Intrathecal delivery of AAVrh10-mHexa combined with anti-inflammatory treatment reduces neuropathological markers and extends the lifespan of mice with early-onset Tay-Sachs disease

Abstract

Tay-Sachs disease is a lysosomal storage disorder caused by mutations in the HEXA gene, which encodes the α-subunit of β-hexosaminidase A—an enzyme that breaks down GM2 ganglioside. Recently, a mouse model of Tay-Sachs, the DKO, with deficiencies in both Hexa and Neu3 genes, showed severe neurological symptoms and neuroinflammation, surviving up to 20 weeks. In this study, we evaluated the therapeutic potential of intrathecal AAVrh10-mediated delivery of mouse Hexa, in combination with istradefylline treatment, in DKO mice. Using molecular, immunohistochemical, and behavioral methods, we found that the mice’s lifespan increased to 30 weeks after receiving AAV alone or with istradefylline. Molecular analyses revealed increased Hexa activity, accompanied by reduced levels of the lysosomal marker Lamp-1 and pro-inflammatory cytokines, such as CCL2 and CCL3, in the cortex, cerebellum, and various organs, including the kidney, liver, and spleen. Immunohistochemistry revealed clearance of GM2 accumulation, fewer lysosomes, decreased active astrocytes, and improvements in neurons and oligodendrocytes in the brains of DKO mice. Correspondingly, their motor activity also improved. These results suggest that AAVrh10-based intrathecal delivery combined with istradefylline provides a promising therapeutic strategy for treating Tay-Sachs disease.

Supplementary information

The online version contains supplementary material available at 10.1007/s11011-026-01802-2.

Introduction

Tay-Sachs disease, a subtype of GM2 gangliosidosis, is caused by mutations in the HEXA gene, which encodes the α-subunit of the enzyme β-hexosaminidase A (Yamanaka et al. 1994). This lysosomal enzyme breaks down GM2 ganglioside into GM3 ganglioside. A deficiency in HEXA leads to the accumulation of GM2 in lysosomes, resulting in intellectual disability, severe neurological problems, motor impairments, muscle weakness, and often death between ages 2 and 4 (Bley et al. 2011). To investigate Tay-Sachs, Hexa-/- mice have been generated; however, they exhibit only mild symptoms and minimal GM2 accumulation in the central nervous system (Yamanaka et al. 1994). A new model, Hexa-/-Neu3-/- (DKO), which lacks both Hexa and Neu3, exhibits severe GM2 accumulation, increased lysosomal activity, and neuroinflammation, ultimately leading to death at 20 weeks (Seyrantepe et al. 2018). Our previous research also identified alterations in secondary lipid metabolism, including phosphatidylcholine (PC), phosphatidylinositol (PI), and sphingomyelin (SM), as well as impaired autophagy in the brains of DKO mice (Can et al. 2022; Sengul et al. 2023). Recently, Basirli et al. (2024) demonstrated that lithium treatment can regulate and restore autophagic flux in cell models of Tay-Sachs disease.

Adeno-associated virus (AAV)-based gene therapy is an innovative approach for treating lysosomal storage disorders (LSDs), offering advantages such as low immunogenicity, non-pathogenicity, broad tissue targeting, and long-lasting gene expression (Park et al. 2017). Unlike lentiviral vectors, AAV vectors remain episomal rather than integrating into the host genome, which reduces the risk of insertional mutagenesis while allowing target gene expression (Li and Samulski 2020). Different AAV serotypes enable transduction of various tissue types. The recently identified AAVrh10 serotype, isolated from Rhesus Macaques, is neurotropic and selectively transduces neurons (Park et al. 2017). It also efficiently targets sensory neurons in the bone marrow, sensory nerve fibers, and myelinated sensory neurons (Park et al. 2017). These features make AAVrh10 an excellent vector for gene delivery in diseases affecting the central nervous system. Previous studies have shown the therapeutic potential of AAVrh10 in models of LSD, such as Krabbe disease (Piguet et al. 2012), neuronal ceroid lipofuscinosis (Sondhi et al. 2007), metachromatic leukodystrophy (Piguet et al. 2012), diabetic neuropathy (Homs et al. 2014), and amyotrophic lateral sclerosis (Wang et al. 2014). These studies employed intracranial and intrathecal administration of AAVrh10, demonstrating its effectiveness in targeting neurons and oligodendrocytes and significantly reducing symptoms of CNS disease.

Istradefylline’s anti-inflammatory effects were observed in a mouse model of Sandhoff disease (Hexb-/-). In these mice, A2A receptor levels in astrocytes correlated with cortical inflammation from 4 to 10 weeks. When 2.5-month-old Hexb-/- mice received intraperitoneal injections of istradefylline, they showed improved motor function, reduced Iba1 + microglia, and decreased levels of inflammation-related chemokines (Ccl2, Cxcl10) and cytokines (Il1a and Il1b) (Ogawa et al. 2018).

This study explored the potential of intrathecal AAVrh10-mHexa delivery, which includes the mouse HEXA gene encoding the α-subunit of β-hexosaminidase A, in an early-onset Tay-Sachs mouse model. We also assessed whether combining istradefylline with AAVrh10-mHexa could help reduce neuroinflammation in these mice.

Materials and methods

Vectors

The mammalian expression vector encoding mouse Hexa (pAAV-CMV-mHexa-3xGGGGs-mCherry-oPRE) was designed and obtained from Vector Builder. To produce infectious virus particles, we purchased pAAV2/rh10, a packaging vector (Rep/Cap) with CNS tropism, and pAdDeltaF6, a helper vector expressing adenovirus genes, from Addgene. Vector verification was performed through restriction enzyme mapping. EcoRI (Biolab NEB Cat# 0101 S), KpnI (Biolab NEB Cat# 3142 S), and PstI (Biolab NEB Cat# 3140 S) enzymes were used for the mammalian expression vector. In contrast, BamHI (Biolab NEB Cat# 0136 S) was used to confirm the packaging vector pAAV2/rh10 and the helper vector pAdDeltaF6. The verified vectors were then amplified using the MaxiPrep kit following the manufacturer’s protocol (Thermo Scientific Cat# K0491).

Production of AAVrh10-mHexa virions

AAVrh10-mHexa virions were produced and purified according to the protocol by Fripont et al. (2019). HEK293T cells were triple-transfected with 10,000 ng of the pAAV2/rh10 expression vector and 20,000 ng of pAdDeltaF6, using a 1:5 DNA to PEI (Merck Cat# 408727) ratio across 50 and 150-mm culture plates (Greiner Bio-One Cat# 639160). After 72 h, mCherry expression was confirmed with fluorescence microscopy. The cells were then harvested, centrifuged, and stored at − 80 °C for future freeze–thaw processing. The cell pellets were resuspended and lysed in a lysis buffer, then centrifuged at 1,167 × g for 15 min at 4 °C. DNase I (Merck Cat# 10104159001) was added, followed by filtration through a 0.45-µm syringe filter. The samples were prepared for purification using an Optiprep density gradient, creating layers of 15%, 25%, 40%, and 60% iodixanol solutions (Merck Cat# D1556) in QuickSeal centrifuge tubes (Beckman Cat# 10736641). Viral lysate was added, and ultracentrifugation was performed at 200,000 × g for 2 h at 12 °C. Viral particles were collected from the 40% iodixanol layer using a syringe. After purification, the virus was washed and concentrated with a 100-kDa Millipore centrifugal filter (Merck Cat# UFC810008), then stored at −20 °C for quantification.

Quantification of AAVrh10-mHexa virions

To determine the virus titer, a mammalian expression vector driven by the CMV promoter was prepared as a standard at various concentrations: 10¹⁰ vg/µl, 10⁹ vg/µl, 10⁸ vg/µl, 10⁷ vg/µl, and 10⁶ vg/µl (viral genomes per µl), following the protocol from Fripont et al. 2019. The purified and concentrated virus particles, along with these standards, were used for RT-PCR. The reaction mixture for CMV included 1X Roche LightCycler 480 SYBR Green I Master Mix (Roche Cat# 04707516001) and 0.4 µM of CMV forward primer (5‘-TCATATGCCAAGTACGCCCC-3’) and reverse primer (5’-CCCGTGAGTCAAACCGCTAT-3’). This mixture was prepared for each standard and purified virus sample. RT-PCR was performed using the Roche LightCycler^® 96 System. The virus titer (vg/µL) was calculated from CMV expression, resulting in a concentration of 2 × 10^9 vg/µL.

Animals

A DKO mouse model with a combined deficiency of the Hexa and Neu3 genes was created as described previously (Seyrantepe et al. 2018). All mice were housed at the Izmir Institute of Technology Animal Research Center in cages that hold up to five animals, maintained at a constant temperature (21 ± 1 °C) and humidity, under a 12-hour light/dark cycle. Food and water were available freely. The Institutional Animal Care and Use Committee of Izmir Institute of Technology approved all animal procedures. The study followed the guidelines established by the Turkish Institute of Animal Health for the use of laboratory animals. Mice were weaned at 4 weeks old for genotyping, which was performed using Polymerase Chain Reaction (PCR) with specific primers for the Hexa and Neu3 genes, as previously described (Seyrantepe et al. 2018).

Intrathecal injection of AAVrh10-mHexa virions

Intrathecal injections were performed on 8-week-old DKO mice to deliver AAVrh10 virions encoding mouse Hexa directly to the brain, bypassing the blood-brain barrier. As a control, 50 µl of PBS was injected into age-matched DKO mice to assess any effects from the solvent. The mice were anesthetized through intraperitoneal injection of xylazine and ketamine. After shaving and sterilizing the spine with Betadine and 70% ethanol, a small incision was made to expose the L5-6 region. An insulin syringe was used to inject 50 µl of AAV (containing 1 × 10¹¹ viral genomes), and the incision was closed with URGO Pansement spray. Each mouse was then housed individually.

Istradefylline injection

Istradefylline, as an anti-inflammatory agent, was administered to 10-week-old DKO mice via intraperitoneal injection (i.p.). Istradefylline (5 mg/mL, Sigma-Aldrich Cat# SML0422) was dissolved in dimethyl sulfoxide (DMSO) and administered intraperitoneally for 21 days.

Thin-Layer chromatography (TLC)

Samples from 50 mg of mouse brain tissue from 20- and 30- week- old WT, 20- week- old DKO, and DKO mice treated with AAV and istradefylline (n = 3) were used for ganglioside extraction, following Sandhoff et al. (2002) and Holm et al. (1972). Cortex sections were weighed, then immersed in 2 mL of distilled water and placed in borosilicate tubes. Homogenization was performed using an Ultra-Turrax at 6000 rpm for 45 s, followed by three 1.5-minute sonication cycles with a Bandelin Sonopuls. After sonication, the samples were dried with N₂ in a Reacti-Thermo Heating Module. Dried samples were then processed for ganglioside extraction. Three milliliters of acetone were added, and each sample was centrifuged at 2000 rpm for 5 min, with the supernatant discarded after each of three repetitions. Pellets were resuspended in 1.5 mL of a methanol: chloroform: water mixture (10:10:1). The samples were centrifuged at 2000 rpm for 5 min, and the supernatants were transferred to new tubes. This process was repeated twice. Next, 2 mL of a methanol: chloroform: water mixture (60:30:8) was added, the mixture was centrifuged, and the supernatants were pooled with the previous ones. The final supernatant, containing both neutral and acidic gangliosides, was processed for separation using freshly prepared DEAE- Sephadex A-25 resins (Sigma Cat# A 25120- 10 G). Gangliosides were loaded with 4 mL of methanol; neutral gangliosides were discarded through the initial flow, and acidic gangliosides were eluted with 5 mL of potassium acetate in methanol. Desalting was performed using Superclean LC-18 columns (Merck Cat# 57012) and a Chromabond Vacuum System, with sequential washes in methanol, potassium acetate, and water, followed by elution with a methanol: chloroform mixture. The eluates were evaporated under nitrogen, stored at 4 °C, and then dissolved in a methanol: chloroform: water (10:10:1) mixture. Gangliosides were analyzed on silica plates using the Linomat 5 applicator, developed with a mobile phase of chloroform, methanol, and CaCl₂, and visualized with orcinol staining. After heating at 120 ° C, images of the stained plates were scanned and quantified using ImageJ (NIH).

Enzyme activity assay

Hexa enzyme activity was measured using a substrate specific to Hexa in various tissues, including the cortex, cerebellum, kidney, liver, spleen, and serum. Samples were collected from 20- and 30-week-old WT mice, 20-week-old DKO mice, and DKO mice treated with AAV or istradefylline at 20 and 30 weeks (n = 3). For each sample, except serum, 50 mg was weighed, mixed with 150 µl of distilled water, and homogenized using an Ultra-Turrax homogenizer (IKAT10, Merck). Homogenates were sonicated twice for 5 s with a Bandelin Sonopuls sonicator. Serum samples (20 µl) were used directly without homogenization or sonication. To each homogenized sample, 50 µl of Hexa substrate (4-Methylumbelliferyl 6-Sulfo-2-acetamido-2-deoxy-β-D-glucopyranoside potassium salt, Santa-Cruz Cat# 223640) and 40 µl of 0.1 M sodium acetate buffer (pH 4.5) were added. The mixtures were incubated in a water bath at 37 °C for 30 min. To stop enzyme activity, 3.9 ml of 0.4 M glycine buffer (pH 10.4) was added. Absorbance was measured with a spectrophotometer (RF-5301). Enzyme activity was determined by measuring absorbance and protein content, expressed as µmol/mg of protein.

RT-PCR analyses

RNA was also extracted from the brain and visceral organs of 20- and 30-week-old WT, 20-week-old DKO mice, and DKO mice treated with AAV, with or without istradefylline (n = 3). Fifty milligrams of each tissue were weighed, then 1 mL of Genezol reagent (Geneaid Cat# GZR050) was added. Samples were homogenized using a Retsch homogenizer (MM100) and incubated at room temperature for 5 min. The homogenates were centrifuged for 5 min at 12,000 × g at 4 °C, and the supernatants were transferred to Eppendorf tubes. To achieve phase separation, 200 µL of chloroform was added, and the mixture was briefly mixed for 10 s. After 2 min at room temperature, the samples were centrifuged at 12,000 × g for 15 min at 4 °C. The aqueous phase was transferred, and 500 µL of isopropanol was added to precipitate the RNA. After incubating for 10 min at room temperature, samples were centrifuged at 12,000 × g for 10 min at 4 °C. The supernatant was discarded, and 1 mL of 75% ethanol was added to the RNA pellet. The tubes were centrifuged at 7500 × g for 5 min at 4 °C. The supernatant was removed, and the RNA pellet was dried at 55 °C. The dried RNA was dissolved in RNase-free water and incubated at 55 °C for 10 min. RNA concentration was measured using a NanoDrop ND-1000 spectrophotometer. The RNA was reverse-transcribed into cDNA with the iScript cDNA Synthesis Kit (Bio-Rad Cat# 1708890). Expression levels of mouse Hexa, inflammation- and lysosome-related genes (Table 1) were quantified with the Roche LightCycler 96 using LightCycler 480 SYBR Green I Master Mix (Roche Cat# 04707516001). GAPDH was used as an endogenous control.

Table 1.

Forward and reverse primer sequences in RT-PCR

Gene	Primer Sequences
mHexa	F: 5’- CACCAGGGCTGGCTTCC − 3’ R: 5’- CATGAAACGCCAGGGGCT − 3’
CCL2	F: 5’-ATGCAGTTAATGCCCCACTC-3’ R: 5’-TTCCTTATTGGGGTCAGCAC-3’
CCL3	F: 5’-TCTGTACCATGACACTCTGC-3’ R: 5’-AATTGGCGTGGAATCTTCCG-3’
CCL5	F: 5’-AGTGCTCCAATCTTGCAGTC-3’ R:5’-AGCTCATCTCCAAATAGTTG-3’
Cxcl10	F: 5’-CACCATGAACCCAAGTGCTGCCGT-3’ R: 5’-AGGAGCCCTTTTAGACCTTTTTTG-3’
IL-1β	F: 5’-TGAGTCACAGAGGATGGGCTC-3’ R: 5’-CCTTCCAGGATGAGGACATGA-3’
Iba1	F: 5’-TCTGCCGTCCAAACTTGAAGCC-3’ R: 5’-CTCTTCAGCTCTAGGTGGGTCT-3’
GFAP	F: 5’-AGTAACATGCAAGAGACAGAG-3’ R: 5’-TAGTCGTTAGCTTCGTGCTTG-3’
Lamp1	F: 5’-CCAGGCTTTCAAGGTGGACAGT-3’ R: 5’-GGTAGGCAATGAGGACGATGAG-3’
GAPDH	F: 5’-CCCCTTCATTGACCTCAACTAC-3’ R: 5’-ATGCATTGCTGACAATCTTGAG-3’

Western blot analyses

Total proteins were extracted from the brain (cortex and cerebellum) and visceral organs (kidney, liver, spleen) of 20- and 30-week-old WT, 20-week-old DKO mice, as well as 20- and 30-week-old DKO mice treated with AAV or AAV combined with istradefylline (n = 3). Proteins were isolated using a lysis buffer (50 mM Tris-HCl, 150 mM NaCl, 1% Triton X-100, 50 mM HEPES, 10% glycerol, protease inhibitors) and measured with the Bradford Protein Assay (Merk Cat# B6916). Equal amounts of protein were separated by SDS-PAGE and transferred onto nitrocellulose membranes (Bio-Rad Cat# 1620115). The membranes were incubated overnight at + 4 °C with primary antibodies: anti-IL-1β (Cell Signaling Technology Cat# 31202, RRID: AB_2799001, 1:1000) and anti-β-Actin (Cell Signaling Technology Cat# 13E5, RRID: AB_2223172, 1:1000). A HRP-conjugated secondary antibody (Santa Cruz Cat# sc-2370, RRID: AB_634837, 1:10000) was applied for one hour at room temperature. Luminata™ Forte Western HRP Substrate (Abcam, Cat# ab5801) was added, and bands were visualized using a digital imaging system (Fusion SL, Vilber). Band intensities were quantified with NIH ImageJ (version 1.48v) and normalized to β-actin.

Immunohistochemistry analyses

Immunohistochemistry was performed on 20- and 30-week-old WT mice, 20-week-old DKO mice, and 20- and 30-week-old DKO mice treated with AAV or AAV plus istradefylline (n = 3). Mice were anesthetized via intraperitoneal injection of Xylazine and Ketamine. Transcardial perfusion started with 10 ml of 0.9% NaCl, followed by 4% paraformaldehyde in 1X PBS. The brains were fixed overnight at 4 °C in 4% paraformaldehyde, then placed in a sucrose gradient (10%, 20%, 30%) in PBS. Each brain was embedded in OCT (Sakura), sectioned coronally at 10 μm using a Leica Cryostat at −20 °C, mounted on adhesive HistoBond slides, and stored at −80 °C. For staining, sections were incubated with primary antibodies: anti-GM2 (KM966, 1:500), anti-Lamp1 (Cell Signaling Technology Cat# 46843, RRID: AB_2134478, 1:500), anti-NeuN (Cell Signaling Technology Cat# 24307, RRID: AB_2651140, 1:500), anti-CNPase (Abcam Cat# ab6319, RRID: AB_2082593, 1:500), and anti-GFAP (Abcam Cat# ab68428, RRID: AB_1209224, 1:500). They were then incubated with secondary anti-rabbit Alexa Fluor 488 (Abcam Cat# ab150077, RRID: AB_2630356, 1:500). Nuclei were stained with DAPI in Fluoroshield mounting medium (Abcam Cat# 104139), and coverslips were applied. Images were captured using an Olympus BX53 fluorescence microscope.

Rotarod and footprint test

Motor coordination and balance were assessed using the rotarod test (Seyrantepe et al. 2018). Each mouse was placed on a rotating rod and trained to walk on it. After training, each mouse was placed on the accelerating rotating rod (Pan-Lab Harvard Apparatus), and the time to fall was recorded with Sedacom version 2.0 (Harvard Apparatus). The rotation speed started at 4 rpm and increased to 40 rpm over 3 min. The mice’s gait was analyzed using the Footprint test. Non-toxic ink was applied to both their forepaws and hindpaws. The mice then walked along a white paper-lined runway inside a closed box. Their footprints were scanned with an HP scanner, and stride, sway, and stance length were measured. Six independent mice were used per group for both the rotarod and footprint tests.

Statistical analysis

All statistical analyses were performed using GraphPad Prism 7 (version 7.0a, GraphPad Software, Inc.). P-values were calculated with a one-way ANOVA followed by Tukey’s post-hoc test for multiple comparisons. Data are presented as mean ± SEM.

Results

Extension of the lifespan and improvements in locomotor activity of DKO mice following intrathecal administration of AAV and istradefylline treatment

Intrathecal AAV injections were administered to DKO mice at 8 weeks of age. The in vivo effects of the AAV-based gene therapy were then evaluated (Fig. 1A). Body weight data showed that DKO mice experienced significant and rapid weight loss after 16 weeks, compared to WT mice (Fig. 1B). Treatment with AAV gene therapy alone or combined with istradefylline effectively restored body weight in DKO mice (Fig. 1B, C). Although untreated DKO mice and DKO mice treated with istradefylline survived until 20 weeks, those treated with AAV, either alone or with istradefylline, had extended survival until 30 weeks (Fig. 1D). DKO mice administered with AAV, and AAV with istradefylline, were monitored daily after 20 weeks of age. During these observations, we assessed phenotypic features related to disease progression, such as slow movement, ataxia, and tremor, as well as daily body weight loss. Based on these observations, mice were sacrificed at 30 weeks of age as a humane endpoint. Before delivering the vector intrathecally, successful entry into HEK293T cells was confirmed by detecting mCherry fluorescence in both single and co-transfected cells (Supplementary Fig. 1). The presence of mCherry also indicates that mHexa is expressed, as both genes are linked via a linker within the vector. The impact of AAV-mediated mHexa expression on GM2 ganglioside levels was evaluated using anti-GM2 immunostaining in vitro. Neuroglial cells from 20-week-old DKO mice, lacking Hexa and Neu3 genes, showed increased GM2 accumulation (Supplementary Fig. 2). Conversely, cells transduced with AAV-mHexa demonstrated a significant decrease in GM2 levels, as shown by anti-GM2 staining, suggesting that mHexa effectively degrades GM2 ganglioside in vitro (Supplementary Fig. 2). In the rotarod test, 20-week-old DKO mice showed a shorter latency to fall, indicating deficits in motor function and balance compared to age-matched WT mice (Fig. 1E). Both AAV gene therapy alone and AAV combined with istradefylline improved locomotor and motor activity in DKO mice at both 20 and 30 weeks of age, relative to WT mice (Fig. 1E). The Footprint test assessed mice walking patterns by measuring stride, sway, and stance lengths. In 20-week-old DKO mice, these measurements were notably reduced compared to age-matched WT mice (Supplementary Fig. 3). Furthermore, 30-week-old DKO mice treated with AAV and those receiving both AAV and istradefylline also exhibited shorter stride lengths than their WT counterparts (Supplementary Fig. 3 A, B).

Fig. 1.

Improvements in body weight, lifespan, and motor activities in DKO mice after intrathecal AAV treatment. A Intrathecal injection of virions carrying mHexa into DKO mice. B Body weights of (n = 10; 5 females, 5 males), DKO mice (n = 8; 4 females, 4 males), DKO mice injected with AAV (n = 8; 4 females, 4 males), and DKO mice treated with AAV and Istradefylline (n = 8; 4 females, 4 males). Each point shows the average for representative mice of each genotype. C Body weights of 20-week-old WT (n = 10; 5 females, 5 males), DKO mice (n = 8; 4 females, 4 males), DKO mice treated with AAV (n = 8; 4 females, 4 males), and DKO mice treated with AAV and Istradefylline (n = 8; 4 females, 4 males). D Kaplan-Meier survival curves for WT, DKO mice, DKO mice with AAV, and DKO mice treated with AAV and Istradefylline. E Running time on the rotarod (seconds) for 20- and 30-week-old WT mice, 20-week-old DKO mice, and 20- and 30-week-old DKO mice treated with AAV, and with or without Istradefylline (n = 6 mice; *p < 0.05, **p < 0.01, ****p < 0.001)

Degradation of accumulating GM2 ganglioside in the brain of AAV-treated DKO mice

TLC analyses showed increased GM2 ganglioside accumulation in the cortex of 20-week-old DKO mice. In contrast, WT mice did not exhibit such accumulation at either 20 or 30 weeks (Fig. 1). Treatment with AAV alone or combined with istradefylline significantly reduced GM2 ganglioside levels in the cortex of DKO mice at both 20 (Fig. 2A, B) and 30 weeks (Fig. 2C, D), compared to untreated DKO mice at 20 weeks. This indicates that AAV-based gene therapy effectively enhances GM2 ganglioside degradation by expressing mHexa.

Fig. 2.

Degradation of accumulated GM2 ganglioside in DKO mice following intrathecal AAV administration. (A, C) TLC analysis was performed on the cortex of 20-week-old WT, 20-week-old DKO mice, as well as 20-week-old DKO mice treated with AAV, and 20- and 30-week-old DKO mice treated with AAV and Istradefylline. (C) TLC analysis was performed on the cortex of 30-week-old WT, 20-week-old DKO mice, as well as 30-week-old DKO mice treated with AAV, and 30-week-old DKO mice treated with AAV and Istradefylline. (B, D) Intensity analyses for GM2 ganglioside were conducted using ImageJ (version 1.48v), and One-way ANOVA was used to determine p-values in GraphPad. Data are presented as means ± SEM (n = 3; ****p < 0.001)

Elevated Hexa enzyme activity in the brain and visceral organs of DKO mice treated with AAV

Hexa enzyme activity levels in the cortex and cerebellum of 20-week-old DKO mice showed residual Hexa activity (Fig. 3A, B). In contrast, Hexa activity was detectable in 20- and 30-week-old WT mice. Furthermore, Hexa activity significantly increased in DKO mice treated with AAV, as well as in those treated with AVV and istradefylline, at both 20 and 30 weeks compared to 20-week-old DKO mice (Fig. 3A, B). No Hexa enzyme activity was observed in the kidney, liver, spleen, and serum of 20-week-old DKO mice (Fig. 3C-F). However, administering AAV, either alone or with istradefylline, significantly increased Hexa enzyme activity in these tissues of both 20- and 30-week-old DKO mice (Fig. 3C-F). Consistent with the Hexa enzyme activity assay, RT-PCR analysis also showed a significant increase in mHexa gene expression in the cortex, cerebellum, and visceral organs including the kidney and spleen, but not the liver, of DKO mice treated with AAV alone or with AAV combined with istradefylline, at both 20 and 30 weeks, compared to untreated 20-week-old DKO mice (Supplementary Fig. 4).

Fig. 3.

Elevated Hexa enzyme activity in the brain and visceral tissues of AAV-treated DKO mice. Hexa enzyme activity was measured in the cortex (A), cerebellum (B), kidney (C), liver (D), spleen (E), and serum (F) of 20-week-old WT, 20-week-old DKO mice, and 20- and 30-week-old DKO mice treated with AAV, as well as 20- and 30-week-old DKO mice treated with AAV and Istradefylline. One-way ANOVA was performed in GraphPad to determine p-values. Data are presented as means ± SEM (n = 3); *p < 0.05, **p < 0.025, ***p < 0.01, and ****p < 0.001)

Altered levels of inflammatory cytokines and chemokines in DKO mice

Gene expression of inflammation- and lysosome-related markers was measured by RT-PCR. In the cortex, 20-week-old DKO mice showed higher levels of CCL3, CCL5, and GFAP compared to age-matched WT mice (Fig. 4B, C, G). Conversely, other neuroinflammatory markers, such as CCL2, Cxcl10, IL-1β, and Iba1, showed no significant differences between groups (Fig. 4A, D, E, F). The CCL3 expression ratio was lower in the cortex of 20- and 30-week-old DKO mice treated with AAV and AAV plus istradefylline than in untreated DKO mice at 20 weeks (Fig. 4B). A reduction in CCL5 and GFAP expression was observed only at 20 weeks (Fig. 4C, G). Elevated lysosomal activity, indicated by increased Lamp1 levels, was seen in the cortex of DKO mice compared to WT at 20 weeks (Fig. 4H). Lysosome buildup decreased in DKO mice treated with AAV alone or with istradefylline at 20 weeks, compared to untreated DKO mice (Fig. 4H).

Fig. 4.

Reduction of neuroinflammation in the cortex (A-H) and cerebellum (I-P) of AAV-treated DKO mice. The expression levels of CCL2, CCL3, CCL5, Cxcl10, IL-1β, Iba1, GFAP, and Lamp1, a lysosome-associated protein, were measured in 20-week-old WT mice, 20-week-old DKO mice, and DKO mice treated with AAV at 20 and 30 weeks, with or without Istradefylline. Expression levels were assessed using the ΔCT method, and p-values were calculated with One-way ANOVA using GraphPad. Data are shown as means ± SEM (n = 3; *p < 0.05, **p < 0.025, ***p < 0.01, and ****p < 0.001)

The levels of neuroinflammatory markers, including CCL2, CCL3, CCL5, Cxcl10, IL-1β, Iba1, and GFAP, as well as the lysosome marker Lamp1, were notably higher in the cerebellum of 20-week-old DKO mice compared to age-matched WT mice (Fig. 4). After treatment with AAV alone or AAV combined with istradefylline, these levels—CCL2, IL-1β, Iba1, and Lamp1—significantly decreased in the cerebellum of DKO mice compared to untreated 20-week-old DKO mice (Fig. 4I, M, N, P).

The analysis of DKO mice treated with istradefylline revealed no significant differences in the expression of inflammatory markers, including CCL2, CCL3, CCL5, Cxcl10, IL-1β, and GFAP, in the cortex and cerebellum compared with 20-week-old DKO mice, except for IL-1β expression in the cerebellum (Supplementary Fig. 6).

Along with the cortex and cerebellum, other visceral organs were examined for inflammatory markers, including CCL2, CCL3, CCL5, Cxcl10, and IL-1β. In the kidneys of 20-week-old DKO mice, the levels of CCL2, CCL3, CCL5, and IL-1β were significantly higher than those in 20-week-old WT mice (Supplementary Fig. 5 A, B, C, E). These inflammatory marker levels decreased after treatment with AAV alone and in combination with istradefylline in DKO mice at 20 and 30 weeks compared to 20-week-old DKO mice (Supplementary Fig. 5 A, B, C, E). Furthermore, significantly higher expression levels of CCL2, CCL3, and CCL5 were observed in the liver of DKO mice compared to WT mice at 20 weeks (Supplementary Fig. 5G, H, I). Regarding CCL3 and CCL5 in the liver, reduced levels were detected in DKO mice treated with AAV and AAV combined with istradefylline at 20 weeks, compared to age-matched DKO mice (Supplementary Fig. 5H, I). In the spleen, the expression levels of CCL5, Cxcl10, IL-1β, and Lamp1 were significantly higher in 20-week-old DKO mice compared to age-matched WT mice (Supplementary Fig. 5 O, P, Q, R). The administration of AAV alone and in combination with istradefylline in DKO mice resulted in lower expression of these genes at both 20 and 30 weeks compared to untreated DKO mice (Supplementary Fig. 5 O, P, Q, R).

Clearance of accumulating GM2 ganglioside and lysosomes, along with the recovery of neurons and oligodendrocytes in DKO mice treated with AAV

Immunohistochemical analyses evaluated how AAV-based gene therapy, alone and combined with istradefylline, affected various factors in mouse groups, including GM2 levels, lysosomal activity, neuronal density, demyelination, and astrocyte activation. In 20-week-old DKO mice, a significant accumulation of GM2 ganglioside was seen in the cortex, hippocampus, thalamus, cerebellum, and pons compared to age-matched controls (Fig. 5). After treatment with AAV alone and combined with istradefylline, there was a notable decrease in GM2 ganglioside accumulation across all examined brain regions in DKO mice (Fig. 5). Alongside measuring GM2 ganglioside levels, lysosome levels were evaluated using Lamp1 staining to assess lysosomal buildup in brain sections (Fig. 6). Higher lysosome levels were observed in the cortex, hippocampus, thalamus, cerebellum, and pons of DKO mice at 20 weeks compared to age-matched controls (Fig. 6). Although there were slight decreases in lysosome levels in the cortex, hippocampus, and cerebellum of AAV-treated and AAV with istradefylline-treated DKO mice at both 20 and 30 weeks (Fig. 6A, B, C, E), significant reductions were detected in the thalamus and pons of these mice relative to 20-week-old DKO mice (Fig. 6A, D, F). Neuronal density was assessed using anti-NeuN staining to examine how mHexa expression, following AAV-based gene therapy, affects neurons. Neuron loss was observed, indicating neurodegeneration in the cortex, hippocampus, thalamus, and pons of DKO mice at 20 weeks compared to age-matched controls (Fig. 7). No significant improvement in neuronal density was observed in the cortex, hippocampus, thalamus, or cerebellum of 20-week-old DKO mice following istradefylline treatment alone (Supplementary Fig. 7). Notably, AAV-mediated gene therapy, combined with AAV and istradefylline treatment, restored neuronal density, resulting in increased neuronal numbers in the cortex, hippocampus, thalamus, and pons of DKO mice at 20 and 30 weeks compared to their 20-week state (Fig. 7B-D, F). However, neuron density in the cerebellum showed no significant change across all mouse groups (Fig. 7E). Immunostaining with anti-CNPase revealed a loss of oligodendrocytes, indicating demyelination in the cortex, thalamus, cerebellum, and pons of DKO mice at 20 weeks compared to age-matched WT mice (Fig. 8A-E). Furthermore, DKO mice treated with istradefylline showed oligodendrocyte loss in the cortex, thalamus, and cerebellum, as seen in 20-week-old DKO mice, compared with age-matched WT mice (Supplementary Fig. 8), indicating that istradefylline did not affect oligodendrocyte density. An increased level of oligodendrocytes, as shown by anti-CNPase intensity, was observed in both 20- and 30-week-old DKO mice treated with AAV and AAV with istradefylline, compared to 20-week-old DKO mice across all examined brain regions, including the cortex, thalamus, cerebellum, and pons (Fig. 8A-E). According to anti-GFAP staining results, activated astrocytes, identified by their morphology, were found in the cortex, hippocampus, and pons of 20-week-old DKO mice (Fig. 8F). In the cerebellum of these mice, active astrocytes were present but less numerous compared to other brain regions such as the cortex, hippocampus, and pons (Fig. 8F). Istradefylline treatment had no detectable effect on astrogliosis in the brains of istradefylline-treated DKO mice, compared to DKO mice (Supplementary Fig. 9). However, no active astrocytes were observed in the cortex and cerebellum of 20-week-old DKO mice treated with AAV and AAV with istradefylline. In contrast, activated astrocytes were found in the pons and cerebellum of 30-week-old DKO mice treated with AAV and AAV with istradefylline (Fig. 8F).

Fig. 5.

Clearance of GM2 ganglioside in the brains of AAV-treated DKO mice. Immunostaining of GM2 in coronal brain sections (A) includes regions such as the cortex, hippocampus, thalamus, cerebellum, and pons of 20- and 30-week-old WT mice, 20-week-old DKO mice, and 20- and 30-week-old DKO mice treated with AAV, as well as DKO mice treated with AAV and Istradefylline. Images were captured at 20X magnification under consistent light intensity, which varies for DAPI and anti-KM966. Sections were stained with anti-KM966 antibody (green; GM2 marker) and DAPI (blue; nucleus). GM2 intensity was quantified using ImageJ (B-F). One-way ANOVA was used to calculate p-values with GraphPad. Data are shown as means ± SEM (n = 3; *p < 0.05, **p < 0.025, ***p < 0.01, ****p < 0.001)

Fig. 6.

No change in Lamp1 levels in the brains of DKO mice after AAV treatment, except in the thalamus and pons. Immunostaining of Lamp1 (lysosome marker) in coronal brain sections (A) from cortex, hippocampus, thalamus, cerebellum, and pons of 20- and 30-week-old WT, 20-week-old DKO mice, and 20- and 30-week-old DKO mice treated with AAV, as well as DKO mice treated with AAV and Istradefylline. Images were taken at 20X magnification under identical light conditions, using different filters for DAPI and anti-Lamp-1. Sections were stained with anti-Lamp1 antibody (green; lysosome marker) and DAPI (blue; nucleus). Lamp1 intensity was quantified using ImageJ (B-F). One-way ANOVA was used to calculate p-values with GraphPad. Data are shown as means ± SEM (n = 3; *p < 0.05, **p < 0.025, ***p < 0.01, ****p < 0.001)

Fig. 7.

Increase in the number of neurons in the brains of AAV-treated DKO mice. Immunostaining of NeuN (a neuron marker) in coronal brain sections (A-F) includes the cortex, hippocampus, thalamus, cerebellum, and pons of 20- and 30-week-old WT, 20-week-old DKO mice, as well as 20- and 30-week-old DKO mice treated with AAV, and with or without Istradefylline. Images were captured at 20X magnification under identical lighting conditions. NeuN intensity was quantified using ImageJ (B-F). One-way ANOVA determined p-values with GraphPad. Data are shown as means ± SEM (n = 3; *p < 0.05, **p < 0.025)

Fig. 8.

Restoration of oligodendrocyte density and reduction of astrogliosis in the brains of AAV-treated DKO mice. Immunostaining of CNPase (an oligodendrocyte marker) in coronal brain sections, including the cortex, thalamus, cerebellum, and pons (A-E) and GFAP (an astrocyte marker) (F) in coronal brain sections, including the cortex, hippocampus, cerebellum, and pons of 20- and 30-week-old WT, 20-week-old DKO mice, and 20- and 30-week-old DKO mice treated with AAV, and with or without Istradefylline. Images were captured at 20X magnification under consistent light intensity, although the light settings differ for DAPI, anti-CNPase, and anti-GFAP staining. Sections were stained with anti-CNPase and anti-GFAP (green) and DAPI (blue; nuclei). The CNPase intensities were measured using ImageJ (B-E). One-way ANOVA was used to determine p-values with GraphPad. Data are presented as means ± SEM (n = 3; *p < 0.05, **p < 0.025, ***p < 0.01, and ****p < 0.001)

Discussion

Various therapeutic strategies have been explored for LSDs, including enzyme replacement therapy (Matsuoka et al. 2010; Tsuji et al. 2011), substrate reduction therapy (Andersson et al. 2004), pharmacological chaperone therapy (Chiricozzi et al. 2014; Maegawa et al. 2007), hematopoietic stem cell transplantation, and gene therapy (Fernández-Pereira et al. 2021). As previously noted, monogenic LSDs are suitable for gene therapy because the effective gene is well understood, and delivering this gene promotes the expression of the deficient enzyme. However, a more challenging issue is delivering substances to the CNS to reach affected tissues. Advances in vector engineering and delivery methods have improved the feasibility of CNS-targeted gene therapy, which remains crucial for LSDs with severe neurological symptoms (Massaro et al. 2021).

Adeno-associated virus (AAV)-based gene therapy provides an effective delivery platform, with AAVrh10 frequently used in treatments targeting the central nervous system. Intra-cerebrospinal fluid injections of AAVrh10-based LYS-GM101 analogs have demonstrated success in mouse and cat models of GM1 gangliosidosis, facilitating GM1 ganglioside breakdown in the CNS without notable adverse effects (Hocquemiller et al. 2022). In a mouse model of mucopolysaccharidosis VII, a lysosomal storage disorder, a single intrathecal injection led to stable transduction of CNS structures, improved behavior, and longer lifespan (Pagès et al. 2019). Previously, a bicistronic vector carrying HEXA and HEXB (AAV9-HEXA/HEXB) was shown to degrade GM2 ganglioside and double survival rates in mouse models of GM2 gangliosidosis, Sandhoff disease (SD), after intravenous delivery (Woodley et al. 2019). More recently, the optimal dose for this bicistronic vector was identified among three options—2.5 × 10^11, 1.25 × 10^11, and 0.625 × 10^11 vector genomes—after testing in a mouse model of SD (Ryckman et al. 2023).

Unlike most methods that deliver the vector into the lateral ventricle (ICV) or the cisterna magna, intrathecal delivery through lumbar puncture is easier to perform, less invasive, and covers a larger area of the CNS. This technique is already commonly used in clinical practice for anesthesia and for delivering recombinant enzymes into the CNS for various conditions (Pagès et al. 2019). Intrathecal administration allows broad distribution of AAV vectors throughout the CNS while avoiding direct injection into brain tissue, which reduces surgical risks and procedure-related complications. For these reasons, we chose intrathecal delivery for our study. Additionally, lumbar intrathecal injection is a standard and straightforward procedure in humans, making it a highly relevant and translatable method for CNS-targeted gene therapy (Boitnott et al. 2025). Repeated intrathecal administration is already employed in both animal models and patients for various treatments. Repeated intrathecal enzyme replacement therapy in mucopolysaccharidosis models has shown that recombinant human 4-sulphatase can effectively reduce storage in mature MPS VI cats (Auclair et al. 2010). Furthermore, repeated intrathecal idursulfase has been administered monthly to patients with neuronopathic MPS II, demonstrating tolerance, decreased CSF glycosaminoglycan levels, and stabilized cognitive function (Muenzer et al. 2022). It has also been shown that as AAV gene therapies develop, repeated injections via intrathecal, intracerebroventricular, or intracisternal routes for CSF delivery may offer new options for treating neurodegenerative diseases. However, successful redosing may require carefully planned immunosuppressive strategies to prevent immune responses against capsid proteins of AAV serotypes (McElroy et al. 2022).

Despite testing various approaches, effective treatments for Tay-Sachs disease remain unavailable. In this study, DKO mice received an intrathecal injection of AAVrh10 expressing mHexa (containing 1 × 10¹¹ viral genomes) at 8 weeks of age. The in vivo effects of combined AAV-based gene and anti-inflammatory therapies were also evaluated (Fig. 1). In our study, we used AAV-based gene therapy in a mouse model of Tay-Sachs disease. Our initial dose choice was based on the scAAV9-Hexb study in Sandhoff (Hexb-/-) mice, which employed very high systemic vg doses during the neonatal period. This research guided our rationale for selecting the initial dose (Niemir et al. 2018). We first tested approximately 1 × 10¹³ vg/kg, equivalent to 1 × 10¹¹ vg for a 10-gram mouse. Importantly, this dose caused no adverse effects after injection. Consequently, we optimized the dose to 1 × 10¹¹ vg/mouse, and the studies reported here used this dose, which was well tolerated.

As previously noted (Seyrantepe et al. 2018), these DKO mice exhibited reduced size, lower body weight, and a high incidence of sudden death around 20 weeks of age. Consistent with this, body weight measurements showed a rapid, significant decline in DKO mice after 16 weeks compared to WT mice (Fig. 1B, C). The AAV therapy, alone or combined with istradefylline, effectively restored the body weight of DKO mice (Fig. 1B, C). While DKO mice typically survived until 20 weeks, AAV treatment—alone or with istradefylline—extended survival up to 30 weeks (Fig. 1A, D), demonstrating that AAV-mediated mHexa expression, together with the anti-inflammatory drug istradefylline, improves body weight and prolongs lifespan in DKO mice. In an early-onset Tay-Sachs disease mouse model, impaired locomotor activity and motor coordination have been previously observed. Reduced locomotor activity was also confirmed in these mice using an open-field test. In this study, as expected, DKO mice at 20 weeks of age could not stay on the rotating rod, indicating severe impairments in motor activity (Fig. 1E). However, motor activity improved in DKO mice treated with AAV alone and in combination with istradefylline (Fig. 1E).ombination with istradefylline (Fig. 1E).

In GM2 gangliosidoses, including Tay-Sachs and Sandhoff diseases, excessive accumulation of GM2 ganglioside has been previously observed in both patients and mouse models (Seyrantepe et al. 2018; Solovyeva et al. 2018). Here, TLC analyses showed a progressive buildup of GM2 ganglioside in the cortex of DKO mice (Fig. 2A, C). Administration of either AAV alone or combined with istradefylline resulted in a significant decrease in GM2 ganglioside levels in the cortex of DKO mice at both 20 (Fig. 1B) and 30 weeks of age (Fig. 1D), demonstrating the effectiveness of AAV-based gene therapy in reducing GM2 ganglioside through mHexa expression. GM2 ganglioside levels were also measured using anti-GM2 staining. Consistent with previous studies, excessive GM2 accumulation was observed in the cortex, hippocampus, thalamus, cerebellum, and pons of 20-week-old DKO mice compared to age-matched controls (Fig. 5). The effectiveness of scAAV9.47 expressing HEXM (which encodes the α subunit and stabilizes the β subunit interface) was confirmed in Hexa-/- mice, where intracranial delivery of this vector degraded GM2 ganglioside (Tropak et al. 2016). In this study, treatment with AAV alone and AAV combined with istradefylline resulted in a significant reduction of GM2 accumulation across all brain regions examined in DKO mice at 20 and 30 weeks (Fig. 5). This indicates that AAV-based gene therapy decreases GM2 buildup through mHexa expression. It should be noted that the level of GM2 reduction varies between TLC and IHC analyses. These two assays measure GM2 levels using different methods, leading to variability in results. For IHC, we used the anti-GM2 monoclonal antibody KM966, which has been shown to specifically recognize GM2 without cross-reacting with other gangliosides (Tordo et al. 2018). IHC allows for region-specific assessment of GM2 accumulation in the cortex, hippocampus, thalamus, cerebellum, and pons. In contrast, TLC provides a biochemical measurement that reflects the average GM2 content of the entire cortex rather than individual regions. Therefore, the level of GM2 reduction in the brains of treated DKO mice, as shown by IHC (Fig. 5), is significantly higher than that seen with TLC (Fig. 2).

In mouse models of various lysosomal storage disorders such as Krabbe disease (Rafi et al. 2012), neuronal ceroid lipofuscinosis (Sondhi et al. 2007), metachromatic leukodystrophy (Piguet et al. 2012), diabetic neuropathy (Homs et al. 2014), and amyotrophic lateral sclerosis (Wang et al. 2014), AAVrh10 gene therapy has been tested through intracranial and intrathecal delivery. These studies demonstrate that AAVrh10-related gene therapy effectively targets neurons and oligodendrocytes, leading to reductions in disease-related features within the central nervous system. Increased expression of the target gene led to higher levels of its enzyme products in the brains of AAVrh10-treated mice. In addition to the central nervous system, AAVrh10 also shows a preference for various tissues, including the kidney, liver, spleen, lung, muscle, and eye (Hoshino et al. 2019). AAVrh10 shows tropism for neurons, astrocytes, and oligodendrocytes (Hoshino et al. 2019), and systemic administration of this vector results in higher transgene expression in both the brain and spinal cord of neonatal mice compared to AAV9 (Tanguy et al. 2015). In mice, intrathecal administration of AAVrh10 efficiently transduces Schwann cells in the peripheral nervous system, with a gradient of vector biodistribution from the injection site (Kagiava et al. 2021). In our study, Hexa enzyme activity was assessed using an enzyme activity assay in the brain and other visceral organs (kidney, liver, and spleen) after mice were injected with AAVrh10 expressing mHexa (AAV). In the brain, both the cortex (Fig. 3A) and cerebellum (Fig. 3B) showed significantly increased Hexa enzyme activity in 20- and 30-week-old DKO mice administered with AAV alone or combined with istradefylline. These results indicate successful expression of mHexa (Supplementary Fig. 4) and increased Hexa enzyme activity in the brains of DKO mice. Furthermore, the absence of age-dependent changes in DKO mice treated with AAV alone or with istradefylline suggests that AAVrh10 facilitates sustained long-term expression of mHexa (Fig. 3). It should be noted that a significant increase in HexA enzyme activity was observed in AAV-treated DKO mice; however, this increase did not reach the levels seen in WT mice. HexA is a heterodimer enzyme composed of an α (HEXA) and a β (HEXB) subunit, and in our study, we administered AAVrh10 encoding the Hexa gene. Administering only the Hexa gene, which encodes the α subunit, may not produce a substantial increase in functional HexA heterodimers. As shown in a previous study, using a bicistronic vector expressing β-hexosaminidase A to deliver both subunits can further increase HexA enzyme activity in the mouse model of Sandhoff disease (Ryckman et al. 2023). In addition to the CNS, the heightened Hexa enzyme activity in visceral organs — kidney (Fig. 3C), liver (Fig. 3D), and spleen (Fig. 3E) — and in the serum (Fig. 3F) of DKO mice receiving AAV and istradefylline demonstrates AAVrh10’s tropism for other organs and the presence of AAVrh10 in the bloodstream after intrathecal administration.

Elevated neuroinflammation has been observed in several lysosomal storage disorders, including Niemann-Pick disease type C (Vruchte et al. 2004), sialidosis (D’Azzo, Machado, and Annunziata 2015), and GM1 gangliosidosis (Jeyakumar et al. 2003). Previous studies detected increased levels of cytokines and chemokines associated with neuroinflammation—such as Ccl2, Ccl3, Ccl4, and Cxcl10—along with decreased anti-inflammatory markers in the brains of DKO mice (Demir et al. 2020). Istradefylline, an anti-inflammatory drug, was also shown to reduce pro-inflammatory cytokines, chemokines, and microglia activation (Iba1) in the brains of Hexb-/- mice, a Sandhoff disease model (Ogawa et al. 2018). Consistent with these findings, 20-week-old DKO mice showed increased neuroinflammatory markers—including CCL3, CCL5, and GFAP in the cortex (Fig. 4B, C, G), and CCL2, CCL3, CCL5, Cxcl10, IL-1β, Iba1, and GFAP in the cerebellum (Fig. 4I-O)—as demonstrated by RT-PCR. These results confirm the presence of neuroinflammation in the brains of these mice. Notably, DKO mice administered istradefylline at 20 weeks showed no differences in the expression of CCL2, CCL3, CCL5, CXCL10, and GFAP in the cortex and cerebellum compared with age-matched DKO mice (Supplementary Fig. 6). These findings indicate that istradefylline treatment alone has no significant impact on neuroinflammatory gene expression. After administering AAV alone or combined with istradefylline, there was a notable reduction in CCL3, CCL5, and GFAP in the cortex (Fig. 4B, C, G), and CCL2 and IL-1β in the cerebellum (Fig. 4I, M), especially at 20 weeks, compared to untreated age-matched DKO mice. These findings suggest that AAV-based gene therapy expressing mHexa can mitigate neuroinflammation in DKO mice. Additionally, DKO mice treated with AAV and istradefylline showed significantly lower Iba1 levels in the cerebellum at 20 weeks, indicating an anti-inflammatory effect of istradefylline (Fig. 4M). However, CCL2 levels were higher in the cortex of DKO mice treated with AAV and istradefylline compared to untreated WT mice at 30 weeks of age (Fig. 4A). This difference reflects the variation in inflammatory response between WT and DKO mice at 30 weeks. Furthermore, CCL2 levels in DKO mice treated with AAV and those treated with AAV and istradefylline are not significantly different at 30 weeks (Fig. 4A). Therefore, there is no clear evidence that istradefylline causes an additional increase in inflammation. The slight rise observed in the DKO mice treated with AAV and istradefylline may reflect variability at 30 weeks rather than a pro-inflammatory effect of istradefylline.

Prior research has also reported excessive expression of inflammatory markers, including IL-1β, TNFα, and IL-6, across multiple organs in a Gaucher disease mouse model (Pandey 2023). Western blot analyses showed no significant change in IL-1β levels in the brains of these mice (Supplementary Fig. 10 A, B). Conversely, decreased IL-1β expression was observed in the kidney of AAV-treated DKO mice and in the spleen of AAV- and istradefylline-treated DKO mice at 20 weeks (Supplementary Fig. 10 C, E). In the liver, elevated IL-1β levels were significantly reduced in DKO mice treated with AAV alone or combined with istradefylline at both 20 and 30 weeks (Supplementary Fig. 10D). RT-PCR analysis of 20-week-old DKO mice showed that pro-inflammatory markers, such as CCL2, CCL3, and CCL5, were significantly elevated in the kidney (Supplementary Fig. 5 A-C) and liver (Supplementary Fig. 5G-I), while CCL5, Cxcl10, and IL-1β levels were higher in the spleen (Supplementary Fig. 5P-S) compared to wild-type controls. These findings indicate increased inflammation in visceral organs. Treatment with AAV alone or combined with istradefylline led to decreased levels of CCL3 and CCL5 in the kidney and liver, as well as Cxcl10 and IL-1β in the spleen, suggesting reduced inflammation in these organs.

Immunohistochemistry was used to evaluate the effects of AAV-based gene therapy alone and in combination with istradefylline. Lysosomal levels, assessed with anti-Lamp1 staining, showed lysosomal accumulation in the cortex, hippocampus, thalamus, cerebellum, and pons of 20-week-old DKO mice (Fig. 6). Previous research demonstrated a decrease in lysosomes in a mucopolysaccharidosis VII mouse model following intrathecal injection of AAVrh10 encoding human β-glucuronidase (Pagès et al. 2019). Here, we observed significant reductions in lysosomes in the thalamus (Fig. 6D) and pons (Fig. 6F) of DKO mice treated with AAV alone and AAV combined with istradefylline.

Neurodegeneration is a common feature of various lysosomal storage disorders and neurodegenerative diseases. Reduced neuronal density and increased apoptosis have been observed in the brains of early-onset mouse models of Tay-Sachs disease (Demir et al. 2020). In this study, neuronal loss, a sign of neurodegeneration, was found in the cortex, hippocampus, thalamus, and pons of DKO mice at 20 weeks compared to age-matched WT mice (Fig. 7). Although istradefylline treatment alone did not significantly improve neuronal density (Supplementary Fig. 7), AAV-based gene therapy restored neuronal density in these brain regions of DKO mice at both 20 and 30 weeks (Fig. 7). These findings suggest that GM2 clearance in the brain prevented neurodegeneration and helped restore neuronal density in DKO mice.

Oligodendrocytes produce a myelin sheath that coats axons, facilitating signal transmission (Onyenwoke and Jay 2015). Some gangliosides, including GD1a and GT1b, are localized on neuronal axons and promote the myelination process (Can et al. 2022). However, excessive lipid and ganglioside buildup can disrupt these myelination processes. In addition to GM2 ganglioside accumulation, increased neuroinflammation further makes oligodendrocytes more vulnerable, leading to demyelination (Peferoen et al. 2014). It has been previously shown that patients with Tay-Sachs and Sandhoff diseases exhibit lower levels of myelin basic protein (MBP), a marker of myelination (Myerowitz et al. 2002). Moreover, higher levels of demyelination, caused by abnormal GM2 ganglioside accumulation, were observed in the brains of DKO mice using anti-myelin basic protein staining and Luxol Fast Blue staining (Can et al. 2022). Our results showed a reduced number of oligodendrocytes in the cortex, thalamus, cerebellum, and pons of DKO mice at 20 weeks (Fig. 8A-E). Istradefylline alone did not significantly change oligodendrocyte numbers in the cortex, thalamus, and cerebellum, which remained similar to those in untreated DKO mice, indicating that istradefylline treatment is insufficient to restore oligodendrocyte density (Supplementary Fig. 8). In a mouse model of Metachromatic Leukodystrophy, a lysosomal storage disorder, AAVrh10 encoding the human ARSA gene, controlled by the cytomegalovirus/β-actin hybrid (CAG/cu) promoter, was injected intrastriatally, resulting in restoration of oligodendrocyte density (Piguet et al. 2012). Consistently, all examined brain sections from DKO mice treated with AAV alone or in combination with istradefylline showed increased oligodendrocyte levels compared to untreated DKO mice, indicating an enhancement of the myelination process (Fig. 8A-E).

Active astrocytes, a condition known as astrogliosis, have been observed in the cerebral region of Niemann-Pick A mouse models, in the cortex of the Sandhoff mouse model (Sango et al. 1995), and in the hippocampus and cerebellum of the Pompe mouse model (Clarke et al. 2021). Additionally, in the cortex, hippocampus, and cerebellum of the Tay-Sachs disease mouse model, active astrocytes were visualized using anti-GFAP staining (Demir et al. 2020). In this study, active astrocytes were identified by morphology. It was demonstrated that active astrocytes are present in the cortex, hippocampus, and pons of 20-week-old DKO mice (Fig. 8F). In DKO mice, istradefylline treatment alone did not significantly reduce active astrocytes in the cortex, thalamus, and cerebellum compared with age-matched DKO mice (Supplementary Fig. 9). In contrast, no active astrocytes were detected, particularly in the cortex and cerebellum, in 20-week-old DKO mice treated with AAV or AAV combined with istradefylline (Fig. 8F).

In summary, this study is the first to demonstrate the clearance of accumulated GM2 ganglioside and a reduction in neuroinflammation in the brains of mouse models with early-onset Tay-Sachs disease. The success of intrathecal AAVrh10-based gene therapy, combined with istradefylline treatment, may suggest a new therapeutic option for patients with Tay-Sachs disease.

Supplementary Information

Below is the link to the electronic supplementary material.

Author contributions

MC conducted the experiments, analyzed and interpreted the data, and assisted in preparing the manuscript. JA provided helpful comments and participated in discussions about the results. VS designed the project, secured funding, drafted and revised the manuscript. All authors read and approved the final version of the manuscript.

Funding

Open access funding provided by the Scientific and Technological Research Council of Türkiye (TÜBİTAK). This study was funded by TUBITAK-FRANCE Project (Grant No: 120N552).

Data availability

The data supporting the results of this study are available from the corresponding author upon reasonable request.

Declarations

Competing interests

The authors declare no competing interests.