Abstract
Graziano B, Wang L, White OR, Kaplan DH, Fernandez-Abascal J, Bianchi L. Neuron. 2024;112(11):1832–1847.e7. doi: 10.1016/j.neuron.2024.02.013. Epub 2024 Mar 8. PMID: 38460523; PMCID: PMC11156561. KCNQs are voltage-gated K+ channels that control neuronal excitability and are mutated in epilepsy and autism spectrum disorder. KCNQs have been extensively studied in neurons, but their function in glia is unknown. Using voltage, calcium, and GABA imaging, optogenetics, and behavioral assays, we show here for the first time in Caenorhabditis elegans (C. elegans) that glial KCNQ channels control neuronal excitability by mediating GABA release from glia via regulation of the function of L-type voltage-gated Ca2+ channels. Further, we show that human KCNQ channels have the same role when expressed in nematode glia, underscoring conservation of function across species. Finally, we show that pathogenic loss-of-function and gain-of-function human KCNQ2 mutations alter glia-to-neuron GABA signaling in distinct ways and that the KCNQ channel opener retigabine exerts rescuing effects. This work identifies glial KCNQ channels as key regulators of neuronal excitability via control of GABA release from glia.
Commentary
KCNQ (Kv7 family) channels, particularly KCNQ2 and KCNQ3, are arguably the most important class of K+ channels in epilepsy. Positional cloning of patients with a rare self-limiting pediatric epilepsy disorder in 1998 led to the discovery of KCNQ2 and KCNQ3, making them one of the earliest genes linked to genetic epilepsy. 1 The discovery that KCNQ2 and KCNQ3 channels form heteromers that mediate the M-current, a slowly activating, noninactivating K+ current, soon followed. 2 Subsequent human research has expanded the genetic landscape of KCNQ variants in epilepsy by identifying both loss-of-function (LOF) and gain-of-function (GOF) variants that cause developmental epileptic encephalopathies (DEE). A large body of basic research has also established neuronal roles for KCNQs, showing they are expressed early in development, activate before the action potential (AP) threshold to act as a brake to incoming activity, and are highly expressed in axons. 3
Previous work has shown that KCNQ LOF in principle neurons leads to hyperexcitability and increased AP firing, 4 whereas LOF in inhibitory neurons leads to potentiation of excitatory synaptic activity. 5 More recent studies on neurons from KCNQ GOF mice have shown that GOF can lead to both hyper- and hypoexcitability of distinct pyramidal neuron populations. 6 Given the rich repertoire of KCNQ channels function in neurons, it is understandable that these excitable cells have been the focus of the physiological roles of KCNQs and the mechanisms by which LOF and GOF might lead to epilepsy. But KCNQ channels are also expressed in nonneuronal brain cells. The Bianchi lab specifically previously showed that the worm (Caenorhabditis elegans) KCNQ channel kqt-2 is needed in a specific type of glial cell called AMsh for the cellular and behavioral response to the aversive odorant 1-octanol. 7 In the current work, they build upon this observation by leveraging the relative simplicity of the worm nervous system and its amenability to genetic manipulation to provide a shockingly complete model of how kqt-2 channels function in a glial cell to regulate GABA release and shape neural circuit behavior. 8
The amphid is a bilateral sensory organ located in the head of C. elegans that is comprised of 12 neurons and two glial cells, including AMsh. It mediates, among other things, the sensation of aversive odorants such as 1-octanol. kqt-2 in AMsh are necessary for 1-octanol-sensing, but how? The authors use a series of imaging experiments in combination with genetic manipulations and the behavioral readout to answer this question. With voltage imaging and genetic manipulations, they show that the first step in the sequence is that kqt-2 regulates the resting potential of AMsh glia and its response to 1-octanol. With Ca2+ imaging, they show that this regulation of the glial membrane potential controls Ca2+ influx into AMsh glia via voltage-gated L-type Ca2+ channels. Then, using GABA imaging, they show that this control of Ca2+ influx regulates the release of GABA from the AMsh glia, likely through both vesicle release and bestrophin channels. The effect of this glial-sourced GABA is to act on a 1-octanol-sensitive neuron called ASH. When GABA release from AMsh is impaired, the ASH neuron has increased baseline excitability and short-term plasticity, which presumably underlies the impaired behavioral response to 1-octanol.
In addition to delineating the role of kqt-2 in this circuit, the authors also show that human KCNQ2, KCNQ3, and KCNQ5 channels (but not KCNQ4) can functionally substitute for kqt-2 in AMsh glia, suggesting that these channels could play a similar role in human glia. Likewise, they express missense hKCNQ2 variants that cause DEE in AMsh glia and show that these variants cause distinct effects on the glial and neuronal Ca2+ responses to 1-octanol. The hKCNQ2 LOF variant, however, does a good job of rescuing the behavioral response to 1-octanol, whereas the GOF variant still has a severely impaired response to this stimulus.
One of the most impressive aspects of this paper is the thoroughness with which the authors test their hypothesis linking molecular function to behavioral output. They test many more genetic constructs than even mentioned in the synopsis above to convincingly demonstrate the molecules and mechanisms at play and their impact on the avoidance response. Rigorous science notwithstanding, the most obvious question that this paper raises in relation to epilepsy is whether disruption of GABA release from glial cells might also play a role in human epilepsies caused by KCNQ variants, or in KCNQ mouse models. The main caveats to making this jump are that the system studied in this paper is in the peripheral nervous system and the species differences between worms, humans, and other mammals. We do not know if KCNQ channels regulate GABA release from astrocytes in the mammalian central nervous system, and the final verdict on this will have to await studies designed to explicitly test this hypothesis. If they do, however, it is possible that this could play a role in epilepsy, as reduced GABA levels would mean less inhibition to counter the neuronal excitability caused by pathogenic KCNQ variants. Also, previous studies in mouse models of temporal lobe epilepsy have found that reactive astrocytes can increase their GABA production and tonic GABA release, thereby inhibiting neurons and reducing network excitability. 9 This raises the intriguing hypothesis that neuronal and astrocytic KCNQ disruption could be synergistic, with neuronal firing alterations leading to seizures followed by a failure of compensatory astrocytic GABA release.
A second takeaway from this paper in relation to epilepsy is their investigation into the mechanisms through which the LOF (or complete loss) and GOF KCNQ2 variants affect glial physiology. For many ion channels, both LOF and GOF variants can lead to epilepsy. Here the authors show that LOF and GOF have opposite effects on the glial membrane potential, as one might expect. However, their effect on the relative increase in Ca2+ levels in the AMsh caused by 1-octanol exposure is similar, due to the interaction of the glial membrane potential with the voltage-gated Ca2+ channels that couple changes in membrane potential to gliotransmitter release. This provides a plausible mechanism underlying the similar phenotypic effects of LOF and GOF variants, and for why GOF variants tend to have more severe phenotypical effects than LOF variants.
Altogether, this paper provides a superb example of how basic science investigations in a nonmammalian model system can generate new ideas and hypotheses that can be further explored in other systems in order to uncover links between the molecular causes of epilepsy and the manifestation of the disease.
