ALS-linked FUS mutations confer loss and gain of function in the nucleus by promoting excessive formation of dysfunctional paraspeckles
Mutations in the FUS gene cause amyotrophic lateral sclerosis (ALS-FUS). Mutant FUS is known to confer cytoplasmic gain of function but its effects in the nucleus are less understood. FUS is an essential component of paraspeckles, subnuclear bodies assembled on a lncRNA NEAT1. Paraspeckles may play a protective role specifically in degenerating spinal motor neurons. However it is still unknown how endogenous levels of mutant FUS would affect NEAT1/paraspeckles. Using novel cell lines with the FUS gene modified by CRISPR/Cas9 and human patient fibroblasts, we found that endogenous levels of mutant FUS cause accumulation of NEAT1 isoforms and paraspeckles. However, despite only mild cytoplasmic mislocalisation of FUS, paraspeckle integrity is compromised in these cells, as confirmed by reduced interaction of mutant FUS with core paraspeckle proteins NONO and SFPQ and increased NEAT1 extractability. This results in NEAT1 localisation outside paraspeckles, especially prominent under conditions of paraspeckle-inducing stress. Consistently, paraspeckle-dependent microRNA production, a readout for functionality of paraspeckles, is impaired in cells expressing mutant FUS. In line with the cellular data, we observed paraspeckle hyper-assembly in spinal neurons of ALS-FUS patients. Therefore, despite largely preserving its nuclear localisation, mutant FUS leads to loss (dysfunctional paraspeckles) and gain (excess of free NEAT1) of function in the nucleus. Perturbed fine structure and functionality of paraspeckles accompanied by accumulation of non-paraspeckle NEAT1 may contribute to the disease severity in ALS-FUS.
KeywordsAmyotrophic lateral sclerosis (ALS) Fused in sarcoma (FUS) NEAT1 Paraspeckle
(Fluorescent) in situ hybridization
Acid guanidinium thiocyanate-phenol-chloroform extraction
Amyotrophic lateral sclerosis
Fragments per kilobase of transcript per million mapped reads
Nuclear Paraspeckle Assembly Transcript 1
Nuclear localization signal
Soluble nuclear extract
Amyotrophic lateral sclerosis (ALS) is a severe adult-onset neurodegenerative disease affecting motor neurons. More than 20 genes have been linked to familial (f)ALS, and many of them encode RNA-binding proteins, including FUS . Over 50 mutations in the FUS gene have been found in fALS and sporadic (s)ALS patients, the vast majority being heterozygous mutations with autosomal dominant inheritance; most of them affect the nuclear localization signal (NLS) of the protein [31, 33, 34, 65]. Mutations in the FUS gene cause an aggressive, sometimes juvenile-onset disease .
The histopathological hallmark of ALS-FUS is partial mislocalisation of this predominantly nuclear protein to the cytoplasm in neurons and glial cells of the spinal cord and formation of FUS-positive inclusions [23, 31, 65]. It should be noted, however, that significant FUS mislocalisation is seen only in a subset of ALS-FUS cases and only in a subset of neurons in the latter cohort [23, 29, 39], suggesting that altered nuclear function(s) of mutant FUS can drive pathological changes sufficient to cause the disease. Indeed, FUS mutations having only a minor effect on its nuclear import, such as R521G(H), are detrimental in in vitro and in vivo models [47, 49, 51, 66]. In addition, ALS-linked FUS mutations outside its NLS have been identified , and they also cause pathological cellular phenotypes [45, 46]. Finally, most recent studies from mouse models of FUS pathology revealed that mutant FUS is able to cause neurodegeneration in the absence of cytoplasmic pathology and even significant mislocalisation, strongly suggesting that nuclear gain of toxic function by mutant FUS represents an important disease mechanism [15, 37]. Despite significant progress in our understanding of cytoplasmic gain of function by mutant FUS , nuclear mechanisms of mutant FUS toxicity are still poorly understood.
Paraspeckles are RNA granules formed in the nuclear interchromatin space, in close proximity to splicing speckles . Paraspeckles contain several core and multiple secondary proteins that are assembled on a scaffold long non-coding RNA (lncRNA) NEAT1 [12, 42, 50, 58]. The NEAT1 gene produces two transcripts, NEAT1_1 and NEAT1_2; only the longer isoform, NEAT1_2, is capable of forming paraspeckles . Established functions of paraspeckles include sequestration of some RNAs and transcription factors, and thus regulation of gene expression, in response to certain stimuli such as proteasomal inhibition and viral infection [1, 10, 24, 26, 73]. Most recently, roles for paraspeckles in enhancing microRNA biogenesis and regulation of mitochondrial function have been identified [27, 67]. Dysfunction of paraspeckles or their components is implicated in the increasing number of human diseases, including cancer, autoimmune and neurodegenerative disorders .
FUS is involved in multiple processes related to cellular RNA metabolism . The protein possesses a low-complexity prion-like domain responsible for its ability to phase-separate and to be recruited into RNA granules in the nucleus or cytoplasm [6, 56]. Although normal and mutant FUS are incorporated into a variety of RNA granules and can even nucleate RNA granules when accumulated [3, 18, 54, 72], the paraspeckle is the only type of physiological RNA granule which requires FUS as a structural component. FUS is defined as an essential paraspeckle protein, in that its knockdown eliminates paraspeckles [42, 55].
Paraspeckles likely play an important role in ALS pathogenesis. Indeed, paraspeckle proteins are enriched in the pool of proteins affected by ALS-causative mutations . Although healthy mammalian neurons lack NEAT1_2 expression and hence paraspeckles in vitro and in vivo [43, 53], de novo paraspeckle formation is typical for spinal motor neurons of sALS and fALS patients and as such can be considered a hallmark of the disease [44, 53]. Previously, we reported pathological aggregation of a core paraspeckle protein, NONO, in cellular and mouse models of FUS pathology as well as in the spinal cord of ALS-FUS patients . Since both FUS and NONO are required to build paraspeckles, formation of these RNA granules was expected to be disrupted in ALS-FUS. However, this assumption has not been tested experimentally.
In the current study, using novel cell lines expressing endogenous mutant FUS, patient fibroblasts and human post-mortem tissue, we have identified excessive assembly of dysfunctional paraspeckles as a novel nuclear pathology caused by FUS mutations.
Materials and methods
Generation of cell lines with targeted modification of the FUS gene
Guide RNA target sequences within the FUS gene were identified using Feng Zhang lab’s Target Finder (https://zlab.bio/guide-design-resources). Respective forward and reverse oligonucleotides were annealed and cloned into pX330-U6-Chimeric_BB-CBh-hSpCas9 (pX330) vector (Addgene) according to the previously described protocol . SH-SY5Y human neuroblastoma cells were split onto a 35 mm dish at 50–60% confluency one day prior to transfection. Equal amounts of plasmids (3.6 μg each) carrying upstream and downstream gRNA target sequence (or one plasmid for FUS knockout) were delivered into cells by calcium phosphate transfection. After 24 h, cells were resuspended at ~ 10–20 cells/ml and plated onto 10 cm dishes. Single-cell derived clones were expanded and screened by immunofluorescence and PCR. For sequencing of the edited portion of FUS gene, the PCR product corresponding to the edited allele was cloned into Zero Blunt® TOPO® vector (Life Technologies), and at least four colonies were sequenced. Primers used for PCR screening and TOPO® cloning: ΔNLS lines: 5’-TGGGGACAGAGGTGGCTTTG-3′ and 5’-CCTTCCTGATCGGGACATCG-3′; FUS KO: 5’-ACCATTTGAGAAAGGCACGCT-3′ and 5’-CACGGATTAGGACACTTCCAGT-3′.
Cell line maintenance, differentiation, transfection and treatments
SH-SY5Y neuroblastoma cells were maintained in 1:1 mixture of Dulbecco’s Modified Eagle’s Medium and F12 medium supplemented with 10% fetal bovine serum (FBS), penicillin-streptomycin and glutamine (all Invitrogen). Cells were transfected in 24-well plates with plasmid DNA (200 ng/well), poly(I:C) (Sigma, 250 ng/well) or siRNA (AllStars Negative Control from Qiagen or NEAT1 Silencer Select®, n272456 from Life Technologies) using Lipofectamine2000. Final concentrations of MG132 and sodium arsenite (both Sigma) were 1 μM and 0.05 mM, respectively. Cells were treated with actinomycin D for 3 h to induce nucleolar caps. Plasmids for expression of GFP-tagged FUS variants are described elsewhere . Plasmids for NONO and SFPQ expression were prepared by inserting respective ORFs into pEGFP-C1 vector. The protocol for obtaining human fibroblasts from a control subject and a patient with FUS P525L mutation [9, 36] was approved by the University of Palermo Review Board (prot.07/2017). Human fibroblasts were cultured under the same conditions as SH-SY5Y cells. Primary murine hippocampal cultures were prepared and transfected as described .
Immunocytochemistry, RNA-FISH and proximity ligation assay (PLA) on cultured cells
Cells were fixed on coverslips with 4% paraformaldehyde for 15 min, washed with 1xPBS and permeabilized in cold methanol (or 70% ethanol in case of RNA-FISH). For immunostaining, coverslips were incubated with primary antibodies diluted in blocking solution (5% goat serum/in 0.1% Tween 20/1xPBS) for 1 h at RT or at 4 °C overnight. Secondary Alexa488- or Alexa546-conjugated antibody was added for 1 h at RT. For RNA-FISH, commercially available NEAT1 probes (Stellaris® Quasar® 570-labelled against 5′ or middle segment of human NEAT1, Biosearch Technologies) and Cy5-labelled oligo(dT)30 probe (for polyA+ RNA detection, Sigma) were used as per standard Biosearch Technologies protocol. For colocalisation studies of NEAT1 and NONO, RNA-FISH was followed by 30 min incubation in anti-NONO antibody and Alexa488-conjugated secondary antibody. PLA was performed using Duolink® In Situ Orange Starter Kit Mouse/Rabbit (DUO92102, Sigma) using anti-FUS (mouse monoclonal, Santa Cruz, sc-47711) antibody in combination with rabbit anti-NONO or SFPQ (A301-322A, Bethyl) antibody. To detect FUS and NONO interaction in paraspeckles, 1:10,000 antibody dilutions were used. Fluorescent images were captured using BX61 microscope equipped with F-View II camera and processed using CellF software (all Olympus). Quantification of paraspeckle numbers/NEAT1-positive area and PLA results was performed using ‘Analyze particles’ tool of ImageJ software. Images were prepared using Photoshop CS3 or PowerPoint 2010 software.
Analysis NEAT1_2 and MALAT1 extractability was performed as described . Briefly, one set of samples lysed in QIAzol (Qiagen) was heated at 55 °C for 10 min and the second set of samples prepared in parallel was left at room temperature. RNA was extracted from both sets as per standard QIAzol protocol. Fold extraction of NEAT1_2 or MALAT1 was calculated as a ratio between levels of these RNAs, measured by qRT-PCR, in heated versus non-heated samples. For obtaining nuclear soluble extract (SNE), a protocol by Werner and Ruthenberg was followed . For standard gene expression and miRNA analysis by qRT-PCR, total RNA was extracted from cells using QIAzol with a heating step (55 °C for 10 min). First-strand cDNA synthesis was performed using random primers (or oligo(dT) primers for NEAT1_1 analysis in SNE) and Superscript IV (Invitrogen) or miScript II RT (Qiagen). Quantitative RT-PCR was performed as described ; to measure miRNA levels, forward miRNA-specific primer was used in combination with the universal reverse primer (unimiR). All primer sequences are given in Additional file 1: Table S1. For RNA-Seq, total RNA was extracted using PureLink total RNA extraction kit (Life Technologies) and possible DNA contamination was removed using RNase free DNase kit (Qiagen). RNA-Seq analysis was performed at School of Biosciences Genomics Research Hub. Libraries were prepared using the TruSeq stranded mRNA kit (Illumina) and single-end sequencing was performed on Illumina NextSeq500 (read length: 75 bp; coverage ~ 20 million reads/sample). Reads were aligned to the human reference genome (GRCh38) using STAR , and FPKM values were obtained using DESeq2 . Reads were viewed in the IGV browser .
Nuclear-cytoplasmic fractionation was performed according to a published protocol (REAP) . Total cell lysates and cytoplasmic fractions were prepared for Western blot by adding 2xLaemmli buffer followed by denaturation at 100 °C for 5 min. SDS-PAGE and detection of proteins were carried out as described elsewhere . Quantification of Western blots was done using Image J and protein levels were normalised to beta-actin.
The following commercial primary antibodies were used: FUS full protein (rabbit polyclonal, 11,570–1-AP); FUS N-terminus (rabbit polyclonal, Abcam, ab84078; aa. 1–50); FUS C-terminus (Bethyl, A300-294A; aa. 500–526); p54nrb/NONO (rabbit polyclonal C-terminal, Sigma); SFPQ (rabbit monoclonal, ab177149, Abcam; rabbit polyclonal, A301-322A, Bethyl); beta-actin (mouse monoclonal, A5441, Sigma). Antibodies were used at 1:500–1:1000 dilution for all applications unless stated otherwise.
Analysis of human tissue samples
Human spinal cord paraffin sections from clinically and histopathologically characterised ALS cases and neurologically healthy individuals were obtained from the MRC London Neurodegenerative Diseases Brain Bank (Institute of Psychiatry, Kings College, London) and Sheffield Brain Tissue Bank. Consent was obtained from all subjects for autopsy, histopathological assessment and research in accordance with local and national Ethics Committee approved donation. Human spinal cord sections for immunohistochemistry were 7 μm thick. Immediately after antigen retrieval in citrate buffer, slides were washed several times in 2xSSC prepared with DEPC-treated water. Slides were incubated with NEAT1 (5′ segment) Stellaris® probe diluted in hybridisation buffer (10% formamide/2xSSC; 5 μl probe in 200 μl buffer per slide under a 24 × 60 mm coverslip) in a humidified chamber at 37 °C overnight. Nuclei were stained with DAPI. Paraspeckles were analysed using the same microscope and camera as above (× 100 magnification). For RNAscope® ISH analysis, Hs-NEAT1-long (411541) probe (Advanced Cell Diagnostics) was used according to manufacturer’s instructions. SFPQ immunohistochemistry on spinal cord sections was performed using SFPQ IHC-00304 antibody (Bethyl) as described earlier .
Quantifications and statistics
N in all cases indicates the number of biological replicates. On all graphs, error bars represent SEM. Statistical analysis was performed using GraphPad Prism 6 software. Mean values of biological replicates were compared using appropriate tests (stated in figure legends). Significance levels are indicated with asterisks (*p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001).
Generation and characterisation of cell lines expressing endogenous mutant FUS.
The requirement of FUS for paraspeckle assembly limits the use of cell models with FUS overexpression or knockdown. Moreover, patient derived pluripotent cells and neurons differentiated from these cells were also unsuitable for this study since both of these cell types lack paraspeckles [8, 43]. Therefore we chose to generate human neuroblastoma SH-SY5Y cell lines expressing endogenous mutant FUS.
Analysis by qRT-PCR showed a small increase of FUS mRNA in FUS ΔNLS lines, consistent with the ability of FUS to autoregulate its own levels, and confirmed the absence of WT FUS mRNA in the homozygous lines (Fig. 1e). Western blot with an antibody against FUS N-terminus showed normal levels of FUS protein (Fig. 1f). As expected, an antibody specific to the extreme C-terminus of FUS (aa.500–526) detected no FUS protein in the homozygous lines (such as ΔNLS1) and its decreased levels in the heterozygous lines (such as ΔNLS2) (Fig. 1f). Western blot also confirmed the absence of detectable FUS protein in the FUS KO line (Fig. 1f).
We noticed that FUS redistribution to the cytoplasm was very modest in the heterozygous FUS ΔNLS lines. In contrast, homozygous cells displayed dramatic FUS mislocalisation, with the border between the nucleus and cytoplasm in the FUS-immunostained cells often indistinguishable (Fig. 1g). Subcellular fractionation confirmed almost normal retention of FUS in the nucleus in the heterozygous lines (Fig. 1h). This pattern is different from the predicted two-fold increase in the cytoplasmic mislocalisation in the homozygous as compared to heterozygous FUS ΔNLS lines and suggests that the presence of non-mutated, nuclear localised FUS partially protects mutant FUS from mislocalisation. Consistent with previous literature, mutant FUS was readily recruited to cytoplasmic stress granules induced by oxidative stress (Additional file 1: Figure S2).
Thus, we established cell lines with mild and severe mislocalisation of endogenous FUS to the cytoplasm suitable for the analysis of paraspeckles.
Mutant FUS induces the accumulation of NEAT1 isoforms and excessive paraspeckle formation
We next used NEAT1 RNA-FISH to image paraspeckles in the lines generated. In our analysis, we included three homozygous and three heterozygous (hereafter ΔNLS_ho and ΔNLS_het, respectively) FUS ΔNLS lines as well as FUS KO cells.
Paraspeckle assembly is directly correlated with the expression of the longer NEAT1 isoform, NEAT1_2, whereas NEAT1_1, although recruited to paraspeckles, is not required for their integrity . NEAT1_2 was recently reported to be “semi-extractable” meaning that heating or shearing steps are required to efficiently extract it by conventional AGPC-based methods . In order to measure NEAT1_2 levels accurately, we included a heating step during RNA extraction with QIAzol. NEAT1_2, quantified by qRT-PCR, was upregulated in ΔNLS_het lines thus providing grounds for the enhanced paraspeckle assembly; however, it was similarly upregulated in ΔNLS_ho lines (Fig. 2d). NEAT1_1 completely overlaps with NEAT1_2 in its 5′ end and cannot be measured separately by qRT-PCR in total RNA samples. NEAT1_1 but not NEAT1_2 is polyadenylated. RNA-Seq analysis of poly(A)-captured RNA which only detects NEAT1_1 showed that this isoform was also significantly elevated in FUS ΔNLS lines (Fig. 2d). SFPQ and NONO are known to regulate NEAT1_2 levels and hence paraspeckle formation . However, mRNA and protein levels as well as distribution of both proteins were similar in WT and FUS ΔNLS lines (Additional file 1: Figure S3B-D). Interestingly, FUS KO cells, which lack paraspeckles, displayed normal NEAT1 levels (Fig. 2d), suggesting that NEAT1 accumulation was caused by the presence of mutant FUS and not by compensatory NEAT1 upregulation in response to paraspeckle disruption. Consistent with the finding that NEAT1 is accumulated in FUS ΔNLS lines, the NEAT1-repressed mRNA ADARB2  was found to be dramatically downregulated in these cells (Fig. 2e), while NEAT1 knockdown was able to elevate ADARB2 both in WT and FUS ΔNLS cells (Additional file 1: Figure S3E).
FUS itself does not stabilise NEAT1_2 and instead is involved in paraspeckle maturation downstream of NEAT1_2 synthesis [42, 70]. We next investigated whether exogenously expressed mutant FUS could restore paraspeckle assembly in FUS KO and ΔNLS_ho lines. Cells were transfected with plasmids to express GFP-tagged FUS WT, FUS ΔNLS (predominantly cytoplasmic), and ALS-linked FUS mutants R524T and R518K (predominantly nuclear) . Overexpression of all FUS variants led to the appearance of bright NEAT1-positive foci in the majority of FUS KO and ΔNLS1_ho cells (Fig. 2f-h), which coincided with the disappearance of paraspeckle precursors (Fig. 2g). There were no significant differences between FUS variants in their ability to nucleate paraspeckles (Fig. 2h) – despite the fact that in cells expressing GFP-tagged FUS ΔNLS, the level of ectopic protein in the nucleus was much lower than in cells expressing other FUS variants (Fig. 2f). This suggests that a certain threshold for nuclear FUS level is required for paraspeckle assembly and that FUS mutants can maintain the formation of visible paraspeckles.
To summarise, the presence of endogenous levels of mutant FUS is accompanied by NEAT1 upregulation. This leads to increased paraspeckle numbers in cells with sufficient nuclear levels of FUS. However, more pronounced FUS mislocalisation, seen in cells expressing two mutant copies of FUS, disrupts paraspeckles.
Mutant FUS is deficient in maintaining the integrity and functionality of paraspeckles
FUS has been shown to be responsible for low NEAT1_2 extractability (“semi-extractability”) . Weakened interaction of FUS with SFPQ/NONO implied its reduced binding to NEAT1_2 in FUS ΔNLS lines. We tested whether NEAT1_2 extractability is altered in cells expressing mutant FUS by comparing typical RNA extraction using QIAzol with a parallel sample subjected to an additional heating step. We first confirmed that heating increases NEAT1_2 extractability ~ 3.5-fold in WT neuroblastoma cells, whereas extractability of another lncRNA, MALAT1, is not affected (Additional file 1: Figure S4B). In FUS KO cells that do not form paraspeckles, NEAT1_2 was almost fully extractable (e.g. its semi-extractability was lost - heated/non-heated ratio close to 1) (Additional file 1: Figure S4B). We further found that NEAT1_2 extractability was significantly increased not only in ΔNLS_ho lines almost lacking visible paraspeckles but also in ΔNLS_het lines, albeit to a lesser extent (Fig. 3b).
It has been reported that FUS CLIP-Seq reads map predominantly to the 5′ region of NEAT1, with the read density being highest in the portion of NEAT1 gene encoding the short NEAT1_1 isoform . This raises the possibility that FUS mediates the recruitment of NEAT1_1 into paraspeckles during higher-order assembly of paraspeckle precursors into mature paraspeckles, whereas the deficiency of mutant FUS in paraspeckle formation would lead to NEAT1_1 release from paraspeckles. NEAT1 gene products were shown to be enriched ~ 10-fold in chromatin-bound fraction  indicating that paraspeckles are co-pelleted with chromatin. We obtained nuclear soluble extract (SNE) using this protocol  and prepared cDNA using oligo(dT) primer in order to amplify only polyadenylated transcripts and hence only NEAT1_1 but not NEAT1_2. Indeed, NEAT1_1 levels in SNE, as quantified by non-saturated PCR and qRT-PCR, were significantly higher in FUS ΔNLS lines as compared to WT cells (Fig. 3c), indicating abnormal release of NEAT1_1 from paraspeckles in mutant FUS expressing cells.
We speculated that compromised ability of mutant FUS to maintain paraspeckle formation might become more evident under stress conditions. To test this, we used a viral infection mimic, synthetic dsRNA poly(I:C), a pathophysiological stimulus reported to enhance NEAT1 synthesis and paraspeckle formation . In ΔNLS_het lines, a significant proportion of poly(I:C)-treated cells had a diffuse NEAT1 signal, as opposed to well-defined paraspeckles in all WT cells (Fig. 3d), indicating that stress-induced paraspeckle assembly is indeed impaired in cells expressing mutant FUS. Similar results were obtained with another paraspeckle-inducing stressor, proteasome inhibitor MG132  (Additional file 1: Figure S4C).
Structural deficiencies in paraspeckles revealed in FUS ΔNLS lines suggested their compromised functionality. One established function of paraspeckles is positive regulation of miRNA biogenesis; in particular, paraspeckles regulate processing of pri-miR-17~92 transcript by enhancing the Microprocessor activity . We found a significant decrease in the levels of six miRNAs produced from this miRNA precursor not only in homozygous but also in heterozygous FUS ΔNLS lines (Fig. 3e).
Overall, the above results indicate that the capability of mutant FUS to maintain structural integrity and functionality of paraspeckles is impaired even in cells with minor cytoplasmic redistribution of the protein.
Paraspeckles are formed in spinal neurons and glia of ALS-FUS patients
In our previous study, we found that NONO is mislocalised and aggregated in ALS-FUS . We examined SFPQ distribution in the same ALS-FUS cases. Although SFPQ was accumulated in the nucleus of neurons and glial cells in ALS-FUS cases, its mislocalisation or aggregation was not observed (Additional file 1: Figure S6). We also studied the behaviour of overexpressed GFP-tagged SFPQ and NONO in primary mouse neurons. In agreement with the post-mortem data, overexpressed SFPQ was confined to the nucleus, whereas NONO often mislocalised and aggregated in the cytoplasm of neurons (Additional file 1: Figure S7). Therefore nuclear SFPQ distribution is preserved in ALS-FUS allowing enhanced NEAT1 accumulation and paraspeckle assembly.
In the current study we provide evidence that accumulation of structurally and functionally compromised paraspeckles may serve as a novel pathomechanism in ALS-FUS. Our study reinforces the notion of enhanced paraspeckle assembly in spinal neurons and glia as a hallmark of ALS. Indeed, we show that paraspeckle formation is typical even for ALS cases with the pathology of a structural paraspeckle protein.
Paraspeckles exert anti-apoptotic activity and increase viability of cells under stressful conditions [24, 53, 67], therefore their formation in motor neurons at the early stages of pathological process in ALS may serve as a mechanism to prolong neuronal survival. However, although cells expressing mutant FUS, similar to TDP-43 depleted cells , are characterised by paraspeckle hyper-assembly, FUS mutations would impact on paraspeckle functionality. Disruption of paraspeckle-dependent neuroprotection may thus contribute to the particularly aggressive disease phenotype (early onset and fast progression) typical for ALS-FUS .
Comparison of our homozygous and heterozygous FUS ΔNLS cell lines revealed that the presence of WT FUS ameliorates mislocalisation of mutant FUS, possibly by retaining the mutant protein in the nucleus via interactions between normal and mutant FUS. It remains to be established whether nuclear retention of mutant FUS is protective or rather detrimental – e.g. by exacerbating toxic gain of function in the nucleus, including via paraspeckles. Results of previous studies of nuclear RNA granules also support gain of nuclear toxicity by mutant FUS as a disease mechanism. For example, a negative effect of mutant FUS on nuclear bodies Gems independent of its cytoplasmic mislocalisation has been demonstrated [57, 71]. In addition, FUS mutations may impact on its nuclear functions by affecting target gene expression directly  or via altered chromatin structure . Of note, components of the chromatin remodelling complex can be recruited to paraspeckles . Our results suggest that nuclear gain of function by mutant FUS may play a more important role in ALS-FUS pathogenesis than previously believed.
We found that in contrast to NONO, SFPQ does not mislocalise or aggregate in ALS-FUS, moreover, its nuclear levels are increased compared to control cases. This accumulation might play a compensatory role and serve to ameliorate the effects of NONO and FUS loss of function. Elevated SFPQ levels would also promote NEAT1 accumulation, however, since SFPQ is not significantly upregulated in FUS ΔNLS lines which nevertheless accumulate NEAT1, additional mechanisms are likely to be involved. Our transcriptomic analysis of FUS ΔNLS did not highlight any significantly dysregulated cellular pathways which could explain for NEAT1 upregulation (data not shown). It is plausible that small changes in the function of multiple pathways in mutant FUS expressing cells synergise to affect NEAT1 expression. In addition, our RNA-Seq analysis provided relatively low read coverage (~ 20 M reads/sample) and thus did not capture possible changes in the levels of low-abundance transcripts which may have impacted on NEAT1 levels. Alternatively, abnormal NEAT1 regulation can be realised at the level of posttranslational protein modifications .
An immediate consequence of altered structural integrity of paraspeckles in cells expressing mutant FUS is the release of NEAT1_1. NEAT1_1 is among the most abundant lncRNAs in human cells [22, 35] including those lacking paraspeckles, such as neurons. It functions to modulate transcription, including via regulation of chromatin active state [7, 35, 69]. It is highly likely that elevated levels of NEAT1_1 in neurons will cause wide-spread changes in gene expression. Recently, NEAT1_1 has been shown to interact with the p53 pathway [1, 40] and modulate neuronal excitability . The latter study is especially intriguing because it suggests that elevated neuronal NEAT1_1 levels in ALS may directly contribute to their abnormal excitability . Further studies are required to decipher molecular mechanisms responsible for NEAT1 upregulation and to establish whether accumulated NEAT1_1 is a significant driver of global gene expression changes in mutant FUS expressing cells.
Another important finding of our current study relevant to ALS pathogenesis is the severe repression of ADARB2 expression in mutant FUS expressing cells. ADARB2 is mainly expressed in the nervous system and was shown to be sequestered into C9ORF72 foci suggesting loss of its function in ALS-C9 , although possible functional consequences of this effect are yet to be addressed. ADARB2 depletion, despite mediated by a different upstream mechanism, might be a converging phenotype in ALS-FUS and ALS-C9.
Finally, our results suggest that the role of FUS in miRNA biogenesis  can be at least in part mediated by paraspeckles, and now it needs to be addressed to what extent pri-miRNA processing relies on the assembly of mature paraspeckles by FUS.
In conclusion, our study identifies a novel molecular phenotype driven by loss and gain of nuclear function of mutant FUS which may contribute to the disease severity in ALS-FUS.
We acknowledge the London Neurodegenerative Diseases Brain Bank and Sheffield Brain Tissue Bank and for providing human materials. We also thank Angela Marchbank from School of Biosciences Genomics Research Hub for the help with RNA-Seq analysis.
The study was supported by fellowships from Medical Research Foundation and Motor Neurone Disease Association (Shelkovnikova/Oct17/968–799) to TAS. The study was also funded by Research Grants from Russian Science Foundation (project No.18–15-00357) and Motor Neuron Disease Association to VLB (Buchman/Apr13/6096). HA is a recipient of Cardiff University/China Council PhD studentship.
Availability of data and materials
The datasets used and/or analysed during the current study are available from corresponding authors on reasonable request.
TAS conceived research; TAS, HA and VLB designed experiments and analysed data; HA, LS and TAS performed experiments; AHF, JRH, AN and VLaB contributed research tools and analysed data; TAS wrote manuscript with input from all authors. All authors read and approved the final version of the manuscript.
Ethics approval and consent to participate
Human samples were from clinically and histopathologically characterised ALS cases and neurologically healthy individuals. Samples were provided by the Sheffield Brain Tissue Bank and MRC London Neurodegenerative Diseases Brain Bank (Institute of Psychiatry, King’s College London). Consent was obtained from all subjects for autopsy and histopathological assessment, and research were performed in accordance with local and national Ethics Committee approved donation.
Consent for publication
Authors declare no competing interests.
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
- 3.Andersson MK, Stahlberg A, Arvidsson Y, Olofsson A, Semb H, Stenman G, Nilsson O, Aman P (2008) The multifunctional FUS, EWS and TAF15 proto-oncoproteins show cell type-specific expression patterns and involvement in cell spreading and stress response. BMC Cell Biol 9:37PubMedPubMedCentralCrossRefGoogle Scholar
- 10.Choudhry H, Albukhari A, Morotti M, Haider S, Moralli D, Smythies J, Schodel J, Green CM, Camps C, Buffa F et al (2015) Tumor hypoxia induces nuclear paraspeckle formation through HIF-2alpha dependent transcriptional activation of NEAT1 leading to cancer cell survival. Oncogene 34(34):4546PubMedCrossRefGoogle Scholar
- 26.Imamura K, Imamachi N, Akizuki G, Kumakura M, Kawaguchi A, Nagata K, Kato A, Kawaguchi Y, Sato H, Yoneda M et al (2014) Long noncoding RNA NEAT1-dependent SFPQ relocation from promoter region to paraspeckle mediates IL8 expression upon immune stimuli. Mol Cell 53(3):393–406PubMedCrossRefGoogle Scholar
- 30.Kukharsky MS, Quintiero A, Matsumoto T, Matsukawa K, An H, Hashimoto T, Iwatsubo T, Buchman VL, Shelkovnikova TA (2015) Calcium-responsive transactivator (CREST) protein shares a set of structural and functional traits with other proteins associated with amyotrophic lateral sclerosis. Mol Neurodegener 10:20PubMedPubMedCentralCrossRefGoogle Scholar
- 35.Li R, Harvey AR, Hodgetts SI, Fox AH (2017) Functional dissection of NEAT1 using genome editing reveals substantial localisation of the NEAT1_1 isoform outside paraspeckles. RNA 23(6):872–881.Google Scholar
- 37.Lopez-Erauskin J, Tadokoro T, Baughn MW, Myers B, McAlonis-Downes M, Chillon-Marinas C, Asiaban JN, Artates J, Bui AT, Vetto AP et al (2018) ALS/FTD-linked mutation in FUS suppresses intra-axonal protein synthesis and drives disease without nuclear loss-of-function of FUS. Neuron 100(4):816–830.e7.Google Scholar
- 44.Nishimoto Y, Nakagawa S, Hirose T, Okano HJ, Takao M, Shibata S, Suyama S, Kuwako K, Imai T, Murayama S et al (2013) The long non-coding RNA nuclear-enriched abundant transcript 1_2 induces paraspeckle formation in the motor neuron during the early phase of amyotrophic lateral sclerosis. Mol Brain 6(1):31PubMedPubMedCentralCrossRefGoogle Scholar
- 48.Ratti A, Buratti E (2016) Physiological functions and pathobiology of TDP-43 and FUS/TLS proteins. J Neurochem 138 Suppl 1:95–111.Google Scholar
- 56.St George-Hyslop P, Lin JQ, Miyashita A, Phillips EC, Qamar S, Randle SJ, Wang G (2018) The physiological and pathological biophysics of phase separation and gelation of RNA binding proteins in amyotrophic lateral sclerosis and fronto-temporal lobar degeneration. Brain Res 1693(Pt A):11–23.Google Scholar
Open AccessThis article is distributed under the terms of the Creative Commons Attribution 4.0 International License (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made. The Creative Commons Public Domain Dedication waiver (http://creativecommons.org/publicdomain/zero/1.0/) applies to the data made available in this article, unless otherwise stated.