Success of Immune Tolerance Induction after AAV Gene Therapy in High-Responding Inhibitor Hemophilia a Dogs May be F8 Genotype Dependent

Background

The development of neutralizing alloantibodies (inhibitors) remains the major side effect of factor VIII (FVIII) replacement therapy for hemophilia A (HA). Inhibitors significantly increase morbidity and mortality. Though the advent of non-factor therapies for HA has dramatically improved hemostatic prophylaxis for inhibitor patients, inhibitor eradication and immune tolerance induction (ITI) remain essential therapeutic goals. Recombinant adeno-associated viral (rAAV) gene therapy for HA has almost completely excluded patients with current or a history of inhibitors. We have previously published that rAAV gene therapy with canine (c) FVIII induced immune tolerance in 5 HA dogs with inhibitors (Finn et al 2010, Doshi et al 2024). HA dogs are an outbred, long-lived, naturally occurring, large animal model that both recapitulates the HA bleeding phenotype and consistently develops inhibitors in a species-specific manner. These 5 previously described dogs all had F8 intron 22 inversion (F8-INV22), which is the most common genotype in severe HA patients and dogs. Both their inhibitor titer at the time of rAAV administration and their peak titer prior to rAAV was ≤21 Bethesda Units (BU). The success rate of ITI in inhibitor patients with large gene deletions is <20% (Coppola et al 2009), threefold lower than the success rate in patients with F8-INV22 and similar translocations. Here, we report results of an ongoing preclinical study evaluating the hypothesis that rAAV gene therapy can induce immune tolerance in these more challenging clinical scenarios, including higher historical inhibitor titers and large F8 gene deletions.

Methods

To test this hypothesis, we developed 2 novel HA canine models with high-responding inhibitors: a) 2 male littermates with F8-INV22 (Duck Toller/Spaniel cross) with high-responding inhibitors; b) an HA dog (Mountain Cur) with a previously undescribed large F8 multi-exon deletion (exons 22-25) and high-responding inhibitors. All studies are approved by the UNC Institutional Animal Care and Use Committee. rAAV vectors were serotype 8 with a previously reported high-expressing cFVIII variant (Doshi et al 2024).

Results

High-Responding F8-INV22: T01 and T02 received 9E12 vg/kg rAAV when their inhibitor titers were >80 BU. Within 1 year, both exhibited inhibitor eradication and anti-cFVIII IgG2 normalization. This inhibitor eradication was consistent with stringent immune tolerance, as evidenced by the absence of inhibitor recurrence despite repeated challenges with cFVIII protein. cFVIII activity has been between 2-7% normal with significant reduction in bleeds for >8 years (ongoing study).

F8 multi-exon deletion: This model’s proband S38 received 1E13 vg/kg rAAV when his inhibitor titer was 600 BU. He had initial detectable spike of cFVIII antigen level, but his course was prolonged by an anamnestic response (peak 745 BU) that occurred after a life-threatening bleed (hemoglobin 3 g/dl) requiring whole blood transfusion. Subsequently, he demonstrated steadily declining inhibiter titer and anti-cFVIII IgG2. However, he was euthanized for humane reasons after another life-threating hemorrhage; at the time of euthanasia (2 years after rAAV), his inhibitor titer and declined >90% its peak value.

5 puppies produced by cloning of S38 were challenged weekly with cFVIII 25 U/kg. All 5 developed inhibitors > 100 BU after the 3rd dose. One (Z10) received 6E12 vg/kg rAAV when his inhibitor titer was 438 BU. He demonstrated >50% reduction in his inhibitor titer until day 73 when he had a life-threatening bleed requiring whole blood transfusion. He then had an anamnestic response to >1000 BU that by day 147 had fallen to 254 BU in this ongoing study.

Conclusions

rAAV gene therapy with a single-administration can provide HA dogs with F8-INV22 genotypes—even with preexisting high-responding, high-titer inhibitors: 1) inhibitor eradication, 2) sustained immune tolerance, and 3) continuous FVIII expression and phenotypic amelioration. However, as is observed with current protein-based ITI approaches, rAAV-based ITI in HA dogs with large F8 deletions remains very challenging. These data may inform ongoing clinical trials evaluating rAAV gene therapy for HA inhibitor patients.