B4Galnt1 Deficiency Reverses Severe Neurological Symptoms in a Mouse Model of Tay-Sachs Disease

TL;DR

Removing the B4Galnt1 gene in a mouse model of Tay-Sachs disease reversed severe neurological symptoms and extended lifespan.

Contribution

B4Galnt1 deficiency is shown to reverse disease symptoms by preventing GM2 ganglioside accumulation in a Tay-Sachs mouse model.

Findings

TKO mice showed halted GM2 ganglioside accumulation.

Neuroinflammation and neuronal death were reduced in TKO mice.

Lifespan was extended by over 18 months in the TKO model.

Abstract

Tay-Sachs disease is a severe neurodegenerative disorder caused by mutations in the HEXA gene, which encodes the α-subunit of the β-hexosaminidase A (HexA) enzyme. HexA deficiency leads to abnormal GM2 accumulation, eventually causing cell death and neurodegeneration. A double-knockout mouse model lacking both Hexa and Neu3 genes (Hexa-/-Neu3-/-, DKO) exhibits neuropathological and clinical features similar to those of the disease, including neuroinflammation. B4Galnt1 (ß-1,4-N-acetyl-galactosaminyltransferase 1) is involved in lipid biosynthesis in mice. We hypothesized that creating a triple knockout model (Hexa-/-Neu3-/-B4Galnt1-/-, TKO) could prevent excessive GM2 ganglioside accumulation and reduce disease symptoms. Molecular biology and immunohistochemistry analyses showed that GM2 ganglioside accumulation was halted in TKO mice. Preventing GM2 ganglioside accumulation alleviated neuroinflammation and neuronal death, extending lifespan by more than 18 months. Our findings suggest that knocking out B4Galnt1 to block GM2 ganglioside accumulation may reverse disease symptoms in the DKO mouse model, indicating a promising, safe target for substrate-reduction therapy via siRNA silencing.

The online version contains supplementary material available at 10.1007/s12017-026-08916-x.

Figures

Body weights/sizes measurements, survival paths, and TLC analysis. (A) Body weights of 5-month-old *WT* (*n* = 12), *Hexa-/-* (*n* = 10), *B4Galnt1-/-* (*n* = 14), *DKO* (*n* = 12), and *TKO* (*n* = 13) mice over 20 weeks are shown. Each symbol represents the mean body weight for mice of the corresponding genotype. (B) Gross appearances of 5-month-old *WT*, *DKO*, and *TKO* mice. (C) Kaplan-Meier survival path of *WT*, *DKO*, and *TKO* mice. Acidic sphingolipids were separated and stained from the cortex (D) and cerebellum (E) brain regions of 5-month-old *WT*, *Hexa-/-*, *B4Galnt1-/-*, *DKO*,

GM2 ganglioside levels in brain sections of 5-month-old mice. Cortical, hippocampal, and cerebellar sections from 5-month-old *WT* (A, F, and K, respectively), *Hexa-/-* (B, G, and L, respectively), *B4Galnt1-/-* (C, H, and M, respectively), *DKO* (D, I, and N, respectively), and *TKO* (E, J, and U, respectively) mice were stained with anti-GM2 antibody (green) to detect GM2 ganglioside levels and with DAPI (blue). The scale bar is 50 μm. The data represent three independent experiments

Neuroinflammation-related gene expression analysis in the brain regions of 5-month-old mice. *Ccl2*, *Ccl3*, *Ccl5*, *Cxcl10*, *Il1ß*, *Gfap*, and *Iba1* gene expression levels in the cortex (A-G) and cerebellum (H-N) of 5-month-old *WT*, *Hexa-/-*,* B4Galnt1-/-*, *DKO*, and *TKO* mice were analyzed by qRT-PCR. Gene expression ratios relative to WT mice were calculated using the ΔCT method, and fold changes were reported. Data represent three independent experiments as the mean ± SEM. Statistical analyses were performed using One-way ANOVA (p* < 0.05, *p* < 0.025, p* < 0.01, and *p* < 0.001)

A, B, C, and D). *DKO* mice took longer to complete the platform than age-matched littermates. Additionally, *DKO* mice had significantly shorter strides, sway, and stance (Supplementary Fig. 4A, B, C, and D). As previously reported (Seyrantepe et al., [2018](#CR28)), paw prints indicate that the neurons controlling the hind legs are affected in these mice. In contrast, 5-month-old *TKO* mice took as long to complete the platform as *WT* mice, and there were no significant alterations in stride, sway, or stance length in these mice (Supplementary Fig. 4A, B, C, and D). To assess whether there

Immunofluorescence analysis of activated microglia. Sections of the cortex (A, F, K, and P), hippocampus (B, G, L, and R), cerebellum (C, H, M, and S), and pons (D, I, N, and T) from 5-month-old *WT*, *Hexa-/-*,* B4Galnt1*-/-, *DKO*, and *TKO* mice were stained with an anti-Moma2 (red) antibody to detect active microgliosis and with DAPI (blue). Histograms show Moma2 intensity in the cortex (V), hippocampus (W), cerebellum (X), and pons (Z). The scale bar is 50 μm. Data represent three independent experiments, with mean ± SEM. Statistical analyses were performed using One-way ANOVA (p* < 0.05,

Density of neuronal cells in brain sections from 5-month- and 18-month-old mice. Cortex, hippocampus, thalamus, and cerebellum sections from 5-month-old *WT* (A, F, K, and P, respectively), 5-month-old *DKO* (B, G, L, and R, respectively), 5-month-old *TKO* (C, H, M, and S, respectively), 18-month-old *WT* (D, I, N, and T, respectively), and 18-month-old *TKO* (E, J, O, and U, respectively) mice were immunostained with an anti-NeuN antibody (red) to detect neuronal loss. Histograms show NeuN intensity measurements in the cortex (V), hippocampus (W), thalamus (X), and cerebellum (Z). The scale

Full text

Introduction

Tay-Sachs disease (TSD), also called Gangliosidosis GM2 Type 2, is a severe autosomal recessive lysosomal storage disorder (LSD) that mainly impacts the central nervous system (CNS). It leads to neurological problems caused by mutations in the HEXA gene, which encodes the α-subunit of the enzyme β-hexosaminidase A (HexA). This enzyme breaks down GM2 into GM3 ganglioside by removing N-acetylgalactosamine. Mutations in HEXA cause GM2 ganglioside to build up in patients. TSD is one of three GM2-gangliosidoses, along with Sandhoff disease (SD) and GM2AP deficiency (Toro et al., 2021).

Hexa-/- and Hexb-/- mouse models have been created to mimic the different forms of human GM2 gangliosidosis: TSD and SD, respectively (Phaneuf et al., 1996). In the Hexb-/- model, significant GM2 ganglioside accumulation was observed, along with stiffness, tremors, ataxia, muscle weakness, and spasticity. In contrast, the Hexa-/- model showed only minor GM2 accumulation and a largely normal phenotype, unlike TSD patients (Phaneuf et al., 1996; Yamanaka et al., 1994; Sango et al., 1995). This difference arises from a sialidase-mediated metabolic bypass in Hexa-/- mice, where GM2 is transformed into GA2 ganglioside by Neu3 sialidase, allowing further degradation by β-hexosaminidase B (Phaneuf et al., 1996; Yamanaka et al., 1994; Sango et al., 1995; Yuziuk et al., 1998). A double-knockout mouse model deficient in both Hexa and Neu3 genes (Hexa-/-Neu3-/-, DKO) was developed to investigate the neuropathological aspects of TSD.

DKO mice exhibit key human symptoms, including tremors, ataxia, gait and motor deficits, cytoplasmic vacuolization of neurons, Purkinje cell loss, and neuronal death, all caused by excessive GM2 ganglioside accumulation. Although healthy at birth, DKO mice typically die around 5 months, a lifespan reduction similar to that seen in Tay-Sachs patients due to GM2 accumulation. In addition to elevated GM2 in the brain, DKO mice also accumulate GM3, GA2, and LacCer (Seyrantepe et al., 2018). We found that DKO mice exhibit increased pro-inflammatory cytokines and chemokines, along with activated microglia and astrocytes, which are linked to cognitive and motor deficits (Demir et al., 2020). Gene expression analysis of pro- and anti-apoptotic markers and TUNEL assays across brain regions demonstrates ongoing neuronal loss and neurodegeneration in this model (Seyrantepe et al., 2018).

ß-1,4-N-acetyl-galactosaminyltransferase 1 (B4Galnt1, also called GalNAcT) is an enzyme involved in synthesizing complex gangliosides. It transfers N-acetylgalactosamine (GalNAc) from UDP-GalNAc to GM3 and GD3 gangliosides, resulting in GM2 and GD2. Because of this function, it is also known as the GM2/GD2 synthase (Takamiya et al., 1996). B4Galnt1-/- mice cannot produce complex gangliosides but maintain high levels of simple gangliosides GM3 and GD3. Despite this, they are born normally, develop brains typically, and have normal lifespans, though male sterility occurs (Liu et al., 1999). However, at 3 months of age, they exhibit axonal damage and demyelination in both the central and peripheral nervous systems, likely due to the importance of complex gangliosides at the axon-myelin interface (Sheikh et al., 1999). They also develop motor problems, including balance, coordination, and strength issues, as well as reflex impairments. From 8 months of age, tremors and catalepsy can also be observed (Chiavegatto et al., 2000).

Mutations in the B4GALNT1 gene are linked to spastic paraplegia 26 (SPG26), a form of Hereditary Spastic Paraplegias (HSPs). These rare neurodegenerative conditions feature symptoms like limb weakness, spasticity, intellectual disability, ataxia, and peripheral neuropathy (Wang et al., 2021; Alecu et al., 2022). Recent studies indicate that a deficiency in complex gangliosides, caused by new mutations in B4GALNT1, may underlie the clinical signs of SPG26 (Inamori et al., 2024).

In this study, we hypothesized that knocking out the B4Galnt1 gene in DKO mice could prevent potentially GM2 ganglioside accumulation and ease TSD symptoms.

Materials and Methods

Generation of Triple Knock-Out Mice

Hexa-/-, Neu3-/-, DKO mice models were generated as previously described (Seyrantepe et al., 2018). B4Galnt1-deficient (B4Galnt1-/-) mice were donated by Prof. Dr. Roger Sandhoff (Deutsches Krebsforschungszentrum (DKFZ), Heidelberg, Germany) as part of the collaborative project “Intensified Cooperation (IntenC): Promotion of German-Turkish Higher Education Research.” To generate Hexa-/-Neu3-/-B4Galnt1-/- (TKO) mice, Hexa+/-Neu3-/- mice, obtained from breeding Hexa-/- and Neu3-/- mice (Seyrantepe et al., 2018), were crossed with B4Galnt1+/- mice. The genotypes of Hexa, Neu3, and B4Galnt1 were determined by PCR using genomic DNA from the mice’s tails. PCR identified the wild-type (WT) and mutant alleles of Hexa and Neu3 with Taq DNA polymerase (Invitrogen Life Technologies, USA) and the primers previously described (Seyrantepe et al., 2018). Similarly, the alleles of B4Galnt1 were genotyped by PCR using Taq DNA polymerase and the primers listed in Supplementary Table 1. The combination of PCR procedures for the Hexa, Neu3, and B4Galnt1 alleles enabled unambiguous identification of all nine potential genotypes.

The mice were housed in groups of five per cage and maintained at a constant temperature with an alternating 12-h light/dark cycle. Food and water were available ad libitum. The Turkish Institute of Animal Health conducted all animal experiments in accordance with the Guide for the Care and Use of Laboratory Animals. The Institutional Animal Care and Use Committee of Izmir Institute of Technology approved the animal studies.

Ganglioside Extraction and Thin-Layer Chromatography (TLC)

Total lipid extraction was performed on brain tissue lysates from 5-month-old mice, following a protocol adapted from a previously described method (Seyrantepe et al., 2010). Briefly, total lipids were extracted from frozen brain samples by adding a methanol/chloroform mixture (1:1, v/v). The lysates were homogenized with an Ultra-Turrax homogenizer (IKA, Germany) and sonicated with a Bandelin Sonicator (Germany). Phosphate-buffered saline was added to the homogenates to induce phase separation, and the mixture was then centrifuged. The collected phases were passed through a Chromabond C18 (Macherey-Nagel, Germany) column. Gangliosides were eluted with methanol and a methanol/chloroform mixture. The samples were evaporated under a nitrogen stream, the residues were dissolved in methanol/chloroform, and the solution was applied to silica-coated TLC plates (Merck, USA) using a Linomat V (CAMAG, Switzerland). The plates were developed in methanol/chloroform/0.22% CaCl2 and stained with orcinol.

Quantitative Real-Time PCR and Western Blot Analyses

Total RNA was extracted from the cortex and cerebellum of 5-month-old WT,* Hexa-/-, B4Galnt1-/-, DKO*, and TKO mice (n = 3) using TRIzol Reagent (Geneaid, Germany). cDNA synthesis was performed with a High-Capacity cDNA Reverse Transcription Kit (Applied Biosystems, USA). The relative expression levels of Ccl2, Ccl3, Ccl5, Cxcl10, Il1ß, Iba1, Gfap,* and Hexb* genes were quantified using Real-Time 480 SYBR Green I PCR master mix (Roche, Swiss) on a Roche LightCycler 96. PCR conditions included one cycle at 95 °C for 10 min, followed by 45 cycles at 95 °C for 20 s, 57 °C for 15 s, and 72 °C for 22 s. Expression levels were normalized to Gapdh, the internal control. Primer pairs for each gene are listed in Supplementary Table 1. Proteins from the same brain regions of these mice were isolated using RIPA buffer for Western blotting. Total protein concentration was measured using the Bradford assay, and equal amounts were loaded onto 10% SDS-PAGE gels, which were then transferred to nitrocellulose membranes (Bio-Rad, USA). Membranes were blocked in PBS-T with 5% non-fat dry milk for 1 h. Primary antibodies—anti-Cleaved Caspase-3 (ab231289, Abcam, USA), anti-HexB (sc-134581, Santa Cruz Biotechnology, USA), and anti-ß-Actin (4970 S, Cell Signaling Technology, USA)—were incubated overnight at + 4 °C. HRP-conjugated secondary antibodies (Jackson ImmunoResearch Lab, USA) were incubated for 1 h at room temperature. Detection was performed using a digital imaging system (Fusion SL, Vilber) with LuminataTM Forte Western HRP Substrate (Millipore, USA). Band intensities were normalized to ß-actin and quantified using NIH ImageJ.

Immunohistochemical Analyses and in situ Apoptosis Detection

Coronal brain sections from 5- and 18-month-old mice were prepared using a previously described protocol (Seyrantepe et al., 2018). Sections from these ages and genotypes were incubated in ice-cold acetone. Tissue was blocked in a humidified chamber at room temperature for 1 h with a solution containing 10% goat serum, 4% BSA, 0.3 M glycine, and 0.1% Triton X-100 in PBS. Primary antibodies—anti-GM2 (1:200; KM966), anti-GFAP (1:200; 12389, Cell Signaling Technology, USA), anti-Moma2 (1:50; ab33451, Abcam, USA), and anti-NeuN (1:50; 24307, Cell Signaling Technology, USA)—were diluted in blocking solution and incubated overnight at + 4 °C in a humidified chamber. Secondary antibodies conjugated with Alexa Fluor dyes were then applied for 1 h at room temperature: goat anti-human DyLight 488 (Thermo, USA) for anti-GM2, goat anti-rabbit Alexa Fluor 488 (Abcam, USA) for anti-GFAP, goat anti-rat Alexa Fluor 568 (Abcam, USA) for anti-Moma2, and goat anti-rabbit Alexa Fluor 568 (Abcam, USA) for anti-NeuN. Slides were mounted with Fluoroshield mounting medium containing DAPI (Abcam, USA). TUNEL analysis was conducted using the ApopTag Fluorescein In Situ Apoptosis Detection Kit (S7110, Millipore, Germany) following the manufacturer’s instructions. Briefly, coronal brain sections from 5-month-old mice were fixed in 1% paraformaldehyde in PBS and pre-treated with ethanol: acetic acid (2:1). The kit’s equilibration buffer was applied for 30 min at room temperature, followed by incubation with TdT enzyme at 37 °C for 1 h and with anti-digoxigenin conjugate at room temperature for 30 min. Nuclear counterstaining was performed with propidium iodide (0.5 µg/ml). Fluorescent images were captured using a BX53F fluorescence microscope (Olympus, Germany), and image analysis was performed with cellSens Entry software (Olympus). Fluorescence intensities were quantified using ImageJ.

Rotarod Test

5-month-old WT (n = 6), Hexa-/- (n = 5), B4Galnt1-/- (n = 7), DKO (n = 6), and TKO (n = 7) mice underwent the Rotarod test to evaluate locomotor activity and coordination. Before testing, mice had a brief 5-minute training session at 4 rpm. During the test, they walked on a rotating rod that gradually accelerated from 4 to 40 rpm over 5 min. The time to each mouse’s fall was recorded using Sedacom version 2 (Harvard Apparatus, USA). Each mouse performed three trials, with at least 15 min of rest between each trial. The mean duration spent on the rod across the three trials was calculated and plotted in the graph.

Footprint Analysis

Footprint analysis was conducted in mice of various ages and genotypes: 5-month-old WT (n = 4), DKO (n = 5), TKO (n = 5), as well as 10-month-old TKO (n = 3) and 15-month-old TKO (n = 3). Non-toxic, washable ink was applied to their fore and hind paws. The mice were placed in a 60 cm by 10 cm by 20 cm box and walked across white paper. Their walking patterns were recorded by scanning the paper. Stride length, sway length, and stance length were measured from the scans and displayed in the graphs.

Results

Extended lifespan and normal body appearance in TKO mice with the blocking of GM2 ganglioside accumulation

At birth, no phenotypic differences were observed among WT, Hexa-/-,* B4Galnt1-/-, DKO, and TKO mice. DKO mice maintained normal body weight until week 14, after which they began to lose weight rapidly. In contrast, B4Galnt1-/- and TKO mice had body weights similar to WT mice up to week 20 (Fig. 1A). TKO mice showed normal body weight and appearance comparable to WT controls. In contrast, DKO mice were noticeably smaller due to severe TSD pathology at 5 months (Fig. 1A, B). DKO mice died within 5 months due to excessive GM2 buildup. In contrast, TKO mice had a significantly longer lifespan, at least 18 months (Fig. 1C). Despite lacking complex ganglioside synthesis, B4Galnt1-deficient TKO mice were apparently normal at birth. Most TKO mice we obtained during the study lived for more than 18 months, whereas some died around 18 months for unknown reasons. Phenotypic findings in TKO mice that died around 18 months of age included tremor, ataxia, gait issues, posture abnormalities, hindlimb paresis, and motor problems. TLC analysis was performed on the cortex (Fig. 1D) and cerebellum (Fig. 1E) of 5-month-old WT, Hexa-/-, B4Galnt1-/-*, DKO, and TKO mice for acidic gangliosides. As shown in Fig. 1D and E, DKO mice had excessive GM2 ganglioside accumulation in both brain regions compared with age-matched controls, whereas GM2 was not detected in the cortex or cerebellum of B4Galnt1-/- and TKO mice, as expected. In addition to GM2, other complex gangliosides, such as GM1, GD1a, GD1b, and GT1b, were not detected in the cortex or cerebellum of B4Galnt1-/- and TKO mice. Gangliosides GM3 and GD3, precursors of complex gangliosides, were detected in both brain regions of B4Galnt1-/- and TKO mice. However, in both B4Galnt1-deficient mouse models, GM3 and GD3 ganglioside levels were significantly elevated in the cortex compared with DKO mice, whereas only GD3 was higher in the cerebellum. There was no difference in lactosylceramide levels between genotypes.

Fig. 1. Body weights/sizes measurements, survival paths, and TLC analysis. (A) Body weights of 5-month-old WT (n = 12), Hexa-/- (n = 10), B4Galnt1-/- (n = 14), DKO (n = 12), and TKO (n = 13) mice over 20 weeks are shown. Each symbol represents the mean body weight for mice of the corresponding genotype. (B) Gross appearances of 5-month-old WT, DKO, and TKO mice. (C) Kaplan-Meier survival path of WT, DKO, and TKO mice. Acidic sphingolipids were separated and stained from the cortex (D) and cerebellum (E) brain regions of 5-month-old WT, Hexa-/-, B4Galnt1-/-, DKO, and TKO mice by TLC analysis. Bands were characterized with red arrows. The data represent three independent experiments

Anti-GM2 staining, which immunohistochemically analyzed GM2 ganglioside levels in the cortex, cerebellum, and hippocampus of mouse brains (Fig. 2), also confirmed findings from TLC analysis. In the brain sections of 5-month-old DKO mice, GM2 ganglioside accumulation was strongly evident in the cortex, hippocampus, and cerebellum (Fig. 2D, I, and N, respectively).

There was limited GM2 ganglioside accumulation in the brain sections of Hexa-/- mice (Fig. 2B, G, and L). Due to the lack of complex ganglioside synthesis, no positive signal for GM2 ganglioside was observed in the brain sections of B4Galnt1-/- (Fig. 2C, H, and M) and TKO mice (Fig. 2E, J, and O).

Fig. 2GM2 ganglioside levels in brain sections of 5-month-old mice. Cortical, hippocampal, and cerebellar sections from 5-month-old WT (A, F, and K, respectively), Hexa-/- (B, G, and L, respectively), B4Galnt1-/- (C, H, and M, respectively), DKO (D, I, and N, respectively), and TKO (E, J, and U, respectively) mice were stained with anti-GM2 antibody (green) to detect GM2 ganglioside levels and with DAPI (blue). The scale bar is 50 μm. The data represent three independent experiments

Rescued expression of neuroinflammatory genes in TKO mice

Our previous research (Demir et al., 2020; Inci & Seyrantepe, 2025) showed that DKO mice exhibited elevated levels of pro-inflammatory chemokines, cytokines, and glial cell markers, all of which are associated with neuroinflammation. In this study, we analyzed the expression of neuroinflammation-related genes in the cortex and cerebellum of 5-month-old mice. The gene expression levels in Hexa-/-, B4Galnt1-/-,* DKO*, and TKO mice are shown as fold changes relative to WT mice.

Our results indicated that DKO mice exhibited significantly increased expression of the Ccl2, Ccl3, Ccl5, Cxcl10, Il1ß, Gfap, and Iba1 genes in both the cortex and cerebellum (Fig. 3). Conversely, gene expression levels in the brain regions of TKO mice were nearly identical to those of WT mice, except for Cxcl10 in both areas, Iba1 in the cortex, and Ccl5 in the cerebellum (Fig. 3). In DKO mice, Ccl2 expression was significantly increased, showing a 31.4-fold rise in the cortex and a 5.3-fold rise in the cerebellum (Fig. 3A and H, respectively). Its expression increased significantly in DKO mice, with a 17.2-fold rise in the cortex and a 29.4-fold rise in the cerebellum (see Fig. 3B and I, respectively). In DKO mice, Ccl5 expression was increased, with a 4.1-fold rise in the cortex and a 2.8-fold increase in the cerebellum (Fig. 3C and J). Upregulated expression of Cxcl10 (18.2-fold in the cortex, 1.3-fold in the cerebellum) (Fig. 3D and K, respectively) was observed in DKO mice. In the brains of DKO mice, a markedly higher expression of Il1ß was observed—1.7 times higher in the cortex and three times higher in the cerebellum (Fig. 3E and L, respectively). In contrast, a deficiency of the B4Galnt1 enzyme in TKO mice led to decreased expression levels of the genes compared to DKO mice; Ccl2 (15.6-fold in the cortex, 6.2-fold in the cerebellum) (Fig. 3A and H, respectively), Ccl3 (15.84-fold in the cortex, 20.45-fold in the cerebellum) (Fig. 3B and I, respectively), Ccl5 (4.3-fold in the cortex, 1.3-fold in the cerebellum) (Fig. 3C and J, respectively), Cxcl10 (5.6-fold in the cortex, 4.1-fold in the cerebellum) (Fig. 3D and K, respectively), and Il1ß (1.6-fold in the cortex, 2-fold in the cerebellum) (Fig. 3E and L, respectively).

The expression level of the Hexb gene in DKO mice was notably higher than that in age-matched littermates, particularly in the cortex and cerebellum of the brain. On the other hand, B4Galnt1-deficient TKO mice exhibited markedly reduced expression of Hexb in the brain regions compared with DKO mice (Supplementary Fig. 1A and 1B). Similarly, DKO mice showed significantly higher levels of the HexB enzyme in the cortex and cerebellum than other mice in the same age group. In contrast, the HexB protein level in TKO mice, compared with DKO mice, was slightly decreased in the cortex and significantly reduced in the cerebellum (Supplementary Fig. 1C and D, E and F, respectively).

Fig. 3. Neuroinflammation-related gene expression analysis in the brain regions of 5-month-old mice. Ccl2, Ccl3, Ccl5, Cxcl10, Il1ß, Gfap, and Iba1 gene expression levels in the cortex (A-G) and cerebellum (H-N) of 5-month-old WT, Hexa-/-,* B4Galnt1-/-*, DKO, and TKO mice were analyzed by qRT-PCR. Gene expression ratios relative to WT mice were calculated using the ΔCT method, and fold changes were reported. Data represent three independent experiments as the mean ± SEM. Statistical analyses were performed using One-way ANOVA (*p < 0.05, **p < 0.025, ***p < 0.01, and ****p < 0.001)

Reversal of activated astrogliosis and microgliosis observed in DKO mice as a consequence of B4Galnt1 deficiency

Expression of the Gfap gene was significantly increased in the cortex and cerebellum of DKO mice (39.3-fold in the cortex, 17.5-fold in the cerebellum); however, in B4Galnt1-deficient TKO mice, expression reverted to normal and was significantly downregulated compared with DKO mice (51.4-fold in the cortex, 16.8-fold in the cerebellum) (Fig. 3F and M, respectively). Immunohistochemical analysis revealed increased GFAP expression and elevated astrogliosis in the cortex (Fig. 4D), hippocampus (Fig. 4I), cerebellum (Fig. 4N), and pons (Fig. 4T) of DKO mice. In TKO mice, GFAP levels were significantly lower than in DKO mice across all brain regions (Fig. 4E, J, O, and U, respectively). The severe astrogliosis observed in DKO mice, resulting from excessive GM2 ganglioside accumulation, was absent in TKO mice due to the elimination of B4Galnt1 enzyme activity.

The microglial activation marker Iba1 (Wittekind et al., 2022) gene expression level in DKO mice was substantially upregulated in the cortex and cerebellum (2.9-fold in the cortex, 4.3-fold in the cerebellum). In contrast, Iba1 gene expression was markedly reduced in both brain regions of TKO mice compared with DKO mice (1.8-fold in the cortex, 4.2-fold in the cerebellum), nearly returning to the level observed in WT mice (Fig. 3G and N, respectively). Anti-Moma2 staining also revealed that the number of Moma2-positive cells in the cortex (Fig. 5D), hippocampus (Fig. 5I), cerebellum (Fig. 5N), and pons (Fig. 5T) of DKO mice was significantly higher. By contrast, the number of Moma2-positive cells in the brain regions of B4Galnt1-deficient TKO mice was considerably decreased (Fig. 5E, J, O, and U, respectively). These results suggest that GM2 ganglioside accumulation drives the activated microglia/macrophage network in DKO mice; conversely, activation of microglia/macrophages was significantly repressed in TKO mice due to the elimination of the B4Galnt1 gene.

Fig. 4. Immunofluorescence analysis of astrocytes in brain sections from 5-month-old mice. Cortex, hippocampus, cerebellum, and pons sections from 5-month-old WT (A,** F**,** K**, and P, respectively), Hexa-/- (B,** G**,** L**, and R, respectively), B4Galnt1-/- (C,** H**,** M**, and S, respectively), DKO (D,** I**,** N**, and T, respectively), and TKO (E,** J**,** O**, and U, respectively) mice were stained with an anti-GFAP antibody (green) to detect reactive astrocyte levels and with DAPI (blue). Histograms show GFAP intensity measurements in the cortex (V), hippocampus (W), cerebellum (X), and pons (Z). The scale bar is 50 μm. Data represent three independent experiments, with mean ± SEM. Statistical analyses were performed using One-way ANOVA (*p < 0.05, **p < 0.025, ***p < 0.01, and ****p < 0.001)

Fig. 5. Immunofluorescence analysis of activated microglia. Sections of the cortex (A,** F**,** K**, and P), hippocampus (B,** G**,** L**, and R), cerebellum (C,** H**,** M**, and S), and pons (D,** I**,** N**, and T) from 5-month-old WT, Hexa-/-,* B4Galnt1*-/-, DKO, and TKO mice were stained with an anti-Moma2 (red) antibody to detect active microgliosis and with DAPI (blue). Histograms show Moma2 intensity in the cortex (V), hippocampus (W), cerebellum (X), and pons (Z). The scale bar is 50 μm. Data represent three independent experiments, with mean ± SEM. Statistical analyses were performed using One-way ANOVA (*p < 0.05, **p < 0.025, ***p < 0.01, and ****p < 0.001)

Delayed neuronal loss with no signs of apoptosis in TKO mice

As shown in Figs. 5 and 6-month-old DKO mice exhibit neuronal death in the cortex, hippocampus, thalamus, and cerebellum (Fig. 6B, G, L, and R, respectively) compared with age-matched WT mice. In addition, the neural density of 5-month-old TKO mice was significantly higher in the cortex and thalamus (Fig. 6C and M, respectively) and slightly higher in the hippocampus and cerebellum (Fig. 6H and S, respectively) than in DKO mice. Compared with WT mice, 5-month-old TKO mice had fewer NeuN-positive cells in each brain region, particularly in the cerebellum. However, at 18 months of age, TKO mice showed significant neuronal loss, except in the hippocampus, compared with age-matched WT mice, and the neuronal density of 18-month-old TKO mice closely resembled that of 5-month-old DKO mice in all analyzed brain regions.

TUNEL staining was performed to measure apoptotic cell death in the cortex, hippocampus, and cerebellum of 5-month-old mice (Supplementary Fig. 2). As shown in our previous study (Seyrantepe et al., 2018), TUNEL-positive signals were significantly higher in the cortex, hippocampus, and cerebellum of 5-month-old DKO mice. In contrast, TUNEL-positive signals were not observed in B4Galnt1-/- and TKO mice, both of which lacked B4Galnt1 enzyme activity. TUNEL analysis revealed robust apoptotic cell death in multiple brain regions of 5-month-old DKO mice. In contrast, TKO mice exhibited no evidence of apoptosis (Supplementary Fig. 2). These findings suggest that eliminating B4Galnt1 activity in DKO mice prevents GM2 ganglioside accumulation-induced programmed cell death.

Western blot analyses were performed to detect the active/cleaved form of caspase-3 in the cortex and cerebellum of 5-month-old WT, Hexa-/-, B4Galnt1-/-, DKO, TKO, and 18-month-old WT and TKO mice (Supplementary Fig. 3). In both brain regions, the cleaved caspase-3 protein abundance in 5-month-old DKO mice was significantly higher than in age-matched littermates. Cleaved caspase-3 protein levels in 5-month-old TKO mice were similar to those in WT mice. Surprisingly, the cleaved caspase-3 level in the cortex of 18-month-old TKO mice was significantly lower than in 18-month-old WT and 5-month-old TKO mice (Supplementary Fig. 3). Our results align with the TUNEL findings, showing that high cleaved caspase-3 levels in 5-month-old DKO mice are normalized when B4Galnt1 activity is eliminated, which suppresses apoptosis in this age group.

Fig. 6. Density of neuronal cells in brain sections from 5-month- and 18-month-old mice. Cortex, hippocampus, thalamus, and cerebellum sections from 5-month-old WT (A, F, K, and P, respectively), 5-month-old DKO (B, G, L, and R, respectively), 5-month-old TKO (C, H, M, and S, respectively), 18-month-old WT (D, I, N, and T, respectively), and 18-month-old TKO (E, J, O, and U, respectively) mice were immunostained with an anti-NeuN antibody (red) to detect neuronal loss. Histograms show NeuN intensity measurements in the cortex (V), hippocampus (W), thalamus (X), and cerebellum (Z). The scale bar is 50 μm. Data represent three independent experiments, with mean ± SEM. Statistical analyses were performed using One-way ANOVA (*p < 0.05, **p < 0.025, ***p < 0.01, and ****p < 0.001)

Early maintenance and late impairment of gait and motor function in TKO mice

We performed a footprint test in 5-month-old WT, DKO, and TKO mice, as well as in 10- and 15-month-old TKO mice, to evaluate how B4Galnt1 enzyme deficiency affects gait patterns (Supplementary Fig. 4A, B, C, and D). DKO mice took longer to complete the platform than age-matched littermates. Additionally, DKO mice had significantly shorter strides, sway, and stance (Supplementary Fig. 4A, B, C, and D). As previously reported (Seyrantepe et al., 2018), paw prints indicate that the neurons controlling the hind legs are affected in these mice. In contrast, 5-month-old TKO mice took as long to complete the platform as WT mice, and there were no significant alterations in stride, sway, or stance length in these mice (Supplementary Fig. 4A, B, C, and D). To assess whether there was a time-dependent change in the gait patterns of TKO mice, 10- and 15-month-old mice were also tested. The time it took for TKO mice to complete the platform was slightly longer in 10-month-old mice and significantly longer in 15-month-old mice. With increasing age, a slight increase in step parameters was observed in late-age TKO mice. The Rotarod test was performed to assess balance and motor coordination in 5-month-old WT, Hexa-/-,* B4Galnt1-/-*, DKO, TKO (Supplementary Fig. 4E), and 18-month-old WT and TKO mice. DKO mice spent significantly less time on the accelerating rod than their counterparts. In contrast, B4Galnt1-deficient B4Galnt1-/- and TKO mice performed almost the same as WT mice on the rod. We observed dramatic impairment in the motor function of 18-month-old TKO mice, which were virtually unable to stand on even a constant-speed rod due to severe weakness and paralysis in their hind limbs. In contrast, age-matched WT mice performed normally on the accelerating rod.

Discussion

DKO mice exhibit the clinical and neuropathological signs of TSD and die at 5 months of age from severe neurodegeneration and neuroinflammation caused by abnormal accumulation of GM2 ganglioside (Seyrantepe et al., 2018; Demir et al., 2020). In this study, we bred DKO mice with B4Galnt1-/- mice to create Hexa-/-Neu3-/-B4Galnt1-/- triple-knockout (TKO) mice, aiming to block GM2 synthesis and reduce TSD symptoms.

Substrate reduction therapy (SRT) aims to prevent substrate accumulation by inhibiting its biosynthesis through targeting the enzymes responsible for its synthesis (Coutinho et al., 2016). Glucosylceramide synthase (GCS) inhibitors, such as Miglustat, which is approved for Gaucher disease and NPC, decrease glucosylceramide synthesis and consequently reduce glycosphingolipid production (Canini et al., 2025). Administration of the specific GCS inhibitor Genz-682,452 to the SD mouse model (Hexb-/-) reduced GM2 ganglioside levels, diminished microgliosis, delayed the onset of motor function deficits, and prolonged lifespan (Marshall et al., 2019). Genistein, an isoflavone, has demonstrated efficacy in inhibiting glycosaminoglycan (GAG) synthesis in various fibroblasts derived from MPS patients, thereby reducing GAG levels in these cells (Piotrowska et al., 2006). Furthermore, it has been reported that Genistein treatment significantly reduces GAG levels in the urine and tissues of MPS II mice, likely due to its ability to cross the blood-brain barrier (Friso et al., 2010). Abidi et al. (2025) characterized recombinantly produced human B4GALNT1 in vitro and tested the inhibitory effects of various bisimidazolium (QT) compounds on the purified enzyme. The QT compounds inhibited human B4GALNT1 activity by 40–94%, supporting their potential as promising leads for the development of therapeutic options targeting B4GALNT1 in GM2 gangliosidosis.

In TKO mice, B4Galnt1 deficiency prevented GM2 accumulation and significantly increased lifespan to 18 months. TLC analysis of brain tissue showed excessive GM2 in DKO mice; however, GM2 and other complex gangliosides were absent in TKO mice, with only GM3 and GD3 remaining. An extra band between GM1 and GM2 was detected in B4Galnt1-deficient mice and identified as the O-acetyl GD3 ganglioside (Liu et al., 2000). Anti-GM2 staining of brain regions confirmed these results: high GM2 levels in the cortex, hippocampus, and cerebellum in DKO mice, but no GM2 in B4Galnt1-/- and TKO mice.

The accumulation of GM1 and GM2 gangliosides activated microglia and astrocytes, resulting in a notable rise in inflammatory cytokines in Hexb-/- mice, a model for GM1 gangliosidosis (ßgal-/-), and in patients with SD (Jeyakumar et al., 2003; Wada et al., 2000). Similarly, we observed that pro-inflammatory cytokine gene expression in the brains of DKO mice was markedly higher than in age-matched controls in our previous study. Ccl2 is a chemokine that facilitates macrophage and monocyte infiltration into the CNS by binding to its receptor, Ccr2, during neuroinflammation. It is produced by various CNS-resident cells, including neurons and glial cells (Joly-Amado et al., 2020; Cherry et al., 2020). Ccl3, another pro-inflammatory chemokine, plays a role in neuroinflammation by affecting macrophages and astrocytes through interactions with Ccl2 (Pelisch et al., 2020; Reichel et al., 2009). Astrocytes and microglia constantly produce Ccl5 and interact with its receptor Ccr5 in the CNS, where Ccl5 also functions in chemotaxis. Multiple studies indicate that the Ccl5/Ccr5 axis is involved in inflammatory and neurodegenerative conditions (An et al., 2025). C-X-C motif chemokine ligand 10 (Cxcl10) is a pro-inflammatory cytokine produced by CNS-resident cells and has chemotactic effects, enhancing the migration of inflammatory cells (Bufi et al., 2025; Zhang et al., 2014). Il1ß is a pro-inflammatory cytokine that has been demonstrated to play a role in the inflammatory response and worsen the severity of various CNS diseases, including neurodegenerative disorders (Mendiola & Cardona, 2018). Although Ccl5 expression in the cerebellums of B4Galnt1-/- and TKO mice did not differ significantly from that in DKO mice, it was still notably higher than in WT mice. This suggests that the altered ganglioside profile resulting from B4Galnt1 enzyme deficiency may play a role. In TKO mice, pro-inflammatory gene levels were significantly reduced in both regions compared with DKO mice. Additionally, apart from Cxcl10 and Iba1 expression in the cortex and Ccl5 in the cerebellum, gene expression levels in these areas were similar to those in WT mice. Our findings indicate that the increased expression of neuroinflammation-related genes observed in DKO mice, associated with excessive GM2 ganglioside accumulation in lysosomes, is restored primarily in TKO mice lacking the B4Galnt1 gene.

We showed that DKO mice expressed higher levels of the Gfap gene than their age-matched counterparts in both the cortex and the cerebellum. Additionally, anti-GFAP staining demonstrated severe astrogliosis in DKO mice. In contrast, Gfap gene expression was significantly reduced in the cortex and cerebellum of TKO mice, approaching WT levels. Expression of Iba1 was markedly elevated in the cortex and cerebellum of DKO mice. As a parallel, the number of Moma2-positive cells, a marker of activated microglia/macrophages (Ohmi et al., 2003), was significantly increased in the brains of DKO mice. Reactive astrocyte levels in TKO mice were abolished in the cortex, hippocampus, cerebellum, and pons. TKO mice exhibited significantly lower Iba1 expression than DKO mice in both brain regions. Anti-Moma2 staining revealed that Moma2-positive cells were greatly diminished and normalized in the brain regions of TKO mice. These results also suggest that removing the B4Galnt1 gene in DKO mice alleviated the severe astrogliosis and microgliosis.

The number of NeuN-positive cells in 5-month-old TKO mice was notably higher in the cortex and thalamus. There was also a slight increase in these cells in the hippocampus and cerebellum compared with DKO mice. Remarkably, the neuronal density in the cerebellum of 5-month-old TKO mice was significantly lower than that of WT mice. In the older age group (18 months), we observed a significant decrease in neuronal density in TKO mice, except in the hippocampus, compared with both age-matched WT mice and 5-month-old TKO mice. These findings suggest that the mortality observed in a subset of 18-month-old TKO mice is likely linked to substantial neuronal loss, eventually leading to neurodegeneration. Our results show that removing B4Galnt1 enzyme activity in DKO mice considerably slowed neural death.

The TUNEL assay, performed to analyze programmed cell death, showed TUNEL-positive signals in 5-month-old DKO mice, while no evidence of apoptosis was found in TKO mice of the same age group. In parallel, our immunoblot analyses of Caspase-3 also confirmed the TUNEL findings, showing a significant increase in active/cleaved caspase-3 levels in 5-month-old DKO mice; in contrast, no difference was observed in age-matched TKO mice compared to WT mice. Although markedly elevated neuronal loss was revealed in 18-month-old TKO mice, surprisingly, no significant increase in active caspase-3 levels was detected at this age. Overall, these findings imply that the apparent neuronal death observed in TKO mice at advanced ages may not be primarily driven by canonical caspase-3-dependent apoptosis. Instead, alternative regulated cell death pathways, such as necroptosis and pyroptosis, may contribute to reduced neuronal density.

Gait disturbances have been reported in various LSDs, including sialidosis type I (Khan & Sergi, 2018), aspartylglucosaminuria (Goodspeed et al., 2021), and Tay-Sachs (González-Sánchez et al., 2025). In the footprint test assessing gait abnormalities, 5-month-old DKO mice displayed a slow gait with shortened stride, sway, and stance lengths, whereas 5-month-old TKO mice did not show these gait issues. Although TKO mice also showed a slower gait at later ages, their step lengths increased with age, likely due to their larger body size at 5 months. The rotarod test revealed impaired motor coordination and balance in 5-month-old DKO mice, whereas TKO mice of the same age exhibited no motor impairment. However, by 18 months, TKO mice experienced significant declines in motor function and balance. Chiavegatto et al. (2000) found that B4Galnt1-deficient mice displayed motor coordination deficits, balance problems, and a slower gait starting at 8 months. We suggest that severe cerebellar neuronal death in aging TKO mice may contribute to these motor and gait issues. Overall, the motor and gait disturbances seen in 5-month-old DKO mice were notably delayed in TKO mice.

In summary, we created B4Galnt1-deficient DKO mice, called TKO mice. These mice exhibited normal overall appearance and development, better motor skills, significantly decreased levels of pro-inflammatory chemokines and cytokines, and reduced astrogliosis and microgliosis. Additionally, removing B4Galnt1 in DKO mice slowed neuronal death. Our results indicate that silencing B4Galnt1 may reduce TSD symptoms and pathology, making it a promising, potentially safe target for SRT treatments, not only in DKO mice but also in TSD patients. Disrupted ganglioside metabolism is associated with cognitive deficits, developmental delays, and neurological conditions like HSP. While complex and straightforward gangliosides are vital for neuronal growth and mature function, targeted gene silencing of their biosynthesis may offer therapeutic benefits for TSD patients.

Supplementary Information

Below is the link to the electronic supplementary material.

Supplementary Material 1

Supplementary Material 2

Supplementary Material 3

Supplementary Material 4

Supplementary Material 5