**3.2 Ataxin-2 (ATXN2)**

In a screen for toxicity modifiers of TDP-43 in yeast, ataxin-2 (ATXN2) was identified (Elden et al., 2010). ATXN2 and TDP-43 form a RNA-dependent complex and are mislocalized in spinal cord motor neurons in ALS patients. *ATXN2* has a polyglutamine (polyQ) region which is normally 22-23 repeats long. Expansion of this region of the protein to 34 repeats causes spinocerebellar ataxia type 2 (SCA2) (Imbert et al., 1996; Pulst et al., 1996; Sanpei et al. 1996). The polyQ repeat length of *ATXN2* was determined in 915 ALS patients and 980 controls and intermediate length polyQ repeats (23-34) were found to be more common in ALS patients and thus may be a risk factor for ALS (Elden et al., 2010). This finding was replicated in several studies with ALS patients from different populations. Interestingly, the exact length of the polyQ repeat region seems to vary between populations (Chen et al., 2011; Daoud et al., 2011; Lee et al., 2011; Ross et al., 2011; Van Damme et al., 2011).

Longer polyQ repeats in ATXN2 possibly stabilize the protein and enhance its interaction with TDP-43. Under stress conditions, increased mislocalization of TDP-43 to the cytoplasm was observed in cells harbouring expanded polyQ repeats in ATXN2 (Elden et al., 2010). ATXN2 was shown to be part of stress granules and interacts with poly-A-binding-protein 1 (PABP), which is involved in poly(A) shortening and translation initiation (Ralser et al., 2005). ATXN2 was also shown to interact with endophilin A1 and A3, which are involved in synaptic vesicle endocytosis (Nonis et al., 2008).

### **3.3 Chromatin modifying protein 2B (CHMP2B)**

A mutation in a splice-site of *CHMP2B* was first identified in a large Danish family with FTD and mutations have since been detected at low frequency in other FTD patients (Skibinski et al., 2005). Screening of the *CHMP2B* gene in ALS patients identified two mutations in two fALS patients. These patients displayed a predominant lower motor neuron phenotype and one of the patients showed signs of FTD (Parkinson et al., 2006). Sequencing of the *CHMP2B* gene in 433 ALS patients identified three missense mutations in one fALS case and three sALS cases (Cox et al., 2010).

The exact function of CHMP2B is unknown but its yeast homologue, vacuolar protein sorting 2 (VPS2), is a component of the ESCRTIII complex (Skibinski et al., 2005). This complex is involved in the trafficking of proteins between plasma membrane, trans-Golgi network, and lysosomes. The *CHMP2B* mutation identified in FTD results in dysmorphic endosomal structures similar to what is seen in ALSin overexpression (Skibinski et al., 2005). In cortical neurons, overexpression of the FTD related CHMP2B splice-site mutant leads to dendritic retraction prior to cell death and the accumulation of autophagosomes (Lee et al., 2007). In hippocampal neurons, the same FTD related CHMP2B mutant causes a decrease in large dendritic spines suggesting that CHMP2B is needed for dendritic spine growth and maturation (Belly et al., 2010).

### **3.4 Haemochromatosis (HFE)**

488 Amyotrophic Lateral Sclerosis

APEX1 is involved in DNA repair and maintains and stimulates the DNA binding activity of transcription factors (Fishel and Kelley, 2007). Frontal cortical levels and activity of APEX1 were significantly reduced in 11 ALS patients as compared to six controls (Kisby et al., 1997). However, in a different study, increased expression levels and activity in ALS brain and

In a screen for toxicity modifiers of TDP-43 in yeast, ataxin-2 (ATXN2) was identified (Elden et al., 2010). ATXN2 and TDP-43 form a RNA-dependent complex and are mislocalized in spinal cord motor neurons in ALS patients. *ATXN2* has a polyglutamine (polyQ) region which is normally 22-23 repeats long. Expansion of this region of the protein to 34 repeats causes spinocerebellar ataxia type 2 (SCA2) (Imbert et al., 1996; Pulst et al., 1996; Sanpei et al. 1996). The polyQ repeat length of *ATXN2* was determined in 915 ALS patients and 980 controls and intermediate length polyQ repeats (23-34) were found to be more common in ALS patients and thus may be a risk factor for ALS (Elden et al., 2010). This finding was replicated in several studies with ALS patients from different populations. Interestingly, the exact length of the polyQ repeat region seems to vary between populations (Chen et al., 2011; Daoud et al., 2011; Lee et al., 2011; Ross et al.,

Longer polyQ repeats in ATXN2 possibly stabilize the protein and enhance its interaction with TDP-43. Under stress conditions, increased mislocalization of TDP-43 to the cytoplasm was observed in cells harbouring expanded polyQ repeats in ATXN2 (Elden et al., 2010). ATXN2 was shown to be part of stress granules and interacts with poly-A-binding-protein 1 (PABP), which is involved in poly(A) shortening and translation initiation (Ralser et al., 2005). ATXN2 was also shown to interact with endophilin A1 and A3, which are involved in

A mutation in a splice-site of *CHMP2B* was first identified in a large Danish family with FTD and mutations have since been detected at low frequency in other FTD patients (Skibinski et al., 2005). Screening of the *CHMP2B* gene in ALS patients identified two mutations in two fALS patients. These patients displayed a predominant lower motor neuron phenotype and one of the patients showed signs of FTD (Parkinson et al., 2006). Sequencing of the *CHMP2B* gene in 433 ALS patients identified three missense mutations in

The exact function of CHMP2B is unknown but its yeast homologue, vacuolar protein sorting 2 (VPS2), is a component of the ESCRTIII complex (Skibinski et al., 2005). This complex is involved in the trafficking of proteins between plasma membrane, trans-Golgi network, and lysosomes. The *CHMP2B* mutation identified in FTD results in dysmorphic endosomal structures similar to what is seen in ALSin overexpression (Skibinski et al., 2005). In cortical neurons, overexpression of the FTD related CHMP2B splice-site mutant leads to dendritic retraction prior to cell death and the accumulation of autophagosomes (Lee et al., 2007). In hippocampal neurons, the same FTD related CHMP2B mutant causes a decrease in large dendritic spines suggesting that CHMP2B is needed for dendritic spine growth and

spinal cord motor neurons were observed (Shaikh and Martin, 2002).

**3.2 Ataxin-2 (ATXN2)** 

2011; Van Damme et al., 2011).

maturation (Belly et al., 2010).

synaptic vesicle endocytosis (Nonis et al., 2008).

**3.3 Chromatin modifying protein 2B (CHMP2B)** 

one fALS case and three sALS cases (Cox et al., 2010).

Mutations in the *HFE* gene are a cause of hereditary haemochromatosis and have been associated with Alzheimer's disease and PD (reviewed by Nandar and Connor, 2011). The first report examining the presence of *HFE* mutations in ALS found no association between two mutations (H63D and C282Y) and ALS patients from the USA (Yen et al., 2004). However, several subsequent studies in a total of 1133 ALS patients and almost 7000 controls individuals from the USA, Ireland, UK, Italy, The Netherlands, and China reported association between the *HFE* H63D polymorphism and an increased risk for ALS (Goodall et al., 2005; He et al, 2011; Restagno et al., 2007; Sutedja et al., 2007; Wang et al., 2004).

The most important function of HFE is the regulation of iron homeostasis by binding to the transferrin receptor and reducing the transport of iron molecules (Feder et al., 1998). When HFE with the H63D mutation binds to the transferrin receptor, iron transport is reduced leading to iron accumulation and increased oxidative stress. In addition, it has been shown that in neuronal cell lines the H63D mutation induces increased oxidative stress, altered glutamate regulation and prolonged ER stress, all cellular processes affected in ALS (Liu et al., 2011; Mitchell et al. 2011).

### **3.5 Neurofilaments (NEFL, NEFM, NEFH)**

One of the pathological hallmarks of ALS is the presence of neurofilament-containing inclusions in the cell body and proximal axon of spinal motor neurons (Delisle and Carpenter, 1984). Neurofilaments are intermediate filaments that constitute the most abundant cytoskeletal element in large myelinated axons. Neurofilaments are formed by the co-polymerization of light (NEFL), medium (NEFM), and heavy (NEFH) subunits, which are each encoded by different genes.

Several lines of evidence suggest a role for neurofilaments in neurodegeneration. Initial evidence came from mouse models overexpressing or deficient for neurofilaments (reviewed in Lariviere and Julien, 2004). Overexpression of NEFL or NEFH resulted in an abnormal accumulation of neurofilaments, as seen in ALS patients, and in axonal atrophy and motor dysfunction but not degeneration. Surprisingly, both overexpression and knockout of neurofilaments in transgenic mutant SOD1 mice increases life span (Couillard-Després et al., 1998; Williamson et al., 1998). This indicates that the role of neurofilaments in ALS is complex and more research is needed to examine the possible contribution of neurofilaments to ALS pathogenesis.

Additional evidence for a role for neurofilaments in ALS comes from genetic studies. Mutations in *NEFL* have been identified in some forms of the sensory and motor neuropathy Charcot-Marie-Tooth disease (Mersiyanova et al., 2000; Shin et al., 2008). The C-terminal tail region of NEFH contains phosphorylation motifs known as KSP repeats. In humans there are two common polymorphic variants of 44 (short) or 45 (long) repeats. Homozygosity for the short repeat allele is associated with Russian sporadic motor neuron disease patients (Skvortsova et al., 2004). Deletions and insertions in the KSP repeats of *NEFH* were detected in ALS patients (Al-Chalabi et al., 1999; Figlewicz et al., 1994; Tomkins et al., 1998). However, another study in 117 unrelated fALS patients could not identify deletions or insertions in the KSP repeats of *NEFH* (Rooke et al., 1996). A missense mutation in the *NEFH* gene was identified in a sALS case and not in controls (Garcia et al., 2006). Moreover, in a recent candidate gene sequencing study, three missense mutations were identified in the *NEFH* gene in two sALS and one fALS case. However, co-segregation of the mutation in the

Genetics of Amyotrophic Lateral Sclerosis 491

ALS patients was investigated. Two missense mutations and a frameshift deletion in the PRPH gene have been identified in sALS patients (Corrado et al., 2011; Gros-Louis et al., 2004; Leung et al., 2004). Additional screening of the *PRPH* gene for mutations in larger cohorts of ALS patients and controls is needed to determine the frequency and

Expression of abnormal peripherin splice variants has also been suggested to play a role in ALS pathogenesis. A toxic splice variant of peripherin (Per61) was found in motor neurons of mutant SOD1 transgenic mice but not wild-type mice (Robertson et al., 2003). Expression of Per61 has more recently also been observed in mutant TDP-43 transgenic mice but not in wild-type TDP-43 transgenic mice (Swarup et al., 2011). In addition, Per61 specific antibodies stain aggregates in human ALS but not in control spinal cord (Swarup et al., 2011). The presence of abnormal peripherin splice variants (Per28) has also been shown in humans (Xiao et al., 2008). Per28 overexpression results in peripherin aggregation and an upregulation of peripherin expression at the mRNA and protein levels in ALS patients as compared to controls (Xiao et al., 2008). A different study showed expression of Per28 in lumbar spinal cord lysates of ALS patients but not control cases (McLean et al., 2010). Although the functional significance of these abnormal splice forms is unknown they seem

Two highly homologous copies of the survival motor neuron gene exist in humans, telomeric *SMN1* and centromeric *SMN2*. *SMN2*, which lacks exon 7 due to a nucleotide difference in a splice enhancer site, produces a less stable SMN protein and has only 20% of the biological function of SMN1 (Lorson et al., 1998). It has been shown that TDP-43 overexpression regulates the inclusion of exon 7 during pre-mRNA splicing of *SMN2* (Bose

Deletions or mutations in *SMN1* cause the autosomal recessive disorder spinal muscular atrophy (SMA), whereas variation in *SMN2* copy number affects SMA disease severity (Lefebvre et al., 1997). SMA patients with a higher copy number of *SMN2* generally have a milder form of the disease (Gavrilov et al., 1998). SMN1 is widely expressed and functions in the assembly of the spliceosome as part of the SMN complex. SMN1 also interacts with several proteins involved in mRNA editing, transport, splicing, transcriptional regulation, and post-transcriptional processing and modification of rRNA (Eggert et al., 2006). The impaired assembly of the spliceosome could lead to neuronal

Thus far, five different studies have failed to detect homozygous *SMN1* deletions in ALS patients (Gamez et al., 2002; Jackson et al., 1996; Moulard et al., 1998; Orrell et al., 1997; Parboosingh et al., 1999). However, an increased frequency of abnormal copy number (one or three copies) of *SMN1* was found in ALS patients compared to controls (Corcia et al., 2002). However, these results were inconsistent with other reports (Corcia et al., 2006; Veldink et al., 2001; Veldink et al., 2005). Recently, a large study was published including new samples of 847 sALS patients and 984 controls, showing that *SMN1* duplications were associated with ALS susceptibility (odds ratio [OR] = 2.07, 95% confidence interval [CI] = 1.34 - 3.20. (Blauw et al, 2011)). A meta-analysis of all previously published data, taking possible heterogeneity between studies into account, confirmed this association with *SMN1* duplications. Other work has shown that homozygous deletions of *SMN2* are associated

pathogenecity of *PRPH* mutations.

to play a role in the development ALS.

et al., 2008).

degeneration.

**3.8 Survival motor neuron (SMN) 1 and 2** 

fALS case could not be tested and none of the missense mutations were predicted to be deleterious (Daoud et al., 2011). One study did not identify ALS specific variation in the *NEFH* gene in fALS and sALS samples (Vechio et al., 1996).

### **3.6 Paraoxonase genes (PON)**

The paraoxonase gene cluster consists of 3 genes (*PON1, PON2,* and *PON3*) and is located in an 80-kb block on chromosome 7q21.3-22.1. PON1 and PON3 are primarily expressed in liver where they are associated with high-density lipoproteins, whereas PON2 is ubiquitously expressed (Costa et al., 2005; Draganov et al., 2000; Ng et al., 2002). Both PON1 and PON2 expression has been shown in mouse brain (Giordano et al., 2011). All PON proteins are able to hydrolyze lactones and PON1 is able to detoxify organophosphate pesticides and neurotoxins. Since neurotoxins are not normally present in the body the biological function of PON1 is thought to be protection of low-density lipoproteins from oxidation (Mackness et al., 1991). PON2 and PON3 share this function (Draganov et al., 2000; Ng et al., 2001). A higher incidence of ALS among Gulf war veterans and farmers suggested that chemical exposure may be a risk factor for ALS (Chió et al., 1991; Horner et al., 2003). Because PON proteins reduce oxidation and are able to detoxify neurotoxins these proteins have been investigated for association with ALS.

Polymorphisms in *PON1* and *PON2* as well as a haploblock spanning *PON2* and *PON3* were found to be associated with sALS (Saeed et al., 2006; Slowik et al., 2006). Since then several other studies in different populations have reported association of SNPs in the *PON* genes with sALS (Cronin et al., 2007; Landers et al., 2008; Morahan et al., 2007; Valdmanis et al., 2008). However, a meta-analysis including 4037 cases and 4609 controls from five casecontrol studies and several genome-wide association studies showed no significant association between *PON* polymorphisms and ALS (Wills et al., 2009). More recently, two other studies failed to detect association between *PON* polymorphisms and ALS (Ricci et al. 2011; Zawislak et al., 2010). In a recent sequencing study, eight mutations in all three *PON* genes were identified in fALS and sALS patients (Ticozzi et al., 2010). Mutations in the *PON* genes might play a role in ALS but additional sequencing is needed to confirm this.

Interestingly, PON1 activity can vary greatly depending on polymorphisms in its coding region (Costa et al., 2005). Thus, mutations in the *PON* genes could affect PON activity and thereby contribute to ALS pathogenesis. Toxicity in neurons caused by oxidative stress was higher in cells from PON2 knockout mice than in wild-type mice, suggesting that PON2 has a protective effect against neurotoxicity caused by oxidative stress (Giordano et al., 2011).
