory element (WPRE) to prevent expression of the potentially oncogenic “protein x.” Finally, we compared several different promoters and a self-complimentary AAV for expression in cultured cells and in mice. We selected the vector yielding the highest levels of expression, termed VNX-101, as the lead. GP101 purified after cell transfection using this construct was functionally equivalent to blinatumomab in a co-culture cell killing assay (both EC50s < 0.08 ng/mL). A single intravenous dose of GMP VNX-101 at 5.0E12 vg/kg in male and 2.5E13 vg/kg in female NSGS mice combined with human peripheral blood mononuclear cells were effective in eliminating leukemic cells in widely disseminated human xenograft mouse models (Raji and Tanoue) and increased progression-free and overall survival compared to controls. To understand the dose-serum level relationship, we measured serum GP101 levels using a qualified ELISA at 4, 6, and 8 weeks following 7 dose levels of VNX-101. Every 3-fold increase in dose resulted in approximately 10-fold increase in serum levels. As previously described for AAV in mice, ~3-fold higher levels in females were required to achieve similar levels in males. Serum levels of GP101 were consistently above the lower threshold of the human therapeutic target range (300 pg/mL) at doses of 3.0E10 vg/kg (females) and 1.0E10 vg/kg (males). We also conducted a single-dose 12-week toxicity study of 3 different dose levels in mice with PK and biodistribution under GLP conditions. We observed a dose-dependent increase in serum GP101, with the lowest tested dose of 1.0E11 vg/kg resulting in GP101 concentration well above the target threshold of 300 pg/mL. At that dose, there were 0.007+0.001 AAV genomes per diploid liver cell as quantified by ddPCR and <0.002 in other organs. There were no safety signals and the No Observed Adverse Event Level was >2.7E13 vg/kg. We also undertook a 12-week study in male and female hamadryas baboon monkeys to evaluate the PK, biodistribution, in-life safety, gross pathology, and histopathology following a single IV administration of VNX-101 at 2.7E12 vg/kg. VNX-101 treatment was well tolerated with no significant changes in body weight or clinical signs reported. Collectively, these preclinical data support the efficacy and safety of VNX-101 as a potential AAV-based treatment for cancer. Based on our findings, we have received a “safe to proceed” from the FDA to conduct a Phase 1/2 clinical trial of VNX-101 in patients with relapsed or refractory B cell precursor ALL.