Synaptic transmission is a complex process comprised of several steps. These include the arrival of action potentials at presynaptic terminals, the activation of presynaptic Ca2+ channels, the binding of Ca2+ ions to the sensors of exocytosis, the fusion of synaptic vesicles with the presynaptic plasma membrane, the release of transmitter into the synaptic cleft, and ultimately the activation of postsynaptic receptors. Despite this large number of steps, synaptic transmission is generally believed to be very rapid. At many synapses, and at physiological temperature, the entire sequence of events takes place in < 1 ms. In particular, one subtype of GABAergic cells, the fast-spiking, parvalbumin (PV)-expressing interneuron, releases the neurotransmitter very rapidly and with high temporal precision (Kraushaar & Jonas, 2000).
Other types of synapses release neurotransmitters more asynchronously, especially during and after trains of stimuli (Barrett & Stevens, 1972; Goda & Stevens, 1994; Atluri & Regehr, 1998). In particular, asynchronous release is very prominent at the output synapses of hippocampal interneurons containing the neuropeptide cholecystokinin (CCK) (Hefft & Jonas, 2005; Daw et al., 2009; Karson et al., 2009). Thus, whereas synchrony of transmitter release is a hallmark property of transmission at PV-interneurons, asynchrony of release characterizes CCK-interneurons.
In addition to these differences in the time course of neurotransmitter release, CCK-interneurons differ from PV-interneurons in several ways. Whereas PV-interneurons exclusively use P/Q-type Ca2+ channels for transmitter release, CCK-interneurons rely on N-type Ca2+ channels (Hefft & Jonas, 2005). Furthermore, whereas PV-interneurons have presynaptic terminals endowed with M2 muscarinic acetylcholine and μ opioid receptors, the terminals of CCK-interneurons express cannabinoid (CB1) receptors (Freund & Katona, 2007). Importantly, CB1 receptors situated on the presynaptic terminals of CCK-interneurons mediate depolarization-induced suppression of inhibition (DSI), a form of short-term synaptic plasticity induced by depolarization of postsynaptic cells (Pitler & Alger, 1994; Wilson et al., 2001). This depolarization induces endocannabinoid synthesis and release from the postsynaptic cell, leading to activation of CB1 receptors, which transiently blocks transmitter release from the presynaptic terminals.
Asynchronous GABA release was originally reported at output synapses of CCK-interneurons on principal cells (Hefft & Jonas, 2005). Whether asynchronous release also occurs at connections between CCK-interneurons and other interneurons has remained unclear. Three recent publications shed light on this question, using paired recordings between synaptically connected neurons. Daw et al. (2009) examined synapses between CCK-interneurons in the hippocampal CA3 and CA1 region. Karson et al. (2009) demonstrated asynchronous release at synapses between CCK-interneurons and PV-interneurons in CA1. Finally, in this issue of EJN, Ali & Todorova (2010) studied synapses between CCK-interneurons in the stratum lacunosum moleculare-radiatum (LM-R) of the CA1 subfield, a region highly enriched in CCK-immunoreactive cells. The findings described in all three papers converge to support the conclusion that output synapses of CCK-interneurons at other interneurons, like those formed on principal cells, exhibit asynchronous release (Hefft & Jonas, 2005). However, different conclusions were reached concerning the ratio of synchronous to asynchronous release (synchronicity ratio) and its dependence on the identity of the postsynaptic target cell. Whereas Daw et al. (2009) and Karson et al. (2009) suggested that the synchronicity ratio is independent of the identity of the postsynaptic target cell, Ali and Todorova report that this ratio is larger for synapses formed between CCK-interneurons than for synapses between CCK-interneurons and pyramidal neurons. Accordingly, they suggest that factors governing asynchronous GABA release are synapse-specific and determined in part by the postsynaptic target. Alternatively, these divergent results may be explained by differences in experimental conditions (room versus physiological temperature, number of presynaptic action potentials, current-clamp versus voltage-clamp recording, and/or age of the animals) and the methods used to quantify asynchronous release. Despite these differences, all three papers unequivocally demonstrate asynchronous release at interneuron-interneuron synapses.
Asynchronous transmitter release and modulation of synaptic transmission by presynaptic CB1 receptors are hallmarks of the function of synapses formed by CCK-interneurons. How are these two properties interrelated? Ali & Todorova (2010) found that the CB1 receptor inverse agonist AM-251 increased the synchronicity ratio, whereas the endocannabinoid anandamide decreased it. This finding raises the interesting possibility that synchronous and asynchronous release are differentially affected during DSI.
Whether other presynaptic receptors on the terminals of CCK-interneurons have similar effects needs to be determined. Furthermore, the computational significance of asynchronous GABA release in principal neuron-interneuron networks remains to be elucidated. Ali & Todorova (2010) suggest that asynchronous GABA release modulates the time windows of inhibition, thereby controlling spike timing among local circuit interneurons.