The ins and outs of interneurons in epileptic neocortex.

Commentary Does impaired inhibition cause epilepsy? Does epilepsy alter inhibitory circuits? These longstanding questions have few clear answers. A recent study by Jin et al. provides a fascinating view of the inputs and outputs of an inhibitory circuit in chronically epileptic neocortex and highlights the complexity of circuit reorganization following injury. The choice of experimental model is an important strength of this investigation. Brain trauma is a common cause of human epilepsy, and the cellular mechanisms linking injury and seizures are obscure. For nearly 2 decades, Prince and his colleagues have skillfully and doggedly investigated a rodent model of chronic injury—the undercut neocortex—using physiological, anatomical, and molecular techniques (1). Partial surgical isolation of a cortical region leads, after a delay of about 10 days, to ongoing epileptiform activity. Paroxysms are accompanied by a wide variety of changes in structure and function. The cortex thins. Pyramidal cells decrease in number and size, but the remaining cells increase their intrinsic excitability. Axons of pyramidal cells sprout, and electrophysiology suggests that this increases their excitatory synaptic connections. Hyperconnected, hyperexcitable pyramidal cells sound like a recipe for seizure activity, but inhibitory systems of neurons are also altered by chronic injury (2). The notion that dysfunctional synaptic inhibition is involved in human seizure disorders is at least as old as the identity of inhibitory neu-rotransmitters themselves (3). The inhibition hypothesis has much to recommend it. Normal activity in the cerebral cortex reflects a constantly shifting balance of synaptic excitation and inhibition (4). Reducing inhibition, even modestly, is one of the simplest ways to induce seizures (5). Increasing inhibition, for example by transplanting inhibitory interneurons, can ameliorate seizures (6), and a wide variety of genetic mutations that impair inhibition have seizures as a phenotype (7). Indeed, earlier studies from the Prince group suggested that pyramidal neurons in the undercut cortex receive lower rates of spontaneous inhibitory inputs compared with uninjured neurons (8). Other measures of inhibition showed mixed effects after chronic injury, however; numbers of inhibitory interneurons and synapses were stable, but their dendrites and axons were thinner and shorter, and their synapses were smaller (9). This set the stage for the experiments of Jin et al. They knew that injured pyramidal cells in layer 5 sprout new axons in response to injury; but would that lead to increased excitation of pyramidal cells and inhibitory interneurons? If so, would the additional excitation of the two cell …