Abstract
TAR deoxyribonucleic acid-binding protein 43 (TDP-43) is a key protein in the pathogenesis of amyoptrophic lateral sclerosis (ALS) and frontotemporal lobar degeneration (FTLD). Recent studies suggest that mutations in the TDP-43 coding gene, TARDBP, as well as variations in TDP-43 protein expression levels may disrupt the dynamics of stress granules (SGs). However, it remains unclear whether the pathogenetic effect of the TDP-43 protein is exerted at the cytoplasmic level, through direct participation to SG composition, or at nuclear level, through control of proteins essential to SG assembly. To clarify this point, we investigated the dynamics of SG formation in primary skin fibroblast cultures from the patients with ALS together with the A382T mutation and the patients with ALS and healthy controls with wild-type TDP-43. Under stress conditions induced by sodium arsenite, we found that in human fibroblasts TDP-43 did not translocate to the SGs but instead contributed to the SG formation through a regulatory effect on the G3BP1 core protein. We found that the A382T mutation caused a significant reduction in the number of SGs per cell (P < 0.01) as well as the percentage of cells that form SGs (P < 0.00001). Following stress stimuli, a significant decrease of viability was observed for cells with the TDP-43 A382T mutation (P < 0.0005).
We can therefore conclude that the A382T mutation caused a reduction in the ability of cells to respond to stress through loss of TDP-43 function in SG nucleation. The pathogenetic action revealed in our study model does not seem to be mediated by changes in the localization of the TDP-43 protein.
Introduction
Amyotrophic Lateral Sclerosis (ALS) is a fatal neurodegenerative disease of the upper and lower motor neurons that results in progressive paralysis and death, usually occurring 1 to 5 years from the onset of symptoms (1). Alongside the genetic and environmental risk factors that have been shown to contribute to the development of ALS, several potential stress factors such as oxidative stress, mitochondrial dysfunction, excitotoxicity and endoplasmic reticulum stress are thought to be involved in the underlying pathogenetic process (2,3). Disease heterogeneity is a well-known characteristic of ALS, but almost all affected patients have in common the formation of cytoplasmic aggregates containing the TAR deoxyribonucleic acid-binding protein 43 (TDP-43) (4,5). It has now been well established that TDP-43 is a ribonuclear protein (RNP) involved in several aspects of RNA metabolism including transcription, alternative splicing, pre-mRNA stability and mRNA transport (6–8). Because mutations in the encoding gene, TARDBP, have been found in patients with familial ALS (FALS), TDP-43 is widely considered to be a key protein in the pathogenesis of ALS as well as other neurodegenerative diseases (9). However, little is known about the mechanisms by which TARDBP mutations exert their pathogenetic effect. Several studies investigating the toxic increment of the protein in the cytoplasm have brought to light results compatible with a gain of function in the cytoplasm (10,11). Pathogenic mutations of TDP-43 could promote cytotoxic accumulation of the protein either by increasing its stability compared to the wild-type (WT) (12) or by inducing structural variations that favour the formation of cytoplasmic aggregates (13). An alternative hypothesis is represented by a loss of function of the TDP-43 protein. It has been shown in studies of TDP-43 knockout or knockdown animal models that down-regulation of TDP-43 leads to neuronal death (14–19). Because it has been shown that TDP-43 binds more than 6000 RNA targets, including those encoding for proteins involved in neurodegenerative diseases or that mount protective response to stress and neuronal death (20,21), it is conceivable that one or more of the aforesaid proteins may fail to properly exert their physiological functions.
Under stress conditions, TDP-43 has been found localized in specific cytoplasmic structures known as stress granules (SGs) (22). These transient cytoplasmic complexes embrace many structural and functional properties that fit in well with what we know so far about ALS pathogenesis. SGs form in the cytoplasm of cells under stress conditions and have a central role in the storage and sorting of transcriptionally inactive mRNA (23). Indeed, it is their presence that allows cells to block or delay the activation of apoptotic pathways, repair the damage and survive the stress (24,25). SG assembly starts with the phosphorylation of the eukaryotic initiation factor (eIF2a). This phosphorylation event inhibits mRNA translation through depletion of the eIF2-met-tRNA-GTP ternary complex, thus permitting the ribonuclear protein TIA-1 to bind to the 48S complex instead of the ternary complex. This in turn promotes polysome disassembly and the consequent recruitment of mRNAs to SGs (26). The composition of SGs varies according to the type of cell and stress stimulus (27).
Several proteins, if overexpressed, may independently generate SG nucleation but two proteins in particular seem to have a prominent role in nucleation, maintenance and dissociation of SGs: RasGAP SH3 domain binding protein1 (G3BP1) and T-cell intracellular antigen (TIA-1) (28–30). Knockdown of G3BP1 or its mutations severely impairs SG assembly (28). It has recently been demonstrated that the depletion of G3BP1 disturbs normal interactions between SGs and processing bodies (PBs) (31). The assembly and disassembly of SGs is a highly controlled process that is critical for cell survival. Several studies have suggested that prolonged SG activity may lead to persistence of the stressful environment with prolonged translational repression, all of which can have potentially deleterious effects (32,33). One study in particular found that SGs inhibited apoptosis by reducing the production of reactive oxygen species (ROS) and that this function was regulated by G3BP1 and ubiquitin-specific protease 10 (USP10) (25). A number of studies have recently investigated the role of TDP-43 in the mechanisms underlying assembly and disassembly of SGs. Dewey et al. (13) found that cells overexpressing mutant constructs of TDP-43 form significantly larger SGs and incorporate mutant TDP-43 earlier into SGs in contrast to WT TDP-43 that forms more stress granules over time, but with approximately the same granule size. These authors propose that mutant TDP-43 alters stress granule dynamics and in this way contributes to the progression of TDP-43 proteinopathies. In another study using the same experimental model, it was demonstrated that mutant TDP-43 promotes cytoplasmic self-aggregation and positioning in the SGs through a specific link with TIA-1 (34). Taken together, these observations suggest that the localization of TDP-43 in the granules might be a precondition in the pathogenetic process mediated by TDP-43. Although the composition and morphology of SGs vary according to the types of cells and stressor methods used, it is still unclear how these variables influence the results since all the studies performed so far describe cell models and stressor methods in which TDP-43 is localized in SGs. TDP-43 is not a mandatory component of SGs. For example, it has not been found in SGs of neural cells (35) and in the HEK293 cell line. Moreover, it has been shown that TDP-43 translocates to SGs when stress is osmotic (sorbitol) but not when stress is induced by SA (13).
Studies more in line with a hypothesis of loss of function have demonstrated that the expression of two SG core proteins, G3BP1 and TIA-1, are under the control of TDP-43 and that a reduced expression of TDP-43 such as that obtained after silencing of the protein by interfering RNA, is able to delay the formation of SGs under oxidative stress conditions produced by SA (36). It is particularly relevant that these authors replicate these findings in a lymphoblast cell model originating from the patients with ALS positive for the R361S mutation of the TDP-43 protein.
Here we describe the results of a study in which we used primary fibroblast cell lines from human skin of patients with ALS and controls to investigate the role of the A382T mutation of TDP-43 in SG dynamics. In cultures of cells with the A382T mutation, we found a significant reduction in the number of cells that formed SGs compared to WT cells from healthy individuals and the patients with ALS. We also observed a significant decrease in cell viability after stress treatment.
We can conclude that in our experimental model the A382T mutation reduces the ability of cells to respond to stress through loss of TDP-43 function in SG nucleation. The pathogenetic action mediated by this mutation does not seem to be influenced by changes in the localization of the TDP-43 protein.
Results
In skin fibroblasts, TDP-43 is essential to SG formation but does not become part of their structure
Cellular localization of TDP-43 after stress induced by sodium arsenite (SA) in human fibroblasts and NSC-34 cells. After SA treatment, TDP-43 immunoreactivity was localized exclusively in the nuclear compartment of fibroblasts from both Controls (A) and the patients with ALS with the A382T mutation of TDP-43 protein and never in SGs. NSC-34 cells were used for positive control of translocation of TDP-43 to SGs.HuR: Alexa488, green labelling; TDP-43: Cy3, red labelling; 60 min: time of SA treatment; Scale bar: 30µm.
TDP-43 silencing and its impact on SG formation. (A) qPCR of TDP-43 mRNA levels in control (siControl) and silenced (siTDP-43) human fibroblasts. The bar graph shows the mean (±SEM) of three independent readings. * = P <0.02. (B) Western blot analysis showed a significant reduction of TDP-43 protein expression in silenced fibroblasts. (C) TDP-43 silenced fibroblasts (siTDP-43) showed a reduction in the ability to form SGs compared to those transfected with control siRNA . Fibroblasts were immunostained with HuR, after treatment with SA for 30 min. Scale bar: 30µm. (D) The bar graph represents the mean percentage (± SEM) of cells with SGs in TDP-43 silenced and control fibroblasts.
Impaired stress granule formation in fibroblasts of the patients with ALS with the A382T mutation
In order to test the hypothesis that mutations in the TARDBP gene may interfere with mechanisms of response to cellular stress, we performed comparisons between primary fibroblast cultures of the patients with ALS with the A382T mutation and cultures of primary fibroblasts obtained from control individuals. The following parameters were evaluated: a) the percentage of cells that formed SGs, b) the size of the SGs and c) the number of SGs per cell.
TARDBP A382T mutation dysregulated stress granule formation. (A): SG formation in TDP-43 A382T mutated and control fibroblasts with Wild Type TDP-43. In untreated cells, no SGs were observed. After 30 min treatment with 0.5 mM SA, stress granule formation was observed in both groups but with a significant difference in the number of SGs per cell and the percentage of SG forming cells. These differences became more evident after 60 min of treatment. When SA was removed, SGs gradually disappeared in both control and TDP-43 mutated fibroblasts. SGs were identified by indirect immunofluorescence using an anti-HuR specific antibody. Scale bar: 50µm. B and C represent the mean percentage (± SEM) of cells with SGs after 30 (B) and 60 (C) min of 0.5 mM SA treatment and subsequent recovery. Comparisons were made between two TDP-43 mutated (A382T) and three WT control fibroblast lines. The data are representative of 8 experiments in which at least 60 cells per condition were observed. *P < 0.01; **P < 0.00001.
In order to clarify whether the decrease in the percentage of cells that form SGs was associated with the A382T mutation of TARDBP rather than different factors attributable to the disease, we repeated the tests using two additional cultures of fibroblasts obtained from the patients with ALS with a WT genotype at the TARDBP, FUS and C9orf72 loci. In these cultures, the ability of cells to form SGs was similar to that observed in cells of healthy controls, thereby indicating that the observed disruption in the dynamics of SGs was specifically attributable to the mutant TDP-43 protein (Supplementary Material, Fig. S2).
Under stress, the A382T mutation causes downregulation of the G3BP1 protein
(A) Fibroblasts with the TDP-43 A382T mutation showed significant down regulation of G3BP1 mRNA levels compared to cells with WT TDP-43 (two fibroblast lines from the patients with ALS and two from healthy controls). The data (mean ± SD) are representative of three independent real time reactions. *P < 0.01. (B) Western blot analysis showed increased G3BP1 protein levels after SA treatment in WT fibroblasts from healthy individuals and the patients with ALS while they were unchanged in cells with the A382T mutation. (C) TDP-43 A382T cells showed reduced viability after oxidative stress. MTT assay revealed different cell viability between TDP-43 mutant and control fibroblasts following 60 min of SA treatment (70.1% vs 78.4%, respectively). After recovery of basal conditions, the control cells displayed an increased viability while TDP-43 A382T mutated cells showed decreased viability. This difference became more apparent after 120 min of recovery (59.9% vs 88.4%). The data are representative of six experiments and are expressed as percentage of the untreated cells. *P < 0.02; **P < 0.03; ***P < 0.0005.
After oxidative stress ALS cells with the A382T mutation show significant reduction in viability compared to control cells
SGs are considered as structures with anti-apoptotic functions (25). Disruption in the mechanisms underlying the formation, maintenance and dissolution of SGs may cause cell death (25,36,38–40). To test whether the reduced formation of SGs observed in cells with the TARDBP mutation correlated with an increased rate of cell death, we performed viability tests in fibroblast cell lines with the A382T mutation and those of WT controls. The cells were analyzed immediately after 60 min of stress induced by SA and then after 1 and 2 h of recovery of basal conditions (Fig. 4C).
Cell viability, using MTT assay, was significantly reduced in ALS cells with the A382T mutation compared to WT control cells after 60 min of treatment with SA (70.1% vs 78.4%; P <0.02). A similar difference was observed for cultures in which the medium was replaced to restore basal conditions following SA treatment. In these cultures, after 60 min of recovery time, a significant difference in cell viability could still be observed (66.1% vs 80.8%; P <0.03). As expected, after 2 h of reconditioning, the control cultures displayed a marked increase in viability whereas in A382T cells the viability remained significantly lower (59.9% vs 88.4%; P <0.0005. Overlapping results for cell viability were also obtained using another test method (AlamarBlue; Supplementary Material, Fig. S5). These data show that the reduction of the number of SGs in cells that carry the A382T mutation is an event able to produce a significant increase in cell death and that this pathogenic impact is maintained over time despite restoration of normal culture conditions. Overall, these data support the hypothesis that the A382T mutation of the TARDBP gene causes a loss of function in the ability of cells to respond to stress through the formation of SGs.
Discussion
Because many variants of the TARDBP gene have been associated with ALS, it becomes important to understand whether missense mutations induce different properties in the mutated protein respect to those of the WT TDP-43 protein, since such knowledge could hold the key to the pathogenetic mechanism of TDP-43.
A382T is the most common mutation of the TARDBP gene. It was initially described in two familial patients with ALS of French origin (17) and subsequently identified in large studies carried out on patients of Italian and French origin (41,42). However, it is in Sardinia that a founder effect accounts for the highest frequency in the world, involving approximately 30% of patients with ALS (43,44) and 20% of familial cases of frontotemporal dementia (FTD) (45). The pathogenic nature of A382T has been demonstrated in several studies (17,41,43,44,46–49). However, this missense mutation has shown an incomplete penetrance (60% at 70 years) with an increased risk for males (44). Recently, this mutation was reported to cause reduction of endoplasmic reticulum Ca2 + signalling since it was found associated with an increased translocation of TDP-43 in cytoplasm after endoplasmic reticulum (ER) stress induced by the calcium modifying drug thapsigargin (49). The A382T mutation has also been analyzed for its potential effect on protein turnover but the results reported in the literature are contradictory. One study performed on the large majority of ALS-associated mutations, found that A382T together with seven other mutations extended the half-life of the protein and caused cytotoxicity as a result of accumulation, cleavage and insolubility in the cytoplasm (48). Conversely, another more recent study found faster protein turnover for the A382T mutation whereas subcellular analysis of TDP-43 distribution did not reveal significant differences in comparison to the distribution of the WT protein (47). Another recent study of the TDP-43 distribution in primary fibroblasts of patients with ALS found that cell lines with the A382T mutation had significantly increased quantities of the protein in both the nucleus and cytoplasm compared to controls and cells from the patients with ALS with the SOD1 mutation (50).
It has been well established that TDP-43 has an important effect on the dynamics of SG assembly and dissociation (27,32,51) but it remains unclear whether it exerts a pathogenetic action at the cytoplasmic level by disrupting the correct assembly/disassembly of SGs through direct participation to SG composition, or at nuclear level through control of proteins essential to SG formation. In our experimental approach, we used primary fibroblast cell lines from human skin to investigate the role of the A382T mutation of TDP-43 in SG assembly mechanisms.
Very few studies have been performed on primary cells originating from the patients with ALS, and this is the first to address the role of TDP-43 in SG formation. The advantages and disadvantages of using primary skin fibroblasts as a study model for neurodegenerative disease have already been extensively discussed (52,53). In our study, the main advantage of these cells was represented by the availability of a study model with an unaltered genomic structure and a protein expression pattern very similar to the original. More specifically, in studies of the pathogenetic role of mutations, these cells are able to avoid the potential artefacts caused by protein overexpression, typically produced by plasmid constructs. TDP-43, when overexpressed, has been shown to be toxic in a variety of cellular and/or animal models (10,11,54,55). Moreover, overexpression of TDP-43 is capable of disturbing the physiological dynamics of SG assembly and maintenance since it influences the expression of SG nucleation factors (31,36,51). Such overexpression may cause SGs to form also in the absence of stress stimuli (34).
In primary fibroblasts obtained from skin biopsies, we found that SA was able to determine the formation of numerous SGs positive for two of their essential components: TIA-1 and HuR, but not TDP-43. Although in our study TDP-43 did not translocate to SGs, it still resulted to have a role in their formation since TDP-43 silenced cells had a significantly reduced ability to form SGs. Furthermore, the A382T mutation caused a significant reduction in the number of SGs per cell as well as the percentage of cells that form SGs. Additionally, cell death under stress conditions was significantly increased in cells with the TDP-43 mutation compared to WT cells. Based on these findings, we can conclude that the significant increase observed for cell death can directly be attributed to the inability of these cells to compensate SA-induced oxidative stress with the formation of SGs. Interestingly, throughout the entire process, TDP-43 was never found located in SGs or other cytoplasmic aggregates. Therefore, this is the first study to have analyzed alterations in SG dynamics in a cellular model in which TDP-43 does not translocate to SGs. Fibroblasts with the A382T mutation displayed the same phenotype as fibroblasts silenced for WT TDP-43, clearly showing loss of protein function exclusively within the nucleus.
Recently, it has been demonstrated that TDP-43 is involved in SG regulation via nucleating factor G3BP1 and that depletion of TDP-43 or its R361S mutation are able to alter the dynamics of SGs by down regulating G3BP1 mRNA levels (36). Our study demonstrated the same mechanism in human fibroblasts. Differences in G3PB1 expression were observed under stress but not basal conditions. This may be explained by the different behaviour of the two mutations in their function towards G3BP1 and suggest that A382T TDP-43 mutation induces lower pathogenicity patterns than R361S mutation. Alternatively, this difference may be the result of different levels of G3BP1 expression in lymphoblasts and fibroblasts. Therefore, the results of our search for a possible mechanism through which the A382T mutation of TDP-43 exerts its pathogenic role clearly point to a loss of TDP-43 nuclear function in the quantitative control of the G3BP1 mRNA transcript.
Ours is the second study to demonstrate that a TDP-43 mutation interferes with the correct dynamics of SGs by exerting a dysregulatory effect on G3BP1 mRNA levels. However, considering the complexity of TDP-43 interactions it is likely that G3BP1 is not the only protagonist. Although we are not aware of the mechanism by which TDP-43 controls G3BP1 levels, our cellular model clearly indicates that it occurs within the nucleus. The two mutations, R361S and A382T, are only a few amino acid residues away from each other and may perhaps delineate a region critical for G3BP1 transcript levels. Different to the A382T and R361S mutations, which reduce cell response to acute stress and their ability to form SGs, the G294A, A315T, Q343R and G348C mutations have been reported to increase the tendency of TDP-43 to aggregate with proteins of SGs, such as TIA-1. For these mutations, the pathological protein would seem to acquire a tendency to translocate to the cytoplasm and aggregate within SGs, thereby increasing their number (34) or increasing their size (13,56). Based on these studies, the increased size of SGs would be the main factor responsible for SG dysregulation induced by overexpressed or mutated TDP-43. This led the authors to hypothesize that the association of TDP-43 with SGs could be a triggering event that after chronic stress eventually leads to irreversible pathological aggregation (13,34). However, caution is warranted, since their cell culture experiments were based on overexpression of TDP-43 (20).
It is well known that TDP-43 is not an essential component of SGs and that its presence in the granules is dependent on the type of cell and stress stimulus used (27). The mechanisms that regulate this difference in behaviour of TDP-43 remain unknown as well as the role that mutations may have in this process. Our study shows that even in the absence of translocation of the protein to SGs, mutant TDP-43 is capable of producing a pathogenic effect on the ability of cells to generate SGs under stress. However, we cannot exclude different consequences, including morphological changes in cells that respond to stress stimuli by translocation of TDP-43 to SGs. Further studies aimed at identifying all the factors by which TDP-43 controls the formation of SGs in different types of cells and tissues, will be critical to our understanding of this pathogenetic pathway leading to neurodegenerative disease.
Materials and Methods
Cell culture and treatments
Primary fibroblast cell cultures were derived from skin biopsies taken with informed consent from (i) The patients with ALS with heterozygous missense mutation (TARBDP-A382T; n = 2), (ii) The patients with ALS with a wild type (WT) genotype at the TARDBP, FUS and C9orf72 loci (n = 2) and (iii) wild type (WT) healthy control subjects (n = 3). The study was approved by the local ethics committee. Cells were grown in high-glucose DMEM supplemented with 20% (vol/vol) foetal bovine serum, 1% penicillin/streptomycin (10,000 units penicillin and 10 mg streptomycin per mL in 0.9% NaCl) (all from Sigma Aldrich, ST). In order to induce stress granule assembly, cells were exposed to 0.5 mM SA (Sigma Aldrich) for 30 and 60 min at 37 °C. Then the medium was replaced to permit cells to recover for further 30–120 min. Finally, cells were fixed for immunocytochemistry or processed for an RNA extraction or viability assay.
The immortalized motoneuronal NSC-34 cell line was used to confirm TDP-43 localization in SGs. The cells were maintained in Dulbecco’s modified Eagle’s medium (DMEM) supplemented with 10% foetal calf serum (FCS) and 0.5% penicillin/streptomycin solution and subcultured every 5 days.
Immunocytochemistry
Cells were grown on glass coverslips and fixed after treatment in 4% w/v paraformaldehyde for 15 min, permeabilized with cold methanol for 5 min and 0.2% Triton X-100 in PBS for 20 min. Different fixation conditions were tested to optimize the immunostaining (1% paraformaldehyde and 2% paraformaldehyde, 2% sucrose).
Cells were then incubated for 2h at room temperature with different primary antibodies. To identify SGs, a mouse anti HuR (1:500, Santa Cruz) and/or goat anti TIA-1 antibody (1:300, Santa Cruz) were used in association with an anti-human TDP-43. We tested three different TDP-43 antibodies produced against various TDP-43 sequences (TDP-43 N-terminal, 1:500, Proteintech; 10782-2-AP; TDP-43 C-terminal 1:500, Proteintech; 12892-1-AP and TDP-43,N-terminal, 1:300, SAB3500236 Sigma-Aldrich) and we obtained the equivalent immunostaining. All antibodies were diluted in phosphate buffered saline (PBS) containing 3% normal donkey serum. The relevant species-specific donkey secondary antisera, conjugated with either Alexa-488 (emitting in green; 1:200) and or cyanin 3.18 (yellow/red; 1:300) (Jackson Immunoresearch Laboratories, West Grove, PA) were used (1h) to reveal immunoreactivity of the primary antisera. Nuclei were counterstained with Hoechst 33342. Coverslips were mounted with Glycerol/PBS (1:1).
Negative controls were routinely performed for each experiment, incubating the samples with non-immune serum and then with appropriate secondary antibody. Imaging was carried out using an Olympus BX41 fluorescence microscope.
Transfection
Cells at 60–80% confluency were transfected using Lipofectamine RNAi-Max (Invitrogen Life Technologies), with 125 pmol of TDP-43 siRNA (Silencer Select, Ambion) or siRNA control (BLOCK-iT™ Alexa Fluor® Red Fluorescent Control, Life Technologies), both diluted in Opti-MEM (Gibco) according to the manufacturer's protocol. After 24h, the medium was replaced with fresh regular media. 72 h post-transfection cells were harvested for western blotting or qPCR to verify silencing.
qPCR
Cells grown in monolayer were quickly rinsed with PBS for 1 min. Total RNA was isolated using Trizol Reagent (Invitrogen) according to the Manufacturer’s instructions. From 1 µg of RNA, the corresponding cDNA was obtained using the SuperScript® III First-Strand Synthesis System (Invitrogen). Quantification of target mRNA was obtained using the LightCycler® FastStart DNA MasterPLUS SYBR Green I (Roche Diagnostics). Relative quantification with respect to the Beta-actin reference gene was performed using Rel Quant software (Roche Diagnostics). Data were normalized to the untreated WT control cells. The set of primers used for qPCR is described in Reference 36.
Western blot
72 h after transfection, cells were lysed in 2% sodium dodecyl sulphate (SDS). The protein concentration was obtained by the DC Protein Assay (Biorad); loading buffer (125mM Tris-HCl, pH 6.8, 10% glycerol, 0.04% SDS, 4% β-mercaptoethanol, 0.2% bromophenol blue) was subsequently added and samples boiled for 5 min. 15 μg of protein were run on 10% SDS-polyacrylamide gel and blotted onto polyvinylidene fluoride (PVDF) membrane (Hybond-P, Amersham). Membranes were blocked with 5% non-fat dry milk 1h at RT and then incubated with the primary antibody (anti TDP-43 N-terminal, 1:1000, Proteintech or anti G3BP1, 1:2500, Proteintech) overnight at 4 °C; membranes were then incubated with horseradish-peroxidase-conjugated antirabbit IgG (1:5000, Invitrogen, Life Technologies) 1h at RT. After washing, protein bands were detected with a chemiluminescent substrate (Biorad). Filters were reprobed with an anti tubulin (1:1000, Abcam) or anti GAPDH (1:1000 Millipore MAb 374) antibodies to confirm equal protein loading. The experiment was repeated three times.
Quantification of cells forming SGs and SG size
For quantification, ten areas of each coverslip were selected. SGs were identified by TIA-1 and/or HuR immunolabelling and cells were scored as positive when they had at least two foci of a minimal size of 0.75 µm2. Cell counting was performed in a double-blind on 40x magnification.
Stress granule size was measured using the ImageJ software. The area of 80 SGs (ranging from 0.75 to 5 µm2) was measured in at least 13 cells per condition.
Cell viability assays
Fibroblasts from TDP-43 mutated patients with ALS and controls were plated in parallel at 50,000/well in 24-well plates. After 24 h, cells were treated with 0.5mM SA (1h) and recovered for 1 and 2h. To verify cell viability, a colorimetric assay based on the MTT labeling reagent (Sigma) was used. Spectrophotometrical absorbance was measured using an EnVision®Multilabel Plate Reader at 570 nm. Data were expressed as percentage of the respective untreated samples (100%). AlamarBlue cell viability test (Invitrogen, Life Technologies) was used to confirm MTT results, the resulting fluorescence intensity was read (560EX nm/590EM). The assays were performed according to the manufacturer’s instructions.
Statistical analysis
Statistically significant changes were calculated using the Student’s t test.
Supplementary Material
Supplementary Material is available at HMG online.
Acknowledgements
We are grateful to Anna Maria Koopmans for professional writing assistance.
Conflict of Interest statement. None declared.
Funding
This study was financially supported by the “Comitato EUGENIO for ALS reasearch”; Sardinian Regional Government (Sardinia POR FSE 2007-1013’ grant N. G31J12002240002); Fondazione Banco di Sardegna (grant N. U1037.2014/AI.919.MGB).
