Combination Therapy to Enable AAV Re-Dosing and Mitigate Transgene Immunity in Liver-Targeted Gene Therapy

Liver directed gene delivery with AAV has made tremendous strides in the last decade and is poised to cure disorders such as hemophilia A and B. Although AAV is typically considered to be relatively innocuous, immunogenicity to the vector, its genome, and encoded transgene has resulted in innate and adaptive immunotoxicities as well as loss of the transgene product, which not only limit efficacy, but also raise important safety concerns. A major impediment to patient re-dosing with AAV is the formation of high-titer neutralizing antibodies (NAb) to the viral capsid following first administration. The right combination of immunosuppressive (IS) drugs have the capacity to control NAb formation, as well as mitigate innate and adaptive immune responses to both the vector and the transgene product. We tested hepatic gene delivery with AAV8 in C57BL/6 mice (n=8-10), using a dose (1X1011 vg/kg) that was previously shown to induce CD8+ T cells to the model transgene, ovalbumin (OVA). We compared (i) B cell depletion with α-CD20, 2 X 250μg doses spaced 3 weeks apart), (ii) α-CD20 combined with the mTOR inhibitor rapamycin (α-CD20+rapa, oral gavage with 4mg/kg rapa, 3X/week/5 weeks), (iii) α-CD20 combined with an antibody to B cell activating factor (α-CD20+α-BAFF, 2 X 60μg doses spaced 10 days apart), a cytokine that we confirmed to be crucial for B cell survival [(J Clin Invest 2021, 15;131(8)]. IS therapies administered concurrent with AAV8 gene therapy caused a high frequency of animals (40 - 75%) to develop transgene immunity (H-2 Kb SIINFEKL tetramer+), corresponding with loss of OVA expression in the liver and in circulation. OVA-CD8+ T cells had a short-lived effector memory phenotype in blood and spleen, characterized as CD44+CD62L-CCR7-KLRG1+CX3CR1+. Interestingly, there were much higher frequencies of OVA-CD8+ T cells in the liver, even when they were undetectable in blood and spleen, and their presence in the liver did not completely correlate with loss of OVA expression. In line with this observation, OVA-CD8+ T cells in the liver expressed multiple exhaustion markers PD1, 2B4, LAG3, TIM3, highly upregulated transcription factors TBet, STAT3, and BATF, but not TCF1, and did not express KLRG1 and CX3CR1 expression. All control and IS treated groups of animals developed NAb (IgM and IgG2c, equivalent to IgG1 in human) to the AAV8 vector, which precluded re-administration with the same AAV serotype (5X1012 vg/kg of AAV8-hFIX). In contrast, when IS therapies were initiated 3 weeks prior to AAV8 gene therapy (1X1011 vg/kg), we were successfully able to re-administer AAV8-hFIX (n=6-10) to 70% of α-CD20, 75% of α-CD20+rapa, and 80% of α-CD20+α-BAFF treated animals without transgene immunity development. Re-administration was unsuccessful in animals that did not receive immune modulation. We then tested pre-IS therapy with a clinically relevant primary administration dose (5X1012 vg/kg of AAV8-OVA). We were able to re-administer AAV8-hFIX to 44% of α-CD20, 88% of α-CD20+rapa, and 67% of α-CD20+α-BAFF treated animals. Mice that were precluded from re-dosing with AAV8-hFIX in the IS treatment groups had high circulating BAFF levels, which correlated with the development of α-AAV8 IgM NAbs. To validate this finding and mitigate α-AAV8 IgM NAbs, we extended α-BAFF injections for an additional 4 weeks in the α-CD20+α-BAFF group (n=10). Extending α-BAFF treatment prevented both BAFF and IgM production, allowing not only successful re-dosing in 100% of animals but also higher circulating FIX levels (12901 ± 686.7ng/mL) as compared to mice that only received AAV8-FIX (9725 ± 415.5 ng/mL). Finally, we tested the practicality of α-CD20+extended α-BAFF treatment in the event of delayed AAV8 re-dosing, once immune compartments have completely repopulated (~14 weeks based on flow cytometry analysis). We were successfully able to re-dose 100% of α-CD20+extended α-BAFF pre-IS treated mice with AAV8-FIX following a delayed re-administration schedule (15 weeks). FIX expression levels remained stable until the last timepoint tested (16 weeks post re-dosing). We conclude that timing and duration of IS, vector dose and transgene immunogenicity are important considerations in developing effective IS therapies. We propose that pre-IS therapy with αCD20+ extended αBAFF is an effective candidate to enable re-dosing with clinically relevant doses of AAV.