Keywords

5.1 Introduction

In 1991, expansion mutations of repeat sequences in the genome were discovered to cause human hereditary diseases, namely, a CGG trinucleotide repeat expansion in the fragile X mental retardation 1 (FMR1) gene causing fragile X syndrome (FXS) and a CAG trinucleotide repeat expansion in the androgen receptor (AR) gene in spinal-bulbar muscular atrophy (SBMA) (La Spada et al. 1991; Verkerk et al. 1991). Since these initial findings, more than 23 expansion mutations of 3 or more nucleotide repeats were found to cause various inherited neurological and neuromuscular diseases (Table 5.1) (La Spada and Taylor 2010). These repeat expansion disorders are largely classified into two groups depending on the location of the repeat sequences in the genome, i.e., the coding region or the noncoding region.

Table 5.1 Clinical and molecular characteristics of repeat expansion disorders in which repeat sequences are located in the coding region of a gene
Full size table

In the former group, expanded CAG repeats produce proteins containing an expanded polyglutamine (polyQ) tract, triggering neurodegeneration via toxic gain-of-function mechanisms in Huntington’s disease (HD); spinocerebellar ataxia (SCA) types 1, 2, 3, 6, 7, and 17; dentatorubral-pallidoluysian atrophy; and SBMA, which are collectively called the polyQ diseases (Katsuno et al. 2014; Takeuchi and Nagai 2017). Expansions of the polyQ tract are thought to trigger misfolding and aggregation of these causative proteins, eventually causing neurodegeneration. Expansion mutations of GCN repeats encoding a polyalanine (polyA) tract have also been reported in oculopharyngeal muscular dystrophy (OPMD) and other diseases, which can lead to both gain-of-function and loss-of-function pathogenic mechanisms (Messaed and Rouleau 2009).

In the latter group, the repeat sequences are located in the noncoding region, such as the 5′-UTR, 3′-UTR, or introns in the genome in FXS and fragile X tremor ataxia syndrome (FXTAS); fragile XE syndrome (FRAXE); myotonic dystrophy (DM) types 1 and 2; Friedreich ataxia; SCA8, 10, 12, 31, 36, and 37; C9orf72-linked amyotrophic lateral sclerosis and frontotemporal dementia (C9-ALS/FTD) ; and Huntington’s disease-like 2 (Table 5.2) (Orr and Zoghbi 2007; Rohilla and Gagnon 2017; Seixas et al. 2017). Since these repeat sequences do not directly encode amino acid sequences in proteins, their pathogenic mechanisms are much more complicated. At least three molecular mechanisms underlying the pathogenesis of these noncoding repeat expansion diseases have been proposed (Nelson et al. 2013; Rohilla and Gagnon 2017). First, loss-of-function of the mutant genes due to silencing of or reduction in gene expression by the repeat expansion mutation has been suggested in FXS, FRAXE, and Friedreich ataxia (Pieretti et al. 1991; Bidichandani et al. 1998). Second, gain-of-function due to aberrant RNAs containing expanded repeats transcribed from the mutant gene have been suggested in most of these diseases, including DM1 and 2, FXTAS, SCA8, 10, 31, and 36, and C9-ALS/FTD . These expanded repeat-containing RNAs were shown to be accumulated as RNA foci in affected tissues and to recruit their corresponding RNA-binding proteins (RBPs), resulting in their loss-of-function (Miller et al. 2000; Mankodi et al. 2001; Jin et al. 2007). Furthermore, a third mechanism has emerged from recent studies, in which expanded repeat RNAs were surprisingly shown to be translated into aberrant repeat polypeptides despite the lack of an initiation codon, via unconventional translation, so-called repeat-associated non-ATG (RAN) translation (Zu et al. 2011; Mori et al. 2013; Ash et al. 2013; Pearson 2011). Subsequent studies demonstrated that these repeat polypeptides produced by RAN translation cause toxicity via gain-of-function mechanisms (Kwon et al. 2014; Mizielinska et al. 2014). However, the molecular mechanisms of RAN translation still remain to be understood, and research toward elucidation of the pathogenic mechanisms of these disorders is still ongoing.

Table 5.2 Clinical and molecular characteristics of repeat expansion disorders in which repeat sequences are located in the noncoding region of a gene
Full size table

In this chapter, we will introduce a number of studies using a variety of fly models to elucidate the pathogenic mechanisms of these repeat expansion disorders. We will also discuss the advantages of fly models as human disease models for studying pathogenic mechanisms and investigating potential therapies for these disorders.

5.2 Fly Models of Repeat Expansion Disorders

5.2.1 Huntington’s Disease

Huntington’s disease (HD) is an autosomal dominant neurodegenerative disorder characterized by loss of neurons mainly in the striatum and cortex, leading to progressive motor impairments, cognitive decline, and psychiatric symptoms. HD is caused by an abnormal expansion of CAG repeats encoding the polyQ tract in exon 1 of the huntingtin (Htt) gene. In the polyQ diseases, such as HD, there is a threshold length of polyQ repeats for clinical manifestation of approximately 35 to 40 repeats in general, and longer repeats are associated with earlier age of onset and severity of disease. The abnormal aggregation of mutant proteins into nuclear inclusions (NIs) is also commonly observed in the brains of patients with the polyQ diseases (DiFiglia et al. 1997; Becher et al. 1998; Paulson et al. 1997).

Jackson et al. (1998) first generated fly models of HD, each expressing the exon 1 fragment of the Htt protein with tracts of either 2, 75, or 120 glutamine residues (Httex1-Q2, Q75, or Q120, respectively) in photoreceptor neurons of the eye. Expression of Httex1-Q2 had no effect on the fly eyes, whereas expression of Httex1-Q75 or Q120 caused repeat length- and age-dependent degeneration of photoreceptor neurons. Although the mutant Htt protein accumulated in the cytoplasm and nucleus of the photoreceptor neurons just after eclosion, nuclear accumulation of mutant Htt was observed in aged HD flies, suggesting that accumulation of the mutant Htt protein in the nucleus plays a crucial role in neurodegeneration. Gunawardena et al. (2003) also established HD fly models expressing the exon 1 fragment of the Htt protein with a 93 polyQ tract and showed that overexpression of this mutant Htt causes axonal transport defects accompanied by accumulation of the pathogenic Htt protein. Lee et al. (2004) established other HD fly models expressing the longer 548 amino acids fragment of the Htt protein with a 128 polyQ tract and also reported the disruption of axonal transport and accumulation of aggregates at synapses, indicating that cytoplasmic accumulation of the pathogenic Htt protein leads to neuronal dysfunction. Interestingly, they did not find axonal transport defects in flies expressing an expanded polyQ tract alone, which show only nuclear aggregates. On the other hand, new HD fly models expressing the full-length Htt protein containing a 128 polyQ tract have been established (Romero et al. 2008), and these flies showed behavioral, neurodegenerative, and electrophysiological phenotypes. They found that increased neurotransmission rather than axonal transport defects is at the root of the neurodegeneration caused by full-length mutant Htt during the early stages of pathogenesis (Romero et al. 2008). The results of these studies indicate that pathogenic outcomes can be affected by the protein context of the polyQ proteins.

5.2.2 Spinocerebellar Ataxia Type 1

Spinocerebellar ataxia type 1 (SCA1) is a dominantly inherited ataxia characterized by progressive cerebellar ataxia, dysarthria, dysphagia, and variable neurological symptoms and is caused by an abnormal expansion of the CAG trinucleotide repeat in the coding region of the ataxin-1 gene.

Fernandez-Funez et al. (2000) created a fly model of SCA1 by introducing transgenes encoding the full-length human ataxin-1 with a normal (SCA1-Q30) or expanded (SCA1-Q82) length polyQ repeats. Expression of SCA1-Q82 caused progressive neurodegeneration, as expected, and notably, flies expressing SCA1-Q30 at a high level also showed neurodegenerative phenotypes, indicating that even wild-type ataxin-1 can cause neurodegeneration. Genetic modifier screening using the SCA1 fly models identified several modifiers involved in protein folding/degradation, RNA processing, transcriptional regulation , and cellular detoxification. These findings shed light on a previously unrecognized new pathogenic mechanism of SCA1: the normal function of ataxin-1 could contribute to SCA1 pathogenesis . Subsequent studies also clarified modifiers involved in the signal transduction pathways by genetic interaction analyses using SCA1 fly models, in combination with mammalian-based genetic and proteomic analyses (Chen et al. 2003; Tsuda et al. 2005; Lam et al. 2006; Park et al. 2013).

In addition, genetic interaction between ataxin-1 and ataxin-2 was demonstrated using the SCA1 fly model (Al-Ramahi et al. 2007). The authors showed that wild-type Drosophila ataxin-2 is a major genetic modifier of the phenotypes of SCA1-82Q flies. They also showed that nuclear accumulation of ataxin-2 contributes to mutant ataxin-1-induced toxicity. Altogether, these findings suggest common mechanisms of neurodegeneration in different types of ataxia.

5.2.3 Spinocerebellar Ataxia Type 3

Spinocerebellar ataxia type 3 (SCA3), also known as Machado-Joseph disease (MJD), is the most common dominantly inherited ataxia and is characterized by progressive cerebellar ataxia and variable neurological symptoms. SCA3 is caused by an abnormal expansion of the CAG trinucleotide repeat in the coding region of the ataxin-3 gene.

The first genetically engineered fly models that were established for human neurodegenerative diseases were the SCA3 models (Warrick et al. 1998). These SCA3 fly models express the C-terminal region of the ataxin-3 protein containing normal (MJDtr-Q27) or pathogenic (MJDtr-Q78) length polyQ repeats. Expression of MJDtr-Q78 in the eye led to late-onset cell degeneration and NI formation (Fig. 5.1), similarly to the characteristics observed in SCA3 patients , whereas the expression of MJDtr-Q27 had no effect (Warrick et al. 1998). In a subsequent study, the same group demonstrated that HSP70, a major stress-induced molecular chaperone, suppresses polyQ-induced neurodegeneration in the SCA3 fly model (Warrick et al. 1999). They also showed that the full-length ataxin-3-Q27, which is a polyubiquitin-binding protein with ubiquitin protease activity , suppresses neurodegeneration and delays NI formation in MJDtr-Q78 flies, depending on its ubiquitin-associated activities and proteasome function (Warrick et al. 2005). These results indicate that the physiological function of the host protein plays a crucial role in SCA3 pathogenesis, as well as indicates the potential therapeutic role of ataxin-3 activity for the polyQ diseases. Moreover, Bilen and Bonini (2007) performed a genetic modifier screen using the SCA3 model fly and identified a set of genes that affects protein misfolding. Importantly, some modifiers of the SCA3 flies also modulated toxicity of tau, which is involved in Alzheimer’s disease and frontotemporal dementia, demonstrating common mechanisms of neurodegeneration between distinct neurotoxic proteins. We also showed that the loss of p62/sequestosome 1, which is involved in selective autophagy, delays the degradation of MJDtr-Q78 protein oligomers and exacerbates eye degeneration , indicating that p62 plays a protective role against polyQ-induced neurodegeneration in the SCA3 fly model (Saitoh et al. 2015). Taken together, these results suggest that chaperone activity and the protein-folding pathway play important roles in the pathogenesis of SCA3.

Fig. 5.1
figure 1

The fly model of SCA3

Expression of MJDtr-Q78 in the eye causes severe eye degeneration as compared to control. Fly genotypes are gmr-Gal4/+ (left) and gmr-Gal4/+; UAS-MJDtr-Q78S/+ (right)

Full size image

It is widely accepted that mutant ataxin-3 proteins containing an expanded polyQ tract cause neurodegeneration. However, Li et al. (2008) provided evidence for a pathogenic role of CAG repeat RNA in polyQ disease pathogenesis using SCA3 fly models. They performed modifier screening for polyQ-induced neurodegeneration and unexpectedly found that muscleblind, a gene implicated in the RNA toxicity of CUG expansion diseases, enhanced eye degeneration in SCA3 flies. Furthermore, they tested the possible role of RNA toxicity by expressing the CAG repeat in the untranslated region , and found that mRNA expression of an untranslated CAG repeat of pathogenic length induced progressive neuronal dysfunction. These results demonstrate the role of RNA toxicity in the pathogenesis of SCA3.

5.2.4 Spinal and Bulbar Muscular Atrophy

Spinal and bulbar muscular atrophy (SBMA), also known as Kennedy disease, is an adult-onset neurodegenerative disorder with an X-linked recessive inheritance. The disease mainly affects motor neurons and is characterized by slowly progressive limb and bulbar muscle weakness and atrophy and gynecomastia. As described in the Introduction section, SBMA is caused by an abnormal expansion of the CAG repeat encoding a polyQ tract in exon 1 of the AR gene (La Spada et al. 1991).

A fly model of SBMA was generated by introducing a transgene encoding the AR protein with a tract of 52 polyQ (AR-Q52) into flies (Takeyama et al. 2002). Although no obvious phenotype was observed in the photoreceptor neurons of the eyes of these flies, administration of androgen or its antagonists led to marked neurodegeneration accompanied with nuclear translocation of the mutant AR. These findings suggest that ligand binding to polyQ-expanded AR leads to its structural alteration and subsequent nuclear translocation , which eventually leads to neurodegeneration in male SBMA patients (Takeyama et al. 2002). Regarding involvement of native AR functions in the pathogenesis of SMBA, Nedelsky et al. (2010) showed that not only the nuclear translocation of AR but also the DNA-binding activity of AR and recruitment of transcriptional coregulators is necessary for its toxicity. These findings indicate that the native functions of AR play a crucial role in the pathogenesis of SBMA.

5.2.5 Oculopharyngeal Muscular Dystrophy

Oculopharyngeal muscular dystrophy (OPMD) is an adult-onset muscular disorder generally with autosomal dominant traits and is characterized by progressive swallowing difficulties, ptosis, and proximal limb weakness. OPMD is caused by a short expansion of the GCG trinucleotide repeat in the coding region of the nuclear poly(A)-binding protein 1 (PABPN1) gene, which encodes a protein that is involved in the polyadenylation of mRNAs and poly(A) site selection (Brais et al. 1998). Whereas the normal PABPN1 allele has a (GCN)10 repeat encoding a 10 polyalanine (polyA) stretch, OPMD patients carry expanded alleles with (GCG)12–17 repeats, encoding expanded polyA tracts in the N-terminal domain of PABPN1 (Brais et al. 1998).

Chartier et al. (2006) established a fly model of OPMD expressing mutant PABPN1 with a 17 polyA tract in muscle and demonstrated progressive muscle degeneration and nuclear inclusions composed of mutant PABPN1 in these flies, which are reminiscent of the characteristics of human OPMD patients. Notably, in this OPMD fly model, the polyA tract was not sufficient to cause muscle degeneration, and the RNA-binding domain (RRM) of PABPN1 was also required. This suggests that OPMD does not only result from polyA toxicity but also from an intrinsic property of mutant PABPN1 that is dependent on the RRM. The authors also identified several suppressors of the muscular phenotype such as the molecular chaperone HSP70 and the anti-apoptotic protein p35 using the OPMD fly model, demonstrating the protective role of molecular chaperones and involvement of apoptosis in mutant PABPN1-induced muscle degeneration.

Recently, Chartier et al. (2015) found that mRNAs encoding mitochondrial proteins are downregulated starting at the earliest stages of progression in fly and mouse models of OPMD. Since the downregulation of these mRNAs correlates with their shortened poly(A) tails, the authors propose that impaired nuclear polyadenylation is an early defect in OPMD.

5.2.6 Spinocerebellar Ataxia Type 8

Spinocerebellar ataxia type 8 (SCA8) is an adult-onset slowly progressive ataxia with autosomal dominant inheritance, which is associated with an expansion of the CTG repeat in the noncoding region of the ataxin-8 opposite strand gene, and possibly the complementary CAG repeat in the ataxin-8 gene. This was the first example of an expansion mutation of a noncoding trinucleotide repeat in SCA, in contrast to most other repeat expansion mutations occurring in the coding regions in other SCAs (Koob et al. 1999). As the CTG trinucleotide repeat is believed to be located in the noncoding region, toxic gain-of-function mechanisms of repeat RNA are thought to be involved in the pathogenesis of SCA8.

To investigate this possibility, Mutsuddi et al. (2004) generated fly models for SCA8 by expressing 9 (normal) or 112 (expanded ) CTG repeats. Both flies expressing normal and expanded CTG repeats in the eye showed late-onset and progressive eye degeneration . Using these SCA8 fly models, they performed a genetic modifier screen and identified four RBPs that are expressed in neurons.

Later, bidirectional expression of CUG and antisense CAG repeat transcripts were reported in an SCA8 mouse model, as well as in SCA8 patients (Moseley et al. 2006). Most surprisingly, the CAG repeat sequence located in the noncoding region was discovered to be translated into repeat polypeptides in the absence of an initiation ATG codon (Zu et al. 2011) in cell and mouse models of SCA8, as well as in SCA8 patients. This unconventional translation was named repeat-associated non-ATG (RAN) translation. These results suggest that toxic gain-of-function mechanisms at both the protein and RNA levels may contribute to the pathogenesis of SCA8.

5.2.7 Spinocerebellar Ataxia Type 31

Spinocerebellar ataxia type 31 (SCA31) is a late-onset autosomal dominant cerebellar ataxia, which is caused by a complex penta-nucleotide (TGGAA)n repeat insertion in the overlapping intron of the brain expressed, associated with Nedd4 gene and the thymidine kinase 2 gene in the antisense strand (Sato et al. 2009). In the brains of SCA31 patients, RNA foci containing UGGAA repeats were observed (Niimi et al. 2013), supporting a toxic gain-of-function mechanism caused by UGGAA repeat RNA in the pathogenesis of SCA31.

To gain insight into the pathogenic mechanisms of SCA31, we generated SCA31 model flies expressing expanded UGGAA repeats (UGGAAexp) and showed that the expression of UGGAAexp causes neurodegeneration accompanied by the accumulation of UGGAAexp RNA foci and pentapeptide repeat proteins produced by repeat-associated translation, as observed in SCA31 patient brains (Ishiguro et al. 2017). Moreover, the ALS -associated RBPs, TAR DNA-binding protein (TDP-43) , fused in sarcoma (FUS), and heterogeneous nuclear ribonucleoprotein A2/B1 (hnRNPA2/B1) bind to UGGAAexp RNA, alter the structure of UGGAAexp RNA, and suppress UGGAAexp-mediated toxicity . These results demonstrate that these RBPs function as RNA chaperones and regulate repeat-associated translation, suggesting that defects of RNA metabolism associated with RBPs contribute to the pathogenesis of SCA31.

5.2.8 Fragile X Tremor Ataxia Syndrome

Fragile X tremor/ataxia syndrome (FXTAS) is a late-onset neurodegenerative disease characterized by kinetic tremor, gait ataxia, parkinsonism, and dementia. FXTAS is caused by a premutation expansion of CGG repeats (55–200) in the 5′-UTR of the FMR1 gene, which is found in FXS carriers and belongs to the FMR1-related disorders, including FXS and FMR1-related primary ovarian insufficiency.

To investigate whether premutation alleles of FMR1 lead to neurodegeneration in vivo, Jin et al. (2003) established FXTAS fly models expressing 60 or 90 CGG repeats. They showed that expression of premutation CGG repeats alone is sufficient to cause neurodegeneration in a dose- and repeat length-dependent manner, suggesting RNA-mediated neurodegeneration in these fly models. In their following study, the authors screened for CGG repeat RNA-binding proteins from mouse brain lysates and identified Pur α and hnRNP A2/B1 as RBPs binding to CGG repeat RNA. They further showed that Pur α suppresses neurodegeneration caused by CGG repeat RNA in the FXTAS fly models, indicating that Pur α plays an important role in the pathogenesis of FXTAS (Jin et al. 2007). Sofola et al. (2007b) also identified RBPs such as hnRNP A2/B1 and CUG-binding protein 1 (CUGBP1) that bind to the CGG repeat and suppresses its toxicity in the FXTAS fly models. These results suggest sequestration of RBPs by CGG repeat RNA as one of the pathogenic mechanisms of FXTAS.

They also reported that co-expression of CGG repeat RNA together with CCG repeat RNA, whose expansion in the FMR2 gene causes another type of X-linked mental retardation, FRAXE, decreases their independent toxicities with each other, by reducing their transcript levels through the RNAi pathway (Sofola et al. 2007a). Furthermore, Sellier et al. (2013) found that the double-stranded RNA-binding protein DGCR8 binds to CGG repeats and is sequestered in CGG RNA aggregates together with its partner, DROSHA, resulting in a reduction in microRNA processing . These results suggest that alteration of the microRNA-processing machinery is involved in the pathogenic mechanisms in FXTAS.

Intriguingly, Todd et al. (2013) demonstrated that CGG repeats work as a template for RAN translation to produce polyglycine-containing proteins, which accumulate in ubiquitin-positive inclusions in the FXTAS fly models and FXTAS patient brains. Moreover, CGG repeat toxicity is suppressed by eliminating RAN translation and is enhanced by increased polyglycine production via ATG-initiated translation, indicating that RAN translation, which produces aberrant polypeptides, is involved in the neurodegeneration in FXTAS.

5.2.9 Myotonic Dystrophy Type 1

Myotonic dystrophy type 1 (DM1) is an autosomal dominant muscular dystrophy characterized by myotonia and muscular dystrophy , together with multisystem impairments, including cataracts, hypogonadism, endocrine dysfunction, heart defects, and cognitive decline. DM1 is the most common muscular dystrophy affected in adulthood, but it also appears as a congenital form. DM1 is caused by an abnormal expansion of CTG repeats in the 3′-UTR of the dystrophia myotonica protein kinase gene (Mahadevan et al. 1992; Brook et al. 1992). In DM1 patients, expanded CUG repeat-containing RNA accumulates as RNA foci in the nucleus of affected tissues and recruit two major RBPs, muscleblind like splicing regulator 1 (MBNL1) and CUGBP1, which bind to the CUG repeat RNA, resulting in their misregulation and alteration of RNA metabolism (Philips et al. 1998; Miller et al. 2000; Timchenko 2013).

To provide further insight into the pathogenic mechanisms of DM1, Houseley et al. (2005) generated DM1 fly models expressing expanded (162), intermediate (48, 56), or normal (11) CTG repeats in the 3′-UTR of a GFP reporter gene. In muscle cells, expanded CUG repeats formed RNA foci and colocalized with muscleblind, which is the Drosophila ortholog of human MBNL1, whereas normal and intermediate CTG repeats did not. However, no pathological phenotype, such as locomotor impairment, shortened life span, or muscular pathology, was detectable in this fly model. Further investigation was conducted by creating a more severe fly disease model with a larger number (480) of interrupted CUG (iCUG) repeats (de Haro et al. 2006). Expressions of this expanded iCUG repeat caused eye and muscle degeneration and the accumulation of expanded iCUG transcripts in nuclear RNA foci. Moreover, expression of MBNL1 was found to suppress expanded iCUG-induced toxicity, whereas expression of CUGBP1 worsened the iCUG-induced toxicity in these DM1 fly models (de Haro et al. 2006). Using this DM1 fly model, de Haro et al. (2013) further identified smaug, which is the Drosophila ortholog of human Smaug1/Samd4A, a translational repressor, as a suppressor of iCUG repeat-induced toxicity. Smaug was found to physically and genetically interact with CUGBP1 and suppresses iCUG-induced myopathy via restoration of the translational activity of CUGBP1 (de Haro et al. 2013).

5.2.10 C9orf72-Linked Amyotrophic Lateral Sclerosis and Frontotemporal Dementia

Amyotrophic lateral sclerosis (ALS) is a progressive neurodegenerative disease predominantly affecting upper and lower motor neurons, resulting in muscle weakness and atrophy, bulbar dysfunction, and eventual respiratory impairment. Frontotemporal dementia (FTD) is a neurodegenerative dementia characterized by cognitive impairment together with behavioral and personality changes. Since the discovery of TDP-43 as a key molecule aggregated in the pathological inclusions of both diseases, these intractable neurodegenerative diseases have been considered to belong to the same disease spectrum with overlapping genetic and neuropathological features (ALS/FTD) (Ling et al. 2013). In 2011, an abnormal expansion of a GGGGCC repeat in the first intron of the C9orf72 gene was identified as the most common genetic mutation of ALS /FTD (C9-ALS/FTD) (Renton et al. 2011; DeJesus-Hernandez et al. 2011). Three hypotheses in the pathogenesis of C9-ALS/FTD have been proposed so far, as follows: loss-of-function of the C9ORF72 protein, toxic gain-of-function of expanded GGGGCC repeat RNAs, and toxic gain-of-function of dipeptide repeat (DPR) proteins generated from expanded repeat RNAs by RAN translation (Ling et al. 2013). However, an FTD patient homozygous for the C9orf72 GGGGCC repeat expansion mutation was reported to demonstrate clinical and pathological features that fit within the range of those of heterozygous patients (Fratta et al. 2013). This fact, together with the lack of C9orf72-coding mutations in ALS patients, excludes the possibility of a loss-of-function mechanism in C9-ALS/FTD (Harms et al. 2013). Moreover, knockout mice for the C9orf72 gene demonstrate immunological defects, but no or mild neurological dysfunction (Koppers et al. 2015; Atanasio et al. 2016; Jiang et al. 2016; Sudria-Lopez et al. 2016). Thus, loss of C9orf72 function may not play a key role in the pathogenesis of C9-ALS/FTD.

Although Drosophila do not have an ortholog of the C9orf72 gene, fly models were employed to explore the toxic gain-of-function mechanisms in the pathogenesis of C9-ALS /FTD. The first C9-ALS/FTD fly model was established by expressing expanded 30 GGGGCC repeats with a CTCGAG interruption (iGGGGCC). Flies expressing iGGGGCC repeats in the eye caused eye degeneration, and those in motor neurons demonstrated motor dysfunction with aging (Xu et al. 2013). To distinguish the toxic gain-of-function mechanisms between expanded repeat RNAs themselves and DPR proteins produced by RAN translation, Mizielinska et al. (2014) generated three C9-ALS /FTD fly models, as follows: (1) flies expressing expanded pure GGGGCC repeats that produce both expanded RNAs and DPR proteins, (2) RNA-only flies expressing stop codon-interrupted expanded GGGGCC repeats that only produce expanded RNAs, and (3) DPR protein-only flies expressing non-GGGGCC RNAs with alternative codons that only produce DPR proteins . They found that flies expressing pure GGGGCC repeats showed neurodegenerative phenotypes, such as rough eye and decreased life span, whereas RNA-only flies showed no apparent phenotype, despite RNA foci formation in both pure GGGGCC repeat and interrupted repeat RNA-only flies. These findings suggest that expanded GGGGCC repeats cause neurotoxicity through the DPR proteins, and RNA foci may not be a direct cause of neurodegeneration in these fly models. The authors further investigated whether expression of the DPR protein alone is sufficient to induce toxicity using DPR protein-only flies. They found that only poly-GR and poly-PR proteins cause eye degeneration , whereas poly-GA and poly-PA proteins do not, indicating that arginine-containing DPR proteins are the major cause of neurodegeneration in C9-ALS/FTD fly models (Mizielinska et al. 2014). Tran et al. (2015) reported a new C9-ALS /FTD fly model expressing 160 GGGGCC repeats flanked by human intronic and exonic sequences. Spliced intronic 160 GGGGCC repeat RNA formed RNA foci in the nucleus of neurons but resulted in low levels of DPRs and no neurodegeneration. These results also indicate that the accumulation of RNA foci is not sufficient to drive neurodegeneration, and the sequences flanking the GGGGCC repeats may modulate RAN translation.

Toward elucidation of the molecular mechanisms underlying the pathogenesis of C9-ALS/FTD, several groups have performed genetic modifier screening using C9-ALS/FTD fly models. Zhang et al. (2015) identified Ran GTPase-activating protein (RanGAP), which is a key regulator of nucleocytoplasmic transport, and showed a genetic interaction between GGGGCC repeats and the nucleocytoplasmic transport machinery. Freibaum et al. (2015) performed genetic modifier screening using flies expressing GGGGCC repeats and GFP in frame to monitor RAN translation and identified 18 genes involved in the nuclear pore complex and nucleocytoplasmic transport. Boeynaems et al. (2016) also discovered genes encoding components of the nuclear pore complex, importins, exportins, Ran-GTP regulators, and arginine methylases as modifiers of C9-ALS /FTD flies. These findings provide evidence that nucleocytoplasmic transport contributes to the pathogenesis of C9-ALS/FTD.

5.3 Perspectives

As introduced above, a number of studies on repeat expansion disorders have been performed using fly models and have contributed toward elucidating the molecular mechanisms of these diseases. In particular, by taking advantage of fly models in rapid and efficient genetic analyses, various modifier genes have been identified by genetic screening , providing insight into the pathogenic mechanisms of these disorders.

The other remarkable advantage of fly models is their short generation cycle, which is useful for research on intergenerational repeat instability. Repeat instabilities are commonly observed in most of the repeat expansion disorders, and further elongation of expanded repeats in the next generation often results in earlier onset and more severe disease phenotypes, which is called anticipation (Mirkin 2007; Orr and Zoghbi 2007). Such elongation of expanded repeats is thought to occur during meiosis in germline cells, whereas repeat instability during mitosis is also known to cause somatic mosaicism (Pearson et al. 2005; Kovtun and McMurray 2008). Jung and Bonini (2007) used a fly model of SCA3 expressing an expanded CAG repeat to clarify the mechanisms underlying repeat instability. They found that repeat instability was enhanced by transcription and was modulated by Rad2/XPG, which is involved in DNA repair mechanisms. Furthermore, repeat instability was increased in SCA3 flies by the loss of CREB-binding protein, which is a histone acetyltransferase, and treatment with trichostatin A, a histone deacetylase (HDAC) inhibitor suppressed this repeat instability. These results clearly indicate the usefulness of fly models to study the mechanisms of repeat instability, which is thought to underlie the fundamental etiology of repeat expansion disorders.

In addition, several studies have shown the usefulness of fly models for the identification of potential drug targets. Using HD fly models, Steffan et al. (2001) first identified HDAC inhibitors, which increase the acetylation levels of histones, as therapeutic candidates for HD. They showed that the administration of sodium butyrate and suberoylanilide hydroxamic acid (SAHA) to HD flies by feeding suppressed neurodegeneration. Based on these findings, the therapeutic potential of HDAC inhibitors for HD were further explored in mouse models, and the therapeutic effects of SAHA were indeed replicated in a HD mouse model (Hockly et al. 2003). Several molecules targeting the misfolding and aggregation of polyQ proteins, such as polyglutamine binding peptide 1 (QBP1), Congo red, and methylene blue, have also been analyzed using fly models for their therapeutic potential (Nagai et al. 2003; Apostol et al. 2003; Sontag et al. 2012).

Although fly models have significantly contributed to extend our knowledge of repeat expansion disorders as mentioned above, we need to recognize the limitations of fly models in studying human diseases, due to the many differences between flies and humans, such as in their development, physiology, metabolism, nervous system, etc. Nevertheless, considering their rapid generation cycle, cost-effectiveness, and advantages in genetic analyses, fly models are powerful tools for studying human diseases (McGurk et al. 2015; Koon and Chan 2017).