AAV-Mediated Artificial miRNA Reduces Pathogenic Polyglucosan Bodies and Neuroinflammation in Adult Polyglucosan Body and Lafora Disease Mouse Models

Abstract

Adult polyglucosan body disease (APBD) and Lafora disease (LD) are autosomal recessive glycogen storage neurological disorders. APBD is caused by mutations in the glycogen branching enzyme (GBE1) gene and is characterized by progressive upper and lower motor neuron dysfunction and premature death. LD is a fatal progressive myoclonus epilepsy caused by loss of function mutations in the EPM2A or EPM2B gene. These clinically distinct neurogenetic diseases share a common pathology. This consists of time-dependent formation, precipitation, and accumulation of an abnormal form of glycogen (polyglucosan) into gradually enlarging inclusions, polyglucosan bodies (PBs) in ever-increasing numbers of neurons and astrocytes. The growth and spread of PBs are followed by astrogliosis, microgliosis, and neurodegeneration. The key defect in polyglucosans is that their glucan branches are longer than those of normal glycogen, which prevents them from remaining in solution. Since the lengths of glycogen branches are determined by the enzyme glycogen synthase, we hypothesized that downregulating this enzyme could prevent or hinder the generation of the pathogenic PBs. Here, we pursued an adeno-associated virus vector (AAV) mediated RNA-interference (RNAi) strategy. This approach resulted in approximately 15% reduction of glycogen synthase mRNA and an approximately 40% reduction of PBs across the brain in the APBD and both LD mouse models. This was accompanied by improvements in early neuroinflammatory markers of disease. This work represents proof of principle toward developing a single lifetime dose therapy for two fatal neurological diseases: APBD and LD. The approach is likely applicable to other severe and common diseases of glycogen storage.

Keywords: AAV9; APBD; EPM2A; EPM2B; GBE1; GYS1; RNAi; miRNA.

Fig. 1

AAV-amiRNA targeting murine Gys1 mRNA. Schematic representation of AAV-amiRNA and target site sequences (a). AAV-amiRNA vector components: CBh, hybrid chicken beta-actin promoter; GFP, enhanced green fluorescent protein; bGH-pA, bovine growth hormone polyadenylation signal (b). Representative images show GFP expression in sagittal brain sections of AAV-amiRNA treated (c) vs PBS-treated mice (inset of c). Scale bar, 2 mm

Fig. 2

Gys1 targeting AAV-amiRNA reduces Gys1 mRNA and protein levels, insoluble glycogen, and PB accumulation in brains of the Gbe1^Y239S APBD mouse model. Neonatal mice (P2) were injected with PBS (N = 9–13 for each experiment) or AAV-amiRNA (N = 9–13 for each experiment), and WT mice (N = 9–13 for each experiment) were used as control. Mice were sacrificed at 3 months for brain tissue analysis. Gys1 mRNA level (a) was measured by ddPCR. Representative brain GYS1 Western blots with stain-free gel as a loading control (b). Quantification of GYS1 Western blots normalized to stain-free gel shown in c. PB quantification in the hippocampus (d) and whole-brain degradation-resistant glycogen content (e). Representative micrographs of PASD-stained hippocampus of PBS (f, h) vs AAV-amiRNA (g, i) treated mice. Scale bar, 300 μm. Data are presented as mean ± SEM. Significance levels are indicated as *, p < 0.05; **, p < 0.01; ***, p < 0.001; ****, and p < 0.0001

Fig. 3

Gys1 targeting AAV-amiRNA reduces Gys1 mRNA and protein levels, insoluble glycogen, and PB accumulation in brains of the Epm2a^−/− LD mouse model. Neonatal mice (P2) were injected with PBS (N = 9–13 for each experiment) or AAV-amiRNA (N = 6 in WB quantification and N = 9–13 for other experiments), and WT mice (N = 9–13 for each experiment) were used as control. Mice were sacrificed at 3 months for brain tissue analysis. Gys1 mRNA level (a) was measured by ddPCR. Representative brain GYS1 Western blots with stain-free gel as a loading control (b). Quantification of GYS1 Western blots normalized to stain-free gel shown in c. PB quantification in the hippocampus (d) and whole-brain degradation-resistant glycogen content (e). Representative micrographs of PASD-stained hippocampus of PBS (f, h) vs AAV-amiRNA (g, i) treated mice. Scale bar, 300 μm. Data are presented as mean ± SEM. Significance levels are indicated as *, p < 0.05; **, p < 0.01; ***, p < 0.001; ****, and p < 0.0001

Fig. 4

Gys1 targeting AAV-amiRNA reduces Gys1 mRNA and protein levels, insoluble glycogen, and PB accumulation in brains of the Epm2b^−/− LD mouse model. Neonatal mice (P2) were injected with PBS (N = 10 for each experiment) or AAV-amiRNA (N = 14 for each experiment), and WT mice (N = 9–13 for each experiment) were used as control. Mice were sacrificed at 3 months for brain tissue analysis. Gys1 mRNA level (a) was measured by ddPCR. Representative brain GYS1 Western blots with stain-free gel as a loading control (b). Quantification of GYS1 Western blots normalized to stain-free gel shown in c. PB quantification in the hippocampus (d) and degradation-resistant glycogen content (e). Representative micrographs of PASD-stained hippocampus of PBS (f, h) vs AAV-amiRNA (g, i) treated mice. Scale bar, 300 μm. Data are presented as mean ± SEM. Significance levels are indicated as *, p < 0.05; **, p < 0.01; ***, p < 0.001; ****, and p < 0.0001

Fig. 5

AAV-amiRNA ameliorates PB-associated immune activation. Neonatal mice (P2) were injected with PBS or AAV-amiRNA and sacrificed at 3 months for brain tissue analysis. WT indicates the wild-type control group. For each panel, from left to right, Cxcl10, Ccl5, Lcn2, and C3 were used as neuroinflammation markers, and relative mRNA expression levels were analyzed by qRT-PCR for Gbe1^Y239S (a), Epm2a^−/− (b), and Epm2b^−/− (c) mice. In panel a: WT (N = 13), PBS (N = 12), and AAV-amiRNA (N = 11). In panel b: WT (N = 8), PBS (N = 13), and AAV-amiRNA (N = 12). In panel c: WT (N = 11), PBS (N = 10), and AAV-amiRNA (N = 13). Data are presented as mean ± SEM. Significance levels are indicated as *, p < 0.05; **, p < 0.01; ***, p < 0.001, ****, and p < 0.0001