Abstract
Function of Inhibitory Micronetworks Is Spared by Na+ Channel-Acting Anticonvulsant Drugs.
Pothmann L, Müller C, Averkin RG, Bellistri E, Miklitz C, Uebachs M, Remy S, Menendez de la Prida L, Beck H. J Neurosci 2014;34:9720–9735.
The mechanisms of action of many CNS drugs have been studied extensively on the level of their target proteins, but the effects of these compounds on the level of complex CNS networks that are composed of different types of excitatory and inhibitory neurons are not well understood. Many currently used anticonvulsant drugs are known to exert potent use-dependent blocking effects on voltage-gated Na+ channels, which are thought to underlie the inhibition of pathological high-frequency firing. However, some GABAergic inhibitory neurons are capable of firing at very high rates, suggesting that these anticonvulsants should cause impaired GABAergic inhibition. We have, therefore, studied the effects of anticonvulsant drugs acting via use-dependent block of voltage-gated Na+ channels on GABAergic inhibitory micronetworks in the rodent hippocampus. We find that firing of pyramidal neurons is reliably inhibited in a use-dependent manner by the prototypical Na+ channel blocker carbamazepine. In contrast, a combination of intrinsic and synaptic properties renders synaptically driven firing of interneurons essentially insensitive to this anticonvulsant. In addition, a combination of voltage-sensitive dye imaging and electrophysiological experiments reveal that GABAergic feedforward and feedback inhibition is unaffected by carbamazepine and additional commonly used Na+ channel-acting anticonvulsants, both in control and epileptic animals. Moreover, inhibition in control and epileptic rats recruited by in vivo activity patterns was similarly unaffected. These results suggest that sparing of inhibition is an important principle underlying the powerful reduction of CNS excitability exerted by anticonvulsant drugs.
Commentary
The effectiveness of sodium channel-targeting anticonvulsants in epilepsy treatment has been well-known for decades, which has inspired numerous insightful investigations into their mechanism of action from both molecular and cellular perspectives. This class of drugs—including carbamazepine, phenytoin, and lamotrigine—preferentially binds to the inactivated conformation of sodium channels and acts in a use-dependent manner, increasing channel block with prolonged or repetitive activation (1, 2). This dependence on high levels of neuronal activity has led to the hypothesis that these drugs inhibit the hallmark pathological high-frequency oscillations prevalent in epilepsy, while leaving normal physiological function relatively (albeit certainly not completely) intact (3). However, there has been a need to bridge the divide between the specific molecular–cellular mechanisms and broad clinical effectiveness by considering the intermediate network context and taking into account the existence of distinct neuronal subpopulations and the different aspects of network function. In particular, while the effects of anticonvulsants have been examined primarily in excitatory cell types (4, 5), their effects on interneurons and inhibition have been largely understudied. It therefore remains an intriguing open question as to whether inhibitory function is also compromised by sodium channel-targeting anticonvulsants.
In a complex series of elegant, carefully conducted experiments, Pothmann and colleagues addressed the question of whether there are cell type-specific effects of carbamazepine arising from the impressive diversity of inhibitory interneurons (6, 7). The authors used a variety of techniques, including slice electrophysiology, voltage-sensitive dye imaging, morphological identification and in vivo juxtacellular recordings from freely moving animals to not only extend the well-established use-dependent blocking action of carbamazepine to the interneuronal population but also to discover striking cell-type specificity in its efficacy. Starting with pyramidal cells in acute hippocampal slices, the authors first showed that, as expected from use-dependent block, reduction in the maximal firing rate of pyramidal cells in response to carbamazepine applied at a clinically relevant dose increased with longer somatic current injections. The drug-induced firing rate changes in response to intracellular current pulses were then quantified again, but this time in three classes of interneurons that innervate different regions of CA1 pyramidal cells: basket cells that target the perisomatic region (BCs), oriens lacunosum-moleculare cells that target the distal dendrites (OLM), and interneurons that target the proximal dendrites (PD). While the BCs and OLM cells showed a similar reduction to that of the pyramidal cells, carbamazepine caused a larger reduction in PD cells. This previously unknown cell-type specificity highlights the valuable information gained by identification of inhibitory neurons by their subclasses when analyzing drug effects.
Critically, however, the widespread reduction in firing observed using current injection was not present when the authors examined synaptically induced firing in distinct cell types, as carbamazepine had little effect on interneuronal firing rates during high-frequency stimulation of the recurrent pyramidal cell axon collaterals located in the alveus. This lack of synaptic effect can be explained in BCs and PD interneurons by the strongly depressing nature of their local excitatory inputs, so that they fire action potentials only briefly at the onset of the stimulus. However, the OLM cell result is surprising given that their inputs are facilitating and, therefore, these cells continue to fire. In any case, this sparing of interneuronal firing when driving synaptic input was a key, novel finding that begs the question of whether the overall inhibitory function in the network is affected, which Pothmann and colleagues also addressed.
Inhibition in the CA1 circuit, as assessed by recording stimulation-evoked inhibitory events from pyramidal cells, was unchanged with the addition of carbamazepine. The feedback inhibition in pyramidal cells was assessed using stimulation of the alveus, while feedforward inhibition was initiated via stimulation of the Schaffer collateral input originating from CA3 pyramidal cells. The peak amplitudes of the inhibitory postsynaptic currents (IPSCs) measured in CA1 pyramidal cells were unaffected for feedback and feedforward inhibition, using high-frequency stimulation of 50 Hz and 100 Hz. This result was supplemented by voltage-sensitive dye imaging showing no impact of carbamazepine on the laminar distribution or strength of both forms of inhibition. To determine if sparing of inhibition was a common characteristic of the sodium channel-targeting class, two additional anticonvulsants—phenytoin and lamotrigine—were also tested. Similar to carbamazepine, phenytoin and lamotrigine reduced the maximal rate of firing for pyramidal cells using current injection, but feedback inhibition was left completely intact in both cases. These results demonstrate that sodium channel-targeting anticonvulsants do not affect inhibition in the control condition.
However, the pathological condition could potentially be differentially affected, as these anticonvulsants have been shown to have altered efficacy in the chronically epileptic hippocampus (4, 5). Several experiments were thus repeated in the pilocarpine model of chronic epilepsy, and the results corroborated the sparing of inhibition found in the control condition. Specifically, the reduction in the maximal rate of pyramidal cell firing in epileptic animals was not significantly different from the control condition, and feedback inhibition was similarly unchanged after carbamazepine addition.
Next, the authors turned to the issue of physiologically relevant patterns of activity; up to this point in their study, inhibition was initiated and measured using an artificial 50 Hz or 100 Hz stimulation paradigm. To better assess the anticonvulsant effects on inhibition under biological network activity, pyramidal cell recordings were obtained in vivo using juxtacellular recordings during physiological (ripple) and pathological (fast ripple) high-frequency oscillations. These activity patterns were then used to stimulate the alveus in control and epileptic animals to determine if feedback inhibition was compromised. The authors found that even with biological high-frequency activity, the addition of carbamazepine had little discernible effect on inhibitory function in both the physiological and pathological networks. One caveat is that no recordings were obtained during bona fide seizure activity, which would have provided the conclusive test for the sparing of inhibition in the pathological state.
In summary, this important study provides a novel, cellular subtype- and network-based explanation for the effectiveness of the class of sodium channel-targeting anticonvulsants; namely, these drugs are able to spare inhibitory function while also acting only during periods of high activity. However, an important consideration is that the clinical efficacy of these anticonvulsants varies depending on the type of epilepsy, such as their potential paradoxical exacerbation of absence seizures (8). In light of the current study, it would be of interest to determine if inhibition is similarly spared by these anticonvulsants in other forms of epilepsy or potentially plays a role in these paradoxical effects. In addition, carbamazepine resistance has been demonstrated in tissue from an experimental model and from drug-resistant patients (9, 10), the mechanism of which provides an avenue of study with important potential clinical implications. Overall, this study serves as a crucial reminder of the information that can be gained through careful identification of neuronal subtypes as well as the dichotomy that can arise between the measured effect on a single cell and the actual impact on its activity in a biological network context.