Ex Vivo Gene Therapy for Haemophilia A: FVIII-Overexpressing Mesenchymal Stromal Cells Fail To Produce Sustained Therapeutic Levels of Circulating FVIII In Vivo.

The use of ex vivo transfected autologous fibroblasts was explored for the delivery of coagulation factor VIII (FVIII) to treat haemophilia A patients, albeit with mitigated results (Roth et al., NEJM 2001). A limiting factor to this approach is the limited expansion potential of somatic cells and cellular dose manufactured for therapeutic use. The use of adult stem or progenitor cells offers a possible remedy to this specific challenge because of their high proliferative capacities, allowing the generation of large number of cells ex vivo. Mesenchymal stromal cells (MSCs) can easily be isolated by bone marrow aspiration and extensively expanded ex vivo using standard tissue culture procedures. They also possess endogenous wound healing properties and are relatively easy to gene engineer. When engineered to secrete therapeutic plasma-soluble proteins, they could be injected with a supporting matrix - as a subcutaneous plug - and used to treat a variety of haematological disorders responsive to secreted proteins such as haemophilia A and B and anaemia for instance. The retrievability of the implanted cells is desirable if unexpected side effects were to occur. The aim of this project was to test a gene engineered MSC platform to treat a clinically relevant model disease, haemophilia A. Canine MSCs (cMSCs) isolated from normal and haemophilia A dogs were genetically modified using a 3rd generation lentiviral vector carrying the B-domain deleted canine FVIII (cFVIII) gene. In vitro production of cFVIII was assessed by ELISA and found to be about 2,4 U/106 cells/24h. 107 of those cells were injected intraperitoneally in NOD-SCID mice and transient expression (8–10 days) of therapeutic levels of cFVIII (up to 10%) could be detected using an ELISA capable of discriminating cFVIII from murine FVIII. However, when these cells were embedded in an injectable human compatible collagen based scaffolding material (Contigen) and implanted subcutaneously into NOD-SCID mice no cFVIII could be detected above background in any animals. Because we previously demonstrated that erythropoietin gene-modified MSCs implanted as described led to an increase in hematocrit for an extended period (Eliopoulos et al. Mol Ther 2004), we believe that the failure of the present approach is attributable to idiosyncrasies of the FVIII system. Of particular relevance is the requirement for the presence of von Willebrand factor (vWF) in stabilizing the FVIII protein and protect it from degradation. We already have evidence that ex vivo transduced endothelial progenitor cells known to produce endogenous vWF can serve as better delivery vehicles for FVIII as part of a matrix embedded subcutaneous plug and lead to prolonged circulating levels of FVIII in mice (Matsui et al., submitted). These data support the hypothesis that FVIII delivery by engineered somatic and stem cells likely requires co-expression of key cofactors, such as vWF, for durable and clinically relevant effects in vivo.