Curative in vivo hematopoietic stem cell gene therapy of murine thalassemia using large regulatory elements

Abstract

Recently, we demonstrated that hematopoietic stem/progenitor cell (HSPC) mobilization followed by intravenous injection of integrating, helper-dependent adenovirus HDAd5/35++ vectors resulted in efficient transduction of long-term repopulating cells and disease amelioration in mouse models after in vivo selection of transduced HSPCs. Acute innate toxicity associated with HDAd5/35++ injection was controlled by appropriate prophylaxis, making this approach feasible for clinical translation. Our ultimate goal is to use this technically simple in vivo HSPC transduction approach for gene therapy of thalassemia major or sickle cell disease. A cure of these diseases requires high expression levels of the therapeutic protein (γ- or β-globin), which is difficult to achieve with lentivirus vectors because of their genome size limitation not allowing larger regulatory elements to be accommodated. Here, we capitalized on the 35 kb insert capacity of HDAd5/35++ vectors to demonstrate that transcriptional regulatory regions of the β-globin locus with a total length of 29 kb can efficiently be transferred into HSPCs. The in vivo HSPC transduction resulted in stable γ-globin levels in erythroid cells that conferred a complete cure of murine thalassemia intermedia. Notably, this was achieved with a minimal in vivo HSPC selection regimen.

Keywords: Gene therapy; Hematology; Hematopoietic stem cells; Mouse models; Therapeutics.

Figure 1. Vector structures.

HDAd-short-LCR: This vector contains a 4.3 kb mini-LCR consisting of the core regions of DNase hypersensitivity sites (HS) 1 to 4 and a 0.66 kb β-globin promoter. The length of the transposon is 11.8 kb. HDAd-long-LCR: The γ-globin gene is under the control of a 21.5 kb β-globin LCR (chr11: 5292319-5270789), a 1.6 kb β-globin promoter (chr11: 5228631-5227023), and a 3′HS1 region (chr11: 5206867-5203839) also derived from the β-globin locus. For RNA stabilization in erythroid cells, a γ-globin gene UTR was linked to the 3′ end of the γ-globin gene. The vector also contains an expression cassette for mgmt^P140K allowing for in vivo selection of transduced HSPCs and HSPC progeny. The γ-globin and mgmt expression cassettes are separated by a chicken globin HS4 insulator. The 32.4 kb LCR-γ-globin/mgmt transposon is flanked by inverted repeats (IRs) that are recognized by SB100x and by frt sites that allow for the circularization of the transposon by Flpe recombinase. HDAd-SB: The second vector required for integration contains the expression cassettes for the activity-enhanced SB100x transposase and the Flpe recombinase. ITR, inverted terminal repeat; frt, flippase recognition target; pA, polyadenylation signal; EF1α, elongation factor 1α.

Figure 2. In vivo HSPC transduction with HDAd-long-LCR containing the 32.4 kb transposon and HDAd-short-LCR containing an 11.8 kb transposon.

(A) Treatment regimen: hCD46tg mice were mobilized and IV injected with either HDAd-short-LCR + HDAd-SB or HDAd-long-LCR +HDAd-SB (2 times each 4 × 10¹⁰ viral particles (vp) of a 1:1 mixture of both viruses). Five weeks later, O⁶BG/BCNU treatment was started. With each cycle, the BCNU concentration was increased from 5 mg/kg, to 7.5 mg/kg, and to 10 mg/kg. The O⁶BG concentration was 30 mg/kg in all 4 treatments. Mice were followed until week 20, when animals were sacrificed for analysis. Bone marrow Lin^– cells were used for transplantation into secondary recipients. Secondary recipients were then followed for 16 weeks. (B) Percentage of human γ-globin–positive cells in peripheral red blood cells (RBCs) measured by flow cytometry. Each symbol is an individual animal. In mice that were mock transduced, fewer than 0.1% of cells were γ-globin–positive. The arrows indicate O⁶BG/BCNU treatment. (C) Levels of γ-globin protein chain measured by HPLC in RBCs at week 20 after in vivo HSPC transduction. Shown are the percentages of human γ-globin to mouse α-globin protein chains. (D) Levels of γ-globin mRNA measured by quantitative reverse transcription (qRT-PCR) in total blood at week 20 after in vivo HSPC transduction. Shown are the percentages of human γ-globin mRNA to mouse α-globin mRNA. (E) VCN per cell in bone marrow MNCs, harvested at week 20 after in vivo HSPC transduction. The difference between the 2 groups is not significant. Statistical analyses were performed using 2-way ANOVA.

Figure 3. Analysis of vector integration sites in HSPCs by LAM-PCR/next-generation sequencing.

Genomic DNA was isolated from bone marrow cells harvested at week 20 after in vivo transduction with HDAd-long-LCR + HDAd-SB. (A) Chromosomal distribution of integration sites. The integration sites are marked by vertical red lines. (B) Examples for junction sequences. IR/DR sequences are in red. The chromosomal sequence is in green. The TA dinucleotides used by SB100x at the junction of the IR and chromosomal DNA are highlighted. (C) Integration sites were mapped to the mouse genome, and their location with respect to genes was analyzed. Shown is the percentage of integration events that occurred 1 kb upstream transcription start sites, 3′UTR of exons, protein coding sequences, introns, 3′UTRs, 1 kb downstream from 3′UTR, and intergenic. (D) Integration pattern in mouse genomic windows. The number of integrations overlapping with continuous genomic windows and randomized mouse genomic windows and size was compared. This shows that the pattern of integration is similar in continuous and random windows. Maximum number of integrations in any given window was not more than 3, with 1 integration per window having the higher incidence.

Figure 4. Analysis of secondary recipients.

Bone marrow Lin^– cells harvested at week 20 from in vivo–transduced CD46tg mice were transplanted into lethally irradiated C57BL/6 mice. Secondary recipients were followed for 16 weeks. (A) Engraftment rates based on the percentage of CD46⁺ PBMCs at weeks 4, 8, 12, and 16 after transplantation. The differences between the 2 groups were not significant. (B) Percentage of γ-globin–expressing peripheral blood RBCs measured by flow cytometry. The differences between the 2 groups are not significant. (C) VCN per cell in bone marrow MNCs harvested at week 20 after in vivo HSPC transduction. The difference between the 2 groups is not significant. (D) Analysis of human γ-globin chains by HPLC in RBCs of secondary recipients. Shown is the percentage of human γ-globin to adult mouse α-globin. (E) Levels of γ-globin mRNA in total blood cells relative to mouse α-globin mRNA. (F) Percentage of γ-globin–expressing erythroid (Ter119⁺ cells) in all bone marrow MNCs. *P < 0.05; **P < 0.001. Statistical analyses were performed using 2-way ANOVA.

Figure 5. Human γ-globin expression after in vivo HSPC gene therapy of Hbb^th3 CD46^+/+ mice with HDAd-short-LCR and HDAd-long-LCR.

(A) Treatment regimen: In contrast to the study shown in Figure 2, this study was done with thalassemic Hbb^th3 CD46 mice. (B) Percentage of human γ-globin–positive cells in peripheral RBCs measured by flow cytometry. Each symbol is an individual animal. The arrows indicate O⁶BG/BCNU treatment. (C) γ-Globin protein chain levels measured by HPLC in RBCs at weeks 10 to 16 after in vivo HSPC transduction. Shown are the percentages of human γ-globin to mouse α-globin protein chains. (D) Representative chromatograms of an untreated Hbb^th3 CD46^+/+ mouse (left panel) and a mouse at week 16 after treatment. Mouse α- and β-chains as well the added human γ-globin are indicated. Notably, 2 independent studies were performed with Hbb^th3 CD46^+/+ mice. First study: N = 6 for HDAd-long-LCR and N = 2 for HDAd-short-LCR followed for 21 weeks. Second study: N = 4 for HDAd-long-LCR and N = 5 for HDAd-short-LCR followed for 16 weeks. Figure 5B shows the combined data until week 21. All remaining figures show the combined data until week 16. Statistical analyses were performed using 2-way ANOVA. *P < 0.05; **P < 0.0001.

Figure 6. Phenotypic correction (week 16).

Left panels: Blood smears stained with Giemsa/May-Grünwald stain (5 minutes). Middle panels: Blood smears stained with Brilliant cresyl blue for reticulocytes. Remnants of nuclei and cytoplasm in reticulocytes appear as purple staining. Right panels: Bone marrow cytospins stained with Giemsa/May-Grünwald stain (15 minutes). Upper panel: Normal bone marrow cellular distribution – erythroid lineage is represented by all stages of erythrocyte differentiation. Middle panel: Predominance of erythroid lineage over white cell lineage – erythroid lineage consists mainly of proerythroblasts and basophilic erythroblasts. Bottom panel: Normal bone marrow cellular distribution – erythroid lineage is mainly represented by maturing polychromatic and orthochromatic erythroblasts. Scale bars: 25 μm.

Figure 7. Hematological parameters before and after in vivo HSPC gene therapy of Hbb^th3 CD46^+/+ mice (week 16).

(A) Reticulocyte counts. (B) Hematological parameters. Statistical analyses were performed using 2-way ANOVA. *P < 0.05; **P < 0.0001.

Figure 8. Phenotypic correction of extramedullary hematopoiesis in spleen and liver.

(A) Spleen size at sacrifice (week 16). Left panel: representative spleen images. Right panel: summary. Each symbol represents an individual animal. Statistical analysis was performed using 1-way ANOVA. **P < 0.0001. The difference between the 2 vectors is not significant. (B) Extramedullary hemopoiesis by H&E staining in liver and spleen sections. Clusters of erythroblasts in the liver and megakaryocytes in the spleen of Hbb^th3 CD46^+/+ mice are indicated by black arrows. Scale bars: 20 μm. Representative images are shown.

Figure 9. Phenotypic correction of hemosiderosis in spleen and liver (week 16).

Iron deposition is shown by Perls’ staining as cytoplasmic blue pigments of hemosiderin in spleen and liver sections. Scale bars: 20 μm. Representative sections are shown.

Figure 10. In vitro studies with human CD34⁺ cells.

(A) Schematic of the experiment: CD34⁺ cells were transduced with HDAd-long-LCR + HD-SB or HDAd-short-LCR + HDAd-SB and subjected to erythroid differentiation (ED). In vitro selection with O⁶BG-BCNU was started at day 5 of ED. At day 18 cells were analyzed by flow cytometry (B) and HPLC (C). (D) VCN at day 18. Statistical analyses were performed using 2-way ANOVA. *P < 0.05; **P < 0.0001.

Figure 11. mgmtP140K mRNA expression levels in bone marrow MNCs at week 16 after in vivo transduction.

Human mgmt^P140K and mouse mRPL10 mRNA levels were measured by qRT-PCR in total bone marrow MNCs. (mRPL10 is a mouse housekeeping gene.) The relative levels were further divided by the VCN (see Supplemental Figure S4). Statistical analyses were performed using 2-way ANOVA.

Figure 12. Effect of SB100x-mediated integration on the transcriptome of CD34⁺ cells.

(A) Schematic of experiment. CD34⁺ cells were infected with a HDAd5/35++ vector containing a GFP/mgmt cassette under control of the EF1α promoter alone or in combination with HDAd-SB. Transduced cells were expanded in ED medium for 16 days. Two rounds of O⁶BG/BCNU selection (50 μM O⁶BG + 35 μM BCNU) were enriched for GFP-positive cells with integrated transposons. At day 16, GFP-positive cells were FACS sorted (sample 6). For comparison (sample 5), CD34⁺ cells that were transduced with the mgmt/GFP vector alone and subjected to selection were used. Because the control cells did not express SB100x, they lost the episomal mgmt/GFP vector and were therefore GFP negative. Total RNA from both samples was subjected to RNA-Seq performed by Omega Bioservices. (B) Genes with altered mRNA expression (log₂ fold change) ranked based on their P value.