Development of a Drosophila seizure model for in vivo high-throughput drug screening Geoff E. Stilwell, 1 Sudipta Saraswati, 2 J. Troy Littleton 2 and Scott W. Chouinard 1 1 Cambria Biosciences, 8A Henshaw St., Woburn, MA 01801, USA 2 The Picower Institute for Learning and Memory, Department of Biology and, Department of Brain and Cognitive Sciences, Massachusetts Institute of Technology, Cambridge, MA 02139, USA Keywords: anti-epileptic drugs, Drosophila melanogaster, epilepsy research, GABA A receptor, neuronal excitability, seizure model Abstract An important application of model organisms in neurological research has been to identify and characterise therapeutic approaches for epilepsy, a recurrent seizure disorder that affects > 1% of the human population. Proconvulsant-treated rodent models have been widely used for antiepileptic drug discovery and development, but are not suitable for high-throughput screening. To generate a genetically tractable model that would be suitable for large-scale, high-throughput screening for antiepileptic drug candidates, we characterized a Drosophila chemical treatment model using the GABA A receptor antagonist picrotoxin. This proconvulsant, delivered to Drosophila larvae via simple feeding methods suitable for automated screening, generated robust generalised seizures with lethality occurring at doses between 0.3 and 0.5 mg ⁄ mL. Electrophysiological analysis of CNS motor neuron output in picrotoxin- treated larvae revealed generalised seizures within minutes of drug exposure. At subthreshold doses for seizure induction, picrotoxin produced an increased frequency of motor neuron action potential bursting, indicating that CNS GABAergic transmission regulates patterned activity. Mutants in the Drosophila Rdl GABA A receptor are resistant to picrotoxin, confirming that seizure induction occurs via a conserved GABA A receptor pathway. To validate the usefulness of this model for in vivo drug screening, we identified several classes of neuroactive antiepileptic compounds in a pilot screen, including phenytoin and nifedipine, which can rescue the seizures and lethal neurotoxicity induced by picrotoxin. The well-defined actions of picrotoxin in Drosophila and the ease with which compounds can be assayed for antiseizure activity makes this genetically tractable model attractive for high-throughput in vivo screens to identify novel anticonvulsants and seizure susceptibility loci. Introduction Epilepsies are circuit-based seizure syndromes caused by the patho- logical synchronized firing of subpopulations of central nervous system (CNS) neurons. To characterise the effects of epilepsy on neuronal function, multiple animal models have been generated using genetic manipulations, chemical treatments or physical perturbations to induce seizures (Sarkisian, 2001; Upton & Stratton, 2003; White, 2003). Rodents treated with pentylenetetrazol (PTZ) or picrotoxin (PTX) have been widely used to study seizure physiology, to identify susceptibility genes and to screen and test antiepileptic drugs (AEDs; White et al., 2002; Yang & Frankel, 2004). Invertebrate model organisms such as Drosophila melanogaster have also been central to the identification of genes important in epilepsy and nervous system function. In particular, channelopathy mutants, including seizure, slowpoke and ether-a-go-go, were characterized in Drosophila based on their temperature-sensitive (TS) paralytic or ether-induced leg- shaking phenotypes (Ganetzky & Wu, 1986; Ganetzky, 2000). Drosophila has also emerged as a potentially useful model for the analysis of seizure susceptibility and AED action with the bang- sensitive (BS) mutants (Kuebler & Tanouye, 2002; Lee & Wu, 2002; Reynolds et al., 2004; Tan et al., 2004). These mutants show electrophysiological responses similar to electroshock-induced sei- zures (Engel & Wu, 1994; Pavlidis & Tanouye, 1995; Lee & Wu, 2002). Although both BS and TS mutants have been the major Drosophila epilepsy models characterized thus far, each model has drawbacks. For instance, the unknown effects of temperature shifts on nervous system function in the TS mutants have made a direct comparison to vertebrate seizure models difficult. Similarly, although BS phenotypes appear behaviourally and physiologically similar to seizures evoked in mammalian models, most cloned BS loci do not correspond to known mammalian genes involved in seizure suscept- ibility or epilepsy. To help circumvent these difficulties, we have characterized an in vivo Drosophila model using seizure-inducing compounds acting via defined c-aminobutyric acid (GABA) receptor pathways. GABA A receptors are pentameric, ligand-gated ion channels that mediate inhibitory synaptic transmission in the CNS. (Mohler et al., 1997; Jones-Davis & Macdonald, 2003). GABA receptor modulation is central to seizure generation, status epilepticus and chronic epilepsy in mammals (Homanics et al., 1997; DeLorey et al., 1998; Cossette et al., 2002; Jones-Davis & Macdonald, 2003). Multiple proconvul- sants, including PTZ and PTX, act as GABA A receptor antagonists (Dibas & Dillon, 2000; Huang et al., 2001; Jones-Davis & Macdonald, 2003). Mutations within GABA receptors have also been found in human epilepsy syndromes and are the basis for several murine seizure models (Cossette et al., 2002; Jones-Davis & Macdonald, 2003). Similarly, many classes of therapeutics enhance GABA currents and are used for treating multiple types of epilepsy Correspondence: Dr Geoff Stilwell, as above. E-mail: gstilwell@cambriabio.com Received 27 March 2006, revised 21 June 2006, accepted 29 July 2006 European Journal of Neuroscience, Vol. 24, pp. 2211–2222, 2006 doi:10.1111/j.1460-9568.2006.05075.x ª The Authors (2006). Journal Compilation ª Federation of European Neuroscience Societies and Blackwell Publishing Ltd