Simultaneous profiling of activity patterns in multiple neuronal subclasses

Pharmacological determination of the fractional block of Nav channels required to impair neuronal excitability and ex vivo seizures

Voltage-gated sodium channels (Nav) are essential for the initiation and propagation of action potentials in neurons. Of the nine human channel subtypes, Nav1.1, Nav1.2 and Nav1.6 are prominently expressed in the adult central nervous system (CNS). All three of these sodium channel subtypes are sensitive to block by the neurotoxin tetrodotoxin (TTX), with TTX being almost equipotent on all three subtypes. In the present study we have used TTX to determine the fractional block of Nav channels required to impair action potential firing in pyramidal neurons and reduce network seizure-like activity. Using automated patch-clamp electrophysiology, we first determined the IC₅₀s of TTX on mouse Nav1.1, Nav1.2 and Nav1.6 channels expressed in HEK cells, demonstrating this to be consistent with previously published data on human orthologs. We then compared this data to the potency of block of Nav current measured in pyramidal neurons from neocortical brain slices. Interestingly, we found that it requires nearly 10-fold greater concentration of TTX over the IC₅₀ to induce significant block of action potentials using a current-step protocol. In contrast, concentrations near the IC₅₀ resulted in a significant reduction in AP firing and increase in rheobase using a ramp protocol. Surprisingly, a 20% reduction in action potential generation observed with 3 nM TTX resulted in significant block of seizure-like activity in the 0 Mg²⁺ model of epilepsy. Additionally, we found that approximately 50% block in pyramidal cell intrinsic excitability is sufficient to completely block all seizure-like events. Furthermore, we also show that the anticonvulsant drug phenytoin blocked seizure-like events in a manner similar to TTX. These data serve as a critical starting point in understanding how fractional block of Nav channels affect intrinsic neuronal excitability and seizure-like activity. It further suggests that seizures can be controlled without significantly compromising intrinsic neuronal activity and determines the required fold over IC₅₀ for novel and clinically relevant Nav channel blockers to produce efficacy and limit side effects.

Feedforward inhibition ahead of ictal wavefronts is provided by both parvalbumin- and somatostatin-expressing interneurons

Key points

There is a rapid interneuronal response to focal activity in cortex, which restrains laterally propagating activity, including spreading epileptiform activity.

The interneuronal response involves intense activation of both parvalbumin‐ and somatostatin‐expressing interneurons.

Interneuronal bursting is time‐locked to glutamatergic barrages in the pre‐ictal period.

Ca²⁺ imaging using conditional expression of GCaMP6f provides an accurate readout of the evolving firing patterns in both types of interneuron.

The activation profiles of the two interneuronal classes are temporally offset, with the parvalbumin population being activated first, and typically, at higher rates.

Abstract

Previous work has described powerful restraints on laterally spreading activity in cortical networks, arising from a rapid feedforward interneuronal response to focal activity. This response is particularly prominent ahead of an ictal wavefront. Parvalbumin‐positive interneurons are considered to be critically involved in this feedforward inhibition, but it is not known what role, if any, is provided by somatostatin‐expressing interneurons, which target the distal dendrites of pyramidal cells. We used a combination of electrophysiology and cell class‐specific Ca²⁺ imaging in mouse brain slices bathed in 0 Mg²⁺ medium to characterize the activity profiles of pyramidal cells and parvalbumin‐ and somatostatin‐expressing interneurons during epileptiform activation. The GCaMP6f signal strongly correlates with the level of activity for both interneuronal classes. Both interneuronal classes participate in the feedfoward inhibition. This contrasts starkly with the pattern of pyramidal recruitment, which is greatly delayed. During these barrages, both sets of interneurons show intense bursting, at rates up to 300Hz, which is time‐locked to the glutamatergic barrages. The activity of parvalbumin‐expressing interneurons appears to peak early in the pre‐ictal period, and can display depolarizing block during the ictal event. In contrast, somatostatin‐expressing interneuronal activity peaks significantly later, and firing persists throughout the ictal events. Interictal events appear to be very similar to the pre‐ictal period, albeit with slightly lower firing rates. Thus, the inhibitory restraint arises from a coordinated pattern of activity in the two main classes of cortical interneurons.

There is a rapid interneuronal response to focal activity in cortex, which restrains laterally propagating activity, including spreading epileptiform activity.

The interneuronal response involves intense activation of both parvalbumin‐ and somatostatin‐expressing interneurons.

Interneuronal bursting is time‐locked to glutamatergic barrages in the pre‐ictal period.

Ca²⁺ imaging using conditional expression of GCaMP6f provides an accurate readout of the evolving firing patterns in both types of interneuron.

The activation profiles of the two interneuronal classes are temporally offset, with the parvalbumin population being activated first, and typically, at higher rates.

Graphical user interface for simultaneous profiling of activity patterns in multiple neuronal subclasses

We provide notes on how to use a graphical user interface (GUI), implemented with MATLAB, for aligning imaging datasets of biological tissue. The original use was for matching two imaging data sets, where one set was taken of the living preparation and another was taken post-fixation and following immunohistochemical processing. This technique is described in detail in an accompanying paper (Parrish et al., [1], where we also include information about the experimental procedures, and examples of the usage of the GUI.

Pyramidal cell activity levels affect the polarity of activity-induced gene transcription changes in interneurons

Changes in gene expression are an important mechanism by which activity levels are regulated in the nervous system. It is not known, however, how network activity influences gene expression in interneurons; since they themselves provide negative feedback in the form of synaptic inhibition, there exists a potential conflict between their cellular homeostatic tendencies and those of the network. We present a means of examining this issue, utilizing simple in vitro models showing different patterns of intense network activity. We found that the degree of concurrent pyramidal activation changed the polarity of the induced gene transcription. When pyramidal cells were quiescent, interneuronal activation led to an upregulation of glutamate decarboxylase 1 ( GAD1) and parvalbumin ( Pvalb) gene transcriptions, mediated by activation of the Ras/extracellular signal-related kinase mitogen-activated protein kinase (Ras/ERK MAPK) pathway. In contrast, coactivation of pyramidal cells led to an ionotropic glutamate receptor N-methyl-d-aspartate 2B-dependent decrease in transcription. Our results demonstrate a hitherto unrecognized complexity in how activity-dependent gene expression changes are manifest in cortical networks. NEW & NOTEWORTHY We demonstrate a novel feedback mechanism in cortical networks, by which glutamatergic drive, mediated through the Ras/ERK MAPK pathway, regulates gene transcription in interneurons. Using a unique feature of certain in vitro epilepsy models, we show that without this glutamatergic feedback, intense activation of interneurons causes parvalbumin and glutamate decarboxylase 1 mRNA expression to increase. If, on the other hand, pyramidal cells are coactivated with interneurons, this leads to a downregulation of these genes.