Genotype-phenotype correlations of amyotrophic lateral sclerosis

Amyotrophic lateral sclerosis (ALS) is a devastating neurodegenerative disease characterized by progressive neuronal loss and degeneration of upper motor neuron (UMN) and lower motor neuron (LMN). The clinical presentations of ALS are heterogeneous and there is no single test or procedure to establish the diagnosis of ALS. Most cases are diagnosed based on symptoms, physical signs, progression, EMG, and tests to exclude the overlapping conditions. Familial ALS represents about 5 ~ 10 % of ALS cases, whereas the vast majority of patients are sporadic. To date, more than 20 causative genes have been identified in hereditary ALS. Detecting the pathogenic mutations or risk variants for each ALS individual is challenging. However, ALS patients carrying some specific mutations or variant may exhibit subtly distinct clinical features. Unraveling the respective genotype-phenotype correlation has important implications for the genetic explanations. In this review, we will delineate the clinical features of ALS, outline the major ALS-related genes, and summarize the possible genotype-phenotype correlations of ALS.


Background
Amyotrophic lateral sclerosis (ALS) is a devastating and inexorable neurodegenerative disease characterized by progressive neuronal loss and degeneration of upper motor neuron (UMN) and lower motor neuron (LMN) [1]. It is the most widespread type of motor neuron disease and has become the third most common neurodegenerative disease in the world [2]. Patients affected with ALS typically suffer from progressive muscle weakness and atrophy and usually die from respiratory failure 2 to 3 years after the onset [3]. ALS is a relentless and incurable disease. There has been no effective therapeutic approach to halt the progression so far. ALS is heterogeneous in its presentation, course, and progression. No single test for diagnosing ALS exists; most cases are diagnosed based on symptoms, physical signs, progression, EMG, and tests to exclude the overlapping conditions.
The etiology of ALS is not fully understood. Appropriate 5~10 % of ALS is familial (FALS) with a Mendelian pattern of inheritance, suggesting that genetic factors play important roles in the pathogenesis of ALS [4]. To date, more than 20 causative genes have been identified in hereditary ALS. In addition, about 30 potential causative or disease-modifying genes have also been identified. Detecting the pathogenic mutations or risk variants for each ALS individual is therefore challenging. Clinical features do not reliably separate the hereditary ALS from sporadic ALS (SALS) due to phenotypic overlap. However, ALS patients carrying some specific mutations may exhibit subtly distinct clinical features. Unraveling the genotype-phenotype correlations has important implications for the genetic explanations. In this review, we will delineate the clinical features of ALS, outline the major ALS-related genes, and review the possible genotypephenotype correlations. We hope this review will be in favor of improving the accuracy of genetic screenings.

Clinical features of ALS
ALS usually commences in later life, with a mean age at onset (AAO) of 65 year. The onset for FALS tends to be earlier than that of SALS. A small proportion of patients may developed juvenile ALS (JALS), in which onset occurs in the first two decades [5]. Symptoms usually begin in the limbs (termed limb onset), although approximate 25 % of ALS patients have bulbar onset. Associated with poorer prognosis, bulbar onset is more common in elderly patients and women [6]. Upper limb weakness and atrophy at onset are most common, subsequently spreading to involve bulbar, trunk, or respiratory muscle [7]. After initial presentation of symptoms, the disease progresses to include both UMN and LMN symptoms and signs. The UMN disturbance involving limbs leads to spasticity, weakness, and brisk reflexes. The LMN features of limbs comprise muscle atrophy, weakness, fasciculation, and decreased reflexes. As limb function deteriorates, patients gradually lose their ability to walk. Bulbar UMN dysfunction includes spastic dysarthria and brisk jerk of gag and jaw, while bulbar LMN dysfunction includes tongue wasting, weakness, and fasciculation. In the later stage of ALS, most cases develop dysphagia, which are associated with weight loss and malnutrition.
Sensory loss is usually absent, but cognitive impairment is common in ALS cases. A significant proportion of ALS cases develop cognitive dysfunction, and in a minority, overt dementia [7]. With appropriate cognitive assessment, 20~50 % of patients with ALS fulfill the consensus criteria for probable or definite frontal temporal dementia (FTD) [8]. In addition, about 30 % of FTD patients manifest signs of motor system dysfunction [9]. The occurrence of ALS, FTD, or ALS with FTD in intra-familial members discloses a considerable overlap between ALS and FTD.

Causative genes
Since the identification of first causative gene in 1993, a growing number of ALS-causing genes associated with Mendelian inheritance have been identified. Although FALS represents only 5~10 % of ALS cases, investigations of the causative genes have greatly increased our understanding of the etiology of ALS. FALS is generally inherited in an autosomal dominant pattern and rarely inherited as an autosomal recessive or X-linked trait [10]. Adult onset autosomal dominant inheritance is more common than juvenile onset which is usually caused by recessive transmission.
To date, at least 21 chromosomal regions containing 19 identified genes have been linked to ALS, termed as ALS 1-21 respectively (Table 1). In 2011, the chromosome 9 open reading frame 72 (C9ORF72) repeat expansions were identified in a significant proportion of ALS and FTD patients, becoming the most common genetic cause of ALS in the Caucasian population. There has been strong evidence to support the pathogenic role for mutations in Cu/Zn superoxide dismutase 1 (SOD1), Fused in sarcoma (FUS), TAR DNA-binding protein (TARDBP), and C9ORF72. However, mutations in other genes are not common or even rarely seen in ALS cases.
In addition to the aforementioned ALS genes, a constellation of other genes have also been implicated in ALS [11]. But mutations in these genes are only identified in a small fraction of ALS patients, indicating that these mutated genes have little contribution to the development of ALS.
Genotype-phenotype correlations ALS1: SOD1 The first causative gene of ALS was identified as SOD1 in 1993 [12]. Mutations in SOD1 are very common, accounting for about 20 % of FALS and 1~2 % SALS cases [13]. To date, more than 185 disease-associated mutations have been described, spread throughout all the 5 exons of SOD1 [14]. The majority of SOD1 mutations are missense mutations, while small deletions or insertions are also described [15]. The pattern of inheritance is autosomal dominant except for the p.D90A mutation which is recessive in the Scandinavian population and dominant in others [16]. Among the SOD1 mutations, p.D90A is the most common worldwide. However, regional disparity of SOD1 mutations also exists. For example, the most frequent SOD1 mutation in North America is p.A4V [17], but in the UK and Japan, the most common mutations are p.I113T, and p.H46R, respectively [18].
Overall, most patients with SOD1 mutations develop a rapidly progressive ALS, although some cases show a diverse phenotype. AAO and severity may vary significantly depending on the variants involved. Patients with p.A4V, p.H43R, p.L84V, p.G85R, p.N86S, or p.G93A mutations exhibit an aggressive form of ALS with survival shorter than 3 years, while cases with p.G93C, p.D90A, or p.H46R mutations show longer life expectancies [18]. Cognitive impairment is very rare and bulbar onset is less frequent than in other FALS types [19]. Some cases with SOD1 mutations have distinct clinical features. Patients carrying homozygous p.D90A mutation manifest insidious onset and a slow progression of ALS, with bladder involvement at the later stage [20]. In contrast, heterozygous p.D90A mutation is associated with various forms of ALS, including bulbar onset, upper-limb onset and fast progression, and lower-limb onset with fast progression [21,22]. The p.A4V mutation causes a limb-onset aggressive form of ALS, with survival less than 2 years after the onset [23]. Cases with p.A4T mutation have a similar phenotype to that seen in p.A4V mutation [24]. We previously reported SOD1 p.C111Y and p.G147D mutations in 3 Chinese ALS families. The p.C111Y mutation led to a relatively mild ALS phenotype, while the p.G147D was associated with a rapid progressive ALS [25]. The recently reported novel p.R115C mutation was identified in an ALS patient who had an extremely rapid progression and aggressive phenotype  [14]. Another novel p.T137A mutation identified in two unrelated Italian families, however, causes a very slow progression of ALS [26,27].

ALS2: Alsin
Mutations in ALS2 are responsible for autosomal recessive, early-onset forms of upper motor neuron diseases, such as infantile ascending hereditary spastic paraplegia (IAHSP) and juvenile primary lateral sclerosis (PLS) [28,29]. To date, more than 50 patients with early onset (~1 year) of the disease have been reported to harbor Alsin mutations [18]. In the typical adult onset ALS, Alsin gene is rarely mutated [30].
Recently, a novel splice-site mutation (c.3512 + 1G > A) in Alsin was identified in a consanguineous JALS family with early onset anarthria and generalized dystonia [31].

ALS4: senataxin (SETX)
ALS4 is a rare autosomal dominant form of juvenile-onset ALS due to mutations in SETX [32]. It is characterized by slowly evolving distal muscle weakness and atrophy, pyramidal signs, and sparing of bulbar and respiratory muscles [33]. In some cases, normal lifespan or atypical features are also described [34,35]. A patient with SETX p.R2136C mutation presented with coexistence of ALS and inflammatory radiculoneuropathy [36]. We previously reported a missense mutation p.T1118I in a sporadic Chinese ALS patient who developed bulbar symptoms 3 years after the onset, and respiratory failure 2 years later [37]. In addition, recessive SETX mutations are reported to cause ataxia and oculomotor apraxia type 2 (AOA2) [38,39].

ALS5: spatacsin (SPG11)
SPG11 mutations were recently identified in several juvenile-onset ALS cases, with autosomal recessive inheritance and AAO ranging from 7 to 23 years [18,40]. Generally, the SPG11-associated ALS showed a slow progression and in some cases apparent UMN involvement [41]. In addition to ALS phenotype, mutations in SPG11 also cause hereditary spastic paraplegia (HSP) with thin corpus callosum [42].

ALS6: FUS
Mutations in FUS gene have emerged as the second most common cause of ALS, accounting for about 35 % FALS and~1 % SALS [43,44]. Up to now, more than 60 mutations in FUS have been identified in ALS cases (http://www.hgmd.org, accessed in March, 2015). Among these mutations, the majority are clustered in exon 15 which encode the C-terminus of the protein, and the most common one is p.R521C [15]. The inheritance is autosomal dominant aside from an autosomal recessive p.H517Q mutation and several de novo mutations [44].
The phenotypes associated with FUS mutations include adult-onset ALS, JALS, ALS-FTD, and rarely pure FTD [45]. Although most patients carrying FUS mutations exhibit a classical ALS phenotype without cognitive impairment, the clinical course of these ALS cases are diverse, even among carriers of the same mutations. Compared to SOD1 patients, FUS-related ALS have an earlier AAO, more frequent bulbar disease, and a more rapid progression [46]. Some FUS mutations are also observed in patients with juvenile-onset ALS with AAO younger than 25 years [47,48]. Atypical features such as ALS with mental retardation, ALS with parkinsonism and dementia [46], and essential tremor were seen in some patients with FUS mutations [49].
ALS8: vesicle associated membrane protein B (VAPB) VAPB mutation was firstly described in several Brazilian families with motor neuron degeneration of various patterns: late-onset spinal muscular atrophy, atypical ALS, or typical ALS [50]. Subsequently, five other point mutations and a small deletion were described. Overall, VAPB mutations are extremely rare in FALS. The phenotypegenotype correlation remains largely undetermined so far.

ALS9: angiogenin (ANG)
The role of ANG mutations in ALS remains ambiguous. Approximate 30 ANG mutations have been reported in ALS, but only the p.K17II mutation is shown to cosegregate with the disease [51]. Mutations in ANG account for a small fraction of ALS cases. A few FALS cases are identified to have concomitant ANG mutations with FUS [52] or SOD1 [53] mutations. In addition, some ANG mutations are detected in healthy controls [54]. A subset of ANG mutation carriers showed cognitive impairment suggestive of FTD [55], or Parkinson's disease (PD) [56].

ALS10: TARDBP
Mutations in TARDBP encoding TDP-43 account for 4 % of FALS cases and~1 % of SALS [57,58]. More than 50 mutations have been identified so far (http:// www.hgmd.org, accessed in March, 2015), mostly clustered in the C-terminal region encoded by exon 6 of TARDBP. Mutations in TARDBP are predominantly missense with an autosomal dominant inheritance. Although TARDBP mutations are detected in ALS cases worldwide, some regional diversity does exist. For instance, the p.A382T mutation has been found in 28.7 % of all ALS cases in Sardinia [59]. Patients with TARDBP mutations usually exhibit a typical ALS phenotype, with limb or bulbar onset, variable disease course, and no overt dementia [60]. It is reported that the TARDBP-linked ALS has a trend for earlier AAO, more upper limb onset, and a longer duration, compared to SALS patients [52,61]. However, we previously reported a TARDBP p.S292N mutation in a FALS case who developed dysarthria, dysphagia, and atrophy of lingual muscle at the age of 64 years [62]. The progression of ALS in this case seemed to be rapid. Other phenotypes associated with TARDBP mutations include FTD [63], ALS-FTD [64], ALS with extrapyramidal signs [65], FTD with parkinsonism [66], and PD [67].

ALS12: optineurin (OPTN)
Previously identified as the cause of primary open angle glaucoma (POAG) [72], mutations in OPTN have been found in both FALS and SALS cases in either a dominant or recessive manner [73]. Although OPTN mutations are relatively common in Japan ALS cases [73,74], they are rare in Caucasian patients [75,76]. The OPTN-related ALS showed relatively slow progression and long duration before respiratory dysfunction [73]. In addition to ALS phenotype, some patients with OPTN mutations present with extrapyramidal symptoms, aphasia, or FTD [77,78].

ALS13: Ataxin-2 (ATXN2)
Long (more than 33) CAG repeat expansion in ATXN2 gene has been identified as a cause of spinocerebellar ataxia type 2 (SCA2) [79]. Recent studies demonstrated that intermediate expansion (27~33 repeats) of ATXN2 was a significant risk factor for ALS [80][81][82]. However, whether the clinical features of ALS can be affected by ATXN2 intermediate repeats is still controversial, because no correlation between ATXN2 repeat length and AAO or survival was observed [83]. A few case reports have described motor neuron degeneration in SCA2 families, raising the possibility that motor neuron involvement is part of SCA2 [84,85].

ALS14: Valosin-containing protein (VCP)
The gene of VCP is known to be mutated in inclusion body myopathy with Paget disease of bone and frontotemporal dementia (IBMPFD) [86]. Recently, mutations in VCP were identified in patients with FALS or SALS [87]. Actually, VCP mutations are not a major cause of ALS. Although more than 38 mutations in VCP have been discovered (http://www.hgmd.org, accessed in March, 2015), only a few of them are responsible for ALS. The phenotype of patients with VCP mutations shows intra-familial variations from IBMPFD to FALS [88], or from ALS to FTD or ALS-FTD [87].

ALS15: ubiquilin 2 (UBQLN2)
Mutations in UBQLN2 were recently identified in Xlinked dominant FALS [89]. However, UBQLN2 mutations are not a frequent cause of ALS [90]. In the affected cases, incomplete penetrance has been noted in females [89]. The predominant phenotype associated with UBQLN2 mutations is ALS, although several patients have concomitant symptoms of FTD [10]. The AAO has been reported to be significantly younger in males than in females, presumably because males are hemizygous but females are heterozygous for the mutation [89].
ALS16: sigma non-opioid intracellular receptor 1 (SIGMAR1) SIGMAR1 mutations were recently identified in families affected with juvenile ALS [91] or ALS with dementia [92]. However, these findings have not been replicated by other groups, suggesting that SIGMAR1 is a rare causative gene of ALS. Recently a splice-site mutation (c.151 + 1G > T) in SIGMAR1 was reported to cause autosomal recessive distal hereditary motor neuropathy in a consanguineous Chinese family [93].

ALS17: chromatin modifying protein 2B (CHMP2B)
Mutations in the CHMP2B gene were initially identified in patients with FTD [94] and then identified in patients with ALS [95,96]. Since only several ALS cases with CHMP2B mutations were reported, there was no characteristic ALS clinical subtype associated with these patients. In addition, CHMP2B mutation (p.R69Q) was also identified in progressive muscular atrophy (PMA) [96].

ALS18: profilin 1 (PFN1)
Missense mutations in PFN1 are firstly reported in two large ALS families and 7 FALS patients [97]. Later, screenings of sizeable ALS and FTD cohorts from diverse populations demonstrated that PFN1 mutations are a rare cause of ALS [98][99][100]. The reported cases with PFN1 mutations seem to present classical ALS with limb onset and no evidence of FTD [97,99,101].

ALS19: erb-b2 receptor tyrosine kinase 4 (ERBB4)
Recently, ERBB4 was identified a causative gene of FALS19, which was characterized by typical, slowly progressive ALS and a lack of obvious cognitive dysfunction [102]. However, this finding has not been replicated by other groups. The genotype-phenotype correlations are thus not determined.

ALS21: matrin-3 (MATR3)
Several mutations in MATR3 were recently reported to cause ALS [106]. Later, a heterozygous MATR3 p.A72T mutation was identified in a sporadic ALS patient with bulbar onset [107]. In another study, 2 splicing variants and a missense mutation were identified in 3 ALS cases with AAO ranging from 58 to 79 years, and bulbar onset in 2 cases [108].

ALS-FTD: C9ORF72
In 2011, two independent groups reported that the massive GGGGCC hexanucleotide repeat expansion in the non-coding regions of C9ORF72 gene caused chromosome 9p-linked ALS and FTD [109,110]. Currently, C9ORF72 repeat expansions have become the most frequently genetic cause of FALS and familial FTD, accounting for about 40 and 25 % of the cases, respectively [13]. In families with ALS-FTD, the frequency reaches to 50-72 % [19]. Notwithstanding, the mutations seem to be geographically clustered, accounting for one third of FALS cases in Europe and North America, but a small percentage in Asian populations. Haplotype analysis indicates that a common European founder appears to be responsible for all cases [111].
The C9ORF72 repeat expansions are associated with various phenotypes, including typical ALS, PMA, PLS, ALS-FTD, and pure FTD [44]. In patients with C9ORF72 mutations, bulbar onset and cognitive impairment seems to be more frequent, and median survival is relatively lower than in patients carrying TARDBP or SOD1 mutations [19]. However, there is no association between the repeat length and disease phenotype or AAO in C9ORF72 mutation carriers [112]. In addition, the C9ORF72 expansions also underlie a small portion of other neurological diseases, such as Alzheimer's disease (AD) [113], Huntington's disease (HD) [114], and PD [115].

Other genes implicated in ALS
Several other genes are also implicated in ALS. DCTN1 mutations were first identified in a family affected with LMN disease [116] and soon identified in several ALS families [117,118]. Subsequently, a cluster of DCTN1 mutations were identified in pedigrees with Perry syndrome [119], PD [120] or progressive supranuclear paralysis (PSP) [121], suggestive of phenotypic variability of DCTN1 mutations. Mutations in SQSTM1 were initially identified as a cause of Paget disease of bone (PDB) [122]. Recently, SQSTM1 mutations were identified in ALS [123,124] as well as FTD [125]. It is speculated that SQSTM1 mutations are mainly associated with lateonset SALS, because the number of early-onset patients with SQSTM1 mutations is much less than that of lateonset patients with SQSTM1 mutations [126].
In addition, mutations in several other genes such as DAO, UNC13A, NEFH, PRPH, TAF15, and ELP3, have been reported as rare causes of ALS. For some genes, there are no additional reports about the mutations as a cause of ALS since the initial publications. Therefore, the evidence supporting a causative role of ALS is not fully convincing.

Implication of the genotype-phenotype correlations
There is no a standard procedure to test the causative mutation in cases with ALS. Many factors should be considered, such as AAO, clinical features, progression, FTD involvement, inheritance manner, and even ethnicity. In this review, we summarized the possible genotype-phenotype correlations of ALS with the aim of providing some clue to improve the clinical decision. Here, we provided a flow diagram of genetic screening strategy in cases diagnosed with ALS (Fig. 1). AAO is an importance factor for the genetic investigations and can be divided into juvenile onset and adult onset. Juvenile onset is usually present in ALS1, ALS2, ALS4, ALS5, ALS6, ALS8, ALS9, ALS13, ALS15, and ALS16. For these cases with juvenile onset and UMN dominant symptoms, Alsin, SPG11, SIGMAR1, or UBQLN2 might be a causative gene. In contrast, FUS, VAPB, SOD1, and SETX should be considered in cases with juvenile onset and LMN dominant phenotype. For these cases with juvenile onset and FTD symptoms, ANG, UBQLN2, and SIGMAR1 can be investigated. In addition, SPG11 and FUS can be sequenced in cases who also present mental retardation, while SOD1, Alsin, SETX, ATXN2 can be considered in those cases with coexistence of cerebellar ataxia.
In cases with adult-onset ALS, many genes should be considered. Although the majority exhibit typical ALS features with both UMN and LMN symptoms, a fraction of cases present either UMN dominant symptoms or LMN dominant symptoms. SPG11, FIG4, UBQLN2 may be mutated in cases with UMN dominant symptoms. In these cases with LMN dominant symptoms, mutations in SOD1, SETX, FUS, VAPB, VCP, CHMP2B, hnRNPA1, or MATR3 might be identified. Presence of FTD hints possible mutations in ANG, TARDBP, OPTN, VCP, UBQLN2, CHMP2B, hnRNPA1, MATR3, or C9ORF72. Motor neuropathy involvement might occur in cases with mutations in SETX, VAPB, or FIG4. Complications of ataxia may be seen in cases with SOD1 or ATXN2 mutations.
Inheritance pattern also provides critical information as to the potential involvement of one or other specific genes in ALS. However, the diverse clinical features between intra-familial cases or a deceased parent at a young age may results in the appearance of "lack of familial history" in some cases. The presentation of ALS or FTD in first-degree relatives and the co-occurrence of ALS with FTD in some patients with ALS supported a positive family history. In addition, ethnic background should be considered when determining which genes are most likely. For example, C9ORF72 has been regarded as the most common cause of ALS in Caucasians, but very rare in Asian population.

Conclusions
Here, we outline the genotype-phenotype correlations of ALS. We hope this review is helpful for detecting the causative mutation in cases with ALS. Although a number of genes have been reported, many ALS cases still do not carry any mutation in the aforementioned genes. Other more genes might be identified as causative genes of ALS in the future.

Competing interests
The authors declare that they have no competing interests.
Authors' contributions L-HF drafted and W-ZY critically revised the manuscript. Both authors read and approved the final manuscript.