159. Vector Genome Loss Does Not Account for Decline of Plasma cFVIII Protein in Hemophilic Dogs That Demonstrate Long Term Clinical Correction Following Helper-Dependent Adenoviral Gene Therapy

Abstract

Helper-dependent adenoviral vectors (HDV) offer the advantage of a large cloning capacity, reduced toxicity and extremely efficient transduction of target tissues. These vectors remain episomal in host target cells limiting the possibility of insertional mutagenesis but potentially reducing the duration of transgene expression due to vector genome loss. In previous studies using these vectors in various animal models of disease, long-term expression of both intracellular and secreted proteins has been achieved. However, in canine models of both hemophilia A and B, this has not been the case. The level of canine factor VIII (cFVIII) transgene product in the plasma of hemophilic dogs treated with HDV gene therapy declines at a greater rate than would be expected based on data from other models of disease including hemophilic mice. In these experiments we treated two hemophilia A dogs from the colony at the University of North Carolina with a HDV encoding cFVIII under the control of a liver tissue-restricted phosphoenolpyruvate carboxykinase (PEPCK) promoter. One animal received 1 × 1012 vp/kg and a second received 3 × 1012 vp/kg. The whole blood clotting time (WBCT) in both of the animals was significantly improved for up to 10 months. However, the plasma level of cFVIII declined from peaks at two weeks post-treatment of 1.3% normal in the low dose animal and 70% normal in the high dose animal to less than 1% at 2 months in both. To determine the reason for this a third animal was injected with 3 × 1012 vp/kg and underwent survival liver biopsy at days 18 and 79 after vector administration correlating with the high (35% normal) and low (<1% normal) levels of cFVIII in the plasma. Genomic DNA was isolated from these biopsy samples. Real-time PCR and semi-quantitative PCR indicate that there was a less than 2-fold decline in vector genome from the early to the late time point. Immunohistochemistry of liver tissue from these two time-points with an antibody directed at cFVIII reveals that the protein is localized to the endothelium of the hepatic veins and sinusoids. There does not appear to be an overall decline in cFVIII staining at the later time-point. These data suggest that vector genome loss alone is not sufficient to account for the decline of plasma cFVIII levels. The Bethesda assay in these animals was negative for inhibitory antibodies to cFVIII. However, it is possible that antibodies that do not inhibit function, but that do increase clearance of the cFVIII protein may account for the decline of plasma cFVIII.