Axo-axonic cells (AACs) are interneurons that constitute the sole inputs to the axon initial segment of pyramidal cells and do not innervate other cell types. They are traditionally thought to provide only inhibitory input that is mediated by GABA (γ-aminobutyric acid). However, writing in Science, Szabadics and co-workers report a surprising role for these cells: they can trigger GABA-mediated excitatory as well as inhibitory postsynaptic responses in adult pyramidal cells.

As action potential initiation is most likely to occur in axons, these synapses are in a prime position for influencing neuronal output. In the mature cortex, the GABA-mediated inhibitory effects of AACs on the postsynaptic membrane depend on the potassium chloride co-transporter 2 (KCC2), which promotes a net outflow of Cl, thereby reducing the intracellular concentration of Cl. Under normal circumstances, activation of ionotropic GABAA (GABA type A) receptors increases the membrane permeability to Cl, which causes a net flow of Cl into the cell, resulting in hyperpolarizing responses.

To study the function of AACs in the mature cortex, Szabadics and colleagues used high-resolution immunolocalization to reveal the distribution of KCC2 and its influence on the polarity of the postsynaptic response in rat and human layer 2/3 pyramidal cells. In both the rat and human neocortex, there was a substantially lower distribution of KCC2 on the axonal initial segments compared with the somatic and dendritic membranes, and the cytoplasm of pyramidal cells. The higher intracellular Cl concentrations associated with reduced KCC2 expression is likely to support GABAA-mediated outflow of Cl, thereby promoting depolarizing responses.

To test this possibility, these researchers compared the effects of GABA-mediated AAC input to axon initial segments, and basket cell input to perisomatic regions of pyramidal cells in layer 2/3 rat somatosensory cortex. Basket cell input resulted excusively in hyperpolarizing responses that led to inhibitory postsynaptic potentials. By contrast, AACs gave rise to a relatively depolarized response, thereby increasing the likelihood of generating action potentials. Indeed, GABA-mediated AAC input triggered postsynaptic action potentials in the axon initial segments of two pyramidal cells that could not be achieved with basket cell input at any of their target sites.

These findings were confirmed in rat and human supragranular cortical layers in which single spikes generated a series of inhibitory postsynaptic potentials that in some cases were followed by longer lasting disynaptic excitatory postsynaptic potentials.

These results highlight for the first time a dual role for AACs in triggering excitatory as well as inhibitory responses in cortical microcircuitry, and pin-point a mechanism by which GABA-mediated input can have an excitatory effect on postsynaptic cells in the adult brain. It is hoped that future work will shed light on the function of this excitatory drive for pyramidal cells.