Homozygous Glb1 ^em1/RASE-/-Tg(LacZ)C57BL/6 (Glb1^-/- ) mice and C57BL/6 wildtype (Glb1^+/+ ) littermates were generated by breeding of heterozygous animals from a previously established population (21). Animals were housed in individually ventilated cages (Tecniplast Deutschland GmbH, Hohenpeißenberg, Germany) with 12 h light and 12 h darkness at 22 - 24°C and 50 - 60% humidity. Food for maintenance and breeding (ssniff Spezialdiäten GmbH, Soest, Germany) as well as water were provided ad libitum. Enrichment of the cages included mouse houses (Tecniplast Deutschland GmbH) and nesting material (ssniff Spezialdiäten GmbH).

DNA was extracted from ear notch tissue taken from up to three-week-old mice. Standard genotyping of the mice was performed using PCR with the forward primer Glb1 Primer 0 fwd (5’-CTG TTG GCT TGA GAC CAG TGT AGT C-3’) binding in intron 14 and the reverse primer Glb1 Primer 0 rev (5’-GAT GCA TAC CTT GGA CCA CCC AG-3’) binding in exon 15 of the Glb1 gene (21).

The study was approved by the Local Institutional Animal Care and Research Advisory committee and permitted by the appropriate authority (LAVES, Oldenburg, Germany, permission number: 33.8-42502-04-19/3204).

C57BL/6 Glb1^+/+ and homozygous Glb1^-/- mice were intracerebrally infected with 1x10⁵ plaque forming units (PFU) of TMEV-BeAn at 5 weeks of age under general anaesthesia. An equal number of control Glb1^-/- and Glb1^+/+ mice was intracerebrally injected with virus-free cell solution (mock/placebo). General anaesthesia was performed via intraperitoneal injection of Medetomidin (Domitor^®, Pfizer 1.0 mg/ml) and Ketamin (Ketamin 10%, WDT, 100 mg/ml) diluted in NaCl at a dosage of 0.5 mg/kg Medetomidin and 100 mg/kg Ketamin. Starting at 0 dpi, mice were examined weekly for general appearance and posture, behaviour and activity as well as gait. Further examination included measuring of the body weight, parachute test (50), grid-walking test (50), hang test (51) and RotaRod performance test (RotaRod Treadmill, TSE Technical & Scientific Equipment, Bad Homburg, Germany) as described before (52). Mice were placed on a rotating rod, with the rotation speed steadily accelerating from 5 rounds per minute (rpm) to 30 rpm over the course of 300 seconds. Each mouse was tested three times, and the highest rod turn rate (rounds per minute (rpm)), that was tolerated at each trial before mice fell off, was recorded and averaged. All clinical and neurological tests performed are described in detail in Supplementary Table 1 . For euthanasia, mice were first put in general anaesthesia as described above and then humanely killed via a second, undiluted intraperitoneal injection of Medetomidin (Domitor^®, Pfizer 1.0 mg/ml) and Ketamin (Ketamin 10%, WDT, 100 mg/ml) at a dosage of 1 mg/kg Medetomidin and 200 mg/kg Ketamin. After death, mice were perfused with phosphate buffered saline (PBS). Following euthanasia at 4, 7, 14 and 98 dpi, respectively, tissue samples were taken from brain, spinal cord and spleen. Samples were fixed overnight in 10% neutral buffered formalin, decalcified with EDTA (only spinal cord with adjacent vertebrae) and embedded in paraffin. Additional organ samples were embedded in OCT medium, frozen in liquid nitrogen and stored at -80°C. Tissue samples from both brain (front left quarter of cerebrum) and spleen were taken and stored in Iscove’s Modified Dulbecco’s Medium (IMDM) supplemented with 10% FCS, 2 mM L-glutamine, and 100 U/mL penicillin/streptomycin until further flow cytometry analysis. All analyses performed in the study were performed once (number of repeat experiments = 1).

Histology and immunohistochemistry were performed on paraffin sections applying the avidin-biotin-peroxidase complex method (Vector Laboratories Inc., Burlingame, CA, USA) with the chromogen 3,3’-diaminobenzidine-tetrahydrochloride (DAB) as previously described (53–57). Primary antibodies directed against glial fibrillary acidic protein (GFAP), amyloid precursor protein (APP), TMEV, CD3, Iba1 and CD107b (Mac3) were used to detect astrocytes, axonal lesions, virus, T cells, microglia/macrophages and activated microglia/macrophages cells, respectively ( Table 1 ).

Haematoxylin and Eosin (H&E) stains were evaluated for perivascular cuffing, i.e. layered perivascular accumulation of infiltrating leukocytes using a semiquantitative scoring system going from 0 to 3 (0 – normal, 1 = single perivascular infiltrates, 2 = 2-3 layers of perivascular infiltrates, 3 = >3 layers of perivascular infiltrates). Cortex, hippocampus, diencephalon/mesencephalon, cerebellum and medulla oblongata were scored separately, and the scores were combined into an average brain score.

Immunostainings were analysed by counting the number of APP positive spheroids, CD3 positive T cells, and TMEV positive cells in cortex, hippocampus, diencephalon/mesencephalon, cerebellum and medulla oblongata (regions of interest (ROI)) in the brain, resulting in a cumulative, total number of positive cells per brain section. Stainings for GFAP, Iba1 and CD107b were evaluated by measuring the relative positive brain area through the use of the QuPath software (QuPath 0.5) (58). Perivascular infiltration of Iba1-positive macrophages at 4 and 7 dpi was investigated semiquantitatively analogous to the scoring system described above for perivascular cuffing on H&E sections.

RNA was isolated from frozen brain tissue using the RNeasy^® Lipid Tissue Mini Kit (Qiagen, Hilden, Germany) following the manufacturers’ instructions. The Omniscript RT Kit (Qiagen), Random Primers (Promega, Mannheim, Germany) and RNase Out (Invitrogen, Darmstadt, Germany) were used for the generation of cDNA. RT-qPCR was performed using the AriaMx Real-time PCR System (Agilent Technologies Deutschland GmbH) and the Brilliant III Ultra-Fast SYBR^®QPCR Master Mix as described (59). The following primers were used: TMEV forward (GAC TAA TCA GAG GAA CGT CAG C), TMEV reverse (GTG AAG AGC GGC AAG TGA GA), IFN-γ forward (CAC GGC ACA GTC ATT GAA AG), IFN-γ reverse (AAT CTG GCT CTG CAG GAT TT), GAPDH forward (GAG GCC GGT GCT GAG TAT GT) and GAPDH reverse (GGT GGC AGT GAT GGC ATG GA). Tenfold serial dilution standards ranging from 10⁸ to 10² copies/μL were used for quantification and the specificity of each reaction was controlled by melting curve analysis. GAPDH was used as housekeeping gene to normalize the results.

Plaque assay was performed as described (60). Briefly, a frozen section of cerebral tissue was weighted, diluted in DMEM to a concentration of 10%, and homogenized using an Omni Tissue Homogenizer (Süd-Laborbedarf GmbH). Then, serial dilutions of homogenates were added to 6-well culture plates (Merck KGaA, Darmstadt, Germany) containing a confluent layer of L cells for 1 hour at room temperature. Subsequently, cells were covered with methyl cellulose (Merck) and incubated for 72 hours at 37°C. Finally, cells were fixed with 10% buffered formalin and stained with crystal violet (Merck) to allow counting of plaque forming units (PFU).

Spleens and brains were dissected from TMEV- or mock-infected C57BL/6 Glb1^+/+ and homozygous Glb1^-/- mice and stored in IMDM complete (IMDM, 10% FCS, 2 mM L-glu, 100 U/ml penicillin, 100 µg/ml streptomycin) until further processing. For processing, brain isolates (left front quarter of cerebrum) were transferred to a petri dish filled with IMDM complete and dissociated using the back of a 25 ml syringe (Roth, #0058.1). For further homogenization, cell dissociates were re-suspended several times. All isolated cells were filtrated using a cell strainer (40 µm) to obtain a uniform single-cell suspension. The filtered cells were centrifuged (300 x g, 5 min, 4°C), the supernatant was discarded and RBCs in the pellet were lysed in lysis buffer (10% 100 mM Tris pH 7.5, 90% 160 mM NH₄Cl) for 5 min at room temperature (RT). Cells were centrifuged twice with one PBS-washing step in between and the resulting pelleted cells were resuspended in IMDM complete until flow cytometrical staining. For flow cytometrical staining, all cells were blocked with anti-mouse CD16/32 (93, eBioscience) for 10 min at 4°C. Afterwards, splenocytes and brain cells were stained for composition and activation state. To analyse splenocyte composition, cells were either stained with PE-conjugated anti-mouse CD4 (GK1.5, BD Biosciences, Franklin Lakes, NJ), APC-conjugated anti-mouse CD8a (53-6.7, BD Biosciences) and FITC-conjugated anti-mouse CD19 (1D3, eBioscience, Frankfurt am Main, Germany) or APC-conjugated anti-mouse CD11c (N418, eBioscience), PE-conjugated anti-mouse CD11b (M1/70, BD Biosciences) and FITC-conjugated anti-mouse CD19 (1D3, eBioscience, Frankfurt am Main, Germany) for 20 min at 4°C. To analyse brain cell composition, cells were stained with PerCP-Cy5.5-conjugated anti-mouse CD45 (30-F11, eBioscience, Frankfurt am Main, Germany), APC-conjugated anti-mouse CD11c (N418, eBioscience) and PE-conjugated anti-mouse CD11b (M1/70, BD Biosciences) for 20 min at 4°C. To analyse the activation of CD4 and CD8a positive T cells in spleen and brain, cells were either stained with FITC-conjugated anti-mouse CD4 (RM4-5, eBioscience), PE-Cy7-conjugated anti-mouse CD62L (MEL-14, eBioscience) and APC-conjugated anti-mouse CD69 (H1.2F3, eBioscience) or FITC-conjugated anti-mouse CD8a (53-6.7, eBioscience), PE-Cy7-conjugated anti-mouse CD62L (MEL-14, eBioscience) and APC-conjugated anti-mouse CD69 (H1.2F3, eBioscience), respectively, for 20 min at 4°C. All spleen and brain samples were analysed using an Attune NxT Flow Cytometer (Thermo Fisher Scientific, Waltham, MA). The gating strategy is illustrated in the Supplementary Figure 1 . Data analysis was performed using FlowJo (Version 10, FlowJo LLC., Ashland, OR). Due to variability of sample size, cell numbers were normalized to 10⁷ analyzed cells.

Statistical analysis was performed using SAS/STAT software, Version 7.1 of the SAS^® System for Windows (SAS Institute Inc.). Graph plotting was performed using GraphPad Prism 9. Significance values for clinical, histological and immunohistochemical results, except for TMEV, were compared via Wilcoxon Tests with Dwass-Steel-Critchlow-Fligner method as post-hoc tests. Similarly, RT-qPCR and plaque assay data were analysed using Wilcoxon Tests. Both the immunohistochemical results for TMEV as well as the flow cytometry-based assays were compared via Mann-Whitney-U-Test including Bonferroni-correction for the latter. Figures for clinical, histological and immunohistochemical data were generated using GIMP 2.10.32.

LSDs can affect immune responses, even before the onset of clinical disease. Thus, antiviral immune responses might be restricted in Glb1 ^-/- mice making them susceptible to TMEV-IDD. To investigate changes in the susceptibility of Glb1 ^-/- mice to TMEV-IDD, we infected Glb1^-/- and Glb1^+/+ control mice on a C57BL/6 background with the BeAn strain of TMEV.

Immunohistochemistry was used to localize and quantify TMEV antigen in the brain. Moreover, virus load in the brain was quantified via RT-qPCR and plaque assay. At 4 dpi, a slightly higher number of TMEV-positive cells was found in Glb1^+/+ compared to Glb1 ^-/- mice via immunohistochemistry, but generally only few cells contained viral antigen ( Figure 1A1-A3 ). RT-qPCR and plaque assay showed a similar trend towards a higher virus load in Glb1^+/+ mice at 4 dpi but a statistically significant difference was lacking ( Supplementary Figure 2 ). At 7 dpi, Glb1 ^-/- mice showed more TMEV-positive cells than Glb1^+/+ mice via immunohistochemistry, which is consistent with a retarded viral clearance ( Figure 1B1-B3 ). RT-qPCR and plaque assay also demonstrated a trend for a higher viral load in Glb1^-/- mice at 7 dpi but no statistically significant difference was found ( Supplementary Figure 2 ). At 14 and 98 dpi, neither Glb1 ^-/- nor Glb1^+/+ mice displayed TMEV-positive cells in the brain ( Figure 1C1-D3 ). However, RT-qPCR still detected TMEV RNA in two Glb1 ^-/- mice at 14 dpi ( Supplementary Figure 2 ).

Next, to investigate the impact of G_M1 gangliosidosis on inflammatory processes in the brain, perivascular cuffing was quantified using a semiquantitative scoring system. As expected, at 4 dpi, stronger perivascular cuffing was found in TMEV- compared to mock-infected Glb1^+/+ mice. Interestingly, TMEV-infected Glb1^+/+ mice also exhibited a stronger perivascular cuffing than TMEV-infected Glb1 ^-/- mice ( Figure 2A1-A5 ). In contrast, at 7 dpi, TMEV-infected Glb1 ^-/- mice displayed a stronger perivascular cuffing than mock-infected Glb1 ^-/- mice and TMEV-infected Glb1^+/+ mice ( Figure 2 B1-B5 ). At 14 dpi, a stronger perivascular cuffing was present in TMEV-compared to mock-infected Glb1 ^-/- mice, whereas no significant difference was found between TMEV-infected Glb1^+/+ and Glb1 ^-/- mice ( Figure 2C1-C5 ). Cortex, hippocampus and diencephalon/mesencephalon (close to the injection site) were most strongly affected by perivascular cuffing. Finally, perivascular cuffing was low or even absent in mice at 98 dpi without statistically significant differences between the groups ( Figure 2D1-D5 ). This suggests that Glb1^-/- mice mount a delayed inflammatory response to TMEV-infection, which is still present at 14 dpi in the brain. Corresponding to the lack of increased clinical disease in TMEV-infected animals, no inflammatory or demyelinating lesions were present in the spinal cord of TMEV- and mock-infected Glb1 ^-/- and Glb1^+/+ mice at 98 dpi.

T cell infiltration into the CNS was further assessed by quantifying CD3⁺ cells. At 4 dpi, TMEV-infected Glb1^+/+ mice displayed a higher number of CD3⁺ T cells than TMEV-infected Glb1^-/- mice. ( Figure 3A1-A5 ). At 7 dpi, TMEV infection resulted in an increase in T cell numbers in both Glb1^+/+ and Glb1 ^-/- mice, which remained at a higher level in Glb1 ^-/- compared to Glb1^+/+ mice at 14 and 98 dpi ( Figure 3B1-D5 ). At 98 dpi, only Glb1^-/- but not Glb1^+/+ mice showed an increased number of CD3⁺ T cells after TMEV infection ( Figure 3D1-D5 ). From 7 dpi until 98 dpi, mock-infected Glb1^-/- mice displayed increased T cell numbers in the brain compared to mock-infected Glb1^+/+ mice ( Figure 3B1-D5 ). Overall, these data indicate a delayed antiviral T cell response in Glb1 ^-/- compared to Glb1^+/+ mice after TMEV infection but a slightly enhanced T cell infiltration into the CNS due to G_M1 lysosomal storage in older animals. Additionally, to evaluate lymphocyte activity (especially CD8⁺ T cells, but also NK cells), IFN-γ levels were measured at 4, 7 and 14 dpi. Similar to the immunohistochemical results ( Figure 3 ), TMEV-infected Glb1^+/+ mice displayed higher levels of IFN-γ than TMEV-infected Glb1^-/- mice at 4 dpi ( Figure 4 ). At 7 and 14 dpi, TMEV-infected Glb1^+/+ mice and TMEV-infected Glb1^-/- mice displayed equally higher levels of IFN-γ than their respective mock-infected control groups ( Figure 4 ).

Subsequently, flow cytometry was used to assess both the composition and activation of CD4⁺ and CD8⁺ T cells in the brain during TMEV infection. In general, the most prominent effects were observed at 4 and 7 dpi. At 4 dpi, TMEV-infected Glb1^+/+ mice already exhibited an increased number of CD4⁺ and CD8⁺ T cells ( Figure 5A ). Only starting at 7 dpi, these changes were observed for cells from Glb1^-/- mice, while both Glb1^+/+ and Glb1^-/- mice showed robust T cell activation as indicated by the down-regulated CD62L expression and increased CD69 expression by both CD4⁺ and CD8⁺ T cells ( Figure 5B ). From 14 dpi on, the number and activation of brain-infiltrated T cells started to decline ( Figure 5 ). This decline was more prominent in Glb1^+/+ mice, which again indicates a protracted response in Glb1^-/- mice as seen in immunohistochemistry. At 98 dpi, immune cell composition in the brain and their activation had largely returned to basal levels, comparable to mock-infected Glb1^+/+ and Glb1^-/- mice ( Figure 5 ). In the spleen, there was a tendency (sometimes significant) for a higher frequency of CD4⁺ in the Glb1^+/+ mice, which could reflect a larger T cell repertoire in comparison to the knockout mice ( Supplementary Figures 3, 4 ).

To quantify the expansion of microglia/macrophages in the brain, we stained for Iba1 and determined the immunopositive area using morphometry. Furthermore, the perivascular infiltration of macrophages in the early phase of the disease (4 and 7 dpi) was semiquantitatively scored on sections stained for Iba1. At 4, 7 and 14 dpi, there were no statistically significant differences in the relative Iba1-positive area between the study groups ( Figure 6A1-C5 ). In contrast, at 98 dpi, TMEV-infected Glb1 ^-/- mice displayed an increase in the Iba1-positive area compared to TMEV-infected Glb1^+/+ mice indicating a long-lasting effect of TMEV infection on the number and/or size of microglia/macrophages in the brain (maybe due to a transition from a ramified to a hyperramified or amoeboid state of microglial cells), which seems to be enhanced by G_M1 lysosomal storage ( Figure 6C1-D5 ). Moreover, the evaluation of perivascular Iba1-positive cells at 4 and 7 dpi showed a delayed infiltration of macrophages into the CNS of TMEV-infected Glb1^-/- mice compared to TMEV-infected Glb1^+/+ mice similar to the delayed infiltration of T cells described above ( Supplementary Figure 5 ).

Flow cytometry was also used to assess the numbers of antigen-presenting cells in the brain during TMEV infection. Only minor changes were found for the absolute numbers of CD45^intCD11b⁺ cells (predominantly microglia). Similar to T cells, increased numbers of CD45^highCD11b⁺ (most likely macrophage but also natural killer (NK) cell) and CD45^highCD11c⁺ (most likely dendritic cell) populations were found in TMEV-infected Glb1^+/+ mice already at 4 dpi and in Glb1^-/- mice not until 7 dpi ( Supplementary Figure 6 ). At later time points, the total number of CD45^highCD11b⁺ and CD45^highCD11c⁺ cells in the brain returned to basal levels comparable to mock-infected Glb1^+/+ and Glb1^-/- mice ( Supplementary Figure 6 ).

Next, we investigated the microglia/macrophage activation status in the brain during TMEV infection using immunohistochemistry for CD107b (Mac-3). At 4 dpi, there were no statistically significant differences in the relative CD107b⁺ area between the study groups ( Figure 7A1-A5 ). At 7 and 14 dpi, TMEV- and mock-infected Glb1 ^-/- mice displayed a significant increase in CD107b⁺ area compared to TMEV- and mock-infected Glb1^+/+ mice, respectively ( Figure 7B1-C5 ). At 98 dpi, CD107b⁺ brain area was similarly low in all study groups ( Figure 7D1-D5 ). Consequently, G_M1 lysosomal storage seems to amplify microglia/macrophage activation transiently in the brain during TMEV infection. Moreover, increased microglia/macrophage activation in mock-infected Glb1 ^-/- mice might be related to an unspecific and maybe dysregulated reaction of microglia/macrophages caused by G_M1 lysosomal storage.

In addition to microglia/macrophage number and activation, the GFAP-positive area was determined to investigate the influence of G_M1 lysosomal storage and TMEV infection on astrocytes in the brain of C57BL/6 mice. However, there were no differences in the GFAP-positive area between the study groups at any of the investigated time points ( Supplementary Figure 7 ). Moreover, APP positive spheroids were not detected in the brain of TMEV- and mock-infected Glb1^+/+ and Glb1 ^-/- mice.

Glb1 ^-/- mice developed clinical signs of G_M1 LSD at an age of 14-15 weeks (9-10 weeks post infection), which were not altered by TMEV infection ( Figure 8A ). Similarly, there was no difference in RotaRod performance between TMEV- and mock-infected Glb1 ^-/- mice ( Figure 8B ). This finding suggests that while Glb1^-/- mice show clinical signs due to G_M1 gangliosidosis, they are clinically unaffected by TMEV infection.

In summary, we observed a delayed induction of the immune response in Glb1^-/- compared to Glb1^+/+ mice during TMEV infection. However, T cell infiltration and activation as well as microglia/macrophage activation were prolonged in Glb1^-/- mice upon TMEV infection, which may suggest that G_M1 gangliosidosis leads to a dysregulated immune response during infection. However, TMEV elimination was not substantially impaired in Glb1^-/- mice, which were only affected by G_M1-associated pathology at the end of the experimental run.