Using High-Throughput Animal or Cell-Based Models to Functionally Characterize GWAS Signals

Using High-Throughput Animal or Cell-Based Models to Functionally Characterize GWAS Signals Purpose of Review The advent of genome-wide association studies (GWASs) constituted a breakthrough in our understanding of the genetic architecture of multifactorial diseases. For Alzheimer’s disease (AD), more than 20 risk loci have been identified. However, we are now facing three new challenges: (i) identifying the functional SNP or SNPs in each locus, (ii) identifying the causal gene(s) in each locus, and (iii) understanding these genes’ contribution to pathogenesis. Recent Findings To address these issues and thus functionally characterize GWAS signals, a number of high-throughput strat- egies have been implemented in cell-based and whole-animal models. Here, we review high-throughput screening, high-content screening, and the use of the Drosophila model (primarily with reference to AD). Summary We describe how these strategies have been successfully used to functionally characterize the genes in GWAS-defined risk loci. In the future, these strategies should help to translate GWAS data into knowledge and treatments. . . . . . Keywords HCS HTS Drosophila Screen GWAS Alzheimer Introduction disequilibrium patterns involving the sentinel SNP may make it difficult or even impossible to determine which gene is Genome-wide association studies (GWASs) determine groups responsible for the observed association. Hence, identifying of single nucleotide polymorphisms (SNPs) in linkage dis- the causative gene and one or more functional SNP are major equilibrium and which are associated with a particular disease, challenges in the GWAS field. For instance, new technologies trait, or phenotype. The most significantly associated SNP is have been developed to tackle this limitation with the obser- usually not the causative SNP, and (by convention) the signal vation that 85–95% of the GWAS-associated SNPs are signif- is assigned to the closest gene. However, the presence of sev- icantly enriched at cell-type-specific regulatory regions: mas- eral genes in a GWAS locus and complex linkage sively parallel reporter assay (MPRA) allowed to screen thou- sands of variants in order to assess their impact on transcrip- This article is part of the Topical Collection on Neurogenetics and tional activity [1]. A further difficulty (even when the risk Psychiatric Genetics gene is known) relates to the determination of how the caus- ative genes are functionally involved in the disease process. * Pierre Dourlen Indeed, it can be difficult to establish a causal link on the basis pierre.dourlen@pasteur–lille.fr of the literature data alone. To take account of these limitations, gene enrichment path- Julien Chapuis way analyses have been developed from GWAS dataset. The julien.chapuis@pasteur-lille.fr main idea is that genes associated with the disease risk will be Jean-Charles Lambert over-represented in specific pathways involved in the disease jean-charles.lambert@pasteur-lille.fr process. In Alzheimer’s disease (AD), these analyses have pointed to the immune response, the regulation of endocyto- INSERM U1167, RID-AGE-Risk Factors and Molecular Determinants of Aging-Related Diseases, Lille, France sis, cholesterol transport, and protein ubiquitination [2]. However, it must be borne in mind that this type of analysis Institut Pasteur de Lille, Lille, France is subject to major limitations (in addition to methodological University Lille, U1167-Excellence Laboratory LabEx DISTALZ, issues): for example, the defined canonical pathways are far Lille, France Curr Genet Med Rep from complete, and many numerous genes have pleiotropic High-Throughput Screening in Cell-Based functions and are thus nominated in many different pathways. Models It is also possible that a gene may have an unknown function with relevance to the pathophysiological context. According to the amyloid cascade hypothesis (one of the main Given this background, it appears that gene enrichment hypotheses in AD, in which the overproduction of amyloid pathway approaches are not able to optimally exploit the ge- (Aβ) peptides leads to neurotoxicity and then neuronal death), netic data generated by high-throughput genomic strategies. It one expects at least some of the GWAS-defined genes to con- is now possible to develop in silico methodologies by com- trol the production of the Aβ peptides that accumulate in the bining numerous datasets with the objective to define both AD brain. For decades, the impact of AD candidate genes on functional variants and genes [3]. However, there is also a Aβ secretion/production was evaluated one by one. Given the need for alternative, powerful approaches for empirically test- large number of AD candidate highlighted by GWASs, these ing multiple GWAS genes in cell-based or animal models. approaches no longer appear to be appropriate. Indeed, the Here, we review high-throughput functional screening strate- assessment of a putative AD gene’simpacton Aβ peptides gies that use in vitro cell-based models and the in vivo production requires cell-based models that are suitable for use Drosophila model for GWAS-defined hits (primarily in AD in fast, large-scale screening. but also for some other neurological diseases) (Fig. 1). We will One of the first attempts to systematically measure the im- not cover Caenorhabditis elegans—another invertebrate ani- pact of GWAS-defined genes involved the RNAi-mediated mal model [4] that has also been used [5]. silencing of 24 late-onset AD genes in Hela cells stably Fig. 1 Schema describing the main steps to set up systematic screenings for GWAS-genomic data (in blue, specific to HCS/ HTS; in green, specific to Drosophila) Curr Genet Med Rep sw over-expressing APP [6]. The levels of APPs-β and Aβ inhibitor was associated with a specific increase in YFP fluo- 1–38/ secreted into the supernatant were then measured. Bali rescence intensity (corresponding to the APP’s C-terminal 40/42 et al. proposed that late-onset AD genes did not specifically fragment). After customization for automatic image process- alter the Aβ42/40 ratio but probably contributed to AD ing, we screened a genome-wide bank of 18,107 human through a distinct mechanism. However, this study was limit- siRNAs. In total, 832 hits were selected as being likely to have ed by the number of genes tested and the number of an impact on APP metabolism—including 8 genes associated endophenotypes measured. Furthermore, the number of loci with the risk of late-onset AD in the reference GWAS meta- detected by GWAS is also increasing, meaning that this type analysis. These data suggested that the 8 genes are involved in of low-scale study is of less interest. AD process via the regulation of APP metabolism. Technological progress has enabled the large-scale automa- For HTS/HCS data, it is important to bear in mind that the tion of biological experiments. These techniques are particu- choice of statistical methods used to define the initial hit list is larly appropriate for generating large amounts of data in cell- critical; this will determine the relevant secondary analyses based models, and mark the entry of cell biology into the age needed to narrow down the hits and thus investigate biological of high-throughput methodologies (as has already happened mechanisms. To assess the relevance of these genes with re- for genetics, epigenomics, and transcriptomics, for example). gard to Aβ peptide levels, we cross-checked our HCS data Rather than focusing on subsets of genes, progress in genome against an association study of SNPs and CSF Aβ42 peptide annotation enables the systematic performance of high- levels in a large sample of patients (n = 2950). Only SNPs throughput screening (HTS)—for instance by designing within FERMT2 were associated with low Aβ42 peptide RNAi constructs to test the effects of gene silencing on bio- levels; this highlighted FERMT2’s potential role in the AD logical phenotypes. process via the modulation of APP metabolism and Aβ pep- This type of systematic analysis was performed with tide generation. Lastly, the impact of FERMT2 expression on 14,603 siRNA pools in HEK-293 cells stably expressing a APP metabolism was validated in several cell-based models, mutant form of APP (NFEV) designed to enhance Aβ pro- including a primary neuronal culture (PNC) endogenously duction. Conditioned media were used to quantify Aβ40, expressing both APP and FERMT2. Therefore, accurate Aβ42, sAPPα,and sAPPβ levels [7, 8]. A “regulatory land- choice of statistical methods, crosschecking with other scape of APP processing” was generated, in order to identify screens, and validation in low-throughput models enabled most of the pathways involved in the regulation of APP me- identifying FERMT2 as a risk factor modulating APP metab- tabolism. Interestingly, some of the identified pathways con- olism. The low number of final positive hits suggests that the tain genes (CLU, BIN1, CR1, PICALM, TREM2, SORL1, AD genetic risk factors in the other loci may be involved in MEF2C, DSG2, EPH1A) linked to AD in the reference different processes than the APP metabolism. This is a limita- tion of the high-throughput methodologies. They are designed GWAS meta-analysis [9]. It is noteworthy that only the aggre- gate effect of genes in a pathway were considered as real with respect to a defined phenotype. Most of the current effects, due to the high false-positive rate and the low level models are based on neuronal dysfunctions involved in AD of reproducibility after siRNA library screening [10]. processes through APP metabolism and Aβ production. By As mentioned above, assays of Aβ released in the cell using these approaches, it is not possible for example to in- culture medium have been used to identify modulators of vestigate APOE or TREM2 functions which are thought to be APP metabolism. However, the underlying mechanisms con- mainly involved in Aβ clearance. trolling APP maturation and trafficking cannot be character- Other cell-based models compatible with HCS analysis ized solely by studying the amount of Aβ present in culture have been generated to assess other phenotypes and could media. Thus, it was necessary to move to the high-content be used to test multiple GWAS genes—for instance QBI- screening (HCS) of multidimensional phenotypes; this con- 293 cells with Dox-regulated inducible expression of human sists of the cell-based quantification of several processes si- tau carrying the P301L mutation, and a GFP tag attached to multaneously; the combination of cell-based and imaging visualize tau inclusions [13]. More recently, GFP bimolecular methodologies provides a more detailed representation of the and trimolecular fluorescence complementation (biFC and cell’s response to various perturbations than HTS does. In line triFC) has been used to study the localization and mechanisms with this approach, we developed a rapid, cell-based, HCS of protein multimers (tau and TDP-43) in the context of neu- assay for the intracellular APP fragments in HEK293 cells rodegeneration [14]. These technologies may enable new 695WT stably over-expressing a mCherry-APP -YFP construct functional assays of genetic factors in neurodegenerative dis- 695WT [11]. The modified APP protein is known to be metab- eases through siRNA screening. 695WT olized in the same way as APP [12]. Our model enabled Some non-mammalian cell-based models like the the detection of specific APP products differentially tagged Drosophila S2R+ cell line have also been used to perform with mCherry or YFP (for the N- and C-terminal fragments, unbiased, genome-wide screening of RNAi. Drosophila cell lines usefully have a low level of gene redundancy. RNA respectively). For instance, treatment with γ-secretase Curr Genet Med Rep interference is also straight forward in these cells as dsRNA Drosophila in Genetic Screening are added in the medium without transfecting reagent. This approach was exemplified by a screen for regulators of the Moving on from in vitro cell-based models, unbiased, high- translocation of Parkin (a protein whose mutation causes throughput screens of GWAS-defined loci can be performed inherited recessive Parkinsonism) to mitochondria [15]. The in a more integrated biological in vivo context by using rele- genome-wide screen identified 60 genes, which were further vant read-outs in animal models of the target disease. narrowed down to 20 candidate genes with a conserved effect However, there are many obvious constraints associated with on both Parkin translocation and mitophagy in HeLa cells. the use of animals and the inability to systematically screen The top hits belonged to the sterol regulatory element binding hundreds of genes in a murine model of AD. This approach is protein (SREBP)-lipogenesis pathway, including the master nevertheless possible in small invertebrate models, such as the regulator of lipid synthesis sterol regulatory element binding fruit fly Drosophila melanogaster. This fly shows good gene transcription factor 1 (SREBF1). Interestingly, the latter is a conservation with humans and only a little gene redundan- GWAS-detected risk factor for sporadic PD suggesting a cy—enabling easier detection of the effects of loss of function. mechanistic link between inherited recessive PD and sporadic Two thirds of human disease-associated genes are estimated to PD—a long debated question [16]. have a functional homolog in Drosophila [18]. In contrast, In order to explore many potential functions of the gene regulatory regions are much less well conserved, there- genetic risk factors, it is important to highlight the grow- fore it is difficult to use Drosophila to identify a causative non- ing need to measure different phenotypes in different cell coding SNP if it affects gene expression. Drosophila has a types. Indeed, gene expression that are cell-specific, like short life cycle, with a new generation produced every 10 days TREM2 (preferentially expressed in microglia), requests at 25 °C. When the latter advantage is combined with high relevant and adapted cellular model. For this purpose, numbers of progeny and the low cost of housing, it is easy to iPSC will likely play an important role in the future to produce the large number of individuals required for optimal generate multiple cell types like neurons, astrocytes, mi- genetic studies. Furthermore, many genetic tools have been croglia, or endothelial cells susceptible to be involved in developed in more than a century of use in forward and re- pathophysiological mechanisms. However, compared with verse genetics. One can therefore modulate more or less any easy-to-transfect standard cell lines, HCS approach using genes, anywhere and at any point in the life cycle [19, 20]. these iPSCs-derived cells will request viral transduction to Mutants are available for almost all the genes in the modify gene expression and will thus lead to screen Drosophila genome [21]. Collections of RNA interference smaller number of genes in these different models for and overexpression constructs have also been generated at the moment. A lentiviral RNA interference library of the genome-wide scale. Expression of these constructs is usu- 597 shRNAs was already used to screen for novel regu- ally based on the Gal4/UAS system [22]. Modifications can lators of synapse formation [17]. In addition, the develop- make the system conditional, with expression controlled by ment of the CRISPR-CASS9 technology and sgRNA li- changing the temperature (Gal80ts, temporal and regional braries should enable large-scale DNA editing and in- gene expression targeting—TARGET—system, [23]) or by crease even more the power of HCS. adding a drug to the flies’ food (the GeneSwitch system; In conclusion, there is no doubt that HTS/HCS methodol- [24, 25]). Thus, Drosophila offers many tools for tightly con- ogies are very powerful ways of potentially assigning patho- trolling gene expression. Furthermore, Drosophila has a physiological functions to GWAS-defined genes. It is likely wealth of external phenotypes (the size, color, and shape of that their use will increase over the coming years, although the eye or of the wing, etc.) that are easily scorable under a this requires major investments in dedicated, automated tech- dissecting microscope. This is essential for quickly screening nology platforms. It is important to bear the following key the effect of gene modulation. Lastly, Drosophila has higher issues in mind: (i) HTS/HCS models face many challenges, cognitive functions (like memory and sleep) for which behav- including the identification of the best experimental system ioral and electrophysiological assays are available. These fea- and the development of robust, reproducible assays; (ii) tures make Drosophila an ideal model for neurogenetics. HTS/HCS models are often highly sensitive, and can generate a large number of unspecific, biologically irrelevant re- sponses; (iii) in view of the number of analyses to be per- Drosophila for Screening GWAS-Defined formed, HTS/HCS approaches must balance the risk of ob- Genes serving significant results by chance against the risk of rejecting biologically valid hypotheses on purely statistical In this context, Drosophila has beenusedtoscreenGWAS grounds; and (iv) as with all data generated in high- candidate genes associated with AD [26–28]. Starting from throughput assays, the most relevant observations need to be the reference GWAS meta-analysis [9], we screened for the replicated and validated in other, unrelated cell-based models. AD candidate genes that modified tau neurotoxicity [28]. Curr Genet Med Rep Based on the regional association plots, we identified a total of regulate autophagy and tau degradation [30]. PD is character- 148 genes in the 19 AD-associated loci. According to the ized clinically by motor and non-motor symptoms and histo- Drosophila RNAi Screening Center Integrative Ortholog logically by dopaminergic neuronal loss and the formation of Prediction Tool (http://www.flyrnai.org/diopt;[29]), 54 of α-synuclein Lewy bodies. The gene coding for cyclin-G- the human genes had a total of 74 Drosophila orthologs at associated kinase (GAK) was identified in a GWAS as a PD 13 loci. Two hundred seventy-eight RNAi lines from five susceptibility gene. To analyze its PD-related functions, Song collections (the Japan NIG collection, the Vienna GD and et al. turned to Drosophila and studied loss of function of the KK collections, and the Harvard TRiP attP2 and attP40 col- Drosophila GAK ortholog (aux)[31]. Knockdown of aux re- lections) and 17 mutant and overexpression constructs were sulted in age-dependent locomotor impairment (in a climbing selected to modulate gene expression (giving four constructs assay), a shorter lifespan, and progressive dopaminergic neu- per gene, on average). The fly eye was used to assess tau ronal loss—as seen in α-Syn overexpression. Downregulation neurotoxicity. Expression of human tau (2N4R isoform) dur- of aux also enhanced α-synuclein-induced dopaminergic neu- ing Drosophila eye development results in small, rough eyes. ronal death, and sensitized flies to the environmental toxin Each of the 295 constructs was tested singly to see whether the paraquat. All these phenotypes represent a broad spectrum size of the tau-expressing eye was modulated. After quantifi- of parkinsonian-like symptoms—supporting the idea that cation of eye size, only genes for which at least two positive GAK/aux is a PD risk factor with a potential role in PD path- constructs from different collections had the same effect were ogenesis. Lifespan and the loss of dopaminergic neurons in considered to be positive hits. By applying these criteria, we Drosophila were also used to identify an interaction between confirmed that Amph (an ortholog of BIN1) modulated tau the PD risk factors PARK16 and LRRK2 [32]. Another ex- neurotoxicity (Chapuis et al. 2013). We also identified four ample of the use of Drosophila to assess disease-like functions new genes: p130CAS, Eph, Focal adhesion kinase (Fak), and is illustrated by studies of the BTBD9 gene. The latter had Rab3-GEF, which are respectively orthologs of CASS4, been identified in a GWAS as a risk factor for restless legs EPHA1, PTK2B,and MADD. Interestingly, three of the five syndrome (RLS), a sensorimotor neurological disorder char- hits (Fak, p130CAS, and Eph) are directly or indirectly in- acterizedby(i) a compellingurgetomoveduringperiods of volved in the cell adhesion pathway. Another Drosophila rest, (ii) relief with movement, (iii) involuntary movements in screen of tau neurotoxicity modifiers also identified genes sleep, and (iv) fragmented sleep [33]. Functional analysis of involved in this pathway [26]. This is one of the advantages BTBD9 was first performed in Drosophila [34]. Loss of the of unbiased systematic screening in Drosophila; the identifi- Drosophila homolog CG1826 (dBTBD9) markedly disrupted cation of new genes that modulate a process (tau neurotoxic- fly sleep, with concomitant increases in waking and motor ity, in this case) and point toward a new pathway (the cell activity. This study also showed that dBTBD9 regulates brain dopamine levels, which is known to correlate with RLS ex- adhesion pathway, in this case) as being a potentially impor- tant pathway for AD pathogenesis. Of course, one can argue pression in humans [34]. Even for behavior like hyperactivity that the tau-associated fly eye phenotype reflects only one or alcohol consumption, Drosophila has provided functional aspect of AD pathogenesis and the results does not explain evidence to complement the sometimes rather correlative ge- how the hits are involved in the pathogenesis. Further work in netic data. In order to better understand dysfunctional reward Drosophila and mammalian models is required to address processing and the discovery of an association between hy- these questions. In our study, we were able to validate the peractivity and the VPS4A gene in a GWAS in adolescents, a genetic interaction between tau and Fak in a cell adhesion- causal role of the sole Vps4 ortholog for hyperactivity was related fly wing readout. By assessing tau phosphorylation validated in Drosophila [35]. Similarly, after the identification and testing a catalytically mutant form of Fak, we could dis- of the association of the AUTS2 gene with alcohol consump- favor a change in tau phosphorylation (classically considered tion at a genome-wide level of significance, functional evi- as one of the culprits in AD pathogenesis) as the cause of the dence was obtained in Drosophila; downregulation of the sole modulation of tau neurotoxicity by Fak in Drosophila. A tau- AUTS2 homolog (tay) led to a lower alcohol sensitivity [36]. PTK2B interaction was confirmed in the brains of a tau mouse With the increasing use of nucleic acid sequencing in re- model and human patients, since we observed an abnormal search and clinical practice, many rare coding variants with somatic accumulation of PTK2B with the appearance of tau unknown functional consequences—called variants of un- oligomers and neurofibrillary tangles [28]. known or uncertain significance, VUS—are being identified. Along with unbiased medium- to high-throughput in vivo Drosophila happens to be a very useful model for addressing screening with easily scorable read-outs, Drosophila is also this issue. The rationale consists in performing functional used to address more specific questions about the functionality transcomplementation experiments in Drosophila,as success- of a gene with respect to the clinical and biological features of fully illustrated for VUS in the TARDBP gene [37]. TARDBP human diseases. The AD risk factor PICALM has been studied encodes the TDP-43 protein, which forms cytoplasmic inclu- in Drosophila. Its ortholog in the fly (Lap ) was shown to sions in patients with the most frequent form of Curr Genet Med Rep frontotemporal dementia (FTD) and most forms of amyotro- Conclusion phic lateral sclerosis (ALS) and in 60% of patients with AD. It is a common disease-causing factor in FTD and ALS. In the context of neurodegenerative disorders in general and However, how TARDBP mutations cause neurodegeneration AD in particular, high-throughput technologies appear to be is not well known, especially with regard to their loss-of- very useful for characterizing the pathophysiological func- function or toxic gain-of-function properties. Null mutants of tions of GWAS-defined genes (Table 1). The next step will the Drosophila ortholog of TARDBP (TBPH) lose the neurons be likely to develop multidimensional high-throughput in the ventral nerve cord that secrete the neurohormone methods allowing to analyze at the same time both functional bursicon [38]. It was shown that expression of human variants and gene functions in accurate cellular types and TARDBP in TBPH-null Drosophila rescued the bursicon neu- models. However, it is important to keep in mind that the more rons, thus indicating functional transcomplementation [37]. the study is complex, the more the statistical analyses and This made Drosophila a platform for testing TARDBP muta- quality control are essential for the success of such sensitive tions. Expression of two typical ALS-causing mutations methods to avoid false-positive/negative results. Furthermore, (p.G287S and p.A315T) could not rescue neuronal loss as confirmation and validation in complementary low- efficiently as wild-type TARDBP did. One atypical variant throughput assays are systematically required. In addition, (p.D169G) could rescue, and another (p.A90V) could not. among practical issues, the high technicality of the methods These findings suggested a partial loss of function in can make them difficult to be mastered and cost effective in TARDBP mutations [37]. The same strategy was successful terms of equipments. for VUS in the TM2D3 and CACNA1A genes, in the context of Beyond these general comments, another limitation of the late-onset AD and ataxia [39, 40]. This method for assessing current high-throughput methods lies on the statement that pathogenic properties of rare variants has been named “diag- these screens are based on processes already suspected to be nostic strategy” and considered the third main approach to involved in the disease, i.e., tau toxicity and APP metabolism study human diseases using fly models, in addition to forward in AD. This is efficient and useful to identify new actors of and reverse genetics [18]. these processes and these actors may point toward a new Table 1 Positive genes in high-throughput functional screening of AD GWAS-defined loci Gene GWAS locus Functions Hit in high-throughput screening BIN1 BIN1 Nucleocytoplasmic adaptor protein involved in endocytosis Modifier of tau (2N4R) toxicity and membrane recycling, cytoskeleton regulation, DNA in Drosophila eye [28] repair, cell cycle progression, and cell death [41] CASS4 CASS4 Member of the CASS scaffolding protein localized at focal Modifier of tau (2N4R) toxicity adhesions, regulates cell spreading and motility [42] in Drosophila eye [28] CD2AP CD2AP Scaffolding protein involved in the regulation of membrane Modifier of tau (0N4R V337M) receptor endocytosis and signaling, actin cytoskeleton toxicity in Drosophila eye [26] organization, endosomal vesicular trafficking, cell adhesion, and cytokinesis [43–48] CELF1 CELF1 CUGBP Elav-like family member 1, role in RNA processing Modifier of tau (0N4R V337M) (splicing and mRNA stability mainly), role in myotonic toxicity in Drosophila eye [26] dystrophy [49] EPHA1 EPHA1 Founding member of the Eph family of tyrosine kinase receptor, Modifier of tau (2N4R) toxicity Interaction with integrin-like kinase and regulation of cell in Drosophila eye [28] morphology and motility through the ILK-RhoA-ROCK pathway, role of ephrin/EphR in synapse development and plasticity [50–52] FERMT2 FERMT2 Focal adhesion protein involved in integrin activation [53] Modifier of tau (0N4R V337M) toxicity in Drosophila eye [26] modifier of APP metabolism by HCS [11] MADD CELF1 Rab3/Rab27 guanine nucleotide exchange factor, role in synaptic Modifier of tau (2N4R) toxicity vesicle trafficking; interaction with TNF receptor, role in cell in Drosophila eye [28] death/survival signaling [54, 55] PTK2B PTK2B Member of the focal adhesion kinase (FAK) family of protein Modifier of tau (2N4R) toxicity tyrosine kinase. Role in signal transduction. Activated by in Drosophila eye [28] neuronal depolarization, Ca2+, and stressful conditions, role in neurite outgrowth, in synaptic plasticity, in neuronal survival, in astrocyte mobility [56–61] Curr Genet Med Rep 4. Sin O, Michels H, Nollen E a a. Genetic screens in Caenorhabditis pathway like the cell adhesion pathway (as observed in the elegans models for neurodegenerative diseases. Biochim Biophys screening we performed using tau toxicity in Drosophila eyes Acta Mol Basis Dis Elsevier BV. 2014;1842:1951–9. [28]). However, one can argue that the identification of novel 5. Mukherjee S, Russell JC, Carr DT, Burgess JD, Allen M, Serie DJ, and unexpected pathways and processes will be only achieved et al. Systems biology approach to late-onset Alzheimer’sdisease genome-wide association study identifies novel candidate genes by developing models based on novel phenotypes. For exam- validated using brain expression data and Caenorhabditis elegans ple, with the identification of several genes involved in mi- experiments. Alzheimers Dement. 2017;13:1133–42. croglia [3, 9, 62], microglia activation will be likely a readout 6. Bali J, Gheinani AH, Zurbriggen S, Rajendran L. Role of genes of high interest as already recently shown [63]. The advent of linked to sporadic Alzheimer’s disease risk in the production of beta-amyloid peptides. Proc Natl Acad Sci. 2012;109:15307–11. the DNA editing tools and iPSCs will enable the development 7. Majercak J, Ray WJ, Espeseth A, Simon A, Shi X-P, Wolffe C, et al. of models always closer to the disease characteristics. In con- LRRTM3 promotes processing of amyloid-precursor protein by clusion, even though it is essential to bear in mind biases and BACE1 and is a positional candidate gene for late-onset limitations, the systematic development of these methodolo- Alzheimer’s disease. Proc Natl Acad Sci U S A. 2006;103:17967–72. gies in a variety of different models should enable us to (i) 8. Camargo LM, Zhang XD, Loerch P, Caceres RM, Marine SD, Uva P, et al. Pathway-based analysis of genome-wide siRNA screens probe specific pathological events, (ii) build ever more com- reveals the regulatory landscape of app processing. PLoS One. plete databases, (iii) develop complex approaches based on 2015;10:e0115369. systems biology, (iv) systematically assess the potential roles 9. Lambert J-C, Ibrahim-Verbaas CA, Harold D, Naj AC, Sims R, of new genes, and (v) define new therapeutic targets and Bellenguez C, et al. Meta-analysis of 74,046 individuals identifies treatments. 11 new susceptibility loci for Alzheimer’s disease. Nat Genet. 2013;45:1452–8. 10. König R, Chiang C, Tu BP, Yan SF, DeJesus PD, Romero A, et al. Funding Information This work was funded by France Alzheimer asso- A probability-based approach for the analysis of large-scale RNAi ciation (#328 and #350), the Alzheimer’s association (BFG-14-318355), screens. Nat Methods. 2007;4:847–9. the Institute Pasteur de Lille, and the Nord-Pas de Calais Regional Council. This work was also funded by the French National Foundation 11. Chapuis J, Flaig A, Grenier-Boley B, Eysert F, Pottiez V, Deloison on Alzheimer’s disease and related disorders, the Lille Métropole G, et al. Genome-wide, high-content siRNA screening identifies the Communauté Urbaine council, and the French government’s LABEX Alzheimer’s genetic risk factor FERMT2 as a major modulator of DISTALZ program (development of innovative strategies for a transdis- APP metabolism. Acta Neuropathol. 2017;133:955–66. ciplinary approach to Alzheimer’sdisease). 12. Sannerud R, Declerck I, Peric A, Raemaekers T, Menendez G, Zhou L, et al. ADP ribosylation factor 6 (ARF6) controls amyloid precursor protein (APP) processing by mediating the endosomal Compliance with Ethical Standards sorting of BACE1. Proc Natl Acad Sci. 2011;108:E559–68. 13. Guo JL, Buist A, Soares A, Callaerts K, Calafate S, Stevenaert F, Conflict of Interest All authors declare that they have no conflict of et al. The dynamics and turnover of tau aggregates in cultured cells: interest. insights into therapies for tauopathies. J Biol Chem. 2016;291: 13175–93. Human and Animal Rights and Informed Consent This article does not 14. Foglieni C, Papin S, Salvadè A, Afroz T, Pinton S, Pedrioli G, et al. contain any studies with human or animal subjects performed by any of Split GFP technologies to structurally characterize and quantify the authors. functional biomolecular interactions of FTD-related proteins. Sci Rep. 2017;7:14013. Open Access This article is distributed under the terms of the Creative 15. Ivatt RM, Sanchez-Martinez A, Godena VK, Brown S, Ziviani E, Commons Attribution 4.0 International License (http:// Whitworth AJ. Genome-wide RNAi screen identifies the Parkinson creativecommons.org/licenses/by/4.0/), which permits unrestricted use, disease GWAS risk locus SREBF1 as a regulator of mitophagy. distribution, and reproduction in any medium, provided you give appro- Proc Natl Acad Sci. 2014;111:8494–9. priate credit to the original author(s) and the source, provide a link to the 16. Do CB, Tung JY, Dorfman E, Kiefer AK, Drabant EM, Francke U, Creative Commons license, and indicate if changes were made. Mountain JL, Goldman SM, Tanner CM, Langston JW, Wojcicki A, Eriksson N Web-based genome-wide association study identifies two novel loci and a substantial genetic component for Parkinson’s disease. Gibson G, editor. PLoS Genet 2011;7: e1002141. 17. Nieland TJF, Logan DJ, Saulnier J, Lam D, Johnson C, Root DE, References et al. High content image analysis identifies novel regulators of synaptogenesis in a high-throughput RNAi screen of primary neu- 1. Ulirsch JC, Nandakumar SK, Wang L, Giani FC, Zhang X, Rogov rons. PLoS One. 2014;9:e91744. P, et al. Systematic functional dissection of common genetic varia- 18. Ugur B, Chen K, Bellen HJ. Drosophila tools and assays for the tion affecting red blood cell traits. Cell. 2016;165:1530–45. study of human diseases. Dis Model Mech. 2016;9:235–44. 2. Jones L, Lambert JC, Wang LS, Choi SH, Harold D, Vedernikov A, 19. Bellen HJ, Tong C, Tsuda H. 100 years of Drosophila research and et al. Convergent genetic and expression data implicate immunity in its impact on vertebrate neuroscience: a history lesson for the future. Alzheimer’s disease. Alzheimers Dement. 2015;11:658–71. Nat Rev Neurosci. 2010;11:514–22. 3. Huang K-L, Marcora E, Pimenova AA, Di Narzo AF, Kapoor M, 20. Venken KJT, Simpson JH, Bellen HJ. Genetic manipulation of Jin SC, et al. A common haplotype lowers PU.1 expression in genes and cells in the nervous system of the fruit fly. Neuron. myeloid cells and delays onset of Alzheimer’s disease. Nat 2011;72:202–30. Neurosci. 2017;20:1052–61. Curr Genet Med Rep 21. Mohr SE, Hu Y, Kim K, Housden BE, Perrimon N. Resources for 39. Jakobsdottir J, van der Lee SJ, Bis JC, Chouraki V, Li-Kroeger D, Yamamoto S, et al. Rare functional variant in TM2D3 is associated functional genomics studies in Drosophila melanogaster. Genetics. 2014;197:1–18. with late-onset Alzheimer’s disease. Haines JL, editor. PLoS Genet 2016;12:e1006327. 22. Brandt R, Gergou A, Wacker I, Fath T, Hutter H. A Caenorhabditis 40. Luo X, Rosenfeld JA, Yamamoto S, Harel T, Zuo Z, Hall M, et al. elegans model of tau hyperphosphorylation: induction of develop- Members of the UDN Clinically severe CACNA1A alleles affect mental defects by transgenic overexpression of Alzheimer’s synaptic function and neurodegeneration differentially. Lu B, editor disease-like modified tau. Neurobiol Aging. 2009;30:22–33. PLOS Genet 2017;13:e1006905. 23. McGuire SE, Le PT, Osborn AJ, Matsumoto K, Davis RL. 41. Prokic I, Cowling BS, Laporte J. Amphiphysin 2 (BIN1) in phys- Spatiotemporal rescue of memory dysfunction in drosophila. iology and diseases. J Mol Med. 2014;92:453–63. Science (80-. ). 2003;302:1765–8. 42. Singh MK, Dadke D, Nicolas E, Serebriiskii IG, Apostolou S, 24. Burcin MM, Schiedner G, Kochanek S, Tsai SY, O’Malley BW. Canutescu A, et al. A novel Cas family member, HEPL, regulates Adenovirus-mediated regulable target gene expression in vivo. Proc FAK and cell spreading. Mol Biol Cell. 2008;19:1627–36. Natl Acad Sci U S A. 1999;96:355–60. 43. Dustin ML, Olszowy MW, Holdorf AD, Li J, Bromley S, Desai N, 25. Osterwalder T, Yoon KS, White BH, Keshishian H. A conditional et al. A novel adaptor protein orchestrates receptor patterning and tissue-specific transgene expression system using inducible GAL4. cytoskeletal polarity in T-cell contacts. Cell. 1998;94:667–77. Proc Natl Acad Sci. 2001;98:12596–601. 44. Cormont M, Metón I, Mari M, Monzo P, Keslair F, Gaskin C, et al. 26. Shulman JM, Imboywa S, Giagtzoglou N, Powers MP, Hu Y, CD2AP/CMS regulates endosome morphology and traffic to the Devenport D, et al. Functional screening in Drosophila identifies degradative pathway through its interaction with Rab4 and c-Cbl. Alzheimer’s disease susceptibility genes and implicates tau- Traffic. 2003;4:97–112. mediated mechanisms. Hum Mol Genet. 2014;23:870–7. 45. Johnson RI, Seppa MJ, Cagan RL. The Drosophila CD2AP/CIN85 27. Shulman JM, Chipendo P, Chibnik LB, Aubin C, Tran D, Keenan BT, orthologue Cindr regulates junctions and cytoskeleton dynamics et al. Functional screening of Alzheimer pathology genome-wide asso- during tissue patterning. J Cell Biol. 2008;180:1191–204. ciation signals in Drosophila. Am J Hum Genet. 2011;88:232–8. 46. Lynch DK, Winata SC, Lyons RJ, Hughes WE, Lehrbach GM, 28. Dourlen P, Fernandez-Gomez FJ, Dupont C, Grenier-Boley B, Wasinger V, et al. A Cortactin-CD2-associated protein (CD2AP) Bellenguez C, Obriot H, et al. Functional screening of Alzheimer complex provides a novel link between epidermal growth factor risk loci identifies PTK2B as an in vivo modulator and early marker receptor endocytosis and the actin cytoskeleton. J Biol Chem. of tau pathology. Mol Psychiatry. 2017;22:874–83. 2003;278:21805–13. 29. Hu Y, Flockhart I, Vinayagam A, Bergwitz C, Berger B, Perrimon N, 47. Kobayashi S, Sawano A, Nojima Y, Shibuya M, Maru Y. The c- et al. An integrative approach to ortholog prediction for disease- Cbl/CD2AP complex regulates VEGF-induced endocytosis and focused and other functional studies. BMC Bioinform. 2011;12:357. degradation of Flt-1 (VEGFR-1). FASEB J. 2004;18:929–31. 30. Moreau K, Fleming A, Imarisio S, Lopez Ramirez A, Mercer JL, 48. Haglund K, Nezis IP, Lemus D, Grabbe C, Wesche J, Liestøl K, Jimenez-Sanchez M, et al. PICALM modulates autophagy activity et al. Cindr interacts with anillin to control cytokinesis in and tau accumulation. Nat Commun. 2014;5:4998. Drosophila melanogaster. Curr Biol. 2010;20:944–50. 31. Song L, He Y, Ou J, Zhao Y, Li R, Cheng J, et al. Auxilin underlies 49. Gallo J-M, Spickett C. The role of CELF proteins in neurological progressive locomotor deficits and dopaminergic neuron loss in a disorders. RNA Biol. 2010;7:474–9. Drosophila model of Parkinson’s disease. Cell Rep. 2017;18:1132–43. 50. Hirai H, Maru Y, Hagiwara K, Nishida J, Takaku F. A novel puta- 32. MacLeod DA, Rhinn H, Kuwahara T, Zolin A, Di Paolo G, tive tyrosine kinase receptor encoded by the eph gene. Science. McCabe BD, et al. RAB7L1 interacts with LRRK2 to modify 1987;238:1717–20. intraneuronal protein sorting and Parkinson’s disease risk. 51. Yamazaki T, Masuda J, Omori T, Usui R, Akiyama H, Maru Y. Neuron. 2013;77:425–39. EphA1 interacts with integrin-linked kinase and regulates cell mor- 33. Stefansson H, Rye DB, Hicks A, Petursson H, Ingason A, phology and motility. J Cell Sci. 2009;122:243–55. Thorgeirsson TE, et al. A genetic risk factor for periodic limb 52. Lai K-O, Ip NY. Synapse development and plasticity: roles of ephrin/ movements in sleep. N Engl J Med. 2007;357:639–47. Eph receptor signaling. Curr Opin Neurobiol. 2009;19:275–83. 34. Freeman A, Pranski E, Miller RD, Radmard S, Bernhard D, Jinnah 53. Lai-Cheong JE, Parsons M, McGrath JA. The role of kindlins in cell HA, et al. Sleep fragmentation and motor restlessness in a biology and relevance to human disease. Int J Biochem Cell Biol. Drosophila model of restless legs syndrome. Curr Biol. 2012;22: 2010;42:595–603. 1142–8. 54. Marat AL, Dokainish H, McPherson PS. DENN domain proteins: 35. Jia T, Macare C, Desrivières S, Gonzalez DA, Tao C, Ji X, et al. regulators of Rab GTPases. J Biol Chem. 2011;286:13791–800. Neural basis of reward anticipation and its genetic determinants. 55. Miyoshi J, Takai Y. Dual role of DENN/MADD (Rab3GEP) in Proc Natl Acad Sci. 2016;113:3879–84. neurotransmission and neuroprotection. Trends Mol Med. 36. Schumann G, Coin LJ, Lourdusamy A, Charoen P, Berger KH, 2004;10:476–80. Stacey D, et al. Genome-wide association and genetic functional 56. Girault JA, Costa A, Derkinderen P, Studler JM, Toutant M. FAK and studies identify autism susceptibility candidate 2 gene (AUTS2) in PYK2/CAKbeta in the nervous system: a link between neuronal activ- the regulation of alcohol consumption. Proc Natl Acad Sci U S A. ity, plasticity and survival? Trends Neurosci. 1999;22:257–63. 2011;108:7119–24. 57. Ivankovic-Dikic I, Grönroos E, Blaukat A, Barth BU, Dikic I. Pyk2 37. Vanden Broeck L, Kleinberger G, Chapuis J, Gistelinck M, and FAK regulate neurite outgrowth induced by growth factors and Amouyel P, Van Broeckhoven C, et al. Functional complementa- integrins. Nat Cell Biol. 2000;2:574–81. tion in Drosophila to predict the pathogenicity of TARDBP vari- 58. Huang Y, Lu W, Ali DW, Pelkey KA, Pitcher GM, Lu YM, et al. ants: evidence for a loss-of-function mechanism. Neurobiol Aging. CAKbeta/Pyk2 kinase is a signaling link for induction of long-term 2015;36:1121–9. potentiation in CA1 hippocampus. Neuron. 2001;29:485–96. 38. Chapuis J, Hansmannel F, Gistelinck M, Mounier A, Van 59. Hsin H, Kim MJ, Wang C-F, Sheng M. Proline-rich tyrosine kinase Cauwenberghe C, Kolen KV, et al. GERAD consortium. 2 regulates hippocampal long-term depression. J Neurosci. Increased expression of BIN1 mediates Alzheimer genetic risk by 2010;30:11983–93. modulating tau pathology. Mol Psychiatry. 2013 Nov;18(11):1225- 60. Strappazzon F, Torch S, Trioulier Y, Blot B, Sadoul R, Verna J-M. 34. Survival response-linked Pyk2 activation during potassium Curr Genet Med Rep depletion-induced apoptosis of cerebellar granule neurons. Mol TREM2 implicate microglial-mediated innate immunity in Alzheimer’s disease. Nat Genet. 2017;49:1373–84. Cell Neurosci. 2007;34:355–65. 61. Giralt A, Coura R, Girault J-A. Pyk2 is essential for astrocytes 63. Ryan KJ, White CC, Patel K, Xu J, Olah M, Replogle JM, et al. A mobility following brain lesion. Glia. 2016;64:620–34. human microglia-like cellular model for assessing the effects of 62. Sims R, van der Lee SJ, Naj AC, Bellenguez C, Badarinarayan N, neurodegenerative disease gene variants. Sci Transl Med. Jakobsdottir J, et al. Rare coding variants in PLCG2, ABI3, and 2017;9(421):eaai7635. http://www.deepdyve.com/assets/images/DeepDyve-Logo-lg.png Current Genetic Medicine Reports Springer Journals

Using High-Throughput Animal or Cell-Based Models to Functionally Characterize GWAS Signals

Free
9 pages
Loading next page...
 
/lp/springer_journal/using-high-throughput-animal-or-cell-based-models-to-functionally-fURMpgGY1R
Publisher
Springer US
Copyright
Copyright © 2018 by The Author(s)
Subject
Medicine & Public Health; Internal Medicine
eISSN
2167-4876
D.O.I.
10.1007/s40142-018-0141-1
Publisher site
See Article on Publisher Site

Abstract

Purpose of Review The advent of genome-wide association studies (GWASs) constituted a breakthrough in our understanding of the genetic architecture of multifactorial diseases. For Alzheimer’s disease (AD), more than 20 risk loci have been identified. However, we are now facing three new challenges: (i) identifying the functional SNP or SNPs in each locus, (ii) identifying the causal gene(s) in each locus, and (iii) understanding these genes’ contribution to pathogenesis. Recent Findings To address these issues and thus functionally characterize GWAS signals, a number of high-throughput strat- egies have been implemented in cell-based and whole-animal models. Here, we review high-throughput screening, high-content screening, and the use of the Drosophila model (primarily with reference to AD). Summary We describe how these strategies have been successfully used to functionally characterize the genes in GWAS-defined risk loci. In the future, these strategies should help to translate GWAS data into knowledge and treatments. . . . . . Keywords HCS HTS Drosophila Screen GWAS Alzheimer Introduction disequilibrium patterns involving the sentinel SNP may make it difficult or even impossible to determine which gene is Genome-wide association studies (GWASs) determine groups responsible for the observed association. Hence, identifying of single nucleotide polymorphisms (SNPs) in linkage dis- the causative gene and one or more functional SNP are major equilibrium and which are associated with a particular disease, challenges in the GWAS field. For instance, new technologies trait, or phenotype. The most significantly associated SNP is have been developed to tackle this limitation with the obser- usually not the causative SNP, and (by convention) the signal vation that 85–95% of the GWAS-associated SNPs are signif- is assigned to the closest gene. However, the presence of sev- icantly enriched at cell-type-specific regulatory regions: mas- eral genes in a GWAS locus and complex linkage sively parallel reporter assay (MPRA) allowed to screen thou- sands of variants in order to assess their impact on transcrip- This article is part of the Topical Collection on Neurogenetics and tional activity [1]. A further difficulty (even when the risk Psychiatric Genetics gene is known) relates to the determination of how the caus- ative genes are functionally involved in the disease process. * Pierre Dourlen Indeed, it can be difficult to establish a causal link on the basis pierre.dourlen@pasteur–lille.fr of the literature data alone. To take account of these limitations, gene enrichment path- Julien Chapuis way analyses have been developed from GWAS dataset. The julien.chapuis@pasteur-lille.fr main idea is that genes associated with the disease risk will be Jean-Charles Lambert over-represented in specific pathways involved in the disease jean-charles.lambert@pasteur-lille.fr process. In Alzheimer’s disease (AD), these analyses have pointed to the immune response, the regulation of endocyto- INSERM U1167, RID-AGE-Risk Factors and Molecular Determinants of Aging-Related Diseases, Lille, France sis, cholesterol transport, and protein ubiquitination [2]. However, it must be borne in mind that this type of analysis Institut Pasteur de Lille, Lille, France is subject to major limitations (in addition to methodological University Lille, U1167-Excellence Laboratory LabEx DISTALZ, issues): for example, the defined canonical pathways are far Lille, France Curr Genet Med Rep from complete, and many numerous genes have pleiotropic High-Throughput Screening in Cell-Based functions and are thus nominated in many different pathways. Models It is also possible that a gene may have an unknown function with relevance to the pathophysiological context. According to the amyloid cascade hypothesis (one of the main Given this background, it appears that gene enrichment hypotheses in AD, in which the overproduction of amyloid pathway approaches are not able to optimally exploit the ge- (Aβ) peptides leads to neurotoxicity and then neuronal death), netic data generated by high-throughput genomic strategies. It one expects at least some of the GWAS-defined genes to con- is now possible to develop in silico methodologies by com- trol the production of the Aβ peptides that accumulate in the bining numerous datasets with the objective to define both AD brain. For decades, the impact of AD candidate genes on functional variants and genes [3]. However, there is also a Aβ secretion/production was evaluated one by one. Given the need for alternative, powerful approaches for empirically test- large number of AD candidate highlighted by GWASs, these ing multiple GWAS genes in cell-based or animal models. approaches no longer appear to be appropriate. Indeed, the Here, we review high-throughput functional screening strate- assessment of a putative AD gene’simpacton Aβ peptides gies that use in vitro cell-based models and the in vivo production requires cell-based models that are suitable for use Drosophila model for GWAS-defined hits (primarily in AD in fast, large-scale screening. but also for some other neurological diseases) (Fig. 1). We will One of the first attempts to systematically measure the im- not cover Caenorhabditis elegans—another invertebrate ani- pact of GWAS-defined genes involved the RNAi-mediated mal model [4] that has also been used [5]. silencing of 24 late-onset AD genes in Hela cells stably Fig. 1 Schema describing the main steps to set up systematic screenings for GWAS-genomic data (in blue, specific to HCS/ HTS; in green, specific to Drosophila) Curr Genet Med Rep sw over-expressing APP [6]. The levels of APPs-β and Aβ inhibitor was associated with a specific increase in YFP fluo- 1–38/ secreted into the supernatant were then measured. Bali rescence intensity (corresponding to the APP’s C-terminal 40/42 et al. proposed that late-onset AD genes did not specifically fragment). After customization for automatic image process- alter the Aβ42/40 ratio but probably contributed to AD ing, we screened a genome-wide bank of 18,107 human through a distinct mechanism. However, this study was limit- siRNAs. In total, 832 hits were selected as being likely to have ed by the number of genes tested and the number of an impact on APP metabolism—including 8 genes associated endophenotypes measured. Furthermore, the number of loci with the risk of late-onset AD in the reference GWAS meta- detected by GWAS is also increasing, meaning that this type analysis. These data suggested that the 8 genes are involved in of low-scale study is of less interest. AD process via the regulation of APP metabolism. Technological progress has enabled the large-scale automa- For HTS/HCS data, it is important to bear in mind that the tion of biological experiments. These techniques are particu- choice of statistical methods used to define the initial hit list is larly appropriate for generating large amounts of data in cell- critical; this will determine the relevant secondary analyses based models, and mark the entry of cell biology into the age needed to narrow down the hits and thus investigate biological of high-throughput methodologies (as has already happened mechanisms. To assess the relevance of these genes with re- for genetics, epigenomics, and transcriptomics, for example). gard to Aβ peptide levels, we cross-checked our HCS data Rather than focusing on subsets of genes, progress in genome against an association study of SNPs and CSF Aβ42 peptide annotation enables the systematic performance of high- levels in a large sample of patients (n = 2950). Only SNPs throughput screening (HTS)—for instance by designing within FERMT2 were associated with low Aβ42 peptide RNAi constructs to test the effects of gene silencing on bio- levels; this highlighted FERMT2’s potential role in the AD logical phenotypes. process via the modulation of APP metabolism and Aβ pep- This type of systematic analysis was performed with tide generation. Lastly, the impact of FERMT2 expression on 14,603 siRNA pools in HEK-293 cells stably expressing a APP metabolism was validated in several cell-based models, mutant form of APP (NFEV) designed to enhance Aβ pro- including a primary neuronal culture (PNC) endogenously duction. Conditioned media were used to quantify Aβ40, expressing both APP and FERMT2. Therefore, accurate Aβ42, sAPPα,and sAPPβ levels [7, 8]. A “regulatory land- choice of statistical methods, crosschecking with other scape of APP processing” was generated, in order to identify screens, and validation in low-throughput models enabled most of the pathways involved in the regulation of APP me- identifying FERMT2 as a risk factor modulating APP metab- tabolism. Interestingly, some of the identified pathways con- olism. The low number of final positive hits suggests that the tain genes (CLU, BIN1, CR1, PICALM, TREM2, SORL1, AD genetic risk factors in the other loci may be involved in MEF2C, DSG2, EPH1A) linked to AD in the reference different processes than the APP metabolism. This is a limita- tion of the high-throughput methodologies. They are designed GWAS meta-analysis [9]. It is noteworthy that only the aggre- gate effect of genes in a pathway were considered as real with respect to a defined phenotype. Most of the current effects, due to the high false-positive rate and the low level models are based on neuronal dysfunctions involved in AD of reproducibility after siRNA library screening [10]. processes through APP metabolism and Aβ production. By As mentioned above, assays of Aβ released in the cell using these approaches, it is not possible for example to in- culture medium have been used to identify modulators of vestigate APOE or TREM2 functions which are thought to be APP metabolism. However, the underlying mechanisms con- mainly involved in Aβ clearance. trolling APP maturation and trafficking cannot be character- Other cell-based models compatible with HCS analysis ized solely by studying the amount of Aβ present in culture have been generated to assess other phenotypes and could media. Thus, it was necessary to move to the high-content be used to test multiple GWAS genes—for instance QBI- screening (HCS) of multidimensional phenotypes; this con- 293 cells with Dox-regulated inducible expression of human sists of the cell-based quantification of several processes si- tau carrying the P301L mutation, and a GFP tag attached to multaneously; the combination of cell-based and imaging visualize tau inclusions [13]. More recently, GFP bimolecular methodologies provides a more detailed representation of the and trimolecular fluorescence complementation (biFC and cell’s response to various perturbations than HTS does. In line triFC) has been used to study the localization and mechanisms with this approach, we developed a rapid, cell-based, HCS of protein multimers (tau and TDP-43) in the context of neu- assay for the intracellular APP fragments in HEK293 cells rodegeneration [14]. These technologies may enable new 695WT stably over-expressing a mCherry-APP -YFP construct functional assays of genetic factors in neurodegenerative dis- 695WT [11]. The modified APP protein is known to be metab- eases through siRNA screening. 695WT olized in the same way as APP [12]. Our model enabled Some non-mammalian cell-based models like the the detection of specific APP products differentially tagged Drosophila S2R+ cell line have also been used to perform with mCherry or YFP (for the N- and C-terminal fragments, unbiased, genome-wide screening of RNAi. Drosophila cell lines usefully have a low level of gene redundancy. RNA respectively). For instance, treatment with γ-secretase Curr Genet Med Rep interference is also straight forward in these cells as dsRNA Drosophila in Genetic Screening are added in the medium without transfecting reagent. This approach was exemplified by a screen for regulators of the Moving on from in vitro cell-based models, unbiased, high- translocation of Parkin (a protein whose mutation causes throughput screens of GWAS-defined loci can be performed inherited recessive Parkinsonism) to mitochondria [15]. The in a more integrated biological in vivo context by using rele- genome-wide screen identified 60 genes, which were further vant read-outs in animal models of the target disease. narrowed down to 20 candidate genes with a conserved effect However, there are many obvious constraints associated with on both Parkin translocation and mitophagy in HeLa cells. the use of animals and the inability to systematically screen The top hits belonged to the sterol regulatory element binding hundreds of genes in a murine model of AD. This approach is protein (SREBP)-lipogenesis pathway, including the master nevertheless possible in small invertebrate models, such as the regulator of lipid synthesis sterol regulatory element binding fruit fly Drosophila melanogaster. This fly shows good gene transcription factor 1 (SREBF1). Interestingly, the latter is a conservation with humans and only a little gene redundan- GWAS-detected risk factor for sporadic PD suggesting a cy—enabling easier detection of the effects of loss of function. mechanistic link between inherited recessive PD and sporadic Two thirds of human disease-associated genes are estimated to PD—a long debated question [16]. have a functional homolog in Drosophila [18]. In contrast, In order to explore many potential functions of the gene regulatory regions are much less well conserved, there- genetic risk factors, it is important to highlight the grow- fore it is difficult to use Drosophila to identify a causative non- ing need to measure different phenotypes in different cell coding SNP if it affects gene expression. Drosophila has a types. Indeed, gene expression that are cell-specific, like short life cycle, with a new generation produced every 10 days TREM2 (preferentially expressed in microglia), requests at 25 °C. When the latter advantage is combined with high relevant and adapted cellular model. For this purpose, numbers of progeny and the low cost of housing, it is easy to iPSC will likely play an important role in the future to produce the large number of individuals required for optimal generate multiple cell types like neurons, astrocytes, mi- genetic studies. Furthermore, many genetic tools have been croglia, or endothelial cells susceptible to be involved in developed in more than a century of use in forward and re- pathophysiological mechanisms. However, compared with verse genetics. One can therefore modulate more or less any easy-to-transfect standard cell lines, HCS approach using genes, anywhere and at any point in the life cycle [19, 20]. these iPSCs-derived cells will request viral transduction to Mutants are available for almost all the genes in the modify gene expression and will thus lead to screen Drosophila genome [21]. Collections of RNA interference smaller number of genes in these different models for and overexpression constructs have also been generated at the moment. A lentiviral RNA interference library of the genome-wide scale. Expression of these constructs is usu- 597 shRNAs was already used to screen for novel regu- ally based on the Gal4/UAS system [22]. Modifications can lators of synapse formation [17]. In addition, the develop- make the system conditional, with expression controlled by ment of the CRISPR-CASS9 technology and sgRNA li- changing the temperature (Gal80ts, temporal and regional braries should enable large-scale DNA editing and in- gene expression targeting—TARGET—system, [23]) or by crease even more the power of HCS. adding a drug to the flies’ food (the GeneSwitch system; In conclusion, there is no doubt that HTS/HCS methodol- [24, 25]). Thus, Drosophila offers many tools for tightly con- ogies are very powerful ways of potentially assigning patho- trolling gene expression. Furthermore, Drosophila has a physiological functions to GWAS-defined genes. It is likely wealth of external phenotypes (the size, color, and shape of that their use will increase over the coming years, although the eye or of the wing, etc.) that are easily scorable under a this requires major investments in dedicated, automated tech- dissecting microscope. This is essential for quickly screening nology platforms. It is important to bear the following key the effect of gene modulation. Lastly, Drosophila has higher issues in mind: (i) HTS/HCS models face many challenges, cognitive functions (like memory and sleep) for which behav- including the identification of the best experimental system ioral and electrophysiological assays are available. These fea- and the development of robust, reproducible assays; (ii) tures make Drosophila an ideal model for neurogenetics. HTS/HCS models are often highly sensitive, and can generate a large number of unspecific, biologically irrelevant re- sponses; (iii) in view of the number of analyses to be per- Drosophila for Screening GWAS-Defined formed, HTS/HCS approaches must balance the risk of ob- Genes serving significant results by chance against the risk of rejecting biologically valid hypotheses on purely statistical In this context, Drosophila has beenusedtoscreenGWAS grounds; and (iv) as with all data generated in high- candidate genes associated with AD [26–28]. Starting from throughput assays, the most relevant observations need to be the reference GWAS meta-analysis [9], we screened for the replicated and validated in other, unrelated cell-based models. AD candidate genes that modified tau neurotoxicity [28]. Curr Genet Med Rep Based on the regional association plots, we identified a total of regulate autophagy and tau degradation [30]. PD is character- 148 genes in the 19 AD-associated loci. According to the ized clinically by motor and non-motor symptoms and histo- Drosophila RNAi Screening Center Integrative Ortholog logically by dopaminergic neuronal loss and the formation of Prediction Tool (http://www.flyrnai.org/diopt;[29]), 54 of α-synuclein Lewy bodies. The gene coding for cyclin-G- the human genes had a total of 74 Drosophila orthologs at associated kinase (GAK) was identified in a GWAS as a PD 13 loci. Two hundred seventy-eight RNAi lines from five susceptibility gene. To analyze its PD-related functions, Song collections (the Japan NIG collection, the Vienna GD and et al. turned to Drosophila and studied loss of function of the KK collections, and the Harvard TRiP attP2 and attP40 col- Drosophila GAK ortholog (aux)[31]. Knockdown of aux re- lections) and 17 mutant and overexpression constructs were sulted in age-dependent locomotor impairment (in a climbing selected to modulate gene expression (giving four constructs assay), a shorter lifespan, and progressive dopaminergic neu- per gene, on average). The fly eye was used to assess tau ronal loss—as seen in α-Syn overexpression. Downregulation neurotoxicity. Expression of human tau (2N4R isoform) dur- of aux also enhanced α-synuclein-induced dopaminergic neu- ing Drosophila eye development results in small, rough eyes. ronal death, and sensitized flies to the environmental toxin Each of the 295 constructs was tested singly to see whether the paraquat. All these phenotypes represent a broad spectrum size of the tau-expressing eye was modulated. After quantifi- of parkinsonian-like symptoms—supporting the idea that cation of eye size, only genes for which at least two positive GAK/aux is a PD risk factor with a potential role in PD path- constructs from different collections had the same effect were ogenesis. Lifespan and the loss of dopaminergic neurons in considered to be positive hits. By applying these criteria, we Drosophila were also used to identify an interaction between confirmed that Amph (an ortholog of BIN1) modulated tau the PD risk factors PARK16 and LRRK2 [32]. Another ex- neurotoxicity (Chapuis et al. 2013). We also identified four ample of the use of Drosophila to assess disease-like functions new genes: p130CAS, Eph, Focal adhesion kinase (Fak), and is illustrated by studies of the BTBD9 gene. The latter had Rab3-GEF, which are respectively orthologs of CASS4, been identified in a GWAS as a risk factor for restless legs EPHA1, PTK2B,and MADD. Interestingly, three of the five syndrome (RLS), a sensorimotor neurological disorder char- hits (Fak, p130CAS, and Eph) are directly or indirectly in- acterizedby(i) a compellingurgetomoveduringperiods of volved in the cell adhesion pathway. Another Drosophila rest, (ii) relief with movement, (iii) involuntary movements in screen of tau neurotoxicity modifiers also identified genes sleep, and (iv) fragmented sleep [33]. Functional analysis of involved in this pathway [26]. This is one of the advantages BTBD9 was first performed in Drosophila [34]. Loss of the of unbiased systematic screening in Drosophila; the identifi- Drosophila homolog CG1826 (dBTBD9) markedly disrupted cation of new genes that modulate a process (tau neurotoxic- fly sleep, with concomitant increases in waking and motor ity, in this case) and point toward a new pathway (the cell activity. This study also showed that dBTBD9 regulates brain dopamine levels, which is known to correlate with RLS ex- adhesion pathway, in this case) as being a potentially impor- tant pathway for AD pathogenesis. Of course, one can argue pression in humans [34]. Even for behavior like hyperactivity that the tau-associated fly eye phenotype reflects only one or alcohol consumption, Drosophila has provided functional aspect of AD pathogenesis and the results does not explain evidence to complement the sometimes rather correlative ge- how the hits are involved in the pathogenesis. Further work in netic data. In order to better understand dysfunctional reward Drosophila and mammalian models is required to address processing and the discovery of an association between hy- these questions. In our study, we were able to validate the peractivity and the VPS4A gene in a GWAS in adolescents, a genetic interaction between tau and Fak in a cell adhesion- causal role of the sole Vps4 ortholog for hyperactivity was related fly wing readout. By assessing tau phosphorylation validated in Drosophila [35]. Similarly, after the identification and testing a catalytically mutant form of Fak, we could dis- of the association of the AUTS2 gene with alcohol consump- favor a change in tau phosphorylation (classically considered tion at a genome-wide level of significance, functional evi- as one of the culprits in AD pathogenesis) as the cause of the dence was obtained in Drosophila; downregulation of the sole modulation of tau neurotoxicity by Fak in Drosophila. A tau- AUTS2 homolog (tay) led to a lower alcohol sensitivity [36]. PTK2B interaction was confirmed in the brains of a tau mouse With the increasing use of nucleic acid sequencing in re- model and human patients, since we observed an abnormal search and clinical practice, many rare coding variants with somatic accumulation of PTK2B with the appearance of tau unknown functional consequences—called variants of un- oligomers and neurofibrillary tangles [28]. known or uncertain significance, VUS—are being identified. Along with unbiased medium- to high-throughput in vivo Drosophila happens to be a very useful model for addressing screening with easily scorable read-outs, Drosophila is also this issue. The rationale consists in performing functional used to address more specific questions about the functionality transcomplementation experiments in Drosophila,as success- of a gene with respect to the clinical and biological features of fully illustrated for VUS in the TARDBP gene [37]. TARDBP human diseases. The AD risk factor PICALM has been studied encodes the TDP-43 protein, which forms cytoplasmic inclu- in Drosophila. Its ortholog in the fly (Lap ) was shown to sions in patients with the most frequent form of Curr Genet Med Rep frontotemporal dementia (FTD) and most forms of amyotro- Conclusion phic lateral sclerosis (ALS) and in 60% of patients with AD. It is a common disease-causing factor in FTD and ALS. In the context of neurodegenerative disorders in general and However, how TARDBP mutations cause neurodegeneration AD in particular, high-throughput technologies appear to be is not well known, especially with regard to their loss-of- very useful for characterizing the pathophysiological func- function or toxic gain-of-function properties. Null mutants of tions of GWAS-defined genes (Table 1). The next step will the Drosophila ortholog of TARDBP (TBPH) lose the neurons be likely to develop multidimensional high-throughput in the ventral nerve cord that secrete the neurohormone methods allowing to analyze at the same time both functional bursicon [38]. It was shown that expression of human variants and gene functions in accurate cellular types and TARDBP in TBPH-null Drosophila rescued the bursicon neu- models. However, it is important to keep in mind that the more rons, thus indicating functional transcomplementation [37]. the study is complex, the more the statistical analyses and This made Drosophila a platform for testing TARDBP muta- quality control are essential for the success of such sensitive tions. Expression of two typical ALS-causing mutations methods to avoid false-positive/negative results. Furthermore, (p.G287S and p.A315T) could not rescue neuronal loss as confirmation and validation in complementary low- efficiently as wild-type TARDBP did. One atypical variant throughput assays are systematically required. In addition, (p.D169G) could rescue, and another (p.A90V) could not. among practical issues, the high technicality of the methods These findings suggested a partial loss of function in can make them difficult to be mastered and cost effective in TARDBP mutations [37]. The same strategy was successful terms of equipments. for VUS in the TM2D3 and CACNA1A genes, in the context of Beyond these general comments, another limitation of the late-onset AD and ataxia [39, 40]. This method for assessing current high-throughput methods lies on the statement that pathogenic properties of rare variants has been named “diag- these screens are based on processes already suspected to be nostic strategy” and considered the third main approach to involved in the disease, i.e., tau toxicity and APP metabolism study human diseases using fly models, in addition to forward in AD. This is efficient and useful to identify new actors of and reverse genetics [18]. these processes and these actors may point toward a new Table 1 Positive genes in high-throughput functional screening of AD GWAS-defined loci Gene GWAS locus Functions Hit in high-throughput screening BIN1 BIN1 Nucleocytoplasmic adaptor protein involved in endocytosis Modifier of tau (2N4R) toxicity and membrane recycling, cytoskeleton regulation, DNA in Drosophila eye [28] repair, cell cycle progression, and cell death [41] CASS4 CASS4 Member of the CASS scaffolding protein localized at focal Modifier of tau (2N4R) toxicity adhesions, regulates cell spreading and motility [42] in Drosophila eye [28] CD2AP CD2AP Scaffolding protein involved in the regulation of membrane Modifier of tau (0N4R V337M) receptor endocytosis and signaling, actin cytoskeleton toxicity in Drosophila eye [26] organization, endosomal vesicular trafficking, cell adhesion, and cytokinesis [43–48] CELF1 CELF1 CUGBP Elav-like family member 1, role in RNA processing Modifier of tau (0N4R V337M) (splicing and mRNA stability mainly), role in myotonic toxicity in Drosophila eye [26] dystrophy [49] EPHA1 EPHA1 Founding member of the Eph family of tyrosine kinase receptor, Modifier of tau (2N4R) toxicity Interaction with integrin-like kinase and regulation of cell in Drosophila eye [28] morphology and motility through the ILK-RhoA-ROCK pathway, role of ephrin/EphR in synapse development and plasticity [50–52] FERMT2 FERMT2 Focal adhesion protein involved in integrin activation [53] Modifier of tau (0N4R V337M) toxicity in Drosophila eye [26] modifier of APP metabolism by HCS [11] MADD CELF1 Rab3/Rab27 guanine nucleotide exchange factor, role in synaptic Modifier of tau (2N4R) toxicity vesicle trafficking; interaction with TNF receptor, role in cell in Drosophila eye [28] death/survival signaling [54, 55] PTK2B PTK2B Member of the focal adhesion kinase (FAK) family of protein Modifier of tau (2N4R) toxicity tyrosine kinase. Role in signal transduction. Activated by in Drosophila eye [28] neuronal depolarization, Ca2+, and stressful conditions, role in neurite outgrowth, in synaptic plasticity, in neuronal survival, in astrocyte mobility [56–61] Curr Genet Med Rep 4. Sin O, Michels H, Nollen E a a. Genetic screens in Caenorhabditis pathway like the cell adhesion pathway (as observed in the elegans models for neurodegenerative diseases. Biochim Biophys screening we performed using tau toxicity in Drosophila eyes Acta Mol Basis Dis Elsevier BV. 2014;1842:1951–9. [28]). However, one can argue that the identification of novel 5. Mukherjee S, Russell JC, Carr DT, Burgess JD, Allen M, Serie DJ, and unexpected pathways and processes will be only achieved et al. Systems biology approach to late-onset Alzheimer’sdisease genome-wide association study identifies novel candidate genes by developing models based on novel phenotypes. For exam- validated using brain expression data and Caenorhabditis elegans ple, with the identification of several genes involved in mi- experiments. Alzheimers Dement. 2017;13:1133–42. croglia [3, 9, 62], microglia activation will be likely a readout 6. Bali J, Gheinani AH, Zurbriggen S, Rajendran L. Role of genes of high interest as already recently shown [63]. The advent of linked to sporadic Alzheimer’s disease risk in the production of beta-amyloid peptides. Proc Natl Acad Sci. 2012;109:15307–11. the DNA editing tools and iPSCs will enable the development 7. Majercak J, Ray WJ, Espeseth A, Simon A, Shi X-P, Wolffe C, et al. of models always closer to the disease characteristics. In con- LRRTM3 promotes processing of amyloid-precursor protein by clusion, even though it is essential to bear in mind biases and BACE1 and is a positional candidate gene for late-onset limitations, the systematic development of these methodolo- Alzheimer’s disease. Proc Natl Acad Sci U S A. 2006;103:17967–72. gies in a variety of different models should enable us to (i) 8. Camargo LM, Zhang XD, Loerch P, Caceres RM, Marine SD, Uva P, et al. Pathway-based analysis of genome-wide siRNA screens probe specific pathological events, (ii) build ever more com- reveals the regulatory landscape of app processing. PLoS One. plete databases, (iii) develop complex approaches based on 2015;10:e0115369. systems biology, (iv) systematically assess the potential roles 9. Lambert J-C, Ibrahim-Verbaas CA, Harold D, Naj AC, Sims R, of new genes, and (v) define new therapeutic targets and Bellenguez C, et al. Meta-analysis of 74,046 individuals identifies treatments. 11 new susceptibility loci for Alzheimer’s disease. Nat Genet. 2013;45:1452–8. 10. König R, Chiang C, Tu BP, Yan SF, DeJesus PD, Romero A, et al. Funding Information This work was funded by France Alzheimer asso- A probability-based approach for the analysis of large-scale RNAi ciation (#328 and #350), the Alzheimer’s association (BFG-14-318355), screens. Nat Methods. 2007;4:847–9. the Institute Pasteur de Lille, and the Nord-Pas de Calais Regional Council. This work was also funded by the French National Foundation 11. Chapuis J, Flaig A, Grenier-Boley B, Eysert F, Pottiez V, Deloison on Alzheimer’s disease and related disorders, the Lille Métropole G, et al. Genome-wide, high-content siRNA screening identifies the Communauté Urbaine council, and the French government’s LABEX Alzheimer’s genetic risk factor FERMT2 as a major modulator of DISTALZ program (development of innovative strategies for a transdis- APP metabolism. Acta Neuropathol. 2017;133:955–66. ciplinary approach to Alzheimer’sdisease). 12. Sannerud R, Declerck I, Peric A, Raemaekers T, Menendez G, Zhou L, et al. ADP ribosylation factor 6 (ARF6) controls amyloid precursor protein (APP) processing by mediating the endosomal Compliance with Ethical Standards sorting of BACE1. Proc Natl Acad Sci. 2011;108:E559–68. 13. Guo JL, Buist A, Soares A, Callaerts K, Calafate S, Stevenaert F, Conflict of Interest All authors declare that they have no conflict of et al. The dynamics and turnover of tau aggregates in cultured cells: interest. insights into therapies for tauopathies. J Biol Chem. 2016;291: 13175–93. Human and Animal Rights and Informed Consent This article does not 14. Foglieni C, Papin S, Salvadè A, Afroz T, Pinton S, Pedrioli G, et al. contain any studies with human or animal subjects performed by any of Split GFP technologies to structurally characterize and quantify the authors. functional biomolecular interactions of FTD-related proteins. Sci Rep. 2017;7:14013. Open Access This article is distributed under the terms of the Creative 15. Ivatt RM, Sanchez-Martinez A, Godena VK, Brown S, Ziviani E, Commons Attribution 4.0 International License (http:// Whitworth AJ. Genome-wide RNAi screen identifies the Parkinson creativecommons.org/licenses/by/4.0/), which permits unrestricted use, disease GWAS risk locus SREBF1 as a regulator of mitophagy. distribution, and reproduction in any medium, provided you give appro- Proc Natl Acad Sci. 2014;111:8494–9. priate credit to the original author(s) and the source, provide a link to the 16. Do CB, Tung JY, Dorfman E, Kiefer AK, Drabant EM, Francke U, Creative Commons license, and indicate if changes were made. Mountain JL, Goldman SM, Tanner CM, Langston JW, Wojcicki A, Eriksson N Web-based genome-wide association study identifies two novel loci and a substantial genetic component for Parkinson’s disease. Gibson G, editor. PLoS Genet 2011;7: e1002141. 17. Nieland TJF, Logan DJ, Saulnier J, Lam D, Johnson C, Root DE, References et al. High content image analysis identifies novel regulators of synaptogenesis in a high-throughput RNAi screen of primary neu- 1. Ulirsch JC, Nandakumar SK, Wang L, Giani FC, Zhang X, Rogov rons. PLoS One. 2014;9:e91744. P, et al. Systematic functional dissection of common genetic varia- 18. Ugur B, Chen K, Bellen HJ. Drosophila tools and assays for the tion affecting red blood cell traits. Cell. 2016;165:1530–45. study of human diseases. Dis Model Mech. 2016;9:235–44. 2. Jones L, Lambert JC, Wang LS, Choi SH, Harold D, Vedernikov A, 19. Bellen HJ, Tong C, Tsuda H. 100 years of Drosophila research and et al. Convergent genetic and expression data implicate immunity in its impact on vertebrate neuroscience: a history lesson for the future. Alzheimer’s disease. Alzheimers Dement. 2015;11:658–71. Nat Rev Neurosci. 2010;11:514–22. 3. Huang K-L, Marcora E, Pimenova AA, Di Narzo AF, Kapoor M, 20. Venken KJT, Simpson JH, Bellen HJ. Genetic manipulation of Jin SC, et al. A common haplotype lowers PU.1 expression in genes and cells in the nervous system of the fruit fly. Neuron. myeloid cells and delays onset of Alzheimer’s disease. Nat 2011;72:202–30. Neurosci. 2017;20:1052–61. Curr Genet Med Rep 21. Mohr SE, Hu Y, Kim K, Housden BE, Perrimon N. Resources for 39. Jakobsdottir J, van der Lee SJ, Bis JC, Chouraki V, Li-Kroeger D, Yamamoto S, et al. Rare functional variant in TM2D3 is associated functional genomics studies in Drosophila melanogaster. Genetics. 2014;197:1–18. with late-onset Alzheimer’s disease. Haines JL, editor. PLoS Genet 2016;12:e1006327. 22. Brandt R, Gergou A, Wacker I, Fath T, Hutter H. A Caenorhabditis 40. Luo X, Rosenfeld JA, Yamamoto S, Harel T, Zuo Z, Hall M, et al. elegans model of tau hyperphosphorylation: induction of develop- Members of the UDN Clinically severe CACNA1A alleles affect mental defects by transgenic overexpression of Alzheimer’s synaptic function and neurodegeneration differentially. Lu B, editor disease-like modified tau. Neurobiol Aging. 2009;30:22–33. PLOS Genet 2017;13:e1006905. 23. McGuire SE, Le PT, Osborn AJ, Matsumoto K, Davis RL. 41. Prokic I, Cowling BS, Laporte J. Amphiphysin 2 (BIN1) in phys- Spatiotemporal rescue of memory dysfunction in drosophila. iology and diseases. J Mol Med. 2014;92:453–63. Science (80-. ). 2003;302:1765–8. 42. Singh MK, Dadke D, Nicolas E, Serebriiskii IG, Apostolou S, 24. Burcin MM, Schiedner G, Kochanek S, Tsai SY, O’Malley BW. Canutescu A, et al. A novel Cas family member, HEPL, regulates Adenovirus-mediated regulable target gene expression in vivo. Proc FAK and cell spreading. Mol Biol Cell. 2008;19:1627–36. Natl Acad Sci U S A. 1999;96:355–60. 43. Dustin ML, Olszowy MW, Holdorf AD, Li J, Bromley S, Desai N, 25. Osterwalder T, Yoon KS, White BH, Keshishian H. A conditional et al. A novel adaptor protein orchestrates receptor patterning and tissue-specific transgene expression system using inducible GAL4. cytoskeletal polarity in T-cell contacts. Cell. 1998;94:667–77. Proc Natl Acad Sci. 2001;98:12596–601. 44. Cormont M, Metón I, Mari M, Monzo P, Keslair F, Gaskin C, et al. 26. Shulman JM, Imboywa S, Giagtzoglou N, Powers MP, Hu Y, CD2AP/CMS regulates endosome morphology and traffic to the Devenport D, et al. Functional screening in Drosophila identifies degradative pathway through its interaction with Rab4 and c-Cbl. Alzheimer’s disease susceptibility genes and implicates tau- Traffic. 2003;4:97–112. mediated mechanisms. Hum Mol Genet. 2014;23:870–7. 45. Johnson RI, Seppa MJ, Cagan RL. The Drosophila CD2AP/CIN85 27. Shulman JM, Chipendo P, Chibnik LB, Aubin C, Tran D, Keenan BT, orthologue Cindr regulates junctions and cytoskeleton dynamics et al. Functional screening of Alzheimer pathology genome-wide asso- during tissue patterning. J Cell Biol. 2008;180:1191–204. ciation signals in Drosophila. Am J Hum Genet. 2011;88:232–8. 46. Lynch DK, Winata SC, Lyons RJ, Hughes WE, Lehrbach GM, 28. Dourlen P, Fernandez-Gomez FJ, Dupont C, Grenier-Boley B, Wasinger V, et al. A Cortactin-CD2-associated protein (CD2AP) Bellenguez C, Obriot H, et al. Functional screening of Alzheimer complex provides a novel link between epidermal growth factor risk loci identifies PTK2B as an in vivo modulator and early marker receptor endocytosis and the actin cytoskeleton. J Biol Chem. of tau pathology. Mol Psychiatry. 2017;22:874–83. 2003;278:21805–13. 29. Hu Y, Flockhart I, Vinayagam A, Bergwitz C, Berger B, Perrimon N, 47. Kobayashi S, Sawano A, Nojima Y, Shibuya M, Maru Y. The c- et al. An integrative approach to ortholog prediction for disease- Cbl/CD2AP complex regulates VEGF-induced endocytosis and focused and other functional studies. BMC Bioinform. 2011;12:357. degradation of Flt-1 (VEGFR-1). FASEB J. 2004;18:929–31. 30. Moreau K, Fleming A, Imarisio S, Lopez Ramirez A, Mercer JL, 48. Haglund K, Nezis IP, Lemus D, Grabbe C, Wesche J, Liestøl K, Jimenez-Sanchez M, et al. PICALM modulates autophagy activity et al. Cindr interacts with anillin to control cytokinesis in and tau accumulation. Nat Commun. 2014;5:4998. Drosophila melanogaster. Curr Biol. 2010;20:944–50. 31. Song L, He Y, Ou J, Zhao Y, Li R, Cheng J, et al. Auxilin underlies 49. Gallo J-M, Spickett C. The role of CELF proteins in neurological progressive locomotor deficits and dopaminergic neuron loss in a disorders. RNA Biol. 2010;7:474–9. Drosophila model of Parkinson’s disease. Cell Rep. 2017;18:1132–43. 50. Hirai H, Maru Y, Hagiwara K, Nishida J, Takaku F. A novel puta- 32. MacLeod DA, Rhinn H, Kuwahara T, Zolin A, Di Paolo G, tive tyrosine kinase receptor encoded by the eph gene. Science. McCabe BD, et al. RAB7L1 interacts with LRRK2 to modify 1987;238:1717–20. intraneuronal protein sorting and Parkinson’s disease risk. 51. Yamazaki T, Masuda J, Omori T, Usui R, Akiyama H, Maru Y. Neuron. 2013;77:425–39. EphA1 interacts with integrin-linked kinase and regulates cell mor- 33. Stefansson H, Rye DB, Hicks A, Petursson H, Ingason A, phology and motility. J Cell Sci. 2009;122:243–55. Thorgeirsson TE, et al. A genetic risk factor for periodic limb 52. Lai K-O, Ip NY. Synapse development and plasticity: roles of ephrin/ movements in sleep. N Engl J Med. 2007;357:639–47. Eph receptor signaling. Curr Opin Neurobiol. 2009;19:275–83. 34. Freeman A, Pranski E, Miller RD, Radmard S, Bernhard D, Jinnah 53. Lai-Cheong JE, Parsons M, McGrath JA. The role of kindlins in cell HA, et al. Sleep fragmentation and motor restlessness in a biology and relevance to human disease. Int J Biochem Cell Biol. Drosophila model of restless legs syndrome. Curr Biol. 2012;22: 2010;42:595–603. 1142–8. 54. Marat AL, Dokainish H, McPherson PS. DENN domain proteins: 35. Jia T, Macare C, Desrivières S, Gonzalez DA, Tao C, Ji X, et al. regulators of Rab GTPases. J Biol Chem. 2011;286:13791–800. Neural basis of reward anticipation and its genetic determinants. 55. Miyoshi J, Takai Y. Dual role of DENN/MADD (Rab3GEP) in Proc Natl Acad Sci. 2016;113:3879–84. neurotransmission and neuroprotection. Trends Mol Med. 36. Schumann G, Coin LJ, Lourdusamy A, Charoen P, Berger KH, 2004;10:476–80. Stacey D, et al. Genome-wide association and genetic functional 56. Girault JA, Costa A, Derkinderen P, Studler JM, Toutant M. FAK and studies identify autism susceptibility candidate 2 gene (AUTS2) in PYK2/CAKbeta in the nervous system: a link between neuronal activ- the regulation of alcohol consumption. Proc Natl Acad Sci U S A. ity, plasticity and survival? Trends Neurosci. 1999;22:257–63. 2011;108:7119–24. 57. Ivankovic-Dikic I, Grönroos E, Blaukat A, Barth BU, Dikic I. Pyk2 37. Vanden Broeck L, Kleinberger G, Chapuis J, Gistelinck M, and FAK regulate neurite outgrowth induced by growth factors and Amouyel P, Van Broeckhoven C, et al. Functional complementa- integrins. Nat Cell Biol. 2000;2:574–81. tion in Drosophila to predict the pathogenicity of TARDBP vari- 58. Huang Y, Lu W, Ali DW, Pelkey KA, Pitcher GM, Lu YM, et al. ants: evidence for a loss-of-function mechanism. Neurobiol Aging. CAKbeta/Pyk2 kinase is a signaling link for induction of long-term 2015;36:1121–9. potentiation in CA1 hippocampus. Neuron. 2001;29:485–96. 38. Chapuis J, Hansmannel F, Gistelinck M, Mounier A, Van 59. Hsin H, Kim MJ, Wang C-F, Sheng M. Proline-rich tyrosine kinase Cauwenberghe C, Kolen KV, et al. GERAD consortium. 2 regulates hippocampal long-term depression. J Neurosci. Increased expression of BIN1 mediates Alzheimer genetic risk by 2010;30:11983–93. modulating tau pathology. Mol Psychiatry. 2013 Nov;18(11):1225- 60. Strappazzon F, Torch S, Trioulier Y, Blot B, Sadoul R, Verna J-M. 34. Survival response-linked Pyk2 activation during potassium Curr Genet Med Rep depletion-induced apoptosis of cerebellar granule neurons. Mol TREM2 implicate microglial-mediated innate immunity in Alzheimer’s disease. Nat Genet. 2017;49:1373–84. Cell Neurosci. 2007;34:355–65. 61. Giralt A, Coura R, Girault J-A. Pyk2 is essential for astrocytes 63. Ryan KJ, White CC, Patel K, Xu J, Olah M, Replogle JM, et al. A mobility following brain lesion. Glia. 2016;64:620–34. human microglia-like cellular model for assessing the effects of 62. Sims R, van der Lee SJ, Naj AC, Bellenguez C, Badarinarayan N, neurodegenerative disease gene variants. Sci Transl Med. Jakobsdottir J, et al. Rare coding variants in PLCG2, ABI3, and 2017;9(421):eaai7635.

Journal

Current Genetic Medicine ReportsSpringer Journals

Published: May 29, 2018

References

You’re reading a free preview. Subscribe to read the entire article.


DeepDyve is your
personal research library

It’s your single place to instantly
discover and read the research
that matters to you.

Enjoy affordable access to
over 18 million articles from more than
15,000 peer-reviewed journals.

All for just $49/month

Explore the DeepDyve Library

Search

Query the DeepDyve database, plus search all of PubMed and Google Scholar seamlessly

Organize

Save any article or search result from DeepDyve, PubMed, and Google Scholar... all in one place.

Access

Get unlimited, online access to over 18 million full-text articles from more than 15,000 scientific journals.

Your journals are on DeepDyve

Read from thousands of the leading scholarly journals from SpringerNature, Elsevier, Wiley-Blackwell, Oxford University Press and more.

All the latest content is available, no embargo periods.

See the journals in your area

DeepDyve

Freelancer

DeepDyve

Pro

Price

FREE

$49/month
$360/year

Save searches from
Google Scholar,
PubMed

Create lists to
organize your research

Export lists, citations

Read DeepDyve articles

Abstract access only

Unlimited access to over
18 million full-text articles

Print

20 pages / month

PDF Discount

20% off