Glucosamine amends CNS pathology in mucopolysaccharidosis IIIC mouse expressing misfolded HGSNAT

Abstract

The majority of mucopolysaccharidosis IIIC (MPS IIIC) patients have missense variants causing misfolding of heparan sulfate acetyl-CoA:α-glucosaminide N-acetyltransferase (HGSNAT), which are potentially treatable with pharmacological chaperones. To test this approach, we generated a novel HgsnatP304L mouse model expressing misfolded HGSNAT Pro304Leu variant. HgsnatP304L mice present deficits in short-term and working/spatial memory 2-4 mo earlier than previously described constitutive knockout Hgsnat-Geo mice. HgsnatP304L mice also show augmented severity of neuroimmune response, synaptic deficits, and neuronal storage of misfolded proteins and gangliosides compared with Hgsnat-Geo mice. Expression of misfolded human Pro311Leu HGSNAT protein in cultured hippocampal Hgsnat-Geo neurons further reduced levels of synaptic proteins. Memory deficits and majority of brain pathology were rescued in mice receiving HGSNAT chaperone, glucosamine. Our data for the first time demonstrate dominant-negative effects of misfolded HGSNAT Pro304Leu variant and show that they are treatable by oral administration of glucosamine. This suggests that patients affected with mutations preventing normal folding of the enzyme can benefit from chaperone therapy.

Figure S1.

Generation and skeletal phenotype of Hgsnat^P304L mice. (A) Schema showing the Cas9/sgRNA-targeting site in Hgsnat exon 9. The sgRNA-targeting sequence is underlined, and the protospacer-adjacent motif (PAM) sequence is shown in green. The c.911C>T mutation is shown in red and marked with an arrow. The C>T substitution disrupts the NcoI restriction site (shown in bold). The exon sequence is capitalized. (B) Sanger sequencing of single allele fragment obtained by PCR amplification of genomic DNA from the tail clips of the Hgsnat^P304L founder mouse showing the presence of the c.911C>T mutation. (C) Genotyping of Hgsnat^P304L mice. The DNA was extracted from clipped mouse tails and a 988-bp product amplified using a forward primer 5′-ATG​GAG​TGC​CTG​ATG​GGA​GG-3′ and a reverse primer 5′-GAT​CTA​GAA​ACG​GCC​CGA​AGA-3′. The PCR products were further digested with NcoI and analyzed on a 2% agarose gel. The 688- and 300-bp fragments are detected for the WT allele, and an undigested 988-bp fragment, for the targeted Hgsnat^P304L allele. (D) A 763-bp fragment of the Spg7 gene, containing the potential off-target sequence 5′-CTG​TGG​GAA​GAC​GCT​GTT​GGC​CA-3′, was amplified by PCR from DNA extracted from tail clips of Hgsnat^P304L founder mice (KI-1 and KI-2) and a control WT mouse. (E) Sanger sequencing of a PCR product confirms the absence of mutations in the Spg7 gene fragment adjacent to the 5′-CTG​TGG​GAA​GAC​GCT​GTT​GGC​CA-3′ fragment homologous to the PAM sequence. (F) A high-resolution in vivo micro-CT scanner (SkyScan 1176) was used to evaluate skeletal deformities in 4-mo-old Hgsnat-Geo and Hgsnat^P304L mice. The mice were anesthetized by isoflurane flow and the images were taken from the dorsal side. Both Hgsnat-Geo and Hgsnat^P304L mice do not develop abnormalities of skull bones. Panels show typical images of three mice analyzed per genotype.

Figure 1.

Hgsnat^P304L homozygous mice express mutant Hgsnat mRNA and show complete deficiency of HGSNAT activity and greater increase of lysosomal biogenesis, but similar GAG storage compared with Hgsnat-Geo mice. (A) Normal levels of Hgsnat mRNA containing c.911C>T mutation are expressed in the tissues of 4-mo-old Hgsnat^P304L mice. The values are normalized for the level of control RPL32 mRNA. Data, means, and SD of experiments performed with five mice (three male and two female) for each genotype are shown. All amplified PCR fragments were homozygous for c.911C>T mutation (inset). (B) HGSNAT activity toward 4-muf-β-D-glucosaminide in the tissues and in cultured MEF cells of 4-mo-old WT, homozygous Hgsnat^P304L, and Hgsnat-Geo mice is reduced to the background level in all studied tissues except for the brain. (C) HGSNAT activity measured using BODIPY-glucosamine is reduced to the background level in the brains of both Hgsnat^P304L and Hgsnat-Geo mice. (D) Activity of total lysosomal β-hexosaminidase shows a bigger increase in the tissues of Hgsnat^P304L compared with Hgsnat-Geo mice. (E) LAMP2 immunostaining is increased in cortical neurons of 4-mo-old Hgsnat^P304L mice compared with Hgsnat-Geo mice, suggesting higher levels of lysosomal storage. Panels show representative images of the somatosensory cortex (layers 4 and 5) and CA1 region of hippocampus of 4-mo-old Hgsnat^P304L, Hgsnat-Geo, and WT mice. Bars represent 15 µm. Graphs show quantification of LAMP2-stained area by ImageJ software. (F) Levels of disaccharides produced by enzymatic digestion of HS (ΔDiHS-0S and ΔDiHS-NS) were measured by MS/MS in blood serum, urine, and brain tissues of WT, Hgsnat^P304L, and Hgsnat-Geo mice at the age of 2, 4, and 6 mo (MO). (G) Levels of disaccharides produced by enzymatic digestion of DS (ΔDi-0S/4S), and KS and DiS-KS were measured in brain tissues of WT, Hgsnat^P304L, and Hgsnat-Geo mice at the age of 2, 4, and 6 mo. All graphs show individual data, means, and SD of experiments performed using tissues from four to seven mice per genotype per age. MEF cell data show results for three cultures, each obtained from pooled skin samples of three mice. P values were calculated by one-way ANOVA with Tukey post hoc test (A and C), nested one-way ANOVA test with Tukey post hoc test (E), or two-way ANOVA with Tukey post hoc test (B and D–F). *, P < 0.05; **, P < 0.01; ***, P < 0.001; ****, P < 0.0001.

Figure 2.

Pathological phenotypes of Hgsnat-Geo and Hgsnat^P304L mice. (A and B) Significant increase in total distance traveled in the OF (A) and the distance traveled in the central zone (B) by Hgsnat^P304L mice compared with age-matched WT controls. (C) Significant decrease and increase in the percentage of time spent in open arms and closed arms in the elevated plus maze, respectively, by Hgsnat-Geo and Hgsnat^P304L mice compared with age-matched WT controls. (D) Significant increase in the number of open arm entries in the elevated plus maze by Hgsnat-Geo and Hgsnat^P304L mice compared with age-matched WT controls. (E) Significant decrease in the discrimination index in Hgsnat^P304L mice at 4 and 6 mo (MO) in NOR test compared with age-matched WT controls. (F) Hgsnat^P304L and Hgsnat-Geo mice show onset of learning impairment in YM at 4 and 8 mo, respectively. All graphs show individual data, means, and SD of experiments performed with 6–17 mice per genotype. P values were calculated by t test for experiments involving comparison of two groups (A and B), and ANOVA with Tukey post hoc test, when comparing three groups (C–F). (G) Kaplan–Meier plot showing survival of Hgsnat^P304L (n = 43) and Hgsnat-Geo male and female mice (n = 35) and their WT counterparts (n = 28). The significance of survival rate differences between strains was determined by the Mantel–Cox test (P < 0.05). By the age of 45 wk, most Hgsnat^P304L mice had to be euthanized on veterinarian request due to urinary retention, while Hgsnat-Geo mice survived to the average age of 63 wk. (H) Wet organ weight of 8-mo-old Hgsnat^P304L, Hgsnat-Geo, and WT mice is shown as a percentage of BW. Enlargement of visceral organs, compared with age-matched WT controls, is detected in Hgsnat^P304L but not in Hgsnat-Geo mice. All graphs show individual data, means, and SD of experiments performed with five or more mice per genotype. P values were calculated using two-way ANOVA with Tukey post hoc test. *, P < 0.05; **, P < 0.01; ***, P < 0.001; ****, P < 0.0001.

Figure 3.

Synaptic defects in Hgsnat-Geo and Hgsnat^P304L mice. (A–F) Neurotransmission is impaired in Hgsnat-Geo and Hgsnat^P304L mice. Significant decrease in the amplitude (A) and frequency (C) of mEPSCs in Hgsnat-Geo and Hgsnat^P304L mice at the ages of P14–P20 and P45–P60 compared with age-matched WT controls. (B) Representative recordings of mEPSCs from WT, Hgsnat-Geo, and Hgsnat^P304L mice at P14–P20 and P45–P60, and overlay of representative individual mEPSC events from neurons of Hgsnat-Geo, Hgsnat^P304L, and WT mice. Significant decrease in the amplitude (D) and frequency (F) of mIPSCs in Hgsnat-Geo and Hgsnat^P304L mice at the ages of P14–P20 and P45–P60 compared with age-matched WT controls. (E) Representative recording of mIPSCs from neurons of WT, Hgsnat-Geo, and Hgsnat^P304L mice at the ages of P14–P20 and P45–P60, and overlay of representative individual mIPSC events from neurons of Hgsnat-Geo, Hgsnat^P304L, and WT mice. All graphs show individual data, means, and SD of experiments performed with six or more mice per genotype. P values were calculated using one-way Kruskal–Wallis test with Dunn’s multiple comparison post hoc test. (G and H) Reduction of synaptic vesicle densities, areas of PSDs, and length of PSDs in Hgsnat-Geo and Hgsnat^P304L CA1 pyramidal neurons. Density of synaptic vesicles, length (µm), and area (µm²) of PSDs were measured in asymmetrical (G) and symmetrical (H) pyramidal neurons from the CA1 region of the hippocampus. Synaptic terminals on the TEM images are marked with black arrowheads and PSDs with red asterisks. Data show values, means, and SD of the results obtained with three mice per genotype with 10–15 neurons quantified per animal. P values were calculated by nested one-way ANOVA test with Tukey post hoc test. Scale bars equal 200 nm in all panels. *, P < 0.05; **, P < 0.01; ***, P < 0.001; ****, P < 0.0001. MO, mo.

Figure 4.

Aggravated pathological changes in the brains of Hgsnat^P304L mice. (A) Astromicrogliosis in brain hippocampal and cortex regions of MPS IIIC mice is indicative of neuroimmune response. Panels show representative confocal microscopy images of brain tissues of 4-mo-old Hgsnat-Geo and Hgsnat^P304L mice and their age-matched WT controls stained with antibodies against CD68 (red) and GFAP (green) markers for activated microglia and astrocytes, respectively. DAPI (blue) was used as a nuclear counterstain. Graphs show quantification of fluorescence with ImageJ software. Individual data, means, and SD obtained for five mice per genotypes (three areas/mouse) are shown. P values were calculated using nested one-way ANOVA test with Tukey post hoc test. (B) Total brain tissues of Hgsnat^P304L mice show increased expression of inflammation markers, MIP1α, and TNFα compared with Hgsnat-Geo mice. The cytokine mRNA levels are normalized for the RLP32 mRNA content. Data show individual data, means, and SD. Five to eight mice were analyzed for each genotype. P values were calculated using one-way ANOVA with Tukey post hoc test. (C–F) Somatosensory cortices (layers 4–5) of Hgsnat^P304L mice show increased levels of markers of impaired autophagy and proteolysis compared with Hgsnat-Geo and/or age-matched WT mice: cytoplasmic LC3-positive puncta (C), granular autofluorescent ceroid materials (D), amyloid-β protein (AP; E), and misfolded SCMAS (F). Panels show representative confocal microscopy images of brain tissues of 4-mo-old (A, B, and D) or 6-mo-old (C, E, and F) Hgsnat-Geo, Hgsnat^P304L, and WT mice. Bars represent 20 µm in A and C, 100 and 25 µm in D, and 25 µm in E and F. Fluorescence was quantified with ImageJ software. Graphs show individual data, means, and SD obtained for five mice per genotype (three areas/mouse). P values were calculated using nested one-way ANOVA test with Tukey post hoc test. (G) Alteration of sphingolipid levels in the brains of Hgsnat-Geo and Hgsnat^P304L mice. Levels of glycans produced by enzymatic cleavage of total sphingolipid extracts of brain tissues from WT, Hgsnat-Geo, and Hgsnat^P304L 2-, 4-, and 6-mo-old mice were measured by normal HPLC. The values show percentage of the specific lipid. Pooled samples of three mice per age per genotype were analyzed. (H) Increased levels of G_M2 ganglioside in the brains of Hgsnat-Geo and Hgsnat^P304L mice. Confocal microscopy images of brain cortex and hippocampus tissues of individual Hgsnat-Geo, Hgsnat^P304L, and WT 4-mo-old mice stained with antibodies against G_M2 (green) and NeuN (red). DAPI (blue) was used as the nuclear counterstain. Scale bar equals 15 µm. Graphs show results of quantification performed using ImageJ software. Individual data, means, and SD obtained for five mice per genotypes (three areas/mouse) are shown. P values were calculated using nested one-way ANOVA test with Tukey post hoc test. *, P < 0.05; **, P < 0.01; ***, P < 0.001; ****, P < 0.0001. MO, mo.

Figure S2.

ER stress and UPR in the brain tissues of Hgsnat^P304L mice. (A) A higher number of hippocampal genes with altered expression levels is found in Hgsnat^P304L than in Hgsnat-Geo mice. Venn diagram showing the number of genes that were upregulated or downregulated in hippocampal tissues of 4-mo-old Hgsnat^P304L and Hgsnat-Geo mice compared with the age- and sex-matched WT mice. Three mice (two male and one female) were studied for each genotype. (B–E) The expression levels of genes involved in lysosomal biogenesis (B), inflammatory response (C), and ER stress/UPR (E) show a trend for a greater increase, while the expression of genes involved in inhibitory synaptic transmission (D) show a trend for further decrease in Hgsnat^P304L compared with Hgsnat-Geo mice. (F and G) Normal protein levels the ER stress markers, CHOP and BiP, are detected in brain cortex tissues of 6-mo-old (F) and 8-mo-old (G) WT, Hgsnat^P304L, and Hgsnat-Geo mice by immunoblot. Graphs show band intensity values measured using ImageJ software. Individual results, means, and SD of experiments with three mice per genotype, per age are shown. P values were calculated using one-way ANOVA with Tukey post hoc test. Source data are available for this figure: SourceData FS2.

Figure 5.

Aggravated pathological alterations in the gene expression and increased levels of protein markers of UPR and the ER stress in the brains of Hgsnat^P304L mice. (A–C) Hippocampal mRNA profiling in 4-mo-old MPS IIIC mice reveals increased expression of genes involved in lysosomal, lipid synthesis, and proinflammatory processes and reduced expression of genes involved in synaptic transmission, vesicular transport, and neurogenesis. Dot plots (left) show significantly enriched GO terms (biological processes, molecular functions, and cellular components) and heatmaps (right) of the genes significantly upregulated and downregulated in the hippocampi of Hgsnat^P304L mice compared with WT mice (A), Hgsnat-Geo mice compared with WT mice (B), and Hgsnat^P304L mice compared with Hgsnat-Geo mice (C). GO terms are plotted in the order of gene ratios, and each pathway is shown as a circle with the color representing the P values (−log₁₀) and the size representing the number of differentially expressed genes. The heatmap colors and their intensity show changes in gene expression levels. Data were obtained by sequencing mRNA samples extracted from three mice per genotype. (D) Brain cortex of Hgsnat^P304L 6-mo-old mice shows increased levels of O-GlcNAc–modified proteins compared with Hgsnat-Geo and WT mice. Panels show representative images of brain cortex (layers 4–5) immunostained for O-GlcNAc (green). DAPI (blue) was used as a nuclear counterstain. Scale bar equals 25 µm. Graphs show results of quantification performed using ImageJ software. Individual data, means, and SD obtained for five mice per genotype (three areas/mouse) are shown. P values were calculated using nested one-way ANOVA test with Tukey post hoc test. (E) Increased levels of ubiquitinated protein aggregates are detected in the brain homogenates of Hgsnat^P304L mice by immunoblotting. Graphs show combined intensities (individual values, means, and SD) of protein ubiquitin⁺ bands, quantified with ImageJ software and normalized by either intensity of tubulin bands or bands of ubiquitin monomers. Three mice per genotype were analyzed. P values were calculated using ANOVA with Tukey post hoc test. (F) Somatosensory cortex (layers 4–5) of Hgsnat^P304L mice shows increased levels of pyramidal neurons containing cytoplasmic ubiquitin⁺ materials. Panels show representative confocal microscopy images of brain tissues, stained for ubiquitin, of 6-mo-old Hgsnat-Geo, Hgsnat^P304L, and WT mice. Scale bar equals 25 µm. Graph shows results of quantification performed using ImageJ software. Individual data, means, and SD obtained for five mice per genotypes (three areas/mouse) are shown. P values were calculated using nested one-way ANOVA test with Tukey post hoc test. *, P < 0.05; **, P < 0.01; ***, P < 0.001; ****, P < 0.0001. Source data are available for this figure: SourceData F5.

Figure S3.

The missense variant Pro311Leu affects expression, lysosomal targeting, processing, and enzymatic activity of HGSNAT. (A) Pro311Leu HGSNAT mutant lacks enzymatic activity. The N-acetyltransferase activity was measured in homogenates of primary cultured skin fibroblasts of healthy control donor (Control) or fibroblasts transduced with LV vectors encoding for the GFP-tagged WT HGSNAT (LV-HGSNAT) or the Pro311Leu mutant (LV-P311L-HGSNAT). The graph shows individual values, means, and SD of three independent experiments. P values were calculated by one-way ANOVA followed by Tukey post hoc test; ****, P < 0.0001. (B) The 75-kD (with EGFP tag) nonglycosylated precursor is the main HGSNAT form detected in the homogenates of cells transduced with the mutant virus, while the fully glycosylated 83-kD precursor and the cleaved 29-kD α-subunit are detected in cells expressing the WT enzyme. The 50-kD band represents a nonspecific cross-reacting protein also present in nontransduced cells. The panel shows a representative blot from three independent experiments yielding similar results. (C and D) The Pro311Leu HGSNAT mutant protein is not targeted to lysosomes. Representative confocal images show fibroblast cells transduced with LV vectors encoding for the GFP-tagged WT HGSNAT (LV-HGSNAT) and or the Pro311Leu mutant (LV-P311L-HGSNAT). (E) Cells grown on glass slides were labeled with Lysotracker Red for 1 h before fixation (C) or stained for the ER (anti-Calreticulin antibodies; D) or Golgi (anti-P115 antibodies; E; red). Scale bar equals 10 μm. Panels show typical images of triplicate experiments.

Figure 6.

Expression of mouse P304L and human P311L mutant HGSNAT variants aggravates GABAergic synaptic defects in cultured primary hippocampal mouse neurons. (A) Hippocampal neurons from Hgsnat^P304L and Hgsnat-Geo mice show reduction in density of Syn1⁺ puncta in proximity to MAP2⁺ dendrites compared with WT cells. The levels of Syn1⁺ puncta are rescued by expression of WT active human HGSNAT but not of the P311L variant. (B) Hippocampal neurons from Hgsnat^P304L and Hgsnat-Geo mice show an equal reduction in density of PSD-95⁺/VGLUT1⁺ puncta in juxtaposition. The levels of PSD-95⁺/VGLUT1⁺ puncta in juxtaposition are rescued by expression of WT active human HGSNAT but not of the P311L variant. (C) Densities of Gephyrin⁺/VGAT⁺ puncta in juxtaposition are reduced in neurons from Hgsnat^P304L mice but not from Hgsnat-Geo mice compared with WT cells. Primary hippocampal neurons of Hgsnat-Geo mice transduced with LV-P311L-HGSNAT show reduction of Gephyrin⁺/VGAT⁺ puncta in juxtaposition compared with WT and nontransduced Hgsnat-Geo cells. The panels show representative confocal images of neurons and enlargements of selected axonal and dendritic fragments. Scale bars equal 40 μm. Graphs show results of puncta quantification with ImageJ software. Puncta were quantified in 20-μm-long segments of dendrite or axon, 30 μm away from the neuronal soma. Individual data, means, and SD from three independent cultures, each involving pooled embryos from at least three mice per genotype, are shown. For each culture, 5–10 neurons were analyzed. P values were calculated using nested one-way ANOVA test with Tukey post hoc test. *, P < 0.05; **, P < 0.01; ***, P < 0.001; ****, P < 0.0001.

Figure 7.

Hgsnat^P304L mice treated with glucosamine show significant increase of HGSNAT activity in brain and liver tissues, reveal delay in development of deficits in memory and learning, and partial rescue of synaptic protein markers in the CA1 area of the hippocampus. (A) HGSNAT activity, measured using the fluorogenic substrate, muf-β-D-glucosaminide, was increased in cultured MEF cells of homozygous Hgsnat^P304L mice with glucosamine (+GA) for 5 d compared with untreated cells. Graph shows individual results, means, and SD of experiments conducted with four to six different cell cultures, each established from pooled tissues of three mice. (B–D) HGSNAT activity is also increased in the brain (B and C) and liver (D) tissue homogenates of 4-mo-old Hgsnat^P304L mice treated with glucosamine (+GA) for 1 wk (B) or 13 wk (C and D) compared with untreated Hgsnat^P304L mice of the same age. (E and F) No decrease in total β-hexosaminidase activity in both organs was detected for Hgsnat^P304L mice treated with glucosamine for 13 wk. Individual results, means and SD from experiments performed with 6–10 mice per genotype, per treatment are shown. P values were calculated using an unpaired t test. (G–I) Hgsnat^P304L mice, treated with glucosamine, show rescue or a trend for improvement of deficits in spatial/short-term memory in the YM test (G) and short-term memory in the NOR test at the age of 4 mo (H and I) compared with untreated Hgsnat^P304L mice. Individual results, means and SD from experiments performed with 24 mice per genotype, per treatment are shown. P values were calculated using one-way ANOVA with Tukey post hoc test. (J) Deficient levels of protein markers of glutamatergic synaptic neurotransmission, VGUT1 and PSD-95, are rescued in the CA1 hippocampal area of Hgsnat^P304L mice, treated with glucosamine at the age of 4 mo. Panels show representative images of brain cortex (layers 4–5) and CA1 area of the hippocampus, stained for PSD-95 (red) and VGLUT1 (green), of 4-mo-old WT and Hgsnat^P304L mice treated or not with glucosamine. Scale bar equals 15 µm. The graph shows quantification of fluorescence with ImageJ software. Individual results, means, and SD from experiments performed with five mice per genotype (three areas/mouse), per treatment are shown. P values were calculated using nested one-way ANOVA test with Tukey post hoc test. (K) Deficient level of synaptic vesicular protein Syn-1 is rescued in the somatosensory cortex area of Hgsnat^P304L mice, treated with glucosamine at the age of 4 mo. Panels show representative images of brain cortex (layers 4–5) and CA1 area of the hippocampus, stained for PSD-95 (red) and Syn1 (green), of 4-mo-old WT, and Hgsnat^P304L mice treated or not with glucosamine. Scale bar equals 15 µm. The graph shows quantification of fluorescence with ImageJ software. Individual results, means, and SD from experiments performed with five mice per genotype (three areas/mouse), per treatment are shown. P values were calculated using nested one-way ANOVA test with Tukey post hoc test. *, P < 0.05; **, P < 0.01; ***, P < 0.001; ****, P < 0.0001.

Figure S4.

Mice treated daily with 2.0 g/kg BW glucosamine (GA) for 13 wk do not show alterations in growth and BW, blood glucose levels, or astro- and microgliosis in brain tissues. (A and B) BW was measured monthly, between the ages of 1 and 3 mo. Mean values and SD obtained for 12 mice per genotype, per sex, per treatment are shown. (C) The blood glucose levels were tested at the age of 4 mo. Individual data, means, and SD obtained for ≥24 mice per genotype, per treatment are shown. P values were measured using two-way ANOVA (A and B) and one-way ANOVA (C) with Tukey post hoc tests. (D) Levels of activated CD68⁺ microglia and GFAP⁺ astrocytes are not changed in the hippocampus and somatosensory cortex of 4-mo-old Hgsnat^P304L mice treated with glucosamine. Panels show representative images of somatosensory cortex (layers 4–5) and hippocampus of 4-mo-old WT, and Hgsnat^P304L mice treated or not with glucosamine and stained for GFAP (green) and CD68 (red). Scale bars equal 25 µm. Bar graph shows quantification of CD68⁺ and GFAP⁺ area with ImageJ software. Individual results, means, and SD of experiments with five mice per genotype, per treatment are shown. P values were calculated using nested one-way ANOVA test with Tukey post hoc test. **, P < 0.01; ***, P < 0.001; ****, P < 0.0001.

Figure 8.

The levels of biomarkers of CNS pathology in the somatosensory cortex are normalized and brain storage of HS is reduced in Hgsnat^P304L mice treated with glucosamine. (A) Reduction of granular autofluorescent ceroid material in cortical neurons. Panels show representative images of brain cortex (layers 4–5) of 4-mo-old WT and Hgsnat^P304L mice treated or not with glucosamine (GA) showing autofluorescent ceroid inclusions in the neurons (green). Scale bar equals 100 µm (upper panels) and 25 µm (lower panel). The graph shows quantification of autofluorescence with ImageJ software. (B) Reduction of G_M2 ganglioside in cortical neurons. Panels show representative images of somatosensory cortex (layers 4–5) of 4-mo-old WT, and Hgsnat^P304L mice treated or not with glucosamine showing immunostaining for G_M2 ganglioside and NeuN. DAPI was used as a nuclear counterstain. Scale bar equals 100 µm (upper panels) and 25 µm (lower panel). The bar graph shows quantification of G_M2 staining with ImageJ software. (C) Reduction of misfolded SCMAS aggregates in cortical neurons. Panels show representative images of somatosensory cortex (layers 4–5), stained for SCMAS, of 4-mo-old WT and Hgsnat^P304L mice treated or not with glucosamine. DAPI (blue) was used as a nuclear counterstain. Scale bar equals 100 µm (upper panels) and 25 µm (lower panel). The bar graph shows quantification of SCMAS staining with ImageJ software. All graphs show individual data, means, and SD obtained for five mice per genotype per treatment (three areas/mouse). P values were calculated using nested one-way ANOVA test with Tukey post hoc test. (D) Levels of disaccharides produced by enzymatic digestion of HS (ΔDiHS-0S and ΔDiHS-NS), measured by MS/MS, are reduced in brain tissues of 4-mo-old Hgsnat^P304L mice treated with glucosamine. All graphs show individual data, means, and SD obtained for five to six mice per genotype per treatment. P values were calculated using two-way ANOVA test with Tukey post hoc test. (E) Reduction of HS⁺ and LAMP2⁺ puncta in cortical neurons. Panels show representative images of brain cortex (layers 4–5) and CA1 region of hippocampus, stained for LAMP2 (green) and HS (red), of 4-mo-old WT and Hgsnat^P304L mice, treated or not with glucosamine. Scale bar equals 25 µm. Graphs show quantification of HS and LAMP2 staining with ImageJ software. All graphs show individual results, means, and SD from experiments conducted with five mice (three panels per mouse) per genotype per treatment. P values were calculated using nested one-way ANOVA test with Tukey post hoc test. *, P < 0.05; **, P < 0.01; ***, P < 0.001; ****, P < 0.0001.