The antipsychotic medications aripiprazole, brexpiprazole and cariprazine are off-target respiratory chain complex I inhibitors

Abstract

Antipsychotic drugs are the mainstay of treatment for schizophrenia and provide adjunct therapies for other prevalent psychiatric conditions, including bipolar disorder and major depressive disorder. However, they also induce debilitating extrapyramidal syndromes (EPS), such as Parkinsonism, in a significant minority of patients. The majority of antipsychotic drugs function as dopamine receptor antagonists in the brain while the most recent 'third'-generation, such as aripiprazole, act as partial agonists. Despite showing good clinical efficacy, these newer agents are still associated with EPS in ~ 5 to 15% of patients. However, it is not fully understood how these movement disorders develop. Here, we combine clinically-relevant drug concentrations with mutliscale model systems to show that aripiprazole and its primary active metabolite induce mitochondrial toxicity inducing robust declines in cellular ATP and viability. Aripiprazole, brexpiprazole and cariprazine were shown to directly inhibit respiratory complex I through its ubiquinone-binding channel. Importantly, all three drugs induced mitochondrial toxicity in primary embryonic mouse neurons, with greater bioenergetic inhibition in ventral midbrain neurons than forebrain neurons. Finally, chronic feeding with aripiprazole resulted in structural damage to mitochondria in the brain and thoracic muscle of adult Drosophila melanogaster consistent with locomotor dysfunction. Taken together, we show that antipsychotic drugs acting as partial dopamine receptor agonists exhibit off-target mitochondrial liabilities targeting complex I.

Fig. 1

Aripiprazole and its primary active metabolite are selectively toxic to galactose-conditioned SH-SY5Y cells. A Normalised cellular ATP measurements from glucose- and galactose-conditioned SH-SY5Y cells exposed to the indicated antipsychotic drugs (10 μM), piericidin A, antimycin A (5 µM) or 2-deoxyglucose (2-DG) (50 mM) for 18 h (the data are mean averages ± range from 2 independent experiments with 3 technical repeats; asterisks show the results from one-way ANOVA with Dunnett’s multiple comparison test, normalised to control). B Normalised cellular ATP measurements from glucose- and galactose-conditioned SH-SY5Y cells treated with the indicated concentrations of aripiprazole or dehydroaripiprazole for 18 h. Data were fitted to the standard dose–effect relationship (activity (%) =  = Bottom + (Top − Bottom)/(1 + 10^((LogIC50-X) × HillSlope)) using GraphPad Prism version 8.0 (mean ± SEM from 3 independent experiments, normalised to control). C Quantification of cell death in glucose- and galactose-conditioned SH-SY5Y cells treated with increasing concentrations of aripiprazole and dehydroaripiprazole for 18 h (mean ± SEM from 3 independent experiments, asterisks, one-way ANOVA with Dunnett’s multiple comparison test, normalised to control) (corresponds to Additional file 1: Fig. S1). D Quantification of glucose and galactose-conditioned SH-SY5Y cell proliferative rates following treatment with indicated concentrations of aripiprazole or dehydroaripiprazole over an 86 h period. Proliferation rates were calculated from the linear phase of each growth curve (mean ± SEM from 3 independent experiments, asterisks, one-way ANOVA with Dunnett’s multiple comparison test, normalised to control) (corresponds to Additional file 1: Fig. S2)

Fig. 2

Aripiprazole, Brexpiprazole, and Cariprazine directly inhibit respiratory complex I via the Q-channel. A Normalised basal and maximal OCR measurements from SH-SY5Y cells treated for 4 h with indicated concentrations of aripiprazole, dehydroaripiprazole, brexpiprazole, cariprazine or didesmethylcariprazine (mean ± SEM from 4 independent experiments, asterisks, one-way ANOVA with Dunnett’s multiple comparison test, normalised to control) (corresponds to Additional file 1: Fig. S3C). B Normalised OCR measurements driven by complex I substrates (pyruvate/malate) or the complex II substrate (succinate). Permeabilised SH-SY5Y cells were treated with 10 μM aripiprazole, dehydroaripiprazole, brexpiprazole, cariprazine or 5 μM piericidin A, antimycin A (mean ± SEM from 4 independent experiments, asterisks, one-way ANOVA with Dunnett’s multiple comparison test, normalised to control) (corresponds to Additional file 1: Fig. S4B). C A cartoon depiction of mammalian complex I structure with the two substrate-binding sites indicated. D Normalised NADH:O₂ oxidoreduction rates of bovine mitochondrial membranes exposed to varying concentrations of aripiprazole, brexpiprazole, cariprazine, dehydroaripiprazole and didesmethylcariprazine. Respective IC₅₀ values from the standard dose–effect relationship (see “Methods” Section) are indicated in brackets. E Normalised NADH:APAD⁺, F NADH:FeCN, and G NADH:DQ oxidoreduction rates of isolated bovine complex I exposed to 100 μM aripiprazole, brexpiprazole, cariprazine or 1 µM piericidin A (mean ± SEM from 3 independent wells per treatment, asterisks, one-way ANOVA with Dunnett’s multiple comparison test, normalised to control)

Fig. 3

Aripiprazole, brexpiprazole and cariprazine are toxic to primary neonatal mouse neurons. A Representative image of a sagittal section from an embryonic mouse brain (e13.5) stained with an anti-tyrosine hydroxylase (TH) antibody to highlight dopaminergic neurons (white arrow). Dashed yellow boxes represent the ventral midbrain and forebrain areas used to prepare neurons. B Representative confocal images of ventral midbrain and forebrain neurons (e13.5) maintained for 7 days and stained with Hoechst (blue) and an antibody to β3 tubulin (green). C Normalised cellular ATP measurements from mouse embryonic (e13.5) ventral midbrain and forebrain neurons. Neurons cultured for 7 days were exposed to aripiprazole, brexpiprazole, cariprazine or piericidin A (5 μM for 18 h) (mean ± SEM from 5 independent experiments, asterisks, one-way ANOVA with Dunnett’s multiple comparison test, normalised to control). D Normalised basal and maximal OCR measurements from mouse ventral midbrain and forebrain neurons (e13.5) grown for 7 days and treated with aripiprazole, brexpiprazole, cariprazine or piericidin A (5 μM for 4 h) (corresponds to Additional file 1: Fig. S5B) (mean ± SEM from 5 independent experiments, asterisks, one-way ANOVA with Dunnett’s multiple comparison test, normalised to control)

Fig. 4

Aripiprazole causes motor dysfunction and mitochondrial defects in adult flies. A Total activity of adult flies treated with DMSO (0.5% v/v) or aripiprazole (1 mM), measured for indicated periods using a Trikinetics system (n = number of flies per test group, asterisks, two-way ANOVA with the Tukey multiple comparison test). B Climbing performance of adult flies fed with DMSO (0.5% v/v) rotenone or aripiprazole (1 mM) for indicated periods. Total number of flies tested per group are indicated on the graph (mean ± SEM from 3 independent experiments, asterisks, two-tailed unpaired t-test). C Representative TEM images of brain and thoracic muscle from adult flies treated with DMSO (0.5%) or aripiprazole (1 mM) for 28 days or rotenone (1 mM) for 5 days (arrows, examples of the scoring criteria for the morphological assessment, m = mitochondria, my = myofibrils). D Quantification of mitochondrial damage. The total number of mitochondria scored in each sample group is indicated (samples pooled from 2 flies per treatment, chi-squared analysis with Bonferroni correction)