Slamdance: seizing a fly model for epilepsy

Shanker Karunanithi, Bruno van Swinderen

brain seizures, as in epilepsy, are characterized by massively synchronous firing of neurons. Given the extraordinary complexity of the human brain, it is surprising that epilepsy is not a more common complaint. Episodic brain seizures currently affect ∼1% of the human population, and causes for the syndrome range from hereditary effects to brain injuries and infectious diseases (Berg 2011). To better understand seizures and how they may be prevented requires access to animal models where behavior and neurophysiology can be easily studied together, ideally in mutant strains relevant to human epilepsies. A recent study by Marley and Baines (2011) presents just such a model using the fruit fly Drosophila melanogaster, and in the process, this study provides evidence for correction mechanisms in two different neurons in response to seizure-inducing mutations or anti-epileptic drug (AED) therapy.

Drosophila has been well established as a model to study epilepsy, although the fly has been until recently less revealing about neural currents underpinning epileptic phenotypes. In early behavioral screens of Drosophila, Seymour Benzer (1971) noted that a number of mutant strains were transiently knocked out following mechanical stimulation, and these mutants were termed “bang-sensitive (b-s)”. Genetic identification of several of these mutations revealed defective ion channels and synaptic machinery (George 2005; Song and Tanouye 2008) that likely induce seizures in the fly. Limited access to brain electrophysiology in the fly has prevented a better understanding of how these mutations lead to epileptiform activity (Glasscock and Tanouye 2005; Pavlidis et al. 1994). However, Marley and Baines (2011) provide a fresh insight by recording from larval neurons in the brains of slamdance (sda) b-s mutants. Voltage-gated sodium currents are compromised in sda mutants due to a defective aminopeptidase pathway (Zhang et al. 2002), producing an epileptic phenotype in fly larvae following electric shock. Mean behavioral recovery times following seizure episodes in sda flies are over five times longer than for wild-type animals, mimicking similar effects in adults following mechanical shock. Recordings from two motor neurons (termed aCC and RP2) reveal increased amplitude and duration of spontaneous rhythmic synaptic currents in the mutants, and this appears to be associated with altered sodium currents. Whole cell voltage clamp recordings from aCC and RP2 neuronal cell bodies revealed two components of membrane sodium currents, the transient (INat) and persistent (INap) components, produced by different conductance states of the same DmNav channel (Lin et al. 2009) (Fig. 1A). However, only the smaller persistent current was increased in sda mutants, whereas the transient current remained unchanged (Fig. 1A). INap has previously been shown to be relevant in human epilepsy, and is a potential target for drug therapies (George 2005; Segal and Douglas 1997). Hence, an effect of sda on this current component already suggests that this mutant is a potentially suitable fly model for epilepsy.

Fig. 1.

A: shown are the persistent (INap) and transient (INat) current components of the DmNav sodium current. slamdance (sda) currents are shown in gray and wild-type (WT) in black. Scale bar, 10 mV/100 ms. B: the INap/INat ratio is used as a readout of seizure susceptibility. Seizure susceptibility in WT animals is below seizure threshold, whereas in sda mutants, seizure susceptibility is above the seizure threshold. However, 24-h feeding of phenytoin (PHY) to sda mutants lowers seizure susceptibility to below threshold. C: in progeny of gravid seizure-prone sda females, seizure susceptibility is below threshold and near WT levels. In aCC neurons, INap is reduced, whereas in RP2 neurons, INat is increased. In both instances, the INap/INat ratio is below seizure threshold. D: negative feedback control of neural circuit output. The actual circuit output will match the desired output when the INap/INat ratio for all neurons is set to be equal. The error (or difference) signal is generated so long as the INap/INat ratio is different for all neurons in the circuit. However, once the ratio is equal for all neurons, no error signal is generated and the actual circuit output matches the target (desired) circuit output.

Since INap has been a target for drug intervention in human epilepsy therapy, it was hypothesized that AEDs might correct seizure-like symptoms in sda mutants. Indeed, feeding the mutants phenytoin (an AED) was found to dramatically reduce mean recovery times to wild-type levels. This behavioral effect of the drug was associated with reduced INap currents in the mutant while INat remained unchanged. As further evidence linking the persistent sodium current with seizure-like effects in fly larvae, another drug (rATXII) was used to potentiate INap in wild-type flies. Above an INap/INat ratio of ∼35%, increasing INap was found to promote seizure-like behavior in larvae, whereas below this threshold larvae displayed wild-type responses to electric shock. Together with the observations in sda mutants (Fig. 1B), these results firmly implicate the persistent sodium current (INap) as primarily responsible for epileptic behavior in this fly model.

Although effective, anti-epileptic drugs can cause developmental defects. It has been reported that some children born to epileptic mothers who were administered AEDs throughout their pregnancies had impaired cognitive functions (Hanson et al. 1976; Holmes 2009), and genetic predisposition to seizures was still evident in these children as well (Hanson et al. 1976; Holmes 2009). Marley and Baines (2011) conducted an experiment in Drosophila to determine the effects on progeny of feeding AEDs to gravid seizure-prone sda females. Gravid sda females were fed phenytoin, which decreases INap and prevents seizures in those animals, as discussed above. The feeding regime allowed subthreshold amounts of phenytoin to be transferred to the eggs. However, the eggs laid by these mutant females, and the larvae that subsequently hatched, were not exposed to the drug. Instead of observing defects in the offspring (termed sda-treated animals), the authors observed significantly reduced seizure susceptibility in the next generation, and this was associated with a lower INap/INat ratio. Similar drug treatments in rodents have indicated some positive outcomes on offspring, but not to the extent observed here (Blumenfeld et al. 2008). Therefore, why such a clear-cut result in Drosophila? Marley and Baines (2011) hypothesize that imbalances in excitation and inhibition in early stages of neural development could lead to seizures. Early pharmaceutical intervention by administering phenytoin may correct for this imbalance or prevent the imbalance from occurring. The simpler nervous system in Drosophila may be less susceptible to side effects caused by AEDs that are otherwise observed in humans and rodents. For these reasons, the sda mutant may be a promising model for investigating the beneficial effects of administering AEDs preventively.

Lowering neuronal excitability reduces seizure susceptibility. A reduction in sodium channel conductance is associated with lowering both neuronal excitability and seizure susceptibility (George 2005; Segal and Douglas 1997). The lower INap/INat ratio observed in sda-treated larvae compared with seizure-prone sda larvae (Fig. 1B) supports this idea. The authors investigated how this correction was achieved by examining sodium currents in the aCC and RP2 motor neurons. The two motor neurons are structurally, functionally, and electrically different (Schaefer et al. 2010). Therefore, it was of interest to determine whether the two neuron types in sda-treated progeny adapt differently or similarly following phenytoin treatment of their mothers. Surprisingly, the neurons seem to adapt differently to achieve the same result, a lower INap/INat ratio. In RP2 neurons, INat is increased without affecting INap, and in aCC neurons, INap is reduced without affecting INat (Fig. 1C); both changes equivalently reduce the INap/INat ratio. Hence, reduction of excitation in these two neuron types is seemingly cell type specific, achieved by altering different conductance states of the DmNav Na+ channel. Previous work suggests that alternative splice variants of the DmNav Na+ channel gene may account for differences in INap and INat in different neurons (Lin et al. 2009). Hence, the altered INap and INat in epilepsy models may be a consequence of altered DmNav splicing during development. By the same token, exposure to AEDs such as phenytoin during early development might alter DmNav gene expression and splicing in embryonic neurons to potentially lower seizure susceptibility in later life.

It is remarkable how the two neuron types in sda-treated animals adapt so differently in lowering seizure susceptibility. What appears even more surprising is that the INap/INat ratios are maintained at the same level following these adaptations: aCC: 31.4 ± 2.7 (n = 8); RP2: 25.8 ± 2.9 (n = 7) (the ratios are not significantly different, P = 0.18). Maintaining a similar INap/INat ratio appears to be built into the circuit; the ratio for the two neuron types are not different for wild-type (aCC: 24.8 ± 5.0, n = 5; RP2: 16.9 ± 2.9, n = 3; P = 0.30) or sda mutants (aCC: 45.4 ± 3.3, n = 5; RP2: 44.0 ± 9.6, n = 2; P = 0.86) (Marley and Baines, personal communication). Therefore, the conservation of the INap/INat ratios for the two neuron types indicates that their activities are intrinsically linked. In sda-treated animals, changes to INap in aCC and INat in RP2 may not be coincidental, but coordinated. Thus, changes in the activity levels in one neuron may also produce changes in the other.

Why do aCC and RP2 neurons possess the same INap/INat ratios? To answer this question, it is important to bear in mind that aCC and RP2 motor neurons innervate a common target muscle (muscle 1; although RP2 also innervates other targets) (Hoang and Chiba 2001), and, therefore, belong to the same neural circuit (Fig. 1C). Hence, changes in the properties of any one element in the circuit likely affect other elements of that circuit (Schulz et al. 2006). According to homeostatic theory, neurons possess the capacity to tune their activities to target levels through alterations in their intrinsic excitability and/or synaptic drive (Grashow et al. 2010; Marder and Goaillard 2006). Alteration to intrinsic excitability is thought to involve readjustments that produce a balance among different conductance densities to support the target level of activity (Marder and Goaillard 2006; Schulz et al. 2006; Tobin et al. 2009). aCC and RP2 neurons acquiring the same INap/INat ratios in sda-treated animals may be part of such readjustments in the fly larvae. Acquiring the same INap/INat ratio among neurons within the circuit may involve negative feedback mechanisms (Davis 2006; Marder and Goaillard 2006). Once the same INap/INat ratio is achieved among neurons within the circuit, the circuit may be able to produce the target output level (Fig. 1D).

The study by Marley and Baines (2011) demonstrates that the slamdance mutant at the larval developmental stage is an excellent model for studying the underlying pathophysiology of seizure susceptibility. Similar dysfunctions in sodium channels associated with seizure susceptibility are also seen in humans and rodents. However, investigating the underlying mechanisms in Drosophila has advantages because of the ease of gene manipulation in this insect, low costs associated with both the care of animals and behavioral screening, and a simpler nervous system that reduces the complexity in uncovering the underlying pathophysiology of seizure. Other major advantages in working with Drosophila larvae are: 1) that the motor neurons and the targets they innervate are well characterized (Hoang and Chiba 2001); 2) temporal and spatial control of transgene expression of identified neurons and their targets is possible (Nicholson et al. 2008); 3) signals can be recorded from identified motor neuron cell bodies (this work) and their synapses (Chouhan et al. 2010; Karunanithi et al. 2002) with relative ease; and 4) the behavioral consequences of gene manipulations and its link to neurophysiology can be assayed (Marley and Baines 2011). A combination of these techniques can now be used to address, for example, how differential regulation of DmNav expression and splicing might ensure homeostasis within a circuit.

The clear demonstration that seizure susceptibility is significantly reduced in offspring of gravid seizure-prone slamdance females fed with phenytoin is a promising result. It offers hope to those with a genetic predisposition to seizure, that pharmaceutical intervention by administering AEDs in early development could potentially suppress seizure susceptibility in later life. The additional observation that different neuron types adapt in their own way to maintain an appropriate sodium current ratio demonstrates that the adaptations are homeostatic, likely constraining neuronal firing to target levels. This mechanism promotes circuit stability and probably decreases the likelihood of seizure.


B. van Swinderen is supported by the Australian Research Council and S. Karunanithi by a Charles Sturt University Competitive Grant.


No conflicts of interest, financial or otherwise, are declared by the author(s).


We thank Drs. Richard Baines and Richard Marley for providing additional data and discussions throughout the writing of this Editorial Focus.