Myeloid/Microglial driven autologous hematopoietic stem cell gene therapy corrects a neuronopathic lysosomal disease

Abstract

Mucopolysaccharidosis type IIIA (MPSIIIA) is a lysosomal storage disorder caused by mutations in N-sulfoglucosamine sulfohydrolase (SGSH), resulting in heparan sulfate (HS) accumulation and progressive neurodegeneration. There are no treatments. We previously demonstrated improved neuropathology in MPSIIIA mice using lentiviral vectors (LVs) overexpressing SGSH in wild-type (WT) hematopoietic stem cell (HSC) transplants (HSCTs), achieved via donor monocyte/microglial engraftment in the brain. However, neurological disease was not corrected using LVs in autologous MPSIIIA HSCTs. To improve brain expression via monocyte/microglial specificity, LVs expressing enhanced green fluorescent protein (eGFP) under ubiquitous phosphoglycerate kinase (PGK) or myeloid-specific promoters were compared in transplanted HSCs. LV-CD11b-GFP gave significantly higher monocyte/B-cell eGFP expression than LV-PGK-GFP or LV-CD18-GFP after 6 months. Subsequently, autologous MPSIIIA HSCs were transduced with either LV-PGK-coSGSH or LV-CD11b-coSGSH vectors expressing codon-optimized SGSH and transplanted into MPSIIIA mice. Eight months after HSCT, LV-PGK-coSGSH vectors produced bone marrow SGSH (576% normal activity) similar to LV-CD11b-coSGSH (473%), but LV-CD11b-coSGSH had significantly higher brain expression (11 versus 7%), demonstrating improved brain specificity. LV-CD11b-coSGSH normalized MPSIIIA behavior, brain HS, GM2 ganglioside, and neuroinflammation to WT levels, whereas LV-PGK-coSGSH partly corrected neuropathology but not behavior. We demonstrate compelling evidence of neurological disease correction using autologous myeloid driven lentiviral-HSC gene therapy in MPSIIIA mice.

Figures

Figure 1

Comparison of eGFP expression from myeloid promoters. (a) pCCL LVs containing human CD18, CD11b, and PGK promoters driving eGFP expression. (b) WT bone marrow was lineage depleted and the enriched stem cell population transduced with LV-hCD18-eGFP, LV-hCD11b-eGFP, or LV-hPGK-eGFP at an MOI of 40–70 and transplanted into myeloablated WT mice. n = 8–10 (c) Donor and eGFP+ chimerism was determined 24 weeks post-transplantation. The vector copy number (VCN) was determined in WBCs. (d–e) Mean fluorescent intensity of eGFP expression was determined for all leukocytes and for CD11b+, CD19+, and CD3+ cells 24 weeks post-transplantation. (f–h) Mice with most equivalent copy numbers (VCN of 2.3–4.1) in WBCs were further compared. Error bars represent the SEM. Significant differences between the two groups marked with a line are demonstrated with *P < 0.05; **P < 0.01; and ***P < 0.001. CMV, cytomegalovirus; eGFP, enhanced green fluorescent protein; hPGK, human phosphoglycerate kinase; LV, lentiviral vector; MOI, multiplicity of infection; ns, nonsignificant; WBC, white blood cell; WT, wild-type.

Figure 2

LV-CD11b–modified autologous hematopoietic stem cells correct MPSIIIA mouse behavior. (a) pCCL LVs were made containing either human CD11b or hPGK promoters driving normal or coSGSH. (b) Expression of hSGSH and coSGSH was compared by transient transfection of the human CHME3 microglial cell line. (c) MPSIIIA bone marrow was lineage depleted and transduced with LV-hCD11b-coSGSH or LV-hPGK-coSGSH at an MOI of 80, then transplanted into busulfan myeloablated MPSIIIA mice. (d,e) VCN was determined in WBCs (behavior n = 8–12; biochemistry n = 5). (f) 4 months post-transplantation, the donor chimerism in WBCs was determined by flow cytometry (n = 3–7). (g–j) Open-field behavior was evaluated at 6 months (24 weeks) of age (n = 8–15 female mice per group). The measures of hyperactivity were: (g) path length, (i) frequency, and (j) duration spent moving over 100 mm/second, while (h) frequency of centre entries measures thigmotaxis and may be a measure of sense of danger. Error bars represent the SEM. Significant differences to MPSIIIA, between the two groups marked with a line, are demonstrated with *P < 0.05; **P < 0.01; and ***P < 0.001. CMV, cytomegalovirus; coSGSH, codon-optimized human N-sulfoglucosamine sulfohydrolase; hPGK, human phosphoglycerate kinase; hSGSH, human N-sulfoglucosamine sulfohydrolase; LV, lentiviral vector; MOI, multiplicity of infection; MPSIIIA, mucopolysaccharidosis type IIIA; VCN, vector copy number; WBC, white blood cell; WT, wild-type.

Figure 3

LV-CD11b improves brain-specific SGSH activity over LV-PGK and corrects primary HS storage of MPSIIIA mice. At 8 months of age, SGSH enzyme activity was measured in (a) bone marrow, (b) spleen, (c) liver, and (d) brain of perfused mice (n = 5). Secondary upregulation of β-Hex was measured in (e) liver and (f) brain. (g) Total sulfated GAG in the brain was measured with the dye binding Blyscan assay. (h) Total HS storage was measured by 2-aminoacridone (AMAC)–labeled disaccharide analysis. The different HS disaccharides were quantified by (i) AMAC, and (j) the relative proportion of HS that was NAc, NS, 6S, and 2S was also determined. Error bars represent the SEM. Significant differences to MPSIIIA are marked with *P < 0.05 and ***P < 0.001. Treatments that are not significantly different to WT are marked with a line and ns (nonsignificant) (P > 0.05). 6S, 6-O-sulfated; 2S, 2-O-sulfated; β-Hex, β-N-acetyl hexosaminidase; GAG, glycosaminoglycan; HS, heparan sulfate; LV, lentiviral vector; MPSIIIA, mucopolysaccharidosis type IIIA; NAc, N-acetylated; NS, N-sulfated; PGK, phosphoglycerate kinase; SGSH, N-sulfoglucosamine sulfohydrolase; WT, wild-type.

Figure 4

LV-CD11b corrects lysosomal compartment size in MPSIIIA brains. Representative confocal images of (a) brain cortex (layer IV/V) and (b) amygdala (medial nucleus) from approximately −1.18 mm relative to bregma were stained with LAMP2 for lysosomal compartment (green) and NeuN for neuronal nuclei (red), and representative images are shown. Bar = 20 µm. LV, lentiviral vector; MPSIIIA, mucopolysaccharidosis type IIIA; PGK, phosphoglycerate kinase; WT, wild-type.

Figure 5

LV-CD11b normalizes neuroinflammation in the cerebral cortex and amygdala of MPSIIIA mice, while LV-PGK mediates improvements in neuroinflammation. Four brain sections per mouse (n = 5 per group), taken from approximately bregma 0.26, −0.46, −1.18, and −1.94 mm, were stained and quantified with isolectin B4 for microglia number in the (a) cerebral cortex and (b) amygdala or with GFAP for astrocytes in the (c) cerebral cortex and (d) amygdala. Representative images from the cerebral cortex (layer IV/V) and the amygdala, both from approximately −1.18 mm relative to bregma, are shown. Sections were counterstained with Mayer's hematoxylin to highlight the nuclei (blue). Error bars represent the SEM. Significant differences to MPSIIIA, between the two groups marked with a line, are demonstrated with *P < 0.05, **P < 0.01; and ***P < 0.001. Bar = 50 µm. LV, lentiviral vector; MPSIIIA, mucopolysaccharidosis type IIIA; PGK, phosphoglycerate kinase; WT, wild-type.

Figure 6

LV-CD11b and LV-PGK both correct secondary storage and presynaptic vesicles in the cerebral cortex, while LV-PGK is less effective in the amygdala of MPSIIIA mice. Four brain sections per mouse (n = 5 per group), taken from approximately bregma 0.14, −0.58, −1.3, and −2.06 mm, were stained and quantified for GM2 ganglioside in the (a) cerebral cortex and (b) amygdala, and VAMP2 for presynaptic vesicles in the (c) cerebral cortex and (d) amygdala. Representative images from the cerebral cortex (layer IV/V) and the amygdala, both from approximately −1.3 mm relative to bregma, are shown. Error bars represent the SEM. Significant differences to MPSIIIA, between the two groups marked with a line, are demonstrated with ***P < 0.001. Bar = 50 µm. LV, lentiviral vector; MPSIIIA, mucopolysaccharidosis type IIIA; PGK, phosphoglycerate kinase; WT, wild-type.