A conditional pan-neuronal Drosophila model of spinocerebellar ataxia 7 with a reversible adult phenotype suitable for identifying modifier genes

Abstract

Spinocerebellar ataxia 7 (SCA7) is a neurodegenerative disease caused by a polyglutamine (polyQ) expansion in the ataxin 7 (ATXN7) protein, a member of a multiprotein complex involved in histone acetylation. We have created a conditional Drosophila model of SCA7 in which expression of truncated ATXN7 (ATXN7T) with a pathogenic polyQ expansion is induced in neurons in adult flies. In this model, mutant ATXN7T accumulated in neuronal intranuclear inclusions containing ubiquitin, the 19S proteasome subunit, and HSP70 (heat shock protein 70), as in patients. Aggregation was accompanied by a decrease in locomotion and lifespan but limited neuronal death. Disaggregation of the inclusions, when expression of expanded ATXN7T was stopped, correlated with improved locomotor function and increased lifespan, suggesting that the pathology may respond to treatment. Lifespan was then used as a quantitative marker in a candidate gene approach to validate the interest of the model and to identify generic modulators of polyQ toxicity and specific modifiers of SCA7. Several molecular pathways identified in this focused screen (proteasome function, unfolded protein stress, caspase-dependent apoptosis, and histone acetylation) were further studied in primary neuronal cultures. Sodium butyrate, a histone deacetylase inhibitor, improved the survival time of the neurons. This model is therefore a powerful tool for studying SCA7 and for the development of potential therapies for polyQ diseases.

Figure 1.

Expression of ATXN7T in transgenic Drosophila strains. A, Western blot of whole extracts (50 μg) of larvae ubiquitously expressing ATXN7T-10Q with 10 polyglutamine repeats (UAS-SCA7T-10Q/+; da-Gal4/+) and ATXN7T-102Q with an expanded polyglutamine (UAS-SCA7T-102Q/+; da-Gal4/+). The asterisk indicates SDS-insoluble polyglutamine aggregates in the stacking gel. Arrows, ATXN7T proteins expressed at the expected sizes. Note the smaller N-terminal fragments of ATXN7T-102Q. γ-Tubulin labeling served as the loading control. B, Staining of ATXN7T (red) in neurons from the central complex of elavC¹⁵⁵-Gal4; UAS-SCA7T-102Q/+ male flies, labeled with the 1C1 antibody. Cell nuclei are counterstained with DAPI. ATXN7T-102Q forms large inclusions in DAPI-labeled nuclei (blue), whereas ATXN7T-10Q remains diffuse in the nuclei (elavC¹⁵⁵-Gal4; UAS-SCA7T-10Q/+ male flies). Scale bars, 5 μm.

Figure 2.

Induced expression of mutant ATXN7T in adult Drosophila neurons decreases locomotor activity and lifespan. A, Western blot of whole extracts of heads from 15-d-old male flies (UAS-SCA7T-10Q/+; elavGS/+ and UAS-SCA7T-102Q/+; elavGS/+) in which expression of ATXN7T-10Q and ATXN7T-102Q was induced (RU+) or not induced (RU−) with RU486. The asterisk indicates SDS-insoluble polyQ aggregates in the stacking gel. Arrows, ATXN7T proteins expressed at the expected sizes. When overexposed, a thin band of ATXN7T was detected in the absence of RU486, indicating a slight leak in the control of expression. B, Lifespan of flies expressing or not expressing ATXN7T-10Q and ATXN7T-102Q in adult neurons. Note that ATXN7T-102Q expression strongly reduced lifespan (log rank analysis; p < 10⁻⁵⁰), whereas ATXN7T-10Q expression had only a slight effect (p < 10⁻¹⁰). RU486 had no effect on longevity in wild-type flies (data not shown). C, Locomotor activity (performance on RING test) of 16-d-old flies expressing or not expressing ATXN7T-10Q and ATXN7T-102Q transgenes. *Significantly different from the ATXN7T-10Q condition (p = 0.006; Student's t test). Error bars represent SEM.

Figure 3.

Adult neurons in Drosophila expressing mutant ATXN7T do not degenerate massively but contain NIIs. A, B, Neuronal death was assessed by TUNEL analysis 18 d after induction (RU+), or not (RU−), of ATXN7T-10Q (UAS-SCA7T-10Q/+; elavGS/+) and ATXN7T-102Q (UAS-SCA7T-102Q/+; elavGS/+) expression. A, Overview of a brain section from a fly expressing ATXN7T-102Q, showing mutant ATXN7T (red) in nuclei (blue) of neurons. A cluster of TUNEL-positive (green) nuclei can be seen on the edge of the optic lobes (boxed). A low level of autofluorescence can be seen throughout the section. Scale bar, 20 μm. B, Enlarged view of boxed area in A compared with an equivalent area from a SCA7T-102Q transgenic fly in which expression of the transgene was not induced and in which no TUNEL-positive cells (arrowheads) were seen. Scale bars, 10 μm. C, Colocalization of ATXN7T, shown here in the optic lobes of UAS-SCA7T-102Q/+; elavGS/+ male flies, with the 19S proteasome subunit, ubiquitin (Ubi), and HSP70. Images in B and C were obtained by spectral confocal microscopy. Scale bars, 10 μm. D, Ultrastructure of a nucleus containing immunogold-labeled mutant ATXN7T (black dots indicated by black arrowheads) in the Kenyon cell region of UAS-SCA7T-102Q/+; elavGS/+ male flies. The very large (≈2 μm) fibrillary aggregate of mutant ATXN7T (white dashed circle) is composed of thin (≈0.3 μm) interwoven fibrils (large black arrow) of ATXN7T. The nuclear membrane (small black arrow), mitochondria (m), and chromatin (Ch) appear normal. Scale bar, 1 μm.

Figure 4.

Lifespan, locomotor activity, and ATXN7T aggregation improve when expression of ATXN7T-102Q is stopped in UAS-SCA7T-102Q/+; elavGS/+ male flies. A, Lifespan when ATXN7T-102Q was not induced (control: RU−), induced continuously with RU486 (RU+), or induced for 2 d (RU+ 2 days) or 4 d (RU+ 4 days). Lifespan was inversely correlated with the duration of RU486 treatment (p < 10⁻²⁰). B, RING test at 16, 18, 21, and 25 d after induction. Locomotor performance was expressed as the ratio of the height reached in 4 s by flies with induced expression of ATXN7T to that reached by noninduced flies of the same strain. At 18 and 21 d after induction, it was better in flies treated for 4 d with RU486 than in those treated continuously, and it remained at control levels at all time points if treatment lasted only 2 d (RU+ 2 d/RU− = 1 ± 0.1). *p < 0.005, significantly different from RU+ condition (Student's t test). At 25 d after induction in continuously treated flies, a statistical analysis could not be performed, because the continuously treated flies were already dead. C, Proportion of large NIIs in ATXN7T-positive neurons over time, in flies treated with RU486 continuously (black bars) or for 2 d (white bars), in a morphologically defined zone of the central complex (supplemental figure, available at www.jneurosci.org as supplemental material). *Significantly less than when treated continuously (p < 0.05; 2-way ANOVA, followed by the Student–Newman–Keuls test for multiple comparisons). There was a significant difference according to induction state (p = 0.009); the interaction between induction state and time after induction (age) was also significant (p = 0.026). nbr, Number. Error bars represent SEM.

Figure 5.

Proteins involved in the SCA7 phenotype in Drosophila are relocated into NIIs in mammalian neurons expressing mutant ATXN7T, and sodium butyrate alleviates fly neuronal death. A, Spectral confocal microscopic images showing colocalization in NIIs of ATXN7T with protein markers of modifier pathways identified in this study: the 19S proteasome subunit, ubiquitin (Ubi), HSP70, activated caspase 3, and components of the mammalian histone acetylation STAGA complex (GCN5 and SPT3). Scale bars, 5 μm. B, The cytotoxicity of ATXN7T-100Q–EGFP, ATXN7T-10Q–EGFP, and EGFP alone was evaluated in the presence or the absence of the HDAC inhibitor SB. *Significantly different from untreated cells (p < 0.001; 1-way ANOVA, followed by the Student–Newman–Keuls test for multiple comparisons). Error bars represent SEM.