Human miR223 promoter as a novel myelo-specific promoter for chronic granulomatous disease gene therapy

Abstract

Targeting transgene expression to specific hematopoietic cell lineages could contribute to the safety of retroviral vectors in gene therapeutic applications. Chronic granulomatous disease (CGD), a defect of phagocytic cells, can be managed by gene therapy, using retroviral vectors with targeted expression to myeloid cells. In this context, we analyzed the myelospecificity of the human miR223 promoter, which is known to be strongly upregulated during myeloid differentiation, to drive myeloid-restricted expression of p47(phox) and gp91(phox) in mouse models of CGD and in primary patient-derived cells. The miR223 promoter restricted the expression of p47(phox), gp91(phox), and green fluorescent protein (GFP) within self-inactivating (SIN) gamma- and lentiviral vectors to granulocytes and macrophages, with only marginal expression in lymphocytes or hematopoietic stem and progenitor cells. Furthermore, gene transfer into primary CD34+ cells derived from a p47(phox) patient followed by ex vivo differentiation to neutrophils resulted in restoration of Escherichia coli killing activity by miR223 promoter-mediated p47(phox) expression. These results indicate that the miR223 promoter as an internal promoter within SIN gene therapy vectors is able to efficiently correct the CGD phenotype with negligible activity in hematopoietic progenitors, thereby limiting the risk of insertional oncogenesis and development of clonal dominance.

FIG. 1.

Performance of miR223 promoter–containing lentiviral vectors in the human myelomonocytic X-CGD cell line PLB985. (A) FACS analysis of gp91^phox expression in undifferentiated (gray) and differentiated (black) X-CGD-PLB985 cells transduced with 223gpW at MOIs 2 and 0.2 along with nontransduced controls. (B) NADPH oxidase activity in 223gpW-transduced X-CGD-PLB985 cells assessed by the DHR assay. (C) Correlation of VCN and GFP expression levels in undifferentiated (diamonds) and differentiated (circles) PLB985 clonal cell populations transduced with 223GFP. (D) Fold upregulation of GFP expression levels during granulocytic differentiation in clonal populations derived from 223GFP-transduced PLB985 cells. Variation of expression levels is given as percentage of coefficient of variation (%CV). (E) Cumulative long-term stability of GFP expression in 223GFP-transduced PLB985 cells during a 4-month culture period in differentiated and undifferentiated clonal populations. CGD, chronic granulomatous disease; DHR, dihydrorhodamine 123; GFP, green fluorescent proten; MOI, multiplicity of infection; NADPH, nicotinamide adenine dinucleotide phosphate; VCN, vector copy number.

FIG. 2.

Expression of gp91^phox and NADPH oxidase activity after in vitro culture of 223gpW-transduced murine X-CGD Lin– BM cells. (A) 223gpW-transduced murine Lin– cells were analyzed for gp91^phox expression in stem and progenitor cells (lower quartiles) and in granulocytes (upper quartiles). (B) Functional reconstitution of NADPH oxidase activity in X-CGD neutrophils mediated by the recombinant expression of gp91^phox as assessed by the DHR assay. BM, bone marrow; Lin–, lineage-negative.

FIG. 3.

In vivo expression profile of miR223 promoter–driven gp91^phox or p47^phox vectors. (A) Lentivirally (223gpW) transduced Lin– BM cells from gp91^phox–/– (CD45.2) mice were transplanted into irradiated SJL mice (CD45.1). Expression of gp91^phox in different donor-derived hematopoietic lineages was determined 23 weeks after transplantation by flow cytometry. Shown in overlay histograms: gray, gp91–X-CGD-PLB985 cells; black, transduced sample. (B) Analysis of the NADPH oxidase activity in peripheral blood granulocytes derived from the mice shown in (A). (C) Lin– BM cells from p47^–/– mice were gammaretrovirally (miR223 p47^phox) transduced and transplanted into congenic recipient mice. Analysis of p47^phox expression and reconstitution of NADPH oxidase activity (D) was performed 8 weeks after transplantation.

FIG. 4.

miR223 promoter activity in macrophages. Lin– BM cells from p47^phox–/– mice were gammaretrovirally transduced (miR223-ΔLNGFR-2A-p47^phox) and reinfused into lethally irradiated p47^phox–/– mice. FACS analysis of (A) ΔLNGFR expression in peritoneal macrophages and (B) T, B, and dendritic cells after 8 weeks revealed preferential expression in peritoneal macrophages. (C) Frozen blood monocytes from a p47^phox CGD patient were taken into culture overnight and lentivirally (miR223-p47^phox) transduced. The cells were kept in culture for 7 days, followed by FACS analysis of p47^phox expression in macrophages (CD206+ CD14+). (D) Functional reconstitution of ROS production was analyzed in macrophages by nitroblue tetrazolium assay. ROS, reactive oxygen species.

FIG. 5.

E. coli killing activity of ex vivo differentiated human p47^phox–/– granulocytes after gene transfer. Lin– cells from a p47^–/– CGD patient were mock transduced or gammaretrovirally transduced with the self-inactivating SFFV-ΔLNGFR-2A-p47^phox or the self-inactivating miR223-ΔLNGFR-2A-p47^phox vectors and propagated to granulocytes in liquid cell culture. (A) Expression of the surrogate marker ΔLNGFR reflecting p47^phox expression in myeloid differentiated (CD11b+) and undifferentiated (CD11b–) cells. (B) Within the CD11b/ΔLNGFR double-positive populations, 0.1%, 48.7%, and 49.1% responded by ROS production upon stimulation with phorbol myristate acetate as revealed in the DHR assay (upper panels). Stimulation with opsonized E. coli (lower panels) resulted in ROS production in 1.3%, 32.1%, and 35.5% of the double-positive cells. (C) Transduced CD11b+ cells were separated by FACS as indicated by the ΔLNGFR gates shown in (A) (purity: SFFV, 87%; miR223, 93%). The isolated populations were analyzed for their E. coli killing capacity in a standard killing assay using the E. coli strain ML35. An increase in optical density at 420 nm indicates bacterial killing activity.