Why do some brains seize? Molecular, cellular and network mechanisms

Authors


Email: andrew.trevelyan@ncl.ac.uk

In March, 2012, Epilepsy Research UK hosted a focused workshop addressing the basic pathophysiology of epilepsy. The particular question asked was why some brains seize, while most brains appear protected against these cataclysmic episodes. The five accompanying reviews and two research articles provide a sampler of the proceedings, focusing in turn on genetic and molecular influences, neuronal and glial cellular pathology, pathognomonic patterns of electrophysiological activity, and the different ways they impact on cortical function.

How well did we do in answering our question? We are certainly living in exciting times. Time was when we were limited to clinical observations and electrophysiological recordings, but now that we are in the post-human-genome era, with parallel explosions in animal transgenic, molecular biological, microscopic and optogenetic technologies, are the mysteries of epilepsy about to be laid bare? Perhaps not just yet. A major obstacle facing us is the heterogeneity of pathological mechanisms existing under the umbrella term, epilepsy. There is a clear need to simplify our task, by subdividing the disease. Genetics offers the most obvious classification scheme, as outlined in the review by Angela Vincent and colleagues (Lerche et al. 2013). Novel epilepsy-associated mutations continue to be described, while other genotypes may influence the efficacy of particular treatments. We are also getting mechanistic insights regarding these genes, from characterising their expression profile (e.g. interneuronal versus principal cell expression), and from functional studies in transgenic mice. An important recent development has been the recognition of a parallel set of acquired epilepsies which are caused by auto-antibodies against many of the same ion channels. But the review also highlights how our technical advances can throw up new puzzles, such as the failure to identify clear genetic associations in a large screening of single nucleotide polymorphisms in individuals with idiopathic generalized epilepsy, while finding what had been considered disease-causing mutations in other control (healthy) individuals (Klassen et al. 2011). Another area of active investigation concerns the many gene expression changes during epileptogenesis, following different brain insults. These present a confusing mixture of pro-epileptic changes with other anti-epileptic adaptations, which we are now starting to tease apart with functional studies.

Moving up to the cellular level, Pavlov et al. discussed first the dynamic nature of GABAergic neurotransmission, and how in epileptic brains, ‘GABA’ is not necessarily synonymous with ‘inhibition’ (Pavlov et al. 2013). This is a much discussed issue in the field, but the authors present the nuances of the experimental evidence in a particularly clear manner, to explain how positive shifts in GABAergic reversal potential may develop from, and then further exacerbate, hyperactivity. This is one of several important positive feedback loops that arise in epileptic tissues, others being the rise in extracellular K+ and changes in neuronal–glial interactions (see Crunelli & Carmignoto, 2013). Pavlov et al. also discuss the influence of pH, describing evidence that acidosis suppresses neuronal activity, and may be involved in terminating seizures, whereas alkalosis may cause hyperexcitability. The mechanism linking pH and neuronal excitability remains obscure, but it may yet explain clinical phenomena (e.g. the relative seizure risk in diabetic ketoacidosis versus non-ketotic states) and also suggest therapeutic avenues through metabolic and respiratory control of acid–base balance.

The role of glia in epilepsy is the topic covered by Crunelli & Carmignoto (2013). Clinical phenomenology, including the epileptogenic nature of gliomas and glial scars, has always suggested a major role for glia in epileptic pathology, although our understanding of the nature of this role has altered dramatically since the development of Ca2+ imaging technology. Prior to that, glia were thought to have an indirect influence on neuronal activity by regulating the ionic and particularly K+ balance of the extracellular space. But latterly, Ca2+ imaging has alerted us to how active glia really are, their activity mediated through receptors for, and the release of, a range of neuroactive compounds, including glutamate. The Crunelli laboratory further present evidence that raised tonic GABA currents in thalamus, associated with absences seizures, may reflect reduced GABA uptake by glia because of lowered GAT1 expression (Pirttimaki et al. 2013).

Network activity is discussed in the final two reviews, by Jiruska et al. (2013) and Trevelyan et al. (2013). Jiruska et al. focus on the complexities of electrophysiological recordings, and how we might achieve a more nuanced interpretation of the underlying neuronal firing. Clinical recordings in particular present great sampling issues, with subdural electrodes being spaced many millimetres apart, while electrodes capable of recording unit activity have tended to be placed singly. Thus, even when actual firing patterns have been recorded, this has generally been from single locations, and not necessarily representative of the true pathological activity. This though is set to change with new technological developments, aimed at clarifying the spatial patterns of activity during seizures.

Of particular importance with regard to spatial patterns is the nature of the ‘inhibitory surround’ or ‘ictal penumbra’, a key feature of epileptic activity, addressed in the research article from Carmignoto's group (Cammarota et al. 2013) and the review by Trevelyan et al. (2013). GABAA neurotransmission may vary in time and space, so the focal pathology may involve GABAergic dysfunction, co-existing with healthy GABAergic inhibition in surround territories. Carmignoto et al. make a good case that in the particular acute pharmacological model studied, the main inhibitory role is provided by parvalbumin-expressing, fast-spiking interneurons. This fits well with theoretical studies of how inhibition is manifest across the dendritic tree, and also with studies of mice models with pathology in this cell class. Moreover, strikingly similar descriptions of such activity have been made in acute, in vitro and in vivo animal models (Prince & Wilder, 1967; Dichter & Spencer, 1969; Schwartz & Bonhoeffer, 2001; Timofeev et al. 2002; Trevelyan et al. 2006), and most recently also in chronic, naturally occurring seizures, in vivo, in humans (Schevon et al. 2012). Finally, Trevelyan et al. describe how this ictal penumbra may be expected to compromise brain function during epileptic discharges, but in a very different way from the hyperactivity of the core participating territories. This will be an important clinical distinction to make, since the focal neuronal firing represents genuinely pathological activity, and the other (the surround) represents the physiological response to this.

These reviews thus present a great range of detail concerning the current state-of-the-art of epilepsy research. But like subdural recordings, they represent an impoverished sampler, partly because of the vast body of knowledge on the topic, but more importantly, because there remain many fundamental mysteries about epilepsy. One thing is clear, that young scientists just entering the field need not worry for topics to study, and novel and exciting research tools to use.

Ancillary