In Vivo Selection Of Genetically Manipulated Hematopoietic Stem Cells For Platelet Gene Therapy Of Hemophilia A

Our previous studies have demonstrated that targeting FVIII expression to platelets (2bF8) by lentiviral (LV) gene delivery to hematopoietic stem cells (HSCs) corrects bleeding diathesis in hemophilia A mice with or without inhibitors. Although the bleeding diathesis is improved in transduced recipients, the transduction efficiency using our current 2bF8 LV, is only about 10%, resulting a median level of platelet-FVIII (Plt-F8) of 1.5 mU/10⁸ platelets even thought a myeloablative conditioning regimen was employed. It has been shown in clinical trials that efficient stem cell gene transfer and myeloablation is not required when there is a powerful selective advantage to the genetically modified cells. We hypothesize that incorporating a drug-resistance gene into the 2bF8 LV construct will allow for in vivo selection of 2bF8 LV-transduced cells which will result in the increase of therapeutic levels of Plt-F8 for hemophilia A gene therapy and reduce the potential for genotoxicity. To address our hypothesis, we constructed a new lentiviral vector, pWPT-2bF8/MGMT, which harbors dual genes, the 2bF8 gene and a drug-resistance gene, the MGMT^P140K cassette. To explore the feasibility of the MGMT-based in vivo selection system, HSCs from FVIII^null mice were transduced with 2bF8/MGMT LV at an MOI (multiplicity of infectious) of 1, which is 1/10 of the MOI used for our regular 2bF8 LV transduction, and transplanted into littermates pre-conditioned with a non-myeloablative regimen, 660 cGy total body irradiation (TBI). We chose a low MOI because one of the goals of using the MGMT selection system is to reduce the potential for genotoxicity. After bone marrow (BM) reconstitution, the recipients were treated with O⁶-benzylguanine (BG) followed by 1, 3-bis-2 chloroethyl-1-nitrosourea (BCNU) monthly for 3 or 4 times. As determined by a chromogenic assay on platelet lysates, functional Plt-F8 expression in recipients was only 0.22 ± 0.15 mU/10⁸ platelets before the drug treatment, but remarkably increased to 4.33 ± 5.48 mU/10⁸ platelets (n = 16) after BG/BCNU drug-selective treatments. The levels of Plt-F8 in the untreated transduced control group remained low over the study period. FVIII activity was not detected in the plasma in any of the recipients even with Plt-F8 as high as 22 mU/10⁸ platelets. The average copy number of 2bF8/MGMT proviral DNA per cell was determined by quantitative real-time PCR. 2bF8 proviral DNA was barely detectable (0.01 ± 0.02 copies/cell) in recipients before drug-selective treatment, but it increased to 0.42 ± 0.15 copies/cell after BG/BCNU treatments, confirming that 2bF8/MGMT genetically modified cells were effectively enriched in vivo after drug-selective treatment. When the tail clip survival test was used to assess phenotypic correction of the FVIII^null coagulation defect, 15 of 16 treated animals survived the tail clip challenge; in contrast, none of the untransduced FVIII^null control mice survived. When ROTEM analysis was used to determine the whole blood clotting time (CT), the CT was shortened from 3043 ± 728 seconds (n = 7) to 931± 273 seconds (n = 6) (P < 0.0001) in treated transduced recipients when compared to FVIII^null mice. There was no significant difference between wild type (722 ± 270 seconds, n = 7) and treated recipients. To ensure sustained Plt-F8 expression in BG/BCNU treated transduced recipients, some primary recipients were sacrificed 9 months after transplantation and BM mononuclear cells were transplanted into secondary recipients. Platelet lysate FVIII activity assays showed that the levels of Plt-F8 in secondary recipients were similar to those in primary recipients, confirming that long-term repopulating HSCs were successfully genetically modified by 2bF8/MGMT LV. When a low intensity pre-conditioning regimen of 440 cGy TBI was used, the levels of Plt-F8 increased from 0.06 ± 0.12 mU/10⁸ platelets to 1.86 ± 2.06 mU/10⁸ platelets after BG/BCNU drug-selective treatment. It is notable that no anti-FVIII inhibitory antibodies were detected in the treated recipients even after rhFVIII challenge, indicating that immune tolerance was induced in the treated animals. In contrast, all FVIII^null mice under the same challenge developed various levels of inhibitors. Taken together, we have established a powerful in vivo selective system that allows us to enrich 2bF8 LV-transduced cells and to enhance platelet-FVIII expression for hemophilia A gene therapy.