Abstract

Abstract 3496

Poster Board III-433

Hemophilia A is the most common, inherited severe bleeding diathesis and is due to deficient FVIII. We and others have shown that targeted delivery of ectopic human B-domainless (hBD) FVIII from within platelet α-granules is effective at improving clotting in several injury models using FVIIInull mice. Importantly, platelet-derived (p) hBDFVIII is ∼100-fold more effective than comparable levels of plasma hFVIII against circulating inhibitors, a significant problem in the hemophilia A population. However, we have also shown in both a cuticular injury model and an in situ cremaster laser arteriole/venule injury model that there are certain limitations to pFVIII therapy, especially concerns of clot stability and risks of clot embolization. In the present study, we examined the efficacy of platelet-delivered inactivation resistant FVIII (IR8), resistant to inactivation by thrombin and activated protein C, and canine (c) BDFVIII, a species of FVIII that has 3-5-fold greater specific activity than hBDFVIII. Both FVIIIs were expressed in platelets using both a transgenic mouse approach and lentiviral gene therapy, and compared to similarly expressed phBDFVIII mice. Multiple founder lines of mice were generated for pIR8 and pcBDFVIII using the platelet-specific glycoprotein Ibα proximal promoter to drive platelet-specific expression. Lentiviral studies involved transducing FVIIInull murine hematopoietic cells with an HIV-1 based self-inactivating lentivirus encoding the previously mentioned FVIIIs driven by either a ubiquitin promoter or by the platelet-specific platelet factor 4 (PF4) proximal promoter; lethally irradiated FVIIInull mice were rescued by injection of virally transduced bone marrow. FVIII antigenic levels in both platelet lysates and releasates were measured, and free, plasma FVIII levels were detected by a FVIII ELISA. Maximum pIR8 levels achieved were comparable to those seen with hBDFVIII in spite of the fact that IR8 binds poorly to von Willebrand factor (vWF), supporting prior observations that pFVIII is stored in α-granules independent of the presence of vWF in α-granules. Surprisingly, maximal pcBDFVIII levels were 1/3rd of those of phBDFVIII. The basis for this observation is presently being determined. The ability of the pFVIII to correct hemophilia A in vivo was tested by a FeCl3 carotid artery injury model and a cuticular bleeding model. In addition, laser injury studies to cremaster arterioles and venules are on-going. FeCl3 carotid artery and cuticular studies demonstrated that both pIR8 and pcBDFVIII were more potent than phBDFVIII in improving outcome in FVIIInull mice. For pcBDFVIII, this occurred despite the low levels of pcBDFVIII. In the cremaster laser injury model, lentiviral-based gene therapy using the three FVIII variants, markedly improved clot formation compared to FVIIInull mice, but further studies are needed to define whether a specific FVIII variant is particularly efficacious. Prior studies of plasma IR8 correction using an adenoviral delivery system showed no improved outcome relative to hBDFVIII. We propose that this difference for IR8 between plasma and platelet expression is due to IR8 binding vWF poorly and that vWF binding is important for maintaining plasma FVIII levels. The higher specific activity of pcBDFVIII appears to more than compensate for its lower level in circulating platelets. We now intend to define the molecular basis for the greater efficacy of pIR8 and pcBDFVIII compared to phBDFVIII and to use these insights to further optimize the efficacy and safety of this delivery strategy in the care of patients with hemophilia A.

Disclosures:

Poncz:Diagnostica Stago: Patents & Royalties.

Author notes

*

Asterisk with author names denotes non-ASH members.