Recent advances in using Drosophila to model neurodegenerative diseases

Abstract

Neurodegenerative diseases are progressive disorders of the nervous system that affect the function and maintenance of specific neuronal populations. Most disease cases are sporadic with no known cause. The identification of genes associated with familial cases of these diseases has enabled the development of animal models to study disease mechanisms. The model organism Drosophila has been successfully used to study pathogenic mechanisms of a wide range of neurodegenerative diseases. Recent genetic studies in the Drosophila models have provided new insights into disease mechanisms, emphasizing the roles played by mitochondrial dynamics, RNA (including miRNA) function, protein translation, and synaptic plasticity and differentiation. It is anticipated that Drosophila models will further our understanding of mechanisms of neurodegeneration and facilitate the development of novel and rational treatments for these debilitating neurodegenerative diseases.

Building neurodegenerative disease models in Drosophila

Neurodegenerative diseases represent a subgroup of diseases sharing certain features. For example, these diseases are often of adult onset and disease pathogenesis usually involves the progressive loss of specific neuronal populations characteristic for each disease type. These diseases were once considered among the most obscure and intractable of all human illnesses. Thanks to the molecular cloning of genes associated with familial forms of such diseases as Alzheimer’s disease (AD) and Parkinson’s disease (PD), the pathogenic mechanisms of neurodegeneration can now be studied at the molecular level. One of the powerful approaches for studying disease mechanism is the establishment of animal models. Drosophila has proven to be an excellent tool for modeling human neurodegenerative diseases. A number of excellent reviews have been written on this topic recently [1–4]. This field is progressing rapidly. The aim of this review is to summarize recent developments in using Drosophila to study various human neurodegenerative diseases.

Animal models offers a number of advantages for studying human neurodegenerative diseases. For example, they allow genetic analysis of disease gene relationships and delineation of cellular pathways involved in disease pathogenesis. They also allow studies of molecular and cellular changes of the relevant brain cell populations at various stages of the disease process, from pre-symptomatic stage to full-blown disease manifestation.

Four approaches have been employed successfully to study neurodegeneration in Drosophila. First, forward genetic screens have been carried out to identify genes that when mutated can cause degeneration of the brain. Examples include the Drosophila mutants such as bub-blegum, swisscheese, and drop-dead [5, 6]. The mammalian homologues of some of these genes also cause neurodegeneration when mutated, thus validating the usefulness of this approach in identifying conserved genes required to maintain nervous system integrity [7]. Second, transgenic overexpression approach has been used to model disease caused by a toxic gain-of-function (GOF) mechanism [8–11]. Third, a genetic inhibition of endogenous gene approach has been used to model the subset of familial diseases transmitted in a recessive fashion, which are likely caused by a loss-of-function (LOF) mechanism. Transposon-mediated mutagenesis, transgenic RNA interference (RNAi), and homologous recombination-based gene knockout can be used for this purpose [12–15]. Fourth, a pharmacological approach can be used to model neurodegenerative diseases and to test candidate therapeutics in animals [16, 17]. Below we summarize the fly models established using the first three genetic approaches and lessons learned from them.

PD models

PD is the most common movement disorder and the second most common neurodegenerative disease. The movement abnormality in PD is largely attributed to the deficiency of brain dopamine content caused by degeneration of dopaminergic neurons in the substantia nigra. A pathological hallmark of the disease is the formation of Lewy bodies, intracytoplasmic inclusions that are composed mainly of α-synuclein and ubiquitin. The most common forms of PD are sporadic with no known cause. Nevertheless, postmortem studies of sporadic cases have identified defective mitochondrial complex I function and oxidative damage in nigrostriatal dopaminergic neurons [18, 19]. It is now well established that genetic factors contribute to PD pathogenesis. At least 10 distinct loci (PARK1-11) have been linked to rare familial forms of PD (FPD) [20, 21]. To date, six FPD genes have been molecularly cloned, including α-synuclein (α-Syn), Parkin, Ubiquitin C-terminal hydrolase-1 (Uchl-1), DJ-1, PTEN induced kinase-1 (PINK-1), and Leucine-rich repeat kinase 2 (LRRK2) [22–28]. In the cases of α-Syn and LRRK2, disease modeling was achieved through transgenic overexpression of wild type or pathogenic forms of proteins, whereas for DJ-1, Parkin, and Pink1, disease models were created by genetic LOF approach.

α-Syn models

α-Syn encodes a small protein whose normal function is likely involved in synaptic transmission. It is a main component of Lewy bodies and its aggregation is frequently found in PD. Although Drosophila does not have a clear α-Syn homologue, overexpression of wild-type and FPD-associated mutant α-Syn in Drosophila has recapitulated key features of PD, including Lewy body–like aggregate formation, selective degeneration of dopaminergic neurons, and locomotor behavior abnormality [10, 29]. The fact that both mutant and wild-type α-Syn form aggregates in fly neurons and cause disease phenotypes supports the notion that improper disposal of aggregation-prone abnormal proteins can result in neurodegeneration. This is also consistent with the finding that genomic triplication of α-Syn gene can cause PD [30].

Consistent with a role for aberrant protein quality control in α-Syn pathogenesis, directed expression of the molecular chaperone HSP70, prevented dopaminergic neuronal loss associated with α-Syn expression in Drosophila [29]. A recent human genetics study indicated that polymorphisms in Hsp70 genes may affect susceptibility to PD [31]. The potential of Hsp70 gene therapy in a mouse model of MPTP-induced PD was also tested. Viral vector-mediated Hsp70 gene transfer to dopamine neurons significantly protected mouse dopaminergic system against MPTP-induced dopaminergic toxicity [32]. Increasing chaperone activity may offer a new avenue for the treatment of PD.

Mammalian studies have previously shown that α-Syn present in Lewy bodies is phosphorylated at serine 129 (Ser129) [33]. Chen and Feany showed that mutating Ser129 to non-phosphorylatable alanine completely suppressed α-Syn toxicity, whereas changing Ser129 to the phosphomimetic aspartate significantly enhanced α-Syn toxicity in the fly model. Interestingly, blocking phosphorylation at Ser129 substantially increased aggregate formation while preventing neurotoxicity [34]. In another study using deletion constructs of α-Syn, however, a positive role for aggregation of α-Syn in mediating dopaminergic neuron toxicity in vivo was proposed [35]. The precise mechanism by which the different aggregated form of α-Syn may exert differential effects on neurotoxicity remains to be determined.

The fly α-Syn model has been used to test candidate factors suspected of playing a role in the pathogenesis of PD. Aging has been firmly established as a risk factor for most neurodegenerative diseases. In one study the role of sirtuins, members of the histone deacetylase family of proteins that participate in aging, in regulating α-Syn toxicity was tested. Pharmacological application of a potent inhibitor of sirtuin 2 (SIRT2) or genetic inhibition of SIRT2 via small interfering RNA rescued α-Syn toxicity in a cellular model. The inhibitors also protected against dopaminergic cell death in a Drosophila α-Syn model of PD [36]. In another study, the role of oxidative stress in α-Syn pathogenesis was tested. It was found that dopaminergic neurons are specifically sensitive to hyperoxia induced oxidative stress and that mutant forms of α-Syn show an enhanced toxicity under these conditions. The co-expression of Cu/Zn superoxide dismutase protects against the dopaminergic neuronal loss induced by mutant α-Syn, thus identifying oxidative stress as an important causative factor in α-Syn pathogenesis [37].

Despite these studies, the exact mechanism by which accumulation of wild type α-Syn or expression of mutant α-Syn causes selective dopaminergic neuron degeneration remains elusive. Systematic genetic studies promise to address this question. In one study initially carried out in yeast, α-Syn was shown to cause a block in endoplasmic reticulum (ER)-to-Golgi vesicular trafficking. A genome-wide screen identified toxicity modifiers, including the Rab guanosine triphosphatase Ypt1p, which associated with cytoplasmic α-Syn inclusions. Elevated expression of Rab1, the metazoan YPT1 homolog, protected against α-Syninduced dopaminergic neuron loss in a fly PD model [38]. This study is beginning to reveal basic cellular functions disrupted by toxic α-Syn expression that may contribute to synucleinopathies. Other genomewide approaches such as DNA microarray analysis of transcriptional profiles or proteomic analysis of protein expression changes in animals ubiquitously expressing α-Syn in the nervous system is beginning to yield useful information [39, 40]. Further refinement of the approaches, such as analyzing purified dopaminergic neurons, the disease-relevant population, at different stages of the disease process will provide more insights into mechanisms of α-Syn pathogenesis.

Parkin and Pink1 models

Parkin encodes a ring-finger domain containing E3 ubiquitin ligase. Its inactivation causes autosomal recessive juvenile parkinsonism. Three labs independently generated parkin null mutant flies [12, 13, 41]. A prominent feature of parkin mutant flies is mitochondrial pathology and apoptotic muscle degeneration. In addition, these mutant flies exhibit sterility, reduced lifespan, reduced cell number and size, and hypersensitivity to oxidative stress. Whole mount confocal analysis also detected notable dopaminergic neuron loss in the protocerebral posterior lateral (PPL) 1 cluster in parkin null flies [42]. Using a different approach, parkin function was inhibited by transgenic RNAi [14]. Parkin RNAi flies exhibited genetic interaction with transgenic flies overexpressing human Pael-R, a Parkin substrate protein [43], in promoting dopaminergic degeneration, consistent with Pael-R being a Parkin substrate that mediates Parkin deficiency-induced PD. Recent studies in mice also showed that Pael-R induces dopaminergic neurons degeneration in the substantia nigra, and that Pael-R toxicity is enhanced under Parkin inactivation condition [44]. In an interesting twist, overexpression of certain mutant forms of human Parkin in Drosophila also led to dopaminergic neurodegeneration and mitochondrial abnormality in indirect flight muscle [45, 46], suggesting that at least some of the FPD mutations in Parkin may have gained certain toxic property or act in a dominant-negative fashion.

Pink1 encodes a Ser/Thr kinase with an N-terminal mitochondrial targeting domain. Drosophila Pink1 models of PD were generated by transposon-mediated mutagenesis and RNAi approaches. Inhibition of Drosophila Pink1 (dPink1) function results in phenotypes strikingly similar to that seen in Parkin mutant or RNAi flies, including male sterility, apoptotic muscle degeneration, dopaminergic neuron loss, defective mitochondrial morphology, and increased sensitivity to multiple stresses including oxidative stress [47–50]. The similar phenotypes observed in Parkin and Pink1 mutant flies prompted genetic epistasis studies to investigate the relationship between these two PD-associated genes. Overexpression of Parkin largely rescued the Pink1 mutant phenotypes, whereas overexpression of Pink1 had no effect on Parkin mutant phenotypes. Further, pink1 parkin double mutants showed phenotypes identical to those observed in either single mutant alone. These observations suggest that Pink1 and Parkin function in the same pathway, with Pink1 acting upstream of Parkin [47–49]. Pink1 has recently been shown to directly phosphorylate Parkin and regulate its mitochondrial localization [51], providing a possible molecular explanation for their genetic relationship.

A breakthrough in understanding the exact mitochondrial process regulated by Pink1 and Parkin is provided by the finding that Pink1 and Parkin interact with the mitochondrial fission and fusion pathway genes [52–55]. Mitochondrial fission and fusion are evolutionarily conserved membrane remodeling processes that regulate the morphology and subcellular distribution of mitochondrial units [56]. Mitochondrial fission/fusion has recently been shown to be important for regulating synaptic structure and plasticity [57], with imbalance of mitochondrial fission/ fusion leading to neurodegeneration [58]. Genetic studies strongly suggest that the mitochondrial morphology defects in Pink1 and Parkin mutant are due to decreased fission or increased fusion. Two dynamin-like GTPases, Drp1 and Opa1, are critical components of the machinery that regulate mitochondrial fission and fusion, respectively. Enhancing mitochondrial fission activity by the overexpression of Drp1 or reducing fusion activity by the loss of one copy of Opa1 can rescue Pink1 mutant phenotypes in dopaminergic neurons [52]. The biochemical mechanisms linking Pink1/Parkin to mitochondrial fission/fusion machinery remain to be determined. Parkin overexpression has been shown to protect against dopaminergic neurodegeneration caused by α-Syn overexpression in Drosophila and mammals [14, 59]. It will be interesting to test whether α-Syn overexpression also affects mitochondrial morphology or dynamics.

It is intriguing that the indirect flight muscle and dopaminergic neurons are particularly vulnerable to Pink1 and Parkin mutations. It is possible that high endogenous levels of oxidative stress in these tissues may confer this vulnerability. Being one of the most metabolically active tissue, fly indirect flight muscle probably produces more ROS than any other tissues and is presumably under constant oxidative stress [60]. Dopaminergic neurons are also constantly under oxidative stress due to the oxidative metabolism of dopamine. In fact, a link between dopamine metabolism and α-Syn toxicity has been observed in Drosophila primary neuronal culture [61]. Together, these conditions may confer particular vulnerability of these tissues to additional oxidative insults caused by mitochondrial dysfunction in Pink1 or Parkin mutants.

DJ-1 models

DJ-1 encodes a highly conserved protein belonging to the ThiJ/PfpI superfamily. Its identified molecular functions include molecular chaperone, RNA binding, and anti-oxidant. There are two DJ-1 homologues in Drosophila, named DJ-1A and DJ-1B. DJ-1A appears to be a closer homologue of human DJ-1 than DJ-1B based on sequence conservation. DJ-1A and DJ-1B knockout flies display altered sensitivity to environmental toxins such as paraquat, H₂O₂, and rotenone [62, 63]. Knock down of DJ-1A by transgenic RNAi in a cell type-specific fashion resulted in accumulation of reactive oxygen species, hypersensitivity to oxidative stress, and dysfunction and degeneration of dopaminergic and photoreceptor neurons [64]. It is unclear at this point whether the differential effects between the genetic deletion mutants and the RNAi flies is due to “off-target” effects of RNAi or that certain non-cell autonomous effects of DJ-1A inactivation may have contributed to the tissue-specific RNAi effect.

Using DJ-1A RNAi-induced cell death as an assay, a systematic genetic interaction study was carried out to identify genetic modifiers. This led to the isolation of components of the PI3 K/Akt signaling pathway as specific genetic modifiers [64]. An independent study also found DJ-1A as a genetic modifier of PTEN-induced small-eye phenotype in Drosophila, and that DJ-1 RNAi or overexpression affects the phosphorylation of PKB/Akt. Interestingly, in primary breast cancer samples, DJ-1 expression correlates negatively with PTEN immunoreactivity and positively with Akt hyper-phosphorylation [65]. DJ-1 thus appears to be a key negative regulator of PTEN, and may serve as a useful prognostic marker for cancer. These studies raise the interesting possibility that cancer and Parkinson’s disease, two seemingly disparate diseases, may share certain underlying biochemical pathways. The specific molecular function of DJ-1 in these processes is unknown. A recent study showed that DJ-1 associates with RNA targets in cells and brain tissues in an oxidation-dependent manner. RNA targets of DJ-1 include mitochondrial genes, genes involved in glutathione metabolism, and members of the PTEN/PI3K cascade [66]. One possibility is that DJ-1 may regulate the translation of these RNA targets.

LRRK2 models

LRRK2 encodes a large protein that contains a leucine rich repeat (LRR) domain, a kinase domain, a RAS-like GTPase domain, and a WD-40 domain. Mutations in LRRK2 are frequently found in familial as well as sporadic PD cases, indicating a general role for LRRK2 dysfunction in disease pathogenesis. Although an initial report claimed no dopaminergic degeneration in transgenic flies overexpressing the wild type or R1069C pathogenic forms of the Drosophila LRRK2 homologue (dLRRK) [67], later studies expressing the I2020T mutant form of dLRRK or the wild type and the G2019S mutant forms of hLRRK2 did observe dopaminergic neurodegeneration [68, 69]. Moreover, both human LRRK2 and dLRRK phosphorylate eukaryotic initiation factor 4E (eIF4E)-binding protein (4E-BP), a negative regulator of eIF4E-mediated protein translation and a key mediator of various stress responses. Modulation of the eIF4E/4E-BP pathway by LRRK2 stimulates eIF4E-mediated protein translation, and attenuates resistance to oxidative stress and survival of DA neuron in Drosophila, suggesting that chronic inactivation of 4E-BP by LRRK2 with pathogenic mutations deregulates protein translation and cause age-dependent loss of DA neurons [69]. This study for the first time directly link deregulation of protein translation to PD pathogenesis.

AD and related tauopathy models

AD is the most common neurodegenerative diseases and the leading cause of dementia. Clinically, it is manifested as a gradual decline of cognitive functions. Two microscopic brain lesions, the amyloid plaque and neurofibrillary tangles (NFT), are diagnostic of AD in the demented patient when present in abnormally high densities for a given age. The main components of the amyloid plaque are the Aβ-40 and Aβ-42 peptides generated by endoproteolysis of the amyloid precursor protein (APP) via the sequential action of β- and γ-secretases. β-secretase activity is provided by the β-site APP-cleaving enzyme (BACE), whereas γ-secretase activity depends on a protein complex consisting of presenilin (Psn), nicastrin, aph-1, and pen-2 [70]. Autosomal dominant mutations in APP, Psn-1, and Psn-2 can accelerate disease onset and progression in familial AD cases. The major component of the NFT is tau protein, a highly soluble, microtubule-binding protein normally enriched in axons [71]. Tau becomes abnormally phosphorylated and insoluble in the NFT [72–74]. Direct evidence of the pathogenicity of tau was provided by the discovery that mutations in tau are associated with frontotemporal dementia with parkinsonism linked to chromosome 17 (FTDP-17) [75–78].

Most of the genes implicated in AD pathogenesis have clear fly homologues. There is a fly homologue of APP (called APP-like or APPL). Flies deficient for APPL exhibit behavioral abnormality, which can be rescued by a human APP transgene, indicating functional conservation between fly APPL and human APP [79]. A recent study found that fly APPL can be processed into Aβ fragments that aggregate into intracellular fibrils, amyloid deposits, and cause age-dependent behavioral deficits and neurodegeneration [80], suggesting that the proteolytic processing and Aβ-induced neurotoxicity are conserved for APP proteins. Fly models of human Aβ peptide-induced amyloid formation and neurodegeneration have also been generated. This was achieved either by constructing triple transgenics co-expressing a wild type human BACE transgene, and human APP and fly Psn transgenes containing FAD-linked mutations, or by overexpressing Aβ-40 and Aβ-42 peptides fused to signal peptides [81, 82]. Overexpression studies in Drosophila have shown that pathogenic APP can cause axonopathy and neurodegeneration, but this effect of full-length APP is independent of Aβ peptides [83], suggesting that other regions of APP protein exert pathogenic roles as well.

The fly Aβ-42 model has been used in genetic modifier screens to identify other genes that act in the same cellular pathway. In one study, the Toll/NFkB signaling pathway was found to mediate the neuropathological effects of human Aβ-42 toxicity in Drosophila, suggesting the involvement of innate immunity/inflammatory pathway in Aβ-42 pathogenesis [84]. In another study, the effect of manipulating Neprilysin on Aβ-42 toxicity in Drosophila was tested. Neprilysin is one of the rate-limiting Aβ-degrading enzymes first found in mammals. In Drosophila, overexpression of Neprilysin reduces Aβ-42-induced neuronal loss and intraneuronal Aβ-42 deposits. However, this manipulation caused a reduction in cAMP-responsive element-binding protein (CEBP)-mediated transcription, age-dependent axonal pathology, and premature death [85]. This finding has therapeutic implications when strategies aimed at enhancing Neprilysin activity are considered.

Tau-induced neurodegeneration has been successfully modeled in Drosophila by either expression of a bovine tau-GFP reporter protein [86], or expression of wild type or pathogenic forms of human tau [87, 88]. Overexpression of a fly homologue of tau (dtau) also compromised neuronal function [89]. Tau phosphorylation/dephosphorylation has attracted considerable interest in tauopathy research, due to the identification of abnormally phosphorylated tau in NFT. A number of protein kinases and phosphatases have been found to regulate tau phosphorylation in vitro [90]; however, surprisingly little is known about the in vivo roles of these enzymes in determining tau toxicity. Through LOF and overexpression genetic studies and biochemical analysis, Drosophila Partioning defective-1 (PAR-1) was found to be a physiological tau kinase that plays a central role in determining tau phosphorylation and toxicity in Drosophila [91]. PAR-1 is a Ser/Thr kinase originally identified in C. elegans for its role in regulating cell polarity and asymmetric cell division [92]. MARK kinases, the mammalian homologues of PAR-1, regulate microtubule dynamics, epithelial cell polarity, and neuronal differentiation under normal conditions and bind to NFT in AD brain [93–95]. In the fly tauopathy model, PAR-1 was found to initiate a multisite phosphorylation process that generates pathogenic tau [91]. These results, together with the findings from a genetic screen that uncovered kinases and phosphatases as major modifiers of tau toxicity [96], established the importance of phosphorylation in determining tau toxicity in vivo.

The importance of individual phosphorylation sites in determining tau toxicity in vivo has not been clearly defined. Post-mortem analysis indicated that in the AD patient brain tau is hyperphosphorylated in more than 20 Ser/Thr sites [97]. In theory, these phosphorylation sites could act in concert to confer tau toxicity. Alternatively, a small subset of sites might play a dominant role in determining tau toxicity. Mutagenesis studies has emphasized the importance of a small subset of phosphorylation events controlled by PAR-1/MARK in determining tau toxicity [91, 98], whereas a large number of SP/TP sites phosphorylated by proline-directed kinases appear to work in concert to promote tau neurotoxicity in vivo [99].

The mechanism by which tau phosphorylation affects its toxicity remains poorly understood. Phosphorylation could affect the protein-protein interaction between tau and its effector proteins, or regulate the turnover, solubility, or proteolytic cleavage of tau. A recent cross species functional genomic study identified the highly conserved puromycin-sensitive aminopeptidase (PSA/Npepps) as a tau modifier. PSA protected against tau-induced neurodegeneration in vivo, whereas PSA loss of function exacerbated neurodegeneration. Further, human PSA was found to directly proteolyze tau [100]. It would be interesting to test whether the cleavage of tau by PSA is regulated by phosphorylation.

The upstream signalling mechanisms that regulate tau phosphorylation under normal and disease conditions are beginning to be defined. The identification of PAR-1 as a physiological tau kinase offers an excellent entry point to dissect the signalling mechanisms. Recent studies show that phosphorylation of PAR-1 by the tumor suppressor protein LKB1 is required for its activation, which in turn promotes tau phosphorylation in Drosophila. Diverse stress stimuli, such as human APP-induced neurotoxicity, can promote PAR-1 activation and tau phosphorylation in an LKB1-dependent manner [101]. These results reveal a new function for the tumor suppressor protein LKB1 in a signalling cascade through which the phosphorylation and function of tau is regulated by AD-related toxic stimuli.

Loss of synaptic function has been recognized as one of the earliest signs of the disease process in AD [102]. Studies in animal models have further supported the role of synaptic dysfunction in disease pathogenesis. APP and APP-like proteins are known to exert normal physiological function at the synapse [79, 103]. The mouse transgenic APP model exhibits learning and memory deficits in the absence of obvious neuronal degeneration [104]. Interestingly, removal of endogenous tau can suppress most of the cognitive deficits in the APP model [105], indicating a critical role of tau in mediating the toxic effects of APP at the synapse. This is consistent with the findings that synaptic toxicity is also observed in Drosophila tau models [106, 107], and that genes involved in synaptogenesis are recovered as tau modifiers in an unbiased genetic screen [108].

It appears that tau is not the only synaptic protein whose abnormality causes synaptic dysfunction in AD. A recent study found that overactivation of PAR-1 in Drosophila causes the delocalization of Dlg from the postsynaptic membrane at the neuromuscular junction (NMJ). This is mediated through the direct phosphorylation of Dlg by PAR-1 [109]. Dlg is the fly ortholog of the key postsynaptic scaffold protein PSD-95, whose synaptic localization has been found to be decreased in AD [110–112]. Given that aberrant activation of PAR-1 like kinases has been implicated in AD conditions, this study offers a new direction for understanding the molecular basis of synaptic dysfunction in AD. Further supporting the role of synaptic dysfunction in AD pathogenesis, a recent study in Drosophila provided evidence of learning and synaptic defects in a Drosophila psn mutant and suggested a presynaptic role for presenilin in normal neuronal function [113].

Compared to early-onset AD, relatively little is known about the genetic control of late-onset AD. Ubiquilin 1 variants have been associated with increased risk for late-onset AD. In Drosophila, silencing of endogenous dUbqln in the central nervous system led to age-dependent neurodegeneration. The dUbqln exhibited strong in vivo functional interactions with the AD-associated genes presenilin and APP, providing insights into the potential role of Ubiquilin 1 in AD pathogenesis [114].

Polyglutamine (polyQ) diseases

A diverse group of dominantly inherited neurodegenerative diseases are caused by CAG trinucleotide repeat expansion within disease-specific proteins. These include Hunting-ton’s disease (HD), spinobulbar muscular strophy (SBMA), spinocerebellar ataxia (SCA) 1, 2, 3, 6, 7, and 17; and dentatorubralpallidoluysian atrophy. These polyQ expansions confer toxic properties to the resident proteins. There is a good correlation between repeat length and disease severity [115]. These polyQ repeats form intranuclear inclusions in neurons and cultured cells and are intrinsically toxic. The Drosophila compound eye has been an excellent system for modeling these polyQ diseases. In almost all cases, expression of truncated, expanded polyQ-containing protein fragments leads to retinal degeneration. In the case of SCA-1, overexpression of the non-expanded wild type ataxin-1 (Atx-1) protein also resulted in retinal degeneration, indicating that the ataxin-1 protein is intrinsically toxic when present at high levels [11]. Studies in Drosophila also revealed that expression of expanded polyQ-containing proteins in glial cells also confers toxicity [116–118], suggesting that degeneration or dysfunction of these cells may also contribute to the disease process.

Significant insights into the pathogenic pathways of polyQ diseases were obtained from comprehensive genetic screens using the SCA models. Loss-of-function and overexpression screens identified genes involved in a number of biological processes, some of which were expected, whereas other are unanticipated, including genes involved in RNA processing, nuclear transport, transcriptional regulation (histone deacetylation), and cellular detoxification [11]. A similar approach applied to the SCA-3 model identified genes involved in protein misfolding and autophagy as modifiers [119]. One interesting finding from the SCA-3 screen is the uncovering of muscleblind (mbl), a gene implicated in the RNA toxicity of CUG expansion diseases, as a SCA-3 modifier, suggesting a toxic role of RNA in polyQ-induced degeneration. Further studies validated this notion. These studies highlight common mechanisms in RNA-based and polyQ-protein-based trinucleotide repeat expansion diseases [120]. Using a candidate gene approach, the miRNA pathway was shown to modulate the neurodegeneration phenotypes of the SCA-3 model [121]. The detailed biochemical mechanism involved in miRNA modulation of SCA-3 toxicity remains to be determined. Given the well established role of the miRNA pathway in regulating protein translation, it is possible that deregulated synthesis of certain key proteins is involved in SCA-3 pathogenesis.

In a biochemical approach using the SCA-1 model, it was found that the 14-3-3 protein could bind to Atx-1 in a phosphorylation-dependent manner, leading to the stabilization and accumulation of the toxic protein [122]. The PI3K/AKT pathway was found to regulate the association between Atx-1 and 14-3-3 and affect Atx-1 toxicity. AKT phosphorylates Atx-1 at a conserved Ser776 residue and this phosphorylation event provides a binding site for 14-3-3 [122]. Other endogenous proteins that bind to Atx-1 also plays an important role in SCA-1 pathogenesis. Interestingly, a recent study revealed opposing effects of polyglutamine expansion on native protein complexes that both contribute to SCA1 pathogenesis. It was found that polyglutamine expansion in Atx-1 favors the formation of a protein complex containing RBM17, contributing to SCA1 neuropathology through a gain-of-function mechanism. Concomitantly, polyglutamine expansion attenuates the formation and function of another protein complex containing Atx-1 and capicua, contributing to SCA1 via a loss-of-function mechanism [123]. Another interesting finding is that different polyQ genes interact to modulate neurodegeneration in Drosophila. Atx-2 was found to mediate the pathogenic effects of SCA-1 and SCA-3. Loss of endogenous Atx-2 function suppressed the toxicity of both SCA-1 and SCA-3 [124, 125]. These findings reveal a previously unknown functional link between these disorders with common clinical features but different etiology, and raise the possibility that therapeutic approaches effective for one disease may benefit others.

PolyQ diseases are caused by the expansion of trinucleotide repeat sequences. Disease occurs only when the normally polymorphic number of repeats expands beyond a critical threshold level. Expanded repeats have a strong tendency to further expand, causing the disease to occur earlier and with greater severity in successive generations, a phenomenon called anticipation. Expression of the SCA-3 polyQ transgene in the Drosophila germline recapitulated several central features of human CAG repeat instability, including the broad range of repeat changes and strong bias for expansion. Genetic studies revealed that germline transcription along with nucleotide excision repair and/or transcription-coupled DNA repair facilitated repeat instability. Further, inhibition of cAMP response element–binding protein (CREB)–binding protein (CBP), a consequence of polyQ cellular toxicity, enhances repeat instability, revealing the presence of a vicious cycle in polyQ pathogenesis [126].

Earlier Drosophila models of HD were built using N-terminal fragment of Htt protein containing polyQ sequence. Although these models have provided important insights into HD, without comparison to a full-length Htt model, it is not clear if these models reliably recapitulated the disease situation. Recent analysis of a full-length Drosophila HD model revealed that expression of expanded full-length Htt leads to increased neurotransmitter release. Resting intracellular Ca²⁺ levels are increased in these flies compared to controls, suggesting a defect in Ca²⁺ homeostasis. Intriguingly, these abnormalities occur before the cleavage and nuclear translocation of Htt protein was evident [127]. This study reveals an unexpected synaptic role for full-length expanded Htt protein in early stages of the disease. Interestingly, in another study, endogenous non-pathogenic Htt was found to protect against the aggregation of polyglutaminated Htt fragments [128]. These studies establish full-length Htt as a relevant and perhaps more physiological system for studying HD pathogenesis.

Another important finding from recent studies is that distinct pathogenic mechanisms underlie different polyQ diseases. For example, despite the isolation of cellular detoxification enzymes as modifiers of SCA-1 pathology, overexpression of SOD has no effect on the HD model [129]. As another example, specific histone deacetylase (HDAC) and Sirtuins promote disease pathogenesis in a Drosophila HD model, with the inhibition of Rpd3 or Sir2 providing neuroprotective effects [130]. In contrast, HDAC6 induces autophagy and protects against neurode-generation in a Drosophila SBMA model [131]. These results suggest that different disease mechanism may underlie the different polyQ diseases, or that different HDACs exert distinct functions. Consistent with these findings, comparative analysis of a collection of genetic modifiers of expanded Ataxin-1-induced neurotoxicity in SCA-1 and HD models revealed that while some modifier genes function similarly in these models, others have model-specific effects. Surprisingly, certain modifier genes modify SCA1 and HD models in opposite directions [132]. Continued comparative analysis of other polyQ models will help uncover genes generally involved in polyQ-induced neurodegeneration as well as genes specific for each disease, which will provide excellent drug targets for the treatment of polyQ diseases. In this respect, the recently developed conditional Drosophila model of SCA7 will be particularly useful [133].

Models for other neurodegenerative diseases

Noncoding trinucleotide repeat diseases

In addition to the polyQ diseases where trinucleotide repeat expansion leads to toxic protein production, there are other diseases caused by trinucleotide repeat expansion within non-coding regions of mRNAs. Examples include SCA-8, −10, −12, Myotonic Dystrophy type 1 and 2, and Fragile × mental retardation (FMR). To understand how the rCGG repeats in the 5′-UTR of the FMR1 gene cause neurodegeneration in premutation carriers, 90 rCGG repeats, which is in-between the number of repeats found in patients (>200) and normal individuals (<60 repeats), was expressed in the fly eye. This 90 non-coding rCGG repeat was sufficient to cause degeneration of photoreceptors [134]. Although no mutant protein was produced from this repeat sequence, the degenerating neurons form neuronal inclusion bodies that are HSP70- and ubiquitin- positive. Further, overexpression of HSP70 was able to suppress rCGG repeats-induced neuronal death. It is possible that abnormal RNAs can cause neurodegeneration by sequestering rCGG repeat binding proteins from their normal function. In a biochemical purification, Pur α and hnRNP A2/B1 were identified as repeat binding proteins. Overex-pression of Pur α in Drosophila could suppress rCGG-mediated neurodegeneration in a dose-dependent manner. These findings support rCGG repeat sequestration of specific binding proteins as the pathogenic mechanism [135]. In another fly model of RNA-induced neurodegeneration, human SCA8 gene, which encodes a non-coding RNA with CUG expansion at the 3′ end, induced a late-onset progressive neurodegeneration phenotype when expressed in the fly eye [136]. A genetic modifier screen recovered four neuronally expressed RNA-binding proteins as modifiers of SCA8-induced neurodegeneration. These studies emphasized the importance of RNA binding proteins in noncoding trinucleotide repeat diseases. It remains to be tested whether the RNA-binding proteins identified from the SCA8 study will modify the toxicity induced by the non-coding rCGG expansion described above, and vice versa. This will reveal whether similar mechanisms are involved in noncoding trinucleotide repeat diseases.

Prion disease and amyotrophic lateral sclerosis (ALS)

Two other prominent neurodegenerative diseases, amyotrophic lateral sclerosis (ALS) and prion diseases, have been successfully modeled in Drosophila. Transgenic flies heterologously expressing disease-associated (P101L) mouse PrP molecules in cholinergic neurons exhibited severe locomotor dysfunction and premature death as larvae and adults. Clinical abnormalities were accompanied by age-dependent accumulation of misfolded PrP molecules, intracellular PrP aggregates, and neuronal vacuoles. These transgenic flies thus exhibited several hallmark features of human Gerstmann–Ströussler–Scheinker syndrome [137].

To model amyotrophic lateral sclerosis (ALS), a motor neuron disease, mutant forms of the gene encoding copper, zinc-superoxide dismutase (SOD1) was ectopically expressed in motor neurons in Drosophila. This resulted in progressive climbing deficits accompanied by defective neural circuit electrophysiology, focal accumulation of human SOD1 protein in motor neurons, and a stress response in surrounding glia [138]. Unfortunately, toxicity was not associated with SOD1 oligomerization or neuronal loss in this model. It is possible that the absence of certain mammal-specific co-factors may have hindered full recapitulation of the disease symptoms in Drosophila.

Dominant mutations in vesicle-associated membrane protein-associated protein B (VAPB) cause ALS8, a slowly progressive motor neuron disease characterized by fasciculation, cramps, and postural tremor. A fly model of ALS8 was established using the corresponding mutation in Drosophila VAPB (dVAP33A). Neuronal expression of disease-associated VAP(P58S) causes phenotypes resembling VAP loss of function mutants and are opposite those of VAP overexpression, suggesting that VAPP58S may function in a dominant-negative fashion. VAP(P58S) formed aggregates and recruited wild type VAP into these aggregates. Further, the ALS8 mutation in dVAP33A was found to interferes with the BMP signaling pathway involved in synaptogenesis at the NMJ [139]. In another study, it was shown that the MSP domains of VAP proteins are cleaved and act as secreted ligands for Eph receptors. The P58S mutation led to a failure to secrete the MSP domain; instead, it induced ubiquitination, accumulation of inclusions in the endoplasmic reticulum, and an unfolded protein response [140]. This study offers new insight into pathogenic mechanisms of ALS.

Spinal muscular atrophy (SMA) and Batten disease

Other rare neurodegenerative disease have also been successfully modeled in Drosophila. Mutations in the survival motor neuron (SMN) gene are associated with SMA, a recessive hereditary neurodegenerative disease characterized by early onset lethality coupled with motor neuron loss and skeletal muscle atrophy. The Drosophila SMN concentrates in the post-synaptic regions and is required for synaptogenesis at the NMJ. A genetic modifier screen resulted in the recovery of modifiers not previously known to be associated with this disease. Among the identified modifiers are components of the BMP signaling pathway previously shown to act in a retrograde signaling process to control synaptogenesis at the NMJ [141]. This study implicates synaptic dysfunction as a major cause of SMA.

Mutations in the gene CLN3 are linked to the neurodegenerative disorder Juvenile Neuronal Ceroid Lipofuscinosis (JNCL), also called Batten Disease. CLN3 encodes a conserved multi-spanning and hydrophobic transmembrane protein whose molecular function is still unclear. A genetic gain-of-function approach was used in Drosophila to identify functional pathways that involve CLN3. Previously unknown genetic interactions between CLN3 and the Notch and JNK signalling pathways were revealed [142]. Further studies of this Drosophila model of JNCL will help understand the pathogenesis and offer strategies for the treatment of this devastating disease affecting young children.

Concluding remarks

The development of highly relevant animal models of human diseases has ushered in a new era in neurodegenerative diseases research. Drosophila proves to be an excellent model organism for studying the normal biological function of genes linked to neurodegenerative diseases and how mutations in these genes leads to neuronal dysfunction and cell death. The true power of fly models lies in the ability to perform large-scale forward genetic screens to dissect the cellular signaling pathways involved in the disease process, without making assumptions about what kind of molecules or mechanisms are involved. Once novel mechanisms or molecules are identified in the fly models, their relevance to mammalian system should be verified. Ongoing and future genetic studies should help answer some of the unresolved questions in the field, such as the molecular nature of the neurotoxic species for each disease type, the key neuronal functions being affected by the neurotoxic agents, and the determinants of the cell type-specific vulnerability underlying each disease. The genetic modifiers identified in flies may also help reveal disease susceptibility genes in humans and ultimately offer new therapeutic targets.

Abbreviations

4E-BP

Eukaryotic initiation factor 4E-binding protein

AD

Alzheimer’s disease

ALS

Amyotrophic lateral sclerosis

APP

Amyloid precursor protein

BACE

β-site APP-cleaving enzyme

CEBP

cAMP-responsive element-binding protein

FMR

Fragile × mental retardation

HD

Huntington’s disease

HDAC

Histone deacetylase

JNCL

Juvenile neuronal ceroid lipofuscinosis

LRRK2

Leucine-rich repeat kinase 2

NFT

Neurofibrillary tangles

NMJ

Neuromuscular junction

Pael-R

Parkin associated endothelin-like receptor

PAR-1

Partioning defective-1

PD

Parkinson’s disease

PINK1

Pten-induced kinase 1

RNAi

RNA interference

SBMA

Spinobulbar muscular strophy

SCA

Spinocerebellar ataxia

SMA

Spinal muscular atrophy

SOD1

Copper, zinc-superoxide dismutase

VAPB

Vesicle-ass ociated membrane protein-associated protein B

SMN

Survival motor neuron