IFNγ drives neuroinflammation, demyelination, and neurodegeneration in a mouse model of multiple system atrophy

Male and female C57BL/6 (#000664 Jackson Laboratories) were used for these studies and maintained on a congenic background. IFNγ/Thy1.1 reporter mice with a C57BL/6 background (generously donated by Dr. Casey Weaver) were also used and have been previously described and characterized [11]. Additionally, male and female Tbet -/- mice (#004648 Jackson Laboratories) were used. Under a C57BL/6 background, these mice have exon 1 of the T-box 21 (Tbx21) deleted. All research conducted on animals were approved by the Institutional Animal Care and Use Committee at the University of Alabama at Birmingham (UAB).

The Olig001 vector is a modified AAV capsid generated via directed evolution that has been characterized previously [21, 22, 26]. Briefly, the Olig001 capsid has a > 95% tropism for oligodendrocytes and the vectors utilized contained the CBh promoter and bovine growth hormone polyA, controlling the expression of either transgene (human α-syn or GFP as control).

Male and female mice aged to 8–12 weeks, were anesthetized with isoflurane applied by an isoflurane vaporizing instrument provided by the Animal Resource Program at UAB. Using a Hamilton syringe and an automatic injecting system, mice were unilaterally (luxol fast blue, DAB and immunofluorescence staining; n = 5) or bilaterally (flow cytometry; n = 3–5 and stereology; n = 7–10) injected with 2 μl of Olig001-GFP (1 × 10¹³ vector genomes (vg)/ml) or Olig001-SYN (1 × 10¹³ vg/ml) into the dorsolateral striatum at a rate of 0.5 μl/min to mimic MSA-P. The needle was left in the injection site for an additional 2 min and then slowly retracted over the course of 2 min. The stereotaxic coordinates used from bregma were AP+ 0.7 mm, ML+ / − 2.0 mm, and DV − 2.9 mm from dura. All surgical protocols and aftercare were followed and approved by the Institutional Animal Care and Use Committee at the University of Alabama at Birmingham.

Four weeks post transduction of the Olig001 virus, mice were anesthetized and transcardially perfused with 0.01M Phosphate-buffered saline (PBS) pH 7.4, followed by a fixation with 4% paraformaldehyde (in PBS, pH 7.4; PFA). Brains were dissected and incubated in 4% PFA solution for 4 h at 4 °C. After PFA fixation, the brains were cryoprotected in a 30% sucrose (in PBS) solution for 3 days until brains were fully saturated. Brains were frozen and cryosectioned coronally at 40 μm on a sliding microtome. Tissue was stored in a 50% glycerol/PBS solution at − 20 °C.

Forty-μm thick free-floating sections were washed in 0.01 M tris-buffered solution (TBS; pH 7.4) 3 X for 5 min. The tissue then underwent a sodium citrate antigen retrieval (0.1M; pH 7.3) process for 30 min at 37 °C. After antigen retrieval the free-floating sections were washed and blocked in 5% normal serum for 1–2 h. The sections were thereafter incubated in 1% serum in TBS-Triton (TBST) primary antibody solution consisting of one of the following antibodies: anti-Iba1 (1:500, WACO), anti-Olig2 (1:250, clone SP07-02; R&D), anti-GFAP (1:500, clone DIF48; Abcam), anti-CD3 (1:500, clone 17A2; Thermo Fisher), anti-CD8 (1:500, clone 4SM15; eBioscience), anti-Thy1.1 (1:500, OX-7; Invitrogen) anti-CD4 (1:500, clone RM4-5; Thermo Fisher), anti-NK1.1 (1:250, clone EPR22990-12; Abcam). After an overnight incubation, the free-floating sections were washed and put into a 1% serum TBST secondary solution for 2 h. Sections were mounted onto coated glass slides, and cover slipped using hard set mounting medium (Vector Laboratories). Fluorescence images were collected on a Ti2 Nikon microscope using a Ci2 confocal system.

Free floating striatal sections were quenched in a 3% hydrogen peroxide/50% methanol in 0.01M TBS (pH 7.4) solution at room temp for 5 min. After three TBS washes, tissue was incubated in antigen retrieval sodium citrate solution for 30 min at 37 °C. Background staining was blocked in 5% serum and incubated with either an anti-MHCII (1:500, M5/114.15.2; Thermo Fisher) or a pSer129 (1:5000, clone EPI53644; Abcam) antibody 1% serum overnight at 4 °C. The following day, sections were incubated with a biotinylated goat anti-rat IgG secondary antibody (1:1000, Vector Labs) in a 1% serum TBST solution. The R.T.U Vectastain ABC Reagent kit and DAB kits (Vector Labs) were used to develop the stain according to manufacturer’s protocol. Striatal sections were mounted onto plus coated slides and dehydrated with a gradient of ethanol solutions. Lastly slides were cover slipped with Permount (Electron Microscopy Sciences). Slides were imaged at 10X on a Zeiss imager M2 brightfield microscope (MBF Biosciences). Data was analyzed in ImageJ and the fold change of mean grey value between the ipsi- and contra-lateral side were calculated. This method of analysis was applied to Figs. 1E and Additional file 1: Fig. S1B, S5B.

To quantify pSer129+ cells, sections were immunostained and dehydrated using the DAB protocol described above. After slides were cover slipped, images were taken at 20× on a Zeiss imager M2 brightfield microscope (MBF Biosciences), and pSer129+ cells were counted as previously published [35]. Briefly, images were imported into ImageJ and a grid was overlayed with a 35,000 per point grid. Five grids within the dorsolateral striatum section were chosen at random to count, and the average was calculated. Three sections were counted per slide, and the average GCI count of the three sections was calculated as the total of pSer129+ cells within the dorsolateral striatum of one mouse.

C57BL/6 mice were treated three days before Olig001-SYN transduction with i.p. injection of either a neutralizing IFNγ antibody (clone XMG1.2; 200 ng; n = 5) or isotype control (IgG1; 200 ng; n = 5). Three days following the initial i.p. injection, mice received an injection of either Olig001-GFP or Olig001-SYN in the dorsolateral striatum. Immediately after vector injection, mice were given an i.p. dose of their respective treatment. To continue their treatment, mice were injected i.p. every three days with 200 ng of either neutralizing IFNγ or the isotype control. After 30 days, mice were anesthetized for their respected endpoints.

Forty-μm thick mounted brain sections were quickly washed with DI water and incubated in a 0.1% luxol fast blue solution at 60 °C for 2 h. Excess dye was removed with running DI water. To differentiate and visualize the myelin from the rest of the tissue, slides were dipped in a 0.05% Lithium Carbonate solution two times for 1 min, followed by three 70% ethanol washes. This differentiation step was repeated until the myelin was stained blue and non-lipid parts of the tissue were clear. Slides were quickly dehydrated and mounted with Permount (Electron Microscopy Sciences). Images were taken at 10X on a Zeiss Axio Imager M2 microscope (MFB Biosciences). Images were quantified with ImageJ. For 4 week timepoints, fold change of contra- to ispi-lateral were calculated based on mean grey value of the luxol stain in the striatum and corpus callosum. Due to the bilateral injections of the 6-month timepoint, the mean grey value of Olig001-SYN and Olig001-GFP were calculated. and the fold change was calculated by comparing the dorsal striatum and corpus callosum of Olig001-SYN to Olig001-GFP.

Four weeks post Olig001 delivery, mice were anesthetized and transcardially perfused with 0.01M PBS pH 7.4. Brain tissue was removed and the striata were dissected. Striatal tissues were triturated and digested with 1 mg/mL Collagenase IV (Sigma) and 20 μg/mL DNAse I (Sigma) diluted in RPMI 1640 with 10% heat inactivated fetal bovine serum, 1% glutamine (Sigma), and 1% Penicillin–Streptomycin (Sigma). After enzyme digestion, samples were filtered through a 70 uM filter and mononuclear cells were separated out using a 30/70% percoll gradient (GE).

For all cell labeling, isolated cells were blocked with anti-Fcy receptor (1:100; BD Biosciences). Cell surfaces were labeled with the following fluorescent-conjugated antibodies against CD45 (clone 30-F11; eBioscience), CD11b (clone M1/70; BioLegend), MHCII (clone M5/114.15.2; BioLegend), Ly6C (clone HK1.4; BioLegend), CD4 (clone GK1.5; BioLegend), CD8a (clone 53.6.7; BioLegend), Thy1.1 (clone HIS51; BD Biosciences), anti-PDGFRa (clone APA5; BioLegend), anti-O4 (R&D Systems), or hCD2 (clone RPA2.10; eBioscience). A fixable viability dye was used to distinguish live cells per manufacturer’s instructions (Fixable Near-IR LIVE/DEAD Stain Kit, Invitrogen).

For intracellular transcription factor labeling, the Foxp3/Transcription Factor Staining Kit (eBioscience) was used accordingly with fluorescent-conjugated antibodies against FOXP3 (clone FJK-16S; eBioscience), T-bet (clone 4B10; BioLegend), GATA2 (clone 16E10A23; BioLegend), RORγt (clone Q31-378; BD Biosciences), or Olig2 (clone 211F1.1; Sigma Millipore). An Attune Nxt (Thermo Fisher Scientific) or a BD Symphony flow cytometer (BD Sciences) were used to analyze samples and FlowJo (Tree Star) software were used for analysis. Mean cell count numbers, percentages, and mean fluorescent intensity (MFI) were measured with FlowJo software to assess for neuroinflammation.

To quantify NeuN+ neurons within the dorsolateral striatum, unbiased stereology was used. Mice used for stereological assessment were bilaterally injected for behavioral time points. Six months post transduction, tissues were harvested and sectioned as described in the immunohistochemistry tissue preparation section above. Free floating sections were stained with NeuN (clone EPR12763; Abcam) and developed with the DAB protocol described above. Once slides were stained and cover slipped, Zeiss Axio Imager M2 microscope and StereoInvestigator Software (version 2021.1.3) (MFB Biosciences) was used to count NeuN+ cells. Due to the localization of α-syn, the contours used for counting were drawn around the dorsolateral striatum (Fig. 4A). To encompass areas of pathology, six serial sections were counted per mouse. Grid size was 300 μm × 300 μm and the counting frame was set at 40 × 40. After dehydration, the thickness of the sections was determined to be 25 μm. Six sections of the striatum were counted and the average counting frame per section was 23. Estimated populations were generated, and a two-way ANOVA and a Tukey post hoc test was used to determine the neuronal loss significance.

The TH+ neurons in the substantia nigra pars compacta (SNpc) were quantified using methods that were previously described [10, 36]. Briefly, SNpc tissue sections were stained with anti-tyrosine hydroxylase (TH) (EP1536Y, Abcam) using the DAB labeling and quantification protocol as described. Once tissue was stained and mounted, the same scope and software used in the stereology for NeuN+ counting was used to analyze TH+ neurons. Five sections were counted per mouse. Grid size of 100 μm × 100 μm with a counting frame of 50 μm × 50 μm was used to count TH+ neurons. The thickness of the sections was 25 μm. Estimated populations were generated and compared to a naïve control. Data was analyzed with a one-way ANOVA with Tukey post test for significance (with 95% confidence and p < 0.05).

At 6 months post Olig001-GFP or Olig001-SYN induction, mice were assessed for motor deficits with open field. Before testing, mice were habituated to testing room for 1 h at the start of the day. Mice were placed in a 44.45 cm × 44.45 cm plastic box and the Noldus Ethovision XT 16 software recorded the activity of the mice for 15 min. After the 15 min, mice were returned to their home cage. The total distance and velocity were measured and recorded.

All graphs and corresponding statistical tests were generated or performed using Prism software (GraphPad). For flow cytometry, data points were compared across time points/antibody treatment using a ranked-summed, non-parametric Wilcoxon test (with 95% confidence and p < 0.05). Stereology analysis in Fig. 4 was analyzed with a two-way ANOVA was used and Tukey’s Honestly-Significant-Difference (with 95% confidence and p < 0.05) was used to determine significance. Stereology analyzed in Additional file 1: Fig. S3 with a one-way ANOVA with Tukey posthoc test for significance (with 95% confidence and p < 0.05).

The authors affirm that the findings of this manuscript are supported by the data therein. Additional information can be requested from the corresponding author.

IFNγ signaling induces MHCII expression on the cell surface of APCs [13]. To determine if α-syn overexpression in oligodendrocytes induced neuroinflammation as a result of IFNγ expression, we utilized mice in which the required transcription factor for IFNγ production in Th1 cells, Tbet, was deleted (Tbet -/-). Without Tbet, a Th1 cell cannot carry out its effector functions, one of those being producing IFNγ. Tbet -/- and WT mice, aged 8–12 weeks, received an intracranial injection of 2uL of Olig001-SYN (or Olig001-GFP as control) in the dorsolateral striatum (Fig. 1A). 4 weeks post transduction, using immunohistochemistry (IHC) and mononuclear cell isolation and flow cytometry, we found that Tbet -/- mice showed a reduction in infiltrating Ly6C+ monocytes (CD11b+ , CD45hi, Ly6C+) (Fig. 1B and C). Within the monocyte population there was no difference in MHCII expression (Fig. 1C). In response to α-syn expression, Tbet -/- mice displayed no change in the number of microglia within the striatum (Fig. 1C), however, in the absence of Tbet, a significant reduction in MHCII expression was observed on the cell surface of microglia (CD11b+ CD45lo) via flow cytometry (Fig. 1D). Consistent with our flow cytometry results, utilizing immunohistochemistry, we observed a decrease in MHCII expression in the dorsolateral striatum (Fig. 1E and F) indicating Tbet expression is required for MHCII expression and peripheral monocyte entry.

To determine whether Tbet-mediated IFNγ is responsible for T cell infiltration in response to α-syn overexpression, we performed mononuclear cell isolation and flow cytometry 4 weeks post vector delivery in WT and Tbet -/- mice (Fig. 2A). Tbet -/- mice displayed a significant reduction in infiltration of CD4+ T cells (Fig. 2B) and an increase of CD8+ T cells (Fig. 2B). Given our previous observations showing that CD4+ T cells are required for MSA pathology, we further investigated the CD4+ T cell response. In the dorsolateral striatum, compared to WT mice, Tbet -/- mice displayed a reduction of CD4+ T cells in close proximity to α-syn (pSer129+) GCI pathology (Fig. 2C). Within the CD4+ T cell population, there were significant changes among the T cell effector subsets Th1, Th17, and T_reg including a reduction in the number of Th1 and Th17 cells (Fig. 2D). Conversely, there was a significant increase in the number of T_reg cells within the CD4+ T cell population, suggesting that loss of Tbet shifts the T cell repertoire from a pro-inflammatory (Th1, Th17) to a restorative (T_reg) state. As IL-17 produced by T cells can also contribute to proinflammatory or autoimmune responses [15, 16, 20], we sought to determine if IL-17 producing T cells, Th17, were contributing to disease pathogenesis. To further understand Th17 cells, we utilized mice deficient in the transcription factor required for Th17 development, RORγT -/-. We performed IHC and flow cytometry on WT and RORγT -/- treated with Olig001-SYN. Four weeks post-transduction, RORγT -/- mice overexpressing α-syn in oligodendrocytes showed no attenuation of MHCII expression, CD4+ and CD8+ T cell entry, and enhanced demyelination compared to WT mice (Additional file 1: Fig. S1) indicating that IL-17 producing Th17 cells do not drive Olig001-SYN mediated neuroinflammation and demyelination.

Demyelination is a key feature of MSA pathology [8] and is observed 4 weeks post transduction in the Olig001-SYN mouse model of MSA. To assess how Tbet contributes to this pathology, Tbet -/- and WT controls, 8–12 weeks of age, were injected with Olig001-SYN in the dorsolateral striatum (Fig. 3A). 4 weeks post-transduction demyelination was assessed via luxol fast blue staining. As a control, naïve, WT and Tbet -/- mice were assessed for luxol fast blue staining and flow cytometry to determine whether Tbet expression affects myelin integrity and oligodendrocyte cell numbers in the striatum. No significant differences were observed between WT and Tbet -/- mice prior to Olig001 induction (Additional file 1: Fig. S2E–G). Following Olig001-SYN transduction, when compared to WT mice, Tbet -/- mice displayed a fourfold increase in the degree of myelination in the dorsolateral striatum and corpus callosum (Fig. 3B and C). The preservation of myelin seen in Tbet -/- is comparable to healthy, non-inflamed mouse striatum (Fig. 3B) indicating that Tbet expression is required for demyelination in the Olig001-SYN mouse model of MSA.

Previous studies have indicated α-syn overexpression in oligodendroglia results in neurodegeneration in the dorsolateral striatum of rats [22]. Additionally, there is NeuN+ cell loss within the dorsolateral striatum of mice when α-syn is overexpressed in oligodendroglia (Additional file 1: Fig. S3). To determine whether IFNγ drives α-syn -induced neurodegeneration, mice were bilaterally injected with either Olig001-GFP or Olig001-SYN. 6 months post transduction, tissue was collected and unbiased stereology was conducted in the dorsolateral striatum (Fig. 4A). Preliminary analysis in 3 month old naïve Tbet -/- mice indicated a significant reduction in the number of NeuN+ cells in the dorsolateral stratum compared to WT mice (Additional file 1: Fig. S2B). 6 months post transduction, WT mice transduced with Olig001-SYN displayed a significant loss of NeuN+ neurons in the dorsolateral striatum (Fig. 4B). However, when compared to the Olig001-GFP control, Tbet -/- mice displayed no significant NeuN+ cell loss between the Olig001-GFP and Olig001-SYN groups (Fig. 4B) indicating Tbet mediates Olig001-SYN induced neurodegeneration. To control for Olig001 vector toxicity, an additional 6-month study was conducted in WT mice transduced with either Olig001-GFP or Olig001-SYN (Additional file 1: Fig. S3). we observed no significant cell loss in the dorsal-lateral striatum between WT naïve and WT Olig001-GFP mice, 6 months post transduction (Additional file 1: Fig. S3B and C).

To determine if α-syn overexpression in oligodendroglia results in TH+ neuron loss in the SNpc, serial sections from WT Olig001-GFP and Olig001-SYN mice were analyzed via unbiased stereology at 6 months post transduction (Additional file 1: Fig. S3D). No significant TH+ cell loss was observed in WT mice transduced with Olig001-GFP control or Olig001-SYN (Additional file 1: Fig. S3D). At 6 months post-transduction, significant neuroinflammation was observed in WT Olig001-SYN injected mice including persistent myeloid activation and infiltrating T cells (Additional file 1: Fig. S4).

To understand if IFNγ drives long term demyelination similar to what is seen in post-mortem MSA tissue, demyelination was assessed at 6 months with a luxol fast blue stain. We observed no change in myelin within the Tbet -/- Olig001-SYN group when compared to, Olig001-GFP control, (Fig. 4C), whereas the WT Olig001-SYN group showed significant demyelination when compared to WT Olig001-GFP (Fig. 4C). Together, these results (Fig. 4B and C) support the hypothesis that IFNγ is crucial neuroinflammation, demyelination, and neurodegeneration in the Olig001-SYN mouse model of MSA.

Our genetic approach established a role for Tbet-mediated IFNγ in Olig001-SYN-mediated MSA pathology. However, as IFNγ expression is not the only pathway mediated by Tbet, a key question remains: is IFNγ driving pathology in the Olig001-SYN model? To specifically determine the role of IFNγ in Olig001-SYN mediated neuroinflammation and demyelination, an IFNγ neutralizing antibody was used to globally decrease IFNγ. WT mice (8–12 weeks old) were pre-treated with either an IFNγ neutralizing antibody (XMG1.2; 200ng) or an isotype control (IgG1; 200ng) intraperitoneally (i.p.) three days prior to Olig001-SYN transduction, and every three days over the course of 4 weeks (Fig. 5A). Four weeks post-transduction, striatal tissue was harvested to assess for neuroinflammation via immunohistochemistry and flow cytometry. Our results indicate XMG1.2 treatment significantly attenuated pro-inflammatory monocyte infiltration and MHCII expression on microglia (Fig. 5B and C) in the ipsilateral striatum compared to isotype control. Additionally, XMG1.2 treatment significantly decreased the number of CD4+ and CD8+ T cells infiltrating the ipsilateral striatum (Fig. 5D). While α-syn expression in oligodendrocytes was unaffected by XMG1.2 treatment (Fig. 5E, Additional file 1: Fig. S5), peripheral T cell entry was attenuated confirming that IFNγ is mediating α-syn-induced neuroinflammation.

Lastly, to determine if IFNγ drives demyelination in Olig001-SYN, luxol fast blue staining was performed on striatal sections from Olig001-SYN treated mice (Fig. 6A). XMG1.2 treatment preserved myelin in the dorsolateral striatum, specifically in the corpus callosum, compared to isotype control in Olig001-SYN transduced mice (Fig. 6B and C). These results indicate that neutralizing IFNγ attenuates Olig001-SYN mediated demyelination, again supporting a role of IFNγ expression in Olig001-SYN-mediated demyelination.

In post-mortem MSA tissue there is evidence of infiltrating CD4+ and CD8+ T cells [35] in the parenchyma and increased IFNγ in the CSF [5], and previous studies in preclinical models have shown that CD4+ T cells are required for neuroinflammation and demyelination [35]. Similarly, the results presented here, using genetic and pharmacological approaches, show that IFNγ is a key mediator of the neuroinflammation, demyelination, and neurodegeneration observed in the Olig001-SYN model. While our previous studies in the Olig001-SYN model suggest IFNγ is produced by T cells [35], it is unclear if other CNS resident and infiltrating immune cells produce IFNγ in response to α-syn expression in oligodendrocytes. To determine which CNS resident and infiltrating immune cells express IFNγ as a result of α-syn overexpression, we used a Thy1.1/IFNγ reporter mouse where Thy1.1 is expressed from the IFNγ promoter [11]. In this mouse, when IFNγ is expressed, Thy1.1 is expressed on the cell surface. Using this reporter model, we isolated mononuclear cells from the dorsolateral striatum of mice treated with Olig001-SYN or Olig001-GFP control (Fig. 7A). Using immunohistochemistry, immune populations known to produce IFNγ (CD4+ T cells, CD8+ T cells, NK cells, astrocytes, and microglia) were investigated for Thy1.1 expression on the cell surface via IHC and flow cytometry (Fig. 7B). These lymphocytes (CD45+) were analyzed, and our results indicate that CD4+ T cells expressed the overwhelming majority of Thy1.1 on the cell surface in response to α-syn overexpression in oligodendrocytes (Fig. 7C). Upon further investigation, the majority of Thy1.1 expression was found on CD45+ TCRb+ CD4+ T cells (Fig. 7D–F), matching results seen in IHC experiments (Fig. 7B). Given that CD4+ T cells are producing the pathogenic IFNγ (Fig. 7F) when α-syn is expressed in oligodendrocytes, these data suggest that the CD4+ T cell effector subtype, Th1 cells, are facilitating the disease process via production of IFNγ. Together, our results show that other immune cell types like CD8+ T cells, B cells, and NK cells do not significantly express IFNγ following α-syn overexpression in oligodendrocytes, but CD4+ T cells drive MSA pathology via IFNγ expression.