Benign neonatal convulsions and spontaneous network activity in the developing brain: is there a link?

Abstract

A striking feature of the immature brain is that spontaneous patterned activity can occur in the absence of afferent activity. In the developing retina, hippocampus, somatosensory cortex and elsewhere, circumstantial evidence strongly suggests that precisely timed and co-ordinated firing of many neurons contributes to establishing the initial synaptic connections among populations of neurons, thereby building a scaffold for subsequent information processing and experience-dependent learning. Although it is well established that spontaneous network activity fades with development, precisely why these bursts of discharges cease is not known. In this issue of The Journal of PhysiologySafiulina et al. (2008) identify a new player in the developmental regulation of spontaneous hippocampal activity, which may fortuitously shed light on why certain neonatal seizure syndromes also resolve after a few weeks of life. Of the various neonatal seizure disorders, perhaps the best understood from a mechanistic viewpoint is benign familial neonatal convulsions (BFNC). This dominantly inherited syndrome is caused by loss-of-function mutations of either KCNQ2 or KCNQ3, which encode Kv7.2 and Kv7.3, respectively (Jentsch, 2000). These K+ channel subunits coassemble to give rise to hetero-tetrameric channels which exhibit slow activation upon modest degrees of depolarization, and which are inhibited by muscarinic receptors through an incompletely understood intracellular cascade (Delmas & Brown, 2005). The discovery of these channels, which followed the positional cloning of candidate genes at two chromosomal loci linked to BFNC, actually resolved a long-standing puzzle as to the identity of the so-called ‘M’ conductance known to be modulated by muscarinic agonists. Because this current normally helps to keep neurons at rest, loss-of-function KCNQ2 or KCNQ3 mutations can be expected to lead to excessive firing and thus explain the associated propensity to seizures. However, unlike many other epileptic channelopathies, seizures in BFNC tend to remit spontaneously. Why? Although Safiulina et al. do not address BFNC directly, their study raises the possibility that the spontaneous remission seen in this disorder is related to a developmental up-regulation of M current density. They examine giant depolarizing potentials (GDPs) in the CA3 hippocampal subfield, a form of spontaneous network activity that can be studied in acute rodent brain slices in vitro (Ben-Ari et al. 1989). GDPs require both glutamate and GABA receptors, and it is tempting to speculate that their disappearance after the first postnatal week is related to a gradual hyperpolarization of the GABAA receptor reversal potential, as the Cl− ion exporter KCC2 starts to be expressed (Rivera et al. 1999). Removal of a GABAergic depolarizing drive to pyramidal neurons could thus diminish circuit excitability. The results of Safiulina et al., however, invoke a second player in the developmental disappearance of GDPs. They show that Kv7.2 expression and M current density both increase over the first 3 weeks of life, and that this is accompanied by a shift in activation threshold to more negative potentials as well as an acceleration of kinetics. All these effects are predicted to decrease the intrinsic excitability of neurons, which may interact with synaptic and non-synaptic forms of communication to affect the occurrence of GDPs. Indeed, Safiulina et al. also show that, although the muscarinic agonist linopirdine has little effect on neonatal neurons, it enhances spontaneous burst firing of older neurons and facilitates the occurrence of GDPs. Conversely, GDPs in neonatal neurons are abolished by the M current activator retigabine. Overall, these results show that intrinsic bursting and GDPs are exquisitely sensitive to the M current, and suggest that the maturation of the M current may contribute to the disappearance of GDPs. In common with other emergent phenomena, GDPs reflect the interplay of many phenomena, and so the two hypotheses to explain their disappearance (an increase in M current and a hyperpolarization of the GABAA reversal potential) are not mutually exclusive. Nevertheless, by drawing attention to the role of Kv7 channels in regulating GDPs, Safiulina et al. also raise the possibility that seizures in BFNC resolve with development because of a similar increase in M current expression. The mutations associated with the disorder are predicted to give only a modest decrease in Kv7/M current density (Schroeder et al. 1998), again arguing for an exquisite sensitivity of the immature brain to this K+ conductance. Although BFNC rarely requires drug treatment, the availability of a relatively selective Kv7 channel opener (retigabine) offers an almost unique tool to reverse the genetic defect. A phase III trial of retigabine as an add-on drug in localization-related epilepsy has recently been completed, although the outcome is not yet known.