Amyotrophic lateral sclerosis (ALS) and frontotemporal dementia (FTD) are part of a clinical, pathological and genetic continuum.
Objectives The purpose of the present study was to assess the mutation burden that is present in patients with concurrent ALS and FTD (ALS/FTD) not carrying the chromosome 9 open reading frame 72 (C9orf72) hexanucleotide repeat expansion, the most important genetic cause in both diseases.
Methods From an initial group of 973 patients with ALS, we retrospectively selected those patients fulfilling diagnostic criteria of concomitant ALS and FTD lacking the repeat expansion mutation in C9orf72. Our final study group consisted of 54 patients clinically diagnosed with ALS/FTD (16 with available postmortem neuropathological diagnosis). Data from whole exome sequencing were used to screen for mutations in known ALS and/or FTD genes.
Results We identified 11 patients carrying a probable pathogenic mutation, representing an overall mutation frequency of 20.4%. TBK1 was the most important genetic cause of ALS/FTD (n=5; 9.3%). The second most common mutated gene was SQSTM1, with three mutation carriers (one of them also harboured a TBK1 mutation). We also detected probable pathogenic genetic alterations in TAF15, VCP and TARDBP and possible pathogenic mutations in FIG4 and ERBB4.
Conclusion Our results indicate a high genetic burden underlying the co-occurrence of ALS and FTD and expand the phenotype associated with TAF15, FIG4 and ERBB4 to FTD. A systematic screening of ALS and FTD genes could be indicated in patients manifesting both diseases without the C9orf72 expansion mutation, regardless of family history of disease.
Statistics from Altmetric.com
Amyotrophic lateral sclerosis (ALS) is a neurodegenerative disease characterised by the degeneration of upper motor neuron (UMN) and lower motor neuron (LMN), which results in death usually within 3–5 years from symptoms onset.1 Frontotemporal dementia (FTD) is associated with the degeneration of the frontal and/or temporal lobes, which leads to a progressive deterioration in behaviour and/or language, with relative preservation of memory.2 To date, there is overwhelming evidence that ALS and FTD are part of a disease continuum with clinical, pathological and genetic overlapping features.3 The hexanucleotide repeat expansion on chromosome 9 open reading frame 72 (C9orf72) is the most prevalent genetic cause of both conditions. Within the European population, it accounts for up to 8% of sporadic and 40% of familial ALS and 6% of sporadic and 25% of familial FTD.4 5 Remarkably, the frequency of this genetic alteration increases to 14% in sporadic and 57% in familial forms in patients with concomitant clinical features of ALS and FTD (ALS/FTD), thus suggesting an enriched genetic load underlying ALS and FTD co-occurrence.4
A systematic evaluation of genes previously related to ALS and/or FTD in patients manifesting both phenotypes who lack the C9orf72 genetic alteration has not yet been performed, probably as a consequence of the low frequency of individuals with concomitant ALS/FTD, the large number of genes that should be evaluated (there are more than 30 genes reported to date), or the high incidence of the C9orf72 expansion mutation that is present in this rare condition.3 6 7
In order to evaluate the overall genetic contribution of genes previously related to ALS or FTD in this complex phenotype, and to uncover possible novel disease-causing variants in these genes, we have performed an in-depth genetic analysis through a whole exome sequencing (WES) approach in a series of patients with concomitant ALS/FTD not carrying the C9orf72 hexanucleotide repeat expansion mutation.
Material and methods
Patients under study were selected from clinical registers through databases from three different centres. All patients were evaluated and diagnosed of ALS by a consultant neurologist according to El Escorial criteria.8 FTD diagnosis was performed following the international criteria for the behavioural variant of FTD9 or the international consensus criteria for primary progressive aphasias (PPA).10 From the motor neuron disease (MND) register at Hospital de la Santa Creu i Sant Pau, among 129 patients with a diagnosis of ALS, 15 fulfilling FTD criteria and not carrying the repeat expansion mutation were identified. After DNA quality control, 13 of them were selected for WES (three with postmortem neuropathological studies). The Hospital 12 de Octubre MND database included 844 ALS individuals. From these, 63 patients fulfilled FTD diagnosis and did not carry the C9orf72 repeat expansion. After DNA quality control, 29 samples were included for WES (one with available neuropathological data). Finally, the Neurological Tissue Bank of the Biobanc-Hospital Clinic-IDIBAPS (Barcelona) database was used to retrospectively identify patients clinically diagnosed with ALS/FTD or cases with a postmortem diagnosis compatible with MND and frontotemporal lobar degeneration (FTLD), which disclosed 12 individuals fulfilling at least one of these criteria.
Our final dataset consisted of 54 patients with concomitant ALS and FTD (44 males and 10 females), all Spanish of European descent. Among them, neuropathological data were available for 15 cases from the Neurological Tissue Bank of the Biobanc-Hospital Clinic-IDIBAPS (Barcelona) and one from the Neuropathology Unit of the Hospital Universitario Fundación Alcorcón (Madrid). Mean age at onset was 62.1 years, ranging from 36 to 82 years. Mean age at death of brain donors was 71 years, ranging from 59 to 84 years. Among all patients, 18.5% had a family history (first or second-degree relative) of ALS and/or FTD.
The C9orf72 hexanucleotide expansion mutation was evaluated as previously described,15 and was excluded in all 54 samples that underwent WES. Genomic regions corresponding to all annotated human exons were enriched by hybridisation using the Agilent SureSelect Human All-Exons V5 kit (Agilent Technologies, Santa Clara, California, USA). Final libraries were sequenced on a HiSeq 2500 (Illumina, San Diego, California, USA) with paired end 100 bp reads. Raw sequencing data processing was performed as described in the online supplementary methods. All coding regions and exon–intron boundaries of 33 genes previously associated with ALS and/or FTD (see online supplementary table 1) were examined. Protein truncating mutations (nonsense, frameshift, canonical +/−1 or 2 splice sites or exon deletion) and non-synonymous novel or rare variants (minor allele frequency (MAF) ≤0.001 in non-Finnish European (NFE) population from the Genome Aggregation Database (gnomAD, http://gnomad.broadinstitute.org), which comprises 126 738 chromosomes), were considered for further analyses. The allele frequency in NFE of 0.001 practically corresponds to the known ALS pathogenic variant in SOD1 (p.D91A).16 We classified these rare genetic alterations into three categories: ‘probable pathogenic’, ‘possible pathogenic’ and ‘of unknown significance’. The ‘probable pathogenic’ category included variants with a MAF ≤0.001 in gnomAD NFE population and previous supporting functional or segregation data, or belonging to a functional domain known to harbour disease-causing mutations, or previously associated with an increased risk of ALS or FTD; the ‘possible pathogenic’ group contained variants with a MAF ≤1×10-5 in gnomAD NFE population without any previous study reporting them; for the ‘unknown significance’ category, we considered all rare variants not belonging to the probable or possible pathogenic groups. Importantly, for recessive genes (see online supplementary table 1), we only took into account homozygous mutations. All probable and possible pathogenic variants were verified through Sanger sequencing. Data derived from WES comprising 267 healthy Spanish individuals17 and 107 subjects from the Iberian population included in the 1000 Genomes Project Consortium (http://www.1000genomes.org) were also used as a reference of genetic variation (table 1). Mean coverage of all the exome targeted region was of 79 reads per base and, on average, 94% of the targeted region was covered with more than 10 reads in each sequenced sample. Sequence data have been deposited at the European Genome-phenome Archive (EGA), which is hosted by the EBI and the CRG, under accession number EGAS00001002439.
Clinical and demographic features of patients carrying a probable or possible pathogenic variant are summarised in table 1.
Probable pathogenic variants
Overall, 11 out of 54 patients with ALS/FTD (20.4%) harboured a probable pathogenic variant which potentially explains the associated phenotype (one patient was a double mutation carrier). Among them, nine cases did not show any family history of ALS and/or FTD.
Screening of the TANK-binding kinase 1 (TBK1) gene revealed one nonsense, two splice-site mutations and two in-frame deletions in five sporadic patients, thus representing 9.3% of our case series. All of them were diagnosed with a limb onset probable ALS and a behavioural variant FTD (bvFTD), except one patient who presented with the non-fluent variant of PPA. This patient carried the p.N254Kfs*4 TBK1 mutation and a missense variant (p.P392L) in the sequestosome 1 (SQSTM1) gene. Among TBK1 mutation carriers, three of them (cases carrying the p.K30_E76del, p.G272_T331del or p.T79del) have been recently reported in an independent study aimed at investigating the genetic contribution of TBK1 in an extended European cohort of patients belonging to the ALS/FTD disease continuum.18 These three patients were included in our exome sequencing project during the European cohort sample recruitment, and TBK1 mutation status was unknown at the time of exome sequencing.
Two mutations (p.P392L and p.A33V) in SQSTM1, previously reported as pathogenic,19 were identified in three patients with a sporadic disease, thus representing the second most important genetic alteration of ALS/FTD in our series (table 1). Neuropathological examination was available in cases 6 (p.P392L carrier) and 7 (p.A33V carrier) and revealed a FTLD type B phenotype with UMN involvement in both cases and LMN in case 6 associated with neuronal cytoplasmic TDP-43 protein aggregates. Both cases showed hippocampal sclerosis and prominent degeneration of basal ganglia. Interestingly, case 7 showed additionally scarce FUS RNA binding protein (FUS), TATA-box binding protein associated factor 15 (TAF15) and Transportin 1 (TNPO1) positive inclusions restricted to dentate gyrus of the hippocampus and temporo-occipital cortex (figure 1). Case 6 was clinically diagnosed with Paget’s disease of bone (PDB) and parkinsonism and had frequent intraneuronal ‘cat-eye’-type TDP-43 protein inclusions and a severe degeneration of the substantia nigra at neuropathological examination, which was the likely substrate of parkinsonism.
The inclusion in the present genetic screening of the FET protein family members FUS, TAF15 and EWS RNA binding protein 1 (EWSR1) (three structurally similar RNA-binding proteins that have been linked with ALS and FTLD) disclosed the presence of two novel missense variants (p. S390T and p.G462S) in the C-terminal RGG domain of TAF15.20 The p.S390T mutation was found in a 49-year-old male diagnosed with bvFTD, ALS (with predominant LMN signs) and dyskinesia. The other mutation found in TAF15 (p.G462S) emerged in a patient diagnosed with a progressive cognitive impairment, behavioural disturbances and psychotic symptoms at the age of 66, followed by parkinsonism and ALS. His mother and two siblings suffered from late onset dementia with prominent behavioural alteration. Samples from affected family members were not available, thus precluding any segregation analysis. Postmortem neuropathological examination revealed a FTLD type B pattern with UMN and LMN involvement with chromatolytic neurons and Bunina bodies in anterior horns associated with neuronal TDP-43 protein aggregates (Type B), with moderate atrophy of basal ganglia and nigral degeneration. The patient had also mild Alzheimer’s disease (AD) neuropathological changes (Braak neurofibrillary stage I, CERAD plaque score B, Thal amyloid phase 5) corresponding to an A3,B1,C2 score with mild amyloid angiopathy. Immunohistochemistry for TAF15 and TNPO1 had an overall very weak staining, but did not stain pathological inclusions in hippocampus and spinal cord and no shift of nuclear-cytoplasmic staining was observed (figure 1).
Analysis of the TAR DNA-binding protein (TARDBP) disclosed the p.A90V mutation, previously described in a patient with concomitant ALS and FTD with a family history of dementia.21 The mutation carrier reported herein was diagnosed with ALS of bulbar onset at the age of 63. After ALS diagnosis, he underwent a cognitive and behavioural screen revealing a bvFTD that started 14 years before the emergence of motor symptoms. Behavioural changes had been previously misdiagnosed as a late-onset bipolar disorder. Family history revealed psychiatric/behavioural symptoms in his father and two sisters. Bulbar weakness rapidly progressed to complete anarthria and severe dysphagia. The patient was reluctant to the placement of gastrostomy and died due to aspiration pneumonia. Neuropathological examination showed classical features of upper and lower MND with prominent corticospinal tract degeneration and loss of motor neurons of anterior horn of the spinal cord. There was a focal and relatively prominent gliosis of amygdala. In residual motor neurons, frequent TDP-43 immunoreactive skein-like neuronal cytoplasmic inclusions were detected, but there were only few inclusions in basal ganglia, thalamus and dentate gyrus of the hippocampus. Topographical distribution corresponded to stage 4 according to Brettschneider et al 11 despite minimal extramotor involvement. There were also frequent oligodendroglial inclusions in primary motor cortex and to a lesser extent in brainstem white matter tracts. Immunohistochemistry for TAF15 and TNPO1 did not stain pathological inclusions in motor neurons (figure 1).
Genetic screening of valosin-containing protein (VCP) revealed a mutation carrier of the p.I27V amino acid exchange, which has been described in patients with different clinical symptoms including FTD.22 23 This patient was diagnosed with bvFTD at the age of 80, after 40 years of a major depression disorder. Amyotrophy, weakness in the lower limbs, bradykinesia and resting tremor were also noticed during physical examination. Neuropathological examination confirmed a MND with predominant LMN involvement associated with scarce TDP-43 neuronal cytoplasmic aggregates. There was additional argyrophilic grain pathology (Saito stage I) and mild AD neuropathological changes: Braak neurofibrillary stage II, CERAD neuritic plaque score A and Thal amyloid phase 3, corresponding to an A2B1C1 score according to current guidelines.13 14
Possible pathogenic variants
A mutation (p.R197H) in the FIG4 phosphoinositide 5-phosphatase (FIG4) gene, which is not present in the NFE population from gnomAD, was found in a patient diagnosed with bvFTD at the age of 60 years and with ALS one year later. Neuropathological examination revealed a common phenotype of MND-TDP, with highly predominant LMN involvement, with prominent cell loss in brainstem and spinal motor nuclei and TDP-43 immunoreactive skein-like cytoplasmic inclusions. Isocortical involvement, other than a mild laminar superficial spongiosis present in multiple cortical regions, was limited to primary motor and premotor cortex. At this level mild to moderate astrogliosis was observed along with scarce TDP-43 positive long and tortuous neurites, glial and neuronal inclusions, mainly in cortical layers II and III. Occasional intranuclear inclusions were also observed. The hippocampus displayed a segmental sclerosis associated with neuronal TDP-43 inclusions in the dentate gyrus.
The other possible pathogenic variant consisted of a missense mutation (p.I666T) in the erb-b2 receptor tyrosine kinase 4 (ERBB4) gene. This rare variant (MAF in gnomAD NFE=1.79×10-5) was carried by a patient who presented with ALS at the early age of 36 and was diagnosed with bvFTD shortly afterwards. No further clinical details are available as this patient declined for follow-up.
We did not find any probable or possible pathogenic variant among the 26 remaining genes that were included in the study. Rare variants of unknown significance are summarised in supplementary table 2.
Our study suggests that one-fifth of patients with concurrent ALS and FTD lacking the C9orf72 hexanucleotide repeat expansion may carry a probable pathogenic variant in any of the 33 bona fide genes with prior evidence to play a role in ALS and/or FTD.
Mutations in TBK1 are the most important cause of ALS/FTD in our series, accounting for 9% of cases. Recently, mutations in TBK1 were found in 4.5% of subjects with concomitant ALS and FTD from Belgium24 and the same frequency was reported in an extended European cohort.18 This frequency reached 10.8% in an independent study performed in French cases.25 Taken together, these data strongly suggest that a comprehensive TBK1 genetic screening should be performed as a first choice in ALS/FTD cases where the C9orf72 expansion has been discarded.
We have identified three patients carrying two previously reported genetic alterations in SQSTM1 (one with the p.A33V and two with the p.P392L variant). Interestingly, one of them also presented a truncating mutation in TBK1. The p.P392L mutation is the most important genetic cause of PDB in Europeans and accounts for 15.6% of patients with sporadic PDB from Spain.26 Therefore, it might be expected that this missense variant could be encountered in a substantial proportion of patients with ALS/FTD from Spanish origin, especially when PDB is also present, as occurred in one of our cases. The high prevalence in the Spanish population (MAF=0.005) may indicate that it is a strong risk factor rather than a disease-causing mutation.
Our genetic screening disclosed two novel mutations in TAF15 (p. S390T and p.G462S), which are placed within the C-terminal RGG domain of the protein. Importantly, both patients showed a similar phenotype which included LMN predominant signs, bvFTD and movement disorders, and in one patient, neuropathology showed a FTLD pattern. Several missense rare variants in this protein domain (ie, p.G391D, p.R395Q, p.R408C, p.G452E and p.G473E) have been identified in sporadic and familial ALS cases.20 Interestingly, mutations such as the p.G391E, p.R408C and p.G437E have been shown to induce cytoplasmic foci in spinal cord neurons, to accelerate protein aggregation in vitro and to exacerbate the deleterious neurodegenerative effect that is present in flies expressing human TAF15.20 27 We were not able to find cytoplasmic aggregates of TAF15 in the case carrying the p.G462S variant. However, this negative finding must be interpreted with caution as postmortem delay, agonic factors, or formalin fixation, among others, might have negatively influenced the results. Our findings support a pathogenic role of TAF15 in ALS and expand its contribution to FTD and suggest that mutations in this gene might be a cause of FTLD with TDP-43 inclusions, in our case following a FTLD-TDP type B pattern.
Nearly all of the missense mutations described so far in TARDBP reside in exon 6, which encodes the C-terminal glycine-rich domain of the TDP-43 protein. The p.A90V variant in TARDBP encountered herein is located in the exon 3 and placed between the bipartite nuclear localisation signal sequence of the protein. This rare mutation has been extensively studied in vitro and has shown to negatively influence cell survival in neurons under stress conditions.28 Furthermore, cell expression of the TDP-43-A90V promoted its sequestration with endogenous TDP-43 as insoluble cytoplasmic pathological aggregates.21 However, the disruption of the nuclear localisation was only observed in a subset of transfected cells, thus suggesting a partial deleterious effect. The fact that this mutation has been described in a patient with a similar phenotype consisting of a very slow ALS/FTD progression21 prompts us to speculate that the p.A90V-driven incomplete disruption of TDP-43 cell localisation may give rise to a slower disease course—14 years in our case. Besides, although this genetic alteration is rare, it has a MAF of 0.0005 in the NFE population, which may suggest that it is a variant with incomplete penetrance, as has been reported for other TARDBP mutations.29 Neuropathological examination showed characteristic features of ALS-TDP with focal gliosis of the amygdala, a region involved in behavioural alterations.
Mutations in VCP have been associated with dissimilar clinical manifestations, including hereditary inclusion body myopathy (IBM) with PDB and early onset FTD (IBMPFD),30 ALS,31 hereditary spastic paraplegia32 and Charcot-Marie-Tooth disease type 2.33 In the present study, we have found the missense p.I27V variant in a patient with a definitive diagnosis of lower MND with accompanying argyrophilic grains and mild AD-pathology, which may justify the clinical bvFTD syndrome as no classical FTLD pattern was observed. This mutation has been reported in a case with sporadic, isolated inclusion body myositis,23 a case with a progressive speech disturbance with no evidence of myopathy and a patient with a typical bvFTD.22 Thus, the VCP gene might have pleiotropic effects and a particular mutation in VCP can give rise to different phenotypes. It is important to stress that the patient did not present the typical FTLD-TDP type D VCP-related pattern,12 thus opening the question as to whether some VCP mutations might give rise to non-TDP type D pathology. The fact that cells expressing this genetic alteration show a disruption of the autophagic properties (a typical consequence related to VCP misfunction) supports the pathological role of this particular rare non-synonymous change.23
FIG4 has been suggested to account for 1%–2% of patients with ALS.34 35 Our study expands the phenotype associated with FIG4 mutations to FTD. Interestingly, a patient with ALS with personality changes harbouring a FIG4 mutation has been previously described34 and pronounced frontoparietal atrophy and cognitive impairment has also been reported in two mutation carriers.35 A recent study has shown FIG4 mutations as the cause of autosomal recessive Yunis-Varón syndrome, a severe disease which is usually lethal at infancy.36 Notably, neuropathology associated with this disorder shows, among other features, severe neuronal loss and diffuse atrophy of the frontal lobes.36 Neuropathological examination of the mutation carrier reported in our study shows UMN and LMN involvement characteristic of ALS, with mild involvement of other isocortical regions. Hippocampal sclerosis, a common finding in frontotemporal lobe degeneration that is by itself associated with TDP-43 neuronal inclusions, was also observed in this case. These data suggest that FIG4 mutations might lead to TDP-43 aggregation.
Mutations in ERBB4 have been described in autosomal dominant familial and sporadic (due to a de novo mutation) forms of isolated ALS.37 To our knowledge, this is the first description of a novel ERBB4 variant in a patient with ALS that also suffered from a concomitant bvFTD. The mutation is only present in 2 out of 111 676 NFE chromosomes and absent from the healthy Spanish population consisting of 748 chromosomes. This finding supports a possible pathogenic role of ERBB4 in ALS and broadens its phenotypic consequences to FTD.
An oligogenic aetiology of ALS has been previously proposed.38 In this study, although discarding the C9orf72 repeat expansion carriers might result in an underestimation of this model, we report a patient with a probable pathogenic mutation in TBK1 and SQSTM1, which might have contributed to the lowest age at onset among TBK1 mutation carriers. Further studies are necessary to elucidate possible epistatic effects between loci and how they can contribute to disease risk and natural course.
It is important to note that 9 out of 11 potentially mutation carriers (81.8%) did not have a family history of ALS or FTD. These data suggest that family history of ALS and/or FTD is not always predictive of a causative mutation in this condition and genetic testing should be also offered to patients with sporadic ALS and FTD.39 Nevertheless, since this is a multicentre study in which clinical data have been ascertained retrospectively, there is a lack of a formal, homogeneous definition and evaluation of family history disease, which could underestimate the presence of disease in family members.39 Finally, from a neuropathological point of view, no bona fide distinctive morphological features suggested yet a specific underlying mutation, in contrast to what has been observed for C9orf72 mutations.40
Our data suggest that high-throughput sequencing approaches in uniform series of patients with concomitant ALS and FTD might be helpful to disentangle the genetic architecture of this two devastating disorders.
Contributors ODI, AGR, RRG, OGR, EG and JC processed and interpreted the data, wrote and revised the manuscript and designed the study. DBH, IIG, JLMB, AR, LCC, AJR, NS, NDL, LG, ECV, JF, RB, AL and JEP interpreted the data and revised the manuscript.
Funding Funding for the project was provided by the Spanish Ministry of Economy and Competitiveness, projects I+D+i 2008, Subprograma de actuaciones Científicas y Tecnológicas en Parques Científicos y Tecnológicos (ACTEPARQ 2009) and ERFD. ODI is funded by Departament de Salut de la Generalitat de Catalunya, Pla estratègic de recerca i innovació en salut (PERIS) 2016-2020 (SLT002/16/00040). PI15/00026 to JC and PI13/00772 and PI15/01618 to RRG jointly funded by Fondo Europeo de Desarrollo Regional (FEDER), Unión Europea, ‘Una manera de hacer Europa’. Telemaratón de RTVE ‘Todos somos raros, todos somos únicos’ (Project num. 29) also funded this study. This work was also supported in part by Generalitat de Catalunya (2014SGR-0235) and Fundació La Marató de TV3 under Grant 201437.10. IIG is supported by i-PFIS grant (IF15/00060) from Instituto de Salud Carlos III. National Registry of Motor Neuron Disease (NMD-ES Project) is partially funded by Centro de Investigación Biomédica en Red de Enfermedades Raras (CIBERER). This study makes use of data generated by the Medical Genome Project. A full list of the investigators who contributed to the generation of the data is available from http://www.medicalgenomeproject.com/en.
Competing interests None declared.
Provenance and peer review Not commissioned; externally peer reviewed.
If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.