Genetics of familial amyotrophic lateral sclerosis

Familial vs sporadic ALS.

From a clinical standpoint, familial (FALS) and sporadic (SALS) cases cannot be distinguished from one another, apart from a mean age at onset for SALS that is 10 years later than for FALS (56 years vs 46 years).¹¹ Thus family history and genetics are the primary factors that discriminate between SALS and FALS. SALS cases comprise the large majority (90%) of all ALS cases. The remaining cases have a positive family history and are considered to be FALS cases; in many instances, however, the disease does not transmit in a typical dominant or recessive manner. Instead, the large part of these “familial” ALS cases consists of a proband with one or two first- or second-degree relatives who are also affected. This is what is observed in families with ALS collected by us and other researchers. A large percentage of family pedigrees include only two or three people with ALS, while only a few pedigrees have many affected individuals (figure 1). The few large and fully penetrant pedigrees of ALS families in the world that are reported, ranging from 4 to 20 affected individuals, are consequently the ones used for linkage analysis studies.^12-18 These identify shared genetic markers in affected individuals of a family and thus indicate regions on a chromosome where a causative gene for ALS exists. In other non-ALS disorders, familial cases are often subdivided into heritable forms with essentially fully penetrant pedigrees and familial forms that have clusters of cases in the same extended family. To date, no controlled studies have been conducted that prospectively evaluate the number of affected individuals in non-SOD1 families with ALS. Controlled studies could serve to estimate the relative risk of developing ALS in primary and secondary relatives of someone with ALS.

• Download figure
• Open in new tab
• Download powerpoint

Figure 1 Number of affected patients in families with ALS

Individuals with familial amyotrophic lateral sclerosis (ALS) were collected by us and collaborators from France and Canada, and the number of additional clinically proven ALS cases in the family was recorded (unpublished results). In more than 50% of familial cases, only one additional family member apart from the proband had ALS. This finding underlies the point that relatively few large families with several individuals with ALS are available for genetic studies.

Twin and adoption studies.

Twin studies are often quite informative with regard to the genetic contribution to a particular trait, especially when this trait is non-Mendelian with no apparent inheritance pattern. A large twin study has been conducted in the United Kingdom, and its results helped to underline the genetic contribution to ALS. In this study, 4 of 26 monozygotic and 0 of 51 dizygotic twins developed ALS, which indicated that the genetic contribution was between 35 and 85%.¹⁹ Altogether, this study provided evidence of the genetic contribution toward ALS cases and warrants follow-up twin studies from additional populations. However, in a disease where a fraction of cases are clearly hereditary, including two of the four monozygotic twins mentioned above, one must be cautious in generalizing these data to all of ALS. It would have been preferable if the concordant monozygotic twins were tested for known ALS genes, especially SOD1. Adoption studies of patients with ALS are required to help delineate the environmental component of the disease. However, these studies would be harder to interpret because disease onset occurs much later in life. In any event, no such study has been conducted.

Genes mutated in ALS: SOD1.

The greatest contribution toward an understanding of ALS thus far has come from the discovery of mutations in the SOD1 gene on chromosome 21q22.11,²⁰ which account for 10–20% of autosomal dominant FALS cases.^21,22 Tremendous efforts have been made not only to understand the multiple likely pathogenic mechanisms of SOD1 in eliciting disease, but also to comprehend the genetic profile of the reported mutations in this gene. A number of conclusions can be drawn from this research:

1. SOD1 mutations result in a toxic gain-of-function pathology. Evidence for this arose largely from mouse studies in which the knockout of the SOD1 gene failed to yield a phenotype²³; conversely, transgenic mice that overexpressed mutant SOD1 did develop a motor neuron phenotype.^24-26

2. The loss of the normal function of SOD1 is not the cause of ALS. Mutations in the SOD1 gene have a broad range of effects on the enzymatic activity of SOD1; however, among the mutated SOD1 proteins reported, some noticeably retained full enzymatic activity. The rest of the mutations were shown to influence, among other things, the stability of SOD1, its ability to dimerize, its hydrophobicity, and its ability to chelate copper ions.²⁷ Different SOD1 mutations may also influence various factors of the disease such as its onset or duration, but they nonetheless cause ALS.

3. Mutations can occur at almost any position in the SOD1 gene. One of the most remarkable aspects of the SOD1 gene is that more than 110 mutations have been reported in nearly 50% of the 153 amino acids in the SOD1 protein. The distribution of these mutations is quite uniform: the largest interval without a reported mutation is only nine consecutive amino acids.

4. Mutations are primarily dominant, with some noteworthy exceptions. The D90A mutation appears in dominant state in certain pedigrees, but in the Swedish population it is actually a polymorphism which primarily causes ALS when in the homozygous state.²⁸ A compound heterozygous case has also been reported with D90A and D96N mutations.²⁹

Considering that SOD1 is the only gene known to be mutated in a substantial number of ALS cases, a thorough comprehension of the mechanism by which SOD1 mutations cause the disease is essential to delineate its pathogenesis.

Non-SOD1 adult-onset dominant FALS.

Novel genes that were recently identified or that remain to be identified might help in a better understanding of novel disease mechanisms for ALS. A mutation in the vesicle-associated membrane protein (synaptobrevin-associated protein) B (VAPB) gene has been identified in seven families on chromosome 20.³⁰ However, only one of the identified families had an individual with classic ALS and three families were instead affected only with spinal muscular atrophy (SMA).³⁰ Furthermore, few reports confirm this gene, which suggests that it is at best a very rare cause of ALS.

The ALS3 locus has been identified on chromosome 18q21.¹³ Genetic analysis of this family pedigree with its 18 affected individuals results in a maximum lod score of 5.36 (P.N.V. and G.A.R., unpublished data). Three other groups reported families mapping to chromosome 16q12 and designated this locus as ALS6.^12,14,15 The ALS7 locus on chromosome 20q13 was reported at the same time, although only two affected individuals were used for genetic analysis and the results should consequently be interpreted with caution.¹⁵ A form of ALS linked to the X-chromosome has also been reported in abstract form.³¹

ALS and frontotemporal dementia.

Frontotemporal dementia (FTD) is the only reported disease which is consistently observed in a subset of ALS families. Estimates of FTD prevalence in ALS cases range between 5 and 15%.^32,33 Examination of pedigrees with both ALS and FTD cases as affected has led to the identification of three additional loci. One locus is at chromosome 9q21-q22, and it is referred to as ALS/FTD.³⁴ Another locus was recently identified on the p-arm of chromosome 9 (9p21.2-9p13.3) in five pedigrees.^16-18 Sequence variants have been identified in the IFT74 gene in ALS patients,³⁵ although mutations have not been identified in other linked pedigrees (P.N.V. and G.A.R., unpublished data). The microtubule-associated protein tau (MAPT) gene on chromosome 17q21 has been implicated in ALS pathology, including its hyperphosphorylated presence in inclusions in ALS patients.³⁶ Mutations in MAPT have also been identified in other neurodegenerative disorders.^37,38 Additional genes have been identified that are mutated FTD. Mutations in the CHMP2B gene on chromosome 3p11.2 cause FTD and contribute to ALS disease,^39,40 while mutations in the progranulin (PGRN) gene on chromosome 17q21.31 which cause FTD^41,42 have been extended to an FTD patient with extensive family history for ALS.⁴³ Both ALS and FTD have similar inclusions which have recently been determined to both be immunopositive for the TAR DNA-binding protein TDP-43.⁴⁴ How mutations in the MAPT, CHMP2B, and PGRN genes can give rise to more than one disorder is not clear. However, their identification in multiple families using traditional genetic approaches validates their deleterious role in FTD, and the handful of ALS cases with mutations implies that they represent either a very rare form of the disease or instead a susceptibility allele for developing ALS.

ALS-related disorders and genes.

Two genes were identified for a juvenile-onset form of ALS, which can be found in both dominant and recessive forms. ALS2 is a form of autosomal recessive ALS in which symptoms occur in the first or second decade of life and then progress slowly for 10 to 15 years. Linkage was established to chromosome 2q33,⁴⁵ and the alsin gene was shown to be causative.⁴⁶ Patients with diagnosed hereditary spastic paraplegia (HSP) and primary lateral sclerosis (PLS) were also found to harbor mutations in alsin.^46-48 ALS4 is an autosomal dominant juvenile-onset form of ALS that maps to chromosome 9q34.⁴⁹ The ALS4 patients display symptoms indicative of ALS, but they also have additional pathologic findings.⁵⁰ Three different coding changes in the senataxin (SETX) gene were identified for ALS4.⁵¹ Another autosomal recessive juvenile-onset locus, ALS5, was identified which maps to chromosome 15q15.1-q21.⁵² An autosomal dominant family with PLS has been reported, which has not mapped to a known locus.⁵³ Finally, a family was studied with a slowly progressing autosomal dominant form of lower motor neuron disease. Mutations in this family were subsequently identified in the p150 subunit of dynactin (DCTN1).⁵⁴ This finding was extended to identify three additional DCTN1 mutations in one SALS and two FALS cases.⁵⁵

Penetrance considerations for Mendelian forms of ALS.

The occurrence of compounding factors such as a missed diagnosis of ALS in a relative or a death at an early age that was due to a non-ALS–related cause may mask clear Mendelian segregation in some families. Even for SOD1, penetrance remains incomplete and individuals who carry SOD1 mutations have an 80% chance of developing the disease by age 85 years.⁵⁶ Other pedigrees that exhibit high penetrance for ALS include those described for ALS3, ALS6, and ALS8.^12-15,57 In families where both ALS and FTD is observed, the rapid progression of FTD with a disease duration of 8 to 10 years in a particular individual may result in death before the onset of ALS symptoms, or it may alternatively mask ALS symptoms, particularly in bulbar-onset forms of the disease. This issue is addressed in pedigrees with ALS and pedigrees with ALS and FTD in which various models of reduced or incomplete penetrance are considered.^13-18 The danger in using reduced-penetrance models is that they can complicate genetic studies and lead to false-positive linkage results.

The genetic component of ALS compared with other similar disorders.

ALS is a unique disease with regard to its clinical and genetic profile. From a pathophysiologic perspective, it resembles HSP, SMA, and spinal and bulbar muscular atrophy (SBMA) since upper or lower motor neurons or corticospinal tracts are mostly affected in these disorders. However, only one gene is primarily responsible for SMA (the survival of motor neuron 1 gene)⁵⁸ and SBMA (the androgen receptor gene).⁵⁹ Conversely, many more loci (33 vs 12 in ALS) and genes (15 vs 5 for typical and atypical ALS) have been implicated in HSP cases.⁶⁰ This is in part due to the slower progression of HSP, which means that multiple generations of affected individuals can be collected simultaneously. If the same number of individuals could be collected in an ALS pedigree, it would not be surprising if the number of loci and genes that are identified to be causative for ALS will be in fact comparable with the number identified for HSP.

The distinction between familial and sporadic forms is often observed in other non–motor neuron neurodegenerative diseases. The familial component to ALS cases at less than 10% is quite comparable with what is observed in Alzheimer disease⁶¹ and Parkinson disease.⁶² The number of causative genes identified for these two diseases provides hope that similar successful gene mutation identifications are possible for ALS. It is important to note that the few mutated genes identified recently in ALS (ALS2, VAPB, SETX), which affect few ALS cases worldwide, might still provide key clues about the mechanisms of disease. An example of this is in Parkinson disease, where the α-synuclein gene accounts for only a handful of cases, yet the knowledge of pathologic mechanisms gleaned from its discovery is paramount.⁶³

Genes involved in hypoxia and oxidative stress.

Compelling evidence for a gene associated with ALS came from studies of the vascular endothelial growth factor (VEGF) gene. Deletion of the hypoxia response element in the promoter region of this gene was shown to cause an ALS-like progressive motor neuron degeneration in mice.⁶⁵ Two haplotypes from three polymorphisms in the human VEGF promoter have similarly been shown to increase the risk of ALS. Furthermore, lentiviral delivery of VEGF was shown to improve the survival time of SOD1 mutated mice.⁶⁶ Related experiments resulted in the detection of seven amino acid changes in the angiogenin (ANG) gene in a group of Irish and Scottish ALS patients.⁶⁷ The rationale for the selection of angiogenin was its similar function to VEGF, and moreover its physical proximity, on chromosome 14, to the apurinic/apyrimidinic exonuclease (APEX) gene. With regard to APEX, a D148E polymorphism was found to be associated with ALS.⁶⁸

The cluster of paraoxonase (PON) genes has also been examined for an increased susceptibility to ALS. In one study, three nonsynonymous single nucleotide polymorphisms (SNPs) which affect function were examined, two in PON1 and one in PON2.⁶⁹ The minor allele at amino acid 192 in PON1 and amino acid 311 of PON2 was found to be significantly associated with Polish ALS cases vs controls. These particular SNPs were not observed to be associated in a second study; however, two additional SNPs in the introns of PON2 and PON3 were indeed found to be associated.⁷⁰ In a third study, a SNP in the promoter region of PON1 was most significantly associated with SALS in the Australian population.⁷¹ While these results are preliminary and require further replication, the identification of coding polymorphisms with an associated risk for ALS lends support to the role of genes involved in oxidative stress in ALS susceptibility.

Genes involved in cytoskeletal structure.

Impaired neurofilament assembly has long been regarded as a hallmark of ALS.⁷² This has been extended to genetic studies whereby mutations in the neurofilament heavy chain subunit (NF-H), and particularly Lys-Ser-Pro (KSP) repeat motif deletions, were identified in SALS cases.^73,74 In addition, a mutation in the peripherin gene was identified in one patient with ALS.⁷⁵ Finally, three mutations in patients with ALS have also been identified in the p150 ^Glued subunit of dynactin (DNCT1).⁵⁵

Genes involved in motor neuron survival.

The identification of mutations and deletions in the SMN gene as responsible for the lower motor neuron disease SMA has led to the examination of this gene in ALS patients. However, mutations were not identified when directly sequencing the gene in 135 SALS and 17 FALS cases.⁷⁶ Homozygous deletions of the SMN1 gene were not identified in ALS patients, and homozygous deletions of SMN2, which are polymorphic in the general population, have been found to be overrepresented in only one of several studies examined.⁷⁷ In an extensive study of Dutch ALS cases, one copy of SMN1 correlated with an increased risk of developing ALS.⁷⁸ In addition, overall SMN (SMN1 + SMN2) expression levels conferred an increased risk of developing ALS as well as increasing the severity of the disease.⁷⁸ Variable levels of SMN1 copy number (one or three copies) have also been considered to be a risk factor for developing SALS.⁷⁹

The discovery that ciliary neurotrophic factor (CNTF) null mice have motor neuron degeneration, and that the additional knockout of the leukemia-inhibitory factor (LIF) gene enhanced this degeneration, led to the examination of these two genes in the context of ALS.^80,81 For CNTF, the gene has been primarily considered as a modifying factor, specifically for the onset and progression of disease in a family bearing a SOD1 mutation.⁸² An association of a V64M variant in LIF has also been reported in ALS patients.⁸³

Genes associated with other neurodegenerative disorders.

The idea has been considered that the apolipoprotein E gene, with its convincing association with Alzheimer disease and increased expression in late-stage G93A mutant SOD1 mouse spinal cords,⁸⁴ might also play a role in ALS. Past studies show that the epsilon-4 allele is associated with the development of bulbar-onset ALS⁸⁵ or age at onset⁸⁶ or survival.⁸⁷ Other studies, however, failed to observe this association.^88-90 Polymorphisms in the hemochromatosis (HFE) gene have also been examined in ALS patients after an association was identified between HFE and Alzheimer disease. A meta-analysis of the association studies performed for the H63D HFE polymorphism showed an overall increased risk (OR of 2.7 for H63D homozygotes) of developing ALS.⁹¹

Gene expression microarrays also have the potential to identify genes which can be further evaluated for association with ALS. These studies also have the important advantage of no a priori selection of candidate genes, and when mouse models are used, they allow an examination of differential mRNA expression at various stages of disease; however, the background associated with these types of experiments is relatively high and often makes the results difficult to reproduce. Gene expression studies have been performed in neuronal-like NSC34 cell lines,⁹² spinal cords of mice,^84,93 and spinal motor neurons and the motor cortex of patients with SALS.^94-96 These studies have proposed multiple sets of genes that are upregulated or downregulated in ALS, such as those involved in vesicle trafficking, cytoskeletal maintenance, proteasome function, apoptosis, and immune response, among others. The number of studies performed is continually increasing, with the hope that a few key pathways and genes will emerge as significant across different experimental conditions and replicates.

De novo mutations.

Another possibility for the high frequency of sporadic individuals is that a new mutation arises over the course of reproduction. If such a mutation was sufficiently deleterious, it would result in the appearance of an apparently sporadic individual who would not have affected parents. This could account for some of the familial cases which do not have an extensive disease history. Given the late age at onset for ALS patients, reproductive fitness is rarely compromised. This means that new mutations introduced into the population would not be under selective pressure and would therefore remain in the gene pool. This is best exemplified in SOD1, where it has taken many generations to accumulate these mutations. Furthermore, in SOD1 the distribution of mutations is quite extensive and only one mutation, A4V, is relatively common, accounting for up to 50% of cases in the North American population.¹¹⁴ Therefore, this is unlikely to be a frequent mechanism for ALS.

Chromosomal rearrangements.

Chromosomal rearrangements and copy number polymorphisms can also be observed as segregating, de novo, or somatic mutations. A few examples of chromosomal rearrangements have been identified in small subsets of sporadic or apparently sporadic ALS and lend credibility to the idea that these chromosomal rearrangements affect gene expression or function and thus cause disease.^115,116 Still, the number of genes implicated in unbalanced translocations and their often drastic consequences on viability and health may explain why cytogenetic abnormalities are not more common in ALS.