Gene therapy augments the efficacy of hematopoietic cell transplantation and fully corrects mucopolysaccharidosis type I phenotype in the mouse model

Abstract

Type I mucopolysaccharidosis (MPS I) is a lysosomal storage disorder caused by the deficiency of α-L-iduronidase, which results in glycosaminoglycan accumulation in tissues. Clinical manifestations include skeletal dysplasia, joint stiffness, visual and auditory defects, cardiac insufficiency, hepatosplenomegaly, and mental retardation (the last being present exclusively in the severe Hurler variant). The available treatments, enzyme-replacement therapy and hematopoietic stem cell (HSC) transplantation, can ameliorate most disease manifestations, but their outcome on skeletal and brain disease could be further improved. We demonstrate here that HSC gene therapy, based on lentiviral vectors, completely corrects disease manifestations in the mouse model. Of note, the therapeutic benefit provided by gene therapy on critical MPS I manifestations, such as neurologic and skeletal disease, greatly exceeds that exerted by HSC transplantation, the standard of care treatment for Hurler patients. Interestingly, therapeutic efficacy of HSC gene therapy is strictly dependent on the achievement of supranormal enzyme activity in the hematopoietic system of transplanted mice, which allows enzyme delivery to the brain and skeleton for disease correction. Overall, our data provide evidence of an efficacious treatment for MPS I Hurler patients, warranting future development toward clinical testing.

Figure 1

Reconstitution of IDUA activity in MPS I mice upon transplantation of wild-type or gene-corrected HSPCs. (A) Experimental scheme. Idua^−/− and Idua^+/+ cells were transduced with IDUA-LV or GFP-encoding LV (in which transgene expression was driven by the human phosphoglycerate kinase promoter) and then transplanted (1 × 10⁶ cells/mouse) into lethally irradiated mice, as indicated. GT, gene therapy–treated Idua^−/− mice; HCT, Idua^−/− mice transplanted with WT HSPC transduced with GFP-LV; MPS I, Idua^−/− mice transplanted with GFP-LV–transduced Idua^−/− HSPCs (mock-transplanted affected controls); WT, Idua^+/+ mice transplanted with GFP-LV–transduced Idua^+/+ HSPCs (mock-transplanted WT controls). (B-D) IDUA activity was measured in the PBMCs (B), in the serum (C), and in the tissues indicated below the x-axis (D) of mice transplanted with either mock-transduced or gene-corrected HSPCs at 4 weeks (B) or 6 months after transplant at sacrifice (B-D). Each dot represents one mouse, and average values are shown (black line). (E) Gene therapy–treated mice were divided into 2 groups according to the IDUA activity measured in total PBMCs (at 6 months from transplantation; left chart) and to the vector copy number per genome (VCN) measured on total BM cells (right chart). IDUA activity measured in the brain is shown for animals having IDUA activity in PBMCs below (<) or above (>) 1500 nmol/mg/h and carrying less (<) or more (>) than 5 LV copies per genome in the BM (1500 nmol/mg/h and 5 LV copies/genome are the average values measured in the entire pool of gene therapy–treated mice). Mean ± min/max are shown. **P < .01; ***P < .001 with Student t test.

Figure 2

Supranormal enzymatic activity allows storage removal in the MPS I–affected tissues. (A) Morphologic analysis of 1-μm sections, toluidine blue stained, from different organs of treated and control animals, as indicated. Representative images show the lysosomal distention in the indicated tissues in MPS I mice (no distention is present in WT animals). Kidney: arrows and arrowheads identify lysosomal storage, which is evident in the tubules (arrows), in the glomeruli (arrowheads; * marks the glomeruli), and in interstitial fibroblasts; liver: arrows highlight storage within hepatocytes, and storage is also present within Kupffer cells; spleen: pathologic storages (arrows) are predominant in the red pulp; heart: pathologic storage is abundant in the endothelium and in myocytes (arrows); frontal cortex: pathologic storage is present within neurons (arrows) and in endothelial cells (v = vessels); subjective score in MPS I: 3-4 in the different examined tissues. Magnification ×100 and ×200. Residual storage is present in all tissues of mice treated with HCTs. Two representative animals (HCT [a] and HCT [b], both having donor-cell engraftment > 70%, as assessed by quantification of donor GFP⁺ cells in peripheral blood by cytofluorimetric analysis) are shown, demonstrating a different grade of storage (subjective score HCT [a]: 1-2 and HCT [b]: 2-3 in the different examined tissues). Magnification ×100 and ×200. A strong reduction of storage is evident in all the examined tissues from a representative GT mouse having a VCN of 5 in the BM (subjective score: 0-1 in the different examined tissues). Magnification ×100. (B) Lysosomal distention/cell engulfment was scored (see “Methods” for details) in the liver, spleen, heart, kidney, and cortex of treated and control mice. Mean ± SEM are shown; n = 4 representative mice analyzed per group (≥ 3 representative images per mouse). *P < .05; **P < .01; ***P < .001 with 1-way ANOVA. (C-D) GAGs were quantified in the urine (C) and in the tissues (D, as indicated below x-axis; the liver, spleen, and kidney were chosen as representative tissues) of treated and control mice. Mean ± SEM are shown; n ≥ 4 representative mice analyzed per group. *P < .05; **P < .01; ***P < .001 with 1-way ANOVA. (E) Western blot for heparin cofactor II–thrombin (HCII-T) complex was performed on the serum of MPS I, GT, and WT mice. Glyceraldehyde 3-phosphate dehydrogenase (GAPDH; bottom image) was used as an internal control.

Figure 3

Differential neurologic outcome of gene therapy and HCT. (A) Repeated open-field test was performed on treated and control mice 6 months after the transplantation. Horizontal activity of MPS I (n = 20), HCT (n = 18), GT (n = 17), and WT (n = 14) mice is reported as a percentage (%) of change between the 1st and the 3rd days. (B) GT mice were divided into 2 groups according to the IDUA activity measured on their brain. The percentage change in horizontal activity between the first and third trials is shown for animals having brain IDUA activity lower (<, n = 8) or higher (>, n = 9) than 20 nmol/mg/h (20 nmol/mg/h is the average activity value measured on the brain of the entire population of GT mice). Mean ± SD are shown; *P < .05; **P < .01; ***P < .001 with 1-way ANOVA in panel A and Student t test in panel B. (C) Purkinje cell frequency (expressed as percentage of the cells counted in WT mice; dense area) and percentage of degenerated Purkinje cells among the total cells (scattered area) in cerebellum slices from affected, WT-, HCT-, and GT-treated mice are shown (see “Methods” for details). Mean ± SD values are shown; n ≥ 4 representative mice analyzed per group (≥ 3 representative sections per mouse). (D) Representative semithin section images from the Purkinje cell layer of treated and control mice, as indicated. The pictures show degenerating Purkinje cells, which, besides accumulating GAGs, display shrunken cell bodies and darkly stained nuclei (arrows, “gl” marks granular layer), in mock-transplanted MPS I mice; in HCT mice, residual storage and Purkinje cell degeneration were seen, whereas in the GT-treated mice we observed a complete rescue of the pathologic phenotype. Magnification ×100 in the images from the top row, ×200 in the bottom row.

Figure 4

Differential skeletal outcome of gene therapy and HCT. (A-B) Pictures (A) and 3-dimensional reconstructions of CT scans (B) from MPS I, WT, HCT, and GT mice 6 months after treatment, showing the different gross appearance of the treated and control mice (the GT mouse shown in panels A and B had a VCN of 5.4 on bone marrow; the HCT mouse had a donor-cell engraftment of 74% on PBMCs). (C-F) Measurements of skull width (C), zygomus volume (D), femur length (E), and humerus width (F) were performed on CT scan images, as shown on the right side of each chart (see “CT” for details) from MPS I (n = 19), HCT (n = 14), GT (n = 15), and WT (n = 14). For avoiding sex biases, the femur length of only the male mice was reported (MPS I n = 10, HCT n = 10, GT n = 8, and WT n = 8); similar results were obtained in females. (G) Gene therapy–treated mice were divided into 2 groups according to the IDUA activity measured on their PBMCs. The femur length is shown for animals (males and females) having PBMC IDUA activity lower (<) or higher (>) than 1500 nmol/mg/h (1500 nmol/mg/h is the average activity value measured in the entire population of gene therapy–treated mice). (H) SSI (calculated as described in the “Peripheral quantitative CT”) was evaluated by pQCT on the diaphysis of the femur (left chart) and tibia (right chart) from MPS I (n = 19), HCT (n = 14), GT (n = 15), and WT (n = 14). Mean and min/max values are shown: *P < .05; **P < .01; ***P < .001 with 1-way ANOVA (C-F,H); *P < .05 with Student t test (G).

Figure 5

Differential effect on the growth plate of gene therapy and HCT. (A) Representative pictures of the proximal epiphysis of the tibiae from mock-transplanted and treated mice (hematoxylin and eosin staining), as indicated. The growth plate is disorganized and has an irregular morphology in both mock-transplanted MPS I and HCT mice (the GT mouse shown in panels A-B had a VCN of 6 on bone marrow; the HCT mouse had a donor-cell engraftment of 80% on PBMCs). Magnifications, ×5 and ×20. (B-C) The ratio between the perimeter and the length of the growth plate (B) was calculated, and the number of chondrocytes aligned in columns perpendicular to the major axis of the growth plate (C) was counted (see supplemental Figure 2 for detailed explanation) for 5 representative MPS I, HCT, GT, and WT mice (≥ 3 representative sections per mouse). Mean and min/max values are shown: *P < .05; **P < .01; ***P < .001 with 1-way ANOVA.

Figure 6

Effect of gene therapy on auditory brainstem responses and retina integrity. (A-B) ABRs were measured on gene therapy–treated (n = 9) and control mice (n = 9 for WT and 5 for MPS I) 6 months after treatment. (A) Representative response after auditory stimulation in WT, MPS I, and GT mice was measured as the following: for each mouse, 3 series of waves (obtained by an average of 500 electrical signals each) were recorded (shown in the top panel). In the bottom panel, the resultant wave obtained by the average of the 3 traces is shown. The latency of waves I and IV was measured as shown by the pink and green lines, respectively. (B) ABR scores of wave I (left chart) and wave IV (right chart) latencies are shown. (C-D) Retinal thickness (D) of MPS1, GT, and WT mice (n = 3) measured on ×20 magnification pictures upon DAPI (4,6 diamidino-2-phenylindole) staining, as shown in the representative images in (C). n = 3 mice analyzed per group (≥ 6 representative images per mouse). PE indicates photoreceptor layer; ONL, outer nuclear layer; INL, inner nuclear layer; and GCL, ganglion cell layer. The white bar in the picture on the left shows the measured thickness. Mean and SD are shown (*P < .05; **P < .01; ***P < .001 with 1-way ANOVA).