Abstract
Woo AM, Fleischel EJ, Patel DC, Sontheimer H. Epilepsia 2025;66(9):3528–3543.
Commentary
When considering mechanisms or sites of seizure generation, ion channels, synapses, glia, and even metabolic pathways readily come to mind. But another mechanism by which neuronal excitability, and hence seizures, can be modulated is the extracellular environment, or so-called extracellular matrix (ECM). The role of the ECM in regulating neuronal excitability has not received as much attention as other mechanisms, yet its roles in modifying neuronal excitability and plasticity have been recognized for decades. 1 The ECM is a complex structure consisting of molecules that alter and are altered by seizure activity. 2 Perineuronal nets (PNNs) are specialized, condensed regions of the ECM that ensheathe neurons. PNNs consist of molecules in the chondroitin sulfate proteoglycan class and related molecules that form a lattice-like structure with molecular links to adjacent proteins and sugars, creating a high density of stationary negative charges that likely contribute to cellular excitability. 3 In both humans and animal models, the PNN meshwork has been shown to anchor and modulate the cells to which it attaches, and it can exert critical roles in synapse formation and homeostasis, synaptic plasticity, and excitation/inhibition balance, as well as neuroprotective and perhaps immunomodulatory functions.4,5 The characteristics of PNNs and their exact roles vary with brain region, cell type, degree of prior seizure activity, and developmental stage. 6 Ultimately, PNNs may comprise a targetable treatment site to enhance seizure control.
Woo et al 7 studied PNNs in a model of status epilepticus (SE) in mice induced by pilocarpine. This long-established model results in hippocampal damage and post-status spontaneous recurrent seizures reminiscent of temporal lobe epilepsy. 8 Classically, PNNs affiliate with fast-spiking inhibitory (GABAergic) neurons of the parvalbumin positive type (PV+), and PNNs are thought to facilitate the high firing rate and hence epileptic propensity of these interneurons (INs). 9 However, using immunohistochemical staining to identify PNNs in this model, Woo et al 7 found that PNNs developed around regular-spiking PV-negative (PV−) INs in certain hippocampal CA1 subregions, namely, stratum lacunosum moleculare and stratum radiatum. All INs encased by newly synthesized PNN had GABAergic markers (especially somatostatin and cholecystokinin); therefore, they surround inhibitory, nonfast-spiking neurons. In addition, the newly synthesized PNN encapsulates INs of the vasoactive intestinal peptide and calretinin types that innervate other INs and are thus disinhibitory. The summed network activity is of particular relevance to epileptic circuits.
The demonstration of de novo synthesis of PNNs around this heterogeneous population of hippocampal CA1 INs implicates both circuit inhibition and disinhibition. The authors used electrophysiological methods to determine whether the atypical PNNs had functional physiological consequences. Indeed, they found that the CA1 PV− INs surrounded by PNNs were hyperexcitable, exhibiting faster firing rates in response to given levels of current injection compared with sham controls not treated with pilocarpine or pilocarpine-treated mice that did not develop SE. Furthermore, while both post-SE and sham control neurons exhibited similar intrinsic membrane properties (eg, input resistance, membrane capacitance, and sag voltages), afterhyperpolarizations (AHPs) were smaller and longer in PNN-encased PV− INs. No single K+ channel subtype was found to account for the AHP changes, but preliminary investigations using transcriptomic analysis to identify genes differentially expressed in sham versus post-SE INs revealed changes in genes related to ECM structure, inflammation, and immune function. In particular, there was upregulation of Kcnk6, a 2-pore domain potassium channel with a leak function and role in membrane potential maintenance. These channels might play a role in long-term epileptogenic changes (increased excitability persisted at least 7 days and perhaps up to 28 days after SE). This complex physiology has not yet been fully deciphered; protein-level validation and longer-term verification are needed.
Prior studies of PNNs revealed that removing PNNs enzymatically can rescue plasticity changes. To see whether enzymatic degradation would restore normal excitability in post-SE CA1 INs, chondroitinase was applied. This resulted in reduced action potential firing in response to current pulses, but the excess firing was not fully ameliorated. Moreover, enzyme treatment resulted in shorter AHPs. Chondroitinase did not otherwise affect membrane properties. These findings suggest that PNNs are, at least in part, contributing to the post-SE hyperexcitability of CA1 INs, but also that factors other than PNNs regulate the excitability of CA1 INs.
Much speculation remains about how PNNs modulate excitability. Further insights await data about how the ECM works—as an ion buffer, structural barrier, ion charge distributor, or modulator of interactions between neurons and glia. The role of de novo PNN synthesis following SE has also not been determined. That is, does the increase in PNNs mediate increased excitability, or does increased excitability occur first, which then drives new PNN formation? While the authors favor the latter explanation based on incomplete efficacy of PNN degradation to normalize activity, other SE models need to be assessed, as do longer-term transcriptomic changes and other contributors to epileptogenesis.
In summary, several aspects of this study expand our understanding of the role of PNNs in cell excitability and epileptogenesis. In the pilocarpine model, SE is associated with the de novo formation of PNNs in a small but specific population of GABAergic, PV− INs in hippocampal area CA1. INs with PNNs are hyperexcitable and fire at faster rates. Whether this increases network inhibition, as would be inferred if affected INs primarily regulate excitatory cells, or whether effects are disinhibitory, as could occur if affected INs primarily inhibit other INs, remains to be elucidated. Notably, PNN presence does not fully explain the post-status hyperexcitability, and the roles of K+ and other ionic, synaptic, and inflammatory contributors require elucidation. The numerous possible mechanistic explanations for the appearance and functions of the novel PNNs are not yet determined, but could reveal novel therapeutic targets. That would surely be a 3-pointer!
