Despite the popularity of adeno-associated virus 2 (AAV2) as a vehicle for gene transfer, its efficacy for liver-directed gene therapy in hemophilia A or B has been suboptimal. Here we evaluated AAV serotypes 2, 5, 7, and 8 in gene therapy of factor VIII (FVIII) deficiency in a hemophilia A mouse model and found that AAV8 was superior to the other 3 serotypes. We expressed canine B domain-deleted FVIII cDNA either in a single vector or in 2 separate AAV vectors containing the heavy- and light-chain cDNAs. We also evaluated AAV8 against AAV2 in intraportal and tail vein injections. AAV8 gave 100% correction of plasma FVIII activity irrespective of the vector type or route of administration.

We reported total correction of blood coagulation plasma factor VIII (FVIII) activity, using adeno-associated virus serotype 8 (AAV8) vectors for liver-specific gene transfer in hemophilia A mice. We now show, irrespective of immunosuppression or route of administration, total long-term correction of hemophilia A mice with pseudotyped AAV8 and AAV9 vectors. We delivered two FVIII vectors, one expressing canine heavy chain and the other expressing canine light chain. Interestingly, when these vectors were given by hepatic portal vein to hemophilia A dogs, only modest FVIII levels were seen despite the species-specific transgene. No dogs treated developed FVIII inhibitors. However, of three dogs treated with AAV8 vector, the single male, given 1.25 x 10(13) genome copies per vector per kilogram (GC/vector/kg), maintained a level of >4.5% for more than 2 years. In contrast, the two female dogs expressed only 2% FVIII activity despite receiving higher doses of 1.52 x 10(13) and 3 x 10(13) GC/vector/kg, respectively. On the other hand, a male dog treated with AAV9 vector at a low dose (6 x 10(12) GC/vector/kg) maintained FVIII levels of 2-2.5% of normal without bleeding for 200 days (observation ongoing). Although hemophilia A mice were not predictive of vector efficacy in dogs, the two treated male dogs became symptom-free for long periods. Even so, translation of these robust vectors either in appropriate large animals or human beings remains challenging.

No abstract