Novel FUS/TLS Mutations and Pathology in Familial and Sporadic Amyotrophic Lateral Sclerosis FREE

Amyotrophic lateral sclerosis (ALS) is a fatal neurodegenerative disease occurring in adult life. It is caused by a degeneration of motor neurons in the cerebral cortex, brainstem, and spinal cord that leads to progressive muscle weakness, wasting, and paralysis.¹ Limb, bulbar, and respiratory muscles are affected. The etiology of motor neuron degeneration in ALS has not yet been elucidated, but there is evidence for a number of pathogenic mechanisms, including glutamate excitotoxicity, oxidative stress, mitochondrial dysfunction, protein misfolding and aggregation, axonal transport defects, endoplasmic reticulum stress, and glial dysfunction.² In 5% to 10% of cases, the disease is familial (familial ALS [FALS]), and causative mutations have been identified in several genes, including SOD1, ANG, and VAPB.³ Recently, mutations in 2 new genes, the transactive response–DNA binding protein gene (TARDBP) and fusion in malignant liposarcoma/translocated in liposarcoma gene (FUS/TLS [OMIM 137070]), were identified in patients with FALS.^4- 6 These genes are particularly interesting because they seem to account for a significant number of FALS cases, approximately 3% to 5%, and both of their products are multifunctional proteins involved in diverse cellular functions, including transcriptional regulation and RNA binding and splicing, which were not previously known to be defective in ALS.

The FUS/TLS protein is a 526–amino acid protein encoded on chromosome 16p11.2. It is widely expressed in most body tissues and is present in both the nucleus and cytoplasm.⁷ The N-terminal of the protein contains multiple SYGQQS degenerate amino acid repeats and functions as a potent transcriptional activator when present in oncogenic fusions with transcription factors in malignant liposarcomas and myeloid leukemia.^8- 9 Full-length FUS/TLS has been shown to regulate the activity of 2 transcription factors in vivo, nuclear factor kB and Spi-1.^10- 11 The C-terminal of FUS/TLS is composed of an RNA recognition motif and a cys4 zinc finger, flanked on either side by multiple RGG amino acid–repeat regions.¹² Together these allow FUS/TLS to bind RNA with sequence specificity. The FUS/TLS protein associates with RNA in both the nucleus and cytoplasm and engages in nuclear-cytoplasmic shuttling.¹³ It has been shown to promote alternative splicing,¹⁴ is present in complexes with heterogeneous ribonuclear proteins,¹⁵ and may be involved in somatodendritic transport of RNA within neurons.¹⁶

The aim of the present study was to determine the frequency of, and clinicopathologic phenotypes associated with, FUS/TLS mutations in a large cohort of ALS cases. Four nonsynonymous changes in the nucleotide sequence, 1 of which is novel, were identified in 6 patients with ALS (4 with FALS and 2 with sporadic ALS [SALS]).

Genomic DNA was extracted from whole blood or fresh-frozen cerebellar tissue samples obtained from ALS patients and controls using the Nucleon BACC Genomic Extraction Kit and Soft Tissue DNA Extraction Kit, respectively. Ethical approval for the study was obtained from the Sheffield Research Ethics Committee and informed consent was obtained in all ALS cases and controls. Intronic polymerase chain reaction primers were designed and optimized to amplify exonic sequences of the FUS/TLS gene (Enstembl transcript ID: ENST00000254108) from genomic DNA, in standard polymerase chain reactions. Primer sequences and optimum conditions are available on request. Polymerase chain reaction products were cleaned up using exonuclease I and shrimp alkaline phosphatase and then subject to direct sequencing using Big Dye Terminator, version 3.1 (Applied Biosystems, Foster City, California) according to the manufacturer's instructions. Sequencing products were read on an ABI 3730 Capillary Analyzer, and the chromatographs obtained were analyzed using Sequencher Software (Gene Codes, Ann Arbor, Michigan).

Cohort 1 was composed of all the ALS donations to the Sheffield Brain Tissue Bank together with clinic-based cases with a known familial history. Cases with SOD1 and TARDBP mutations were excluded, leaving a total of 159 patients with ALS, 42 FALS and 117 SALS cases. Control DNA from 293 individuals with no neurological disease was used for comparison. All samples were from white residents of the United Kingdom. Demographic data on the ALS patients and controls are shown in Table 1. Because most reported mutations in FUS/TLS are found in the C-terminal region, we also screened exons 14 and 15 in a larger clinic-based cohort of SALS (cohort 2; n = 431).

The brain and spinal cord from cases 1, 4, 5, and 6 were donated for research with the consent of the next of kin and with ethics committee approval. For cases 1, 5, and 6, one cerebral hemisphere, half the midbrain and brainstem, a portion of the cerebellum, and segments of the spinal cord at various levels were rapidly frozen at autopsy. The remaining central nervous system tissue was formalin-fixed and selected blocks were processed to paraffin. For case 4, only a portion of the cerebellum was frozen, otherwise the whole brain and spinal cord were formalin-fixed.

In addition to routine neuropathologic examination, immunocytochemistry was performed for p62/sequestosome 1 and TAR DNA-binding protein 43 (TDP-43) using standard protocols with antigen retrieval. Immunocytochemistry for FUS/TLS was performed on paraffin-embedded tissue from the spinal cord using a standard protocol with minor modifications (50 minutes of H₂O₂ to block endogenous peroxidase; 48 hours of primary antibody incubation). In case 5, owing to prolonged formalin fixation, immunocytochemistry was performed on frozen spinal cord tissue for FUS/TLS using the method of Kwiatkowski et al⁵ (primary antibody, 1:100). Comparative immunocytochemistry for FUS/TLS was performed on 1 SOD1-positive familial case, 1 SALS case, and 1 control case with no neurological disease.

In cohort 1, we identified 4 heterozygous nonsynonymous nucleotide changes in 6 patients. In exon 14, c.1520G>A was identified in a patient with SALS (Figure 1A), which is predicted to result in substitution of aspartate for glycine at residue 507 (p.Gly507Asp). In exon 15, c.1570A>T (Figure 1B), which was predicted to result in substitution of a tryptophan for arginine at residue 524 (p.Arg524Trp), was identified in 3 patients from 2 families with ALS. These changes were not identified in 293 healthy age- and sex-matched controls who were screened. In exon 5, c.521_523 + 3delGAGgtg was identified in a FALS patient (Figure 1C), which is predicted to cause deletion of a single glycine at residue 174 (p.Gly174del). This change was previously reported by Kwiatkowski and colleagues⁵ in 3 patients with FALS. Finally, in exon 6, c.684_685insGGC was identified in a SALS patient (Figure 1D), which is predicted to cause insertion of a single glycine in a run of 10 glycine residues (p.Gly228_Gly229insGly). Both the exon 5 and exon 6 changes were also each identified in 1 control case of 293 screened. c.684_685insGGC was identified in a 77-year-old woman, and c.521_523 + 3delGAGgtg was identified in a 72-year-old man.

A list of the sequence changes identified in ALS patients, together with several other novel nonsynonymous polymorphisms identified in controls, is shown in Table 2. No additional mutations in exons 14 or 15 were identified in the clinic-based cohort of 431 SALS patients (cohort 2). In silico analysis of p.Gly507Asp and p.Arg524Trp with PMut software predicts that they are pathogenic. These residues also show a high degree of interspecies conservation (data not shown). We therefore conclude that these changes are likely to represent pathogenic mutations.

A summary of the clinical features of the ALS patients with FUS/TLS mutations is shown in Table 3.

This case had the p.Arg524Trp substitution. This man had FALS and developed the progressive muscular atrophy variant of motor neuron disease (MND), with upper limb onset at the age of 61 years. He presented to the neurology clinic with a striking clinical picture of flail arms and head drop. His brother had presented 9 months earlier with an almost identical pattern of disease at age 58 years; however, DNA was not available from the brother to confirm segregation with the disease. Both the index case and his brother had a history of colonic neoplasia. No other family history of ALS was reported. However, there was a family history of infant deaths, with 4 male siblings of the patient dying within the first few weeks of life. The clincopathologic findings of the index case have been reported previously.¹⁷

These 2 cases had the p.Arg524Trp substitution. They were a brother and sister who had classic limb-onset ALS. The brother's disease was diagnosed at age 55 years and had a disease duration of 3 years, dying at age 58 years in 1991. His sister was affected simultaneously, with a similar disease duration. She died at age 63 years. No further clinical details are now available, as hospital records were destroyed 10 years after the patients' deaths.

This case had p.Gly174del. She had multiple sclerosis as well as FALS. The onset of her multiple sclerosis was at the age of 23 years and was initially relapsing and remitting before becoming secondary progressive in the 15 years leading up to the onset of her FALS. She presented with bulbar-onset ALS at age 62 years. There was a 14-month history of increasing leg weakness, bulbar symptoms, and weakness in her right hand. Her mother had died following a 2-year degenerative disease. Clinical details were unavailable, but the description was compatible with ALS with frontotemporal lobar dementia. On examination, the upper limbs showed asymmetric weakness and occasional fasciculations. Her tongue showed fasciculation, wasting, and weakness. The lower limbs showed marked spastic paraplegia and generalized wasting. All deep tendon reflexes were pathologically brisk, and the plantar responses were extensor. Routine laboratory investigation results were normal. Magnetic resonance imaging of the brain and spine at presentation demonstrated demyelination consistent with the earlier diagnosis of multiple sclerosis. Neurophysiologic testing showed evidence of anterior horn cell disease. Bulbar involvement became more severe, with unintelligible speech and required insertion of a percutaneous endoscopic gastrostomy tube for feeding. There was no evidence of cognitive dysfunction. Deterioration in her respiratory function led to death 26 months after ALS symptom onset.

This case had the p.Gly507Asp substitution. This case was a man who presented at age 69 years with predominantly lower motor neuron (LMN) ALS. He had no family history of neurological disease; his parents died at about 70 years. He developed weakness of his right leg, which progressed over 18 months, followed by severe and disabling weakness in both upper limbs. The upper limbs showed marked bilateral, proximal muscle wasting with fasciculation around the shoulder girdle. Tone was decreased and reflexes were depressed. The lower limbs showed fasciculation in both quadriceps with distal weakness and foot drop on the right, normal tone, brisk knee, and absent ankle reflexes. There was no evidence of bulbar dysfunction or cognitive impairment. His serum creatine kinase level was mildly elevated (789 U/L). Neurophysiology showed severe active diffuse neurogenic changes consistent with an anterior horn cell disorder. Disease progression included increasing upper and lower limb involvement and he died from respiratory failure 42 months from symptom onset.

This case had the p.Gly228_Gly229insGly insertion. He developed bulbar-onset ALS at age 41 years. There was no family history of neurological disease, the individual dying before both of his parents. He presented with 18 months of progressive dysarthria and dysphagia associated with weight loss of 30 kg. He developed weakness in his upper and lower limbs. He had a medical history of epilepsy controlled with valproate. On examination, he had a weak, wasted, and fasciculating tongue. There was wasting of intrinsic hand muscles and of the anterior compartment of the right lower leg. Fasciculations were present in all 4 limbs and reflexes were brisk. Neurophysiologic testing was consistent with anterior horn cell disease. Bulbar dysfunction worsened, resulting in aspiration and chest infections. Respiratory muscle weakness resulted in symptoms of nocturnal carbon dioxide retention. His forced vital capacity 3 months prior to death was 29% of predicted. Limb weakness showed gradual progression that required use of a wheelchair in the last month of the patient's life. He developed respiratory failure 2 days after percutaneous endoscopic gastrostomy tube insertion and died 25 months after symptom onset.

Case 1 (Arg524Trp), described previously,¹⁷ showed marked depletion of LMN throughout the spinal cord with absent ubiquitinated cytoplasmic inclusions in LMN and minimal upper motor neuron pathology. Immunocytochemistry for p62 confirmed the absence of neuronal cytoplasmic inclusions (NCIs) in both the motor and extramotor central nervous system, but glial cytoplasmic and nuclear inclusions (Figure 2A and E) were identified. Immunocytochemistry for TDP-43 showed normal nuclear labeling without neuronal or glial cytoplasmic inclusions (GCIs) (Figure 2I). Immunocytochemistry for FUS/TLS showed predominant nuclear labeling in neurons and glia. However, a number of LMNs showed strong cytoplasmic FUS/TLS expression, including skein-like structures in occasional LMNs (Figure 3B), and GCI were also present (Figure 3G).

Case 4 (Gly174del) showed pathologic evidence of both ALS and multiple sclerosis. Loss of LMNs from the spinal cord and medullary motor nuclei was associated with Bunina bodies and ubiquitin/TDP-43–positive NCI in the motor, frontal, and temporal neocortex; hippocampus (pyramidal and dentate granule cell layers); midbrain; medulla; and spinal cord at multiple levels (Figure 2C and K). Ubiquitinated/p62-positive GCIs were also observed (Figure 2G). There was myelin pallor and a prominent microglial reaction in the corticospinal tracts (Figure 4). Immunocytochemistry for FUS/TLS showed normal appearances without cytoplasmic retention or inclusions (Figure 3D and I). In addition, there were small circumscribed areas of demyelination with relative axon preservation, associated with sparse, predominantly T cell, lymphocytic infiltration in the cerebral peduncle, ventral pons, medullary pyramid, and spinal dorsal columns (Figure 4). There were shadow plaques in the cerebral white matter around the angles of the lateral ventricles.

Case 5 (Gly507Asp) also showed marked loss of LMN at all spinal levels with relative preservation in motor nuclei of the brainstem. Bunina bodies were present in residual LMN. There was preservation of Betz cells in the neocortex and minimal pathology of the corticospinal tracts. Immunocytochemistry for p62 showed infrequent skein-like NCI in LMN on screening multiple sections from spinal cord (Figure 2B), but none in the motor cortex, medulla, or extra motor blocks. Glial cytoplasmic inclusions were also present (Figure 2F). Results of TDP-43 staining were normal (Figure 2J). Immunocytochemistry of frozen spinal cord tissue showed FUS-positive GCI (Figure 3H). Lower motor neurons demonstrated normal FUS labeling (Figure 3C).

Case 6 (Gly228_Gly229insGly) showed spinal LMN loss with a microglial reaction throughout lateral and ventral spinal cord white matter tracts. Residual neurons showed Bunina bodies and ubiquitinated/TDP-43–positive NCIs (Figure 2D and L) and GCI (Figure 2H). The FUS immunocytochemistry showed normal expression (Figure 3E and J).

We report findings from screening the FUS/TLS gene in a large cohort of FALS and SALS cases from the north of England. Four nucleotide sequence changes, 1 of which is novel, were identified in 3 familial cases and 2 sporadic cases. Screening for these changes in healthy controls revealed that 2 of the changes, p.Gly507Asp and p.Arg524Trp, were not present in controls. Pathology of the cases with p.Arg524Trp (case 1) and p.Gly507Asp (case 5) showed marked depletion of LMN, minimal upper motor neuron pathology, and absent ubiquitinated TDP-43 immunoreactive neuronal inclusions. The examination of extensive pathologic material in case 1 revealed infrequent LMNs containing skein-like cytoplasmic inclusions labeling strongly with antibodies to FUS/TLS, but negative for TDP-43, which is consistent with previous findings.^5- 6 Both cases also demonstrated GCIs positive for FUS/TLS and p62.

The other 2 FUS/TLS sequence changes identified in ALS cases, p.Gly174del and p.Gly228_Gly229insGly, were also detected in single control cases. Pathology of these cases was typical of SALS and non-SOD1–associated FALS,¹⁸ including NCI labeling with antibodies to p62 and TDP-43. The FUS/TLS immunocytochemistry in these cases also showed a normal nuclear pattern of labeling without cytoplasmic inclusions. Both of these changes result in alterations in the number of glycine residues in regions containing glycine-repeated motifs. There is a high degree of interspecies variation in the number of glycine residues in these regions. We conclude that these 2 changes, one of which (c.delGAGgtg523) was previously reported by Kwiatkowski and colleagues⁵ in a FALS case and the other (c.684_685insGGC) reported recently by Belzil and colleagues¹⁹ in a non-ALS case may represent benign polymorphisms. Analysis of clinical findings in cases with p.Gly507Asp and p.Arg524Trp indicates a predisposition for limb onset and a predominant LMN phenotype in ALS cases with FUS/TLS mutations.

Based on our findings, we estimate the frequency of FUS/TLS mutations in non-SOD1 FALS to be approximately 5%. We also identified a novel FUS/TLS mutation in a SALS case (p.Gly507Asp). Subsequent screening of this region and exon 15, the most common site for FUS/TLS mutations in ALS cases, in 431 sporadic cases did not reveal any additional sequence changes. This suggests that FUS/TLS mutations are a much less frequent cause of sporadic disease. This is consistent with other FALS-associated genes; mutations in SOD1, TARDBP, and ANG account for 12% to 23.5%, 3.6% to 5%, and 1.5% to 2.3% of FALS cases, respectively, but only 1% to 7%, approximately 0.5%, and approximately 1% of SALS cases.^4,20- 23 Various factors may result in familial ALS being misdiagnosed as apparent sporadic disease, including failure to take a sufficiently detailed family history, incomplete penetrance, and nonpaternity. Therefore, mutations in these genes in truly sporadic disease may be even less frequent than so far reported. However, postmortem central nervous system tissue from patients with SALS shows signs of oxidative stress,²⁴ and TDP-43 and FUS/TLS have been found in NCIs in sporadic forms of ALS¹⁸ and a closely related condition, frontotemporal lobar dementia,²⁵ respectively, suggesting that these genes have an important, though as yet not completely defined, role in the neurodegeneration occurring in sporadic disease.

The FUS/TLS protein is a widely expressed protein involved in diverse cellular functions. The highly conserved C-terminal region, encoded by exons 14 and 15, appears to be the site most commonly mutated in ALS. The C-terminal of FUS/TLS is important in regulating DNA and RNA binding^13,26 and for alternative splicing activity.¹⁴ Interestingly, the C-terminal of TDP-43 is also the site most commonly mutated in ALS and is important for its alternative splicing activity.^27- 28 Deletion of the C-terminal of TDP-43 is known to disrupt nucleocytoplasmic transport, possibly by altering protein solubility.²⁹ In cell models of C-terminal FUS/TLS mutants, there is increased cytoplasmic localization of mutant protein.⁶ Therefore, mutations in the C-terminal of FUS/TLS may exert their pathogenic effects by disrupting the subcellular distribution of the protein. This would have implications for its reported role in somatodendritic RNA transport in neurons¹⁶ and has similarities with findings in another degenerative motor system disorder, spinal muscular atrophy.³⁰ Accumulation of mutant FUS/TLS protein in the cytoplasm may sequester other heterogeneous ribonuclear proteins and RNAs, in the same way that TDP-43, forcibly localized to the cytoplasm by mutation of its nuclear localization sequence, sequesters endogenous TDP-43 from the nucleus.³¹ The implications of this would be a progressive generalized disruption to transcription and RNA processing, and suggests a mechanism by which C-terminal FUS/TLS mutations may exert their effects by a toxic gain of function. Given the similarities between the pathophysiology of FUS/TLS and TDP-43 in ALS, the possibility that mutations, predominantly in the C-terminal region, of FUS/TLS may also exert a deleterious effect on gene transcription and RNA processing requires further investigation.

Correspondence: Pamela J. Shaw, MD, Academic Unit of Neurology, University of Sheffield, Medical School, Beech Hill Road, Sheffield S10 2RX, England (pamela.shaw@sheffield.ac.uk).

Accepted for Publication: September 16, 2009.

Author Contributions:Study concept and design: Kirby, McDermott, and Shaw. Acquisition of data: Hewitt, Highley, Hartley, Hibberd, Hollinger, Williams, Ince, McDermott, and Shaw. Analysis and interpretation of data: Hewitt, Kirby, Highley, Ince, McDermott, and Shaw. Drafting of the manuscript: Hewitt, Kirby, Highley, Williams, Ince, McDermott, and Shaw. Critical revision of the manuscript for important intellectual content: Hewitt, Kirby, Highley, Hartley, Hibberd, Hollinger, Williams, Ince, McDermott, and Shaw. Obtained funding: Shaw. Administrative, technical, and material support: Kirby, Hartley, Hibberd, Hollinger, Williams, and Shaw. Study supervision: Kirby, Ince, McDermott, and Shaw.

Financial Disclosure: None reported.

Funding/Support: This project was funded by the Medical Research Council (Dr Ince), Wellcome Trust (Dr Shaw), and Motor Neurone Disease Association (Dr Shaw, Ms Hartley, and Ms Hibberd). Dr Highley is funded by an MRC/MNDA Lady Edith Wolfson Fellowship. Additional control DNA was extracted from central nervous system tissue provided by the Newcastle Brain Tissue Resource, which is funded in part by grant G0400074 from the UK Medical Research Council.