Successful Gene Therapy of Murine Paroxysmal Nocturnal Hemoglobinuria (PNH) Using Mgmt-Mediated In Vivo Selection of Genetically Corrected, Drug-Resistant Hematopoietic Stem Cells.

PNH is an acquired disorder of the hematopoietic stem cell (HSC) caused by a somatic mutation in the PIGA gene. Blood cells carrying the mutation are deficient in all glycosyl phosphatidylinositol-anchored proteins (GPI-APs). Long-term restoration of GPI-AP has been obtained in vitro using lentiviral vectors stably expressing PIGA in human HSCs. However, gene therapy for PNH might be problematic because of the lack of a growth advantage of phenotypically corrected cells. We hypothesized that providing corrected cells with a selective growth advantage might overcome this possible limitation. We chose to test this in a murine model of PNH that has GPI-AP deficient blood cells because of a somatic Piga gene mutation targeted to the HSCs (LF mice). We designed a novel bicistronic lentiviral vector (MMIP) expressing the human PIGA cDNA and the human alkylating drug resistance mutant O6-methylguanine DNA methyltransferase (MGMT G156A) under the control of the MND promoter. MGMT G156A confers protection against alkylating agents. Eight LF donor mice were injected with 5-FU. Bone marrow (BM) was harvested 5 days later for MMIP transduction performed twice at MOI 100. Transduced and mock transduced BM cells were injected into 15 lethally irradiated C57Bl/6 recipient mice. One month after BM transplantation (BMT) 8 MMIP-transduced mice were treated with the alkylating agents BCNU and O6-benzylguanine (BG) (5mg/kg and 30mg/kg). BCNU /BC treatment was repeated twice, each one month apart. Six months after BMT the percentage of donor engraftment was 75.1 ± 25% for granulocytes (G) and 88.4 ± 26% for lymphocytes (L) in the control group (n=3). Donor chimerism in the MMIP-transduced group not treated with BCNU/BG was 84.6 ± 25 and 87.8 ± 5 % (n=4). The highest donor chimerism was obtained in mice treated with BCNU/BG, 98.9 ± 0.4 and 98.6 ± 0.5 % (n=8). Next we determined the proportion of donor blood cells with a restored expression of GPI-linked proteins. Only a small proportion of GPI-AP expressing cells was found in mice receiving the mock transduced BM six months after BMT, 22±37% GPI-AP expressing RBCs; 3.1± 2% GPI-AP expressing G, and 0.5 ± 0.2% GPI-AP expressing L. In contrast, mice receiving the MMIP-transduced BM cells but not treated with BCNU/BG showed a higher level of GPI-AP+ cells: 87.8±16% for RBCs, 70.2 ±23% for G and 56.3 ± 13% for L (n=4). However, mice receiving the MMIP-transduced BM and BG/BCNU treatment, achieved an almost complete restoration of GPI-AP on peripheral blood cells: 98±3% of RBCs, 96.4 ±4.7% of G and 89.6 ± 13% of L (n=8). Findings in BM, in methylcellulose progenitor assays, and in secondary spleen colony assays confirmed that restoration of GPI-AP and BCNU/BG selection had occurred at the HSC level. This is the first report of long-term restoration of GPI-APs expression in vivo using a murine model for PNH. Our results demonstrate that BCNU/BG treatment is well tolerated (100% survival) leading to a strong in vivo selection of HSCs with restored expression GPI-AP. The selective growth advantage of converted HSCs after dual gene therapy followed by BCNU/BG treatment allows us to achieve and possibly maintain a high level of hematopoiesis with restored GPI-AP expression. The selection for the corrected HSCs might allow them to overcome a possible growth disadvantage in the competition between PNH and normal HSCs.