Objectives Amyotrophic lateral sclerosis (ALS) is a neurodegenerative disease mainly involving cortical and spinal motor neurones. Molecular studies have recently identified different mutations in the ubiquilin-2 (UBQLN2) gene as causative of a familial form of X-linked ALS, 90% penetrant in women. The aim of our study was to analyse the UBQLN2 gene in a large cohort of patients with familial (FALS) and sporadic (SALS) amyotrophic lateral sclerosis, with or without frontotemporal dementia (FTD), and in patients with FTD.
Methods We analysed the UBQLN2 gene in 819 SALS cases, 226 FALS cases, 53 ALS–FTD patients, and 63 patients with a clinical record of FTD. Molecular analysis of the entire coding sequence was carried out in all FALS and ALS–FTD patients, while SALS and FTD patients were analysed specifically for the genomic region coding for the PXX repeat tract. Healthy controls were 845 anonymous blood donors and were screened for the PXX repeat region only.
Results We found five different variants in the UBQLN2 gene in five unrelated ALS patients. Three variants, including two novel ones, involved a proline residue in the PXX repeat region and were found in three FALS cases. The other two were novel variants, identified in one FALS and one SALS patient. None of these variants was present in controls, while one control carried a new heterozygous variant.
Conclusions Our data support the role of the UBQLN2 gene in the pathogenesis of FALS, being conversely a rare genetic cause in SALS even when complicated by FTD.
- Motor Neuron Disease
Statistics from Altmetric.com
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.
Amyotrophic lateral sclerosis (ALS) is a late-onset disease caused by motor neurone degeneration, currently with no effective therapies.1 ALS is the third most common neurodegenerative disease in the western world,2 with an incidence of 1.5–2.7 per 100 000 population/year, and a male : female ratio of 1.5 : 1.
A number of genes (SOD1, TARDBP, FUS, C9ORF72) have been discovered as causative for approximately 50–65% of FALS cases.5–10 Mutations in at least 10 other genes have also been reported as rare causes of ALS or ALS-like syndromes.11–13
Dementia, usually of the frontotemporal lobar type (FTD), may occur in some ALS cases. The identification of a hexanucleotide repeat expansion in the C9ORF72 gene in a high percentage of cases in both ALS and FTD disorders suggests that a common genetic cause may contribute to both phenotypes.9 ,10
Mutations in the PXX repeat region of the UBQLN2 gene have recently been identified in five large families with ALS and ALS/dementia presenting with a dominant X-linked transmission mode.14 Interestingly, aggregates of ubiquilin 2 protein were described in spinal motor neurones not only of patients carrying UBQLN2 mutations, but also in all SALS and FALS cases analysed. Moreover, ubiquilin 2-positive inclusions were frequently associated with TDP-43, FUS and optineurin proteins.14
In order to elucidate further the contribution of ubiquilin 2 to neurodegeneration we screened the UBQLN2 gene in a large cohort of Italian ALS and a smaller group of FTD patients.
Patients and controls
Patients in this study were collected by six ALS centres participating in the Italian SLAGEN Consortium. All patients were of Italian descent. The diagnosis of ALS was made according to the El Escorial revised criteria.15 Familial history was considered positive if the proband had one or more first or second-degree relatives with ALS, according to the recent criteria proposed for FALS classification.3 The diagnosis of FTD was made according to the criteria proposed by Neary et al16 and Strong et al.17 Our cohort included 819 SALS, 226 FALS, 53 ALS–FTD (of which five had a positive familial history) and 63 FTD patients.
SALS patients (61.8% men) had a mean age at onset of 52.9±14.6 years. Of note, the age at onset in our SALS cohort is lower than reported in the literature.1 This finding can be explained by the fact that our cohort has been collected at third level referral centres for ALS care and genetic analysis, and thus has more younger onset cases. FALS patients (n=226) included 96 individuals carrying mutations in other ALS causative genes (FALS–GEN) (age at onset 50.0±9.3 years) and 130 FALS cases genetically undetermined (FALS–UNK) (age at onset 52.7±14.4 years). Patients presenting with FTD without signs of motor neurone disorder (n=63) were 44.4% men and had an age at onset of 63.5±9.5 years. Healthy controls were 845 anonymous blood donors of Italian origin, who were collected after obtaining information on personal and familial history of neurological diseases. Controls were matched for age and geographical origin with cases.
We received approval for this study from the ethics standard committees on human experimentation of the participating institutions. Written informed consent was obtained from all individuals participating in the study.
Amplification of the intronless UBQLN2 gene, including part of the 5′UTR (146 bp) and 3′UTR (311 bp) sequences, was obtained as previously described.14 Sequence analysis was performed using the ABI PRISM 3130 XL genetic analyser.
Nucleotide numbering of UBQLN2 gene mutations reflects complementary DNA numbering with +1 corresponding to the A of the ATG translation initiation codon of the GenBank reference sequence NM_013444.3. The initiation codon is methionine 1.
Molecular analysis was carried out for the entire UBQLN2 gene in all FALS and ALS–FTD patients (n=279), while all SALS and FTD patients (n=882) were tested only for the region spanning the PXX repeat domain (465 bp).
UBQLN2 analysis was performed in all FALS cases independently on male-to-male transmission in order to detect potential de-novo mutational events and/or concurrent UBQLN2 variants in patients carrying mutations in other ALS causative genes.
Genetic analysis of the UBQLN2 gene revealed the presence of five different variants in five ALS patients, four of which were familial (two FALS–UNK and two FALS–GEN) and one sporadic (table 1). The mutations were identified in heterozygous status in three women, and in hemizygous status in two men.
All variants were missense mutations and determined different amino acid substitutions, all occurring in evolutionarily conserved residues in mammals (figure 1). Three mutations involved a proline residue and were located in the PXX repeat region (c.1490C>A (p.P497H); c.1516C>T (p.P506S); and c.1598C>T (p.P533L)), while the other two were just upstream (c.1337T>G (p.M446R)), and downstream (c.1612G>C (p.V538L)) of the PXX motif (table 1 and figure 1).
Among the five variants identified in ALS patients, the p.P497H has previously been reported by Deng et al,14 while the other four are novel.
In our study, the p. P497H mutation was identified in a man also presenting with FTD signs who had developed bulbar ALS at the age of 33 years.
The p.P506S mutation, identified in a man presenting with an early onset (30 years) ALS, alters the proline 506 in the PXX repeat region, a residue that has already been found to be mutated to threonine in a family reported by Deng et al.14 Similarly to p.P506T, p.P506S is predicted by the PMUT software to have a mild pathological effect.
The c.1598C>T mutation resulting in the p.P533L substitution also involves a proline residue closely located to the 12 PXX repeat region (491–526) and, according to the PMUT and SNAP algorithms, is likely to disrupt protein function. This mutation has been identified in a young woman who developed ALS at the age of 34 years. Interestingly, this patient also carried the previously reported mutation p.Q314L in the OPTN gene.23
The c.1337T>G variation resulting in the p.M446R substitution was identified in a woman with a disease onset at the age of 43 years. In-silico analysis using PMUT and SNAP programs predicted that this mutation might have a possibly damaging effect on the mutated protein. In this patient we had previously reported the p.M337V mutation in the TARDBP gene.20
The c.1612G>C p.V538L variant has been detected in a sporadic case presenting with the first symptoms at 57 years. Both PMUT and SNAP programs predicted that this is a neutral variant, which is expected be tolerated at the protein level.
We could not assess the segregation of the identified variants with disease status in FALS pedigrees because no other affected family members were available for DNA testing.
All the variants identified in ALS patients were not present in 845 DNA samples from healthy Italian individuals. Nonetheless, we found a missense mutation (c.1316A>T) leading to the amino acid substitution p.N439I in heterozygous form in one control subject. This mutation was predicted to have a functional pathological effect on the resulting protein by PMUT and SNAP analysis. (table 1 and figure 1). Control subject TO94 is an anonymous blood donor younger than 60 years, with no personal or familial history of neurological diseases.
Our study also detected the two exonic single nucleotide polymorphisms (SNP) c.1383G>A, (p.G461G) and rs45559331 (c.1461C>A (p.T487T)) in a limited number of patients and controls (table 2). These SNP have already been reported at a very low allele frequency in a systematic, large scale re-sequencing screening of X-chromosome coding exons in families with mental retardation25 and in the 1000 Genomes Project database (May 2011 release) (table 2).26
The main clinical features of ALS patients carrying the identified UBQLN2 mutations are reported in table 1. Briefly, FALS patients carrying UBQLN2 mutations showed a very early age of onset (average 35 years, range 30–43) compared to the rest of the FALS cohort (average 52 years, range 20–85). On the contrary, the SALS patient with the possibly neutral p.V538L variant displayed an age of onset similar to the other SALS. The disease progression showed large variability within mutated cases (range 6 months–5 years) (table 1). The clinical phenotype was dominated in all cases by the presence of widespread upper motor neurone signs. In particular, patient B785 developed a slowly progressive ascending spastic paralysis without clinical or neurophysiological signs of lower motor neurone involvement, and was ultimately diagnosed with primary lateral sclerosis (PLS). His mother had classic ALS, while, similar to the proband, the sibling presented with an early-onset, slowly progressive motor neurone disease. Our cohort included 30 patients with PLS (male : female ratio 1 : 1), four of which were familial. The average age at onset was 59±7 years, and 33% of PLS patients had a bulbar onset. Similar to what reported by Deng et al,14 who identified UBQLN2 mutations in eight ALS–FTD patients, we observed cognitive and behavioural disturbances in two out of five of our mutated patients (AB489 and B92). In patient AB489 the neuropsychological evaluation was consistent with a diagnosis of a behavioural variant of FTD, with prominent apathetic and anhedonic features.
Conversely, we did not identify any UBQLN2 mutation in patients with pure FTD.
The recent identification of UBQLN2 gene mutations in FALS patients,14 some of them also presenting with FTD, was a remarkable finding as ubiquilin 2 protein appears to be a prominent component of the cellular aggregates typically occurring in the spinal motor neurones of ALS patients.
In particular, inclusions positive for ubiquilin 2 were found not only in FALS cases carrying UBQLN2 mutations, but also in SALS and FALS individuals carrying mutations in other genes.20–23 Interestingly, ubiquilin 2 was frequently described to co-localise with TDP-43, FUS and optineurin proteins within aggregates.14 A distinct neuropathological pattern, characterised by the presence of cytoplasmic inclusions immunoreactive for ubiquilin 2 has recently also been observed in patients carrying C9ORF72 repeat expansions.27
Ubiquilin 2, a 624 amino acid-long protein, is a member of the ubiquitin-like protein family. In humans, four distinct genes encode separate proteins sharing common domains. The structural organisation of all members of the family is characteristic of proteins that deliver ubiquitinated proteins to the proteasome for degradation. Ubiquilins are characterised by the presence of an N-terminal ubiquitin-like domain and a C-terminal ubiquitin-associated domain. The ubiquitin-like domain binds to the proteasome, while the ubiquitin-associated domain binds to the poly-ubiquitin chains conjugated for degradation by the proteasome. The middle part of ubiquilins is variable and specific for each family member.28 Ubiquilin 2 has a unique repeat region containing 12 PXX tandem repeats.29 ,30 Interestingly, all the previously identified mutations involved a proline residue in the unique PXX repeat region of the protein.14
In this study, we have analysed UBQLN2 in a large Italian cohort of ALS patients, in order to elucidate further the contribution of mutations within this gene to the pathogenesis of ALS and FTD.
We have found five different variants in the UBQLN2 gene in five unrelated ALS patients. These variants affected highly conserved amino acids located within or in the proximity of the PXX repeat domain. Three of them (two novel and one previously reported by Deng et al)14 involved proline residues and were identified in FALS cases. The other two variants, p.M446R and p.V538L, located outside the PXX repeat domain, were found in one FALS and one SALS patient, respectively.
Remarkably, mutations identified in FALS patients (p.P497H, p.P506S, p.P533L and p.M446R) are all predicted to have a deleterious effect on the mutated proteins indicating an important role in the pathogenesis of ALS. By contrast, p.V538L seems to have a milder effect on phenotype, as also suggested by the in-silico analysis that showed a neutral effect on protein function. In addition, p.V538L has been found in a SALS case with a higher age of onset compared to the remaining mutated subjects, possibly indicating a reduced phenotypic expression. For these reasons, we believe that p.V538L needs further validation in other ALS cohorts, the demonstration of segregation in affected pedigrees, and/or of a pathogenic effect in disease models.
The finding of UBQLN2 variants in two FALS women carrying mutations in other ALS causative genes (OPTN and TARDBP) is of particular interest. Although the exact contribution of each mutation to the phenotype cannot be determined because segregation could not be proved for any of them, it is possible that in these two cases ALS might have a complex genetic aetiology involving at least two genes. In fact, recent studies point to an oligogenic basis for some ALS cases.31 ,32 In our FALS cohort, we previously observed the concurrence of pathogenic mutations in different genes in the same patient, such as SOD1 and ANG33 or TARDBP and C9ORF72.1 It must be remarked, however, that heterozygous mutations in UBQLN2 in women may contribute to a lesser extent to the phenotype than heterozygous variants in dominantly transmitted autosomal genes. As segregation cannot be proved, however, we cannot exclude that one of the two mutations represents a benign rare variant.
In an analytical perspective, this finding suggests also including patients carrying known pathogenic mutations when screening for novel ALS genes.
None of the variants identified in our ALS patients were present in 845 control samples of the same geographical origin, nor were they reported in the 1000 Genomes Project database. It is remarkable that the analysis of the control group has disclosed the presence, in a female control subject, of a UBQLN2 gene variant (p. N439I) with a possible pathological effect.
In total, we have identified six missense mutations in the UBQLN2 gene, four of them in a panel of 226 FALS patients. We therefore estimate that approximately 2% of FALS cases within the Italian population are caused by a mutation in the UBQLN2 gene. This frequency is similar to that reported in the original paper first describing the mutations,14 and in the recent study of Williams et al,34 but higher than that reported in other screening on a French FALS cohort35 and on French-Canadian FALS and SALS patients.36
The identification of different variants, four of them with predicted damaging effects on the protein function, supports the hypothesis that UBQLN2 may play an important role in the pathogenesis of FALS. Our results also suggest that patients carrying mutations in the UBQLN2 gene share a distinct clinical phenotype characterised by early age at onset, prominent upper motor neurone dysfunction, and, in some cases, the presence of a cognitive impairment of frontal type.
In summary, our results suggest that UBQLN2 mutations are responsible for approximately 2% of cases in our cohort of Italian FALS. Although the exact role of ubiquilin 2 is still unclear, it seems likely that its association with the ubiquitin proteasome system (UPS) plays an important role in ALS pathogenesis. Interestingly, ALS-associated mutations were found in several genes encoding for UPS-related proteins (UBQLN2, OPTN and VCP) suggesting that different genotypes may converge on the same phenotype through common pathways.37 The discovery of new UBQLN2 mutations may thus provide the basis for functional studies aimed at evaluating the pathological effects of different mutants on the UPS and on inclusion formation in neurodegenerative diseases.
The authors would like to thank the Italian patients with ALS and their caregivers for support.
Contributors Writing team: CiGe, NT, CT, AR, VP, VS, all others received and approved the manuscript. Patient selection and record collection: NT, VS, AR, CiGe, FT, LM, AC, GS, GR, SC, AB, CaGa, GPC. Preliminary genetic screening, sequencing and genotyping: CT, VP, BC, RDB, LC, JA, CC, PM. Data analysis: CiGe, NT, CT, AR, VP, RDB, LC, CC, JA, SDA, BC. Scientific planning and direction: VS, NT, CiGe, FT, AR, SD, GPC.
Funding This work was supported by AriSLA, grant EXOMEFALS 2009, and co-financed with the support of ‘5×1000’—Healthcare Research of the Ministry of Health (CiGe, CT, VP, BC, NT and VS). AB is supported by a fellowship from ‘Amico Canobio’ Association. SDA is partially supported by Fondazione Cariplo, grant no 2010-0728 and PRIN 08.
Competing interests None.
Ethics approval Approval for this study was received from the ethics standard committees on human experimentation of the participating institutions.
Patient consent Obtained.
Provenance and peer review Not commissioned; externally peer reviewed.