Immature brains don't need GABA to get ‘hyper’-excited

It is now well known that nervous system development at prenatal and postnatal stages involves activity-dependent remodelling of neural networks. Activity dependence of neural network development is a widespread phenomenon that has been observed in many different brain areas, including the thalamic nuclei, cortical sensory areas and hippocampus (Moody & Bosma 2005). During the first two postnatal weeks, activity underlies network refinement through hebbian stabilization or removal of synapses. This activity usually takes the form of spontaneous synchronized neuron firing that favours calcium entry, which is necessary for functional and morphological plasticity. One major consequence of this phenomenon is that immature networks are, by nature, more excitable than mature networks. The pathophysiological correlate of this is that immature brains are more susceptible to developing epileptic activity (Sperber et al. 1999). However, whether this hyperexcitability is mainly the consequence of particular excitability properties of individual neurons or of networks and whether it involves mainly excitatory or inhibitory neurons and/or synapses remain to be determined.

In a recent issue of The Journal of Physiology, Shao & Dudek (2009) present a very thorough study of the mechanisms underlying epileptiform activity in the CA3 area of the hippocampus at immature developmental stages. Using CA3 minislice preparations (Miles & Wong, 1983), they analysed the population bursts induced by extracellular stimulation after blockade of fast inhibitory GABAergic transmission (GABA_A receptor blockade) in immature (postnatal days 9–14) and mature preparations (postnatal days 90–100). As previously described, the authors found that the duration of these bursts is almost two orders of magnitudes longer in the immature compared to the mature hippocampus (14 s vs. 180 ms). Moreover, the burst latency is significantly shorter in immature animals compared to mature animals. Using pharmacological tools, the authors exclude the involvement of GABA_B, NMDA and metabotropic glutamate receptors and show that AMPA-mediated glutamatergic transmission is necessary and sufficient to sustain prolonged population bursts. The authors then discovered that the excitatory network has significantly different properties in the immature animal: (i) increased frequency and amplitude of recurrent excitatory synaptic events paradoxically coupled to a lower probability of release of glutamate, (ii) increased intrinsic excitability of excitatory neurons due to an increased membrane resistance and decreased medium AHP (afterhyperpolarization ocurring after a few action potentials).

The interplay between these different network properties explains why the immature CA3 is able to generate long population bursts while the mature CA3 is unable to sustain prolonged recurrent activity. In the immature CA3, when an excitatory neuron is stimulated, it induces large synaptic responses in the postsynaptic neuron, which easily fires action potentials due to its high membrane resistance. This spiking also occurs at a shorter latency than in older animals, presumably due to the shorter axons and higher prevalence of axo-axonic coupling in immature preparations (MacVicar & Dudek, 1980; Gomez-Di Cesare et al. 1997). The postsynaptic neuron is then able to quickly recurrently activate its presynaptic partner, which is able to sustain high frequencies of firing due to its almost non-existent AHP. When this presynaptic neuron releases glutamate for the second time onto the postsynaptic neuron, summation of synaptic events occurs due to synaptic facilitation (a classical consequence of the low probability of release), and therefore the likelihood that the postsynaptic neuron will fire an action potential is even higher than after the first release of glutamate. Thus, these properties create a positive feedback loop that will engage the network in an epileptiform regime of activity. In the mature CA3 the high probability of release and the longer latency of synaptic events are unfavourable to summation of synaptic events, and the lower membrane resistance and larger AHP prevent repetitive firing, such that the recurrent wave of activity will die out after a small number of cycles.

This work is particularly interesting because it underlines the importance of AMPA receptors and intrinsic excitability in the genesis of epileptiform activity in the immature hippocampal network. Over the past twenty years, an increasing interest has been taken in the role of the depolarizing actions of GABA in epileptiform activity in the immature hippocampus (until postnatal week 2) (Ben-Ari et al. 2007). The experimental paradigm Shao and Dudek used rules out the involvement of depolarizing actions of GABA in the initiation of prolonged population bursts, as all experiments were performed in the presence of GABA_A-receptor blockers. Indirectly, this work undermines the role of depolarizing GABA in the genesis of some forms of epileptiform activity in the hippocampus. This timely study echoes the recent results obtained by Rheims et al. (2009) who demonstrated that depolarizing actions of GABA in immature networks may have been overestimated in most studies due to differences in brain metabolism between immature and mature animals, which were usually not taken into account in experimental paradigms. These authors showed that accounting for these metabolic differences significantly diminishes the depolarizing effect of GABA in young networks. Taken together, these two studies support the idea that at least some forms of epileptiform activity observed in the young hippocampus might rely on increased synaptic and intrinsic excitability of excitatory CA3 neurons, rather than on particular properties of GABA neurotransmission.