Introduction

Studies of disease concordance rates among monozygotic and dizygotic twins suggest that 53–84% of amyotrophic lateral sclerosis (ALS) population risk is genetically determined. *

* Refers to the proportion of disease risk that is genetically determined both within and across all individuals of a population, not the proportion of cases for which genetic risk factors are involved.

The presence or absence of a prior family history is used to divide ALS into familial (FALS) and sporadic (SALS) forms. Strict criteria for this division, in terms of minimal level of relatedness or consistency of phenotype, are not in place but the majority of observable FALS pedigrees would include only a small number of affected persons and are not usually associated with striking patterns of Mendelian segregation. In keeping with the elevated recurrence rate among relatives, mutations of major effect are believed to make a more significant contribution to FALS than to SALS burden. However, high-impact mutations are also observed within an important proportion of SALS cases. In these instances, the absence of a prior family history might reflect the small patient family size (providing limited opportunity for the observation of affected relatives), variant penetrance, ^†

† Proportion of individuals carrying a mutation that present with the associated clinical phenotype.

Amyotrophic Lateral Sclerosis Genes

Since the discovery of the first ALS gene in 1993, over 100 disease loci have been proposed to influence ALS susceptibility or clinical phenotype ( http://alsod.iop.kcl.ac.uk ). The pathogenic relevance of many of these is well supported ( Table 3.1 ), but whether and how most contribute to the ALS burden is far from certain. The difficulty in establishing clinical relevance is multifactorial, but stems primarily from the relatively low frequency of disease, the heterogeneity of causative factors, and the fact that every human genome contains a considerable number of potentially disease-related genetic variants. These issues can also complicate the interpretation of mutations observed at well-established disease genes, where factors such as gene size can mean the probability of observing entirely incidental patient variants is not negligible. Another issue concerning multiple ALS genes is that of pleiotropy, which refers to the association of one gene with multiple phenotypes. In the case of ALS, reported genes have also been associated with frontotemporal dementia (FTD), motor neuropathy, spastic paraplegia, progressive bulbar palsy, glaucoma, spinal muscular atrophy, spinocerebellar ataxia, oculomotor apraxia, and schizophrenia. In certain cases, even individual mutations can associate with multiple seemingly distinct clinical presentations. A prime example of this is a point mutation within the gene valosin – containing protein ( VCP) which, even within a single family, associates with variable combinations of ALS, FTD, Paget’s disease of bone, and inclusion body myopathy. Contrary to this, certain ALS mutations associate with very specific clinical profiles, such as the P525L mutation of fused in sarcoma (FUS) that is consistently observed in the context of aggressive early onset disease, and much of the clinical heterogeneity seen in the disease may reflect variation in causative as well as modifying factors.

The importance of individual ALS genes varies considerably according the ancestral background. Cumulatively, mutations within genes of major effect are identifiable in ~11% of patients of European ancestry (38–67% of FALS; 5–15% of SALS). Several common variants associated with ALS susceptibility have also been identified; however, interpretation of these variants is complicated by factors such as linkage disequilibrium ^‡

‡ Refers to the nonrandom association of two or more distinct genetic variants.

Protein Aggregation: Superoxide Dismutase 1

Superoxide dismutase 1 (SOD1) was the first gene identified to be mutated in a Mendelian type of ALS in 1993. The gene spans 9.3 kb on chr21q22.1, is composed of 5 exons, and encodes for the 153-amino-acid-long protein Cu/Zn superoxide dismutase, a cytoplasmic enzyme responsible for the catabolism of superoxide radicals to hydrogen peroxide and molecular oxygen. The protein is constituted of two identical monomers, each one consisting of an eight-stranded beta-barrel, and binding a copper and a zinc ion. SOD1 is ubiquitously expressed, highly conserved, and represents ~1% of all cytoplasmic proteins.

To date, more than 170 different SOD1 mutations have been reported ( http://alsod.iop.kcl.ac.uk ), the vast majority of which are missense substitutions equally distributed throughout the gene. Several indels, mainly clustered in the C-terminal region and leading to a premature truncation of the mature protein have also been described. Although the mutational frequency varies among different ethnicities, SOD1 accounts for ~20% of FALS, and 2–3% of SALS patients. The clinical phenotype is characterized by a considerable interfamilial and intrafamilial variability with regards to the age at onset, site of onset, and disease duration. Conclusive genotype–phenotype correlations are possible only for the most frequent SOD1 mutations. For instance, A4V is consistently associated with a high-penetrant, early-onset lower motor neuron disease with very rapid course. Conversely, H46R and G93D display a very mild phenotype, with carriers often surviving more than 20 years after disease onset. It must be noted, however, that the majority of the SOD1 variants described so far are private mutations, for which no genotype–phenotype correlation can be drawn. Moreover, the pathogenic role of some of these variants has recently started to be questioned. For instance, Felbecker et al. have described four families in which the E100K and D90A mutations do not segregate with the disease. These findings must be taken into serious consideration with regards to genetic testing and counseling in clinical practice. All SOD1 mutations are inherited as dominant traits, with the exception of the D90A variant, observed both in recessive pedigrees in Scandinavia, and in dominant pedigrees in the rest of the world. While dominant families develop classic ALS, recessive individuals have a milder phenotype characterized by upper motor neuron signs and prolonged survival, indicating the possible existence of a protective genetic factor in linkage with D90A in Scandinavian populations.

The pathogenic effects of SOD1 mutations are most likely not secondary to the abolition of the physiological functions of the wild type protein. In fact, several mutants retain full catalytic activity, and is there no correlation between residual enzymatic activity and/or protein stability with the disease phenotype. The observation that SOD1 knockout mice do not develop motor neuron disease, while transgenic animals overexpressing the human mutant SOD1 gene do, also argues against a loss-of-function hypothesis. Conversely, it is believed that SOD1 mutations result into the acquisition of a novel function that is toxic to motor neurons. Several studies showed that mutant SOD1 is prone to misfolding and forms cytoplasmic aggregates. In turn, aggregates may lead to cell death by sequestering other cytoplasmic proteins essential for neuronal survival, by clogging the ubiquitin/proteasome system, by chaperone depletion, or by disrupting mitochondria, the cytoskeleton, and/or axonal transport. Interestingly, there is evidence that posttranslational modifications may induce misfolding and increase aggregation propensities also of wild-type SOD1, thus suggesting a possible pathogenic role also in SALS.

Dysfunction of mRNA Metabolism: TAR DNA-Binding Protein and Fused in Sarcoma

A major step into the understanding of ALS pathogenesis has been the discovery of the protein TAR DNA-binding protein 43 (TDP-43) as the main component of ubiquitinated cytoplasmic inclusions in ALS and in frontotemporal lobe dementia with ubiquitin inclusions (FTLD-U). TDP-43 is a 43-kDa, 414 amino-acid-long multifunctional DNA/RNA binding protein encoded by the TAR DNA-binding protein ( TARDBP ) gene, and belonging to the heterogeneous ribonucleoprotein (hnRNP) family. After this breakthrough discovery, mutations in the TARDBP gene have been found to be a major cause of Mendelian ALS in several populations of different geographic origins. Subsequently, disease-causing mutations were identified also in the FUS gene, which encodes for another RNA binding protein, thus suggesting that an alteration of mRNA homeostasis may represent a key event in ALS pathogenesis. Mutations in TARDBP and FUS account for ~5% of FALS and 1% of SALS cases each.

TDP-43 and FUS share many structural, functional, genetic, and neuropathological similarities. Both proteins have been demonstrated to play a role in several biological processes, including transcriptional regulation, splicing, nucleo-cytoplasmic shuttling, transport and stabilization of mRNAs. In the central nervous system, both proteins are involved in mRNA transport toward the dendrites and in regulating synaptic plasticity.

Given their biologic role, TDP-43 and FUS contain similar protein domains. TDP-43 contains two highly conserved RNA recognition motifs, flanked by an N-terminal domain and a C-terminal tail. The latter element contains a glycine-rich region that is reputed to mediate protein–protein interactions, mainly with others hnRNPs. FUS is composed of an N-terminal transactivating domain, a central domain that contains a RNA recognition and a zinc finger motif, and a C-terminal region rich in arginine and glycine. Both proteins have nuclear localization (NLS) and nuclear exporting (NES) signals, allowing for nucleo-cytoplasmic shuttling.

The vast majority of TARDBP mutations are clustered in the C-terminal glycine-rich region of the protein, while pathogenic FUS variants are located within or disrupt the NLS. The observation of cytoplasmic inclusion immunoreactive for TDP-43 or FUS in carriers indicates that mutations decrease their solubility, thus increasing aggregation propensity . Moreover, while in unaffected neurons both proteins localize in the cell nucleus, they are absent from the nuclei of inclusion-bearing cells, suggesting a nucleo-cytoplasmic redistribution . These observations lead to intense speculation on the pathogenic role of TDP-43 and FUS in ALS: toxicity might be caused by the aggregating protein being sequestered away from its normal nuclear function (loss-of-function hypothesis) or, conversely, insoluble aggregates might have a toxic gain-of-function independent of the protein’s physiological cellular activities. These findings collectively suggest that TARDBP – and FUS -mediated toxicity in ALS and FTD may occur through common cellular pathways, thus underscoring the need for a comprehensive identification of overlaps and differences between the mRNA target binding profiles of the two. Notwithstanding these similarities, it must be noted that TDP-43 positive inclusions are observed not only in mutation carriers, but also in the vast majority of SALS and in ~55% of FTLD-U cases. Conversely, FUS aggregates are observed exclusively in ALS patients harboring mutations and in a minority of nonmutated FTD cases displaying atypical neuropathological phenotypes.

Significant differences between the two genes also exist at the phenotypic level. TARDBP -mutated patients usually develop a motor neuron disease indistinguishable from classic ALS with regards to age at onset, disease duration and distribution of upper and lower motor neuron signs ( http://alsod.iop.kcl.ac.uk ). The onset of the disease is often spinal, with a preferential involvement of the upper limbs. Since most TARDBP mutations are private, it is difficult to establish clear genotype–phenotype correlation. It has been suggested that A382T, which is the variant most commonly observed, may be associated with a low penetrant, predominantly lower motor neuron disease with an asymmetrical onset in the distal muscles of the limbs, subsequently spreading to proximal muscles, with relative sparing of the bulbar muscles. This variant has also been observed in patients with atypical phenotypes including parkinsonism and/or ataxia. Conversely, FUS mutations are consistently associated with a severe type of motor neuron disease characterized by onset in the third or fourth decade, very rapid disease course, and predominant lower motor neuron signs. Mutations affecting the NLS, in particular R521C, may result in an uncommon phenotype characterized by a symmetrical proximal spinal onset, with early involvement of the axial muscles and head drop.