A Druggable Genome Screen Identifies Modifiers of α-Synuclein Levels via a Tiered Cross-Species Validation Approach

Abstract

Accumulation of α-Synuclein (α-Syn) causes Parkinson's disease (PD) as well as other synucleopathies. α-Syn is the major component of Lewy bodies and Lewy neurites, the proteinaceous aggregates that are a hallmark of sporadic PD. In familial forms of PD, mutations or copy number variations in SNCA (the α-Syn gene) result in a net increase of its protein levels. Furthermore, common risk variants tied to PD are associated with small increases of wild-type α-Syn levels. These findings are further bolstered by animal studies which show that overexpression of α-Syn is sufficient to cause PD-like features. Thus, increased α-Syn levels are intrinsically tied to PD pathogenesis and underscore the importance of identifying the factors that regulate its levels. In this study, we establish a pooled RNAi screening approach and validation pipeline to probe the druggable genome for modifiers of α-Syn levels and identify 60 promising targets. Using a cross-species, tiered validation approach, we validate six strong candidates that modulate α-Syn levels and toxicity in cell lines, Drosophila, human neurons, and mouse brain of both sexes. More broadly, this genetic strategy and validation pipeline can be applied for the identification of therapeutic targets for disorders driven by dosage-sensitive proteins.SIGNIFICANCE STATEMENT We present a research strategy for the systematic identification and validation of genes modulating the levels of α-Synuclein, a protein involved in Parkinson's disease. A cell-based screen of the druggable genome (>7,500 genes that are potential therapeutic targets) yielded many modulators of α-Synuclein that were subsequently confirmed and validated in Drosophila, human neurons, and mouse brain. This approach has broad applicability to the multitude of neurological diseases that are caused by mutations in genes whose dosage is critical for brain function.

Keywords: modifier; neurodegeneration; pooled screen; protein dosage; shRNA; α-Synuclein.

Figure 1.

Screening the druggable genome to identify modulators of α-Syn levels. A, Strategy used to identify modulators of α-Syn via pooled RNAi libraries in transgenic reporter cells expressing equimolar amounts of α-Syn:EGFP and DsRed (see Table 1-2). Approach was used for eight distinct libraries in quadruplicate. B, Computational prioritization of top gene candidates via tier classification. Hit ratio, directionality score, and conflict score are presented as violin plots. See also Figure 1-1; and Table 1-1.

Figure 2.

Validation in human cells reveals modulators of endogenous α-Syn levels. A, Validation approach in HEK293T cells (see Table 1-2). Individual lentiviral infection is performed in quadruplicate using three shRNAs per gene target of the Tier 1 candidates (350 genes). Individual samples are collected and tested for α-Syn levels via ELISA (as well as total protein levels). Hits that decrease α-Syn levels are further confirmed using Western blot analysis. B, Summary chart of all shRNAs tested for Tier 1 candidates. Blue bars represent hits that decrease α-Syn levels. Red bars represent those that increase α-Syn levels (n = 4 samples per condition). C, The top 33 hits that decrease α-Syn by >15% using ≥2 shRNAs are retested by Western blot. A representative Western blot for α-Syn following downregulation of candidate targets CDK14, CHMP4B, and DUSP11 is included above the graphical representation (n = 4 samples per condition). *q < 0.1 (one-way ANOVA followed by correction for false discovery rate of 10% via the two-stage step-up method of Benjamini and Yekutieli, 2006). See also Figure 2-1; and Table 1-1.

Figure 3.

Cross-species validation of novel α-Syn modifiers in Drosophila. A, Approach for identifying modifiers of α-Syn toxicity in Drosophila. B, Motor performance as a function of age in control fruit flies (blue line in every panel), flies specifically expressing α-Syn in neurons (black line in every panel), and animals expressing α-Syn and carrying genetic variants reducing the function of the indicated candidate gene in neurons (red line in every panel). Two read-out metrics are represented: speed in mm/s (top panel for each gene) and stumbling events (bottom panel for each gene). Expression of α-Syn in the Drosophila nervous system (using elav-Gal4) leads to progressive motor performance deficits, resulting in decreased average speed and increased stumbling events as the fruit flies age. Reducing levels of each of the indicated candidate genes result in an amelioration of the α-Syn-induced motor impairment, seen as improved average speed and decreased number of stumbles as the animals age, compared with Drosophila expressing α-Syn alone. Error bars indicate SEM. *Statistical significance between α-Syn alone (black line) or together with a candidate gene knockdown (red line) based on a linear mixed-effect model ANOVA (α = 0.05). LOF, Loss of function; hp, hairpin RNAi. The specific genotypes are indicated in Table 1. For Additional characterization of the Drosophila model and results with additional alleles see Figure 3-1.

Figure 4.

Testing candidate α-Syn modulators in human neurons. A, Approach for identifying modulators of α-Syn levels in hESC-derived neurons (see Table 1-2). Cells are infected for 2 weeks with three independent shRNAs. Expression of virus is confirmed by presence of turboGFP (tGFP) signal (B), and candidate knockdown is confirmed using qPCR (see Table 1-1). C, Representative Western blot from human neurons infected with lentiviruses harboring the indicated shRNAs targeting candidate genes LOXL1, SENP8, and CDK14. D, Quantification of each candidate tested (n = 8 samples per condition). Each bar represents an individual shRNA. Three shRNAs were tested per gene, except for HLA-DRB1. Percentage change in α-Syn levels was quantified in relation to a scrambled shRNA. *q < 0.1 (one-way ANOVA followed by correction for false discovery rate of 10% via the two-stage step-up method of Benjamini and Yekutieli, 2006). See also Figure 4-1; and Table 1-1.

Figure 5.

Candidate modulators regulate α-Syn levels in the mammalian brain. A, Approach to testing modulators of α-Syn levels in mouse brain. B, Confirmation of virus expression using UV light on 3-d-old pups (3 d post injection [D.P.I.]) and epifluorescence on cryosectioned tissue from 3-week-old pups (21 D.P.I.). C, Representative Western blots for α-Syn following downregulation of candidate targets Chmp4b, Acsbg1, and Lgals3bp are included above the graphical representation (n = 4–11 samples per condition). D, Each gene was targeted using two independent shRNAs (each bar represents an individual shRNA), except for Chmp4b (one caused perinatal lethality). *q < 0.1 (one-way ANOVA followed by correction for false discovery rate of 10% via the two-stage step-up method of Benjamini and Yekutieli, 2006). See also Table 1-1.

Figure 6.

Mechanistic insight into the regulation α-Syn levels for two candidate modulators: DUSP11 and LGALS3BP. A, Experimental design for testing individual modifiers of protein stability in a bidirectional manner. B, Flow cytometry-based quantification of α-Syn stability following DUSP11 and LGALS3BP siRNA knockdown and overexpression in experimental (DsRed-IRES-SNCA-EGFP; blue) and control (DsRed-IRES-EGFP; white) cell lines (see Table 1-2). Data are presented as the percentage change compared with either siScrambled or a control plasmid (empty Flag-tag vector). C, qPCR for Snca expression following neonatal depletion of Dusp11 and Lgals3bp in mouse brain. D, Immunoprecipitation of endogenous α-Syn from HEK293T cells transfected with indicated flag-tagged constructs. α-Syn interacts with DUSP but not the other proteins; the signal in CDK14 IP is not specific as it is also seen with the IgG control. Western blot is representative of three independent experiments. ***p < 0.001, ****p < 0.0001. B, C, ANOVA followed by Dunnett's multiple-comparisons test.

Figure 7.

Pipeline for identifying potentially druggable modulators of α-Syn levels. Validation pipeline for uncovering modulators of α-Syn levels. Primary cell-based screen is performed using pooled libraries of shRNAs targeting >7500 genes (Fig. 1). Secondary screens in HEK293T cells (Fig. 2) and Drosophila (Fig. 3) help narrow down the list and highlight strong modulators of α-Syn levels and toxicity. Validation in human neurons (Fig. 4) and mouse brain (Fig. 5) ensures robustness of hits and validates targets for preclinical nomination.