Hemophilia A is an X-chromosome-linked bleeding disorder due to a deficiency in clotting factor (F) VIII, affecting ∼1:5000 males. Inhibitor development to circulating FVIII is not an uncommon complication for patients receiving protein replacement therapy. We had proposed that the ectopic expression of FVIII in platelet alpha-granules will be targeted to sites of vascular injury, and be protected from circulating FVIII inhibitors. We and others have confirmed that this platelet (p)FVIII is in fact effective in a number of hemostatic models and is protected from circulating inhibitors. However, pFVIII has different temporal and spatial availability from plasma FVIII that may underlie the limited efficacy of pFVIII observed in some hemostatic models using FVIIInull mice. No group has been able to achieve levels greater than the equivalent of ∼200 mU/ml of whole blood pFVIII levels using standard human (h) B-domainless FVIII (hBFVIII). We therefore sought to improve pFVIII efficacy by expressing higher levels and/or by using pFVIII species with increased activity. We tested canine (c) BFVIII because of its reported higher efficacy in vitro and in mice, and because it is secreted at 1.5–2 fold HIGHER levels than hBFVIII in cell lines. We found that the level of pcBFVIII protein in Megs was actually ∼3-fold LOWER than phBFVIII. In spite of this lower level, pcBFVIII corrected the bleeding diathesis in FVIIInull mice more effectively than phBFVIII. Using Megs from transgenic FVIIInull mice that expressed pcBFVIII or phBFVIII, we found that mRNA levels of the FVIIIs were comparable. However, compared to littermate pFVIIInull mice, FVIIInull mice that expressed either pFVIII had lower ploidy Megs and these Megs showed higher levels of apoptosis. This apoptosis was enhanced by increasing the pFVIII level in the Megs by exposure to sodium butyrate. These effects were more marked in the pcBFVIII- than phBFVIII-expressing Megs. We then tested whether this phenomenon was limited to murine Megs, but found that lentiviral expression of pBFVIII increased Meg apoptosis whether the Megs were derived from mice, canine or human marrow, and that in each, levels of pcBFVIII were ∼3-fold lower than phBFVIII. In vivo studies support these in vitro apoptotic effects of pFVIII on Megs: In transgenic mice, steady-state platelet counts were normal in pFVIII mice, but TPO levels were 4-times higher in the phBFVIII and 8-times higher in the pcBFVIII compared to littermate controls (p<0.02 and p<0.0002 for phBFVIII/FVIIInull and pcBFVIII/FVIIInull vs. FVIIInull, respectively). In FVIII lentiviral transfected marrow/bone marrow transplant (lenti/BMT) studies, platelet counts did not differ in recovering recipient mice to negative controls, but both pBFVIII-expressing recipient lenti/BMT mice showed a peak in recipient platelets contributing to the total platelet mass at 4 weeks that decreased over the following month. This was especially true for the pcBFVIII-expressing lenti/BMT mice wherein >12% of the platelets were recipient-derived at 4 weeks compared to <2% in mice receiving a control lentivirus-transfected marrow (p<0.002). Thus, these studies suggest that there is a limit as to how much pFVIII can be stored in platelets, so efforts should be focused on increasing pFVIII activity. As an example of this approach, we had identified an R1645H Furin/PACE cleavage site difference between hBFVIII and cBFVIII that increases the level of single-chain to two-chain FVIII released from cell lines. The more single-chain FVIII, the slower the loss of FVIII activity, and the more overall activity was seen. We introduced this R1645H substitution into phBFVIII Megs and show antigen levels of this phBFVIIIR1645H comparable to phBFVIII in vitro and in lenti/BMT-reconstituted FVIIInull mice. Excitingly, these phBFVIIIR1645H lenti/BMT FVIIInull mice demonstrated normalization of hemostasis in several hemostatic models. Since pFVIII tends to be released deep inside a growing thrombus from degranulating platelets with limited washout, a more stable version of FVIII such as FVIIIR1645H may be particularly efficacious as pFVIII, and we intend to pursue this hFVIII variant in larger animal studies as it may be especially useful for clinical gene therapy.
No relevant conflicts of interest to declare.
Asterisk with author names denotes non-ASH members.