We report a case of juvenile thrombophilia associated with a substitution of leucine for arginine at position 338 (R338L) in the factor IX gene (factor IX-R338L). The level of the mutant factor IX protein in plasma was normal, but the clotting activity of factor IX from the proband was approximately eight times the normal level. In vitro, recombinant factor IX-R338L had a specific activity that was 5 to 10 times as high as that in the recombinant wild-type factor IX. The R338 substitution causes a gain-of-function mutation, resulting in factor IX that is hyperfunctional.

• The novel FVIII variant (FVIII-RH) has enhanced stability and procoagulant activity in both in vitro and in vivo models.• FVIII-RH is efficacious and safe; thus, it is an attractive molecule for protein replacement and as a transgene in gene-therapy strategies.Recombinant canine B-domain deleted (BDD) factor VIII (FVIII) is predominantly expressed as a single-chain protein and exhibits greater stability after activation compared with human FVIII-BDD. We generated a novel BDD-FVIII variant (FVIII-RH) with an amino acid change at the furin cleavage site within the B domain (position R1645H) that mimics the canine sequence (HHQR vs human RHQR). Compared with human FVIII-BDD, expression of FVIII-RH protein revealed a 2.5-fold increase in the single-chain form. Notably, FVIII-RH exhibited a twofold increase in biological activity compared with FVIII-BDD, likely due to its slower dissociation of the A2-domain upon thrombin activation. Injection of FVIII-RH protein in hemophilia A (HA) mice resulted in more efficacious hemostasis following vascular injury in both the macro-and microcirculation. These findings were successfully translated to adeno-associated viral (AAV)-based liver gene transfer in HA mice. Expression of circulating FVIII-RH was approximately twofold higher compared with AAV-FVIII-BDD-injected mice. Moreover, FVIII-RH exhibits superior procoagulant effects compared with FVIII-BDD following a series of hemostatic challenges. Notably, the immunogenicity of FVIII-RH did not differ from FVIII-BDD. Thus, FVIII-RH is an attractive bioengineered molecule for improving efficacy without increased immunogenicity and may be suitable for both protein-and genebased strategies for HA. (Blood. 2013;121(21):4396-4403)

694 Hemophilia B (HB) is a severe bleeding disorder affecting 1 in 30,0000 men and results from the deficiency of FIX clotting activity. The current treatment involves infusion of either plasma derived or recombinant FIX. We have recently described a novel naturally occurring gain-of-function mutation in the FIX gene found associated with hyperfunctional activity of the protein (N Engl J Med, 2009). The proband presented with 776% FIX clotting activity (FIX:C), but normal FIX antigen levels (FIX:Ag of 92% of normal), and after detailed analysis the causative mutation was found to be a transversion of R338L in FIX (FIX-Padua) that segregates in a classical X-linked fashion. Furthermore, in vitro expression of FIX-Padua confirmed that the protein had a specific activity that was 8 fold higher than FIX-wild type. Arginine at 338 is highly conserved in FIX from mammals but is unique among other human vitamin K-dependent proteins. We sought to determine whether FIX-Padua would be an attractive strategy for novel therapy of hemophilia B by introducing this mutation in the canine FIX gene (cFIX) and testing in HB dog models. The HB colony at Chapel Hill is an outbred, immunocompetent large animal model that have a missense mutation in the cFIX gene and have no detectable circulating levels of cFIX antigen or activity, and its disease phenotype closely correlates with that seen in humans. Adeno-associated viral (AAV) vector encoding the cFIX-Padua and cFIX-wt under control of the CMV promoter was generated and administered to HB dogs. AAV6 was chosen as the serotype to overcome the high prevalence of neutralizing antibodies to AAV2 in the general population. Delivery of AAV6-cFIX wild type by intravascular delivery to the skeletal muscle via anterograde limb perfusion (ALP) coupled with transient immunosuppression (duration of 5 weeks) resulted in FIX:Ag and FIX:C of 2-5% of normal with partial shortening the whole blood clotting time. Post AAV6-cFIX Padua injection we observed a remarkable correction of the WBCT to the normal range (8-10 min) within the first 7 days that has continued for the duration of the study. One dog's (M55) FIX:Ag levels were ∼250 ng/mL (5% of normal) at the peak and 180 ng/mL (3.6%) at the plateau. Notably, his FIX:C levels were 49% at peak and 35% at plateau, indicating a specific activity ∼10 fold higher than wild type cFIX. Treatment of a second dog (M59) resulted in FIX:Ag levels of 175 ng/mL (3.5%) at the peak and 60 ng/mL (1.2%) at plateau. This is in contrast with his FIX:C levels that were 35% at peak and 12% at plateau levels, respectively. Like M55, the specific activity was 10 fold higher than wild type cFIX. An important consideration is the potential for an immune response against the cFIX-Padua neo-antigen. Notably, there is no evidence of formation of inhibitor to cFIX in these dogs for a cumulative observation period of 6 months post-immunosuppression. We have demonstrated that the cFIX-Padua mutant protein has 10-fold higher specific activity in a canine model of HB. This is significant as it allows for lower doses of AAV vectors to be used for the treatment of HB without compromising efficacy, thereby increasing the safety profile of clinical gene therapy for HB. In addition, FIX Padua provides an excellent alternative to overcome the limited ability of skeletal muscle to synthesize fully posttranslational modified functional FIX (Blood, 2001). Thus, FIX-Padua improves both the efficacy and safety of skeletal muscle-mediated FIX resulting in FIX:C levels sufficient to convert the severe HB phenotype to mild hemophilia (10-35% of normal) at vector doses already tested clinical trials. This strategy has the potential of improving other gene and/or cell therapy for HB targeting ectopic targets as well as the liver. Disclosures: High: Genzyme: Patent licensed to Genzyme, but Dr. High has waived all financial interest, Patents & Royalties.

Direct intramuscular injection (IM) of adeno−associated viral (AAV) serotype 2 in humans with hemophilia B (HB) is a promising therapeutic strategy since muscle biopsies obtained >3 years after vector injection demonstrated stable local gene expression. However to achieve therapeutic FIX levels using AAV−2 would required hundreds IM injections. The use of alternate AAV serotypes is an attractive strategy since AAV−1 or AAV−6, resulted in a >10−fold increase in transgene levels compared to AAV−2 in large animals, but the immune response to the transgene product has been consistently demonstrated as a major limitation of this strategy. There is growing evidence that blood proteases play an important role in modulating inflammatory and immune responses through activation of PARs. Mice lacking PAR−1(−/−) or PAR−2(−/−) alleles presented amelioration of immune− or infection−mediated diseases. Here we sought to determine whether inhibition of PARs could be used as a strategy to prevent immune responses to the FIX following AAV−mediated gene transfer to skeletal muscle. We used PAR−1 and PAR−2 knockout mice on C57Bl/6 background and littermate mice received IM injection of AAV1−CMV−hFIX. At dose 5x1011vg/kg, PAR−2 (−/−) mice (n=5) exhibited circulating FIX levels of 500± 99ng/ml (8–10%) which remained stable for the duration of the experiment (10 weeks), and no antibodies for FIX were detected (n=5). In contrast, all PAR−2 (+/+) mice (n=4) developed antibodies to FIX which inhibits FIX clotting activity, as determined by Bethesda assay (2.1± 0.6 BU). However, when similar vector doses were delivered to PAR−1(−/−) or PAR−1(+/+) (n=4/genotype) mice, antibodies to FIX developed in all animals. We next tested a higher vector dose in PAR−2 models. At dose 1x1012vg/kg, PAR−2(−/−) mice(n=7) resulted in FIX levels of 1,500±353ng/ml, and again no antibodies for FIX were detected. At the same dose, 6 out of 10 mice of PAR−2 (+/+)/(+/−) developed inhibitory antibodies (1.8± 0.7 BU). Further increase in the vector dose to 5 x 1012 vg/kg resulted in the development of inhibitor to FIX in both PAR−2 (−/−) (4/11 mice, 36%) and PAR−2(+/+)/(+/−) (10/17 mice, 60%). This suggests a threshold value in the protective effect in the PAR−2 (−/−) model. We sought to assay for FIX−specific T−cell by ELISPOT assay to quantify IFN−γ secretion from splenocytes of PAR−2 (−/−) and PAR−2 (+/+) mice injected at 5x1011 or 1x1012vg/kg. No difference in IFN−γ secretion was observed between PAR−2 (−/−) and their controls. Moreover, upon repeated challenges with FIX protein following vector injection antibody to FIX was detected in only 1/4 PAR−2 (+/+) mouse and none of 5 PAR−2 (−/−). Thus, PAR−2 inhibition does not compromise the tolerance to FIX. In a different model, intravenous injection of FIX protein into normal mice upon simultaneous activation of PAR−2 by using specific agonist peptide the rates of FIX antibody formation were comparable with those of a control peptide group. Thus, PAR−2−mediating antibody formation to FIX may differ among distinct immunologic challenges. Together, these data suggest that PARs play a role in the immune response to FIX and that inhibition of PAR−2 (but not PAR−1) could be a novel target in preventing inhibitor formation in hemophilia gene therapy and potentially for protein−based therapy.

2205 F8 ectopically expressed during megakaryopoiesis is stored in platelet (p) a-granules and released at sites of injury. This F8 is effective in hemostasis even in the presence of circulating inhibitors so that pF8 represents a novel strategy for therapy for patients with hemophilia A with inhibitors. However, cremaster laser injury studies showed that pF8 release from a-granules has a distinct temporal and spatial availability that leads to clot instability. We proposed that F8s, such as canine (c) B-domainless (B) F8, with greater specific activity than human (h) BF8, would prevent pF8 clot instability. We confirmed our thesis; however at the same time, pcBF8 levels were only ∼30% of phBF8 whether studied in transgenic mice models or by lentiviral/bone marrow transplantation (lenti/BMT) into F8null mice. This was surprising because in baby hamster kidney cells, cBF8 was secreted into the media at three times greater than hBF8. This lower level in murine megakaryocytes (Megs) was not due to mRNA levels as shown by qRT-PCR analysis. Using cultured marrow from transgenic mice expressing hBF8 or cBF8, there was a marked decrease in relative number of Megs in both pF8s compared to wildtype (WT) Megs, with the pcBF8 cells showing a greater decrease (49 ± 5% for cBF8 Megs vs. 36 ± 8% for phBF8, p < 0.02). Cultured Megs from pcBF8 and phBF8 mice both showed increased numbers of small, low ploidy Megs, and this was again higher for pcBF8 Megs (58 ± 9% for cBF8 Megs vs. 32 ± 6% for phBF8, p < 0.02). TUNEL studies as an indication of apoptosis showed that Megs expressing either F8s showed significant increased apoptosis than WT Megs with pcBF8 showing 37 ± 22% vs. 19 ± 11% in phBF8. The above data show that pcBF8 is deleterious to Meg development and was supported by a retrospective analysis of lenti/BMT pF8 platelet counts where recipient mice platelets made up a higher % of recovered platelet counts (17% ± 3% for pcBF8 vs. 4 ± 1% for phBF8, p < 0.05). We then tested whether we can separate the greater specific activity affect of pcBF8 from its deleterious affect on megakaryopoiesis. Our group has previously shown that cBF8 may have increased stability because it is predominantly expressed as a single chain, likely involving an R1645H (RH) substitution at a conserved PACE/furin site in most F8 species. Lenti/BMT pF8 studies expressing phBF8RH showed that this pF8 was expressed at the same level in reconstituted mice as phBF8, but was more efficacious in several bleeding models, including near-normal hemostasis in the cremaster laser injury model in F8null recipient mice, thus becoming the first lenti/BMT pF8-expressing F8null mouse with near-normal hemostasis in a F8null setting. Preliminary Meg count and apoptosis studies show that phBF8RH is no more deleterious to Megs than phBF8. Thus, our studies point out that F8 is deleterious to Megs with some species being more deleterious than others. This apoptosis limits F8 levels in Megs, and this deleterious effect can influence post-BMT outcome. We also present a model of how one can take advantage of a F8 variant that had a high specific activity while avoiding its low expression levels in Megs. Thus our studies provide important new insights into the biology of pF8 that may be important in developing this platelet-delivery strategy for the treatment of hemophilia A. Disclosures: Pipe: Baxter BioScience: Consultancy, Honoraria, Membership on an entity's Board of Directors or advisory committees; Novo Nordisk: Honoraria, Membership on an entity's Board of Directors or advisory committees; Inspiration Biopharmaceuticals: Honoraria, Research Funding; CSL Behring: Honoraria.