Hemophilia A is a bleeding disorder due to Factor VIII (FVIII) deficiency. FVIII is a very immunogenic protein, as approximately 30% of patients develop inhibitory antibodies (inhibitors) after FVIII protein replacement therapy. In addition, inhibitors often develop after gene therapy for hemophilia A. Identification of a treatment that could prevent inhibitor formation would be important. In this study, we explored the possibility of inducing tolerance to FVIII by neonatal gene transfer. An amphotropic gamma retroviral vector (RV) expressing human B domain-deleted FVIII (hFVIII) from the human α1-antitrypsin promoter was used to define the level of hFVIII that is necessary to achieve tolerance. Hemophilia A mice were injected with 1010 (high), 109 (medium) or 108 (low) transducing units (TU)/kg of RV at 2 to 3 days after birth, which resulted in expression of 63%, 7.3% or <2% of normal hFVIII antigen, respectively. Animals that received the high or medium dose of RV achieved hemostasis in vivo, although those that received the low dose did not. None of the mice produced antibodies to hFVIII. Similar results were also achieved in normal C3H mice, although some animals with low expression (2% of normal) developed low levels of anti-hFVIII antibodies. In contrast, hemophilia A and C3H mice that were challenged with hFVIII protein made potent inhibitors. Thus, neonatal gene transfer does not induce antibodies to hFVIII if the level of expression achieved is sufficiently high (>2x10−9 M). Mice that are tolerant to gene transfer are being challenged with recombinant BDD-hFVIII protein to determine if they are truly tolerant. Although induction of tolerance with neonatal gene transfer in mice is encouraging, tolerance has been more difficult to achieve in large outbred animals because their immune systems are more mature. Cats appear to have a much more mature immune system at birth than mice, as they mount a potent cytotoxic T lymphocyte response to canine iduronidase after neonatal gene therapy. Cats were therefore chosen as a large animal model in which to study tolerance induction with neonatal gene transfer. Neonatal cats were injected IV with 8.5x108 TU/kg of RV (medium dose) at day 5 after birth. Four cats achieved 37±3 ng/ml hFVIII (19% of normal) at 2 weeks after RV transduction. Three cats maintained hFVIII expression for 4 months. The two cats with the highest level of expression (~40 ng/ml) never developed antibodies, while the cat with a medium level of expression (30 ng/ml) developed only very low levels of anti-hFVIII antibodies. The cat with the lowest level of expression (20 ng/ml) lost expression at 2 months after RV transduction, and developed high titer inhibitors. We conclude that neonatal gene transfer does not induce antibodies to hFVIII in most cats. However, a high level of hFVIII expression may be necessary to achieve tolerance after neonatal gene transfer in cats. RV-treated cats will be challenged with hFVIII protein to determine if they are truly tolerant. In addition, normal cats will be challenged with hFVIII protein to determine the frequency of inhibitor formation in cats. Since the cat immune system is relatively mature at birth, these data are encouraging that neonatal gene or protein therapy might induce tolerance to hFVIII in humans. (This project is supported by National Hemophilia Foundation and the Bayer Hemophilia Awards Program.)