Background
A major challenge in current neurodegeneration research is the identification of effective therapies. Over recent years, simple model organisms, such as the nematode worm
Caenorhabditis elegans, have been increasingly recognised as powerful systems for revealing the conserved molecular mechanisms that underlie neurodegeneration [
1]. Indeed various laboratories have developed and characterised a diverse set of
C. elegans models of various human neurodegenerative diseases, including Alzheimer’s [
2], Parkinson’s [
3] and polyglutamine expansion diseases [
4]. Genetic screens performed in these models have identified a variety of genes that can suppress or increase disease progression and are thus potential therapeutic drug targets. However, relatively few of these genetic modifiers are common to more than one disease model, despite the shared feature of protein misfolding/aggregation [
5,
6].
Complementary to its utility for genetic screens,
C. elegans is a useful pharmacological model for testing potential neuroprotective compounds. Attention has mainly focused on screening existing FDA-approved medications rather than novel compounds, as repurposing of drugs pre-approved for other indications obviates the need for early toxicity trials and thus expedites translation to clinical testing [
7,
8]. For example,
C. elegans Alzheimer’s models expressing human Aβ
1–42 have identified neuroprotective effects of several approved compounds, including antibiotic, antidepressant and antihypertensive drugs [
9]. Similarly, dopamine D2 receptor antagonists have been shown to ameliorate mutant tau-induced functional defects and reduce aggregation in a frontotemporal dementia with parkinsonism-17 (FTDP-17) tauopathy model [
10]. A wide variety of other neuroprotective compounds have also been identified in chemical screens using worm neurodegeneration models including spinal muscular atrophy [
11], Parkinson’s [
12] and Huntington’s diseases [
13].
Most compounds identified in
C. elegans chemical screens to date are effective in only a single neurodegenerative model, suggesting that translational potential may be disease-specific. However, some compounds, such as resveratrol, have been shown to be protective in a range of worm models and also in mammalian systems [
14‐
18]. This demonstrates that it is possible to identify generally neuroprotective compounds that alleviate the functional consequences of protein misfolding common to neurodegeneration. Here we report that ethosuximide, a widely prescribed anti-epileptic drug, improves the phenotypes of multiple neurodegenerative disease models and we reveal a conserved action of the drug in modulating DAF-16/FOXO target gene expression in worms and mammalian neurons.
Discussion
Neurodegenerative diseases are increasingly common and exert large costs on society. However, no disease-modifying therapies for these devastating disorders are currently available. Hence, there is considerable current interest in the idea of repurposing existing medicines for the treatment of neurodegeneration [
8]. Drugs that can mitigate the impact of common pathological processes such as protein misfolding/aggregation that underlie multiple neurodegenerative diseases could be especially useful therapeutics. Ethosuximide may be a candidate for such a repurposing approach. We have shown here that ethosuximide ameliorates the phenotypes of two distinct worm neurodegenerative disease models based on deletion of an endogenous neuroprotective gene (
dnj-14) and pan-neuronal expression of a disease-associated mutant Tau protein. Furthermore, an independent study has recently reported a neuroprotective effect of ethosuximide against human mutant mTDP-43-mediated proteotoxicity in a transgenic
C. elegans model of amyotrophic lateral sclerosis [
15]. Importantly, we have discovered that ethosuximide also reduces polyglutamine protein aggregation in mammalian neuroblastoma cells and acts by modulating DAF-16/FOXO target gene expression in both worms and mammalian neurons. These evolutionarily conserved effects on so many different neurodegenerative disease models suggest that ethosuximide merits consideration for therapeutic applications in patients. Indeed, given the difficulty of distinguishing between tau- and TDP-43-associated frontotemporal dementia in the clinic, a drug with potential to protect against both pathologies could be particularly useful. Future experiments in mouse neurodegeneration models should therefore be given a high priority in order to validate ethosuximide’s repurposing potential for treatment of human neurodegenerative diseases.
The mechanism by which ethosuximide exerts its anti-epileptic action is unclear and controversial. Ethosuximide has long been thought to act by blocking T-type calcium channels; however, more recent work has suggested that actions on other voltage-sensitive ion channels, such as sodium and potassium channels, are involved in the therapeutic effect of ethosuximide [
30]. Our finding that the beneficial effects of ethosuximide on motility and longevity in a frontotemporal dementia model persist in strains harbouring a null mutation in the
C. elegans T-type calcium channel,
cca-1, suggests that ethosuximide’s neuroprotective activity is not mediated by inhibition of T-type channels. This conclusion is supported by previous work in
C. elegans showing that lifespan extension by ethosuximide is unaffected by a different mutant allele of
cca-1 [
39]. Although we cannot rule out redundancy with other channel subunit homologs that share a putative ethosuximide binding site, given that the
C. elegans genome does not encode voltage-sensitive sodium channels [
40], it seems likely that ethosuximide ameliorates neurodegeneration via a distinct mechanism.
Ethosuximide has previously been suggested to extend lifespan in wild type worms via inhibition of chemosensory neurons [
39,
41], resulting in a perceived state of dietary restriction and hence lifespan increase [
39]. However, ethosuximide’s ability to extend lifespan persists even under maximal dietary restriction conditions [
42], thus arguing against this mechanism. In our study, lifespan in wild type control worms was unaffected by ethosuximide, but nevertheless the protective effect of ethosuximide in worm neurodegeneration models is unlikely to involve inhibition of chemosensory neurons, as the
dnj-14 model used in this study already exhibits profound chemosensory impairment, which is actually rescued by ethosuximide.
The transcriptomic analysis reported here suggests instead that ethosuximide acts by modulating the expression of DAF-16/FOXO target genes. DAF-16/FOXO is a conserved transcription factor acting downstream of DAF-2 in the IIS pathway [
36]. Reducing IIS pathway activity by inhibiting DAF-2 causes an increase in DAF-16 activity and the consequent modulation of DAF-16 target gene expression leads to phenotypic changes such as increased lifespan. In addition to its pivotal role in longevity and stress resistance, reduced IIS pathway activity has been shown to confer neuroprotection in nematode neurodegeneration models based on expression of Aβ [
43], TDP-43 [
44] and polyQ [
45]; and also in mouse Aβ models [
46]. Our finding that the neuroprotective effect of ethosuximide
in vivo was dependent on DAF-16 function is consistent with these studies and also with our transcriptomic analysis. It is also consistent with the recent observation that ethosuximide’s protective effect on a
C. elegans amyotrophic lateral sclerosis model requires DAF-16 [
15], but contrasts with earlier work showing that ethosuximide still increases lifespan in
daf-16 mutants [
41]. Although we observed little effect of ethosuximide on
skn-1 pathway targets or on oxidative stress detoxification genes such as catalase, superoxide dismutases and peroxiredoxins, genes that are commonly regulated by DAF-16 and oxidative stress showed some enrichment (Additional file
9: Figure S7), which may be relevant given the established links between oxidative stress and protein aggregation in neurodegenerative disease models. Ethosuximide may mimic some of the effects of reduced IIS by inducing the expression of neuroprotective DAF-16 target genes, in particular those containing the DAE motif [
47].
Analogous to its ability to regulate DAF-16 target gene expression and modulate protein aggregation in
C. elegans, we found that ethosuximide induces transcription of FOXO target genes and confers protection against polyQ aggregation in mammalian neuronal cell lines. This therefore suggests an evolutionary conservation of DAF-16/FOXO-regulated ethosuximide responses. The DAF-16 homologue FOXO3a is highly expressed in adult brain, and plays key regulatory roles in neuronal survival under basal, stress and disease conditions [
48,
49]. Furthermore, FOXO3a contributes to neuroprotection by Sirt1 in a striatal cell model of Huntington’s disease [
50] and moderate FOXO3 activation was recently shown to oppose α-synuclein accumulation and proteotoxicity [
51].
Ethosuximide is cheap to manufacture, has been widely prescribed for several decades, is well tolerated at high doses and has a good safety record – indeed, it is often used in children to treat absence seizures [
52]. Furthermore, it has >95 % bioavailability and freely crosses the blood-brain barrier, with levels of the drug in cerebrospinal fluid and multiple brain regions being similar to those found in plasma [
52]. Optimal improvement in the
C. elegans models described here was seen at an externally applied concentration of 1-2 mg/ml, which equates to a measured internal concentration within the worm of 15-30 μg/ml [
41]; while effects in mammalian neurons were seen at 0.5-1 mg/ml. These values are close to the therapeutic dose range of ethosuximide for epilepsy in humans (40-100 μg/ml) [
52], suggesting that neuroprotective doses could be achieved in patients. Ethosuximide may therefore be a candidate drug for repurposing as a treatment for neurodegenerative diseases.
Competing interests
The authors declare that they have no competing interests.
Authors’ contributions
XC, HVM, SW and SSK performed the experiments. XC, HVM, BK, JWB, RDB and AM analysed and interpreted the data. BK provided essential reagents and materials. AM, RDB and JWB conceived and designed the experiments. AM and XC wrote the manuscript with input from all of the other authors. All authors read and approved the final manuscript.