Nineteen OT-I (12♂ and 7♀), 20 OT-II (8♂ and 12♀) and 17 C57BL/6 - control mice (WT) (10♂ and 7♀) were bred at the Institute for Neuroimmunology and Multiple Sclerosis Research (IMSF) of the University Medical Center Göttingen. Animal experiments were performed at the Department of Pathology, University of Veterinary Medicine Hannover in accordance with the Lower Saxony State Office for Consumer Protection and Food Safety (LAVES) under the permission number 33.8-42502-17/2418. Mice were housed in individually ventilated cages (Tecniplast Deutschland GmbH, Hohenpeißenberg, Germany) with 12 h light and 12 h darkness at 20–24°C and 50%–60% humidity. Food (ssniff Spezialdiäten GmbH, Soest, Germany) and water were provided ad libitum.

All mice were intracerebrally infected with 5,4x10⁵ plaque forming units (PFU) of the Hannover strain of TMEV-BeAn-1-TiHo at 5 weeks of age under general anaesthesia. For general anaesthesia, mice received a intraperitoneal injection with a combination of Medetomidin (0.0005 mg/g bodyweight) and Ketamin (0.1 mg/g bodyweight), both diluted in saline. Mice were examined weekly for general appearance and posture, behaviour and activity, as well as gait. Further examination consisted of determining body weight and RotaRod performance test (RotaRod Treadmill, TSE Technical & Scientific Equipment, Bad Homburg, Germany) as previously described (61).

Animals were euthanized when they reached either the maximum score in any single category (e.g. paralysis of a limb) or the second highest score in two categories, accompanied by a body weight-loss of 20% or more (Table 1).

For euthanasia, mice were put under general anaesthesia as described above. Thereafter, they received an additional intraperitoneal injection with undiluted Medetomidin (0.001 mg/g bodyweight) and Ketamin (0.2 mg/g bodyweight). Mice were euthanized and post mortally perfused with phosphate buffered saline (PBS) at the following planned necropsy dates: 7 dpi (OT-I n=5 (4♂, 1♀), OT-II n=5 (3♂, 2♀), WT n=5♂), 14 dpi (OT-I n=6 (3♂, 3♀), OT-II n=5 (2♂, 3♀), WT n=6(2♂, 4♀)), 42 dpi (OT-II n=2 (1♂, 1♀)) and 147 dpi (OT-II n=3♀, WT n=6 (3♂, 3♀)). Eight OT-I and five OT-II mice were euthanized for humane reasons between 13 and 37 dpi (OT-I: 13 dpi n=1♂, 16 dpi n=3 (1♂, 2♀), 20 dpi n=2♂, 21 dpi n=1♂, 35 dpi n=1♀; OT-II: 14 dpi n=1♀, 18 dpi n=1♀, 26 dpi n=2 (1♂, 1♀), 37 dpi n=1♂). The right cerebral- and cerebellar hemisphere, as well as tissue samples from cervical-, thoracic- and lumbar spinal cord were fixed for 24 h in 10% neutral buffered formalin and thereafter embedded in paraffin.

In addition, the left cerebral- and cerebellar hemisphere and tissue samples from cervical-, thoracic- and lumbar spinal cord were placed in Tissue-Tek^® O.C.T.™ Compound, deep frozen in liquid nitrogen and stored at -80°C.

Formalin fixed, paraffin embedded (FFPE) and O.C.T.-embedded material was cut into 2 μm thick slices. FFPE tissue sections were mounted on glass slides and stained with hematoxylin and eosin (HE) and Luxol Fast Blue (LFB) (62).

Immunohistochemistry was performed on FFPE and O.C.T. frozen tissue samples via the avidin–biotin–peroxidase complex (ABC)-method (Vector Laboratories Inc., Burlingame, CA, USA) as previously described (63–67). A list of antibodies and the respective methodology can be found in Table 2.

Immunohistochemistry using primary antibodies directed against CD4⁺- and CD8⁺ T cells was performed on O.C.T.-embedded tissue. For this purpose, frozen sections were cut into 5 μm thick slices and mounted on glass slides. Slides were then thawed at room temperature and placed in 0.45% hydrogen peroxide with phosphate buffered saline (PBS) for 30 minutes to inhibit endogenous peroxidases. Afterwards, slides were placed in cover plates and covered in 120μl of rabbit serum in a 1:5 dilution with PBS for 20 minutes to prevent unspecific binding. Thereafter, each slide was covered over night with 120μl of primary antibody dilution (Table 2). After washing with PBS, 120μl of a1:200 dilution of secondary antibody were added (Table 2). Staining was then performed via the ABC-method with 3’3-diaminobenzidine (DAB) as described before (63–66).

Perivascular lymphohistiocytic infiltrates was semi quantitatively evaluated on hematoxylin and eosin (HE) stained longitudinal brain sections and transverse sections of cervical-, thoracic and lumbar spinal cord. The applied scoring systems consisted of four categories (0 = no, 1 = single, 2 = 2-3 layers, 3 = >3 layers of perivascular lymphocytes and or microglia/macrophages, respectively) as described previously (69). In the brain, forebrain, cortex, hippocampus, thalamus, hypothalamus, midbrain, cerebellum, pons and medulla were evaluated separately (Figure 1). For evaluation of the whole brain, single area scores were added up into a cumulative total brain score. The olfactory bulb was not included in the evaluation.

For evaluation of immunohistochemical detection of TMEV, CD3, CD4 and CD8 within the brain and spinal cord, positive cells were counted manually for each area of brain and three sections of spinal cord, respectively. The evaluation of immunohistochemical detection of cleaved caspase 3 was performed by manually counting immunolabeled neurons. The evaluation of immunohistochemical detection of β-amyloid precursor protein (βAPP) was performed by manually counting immunolabeled axons. The evaluation of immunohistochemical detection of glial fibrillary acidic protein (GFAP) in the hippocampus was performed manually by counting immunolabeled cells. For evaluation of Iba-1 and NeuN immunolabeled brain- and spinal cord sections, tissue slides were digitalized using an Olympus SLIDEVIEW VS200 scanner unit. The percentage of positive area was calculated for Iba1 and spinal cords immunolabeled for NeuN within the regions of interest (Figure 1) by setting an optical threshold using the QuPath-software (70–72). For hippocampi immunolabeled for NeuN, the number of positive cells in the CA 2 region of the hippocampus was calculated per 10⁴ μm² using the QuPath-software and compared by neuronal density.

Statistical analysis was performed using SAS/STAT software, Version 9.1 of the SAS^® System for Windows (SAS Institute Inc.) and SPSS for Windows, version 27 (IBM^® SPSS Chicago, IL, United States).

Significance values at 7- and 14 dpi were compared via multiple Wilcoxon Tests with Dwass, Steel, Critchlow-Fligner method as a post-hoc test to correct for multiple comparisons. Statistical significances between euthanized OT-I and OT-II mice were calculated via Mann-Whitney U test. Correlations were calculated via Spearman test. Graphs were plotted using GraphPad Prism 9 and combined into figures via GIMP 2.10.32.

OT-I mice started to display signs of early neurological dysfunction characterized by mild ataxia and gait insecurities (Table 1) between 7 and 14 dpi. The clinical signs of these animals progressed over the course of 5 to 8 days and animals showed weight loss equal to or exceeding 20%, as well as final paralysis of one or two hind limbs, leading to euthanasia for humane reasons prior to the planned necropsies (42 and 147 dpi). In total eight OT-I mice were euthanized at 13 dpi (n=1♂, hind limb paresis), 16 dpi (n=1♂, 2♀, hind limb paresis), 20 dpi (n=2♂, hind limb paresis), 21 dpi (n=1♂, hind limb paresis) and 35 dpi (n=1♀, weight loss), respectively (Figure 2A).

Twelve out of 20 OT-II mice (60%) developed clinical disease characterized by mild ataxia and gait insecurities first appearing between 7 and 14 dpi. Three out of twelve OT-II mice with clinical disease (25%) made a complete clinical recovery after mild initial signs of gait insecurities and ataxia. Five OT-II mice (25% of all OT-II mice) displayed severe clinical disease, characterized by hind limb paralysis and weight loss equal to or exceeding 20% and were euthanized for humane reasons at 14 dpi (n=1♀, weight loss), 18 dpi (n=1♀, hind limb paresis), 26 dpi (n=1♂ (weight loss, 1♀ hind limb paresis) and 37 dpi (n=1♂, weight loss) respectively (Figure 2B). The remaining 4 OT-II mice with mild clinical disease were part of the 14 dpi study group and were sacrificed and necropsied as planned at 14 dpi.

TMEV-infected control mice (n=17) did not display any clinical disease until 7, 14 or 147 dpi, respectively (Figure 2C).

There was no significant difference in RotaRod performance between OT-I, OT-II and WT mice up to 35 dpi (Figure 2D). Mice with hind limb paresis/paralysis were not subjected to the RotaRod test on the day of necropsy. Motor function in unaffected or recovered OT-II mice (n= 11) as well as WT controls, assessed by RotaRod performance test, did not decrease significantly until 14 dpi or 147 dpi, respectively (Figure 2D).

All eight OT-I mice (5♂, 3♀) planned for study groups after 14 dpi were euthanized for humane reasons with no significant difference in survival times between males and females. Out of 10 OT-II mice planned for 42 and 147 dpi respectively, five (2♂, 3♀) were euthanized for humane reasons, while the other five survived to 42 and 147 dpi respectively (1♂, 4♀), without any sex-related differences.

At 7 dpi, OT-I and OT-II mice displayed statistically significant decreased levels of perivascular lymphohistiocytic infiltrates compared to WT mice using HE stained sections (Figures 3A1, A2, A4). By 14 dpi, the perivascular immune cell infiltration increased drastically and reached similar levels in OT-I, OT-II and WT mice (Figures 3B1–B4). OT-I mice at the end stage of the disease, characterized by weight loss of 20% or more and hind limb paralysis, showed perivascular lymphohistiocytic infiltrates at a similar level to OT-II mice at the same stage of clinical disease (Figures 3C1–C3).

Looking at the hippocampus only, there was no significant difference in lymphohistiocytic perivascular infiltrates between the study groups at any time point (Supplementary Figure 1). In the spinal cord, OT-II mice displayed an increase in lymphohistiocytic perivascular infiltrates compared to WT mice at 14 dpi (p=0.028, Supplementary Figure 2B).

At 7 dpi, OT-I mice displayed a statistically significant reduction in CD3⁺ T cell infiltration into the brain compared to both OT-II and control animals (Figures 4A1–A4). By 14 dpi, OT-I mice displayed a significant reduction in the infiltration of CD3⁺ T cells only compared to control mice (Figures 4B1–B4). Upon reaching the defined criteria for euthanasia, OT-II mice displayed stronger CD3⁺ T cell infiltration, compared to OT-I mice at a similar stage of disease (Figures 4C1–C3).

OT-II mice at 7 dpi showed significantly increased T cell infiltration into the brain than OT-I mice, but significantly less infiltration than control mice (Figures 4A1–A4). At 14 dpi, the infiltration of CD3⁺ T cells in OT-I and OT-II mice is no longer statistically significant different, but it was still on a significantly lower level compared to control mice (Figures 4B1–B4). Upon reaching the defined criteria for euthanasia, OT-II mice displayed increased CD3⁺ T cell infiltration, compared to OT-I mice at a similar stage of disease (Figures 4C1–C3).

Looking at the hippocampus only, OT-II mice displayed an increased CD3⁺ T cell infiltration at 7 dpi compared to OT-I mice (p=0.043, Supplementary Figure 3A). By 14 dpi, OT-I and OT-II mice displayed a similar degree of CD3⁺ T cell infiltration with WT mice displaying significantly increased infiltration compared to OT-II mice (p=0.046, Supplementary Figure 3B). In the spinal cord, WT mice displayed an increased CD3 reaction at 7 dpi compared OT-I mice (p=0.025, Supplementary Figure 4A). In animals euthanized for humane reasons at 13-37 dpi, OT-II mice displayed an increased CD3⁺ T cell infiltration compared to OT-I mice (p=0.019, Supplementary Figure 4C).

At 7 dpi, the number of CD4⁺ T cells in the brains of OT-I mice was significantly reduced compared to control mice, whereas the number of CD4⁺ T cells in OT-II mice reached an intermediate level between the one observed in OT-I mice and control mice without statistically significant difference to either of them (Figures 5A1–A4). At 14 dpi, the cerebral infiltration of CD4⁺ T cells had reached similar levels across all study groups (Figures 5B1–B4). At the end stage of disease, OT-II mice displayed a significant increase in CD4⁺ T cell infiltration compared to OT-I mice at the same stage of disease (Figures 5C1–C3).

Looking at the hippocampus only, WT mice displayed an increased CD4⁺ T cell reaction compared to OT-I mice at 7 dpi (p=0.036, Supplementary Figure 5A). At other time points, there were no differences between the study groups (Supplementary Figure 5). In the spinal cord, there was no significant difference in CD4⁺ T cell infiltration between the study groups at any time point (Supplementary Figure 6).

At 7 dpi, the number of CD8⁺ T cells in the brains of OT-I and OT-II mice was significantly reduced compared to control mice (Figures 6A1–A4). At 14 dpi, OT-I mice displayed a reduced infiltration of CD8⁺ T cells compared to both OT-II and control mice without significant difference between OT-II and control mice (Figures 6B1–B4). At the end stage of the clinical disease, OT-I as well as OT-II mice showed very low numbers of CD8⁺ T cell infiltration into the brain, with no statistically significant difference between these groups (Figures 6C1–C3).

Looking at the hippocampus only, both OT-II and WT mice displayed an increased CD8⁺ T cell infiltration compared to OT-I mice at 7 dpi (OT-I vs. OT-II p=0.021, OT-I vs. WT p=0.022, Supplementary Figure 7A). WT mice still displayed an increased CD8⁺ T cell infiltration at 14 dpi compared to OT-I mice (p=0.01, Supplementary Figure 7B). In the spinal cord, there was no significant difference in CD8⁺ T cell infiltration between the study groups at any time point (Supplementary Figure 8).

To assess the viral clearance by the immune system, we investigated the cerebral virus antigen expression at the different time points. There was no statistically significant difference between study groups regarding cerebral virus immunoreactivity at 7 dpi (Figures 7A1–A4). At 14 dpi, control mice had almost completely cleared TMEV from the CNS while OT-I mice still displayed a high virus antigen load within the brain. The strongest immunoreactivity was detected in hippocampus (Supplementary Figure 10) and thalamus (data not shown). Although OT-II mice had not completely cleared the virus from the brain by 14 dpi, the cerebral virus immunoreactivity was significantly lower than in OT-I mice and not statistically different from WT mice (Figures 7B1–B4). Furthermore, in contrast to OT-II and WT mice, OT-I mice displayed virus spread to the spinal cord by 14 dpi (Figures 7C1–C4; Supplementary Figure 9). In animals euthanized for humane reasons, OT-I mice still showed significantly higher cerebral virus loads than OT-II mice with the latter having almost cleared the virus (Figures 7D1–D3).

Looking at the hippocampus only, OT-I mice displayed an increased virus immunoreactivity at 14 dpi compared to WT mice (p=0.015, Supplementary Figure 10B). In animals euthanized for humane reasons at 13-37 dpi, OT-I mice displayed an increased virus immunoreactivity compared to OT-II mice (p=0.018, Supplementary Figure 10C).

At 7 dpi, OT-I mice displayed a diminished cerebral microglia/macrophage activation, that lacked statistical significance compared to WT mice (Figures 8A1–A4). By 14 dpi, this trend was reversed; at this time point OT-I and OT-II mice displayed higher numbers of intracerebral microglia/macrophages than WT mice (Figures 8B1–B4). Microglia/macrophage counts in OT-I mice further increased in animals euthanized for humane reasons, displaying significantly higher numbers of microglia/macrophages compared to OT-II mice at the same stage of disease (Figures 8C1–C3).

OT-II mice at 7 dpi displayed microglia/macrophage-levels similar to WT (Figure 8A1–A4), and this level stayed constant between 7 and 14 dpi. Accordingly, at 14 dpi they displayed microglia/macrophage numbers similar to OT-I and significantly higher than WT mice, for which microglia/macrophage numbers declined between 7 and 14 dpi (Figures 8B1–B4). Microglia/macrophage numbers in OT-II mice also remained steady between 14 dpi and later time points on which the animals were euthanized for humane reasons. Accordingly, microgliosis in OT-II mice euthanized for humane reasons was significantly lower than in OT-I mice at the same disease stage (Figures 8C1–C3).

Iba-1 positive microglia in OT-I mice remained phenotypically unchanged compared to controls until 7 dpi (Figure 8A2). Afterwards they exhibited amoeboid and hyper-ramified phenotype (Figures 8B2, C2). Microglia morphology in OT-II mice did not change as drastically over the course of the experiment. At 7 dpi, OT-II mice mainly showed resting-, ramified- and amoeboid microglia in equal measure (Figure 8A3). This morphology remained to be constant throughout the course of the disease (Figures 8B3, C3). WT mice displayed mainly amoeboid microglia until 7 dpi (Figures 8A4). By 14 dpi, parallel to a reduction in overall numbers of microglia/macrophages, their morphology changed to mainly resting microglia (Figure 8B4).

Looking at the hippocampus only, OT-I mice displayed an increased microgliosis and macrophage infiltration compared to WT mice at 14 dpi (p=0.043, Supplementary Figure 11B). In the spinal cord, OT-I mice displayed an increased microgliosis and macrophage infiltration compared to both OT-II and WT mice at 14 dpi (OT-I vs. OT-II p=0.017, OT-I vs. WT p=0.011, Supplementary Figure 12B). In animals euthanized for humane reasons at 13-37 dpi, OT-I mice also displayed an increased spinal cord microgliosis and macrophage infiltration compared to OT-II mice (p=0.015, Supplementary Figure 12C).

OT-I mice at 7 dpi displayed significantly less neuronal apoptosis in the brain compared to OT-II mice, but no difference compared to WT mice (Figures 9A1–A4). At 14 dpi this was switched around, as by then OT-I mice displayed a statistically significant increase in neuronal apoptosis, compared to both OT-II and control mice (Figures 9B1–B4). Upon euthanasia for humane reasons, OT-I and OT-II mice displayed similar levels of cerebral neuronal apoptosis (Figures 9C1–C3). Hippocampus (Supplementary Figure 13) and thalamus (data not shown) displayed the highest numbers of cleaved caspase3-positive neurons. The neuronal apoptosis in OT-II mice at 7 dpi was significantly increased, compared to OT-I mice, but not higher than in WT mice (Figures 9A1–A4). By 14 dpi, OT-II mice showed a lower level of neuronal apoptosis compared to OT-I mice, but not significantly different from WT mice (Figures 9B1–B4).

Neuronal apoptosis in OT-I mice correlated moderately with the amount of detectable TMEV-antigen (coefficient: 0.77, p=0.0001) and, to a lesser degree, with microgliosis (coefficient: 0.586, p=0.008). In OT-II mice, neuronal apoptosis correlated more strongly with infiltration of CD4⁺ T cells (coefficient: 0.741, p=0.00018) and not as strongly with detection of TMEV-antigen (coefficient: 0.6887, p=0.0008).

Looking at the hippocampus only, all study groups displayed a similar degree of neuronal apoptosis at all time points (Supplementary Figure 13). No mouse displayed neuronal apoptosis in the spinal cord at any time point (Supplementary Figure 14).

All study groups displayed a similar degree of glial fibrillary acidic protein (GFAP) positive cells in the hippocampus at all time points (Supplementary Figure 15).

Hippocampal neuronal density in the CA2 region, as measured by expression of NeuN remained similar between all study groups at all time points (Figure 10). Neuronal loss mainly affected the hippocampal CA2 region (Figure 10). In the spinal cord, there were no differences in neuronal NeuN expression between the study groups at any time point (Supplementary Figure 16).

Neither OT-I, OT-II nor control mice displayed axonal damage within the brain, hippocampus or spinal cord at any time point (Supplementary Figures 17–19).