Modelling human brain diseases in Drosophila melanogaster offers several advantages for investigation of molecular and cellular mechanisms underlying human disease. Short life span, large number of offspring, many genetic techniques, a well known anatomical situation and a wide variety of mutants are convenient characteristics of Drosophila as a model organism. Time and tissue specific inducible promoters are available [1–3]. Anatomic divergence between the fruit fly and humans is apparent, which may be not sufficient to recapitulate some morphological features of neurological disease but fundamental molecular pathways are highly conserved [4]. Functional analysis of human disease genes including high–throughput pharmacological screens as well as behavioral assays have become available in Drosophila.

Sequencing of the Drosophila genome revealed about 13.600 genes [5], being less than the human genome [6,7] and comparison of the genomes of humans and Drosophila reveals further differences concerning a higher microsatellite mutation rate in humans, whereas nucleotide diversity is more prominent in Drosophila [8] but this does not compromise Drosophila as a disease model.

One obvious disadvantage of using fly models is the risk that important pathogenetic factors are vertebrate-specific and may be ignored in invertebrate models. For example, immunological diseases such as multiple sclerosis cannot be modelled convincingly in Drosophila melanogaster. Furthermore, brain infarcts and brain hemorrhage cannot be analyzed in Drosophila because vessels are lacking and blood cells are mainly restricted to primitive hemocytes. However, most Drosophila models do reproduce some aspects of human diseases although one has to bear in mind that the differences between mammals and invertebrates represent potential drawbacks in modelling brain diseases.

In Drosophila a wide armamentarium of genetic tools is available. In a reverse genetic approach a candidate gene is tested for its potential functional role. One of the most important genetic systems used in reverse genetic approaches is the GAL4/UAS-system (see Figure 1.). Human proteins or Drosophila proteins can be expressed in a tissue and time dependent manner in the fly [9]. In most Drosophila disease models a human disease related transgene is inserted downstream of a UAS (upstream activating sequence) and can be expressed under the control of the yeast transcriptional activator GAL4. Absence of GAL4 results in inactivity of the transgene. After crossing flies carrying the transgene to flies expressing GAL4 under control of a cell- or tissue specific promoter, human protein expression is subsequently restricted to GAL4-expressing tissues. A large number of GAL4 driver lines including the glial promoter repo (reversed polarity), the pan-neuronal promoter elav (embryonal lethal, abnormal vision) and the eye-specific promoter GMR (Glass Multimer Reporter) are available in Drosophila.

Compared to reverse genetics the forward genetic approach is an unbiased method – as to the function of a gene - for identification of genes based on phenotype. These approaches are usually conducted in one of two ways. On the one hand, a screen for mutations (chemical mutagenesis or insertional mutagenesis techniques (Enhancer-Promoter (EP)-elements and RNAi lines) (see Figure 2.) that reduce life-span, exhibit behavioral abnormalities or induce neuronal degeneration can be carried out. Several genes have been detected using these screens, among those the mutant bubblegum [10–12]. In principle, this approach can be useful to understand diseases whose genetic basis is undetermined yet [13].

On the other hand, modifier screens can be used in order to find modifiers (suppressors/enhancers) of the disease phenotype in a mutant background (see Figure 3). Modifier genes are assumed to encode proteins that are related to pathways involved in the disease phenotype.

Drosophila melanogaster is an insect undergoing metamorphosis, showing different developmental stages: embryo, larva, pupal stages and the adult fly. The central nervous system of the Drosophila embryo is composed of neurons and glial cells (Figure 4). The neurons build commissures in a close association with midline glial cells [14,15]. Glial cells in the Drosophila CNS can be classified either as midline glia or as lateral glia [16]. Surface glial cells form a continuous covering of the CNS and peripheral nerves. They comprise peripheral and exit glial cells, subperineurial glial cells, which enclose the CNS, and channel glial cells lining the channels lancing the ventral nerve cord [17].

In Drosophila, various organs and anatomical structures arise from ten pairs of imaginal discs and the genital disc. The brain of the Drosophila larva is composed of two hemispheres and the subesophageal ganglion (Figure 5).

In brain disease models, brain and the complex eye of adult Drosophila are often used. Especially the mushroom bodies (Figure 6), association areas necessary for olfactory learning and memory [18], composed of about 2,500 Kenyon cells, neurons with small, densely packed cell bodies [19], are relevant in brain disease models. Additionally, the fly visual system is also often used as a model system in brain research [20,21] as it is less complex than the brain, formed by neurons which develop in a very stereotyped manner and can be conveniently investigated. In particular a “rough eye” phenotype caused by pathological processes can be easily studied, enabling screening for genes related to the process of interest. Eye development always begins with the differentiation of R8 photoreceptor neurons at uniformly spaced positions. These neurons signal to neighboring cells to develop ommatidia (unit eyes). Sequential differentiation of the other seven photoreceptor types (R1-R7) follows afterwards [22]. Each R8 neuron recruits one cell of each type, such that seven photoreceptors cluster around each R8 neuron [23] (Figure 6).

Alzheimer’s disease (AD) is the most common neurodegenerative disease characterized by extracellular accumulation of Aß peptide in senile plaques and intracellular accumulation of hyperphosphorylated tau as neurofibrillary tangles and neuropil threads [24,25]. Familial AD, which makes up only less than 1% of cases, is caused by mutations of genes encoding amyloid precursor protein (APP), presenilin 1 and presenilin 2 [26].

Because not all components of APP proteolytic processing are conserved, modelling AD in Drosophila is quite challenging. The Drosophila APP homolog (dAPPL) does not contain the Aß domain and cannot be cleaved in flies [27,28]. Studies in Drosophila overexpressing wildtype human APP (hAPP695), hAPP695 with pathogenic mutations (hAPP-Swedish), full length hAPP with N-terminal myc tag as well as a construct comprising a signal peptide, ßA4 region and a C-terminal myc tag demonstrated γ-secretase activity in flies [29]. BACE activity is not present in Drosophila, whereas γ-secretase, presenilins and nicastrin are conserved [30–32].

Expression of human Aß40 and Aß42 in the fly brain caused age-dependent learning defects but only Aß42 led to formation of diffuse amyloid deposits in the Kenyon cell region positive in the thioflavin S staining but in transmission electron microscopy without clear amyloid fibril structure, neurodegeneration, and decline of locomotor functions [33] Expression of Aß42 in the CNS was found to cause short-term memory impairment [34]. Expression of human Aß42 in the eye caused degeneration of the retina, which could be suppressed by neprilysin overexpression [35]. In another approach human BACE and hAPP were expressed singly or in combination in fly eyes using gmr-GAL4. Expression of hAPP alone or in combination with BACE resulted in degeneration of retinal photoreceptors and age-dependent plaque formation, which could be enhanced by coexpression of the Drosophila presenilin (UAS-DPsn) containing a point mutation corresponding to the FAD mutant L235P. Amyloid deposits were positive for thioflavin S, ß-amyloid as well as Aß42 detecting antibodies in flies expressing hAPP and DPsnL235P or hAPP, BACE and DPsnL235P. Electron microscopy revealed “star-like” deposits in the retina, neuropil regions and optic ganglia [36]. Taken together in this study a sequence of neuronal degeneration followed by plaque formation is described, which does not recapitulate the chronological sequence of AD in humans. Secondly gender differences could be detected with respect to plaque formation being more severe in male than in female fliese, a fact which is not seen in human AD. On the other hand, age dependent progression of neurodegeneration and plaque formation is detected, immunohistochemical staining profile and EM ultrastructure with star-like formation resembles human AD plaques [37].

Further studies using deletion and overexpression of dAPPL, expression of hAPP695 and of hAPP mutants (hAPP-London and hAPP-Swedish) in larval motor neurons found accumulation of organelles along axonal tracts [38], suiting well to axonal transport disturbances seen in transgenic mouse models [39]. The neurofibrillary, tau associated pathology is lacking in Drosophila Alzheimer models, a fact that has also to be considered in the majority of mouse Alzheimer models [40].

Taken together, the fly can be used to study some molecular aspects of Alzheimer disease although the disease models remain more artificial than rodent models.

Neurofibrillary tangles (NFT), neuritic plaques and neuropil threads composed of hyperphosphorylated tau forming intraneuronal inclusions are characteristic of AD, but tau inclusions are also associated with other neurodegenerative diseases such as Pick’s disease, corticobasal degeneration, argyrophilic grain disease, progressive supranuclear palsy and fronto-temporal lobar degeneration (FTLD) [41] characterized by diverse clinico-pathological phenotypes from aphasia to dementia syndromes [42].

Drosophila has one tau gene which is expressed in neurons and localized in axonal processes [43]. By expressing human wildtype tau and FTLD-linked mutant forms of tau (R406W, V337M) in cholinergic neurons a fly model of tauopathies was established showing adult-onset progressive neurodegeneration with vacuolization and reduced life span, particularly in R406W transgenic flies. Abnormal tau protein accumulated but NFT could not be detected [44]. When wildtype human tau (4R isoform) was coexpressed with the fly homolog of GSK-3β (shaggy), presumably involved in hyperphosphorylation of tau, neurodegeneration was enhanced and insoluble intraneuronal NFT formation, positive for AT100 antibody was observed. Ultrastructurally filamentous structures, including paired helical filaments with a characteristic periodicity as well as straight filaments were detected [45]. Antibody AT100 stains both intracellular NFT and extracellular NFT [46]. As a consequence of neuronal expression in this Drosophila model extracellular NFTs cannot be dectected in these flies. Ultrastructural findings in these flies are similar to paired helical filaments and straight tubules found in a wide range tauopathies [47] as well as detergent-resistance, another characteric of tau accumulations [42]. Toxicity of wildtype dtau and htau under control of gmrGAL4 was dosage dependent, and coexpression of these genes with known modifiers of tau differed in approximately half of the investigated Drosophila lines. Comparing dtau and htau in modifier screens in Drosophila model may clarify the degree of functional homology [48].

Antiapoptotic genes (p35, DIAP1 and DIAP2) were shown to reduce tau toxicity [45]. Overexpression of wildtype human tau in motor neurons disrupted axonal transport leading to locomotor phenotypes. Again, these could be enhanced by coexpression of GSK-3β and reversed by GSK-3β inhibitors lithium and AR-A014418 [49]. Modifier screens revealed a number of kinases, among those par-1 [48,50,51] to be involved in tau related neurodegeneration.

Thus, Drosophila tau models replicate several important features of human tauopathies, including tau hyperphosphorylation, accumulation, NFT formation and neuronal degeneration.

Parkinson’s disease (PD) is a movement disorder showing resting tremor, rigidity, akinesia and postural instability, which mostly occurs sporadically, while hereditary cases represent less than 10% of patients [52]. Familial PD cases have been related with mutations, duplications and triplications in SNCA (α-synuclein)/PARK1, parkin/PARK2, UCHL-1 (ubiquitin carboxy-terminal hydrolase L1)/PARK5, DJ-1/PARK7, PINK-1 (PTEN-induced putative kinase)/PARK6, LRRK2/PARK8, ATP13A2 (p-type ATPase)/PARK9, and HTRA2 (HtrA serine peptidase 2)/PARK13. Other PARK loci have been identified, but the mutated gene is unknown [52]. Neuropathological hallmarks are loss of dopaminergic nigrostriatal neurons and typically accumulation of α-synuclein in cytoplasmic inclusions called Lewy bodies (LB) and Lewy neurites [52].

To create a fly model of PD, wildtype human α-synuclein and two familial mutant forms (A30P and A53T) were expressed in dopaminergic neurons. Expression of α-synuclein led to age-related loss of dopaminergic neurons, LB-like accumulations and behavioral deficits. In flies expressing mutant α-synuclein in a pan-neuronal pattern α-synuclein inclusions were also found in non-dopaminergic neuronal cell bodies like in human PD brains [53]. LB-like structures in the fly stained positive for α-synuclein and were ultrastructurally composed of filaments with granular material similar to human Lewy bodies. [53,54]. l-DOPA, a drug used to treat PD patients, could suppress behavioral defects [55].

Mutations of PARK2 encoding parkin are linked to autosomal recessive, juvenile onset parkinsonism [52]. Loss of function mutations of Drosophila parkin increased sensitivity to oxidative stress [56,57] and showed pathological mitochondrial structure although no dopaminergic neurodegeneration could be observed. Other studies suggested parkin protection against α-synuclein damage to dopaminergic neurons [58]. In another model expression of two mutant forms of human parkin, Q311X and T249R in dopaminergic and serotoninergic neurons under control of the ddc-GAL4 driver, selectively caused degeneration of dopaminergic neurons and progressive locomotor dysfunction [59].

In humans mutations in PINK-1 lead to early-onset autosomal PD [52]. Loss-of-function mutations of the Drosophila PINK1 homolog show dopaminergic neuronal degeneration, flight muscle degeneration, locomotor defects and mitochondrial defects. Interestingly, expression of parkin could ameliorate PINK1 phenotypes [60].

Mutations of DJ-1 lead to an early onset autosomal recessive variant of PD [52]. Loss of the two DJ-1 homologs in Drosophila DJ-1α and DJ-1β was investigated and flies with deletions of DJ-1 β and DJ-1β were viable, fertile and showed normal lifespan. Interestingly these flies were selectively sensitive to toxins like paraquat and rotenone, linked to sporadic PD in humans [61]. Loss of function of DJ-1β led to locomotor deficits without loss of dopaminergic neurons [62].

Mutations of LRRK2 cause a late onset autosomal-dominant form of PD [52]. In flies expressing wildtype fly LRRK and the Arg1069Cys mutation corresponding to pathogenic Arg1441Cys mutation in LRRK2 associated with PD [63] under control of various GAL4 driver lines inducing whole body, muscle, eye and dopaminergic neuron specific expression did not show defects, but LRRK loss-of-function mutants showed decreased locomotor activity and reduction of dopaminergic neurons [64]. A second gain-of-function LRRK2-PD model was established using human LRRK2 and LRRK2-G2019S, another mutation associated with PD [65,66]. Expression of wildtype and mutant human LRRK2 in photoreceptor cells by gmr-GAL4 led to neuronal degeneration and expression in dopaminergic neurons resulted in selective loss of those neurons, locomotor dysfunction and reduction of life span [67].

Rotenone treatment of wildtype flies led to loss of dopaminergic neurons and locomotor defects [68]. Treatment of wildtype flies with paraquat led to impaired climbing capability and decreased survival which could be restored by cannabinoid receptor agonists (CP55,940) and a specific inhibitor of stress responsive Jun-N-terminal kinase signalling (SP600125) to different extent. [69].

Taken together, several genes involved in PD as well as PD models utilizing toxins have been investigated in the fly. To a remarkable extent neuropathological hallmarks could be modeled in Drosophila. Interdependence of different genes can be suitably investigated in this organism and is expected to influence understanding of the disease.

Prion diseases are rare fatal neurological diseases of genetic or infectious origin, but most often occur sporadically. In humans, sporadic (sCJD), familial (fCJD) and variant (vCJD) Creutzfeldt-Jakob disease, Gerstmann-Sträussler-Scheinker disease (GSS), fatal familial insomnia (FFI) and kuru are known [70]. Prion diseases are induced by misfolding of prion protein PrP^C into one of several pathogenic isoforms [70,71].

GSS disease, one of the inherited prion diseases can be caused by some mutations of the prion protein gene, but most commonly by a point mutation at codon 102 and methionine at position 129 [72]. Histologically in GSS, widespread, large, multi-centric amyloid plaques with a dense core encircled by satellite globules located predominantly in the cerebral cortex and cerebellum prevail, positively stained by PrP antibodies, accompanied by white matter degeneration and neuronal loss, spongiform changes, gliosis and NFTs. Ultrastructurally amyloid plaques consist of radiating bundles of curvilinear filaments without definite periodicity [73].

The first effort to model a prion disease in Drosophila, an organism which does not have a prion gene, involved the expression of Syrian hamster prion protein under control of a HSP70 promoter [74]. Transgenic flies produced full-length prion protein on heatshock. No phenotypes were observed and the expressed prion protein never appeared to achieve the pathological protease-resistant form [74]. In the only prion fly model so far, a mouse prion protein with a proline to leucine mutation (P101L), homologous to a human mutation (P102L) causing GSS was used [75], leading to locomotor dysfunction, reduced life span, progressive vacuolar pathology and PrP inclusion bodies. Increase in proteinase K resistance (a specific feature of PrP^Sc diseases) was not seen, but biochemical analysis suggested that PrP is aberrantly folded. Transmissibility to other flies and ultrastructure of inclusions were not investigated [75].

This model can be useful with regard to understanding the pathobiology of P102L. GSS is a transmissible disease characterized mainly by proteinase K-resistant PrP positive plaques, amyloidosis and spongiform degeneration whereas the fly model shows vacuolation and inclusions rather than plaque formation, but not proteinase K resistance. Although neuropathology and biochemistry appear to be different in these two species, modifier screens may be of interest to better understand selected molecular pathogenetic pathways.

Several hereditary neurodegenerative diseases known as polyQ diseases result from CAG repeat expansion within the respective disease genes [76]. These include Huntington disease, X-linked spinobulbar muscular atrophy (SBMA, Kennedy disease) and the spinocerebellar ataxias SCA1, SCA2, SCA3 (also known as Machado-Joseph disease), SCA6, SCA7 and SCA17 as well as dentatorubral pallidoluysian atrophy (DRPLA).

Amyotrophic lateral sclerosis (ALS) is characterized by upper and lower motor neuron defects including brisk reflexes, spasticity and pathological reflexes, fasciculations and weakness [99]. More than 90% of ALS cases are sporadic, but the disease can be inherited as an autosomal or X-linked familial condition. 15% of familial cases are caused by mutations in the gene encoding the enzyme superoxide dismutase-1 (SOD1) [100]. Histologically, ALS is characterized by selective loss of motor neurons in the brain and spinal cord. ALS is characterized by ubiquitin-positive inclusions (“skein-like”, “Lewy body-like” or a combination of the two), small eosinophilic inclusions in spinal motor neurons (“Bunina bodies”), hyaline inclusions [101] as well as TDP-43 positive filamentous inclusions [102].

Expression of hSOD1 (WT and A4V and G85R mutants) in motor neurons using the D42 motor neuron driver led to progressive motor dysfunction, whereas expression of Drosophila Superoxide dismutase (Sod) using the same driver line did not show an effect [103]. Loss of climbing ability was not restricted to mutant forms of hSOD1 but was also observed in wildtype hSOD1. Hypothetically hSOD1 may be recognized as toxic mutant form of SOD1 in Drosophila and therefore led to this effect. Interestingly neither loss of motor neurons nor retinal degeneration was observed. HSOD1 forms intracellular inclusions but solubility was not altered. [103].

A recent publication of a Drosophila model of ALS8 [104], established by using the respective Drosophila mutation (dVAP33A) of the human disease causing dominant mutation of VABP (vesicle-associated membrane protein (VAMP)-associated membrane protein B) showed that the the mutant protein aggregates, is ubiquitinated and recruits wild type protein into aggregates. Furthermore a new mechanism underlying ALS pathogenesis at the neuromuscular junction involving BMP signaling pathways could be identified [104]. In Drosophila synapses at the neuromuscular junction are glutamatergic, which provides similarities to the spinal cord synapse being of central interest in ALS in humans [104].

Earlier attempts to model ALS yielded confusing results as expression of hSOD1 transgene and human mutated Gly41-to-Ser SOD in motorneurons by D42-GAL4 led to increased longevity and could rescue the lifespan of a dSod null mutant with shortened lifespan [105,106].

Taken together, publications about ALS in Drosophila did not show motor neuron loss, the observed intracellular inclusions did not reflect characteristics of human ALS inclusions, and investigations focussed on the neuromuscular junction. But despite of apparent discrepancies, these results will constitute the basis of modifier screens and are expected to contribute to further understanding of ALS pathogenesis.

Leigh disease is a progressive mitochondrial encephalopathy characterized by psychomotor delay starting during the first months of life. Life expectancy is 1–4 years. Optic atrophy is also frequently observed [107]. The disease is caused by inherited mutations in both nuclear- and mitochondrial-encoded genes involved in energy metabolism, including mitochondrial respiratory chain complexes I, II, III, IV, and V, crucial to electron transport [108].

Macroscopically, symmetric necrotizing lesions in subcortical areas of the CNS (brain stem, cerebellum, diencephalon and corpus striatum), which are associated with spongiosis, demyelination and astrocytosis, are observed [109]. Ultrastructurally mitochondrial inclusions are found in the neurons of the brain, as well as increase of mitochondrial antigens in neurons, vessels, plexus and astrocytes [110].

In the fly photoreceptor cells mutant for Succinate dehydrogenase (Sdh) develop completely normally and innervate the appropriate synaptic partners. After some time receptor cells degenerate, progressively losing expression of synaptic markers, and undergoing extensive morphological changes [111]. Thus this model may capture some elements of the human disease.

Another model of Leigh disease was established using downregulation of Surfeit1 (Surf1) showing a number of behavioral and electrophysiological abnormalities including reduced photoresponsiveness, reduced locomotor speed and impaired optomotor response as well as abnormal electroretinograms with different driver lines [112]. Surf1 downregulation driven by an actin-GAL4 line showed an underdeveloped CNS, while CNS wide silencing of Surf1 driven by elav-GAL4 led to prolonged lifespan, normal CNS development, slight impairment of locomotor activity and photobehavior; histochemical reaction to COX was reduced in the optic lobes.

Ultrastructural investigation of the body wall muscle fibers in actin-GAL4 driven flies showed larger mitochondria, different distribution and morphological alterations [112], while other key features of human brain pathology were not observed.

Remarkably, mutant flies exhibited increased longevity which does not reflect the course of human disease.

Niemann-Pick disease represents a group of lysosomal lipid storage disorders. Niemann-Pick type C (NPC), an autosomal-recessive disease, shows a wide spectrum of phenotypes with variable begin from perinatal period to adult age. Major neurological symptoms include cerebellar ataxia, dysarthria, dysphagia, seizures and progressive dementia [113]. NPC is characterized by accumulation of cholesterol, glycospingolipids and other lipids.

A defect of organelle trafficking and a failure of lipid homeostasis are prominent [114], caused by mutations of NPC1 or NPC2 [115,116]. Histologically, NPC is characterized by progressive loss of neurons, particularly Purkinje cells in the cerebellum, lipid storage, formation of meganeurites and ectopic dendrites as well as the presence of neurofibrillary tangles [117].

NPC1a null alleles in Drosophila die at an early larval stage, but feeding NPC1a mutants the steroid hormone (molting hormone) 20-hydroxyecdysone (20E) extends lifespan, suggesting that reduced ecdysone synthesis results from NPC1a loss. Feeding with excess cholesterol compounds extends lifespan further till adult stages. [118].

In another Drosophila dnpc1a model flies showed sterol accumulation as in human disease. By 7-dehydrocholesterol treatment life expectancy of dnpc1a mutants could be extended till adulthood. Brain morphology was unremarkable without any neurodegenerative changes [119].

In another study using the same dnpc1a mutants more extensive brain investigation revealed neuronal cholesterol deposits, accumulation of multilamellar bodies and age-dependent vacuolization. Age-dependent neurodegeneration, early lethality and movement disorders could all be completely rescued by neuronal and partially rescued by glial expression of wild-type dNPC1a transgene [120]. Npc2a mutants displayed a shorter life span, but did not show any brain vacuolization. TUNEL staining revealed neurons undergoing apoptosis [121].

Taken together, NPC can be modelled suitably in Drosophila, as cholesterol storage and neurodegenerative aspects of the disease are represented. On the other hand, one has to keep in mind two potential drawbacks of the invertebrate model when interpreting the results. Firstly Drosophila and other insects have redundant npc1 and npc2 genes (npc1a, npc1b, npc2a, npc2b), whereas mammals including humans only possess one NPC1 and one NPC2 gene. Functions of the other npc genes are not understood yet. Secondly, steroid actions in flies and humans are certainly different as flies cannot synthesize sterol, in particular the molting hormone 20E.

Ceroid lipofuscinoses are characterized by variable but mainly pediatric onset, vision loss, motor dysfunction, seizures and decline of intellectual capacities. [122]. Several causative genes, including CLN2-3, 5–8, 10, PPT1 and MFSD8 were identified [123,124]. Infantile neuronal ceroid lipofuscinosis (INCL), caused by loss of palmitoyl-protein thioesterase 1 (PPT1) on 1p32, shows brain atrophy, neuronal swelling, sudanophilic changes, granular osmiophilic deposits, lysosomal accumulation positive for acidic phophatase in neuronal and astrocytic cells, and rarefaction and shrinking of corticabasal and bulbar neurons [122].

Palmitoyl-protein thioesterase 1 (Ppt1) mutant flies have reduced life span and CNS-specific accumulation of autofluorescent storage material, which unlike human granular osmiophilic deposits were homogeneous in structure and composed of concentric layers of material [125]. The deposits may also be biochemically different as they could not be detected with lipophilic stains [125].

Targeted overexpression of Ppt1 in the Drosophila visual system results in apoptotic neuronal cell loss, leading to misorganized ommatidia [126]. Performance of a gain-of-function modifier screen using enhancer-promoter lines could connect Ppt1 function to synaptic vesicle cycling, endolysosomal trafficking and synaptic plasticity [127]. Observations in Drosophila have to be analyzed carefully as downregulation of Ppt1 leads to accumulation of storage material, different from human GRODs without neurodegeneration und thus Ppt1 in Drosophila might have a different function than PPT1 in humans.

Although human disease is caused by loss of PPT1 and not overexpression, data from Drosophila studies reveal that misregulation of PPT1 may lead to neuronal cell loss and thus the correct titration of enzyme activity may be of importance. Nevertheless the possibility of modifier screens may reveal the physiological role of Ppt1 in Drosophila and this may promote understanding human disease.

Neurofibromatosis type 1 (NF1) is an autosomal dominantly inherited neurocutaneous disorder due to mutations in NF1 on 17q11.2 [128]. NF1 is a common disease that mainly affects peripheral and central nervous system (neurofibromas, optic gliomas, astrocytomas, malignant peripheral nerve sheath tumors), the skin (café au lait spots, axillary and inguinal freckling), and may show further neuroendocrine/neuroectodermal tumors, hematopoietic tumors, osseous lesions, iris hamartomas and intellectual handicap [128].

Histologically dermal neurofibromas, well-circumscribed benign tumors composed of Schwann cells, as well as plexiform neurofibromas, producing diffuse enlargement of nerve trunks prevail [129]. Malignant peripheral nerve sheath tumors, highly aggressive tumors characterized by a herringbone pattern of cell growth, are localized within nerve fascicles but invade the adjacent soft tissues [129]. Gliomas are most often pilocytic astrocytomas within the optic nerve and bilateral growth is not uncommon in NF1 patients [129].

Manipulation of NF1 gene was investigated during the last ten years in Drosophila showing that neurofibromin gene plays a role in tissue growth in all developmental stages [130], learning [131], circadian rest-activity rhythm [132], lifespan determination [133], Ras and cAMP interactions [133,134] and stress resistance [133]. Whether these findings will gain therapeutic consequences, for example by using antioxidative drugs, remains controversial. The main point that neurofibromatosis is characterized by appearance of multiple tumors in humans is not represented in all studies yet, thus modelling this disease in Drosophila remains disappointing.

Neurofibromatosis type 2 (NF2) primarily affects the nervous system, bilateral vestibular schwannomas being prominent. Schwannomas of other cranial nerves, meningiomas, ocular abnormalities, meningioangiomatosis, glial hamartomas and neurofibromas may also be present [128]. NF2 is an autosomal dominantly inherited neurocutaneous disorder due to mutations in NF2 on 22q12.2 [128].

In Drosophila, Merlin mutations have been investigated and were shown to regulate cell growth and cell cycle [135]. It could be shown that Merlin is part of the hippo pathway [135]. Tumor formation in Drosophila could not be established by using Merlin mutations [136]. How these results could influence possible therapeutic options in NF2 patients remains to be elucidated.

Tuberous sclerosis (TSC) is an autosomal dominant disease due to heterozygous mutations in TSC1 on chromosome 9q34 or TSC2 on chromosome 16p13 [137]. TSC is a neurocutaneous disorder characterized by brain abnormalities (cortical tubers, subependymal giant cell astrocytomas (SEGAs), subependymal glial nodules, seizures, mental retardation, autism and attention deficit-hyperactive disorders), kidney pathologies (angiomyolipomas, cysts and renal tumors) as well as rhabdomyomas of the heart [138].

Cortical tubers are strongly associated with the development of epilepsy, especially infantile spasms. They consist of giant cells, dysmorphic neurons, disrupted cortical lamination, gliosis and calcifications [139]. SEGAs are well-circumscribed, often calcified tumors with a mixed glioneuronal phenotype.

In Drosophila, gigas (gig) was identified as homolog of TSC2, leading to increase of cell size and imaginal discs as well as abnormal cell cycle progression [140]. While human SEGAs contain giant cells, corresponding Drosophila tissues in gigas mutated flies show hypertrophic changes but no distinct brain tumors.

Furthermore, several fly mutants interfering with asymmetric cell division of neuroblasts exhibit neuronal/neuroblastic tumors which are referred to as “hyperplastic” in case of preserved architecture such as malignant brain tumor (l(3)mbt) [141], or “neoplastic” with loss of architecture and invasion such as brain tumor (brat), raps, numb, pros and mira, the latter four genes encoding for epithelial polarity proteins [142]. Results from these experiments show that maintenance of properly organized apical and basolateral domains is essential to prevent movement of cells out of an epithelium. Mutation of these genes may therefore contribute to metastasis. Different approaches have been pursued in Drosophila to model metastatic spread. Transplantation experiments with brat, dlg and lgl mutated brains into abdomens of adult flies showed that serial transplantations may lead to selection of more aggressive cell clones [143].

Modifier screens have been carried out and Semaphorin-5c and apontic were found to inhibit metastasis in a lgl mutant [144]. These kinds of screens will provide new insights into pathways involved in metastatic events.

Also reverse genetics can be applied to study invasion events as Matrix metalloproteinase 1 (Mmp1) known to promote invasion in humans was upregulated in metastasis models using scribbled (scrib) LOF mutations. Loss of Mmp1 function by RNAi suppressed invasiveness [145].

Although some important anatomical features like vessels are lacking in Drosophila and polarity genes will be more important for carcinoma formation than for endogenous brain tumors, these models may help to understand a variety of basic cell biological features underlying tumor biology and may lead to new concepts in human tumor therapy. Remarkably, fly models of gliomas and medulloblastomas, the most common human brain tumors in adults and childhood, respectively, have not been published yet.