918. Autologous canine bone marrow stromal cells retrovirally engineered with the canine EPO gene and implanted in vivo in a collagen matrix produce and release systemically functional EPO over a prolonged period

Many features of dogs make them suitable animal models to study new therapeutic approaches for human diseases: a large body mass, a long life span and spontaneous diseases such as chronic renal failure (CRF) that leads to anemia. This, and the longevity of canine red blood cells (RBC) comparable to that of humans, render dogs particularly suitable to develop new treatments of CRF-induced anemia. In contrast, rodents differ from humans by a much smaller body mass, shorter longevity of RBC (43 days) and the absence of spontaneous diseases resembling those of people. Current treatment of CRF-induced anemia in people (and dogs) consists of repeated injections of recombinant human EPO (rhEPO), which incurs high costs and specialized personnel. Our hypothesis is that the implantation of an “organoid” composed of autologous cells embedded in an organic matrix and engineered to produce and systemically release EPO can correct CRF-induced anemia in dogs for a long period, replacing the need for repeated injections of rhEPO. As a first step towards this goal, we demonstrated the feasibility of this approach in normal dogs. We used bone marrow stromal cells (BMSCs) for the sustained production of EPO in vivo because these cells are easily accessible, transduceable and expandable in vitro. We isolated and transduced canine BMSCs with the canine EPO gene using a bicistronic retrovector. Using a MOI of 60, the transduction efficiency was 60 to 90% as analysed by FACS; the in vitro production and release of canine EPO was verified by Western blot and measured by ELISA (2 U of EPO/106 cells/24 hrs). The incorporation of engineered cells in a biocompatible solid matrix makes it possible to remove the implant in case of over expression of the transgene or other complications. To demonstrate the functionality of the secreted protein, the canine EPO-engineered BMSCs were embedded in bovine type I collagen (Inamed™), and the resulting organoid was implanted in 5 NOD-SCID mice (7 × 106cells/500 μl collagen/mouse). Four control mice were treated identically except that the implanted BMSCs (from the same dog) were transduced with empty retroparticles. The hematocrit (Ht) of treated mice increased from 53% to 88% in 33 days and remained elevated over 3 months following implantation. The canine EPO-engineered BMSC were embedded within the same matrix and implanted s.c. in the respective donor dogs. The number of embedded BMSCs and volume of implants were optimized: 32 implants were given in 11 injections (3 implants per injection, each implant containing 4 × 106 BMSCs). This protocol has been implemented in 3 dogs whose Ht has increased from 42 (SD: ±6.9) to 55 (SD: ±2) within 12 days and has remained elevated for 4 weeks. These results clearly demonstrate the feasibility of cellular therapy in large immunocompetent mammals, a crucial step towards its application in people. We will next carry out the optimized protocol on canine and feline patients affected by CRF-induced anemia.