Sacsin Deletion Induces Aggregation of Glial Intermediate Filaments

Autosomal recessive spastic ataxia of Charlevoix-Saguenay (ARSACS) is a neurodegenerative disorder commonly diagnosed in infants and characterized by progressive cerebellar ataxia, spasticity, motor sensory neuropathy and axonal demyelination. ARSACS is caused by mutations in the SACS gene that lead to truncated or defective forms of the 520 kDa multidomain protein, sacsin. Sacsin function is exclusively studied on neuronal cells, where it regulates mitochondrial network organization and facilitates the normal polymerization of neuronal intermediate filaments (i.e., neurofilaments and vimentin). Here, we show that sacsin is also highly expressed in astrocytes, C6 rat glioma cells and N9 mouse microglia. Sacsin knockout in C6 cells (C6Sacs−/−) induced the accumulation of the glial intermediate filaments glial fibrillary acidic protein (GFAP), nestin and vimentin in the juxtanuclear area, and a concomitant depletion of mitochondria. C6Sacs−/− cells showed impaired responses to oxidative challenges (Rotenone) and inflammatory stimuli (Interleukin-6). GFAP aggregation is also associated with other neurodegenerative conditions diagnosed in infants, such as Alexander disease or Giant Axonal Neuropathy. Our results, and the similarities between these disorders, reinforce the possible connection between ARSACS and intermediate filament-associated diseases and point to a potential role of glia in ARSACS pathology.


Introduction
The autosomal recessive spastic ataxia of Charlevoix-Saguenay (ARSACS) is a rare, earlyonset neurodegenerative disorder, usually diagnosed at gait initiation (12-18 months) [1]. ARSACS is clinically characterized by cerebellar ataxia, spasticity, axonal demyelination, sensory-motor peripheral neuropathy, amyotrophy, dysarthria, skeletal finger and feet abnormalities, nystagmus, retinal hypermyelination and variable intellectual dysfunction [2]. Its histopathological features include the atrophy of the anterior vermis, associated with the loss of Purkinje cells in the cerebellum [1,2] and deposits of lipofuscin in cerebellar cortical
For experiments, cells were plated on sterile plastic dishes or on sterile glass coverslips and allowed to adhere for 16-24 h before experiments and/or sample preparation. When applicable, C6 and C6 Sacs−/− cells were incubated with rotenone (5 µM) for 4 h or its vehicle (DMSO 0.1% v/v). For cytokine experiments, complete medium was removed, cells were washed with PBS and incubated in serum-free medium for 2 h before adding LIF (200 ng/mL), IL-6 (30 ng/mL) and/or the soluble IL-6 receptor (IL-6R, 60 ng/mL). Cells were then incubated for 2 h with the cytokines in the absence of serum before sample preparation.

Generation of Sacsin KO Cell Lines Using the CRISPR/Cas9 System
A sacsin knockout strain from C6 rat glioma cells was generated using a commercial Sacsin CRISPR/Cas9 knockout plasmid set (sc-404592, Santa Cruz Biotechnologies, Dallas, TX, USA), following manufacturers' instructions. Transfected cells were isolated by fluorescence-activated cell sorting (FACS) and single cells were plated in 96-well plates for clonal expansion. Resulting clones were probed for sacsin expression by Western blot.

Primary Cultures of Astrocytes
Cultures were prepared as described elsewhere [19]. Briefly, astrocyte-enriched cultures were prepared from neonatal Sprague Dawley rat pups' cerebral cortexes (0-2 days). Animals were sacrificed by decapitation and the brains were dissected in ice-cold PBS (NaCl 137 mM, KCl 2.7 mM, Na 2 HPO 4 .2H 2 O 8 mM and KH 2 PO 4 1.5 mM, pH 7.4). Cortexes were isolated, placed in 10 mL of 4.5 g/L glucose DMEM, supplemented with 10% FBS and 1% antibiotic/antimycotic, and dissociated mechanically through up-and-down movements with a serological pipette until no cell clumps were observed. Cell suspension was filtered successively through 230 µm and 70 µm (BD Falcon, NJ, United States) cell strainers and centrifuged at room temperature (RT) at 200× g for 10 min. The final pellet was resuspended in 4.5 g/L glucose DMEM, and cells were seeded according to the required assay. Cultures were kept at 37 • C in a humidified atmosphere (5% CO 2 ) and medium was changed twice a week. At 10 days in vitro (DIV), plates were shaken for 6 h in an orbital shaker at 300 rpm to remove any contaminating microglial cells and obtain astrocytic-enriched cultures.

Cell Viability Assays
For MTT assays, C6 and C6 Sacs−/− cells were seeded onto 96-well plates at a concentration of 10 4 cells/well. After the corresponding treatments, cells were incubated with MTT (0.5 mg/mL) for 2 h. The medium was then replaced with DMSO (100% v/v). After 15 min of incubation in the dark at RT, absorbance was determined using an automatic microplate reader (Tecan Sunrise Microplate Reader) (Tecan, Männedorf, Switzerland) at 490 nm. For LDH assays, cells were seeded in 24-well plates at a concentration of 10 5 cells/well, and the determination of released and total LDH was carried out by means of the LDH cytotoxicity detection kit (Takara), following manufacturer's instructions. DAPI exclusion assay was also used as a complementary cytotoxicity assay by flow cytometry, as described below.

Flow Cytometry
Cells were seeded in 6-well plates (5 × 10 5 cells/well). After treatments, cells were detached by trypsin (0.05% w/v), collected, resuspended in PBS, and incubated with 10 µM DCFH-DA (All reactive oxygen species) or DHE (superoxide radicals) in DMEM medium (without serum) for 20 min at 37 • C in the dark. After washing cells twice with PBS, pellets were resuspended in PBS with DAPI (1 µg/mL) to discriminate between live and dead cells, and fluorescence was immediately analyzed by means of a BD LSRFortessa X-20 cell analyzer (BD Biosciences, San Jose, CA, USA). At least 10,000 events (low velocity) were recorded for analysis with FLOWJO software Version 9 (Emerald Biotech Co., Ltd., Córdoba, Spain).

Western Blot
Western blot analysis was performed as previously described [20]. Briefly, cells were lysed using an NP-40 lysis buffer (150 mM NaCl, 50 mM Tris-HCl pH 7.4/7.5, 1% NP-40 v/v) supplemented with protease inhibitors (NZYTech) (Lisboa, Portugal) and phosphatase inhibitors (Halt Phosphatase Inhibitor Single-use Cocktail, Thermo Fisher Scientifics, Waltham, MA, USA). Samples were sonicated in UP200s sonicator (Hielscher Ultrasonics GmbH, Teltow, Germany) for 8 s. Cell suspension was then centrifuged at 10,000× g for 10 min at 4 • C, and the soluble protein fraction was collected and quantified by the Bradford method, incubating 1 µL of sample with 200 µL of Bradford solution (Alfa Aesar, Ward Hill, MA, USA) for 5 min and reading the absorbance at 595 nm. Thirty micrograms of total protein was separated by SDS-PAGE on 10% (w/v) polyacrylamide gels (for detection of STAT3) or gradient gels 6% + 15% (w/v) (for detection of sacsin and IFs), and transferred to a nitrocellulose membrane. Protein transfer quality was assessed by Ponceau S staining. Membranes were blocked with 5% (w/v) milk in TBS-T (TBS supplemented with 0.1% Tween-20) and probed with primary antibodies in 5% (w/v) Bovine Serum Albumin (BSA) in PBS overnight at 4 • C. Primary antibodies were used at a dilution of 1:1000, except for anti-GAPDH and anti-sacsin antibodies (N-terminal, sc-515118, Santa Cruz Biotechnologies (Dallas, TX, USA); and C-terminal, ABN1019, Merck-Millipore (Burlington, MA, USA) which were used at 1:2000 and 1:200 dilutions, respectively. Membranes were then washed three times with TBS-T for 10 min, followed by 2 h incubation with HRP-conjugated secondary antibodies (1:10,000) in blocking solution. After washing membranes three times with TBS-T for 10 min, chemiluminescence detection was performed using the Pierce ECL Plus Western Blotting Substrate and the Amersham Imager 680 blot and gel imager (Cytiva, Marlborough, MA, USA). The integrated intensity of each band was calculated using computer-assisted densitometry analysis with ImageJ software, normalized to the loading control GAPDH as appropriate.

Filter Trap Assays
Cells were washed with PBS 1X and collected by scraping in native lysis buffer (173 mM NaCl, 50 mM Tris pH 7.4, 5 mM EDTA) supplemented with protease and phosphatase inhibitor cocktail. Samples were sonicated for 10 s at 5 mA using a UP200S Sonicator (Hielscher, Teltow, Germany). Protein extracts were collected after sample centrifugation at 10,000× g for 10 min at 4 • C and quantified by means of the Bradford method. One hundred µg of native protein extracts were diluted in PBS to produce a final volume of 100 µL and SDS was added to a final concentration of 1% (w/v). Samples were loaded on a dot-blotting device and filtered by vacuum through nitrocellulose membranes previously incubated with 1% (w/v) SDS solution in PBS. After filtration, membranes were washed twice with 1% (w/v) SDS solution in PBS and processed for immunoblotting detection, as described above.

Fluorescence Microscopy of Live Cells and Immunocytochemistry
For fluorescent microscopy, C6 and C6 Sacs−/− cells were seeded at the density of 10 5 cells/cm 2 on 35 mm glass-bottom dishes. Twenty-four hours after seeding, cells were incubated with either sirActin kit (Spirochrome, Stein am Rhein, Germany) and Tubulin tracker deep red (Invitrogen, Carlsbad, CA, USA) following manufacturer's instructions. For visualization of intermediate filaments, cells were incubated with MitoTracker Red CMXRos (25 nM) for 30 min at 37 • C in 5% CO 2 atmosphere before processing samples for immunocytochemistry. Cells were fixed in ice-cold methanol for 20 min and blocked with 1% (w/v) BSA in PBST (PBS supplemented with 0.1% Tween-20) for 1 h at RT. Overnight incubation at 4 • C was performed with the following primary antibodies diluted in blocking solution: mouse monoclonal anti-sacsin (1:50), rabbit polyclonal anti-GFAP (1:200), mouse monoclonal anti-nestin (1:100) and mouse monoclonal anti-vimentin (1:100). Cells were incubated with the corresponding goat anti-mouse or anti-rabbit secondary antibodies conjugated with AlexaFluor 488 (Thermo Fisher Scientifics, Waltham, MA, USA) or AlexaFluor 594 (Thermo Fisher Scientifics, Waltham, MA, USA) (1:800) for 2 h at RT, and counterstained with the nuclear marker Hoechst 33342 (Thermo Fisher Scientifics, Waltham, MA, USA). Images were acquired by means of a Leica TCS SPE high-resolution spectral confocal system (Wetzlar, Germany) equipped with a Leica DFC 365 FX camera (Wetzlar, Germany) using a 63X/1.4 oil objective (Wetzlar, Germany) and processed by Leica LAS X Core (Wetzlar, Germany) and ImageJ software (National Institutes of Health, Bethesda, MD, USA). The total numbers of reference C6 cells counted in three independent experiments were 549 (GFAP), 464 (Nestin) and 353 (Vimentin). The total numbers of C6 Sacs−/− cells counted in three independent experiments were 836 (GFAP), 697 (Nestin) and 438 (Vimentin).

Statistical Analysis
Statistical analysis and graphical representation of data were performed using Graph-Pad Prism software Version 8 (GraphPad, San Diego, CA, USA). Sample data are represented as mean ± standard error (SEM) of three independent experiments. For statistical evaluation, one-way or two-way Analysis of Variance (ANOVA) and Tukey's post hoc test were used for multiple comparisons. Student's t-test was applied for comparisons in experiments with two groups. Results were considered significant when p < 0.05.

Astroglia Express Sacsin
Public databases indicated that glial cells contain sacsin mRNA, but data are scarce and there is no empirical evidence that sacsin mRNA is actually translated into protein.
The sacsin protein is easily detected by immunoblotting in rat primary astroglia and C6 rat glioblastoma cells at approximately the same levels ( Figure 1A). Sacsin levels in glial cells were relatively higher than in other human and rodent cell lines, some of them described to express medium-high levels of sacsin mRNA in The Protein Atlas, such as HEK293 or HeLa cells ( Figure 1B,C). Surprisingly, C6 and N9 rat microglial cells had higher sacsin levels than the HT22 mouse cell line, of neuronal origin ( Figure 1C). We failed in our attempts to detect sacsin in adult rat neural precursor cells from the dentate gyrus and the subventricular zone (data not shown). The immunocytochemistry of astrocytes and C6 glioma cells showed the expected cytoplasmic distribution of sacsin with some mitochondrial localization ( Figure 1D). These data suggest that sacsin expression is not exclusive to neurons, but also expressed in glial cells.
We next aimed to develop a glial model of ARSACS, deleting sacsin in C6 cells by means of a CRISPR/Cas9 approach ( Figure 1E). We isolated 96 individual clones, of which 42% (40/96) survived and proliferated. Eighteen clones were probed for sacsin protein expression, and around 17% (3/18) of the clones did not express detectable levels of sacsin. We randomly selected one of them for further studies ( Figure 1F).

Sacsin Loss Induces Higher Sensitivity to Oxidative Challenge
Mitochondrial alterations and oxidative stress are common hallmarks in neurodegenerative disorders, and ARSACS is no exception [7,21]. C6 Sacs−/− cells showed a significantly higher level of basal oxidative stress ( Figure 2A). Next, reference C6 cells and the C6 Sacs−/− strain were treated with rotenone, a mitochondrial Complex I inhibitor, to determine whether sacsin loss could undermine their response to oxidative challenges. Incubation with the rotenone for 4 h had similarly mild toxicity in wild type and C6 Sacs−/− cells ( Figure 2B,C; Supplementary Figure S1). Neither MTT or LDH cytotoxicity assays gave indications of significant toxicity in these conditions in both cell lines ( Figure 2B; Supplementary Figure S1A). However, a flow-cytometry-based DAPI exclusion assay suggested a non-significant tendency to higher toxicity in C6 Sacs−/− cells (Supplementary Figure S1B), and these cells showed a stronger decrease in cell size in C6 Sacs−/− cells (forward scatter, FSC-A, Figure 2D) consistent with a higher degree of damage. C6 Sacs−/− cells also show a significantly higher increase in oxidative stress upon rotenone exposure, often accompanied by a slight increase in DAPI staining, which is indicative of membrane damage ( Figure 2E-I). We next aimed to develop a glial model of ARSACS, deleting sacsin in C6 cells by means of a CRISPR/Cas9 approach ( Figure 1E). We isolated 96 individual clones, of which 42% (40/96) survived and proliferated. Eighteen clones were probed for sacsin   Data were analyzed by means of two-way ANOVA, followed by a Tukey post hoc test, *, significant vs. vehicle; #, significant vs. C6 reference strain, p < 0.05.

Sacsin Deletion Leads to Juxtanuclear Accumulation of Glial IFs
Sacsin deletion in neuronal cells induces the accumulation of neurofilament light, medium and heavy polypeptides (NFL, NFM and NFH, respectively), peripherin, αinternexin and vimentin in the juxtanuclear region, with the concomitant depletion of mitochondria in the same region [9,10]. Immunocytochemistry analysis unmasked a similar profile for vimentin, nestin and GFAP in C6 Sacs−/− cells ( Figure 3A-C; Supplementary Figure S2): juxtanuclear accumulation of the three glial intermediate filaments and depletion of mitochondria from this region. Transformation of widefield images by the Nano J Super-Resolution Radial Fluctuations (SRRF) ImageJ plugin [22] showed more condensed IF networks in this juxtanuclear region ( Figure 3B). However, the impact was different for each IF: approximately 40% of cells showed accumulation of GFAP, 60 % of nestin, and 60 % of Vimentin ( Figure 3C). The protein levels of glial IFs were generally higher in C6 Sacs−/− than in reference C6 cells ( Figure 4A), but the increase in nestin levels did not achieve significance ( Figure 4B). Filter trap assays supported the aggregation of these IFs ( Figure 4C), although the increase in vimentin aggregation did not achieve statistical significance ( Figure 4D). No significant alterations were observed in the organization of actin or microtubule networks (Supplementary Figure S3). These data indicate that sacsin deficiency also disrupted the glial IF networks and induced mitochondrial network remodeling; the process occurs in neurons.

Sacsin Deletion Produces Alterations in the Response to Inflammatory Cytokines
Astroglia participates in neuroinflammation, where cytokine signaling plays a key role [23], and IFs are currently considered important scaffolds actively involved in signal transduction [24]. We hypothesized that sacsin deletion could interfere with specific signaling/inflammatory pathways, such as the Signal Transducer and Activator of Transcription 3 (STAT3) pathway, which play key roles in neuroinflammation [25,26]. Reference C6 cells responds poorly to LIF, IL-6 ( Figure 5A) and Tumor Necrosis Factor alpha (data not shown). However, they show STAT3 activation after 20 min of incubation with IL-6 in combination with its soluble receptor IL-6R ( Figure 5A-E), as determined by key STAT3 post-translational modifications, such as Y705 and S727 phosphorylation or K49 acetylation. This activation is accompanied by a significant increase in total STAT3 levels, which is not observed in C6 Sacs−/− cells ( Figure 5B). The lower levels of total STAT3 C6 Sacs−/− cells could explain why we also observe lower levels of post-translational modifications in these cells, as the corresponding PTM/total STAT3 ratios are not significantly different from reference C6 cells ( Figure 5C-E). These results suggest that the impairment in STAT3 signaling is not by direct regulation of STAT3 modifications, but by regulation of STAT3 levels, a hypothesis that is consistent with sacsin's role as a chaperone.

Sacsin Deletion Produces Alterations in the Response to Inflammatory Cytokines
Astroglia participates in neuroinflammation, where cytokine signaling plays a key role [23], and IFs are currently considered important scaffolds actively involved in signal transduction [24]. We hypothesized that sacsin deletion could interfere with specific signaling/inflammatory pathways, such as the Signal Transducer and Activator of Transcription 3 (STAT3) pathway, which play key roles in neuroinflammation [25,26]. Reference C6 cells responds poorly to LIF, IL-6 ( Figure 5A) and Tumor Necrosis Factor alpha (data not shown). However, they show STAT3 activation after 20 min of incubation with IL-6 in combination with its soluble receptor IL-6R ( Figure 5A-E), as determined by key STAT3 post-translational modifications, such as Y705 and S727 phosphorylation or K49 acetylation. This activation is accompanied by a significant increase in total STAT3 levels, which is not observed in C6 Sacs-/-cells ( Figure 5B). The lower levels of total STAT3 C6 Sacs-/cells could explain why we also observe lower levels of post-translational modifications in these cells, as the corresponding PTM/total STAT3 ratios are not significantly different from reference C6 cells ( Figure 5C-E). These results suggest that the impairment in STAT3 signaling is not by direct regulation of STAT3 modifications, but by regulation of STAT3 levels, a hypothesis that is consistent with sacsin's role as a chaperone.

Discussion
Sacsin mRNA is present in most tissues, although its expression levels show some cell specificity (source: The Protein Atlas). In the central nervous system, it was described to have higher levels in Purkinje cells, pyramidal neurons, thalamic and pontine nuclei and reticular formation [4]. However, public transcriptomics data indicated that sacsin is also expressed in glial cells, including astrocytes, Müller glia, oligodendrocyte precursor cells, mature oligodendrocytes and microglia [17,18]. Astrocytes express sacsin RNA levels as high as neurons, especially in younger individuals [17]. Our results confirm that the sacsin protein is present at high levels in astrocytes, and possibly in microglia, since we could detect sacsin in N9 microglial cells.
To the best of our knowledge, scientific studies of sacsin function focused almost exclusively on neurons, and a possible role of glial cells in ARSACS pathogenesis is completely unknown. We developed a new cellular model to investigate the role of sacsin in glial cells based on C6 rat glioma cells (Figure 1). The fact that they are rodent cells will enable comparative studies in mice models of ARSACS and primary astroglial cells from rats. The loss of sacsin in C6 cells produces the accumulation of intermediate filaments-vimentin, nestin and GFAP-in the juxtanuclear area, where there is a concomitant depletion of mitochondria (Figures 3 and 4). These results are consistent with previous reports where sacsin knockout from human HEK-293T (embryonic kidney) and SH-SY5Y (neuroblastoma) cells induced alterations in vimentin and/or neurofilament networks [9,10]. Sacsin is also expressed in keratinocytes and fibroblasts, and ARSACS patients showed skin alterations and lipofuscin deposits [3]. These data suggest that sacsin could have a far-reaching role in the organization and dynamics of different intermediate filaments beyond neurons, and ARSACS could belong to the growing family of IF-pathies, as AxD or GAN [13,15,27].
Although ARSACS, AxD and GAN are different genetic diseases with their own clinical profiles [2,15,28,29], the similarities between some of their symptoms (e.g., pediatric diagnosis, ataxia, dysarthria, nystagmus) and histopathological features (e.g., white matter loss, IF disruption and, at least in ARSACS and GAN mitochondrial dysmotility) are remarkable [4,27,30,31]. The three pathologies have mixed features of neurodegenerative and neurodevelopmental disorders. Vimentin, nestin and GFAP are also typically found in neural precursor cells [32,33], and C6 cells show properties of both astroglia and neural precursor cells [34,35]. Although we were unable to detect sacsin in adult neural precursor cells from the rat dentate gyrus and the subventricular zone (Data not shown), sacsin could regulate the organization of IFs in precursor cells during development, with potential implications for the disease onset. Our results should encourage further studies in these directions.
Oxidative stress and neuroinflammation are central to neurodegenerative diseases, and astrocytes are major players in these processes. IFs are involved in signal transduction as scaffolds for signaling proteins and mitochondrial motility [24,36], and disruption of GFAP alone in AxD models disrupts various signaling pathways [37]. Our results indicate that sacsin deletion impairs the REDOX status of C6 cells and their response of cells to oxidative challenges and cytokines (Figures 2 and 5). The impairment of STAT3 signaling in C6 Sacs−/− cells seems to be mediated by the regulation of STAT3 levels rather than its post-translational modifications ( Figure 5). Bearing in mind the role of sacsin as a chaperone [5,6], it is possible that it contributes directly or indirectly to STAT3 folding or stability. The consequences of disrupting the response of glia to inflammatory cues are difficult to predict, but could be relevant for disease onset and progression, especially if this disruption affects both astroglia and microglia.
In summary, to the best of our knowledge, our study is the first to show that astrocytes express sacsin at the protein level and that their depletion in glial-like cells causes pathological hallmarks of ARSACS similar to those observed in neuronal cells. Sacsin knockout cells showed a dysregulation in their REDOX balance and altered responses to inflammatory cues. These data support a potential role of astroglia, microglia or even neural precursor cells in ARSACS, which should be further analyzed. Future studies should test our findings in postmortem brain tissues from ARSACS patients and existing mouse models of the disease, and analyze how the functions of primary astroglia are disrupted by sacsin loss. Considering the developmental nature of ARSACS, these studies should probably focus on embryonic or perinatal stages.
Supplementary Materials: The following supporting information can be downloaded at: https:// www.mdpi.com/article/10.3390/cells11020299/s1, Figure S1: Toxicity of short treatment of rotenone in C6 strains; Figure S2: Separation of channels corresponding to Figure 3A; Figure S3: Sacsin knockout does not produce gross changes in actin and microtubule networks; Figure S4: Full membranes and molecular weight markers for western blots in Figure 1; Figure S5: Full membranes and molecular weight markers for western blots in Figure 3; Figure S6: Full membranes and molecular weight markers for western blots in Acknowledgments: The authors thank the Advanced Imaging Unit from the BioISI Microscopy Facility (Faculdade de Ciências, Universidade de Lisboa, FCUL, Portugal) for support with bioimaging and flow cytometry. We thank Adelaide Fernandes, Susana Solá (Faculdade de Farmacia, Universidade de Lisboa) and Sara Xapelli (Instituto de Medicina Molecular João Lobo Antunes) for providing protein samples from N9 cells, mouse embryonic and adult neural precursor cells. We also thank Sandra Tenreiro and Carlos Farinha for productive scientific discussions.

Conflicts of Interest:
The authors declare no conflict of interest. The funders had no role in the design of the study; in the collection, analyses, or interpretation of data; in the writing of the manuscript, or in the decision to publish the results.