ALS lymphoblastoid cell lines as a considerable model to understand disease mechanisms

New evidences switch the hypothesis of amyotrophic lateral sclerosis (ALS) from a “neurocentric” to a “multisystemic” or “non-neurocentric” point of view. From 2006, we focused on the study of non-neural cells, patients’ peripheral blood mononuclear cells (PMBCs) and lymphoblastoid cell lines (LCLs). Here, we characterized LCLs of sporadic ALS and patients carrying SOD1, TARDBP and FUS mutations to identify ALS biologically relevant signature, and whether and how mutations differentially affect ALS-linked pathways. Although LCLs are different from motor neurons (MNs), in LCLs we find out some features typical of degenerating MNs in ALS, i.e. protein aggregation and mitochondrial dysfunction. Moreover, different gene mutations otherwise affect ALS cellular mechanisms. TARDBP and FUS mutations imbalance mitochondrial dynamism toward an increased fusion, while sALS and SOD1 mutations mainly affect fission. As regard protein aggregation and/or mislocalization, TARDBP and SOD1 mutations show the presence of aggregates, while FUS mutation does not induce protein aggregation and/or mislocalization. Finally, all LCLs, independently from mutation, are not able to work in a condition of excessive energy request, suggesting that mitochondria from ALS patients are characterized by a significant metabolic defect. Taken together these data indicate that LCLs could be indicated as a valid cellular model in ALS research to study specific pathological pathways or to identify new ones. D is ea se M o de ls & M ec ha ni sm s • D M M • A cc ep te d m an us cr ip t


Introduction
Amyotrophic lateral sclerosis (ALS) is a complex multi-factorial and multi-systemic disorder, characterized by massive motor neuron (MN) loss in the brainstem, spinal cord and motor cortex (Robberecht and Philips, 2013).Multiple pathways might play a critical role on MN survival, including glutamate induced excitotoxicity, endoplasmic reticulum stress, proteasome inhibition, secretion of toxic factors by nonneuronal cells, oxidative stress, axonal disorganization, neuromuscular junction abnormalities, aberrant RNA processing, protein aggregation and mitochondria-mediated damage (Taylor et al., 2016).
Mitochondrial alterations represent a milestone in all the major neurodegenerative diseases, including ALS (Smith et al., 2017;Lin and Beal, 2006).The evidence that mitochondria are impaired in ALS pathogenesis is clear from various studies involving cellular and animal models as well as ALS patients (Tefera et al., 2017;Cacabelos et al., 2016;Allen et al., 2015).Defects in mitochondria Ca 2+ buffering ability and in the activity of the electron transport protein complex were reported in the spinal cord of mutant SOD1 mice during the pre-symptomatic phase of the disease (Abu-Hamad et al. 2017;Nalbandian et al., 2015;Kawamata and Manfredi, 2010).Furthermore, defective mitochondria were identified in mutant FUS and TDP-43 expressing cells and Drosophila, which display aberrant and non-functional mitochondria (Khalil et al., 2017;Onesto et al., 2016;Deng et al., 2015).Abnormal mitochondrial morphology was observed in neurons and peripheral cells of sporadic and familial ALS patients (Rodriguez et al., 2012); moreover, mitochondrial fragmentation has been also documented in ALS cell and animal models (Song et al., 2013).The morphology of the mitochondrial network is influenced by the delicate balance between two opposing events: fusion and fission (Itoh et al., 2013;da Silva et al., 2014), which are highly coordinated and strictly controlled by specific proteins: dynamin-related protein1 (Drp1) and fission 1 (Fis1) for the fission process, and mitofusin 1/2 (MFN1/2) and optic atrophy protein 1 (OPA1) for the fusion process (Gao et al., 2017).
Mitochondrial fragmentation is strictly dependent on changes in the expression of mitochondrial fusion and fission regulators in experimental models expressing ALS-associated mutant SOD1 (Song et al., 2013).Also, neurons expressing ALS-associated mutant TDP-43 and mutant FUS show mitochondrial fragmentation and alterations in fusion and fission regulators (Deng et al., 2015;Wang et al, 2013).
A common hallmark of neurodegenerative diseases is the presence of misfolded protein aggregates in affected regions of the nervous system (Cuanalo-Contreras et al., 2013).Conformational alterations due to protein misfolding may contribute to disease by either a gain of a toxic function, or by the loss of proper biological activity of a protein.Moreover, the formation of abnormal protein aggregates can inhibit indispensable cellular functions (Takalo et al., 2013).

Disease Models & Mechanisms • DMM • Accepted manuscript
switch the ALS hypothesis from a "neurocentric" to a "multisystemic" or "non-neurocentric" point of view.Starting from 2006, our group extensively addressed its attention to the study of non-neural cells with attention to patients' PMBCs and lymphoblastoid cell lines (LCLs) (Cereda et al., 2013;Guareschi et al., 2012;Gagliardi et al., 2010;Cereda et al, 2006;).
The present study aimed, for the first time, a deep characterization of LCLs of sALS and of mutated patients, to identify any biologically relevant molecular signature associated with ALS pathology.Finally, we aimed to highlight whether and how mutations in SOD1, TARDBP, FUS genes differentially affect ALS-linked pathways.

Results
Increasing evidence suggest that peripheral tissues, such as PBMCs and LCLs, gained the scene in ALS research, sharing some pathological features with degenerating MNs (Cereda et al., 2013;Guareschi et al., 2012;Gagliardi et al., 2010;Cereda et al., 2006).Here, we focused our attention on the characterization of LCLs in ALS; we investigated two of the main ALS-related pathogenic mechanisms: accumulation of protein aggregates and mitochondrial dysfunction.

SOD1, TARDBP and FUS mutations showed alterations in soluble protein levels in LCLs.
To verify if SOD1, TARDBP and FUS mutations affected protein expression levels, western blotting (WB) for SOD1, TDP-43 and FUS have been performed on total soluble protein fractions (Fig. 1A, B, C).SOD1 protein expression levels showed an overall reduction in all mutated patients versus healthy controls (Ctrl); only in patients with SOD1 mutation when compared to Ctrl, this reduction was statistically significant (**p<0.01)(Fig. 1A).Similarly, TDP-43 protein expression levels showed a significant decrease (*p<0.05) in SOD1 mutated patients versus Ctrl; no changes were instead reported in sALS, TARDBP and FUS mutated patients (Fig. 1B).As regards the expression of FUS protein, we reported a significant decrease in sALS and TARDBP mutated patients compared to Ctrl (***p<0.001),while no differences were reported for patients carrying mutations in SOD1 and FUS (Fig. 1C).We also evaluated mRNA levels for SOD1, TARDBP and FUS.We reported an increase in SOD1 mRNA levels in SOD1 mutated patients, while no changes were reported for TARDBP and FUS mRNA levels in both sALS and mutated patients (Fig. S1A, B, C).

SOD1, TARDBP and FUS mutations show altered proteins localization in LCLs.
It is widely accepted that the mislocalization of SOD1, TDP-43 and FUS proteins can ultimately account for various ALS pathological signaling (Ido et al., 2011;Ilieva et al., 2009).To disclose the molecular basis and the downstream cascades of the mislocalized and/or aggregated proteins, we fractionated nucleus and cytoplasm and we evaluated SOD1, TDP-43 and FUS by WB and immunofluorescence in LCLs from Ctrl, sALS and mutated patients.

Disease Models & Mechanisms • DMM • Accepted manuscript
WB results stated that SOD1 protein expression levels are reduced in cytoplasm (Fig. 2A) and decreased significantly in nucleus (***p<0.001) of sALS, SOD1 and TARDBP mutated patients versus Ctrl suggesting that SOD1 did not re-localize from cytoplasm to nuclear compartment (Fig. 2A).
In the nuclear compartment, WB results showed a decrease in TDP-43 protein expression levels in all ALS patients compared to Ctrl, which gained the statistical significance in sALS (***p<0.001) and SOD1 (***p<0.001)mutated patients (Fig. 2B).In the cytoplasm compartment, we reported an increase in TDP-43 protein expression levels only in TARDBP mutated patients compared to Ctrl (Fig. 2B), nevertheless the statistical significance was not reached.
About FUS protein expression levels, no great changes were observed in the nuclear compartment except for SOD1 mutated patients which showed a significant decrease compared to Ctrl (**p<0.01)(Fig. 2C).
Instead, in the cytoplasm compartment we reported a significant reduction in FUS protein expression levels in sALS patient (*p<0.05)compared to Ctrl (Fig. 2C).

SOD1, TARDBP and FUS mutations show protein re-localization and aggregation in LCLs.
In Fig. 3A, SOD1 immunostaining showed a cytoplasm homogenous signal distribution in Ctrl, sALS and TARDBP mutated patients.Instead, in SOD1 mutated patients we evidenced the presence of aggregates at cytoplasm/perinuclear level; also, the intensity profile analysis (along the white line in Fig. 3A) showed an elevated peak right outside the nucleus, pointing at aggregate formation.
In TARDBP mutated patients, TDP-43 immunostaining analysis also highlighted the presence of cytoplasm aggregates (see the white line and the intensity profile analysis in Fig. 2B), thus, suggesting an abnormal protein localization (Fig. 3B), according to WB results (Fig. 2B).
Data from immunostaining analysis showed no cytoplasm mislocalization and/or aggregation of FUS protein (Fig. 3C) in all analyzed patients.Additional immunofluorescence images are reported in supplementary figures Fig. S2, S3 and S4.

Mitochondrial morphological changes in LCLs of SOD1, TARDBP and FUS mutated patients
As is known, mitochondria are highly dynamic organelles which constantly fuse and divide in response to diverse stimuli; moreover, their morphology is strictly related to changes in mitochondrial and cellular functions, and an imbalance in mitochondrial dynamics plays a critical role in various neurodegenerative diseases (Itoh et al., 2013;da Silva et al., 2014).
Here, we investigated mitochondrial morphology by means transmission electron microscopy (TEM).As expected, mitochondria from Ctrl were normal with a longitudinal shape and densely packed cristae (Fig. 4A).sALS patients showed mainly longitudinal mitochondria, even if with signs of degeneration, i.e. increased number of vacuoles (see yellow arrows) and reduced dimensions (Fig. 4B).In SOD1 mutated patients, mitochondria were smaller and mainly round shaped; they presented an increased vacuolization Disease Models & Mechanisms • DMM • Accepted manuscript (yellow arrows) and cristae are distorted by holes with empty matrix (red arrows) (Fig. 4C).TARDBP mutated patients also showed signs of mitochondrial degeneration.Mitochondria are bigger, probably because of a swelling associated with disarrangement of cristae and partial or total cristolysis.The longitudinal shape is lost or severely compromised (Fig. 4D).Finally, in FUS mutated patients, in addition to the signs of degeneration observed in all patients, mitochondria appear giant (megamitochondria) in which the regular spacing of cristae is lost (Fig. 4E, F).
Additional TEM images are reported in supplementary figures Fig. S5.

SOD1, TARDBP and FUS mutations showed differences in mitochondrial dynamic in LCLs.
It is widely accepted that fusion, fission, and trafficking may contribute to mitochondrial dysfunction in neurodegenerative diseases (Burte et al., 2015;Itoh et al., 2013).Here, we characterized LCLs according to changes in mitochondrial dynamism, to pinpoint possible mutation-related differences.We analyzed optic atrophy 1 (OPA1), mitofusin (MFN1) as proteins involved in the fusion process and dynamin-related protein 1 (Drp1) as a protein controlling the fission process.
We did not observe differences in OPA1 protein expression levels in the mitochondrial fraction of ALS patients compared to controls (Fig. 5A); of note FUS mutated patients showed a slight but not significant decrease in OPA1 protein expression levels when compared to Ctrl.About MFN1, which together with OPA1 acts in the fusion process, reported a significant increase in protein expression levels in TARDBP mutated patients (*p<0.05)(Fig. 5B).
Finally, about Drp1, a decrease in protein expression levels was reported in the mitochondrial fractions of TARDBP and FUS mutated patients (Fig. 5C).These data agree with TEM results which showed an altered mitochondrial morphology, i.e. giant or megamitochondria, in TARDBP and FUS mutated patients which suggests an imbalance of the dynamics versus fusion and an increased connectivity among mitochondria (TEM images see Fig. 4D-E-F).Moreover, we observed a significant increase in Drp1 protein expression levels in cytoplasm of sALS patients compared to Ctrl (*p<0.05)(Fig. 5C), suggesting an increased mitochondrial fragmentation as evidenced by the presence of small mitochondria (TEM image see Fig. 4C).
Moreover, the increase in Drp1 in sALS patients and the reduced dimensions of mitochondria could be associated to an alteration of the mitochondria-associated membranes (MAMs) which, it was accepted to regulate mitochondrial dynamism throughout Drp1 (Polo et al., 2017;Watanabe et al. 2016).
Analysis of mitochondrial/cytoplasmic partition of Cyt-C shows increased mean cytoplasmic level, expressed as a percent of the total cytoplasmic and mitochondrial level of the protein, in FUS and TARDBP mutated samples, while no changes were reported in control samples, sALS and SOD1 mutated samples (Fig. 5D).

LCLs.
To assess mitochondrial respiration, the Seahorse XF-24 Bioanalyzer was used to simultaneously measure the oxygen consumption rate (OCR) and the glycolytic flux by measure of the extracellular acidification rate (ECAR) in LCLs of sALS and mutated patients.

Oxygen consumption rate (OCR)
In Fig. 6A we reported normalized OCR measured Seahorse XF-24 Bioanalyzer in LCLs of healthy controls, sALS and mutated patients.The basal measurements of OCR were recorded before drug injection, then oligomycin, FCCP and rotenone were injected sequentially to measure the key forms of mitochondrial function.Basal cellular oxygen consumption rate (bOCR) increased significantly in all patients carrying mutations in SOD1, TARDBP and FUS, while in sALS patients bOCR is like that measured in control subjects (Fig. 6B).To determine whether the increase in the oxygen consumption rate (OCR) was mitochondrialdependent, we added rotenone, a complex I inhibitor, to LCLs.As reported in Fig. 6C, mitochondrial oxygen consumption (mOCR) increased significantly in all mutated patients and in sALS compared to Ctrl.Instead, no significant differences were reported in both the ATP synthase-coupled mOCR (i.e.measured after the addiction of oligomycin) and ATP synthase-uncoupled mOCR (calculated by subtracting mOCR plus oligomycin from mOCR plus rotenone) (Fig. 6D, E).The spare respiratory capacity (SRC), which refers to the mitochondrial ability of cells to generate energy under conditions of great demand, was diminished in all patients (mutated and sporadic) compared to healthy controls (Fig. 6F), of note SRC reach negative values in SOD1 and FUS mutated patients.

Extracellular acidification rate (ECAR)
Simultaneously to the measurement of OCR, the Seahorse XF-24 Bioanalyzer also analyzes ECAR, an index of cellular glycolytic flux.Subsequent injections of glucose, oligomycin and 2-deoxy-glucose (2-DG) were carried out to determine the glycolytic flux ECAR specific, the glycolytic capacity and the glycolytic reserve in ALS patients and control LCLs (Fig. 7A).We reported a down-regulation of the glycolytic flux ECAR specific (of about 32%) in SOD1 mutated patients (Fig. 7B), while no significant differences were reported in the glycolytic capacity and glycolytic reserve (Fig. 7C, D).

Discussion
Peripheral tissues have been progressively gained the scene in the research on neurodegenerative disorders, and peripheral blood has been indicated as a potential source of biomarkers in ALS (Cereda et al, 2013;Gagliardi et al, 2010;Cereda et al, 2006;Nardo et al., 2011).Even though PBMCs may represent a useful tool in the research on ALS pathogenesis to study genetic mechanisms or to unravel whether experimental stress may alter the transcriptional state of the cells, they also display limits.First, PBMCs cannot be grown and stored as a cell line and therefore do not show the right handling flexibility needed

Disease Models & Mechanisms • DMM • Accepted manuscript
for adequate modelling.For this reason, we focused on EBV-immortalized lymphoblastoid cell lines that, while sharing many of the PBMCs' feature, can also be grown in culture.Here, we illustrate for the first time LCLs from sALS patients to categorize potential biological molecular signature associated with the pathology, i.e. protein aggregation and mitochondrial dysfunction.We also take advantage of the availability of LCLs from ALS patients carrying mutations in SOD1, TARDBP, FUS genes to investigate whether and how mutations affect different ALS-linked pathways.
Data from LCLs of sALS patients showed no substantial changes in total protein expression levels of SOD1 and TDP-43, while total FUS protein expression levels decreased significantly, thus suggesting the absence of alterations in total protein expression.A different picture appeared when we fractionated nuclear and cytoplasmic compartments.In sALS patients SOD1, TDP-43 and FUS protein expression levels decreased in both nucleus and cytoplasm when compared to Ctrl.Nevertheless, no protein re-localization and no signs of protein aggregation were reported.Even if protein mislocalization, particularly TDP-43 and FUS, plays a key role in ALS pathology (Boeynaems et al., 2016;Dormann and Haass, 2011), in LCLs from sALS patients we did not report nucleus-cytoplasmic transport defects and/or protein aggregation.The reason of this could be locate, at least in LCLs from sALS patients, in the absence and/or low levels of elements, i.e. mutations and nuclear import regulators, which negatively affect the delicate balance of the nuclear import and export pathways.About mitochondria, in LCLs from sALS patients, they maintained a "normal" longitudinal shape but appeared smaller, this could be related to the increased expression of the fission protein Drp1, which helps to create new mitochondria in order to preserve a healthy mitochondria population (Twig et al., 2008).
Finally, the oxygen consumption rate (OCR measurements) did not differ significantly from healthy Ctrl, further indicating that mitochondria from LCLs of sALS patients are only marginally affected by the pathology.
LCLs from SOD1 mutated patients showed that, despite an increase in SOD1 mRNA levels, protein expression levels both in total soluble fraction and in nuclear/cytoplasm extracts were reduced compared to healthy controls.A similar result was already obtained by our group in PBMCs of sALS patients (Gagliardi et al., 2010), suggesting that either alterations in mRNA stability or the presence of misfolded/aggregated proteins, might cause an increase in SOD1 protein expression levels in the insoluble fraction with a concomitant decrease in protein expression levels in the soluble fraction.Mutated SOD1 has been reported to form intracellular aggregates that may exert a toxic function (Proctor et al., 2016).In LCLs from SOD1 mutated patients, we observed cytoplasm/perinuclear aggregates, suggesting protein recruitment in cytoplasmic aggregates following mutations in SOD1 gene.Since aggregates are a hallmark in ALS pathology, the presence of aggregates in SOD1 mutated LCLs further corroborate the idea that LCLs could be an interesting device for ALS research.About mitochondrial dysfunction we did not find significant changes in fusion and fission protein expression levels.Nevertheless, TEM analysis revealed alterations in Disease Models & Mechanisms • DMM • Accepted manuscript mitochondrial morphology, the presence of smaller and round shaped mitochondria account for favoring the fission over the fusion pathways, as previously observed in cellular and mouse models expressing mutant SOD1 (Magrane et al., 2014).Moreover, the increase in mitochondrial respiration and the decrease in SRC referred to difficulties in the mitochondrial ability of cells to face the increase in energy demand.The altered mitochondrial metabolism was also confirmed by the decrease in the glycolytic ECAR.
In LCLs from mutant TARDBP patients, SOD1, TDP-43 and FUS protein expression levels in RIPA soluble extracts and mRNA expression remained unchanged compared to Ctrl.In fractionated nucleus and cytoplasm, SOD1 maintained its "normal" cytoplasm localization even if its expression is reduced.No changes were also reported in FUS, thus suggesting that the expression levels of these two proteins are not affected by the presence of mutations in TARDBP gene.According to literature (Oberstadt et al., 2017) in TARDBP mutated LCLs, TDP-43 was depleted from the nucleus and accumulated in the cytoplasm, where the protein formed extra-nuclear round-shaped toxic aggregates (see immunofluorescence images).Even if this result need further and deeper examination, our findings established that LCLs recapitulates ALS pathological features in TARDBP mutated patients, and can be indicated as a compelling model of the disease.Mitochondria from TARDBP mutated patients showed abnormal clustering and elevated levels of MFN1; these findings suggested that mutations in TARDBP gene may alter mitochondrial dynamics, we hypothesize an imbalanced dynamism which has a bias towards the fusion pathway, further confirmed by the formation of giant mitochondria containing electrondense globes and vacuoles.However, our results partially disagree with data reported in literature; in cultured motor neurons as well as in the intact sciatic nerve of living mice but also in fibroblasts from patients harboring the A382T mutation in TDP-43, where a fragmented mitochondrial network and an impaired mitochondrial transport was reported (Magrane et al., 2014;Onesto et al., 2016;Xu et al., 2010).These discrepancies may be primarily due to the different model adopted: i.e.LCLs from TARDBP mutated patients versus transgenic mouse models or fibroblasts from ALS patients (Magrane et al., 2014;Onesto et al., 2016;Xu et al., 2010).Finally, about mitochondrial metabolism, TARDBP mutated LCLs refers to an increase in mitochondrial respiration and to a reduction in the SRC, while no changes were reported in the glycolytic flux.Our results allow us to speculate that the abnormal mitochondrial morphology could be caused by alterations in components (fusion elements) of the mitochondrial dynamics machinery, while intrinsic mitochondrial bioenergetics defects seem to be not involved in this process.
No significant alterations in LCLs of FUS mutated patients, i.e. mRNA up or down-regulation, protein expression levels and protein mislocalization, were reported.We could hypothesize that in LCLs FUS mutation is not so relevant to eventually induce a gain and/or a loss of function, which in turn determine protein aberrant localization, as reported by other authors (Ederle and Dormann, 2017; Suzuki and Matsuoka, 2015).What is noteworthy is the presence of megamitochondria, as an index of increase mitochondria damage, thus suggesting a direct effect of FUS mutation on mitochondria.

Disease Models & Mechanisms • DMM • Accepted manuscript
Finally, cytochrome c augmented levels in the cytoplasm may be suggestive of mitochondrial damage and the release of cytochrome c from the mitochondrion to the cytoplasm is an early step in cellular death via apoptotic pathway (Danial and Korsmeyer, 2004), therefore apoptotic pathways seem to be recruited both in FUS and TARDBP mutated samples but not in sALS nor in SOD1 patients.
In conclusion, patient-derived LCLs display features typical of degenerating MNs in ALS, mainly protein aggregation and mitochondrial dysfunction.Consequently, LCLs could be suggested as a valid cellular model of ALS to study specific pathological pathways or possibly to identify new ones.Moreover, LCLs obtained from both sporadic and mutated ALS patients have the advantage of expressing physiological levels of mutant genes, thus preventing any trouble due to the over-production of the related protein.
An important aspect arisen from this study is how the different gene mutations affect the different analyzed cellular mechanisms.TARDBP and FUS mutations mainly imbalance mitochondrial dynamism by increasing fusion process, while sALS and SOD1 mutation mainly affect fission.Thus, at a first glance, we could hypothesize to identify two patient categories: one including patients carrying mutations in RNA binding proteins and the other including sALS and SOD1 mutated patients.However, even if TARDBP and FUS mutations show similarities in terms of mitochondrial dynamism they are different as regard protein aggregation and mislocalization.In fact, we showed a similar pathway between TARDBP and SOD1 mutations, with the presence of protein aggregation; while FUS mutation did not induce protein aggregation and/or mislocalization.As regard the mitochondrial metabolism, all LCLs, independently from mutation, are not able to work in a condition of excessive energy request, suggesting that mitochondria from ALS patients are all characterized by a significant metabolic defect.Morphological differences observed between mutated and sALS patients we suggested that may account for differences in mitochondrial metabolism.In particular, mitochondrial anomalies reported in sALS patients seem to be limited to mitochondrial morphology with no evident defect in metabolism according to what reported in Rodríguez et al. (2012).In mutated patients, TARDBP and FUS, the increased fusion of mitochondria probably concern mainly damaged mitochondria.We can suggest that mutations seem to implicate difficulties in handling and/or repair of damage mitochondria, thus pointing to the observed alteration in mitochondria functionality.
Finally, it is important to stress that when analyzing mutated patients with a rare disease, even if the results did not reach the statistical significance, anyway we should pay close attention to the obtained data because of the small number of available samples and the importance of each single data.On the other hand, since patient's blood is relatively easy to obtain and LCLs can easily be settled, it is possible to collect a considerable number of samples specific for each mutation, thus allowing comparison with statistical power between pathological signatures of different mutation, possibly leading to patient stratification on a molecular basis.

Patients' enrolment
Patients affected by ALS were enrolled at the "C.Mondino" National Neurological Institute, in Pavia.ALS diagnosis was made according to the revised El Escorial Criteria (Brooks et al., 2000).Sex-and age-matched healthy volunteers, free from any pharmacological treatment and pathology, were recruited at the Transfusion Centre of the IRCCS Policlinico S. Matteo Foundation in Pavia.Eleven ALS patients and four Ctrl were immortalized; of the eleven patients 4 were sporadic, 3 harbor mutations in SOD1 gene, 2 harbor mutations in TARDBP and 2 are mutated in FUS gene as reported in table 1.

EBV-immortalization
Peripheral blood mononuclear cells (PBMCs) were obtained after informed consent and in accordance with guidelines approved by the Ethical Committee of the C. Mondino National Neurological Institute (Protocol n°375/04 -version 07/01/2004).
PBMCs were isolated from peripheral venous blood by Histopaque®-1077 (Sigma-Aldrich) following manufacturer's instructions.An aliquot of PBMCs were immortalized into LCLs via Epstein-Barr Virus (EBV) infection (Dreser et al., 2017;Scammel et al., 1997), with minor modifications.Briefly, 5x10^6 cells were resuspended in RPMI 1640 medium (Sigma-Aldrich), supplemented with 20% fetal bovine serum (FBS) (Sigma-Aldrich), 0,3 mg/L L-glutamine, 5% penicillin-streptomycin and cyclosporine A (Sigma-Aldrich).EBVmix, prepared according to Caputo and collaborators, plus RPMI 1640 with cyclosporin A, was added to the cells.Cells were incubated at 37°C in a humidified atmosphere with 5% CO2 for a week.The medium was then changed and cells were left in incubation until clusters of growing cells appear.Lymphoblastoid cells were maintained and expanded in RPMI 1640 medium, supplemented with 20% FBS, 0,3 mg/L L-glutamine, 5% penicillin-streptomycin, humid incubator at 37°C in a with 5% CO2.At need cells were pelleted, washed with 1X PBS and further processed as required.
Cells viability for LCLs was assessed by Trypan Blue assay.Briefly, cells suspension was mixed with 0.4% Trypan Blue (Sigma-Aldrich) and counted with the automated cells counter TC20 (BioRad) to evaluate the percentage of live cells, which was about 75-80%.
Protein-containing supernatants were collected and stored at -80°C.Subcellular fractionation: LCLs subcellular fractionation was performed according to Schreiber and colleagues (Schreiber et al., 1989), with some modifications.Cells were first re-suspended in an ice-cold Disease Models & Mechanisms • DMM • Accepted manuscript hypotonic lysis buffer (10 mM HEPES, pH 7.9, 10 mM KCl, 0.1 mM EDTA, 1 mM dithiothreitol (DTT), 0.5 mM phenylmethylsulfonyl fluoride (PMSF), 1% PIC), and incubated on ice.10% Nonidet NP-40 (Fluka) was added and samples were centrifuged at the maximum speed.The cytoplasm proteins were collected and stored at -80°C until the use.The pellet was re-suspended in an ice-cold hypertonic nuclear extraction buffer (20 mM HEPES, pH 7.9, 0.4 M NaCl, 1 mM EDTA, 1 mM DTT, 1 mM PMSF fluoride, 1% PIC), and incubated on ice with agitation.The nuclear extracts were centrifuged at the maximum speed and the nuclear proteins were collected and frozen at -80°C until the use.
Mitochondrial extraction: Mitochondrial fraction (Mt) was isolated using the Cytochrome C Releasing Apoptosis Assay Kit (Abcam), according to manufacturer instructions.Briefly, cells were centrifuged at 600 x g at 4°C, pellets were washed in ice-cold 1X PBS and centrifuged again.Supernatant were discarded and cells re-suspended in 1X Cytosol Extraction Buffer Mix (CEB), plus 1 mM DTT and 1% PIC.Samples were incubated on ice and then centrifuged at 10.000 x g at 4°C.Supernatant was collected as cytosolic fraction.
The remaining pellet was re-suspended in Mitochondrial Extraction Buffer Mix (MEB), plus 1 mM DTT and 1% PIC, and stored as mitochondrial fraction.
The protein content of each extracted fraction was then quantified by BCA assay (Sigma-Aldrich).

RNA extraction
Total RNA from sALS and mutated patients as well as from control subjects cells was extracted with Trizol® reagent (Invitrogen) according to the manufacturer's recommendations.RNA quality and quantity was determined using NanoDrop spectrophotometer (Celbio).

Reverse transcription
Total RNA was reverse transcribed using the iScript cDNA synthesis kit (BioRad) according to the manufacturer's protocol.The reaction mix was incubated for 5 min.at 25°C, for 30 min.at 42°C and for 5 min.at 85°C.The cDNA samples were stored at -20 °C.

Real Time PCR
Sybr-Green primers were designed with identical annealing and melting temperatures using Primer Express Software (Applera, Italy): SOD1 (FW GGTCCTCACTTTAATCCTCTATCCAG; RW CCAACATGCCTCTCTTCATCC);

Fig. 2 .
Fig. 2. SOD1, TARDBP and FUS mutations showed protein re-localization in LCLs.Representative immunoblots for nuclear and cytoplasm SOD1, TDP-43 and FUS protein expression levels in Ctrl, sALS, SOD1,TARDBP and FUS LCLs (Panels A, B and C).Histone H1 was used for nuclear sample normalization; GAPDH was used cytoplasm sample normalization.Densitometric analysis of WB data are performed by ImageJ software.analysis was carried out on n=4 Ctrl, n=4 sALS, n=3 SOD1, n=2 TARDBP and n=2 and FUS LCLs.are means ± S.E.M. and one-way analysis of variance (ANOVA) followed by Dunnet Multiple Comparison Test as a post-hoc test.*p<0.05 in C (Cytoplasm), **p<0.01 in C (Nucleus) and ***p<0.001 in A and B (Cytoplasm), all relative to Ctrl.

Fig. 3 .
Fig. 3. SOD1, TARDBP and FUS mutations showed protein re-localization and aggregation in LCLs.Representative immunostaining of SOD1, TDP-43 and FUS in LCLs of sALS, SOD1, TARDBP and FUS mutated patients.Immunostaining SOD1 (Panel A) showed a cytoplasm homogeneous distribution in Ctrl, sALS and TARDBP mutated patients; in SOD1 mutated patients aggregates at cytoplasm/perinuclear level were by the white arrows.TDP-43 immunostaining (Panel B) showed presence of cytoplasm aggregates (see white line and intensity profile analysis) in TARDBP mutated patients; Ctrl, sALS and SOD1

Fig. 4 .
Fig. 4. Mitochondria morphology of Ctrl, sALS, SOD1, TARDBP and FUS mutated LCLs.Representative images of transmission electron microscopy analysis from Ctrl (Panel A), sALS (Panel B), SOD1 (Panel C), TARDBP (Panel D) and FUS (Panel E, F) mutated patients.Ctrl mitochondria were normal with a longitudinal shape and densely packed cristae (A).Mitochondria from sALS patients were longitudinal with increased vacuolization (yellow arrows) and reduced dimensions (B).Mitochondria from SOD1 mutated patients were smaller and round shaped, with increased vacuolization (yellow arrows) and paucity of cristae (red arrows) (C).TARDBP mutated patients showed bigger and round shaped mitochondria with disarranged cristae (D).Mitochondria from FUS mutated patients are completely degenerated megamitochondria in these patients (E, F).

Fig. 5 .
Fig. 5. SOD1, TARDBP and FUS mutations showed differences in mitochondrial dynamic in LCLs.Representative immunoblots for mitochondrial fusion proteins (OPA1 and MFN1) and for mitochondrial and cytoplasm fission proteins (Drp1) protein expression levels in Ctrl, sALS, SOD1, TARDBP and FUS LCLs (Panels A, B and C).COX4 was used for mitochondrial sample normalization; GADPH was used for cytoplasm sample normalization.Representative immunoblot for cytochrome c release was reported in panel D. Densitometric analysis of WB data are performed by ImageJ software.Statistical analysis was carried out on n=4 Ctrl, n=4 sALS, n=3 SOD1, n=2 TARDBP and n=2 and FUS LCLs.Data are means ± S.E.M. and one-way analysis of variance (ANOVA) followed by Dunnet Multiple Comparison Test as a post-hoc test.*p<0.05 in B and C (Cytoplasm), all relative to Ctrl.