Serotonergic signalling suppresses ataxin 3 aggregation and neurotoxicity in animal models of Machado-Joseph disease

Abstract

Polyglutamine diseases are a class of dominantly inherited neurodegenerative disorders for which there is no effective treatment. Here we provide evidence that activation of serotonergic signalling is beneficial in animal models of Machado-Joseph disease. We identified citalopram, a selective serotonin reuptake inhibitor, in a small molecule screen of FDA-approved drugs that rescued neuronal dysfunction and reduced aggregation using a Caenorhabditis elegans model of mutant ataxin 3-induced neurotoxicity. MOD-5, the C. elegans orthologue of the serotonin transporter and cellular target of citalopram, and the serotonin receptors SER-1 and SER-4 were strong genetic modifiers of ataxin 3 neurotoxicity and necessary for therapeutic efficacy. Moreover, chronic treatment of CMVMJD135 mice with citalopram significantly reduced ataxin 3 neuronal inclusions and astrogliosis, rescued diminished body weight and strikingly ameliorated motor symptoms. These results suggest that small molecule modulation of serotonergic signalling represents a promising therapeutic target for Machado-Joseph disease.

Keywords: ataxin 3 aggregation; selective serotonin reuptake inhibitor, citalopram; spinocerebellar ataxia type 3; therapy.

Spinocerebellar ataxia type 3 (Machado-Joseph disease) is an incurable neurodegenerative disorder caused by polyglutamine expansion within ataxin-3. In a repurposing drug screen, Teixeira-Castro et al. identify the selective serotonin reuptake inhibitor citalopram as a suppressor of mutant ataxin-3 neurotoxicity in transgenic C. elegans and mouse models of the disease.

Figure 1

Identification of small-molecule suppressors of ATXN3 pathogenesis. (A) Graphical representation of the results of the C. elegans-based screening in which we assessed a subset of 599 compounds of the Prestwick Chemical library. Black lines represent the mean percentage of animals with locomotion impairment for untreated AT3q130, AT3WT and wild-type animals. Grey dots show the mean of percentage of locomotion impaired AT3q130 animals upon treatment for each compound. The red line represents the assay cut-off to a minimum of 50% of effect in animals’ motor behaviour. The red dot represents the percentage of locomotion impaired AT3q130 animals upon citalopram treatment, as obtained in the primary screen. (B) Motility analysis of AT3q130 animals treated with the top 11 hit compounds found in the screen (n = 4, ± SD), **P < 0.01, ***P < 0.001 (Student’s t-test). (C) Quantification of AT3q130 aggregation by confocal imaging and fluorescence intensity (n ≥ 10, ± SD) *P < 0.05, **P < 0.01, ***P < 0.001 (ANOVA, Bonferroni’s test). (D) Human ATXN3 expression in AT3q130 animals treated with the top hit compounds (n = 4–6 ± SEM). *P < 0.05 (Student’s t-test). (E) Motility analysis of AT3q130 animals treated with compounds targeting serotonergic neurotransmission found in the screen (n = 4, ± SD), *P < 0.05, **P < 0.01, ***P < 0.001 (Student’s t-test). WT = wild-type; AT3WT = wild-type ATXN3 expressing animals; AT3q130 = mutant ATXN3 expressing animals.

Figure 2

Early life chronic citalopram treatment suppressed mutant ATXN3 aggregation and neuronal dysfunction in C. elegans. (A) Aggregate quantification in AT3q130 animals upon citalopram (cit) treatment. (B) Representative western blot analysis of ATXN3 protein upon biochemical fractionation of AT3q130 protein extracts (out of n = 3). (C) Human ATXN3 protein levels in AT3q130 cit animals. (D) Motility of citalopram-treated wild-type (WT::cit) and AT3q130 (AT3q130::cit) animals and OFF-drug effect (AT3q130::cit OFF) as disease progressed. (E) Motor behaviour of AT3q130 cit animals treated for 4 days, with treatment initiation at the indicated days and (F) treatment duration for the indicated days. (G) Locomotion impairment and aggregation load of AT3q130 animals in the mod-5 background and upon treatment with citalopram, S-citalopram and fluoxetine. For motor behaviour assays: (n = 3–4 ± SD), *P < 0.05, **P < 0.01, ***P < 0.001 (Student’s t-test). For aggregate quantification: (n ≥ 8 ± SD), *P < 0.05, **P < 0.01, ***P < 0.001 (ANOVA, Bonferroni’s test). For western blot: (n = 4 ± SD), P > 0.05 (Student’s t-test). TX100 = Triton^™ X-100; FA = formic acid; cit = citalopram; S-cit = S-citalopram; fluox = fluoxetine; WT = wild-type.

Figure 3

Serotonergic signalling improves ATXN3 pathogenesis in a G-protein coupled receptor-dependent manner. (A) Schematic of a serotonergic synapse showing that 5-HT is synthesized and released into the synaptic cleft, where it activates postsynaptic 5-HT receptors. Vabicaserin may activate the regulatory signalling coupled with SER-1. Pindolol, lysergol and DHE probably antagonize the 5-HT autoreceptor SER-4, whereas the 5-HT1A receptor agonist buspirone may desensitize the receptor, shutting down the stop production signal mediated by 5-HT autoreceptors in presynaptic neurons. (B) Motor behaviour of AT3q130 animals treated with vabicaserin, buspirone and dihydroergotamine. (C) Motility defects and aggregation load of AT3q130 animals in a ser-1 genetic background, with and without cit, S- citalopram and estrone treatments. (D) Motility performance and mutant ATXN3 aggregation phenotypes of AT3q130 animals in the absence of SER-4. For motor behaviour assays: (n = 3–4, ± SD), *P < 0.05, **P < 0.01, ***P < 0.001 (Student’s t-test). For aggregate quantification: (n ≥ 12, ± SD) *P < 0.05, **P < 0.01, ***P < 0.001 (ANOVA, Bonferroni’s test). DHE = dihydroergotamine; DMSO = dimethyl sulphoxide; cit = citalopram; S-cit = S-citalopram; fluox = fluoxetine.

Figure 4

CMVMJD135 mice show normal levels of 5-HT and 5-HT metabolic turn over at fully symptomatic stages of disease. Levels of 5-HT, 5-HIAA and 5-HT turnover (5-HIAA/5-HT) were measured by HPLC (A) in the cerebellum (n = 5 wild-type; n = 7 CMVMJD135), (B) medulla oblongata (n = 6 wild-type; n = 8 CMVMJD135) and (C) substantia nigra (n = 5 wild-type; n = 5 CMVMJD135) at 24 weeks of age. Data presented as the mean ± SEM. P > 0.05 (Student’s t-test). TO = turnover; WT = wild-type.

Figure 5

Impact of citalopram treatment at 8 mg/kg on the neurological deficits of CMVMJD135 mice. (A) Schematic representation of the preclinical therapeutic trial. Significant differences observed between vehicle (n = 13) and citalopram-treated CMVMJD135 mice (n = 16) in (B) body weight (P = 0.001, 20–34 weeks) and (C) stride length (P = 0.015, 30 weeks). (D) Tremors, (E) limb clasping and (F) gait were evaluated from 18 to 34 weeks of age with phenotype amelioration from 22 to 34 weeks of age. (n = 13–16, ± SD), *P < 0.05, **P < 0.01 and ***P < 0.001 (Mann-Whitney U-test for non-parametric variables and ANOVA, Tukey correction for continuous variables). cit = citalopram.

Figure 6

Citalopram treatment ameliorates balance and motor coordination and suppresses mutant ATXN3 aggregation in Machado-Joseph disease mice. Significant differences observed between vehicle (n = 13) and citalopram-treated CMVMJD135 mice (n = 16) in the (A) square beam (P < 0.001, 20–34 weeks) and (B) motor swimming (P < 0.001, 14–34 weeks) tests. (C) Immunohistochemistry and quantification of GFAP-positive cells per area in substantia nigra from wild-type, vehicle- and citalopram-treated CMVMJD135 mice (n = 5 per group, 34 weeks). (D) Neuronal inclusions in the pontine nuclei, reticulotegmental nucleus of the pons, facial motor nucleus and lateral reticular nucleus of vehicle and citalopram-treated CMVMJD135 mice (n = 4, 34 weeks). (E) Brainstem immunoblots and quantification of total human ATXN3 protein from vehicle and citalopram-treated CMVMJD135 mice (n = 5, 34 weeks). Data presented as mean ± SEM, *P < 0.05 and **P < 0.01 [ANOVA, Tukey correction (A and B) and one-way ANOVA (C–E)]. Scale bars = 20 µm. cit = citalopram; WT = wild-type; SN = substantia nigra; PN = pontine nuclei; RtTg = reticulotegmental nuclei of pons; 7N = facial motor nuclei; LRt = lateral reticular nuclei.

Figure 7

Citalopram treatment shows neuroprotective effects in CMVMJD135 mice. (A) Immunohistochemistry and quantification of ChAT-positive cells per total area in the facial motor nucleus from wild-type, vehicle and citalopram-treated CMVMJD135 mice (n = 4 per group, 34 weeks). (B) Immunohistochemistry and quantification of ChAT-positive cells per total area in the lumbar region of the ventral horn of the spinal cord (LSC) from wild-type, vehicle and citalopram-treated CMVMJD135 mice (n = 3–4 per group, 34 weeks). (C) Immunohistochemistry and quantification of Calbindin D28K-positive Purkinje cells per total area in the cerebellar cortex from wild-type, vehicle and citalopram-treated CMVMJD135 mice (n = 4 per group, 34 weeks). Data presented as mean ± SEM, *P < 0.05 and **P < 0.01 (one-way ANOVA). Scale bars = 200 µm (A) and 100 µm (B and C). cit = citalopram; WT = wild-type; 7N = facial motor nuclei; LSC = lumbar spinal cord; CBX = cerebellar cortex.