Lighting the Fuse: Deconstructing Complex Network Interactions Using On-Demand Seizures

Commentary

Dynamic, Cell Type-Specific Roles for GABAergic Interneurons in a Mouse Model of Optogenetically Inducible Seizures.

Khoshkhoo S, Vogt D, Sohal VS. Neuron 2016;93:291–298.28041880

GABAergic interneurons play critical roles in seizures, but it remains unknown whether these vary across interneuron subtypes or evolve during a seizure. This uncertainty stems from the unpredictable timing of seizures in most models, which limits neuronal imaging or manipulations around the seizure onset. Here, we describe a mouse model for optogenetic seizure induction. Combining this with calcium imaging, we find that seizure onset rapidly recruits parvalbumin (PV+ve), somatostatin (SOM+ve), and vasoactive intestinal peptide (VIP)-expressing interneurons, whereas excitatory neurons are recruited several seconds later. Optogenetically inhibiting VIP interneurons consistently increased seizure threshold and reduced seizure duration. Inhibiting PV+ and SOM+ interneurons had mixed effects on seizure initiation but consistently reduced seizure duration. Thus, while their roles may evolve during seizures, PV+ and SOM+ interneurons ultimately help maintain ongoing seizures. These results show how an optogenetically induced seizure model can be leveraged to pinpoint a new target for seizure control: VIP interneurons.

The unpredictability of seizures severely affects the quality of life of patients with epilepsy and is a major impediment to understanding the microcircuit dynamics in the seconds to minutes preceding a seizure (1). Although pharmacological approaches have been used to elicit and examine seizures, the onset is still unpredictable, and the pharmacology confounds mechanistic analyses. Thus, despite the explosion of techniques ranging from optogenetic manipulation of cellular activity, physiology, and neuronal imaging, the mechanisms underlying transition to seizure have remained elusive due to the inability to repeatedly and reliably predict seizure onset. In particular, the role of GABAergic interneurons has been difficult to decipher due to their diversity, distinct connections, and activity patterns. GABAergic neurons shape patterned activation of neuronal ensembles during essential brain functions (2). Traditionally, a reduction or transient collapse in GABAergic inhibition is thought to promote hyperexcitability and seizures; however, recent studies have shown an aberrant increase in GABAergic signaling at seizure onset suggesting an ictogenic role for GABA, possibly through pathological synchronization of networks (3–5). Nevertheless, the identity and contribution of interneuron classes to seizure initiation is not fully known, largely due to the unpredictable nature of seizures.

Khoshkhoo and colleagues introduced a novel “optical-kindling” technique to evoke seizures on-demand through focal optogenetic activation of neurons in mouse primary motor cortex. They showed that 20–40 Hz optogenetic stimulation evokes seizures with a predictable 1–2-minute delay, enabling analysis of processes underlying seizure onset. Interestingly, the majority of optically evoked seizures were primarily generalized. However a few seizures had a focal origin often with a focus contralateral to the optically kindled hemisphere. The authors exploited this on-demand seizure model and mouse lines with Cre driven by cell specific promoters to examine the contribution of parvalbumin (PV) and somatostatin (SOM) interneurons, which are derived from the medial ganglionic eminence (MGE) and vasoactive intestinal peptide (VIP) expressing “interneuron-targeting interneurons” to seizure initiation and termination.

By combining optical kindling with bilateral EEG and contralateral fiber photometric recordings of bulk calcium signals from neurons expressing a fluorescent reporter GCaMP6f in the primary motor cortex, they showed that all three interneuron classes exhibit rapid increases in calcium signals at seizure onset, approximately 10 sec ahead of excitatory neurons. While PV and SOM neurons showed sustained calcium signals until seizure termination, calcium signals in VIP and excitatory neurons decreased prior to seizure termination, indicating cell-type specific differences in neuronal contribution to seizure maintenance and termination. The authors used cre-driven expression of Archaerhodopsin (eArch 3.0)—a microbial proton pump that hyperpolarizes the membrane on light activation—to silence specific interneurons and determine their role in seizure dynamics. As summarized in Table 1, they showed that silencing interneurons between optical-kindling epochs has complex effects on seizure thresholds and duration. While combined silencing of PV and SOM neurons (using Dlxl12b promoter) reduced seizure threshold regardless of the side of silencing, individually silencing PV neurons contralateral (not ipsilateral) to optical kindling increased seizure threshold. Silencing SOM neurons failed to modulate seizure onset. In contrast, silencing VIP neurons contralateral to optical kindling increased seizure threshold, but had no effect ipsilateral to seizure induction. Silencing interneurons consistently reduced seizure duration regardless of the cell-type or side suggesting that all three classes of interneurons examined contribute to maintenance of seizure activity. Based on findings that silencing VIP neurons increases threshold, albeit only contralateral to seizure induction, and reduces duration of seizures, the authors speculated that suppressing VIP neurons could have therapeutic potential. The authors attributed the anti-seizure effects of VIP neuron suppression to elimination of their inhibitory break on SOM neurons, which they predominantly innervated (6). Fiber photometry also identified prolonged and robust calcium signals in all neuronal types occurring variably following seizure termination, suggesting that optical-kindling may generate cortical spreading depression-like event.

TABLE 1.

Summary of Findings

The optical-kindling model has definite advantages to analyzing seizure dynamics, as it eliminates unpredictability and enables specific and temporally targeted analysis of mechanisms of seizure onset. However, optically evoked seizures have some unique characteristics that merit further consideration. As acknowledged by the authors, optical-kindling studies detailed here do not include structural and functional reorganization including interneuron plasticity, which occurs in networks with spontaneous recurrent seizures (4, 7). Whether epileptic networks will exhibit the 1–2-minute delay between optical-kindling and seizure onset (which is a specific advantage of this model) needs to be examined. That aside, the optical-kindling evokes seizures after 1–2 minutes and, despite focal induction, manifests primarily as generalized seizures. While understanding the mechanisms driving these features is technically challenging, whether these processes are generalizable to spontaneous seizures needs to be considered. Another unexpected feature of the model is the reset of seizure threshold to baseline on subsequent days of kindling, which contrasts with the reliable progression to spontaneous epilepsy following electrical kindling. Long-term studies are needed to determine if optical-kindling contributes to network-level changes and epileptogenesis. The reset of seizure threshold to baseline bears further analysis as they could provide clues to seizures that progress to epilepsy and those that do not.

Mechanistic analysis of interneuron contribution to seizure onset and termination in the “on-demand” seizure model reveal several changes which are interesting yet difficult to interpret. It seems logical that simultaneously suppressing both PV and SOM neurons would aid seizure onset. However, selective and focal suppression of PV neurons paradoxically reduced seizure probability but only on contralateral silencing. Previous studies have supported a role for PV neurons in seizure onset (8, 9), while suppressing ongoing seizures (10). As noted by the authors, depolarizing GABA during the preictal phase could account for PV neuronal support of seizure onset. However, the reasons for lateralization of the role of PV neurons in seizure onset are unclear and may result from the spatially discrete suppression paradigm and proximity to seizure focus (11). Similarly, anti-seizure effects of silencing VIP interneurons are difficult to explain. SOM and some PV neurons are the primary targets of VIP neurons (6). It is therefore surprising that silencing VIP neurons has the same effect in reducing seizure duration as suppressing PV or SOM interneurons. Similarly, the proposal that silencing VIP neurons increase seizure threshold by releasing SOM and possibly PV neurons from inhibition is surprising in light of data showing that suppressing PV neurons also increases seizure threshold and SOM neurons does not impact seizure onset. It is possible that the complex outcomes may reflect distributed network level changes in activity resulting from the focal manipulations rather than direct microcircuit effects. Alternatively, the possibility that VIP neurons have alternate downstream targets needs further examination.

In conclusion, the road to understanding seizure onset is daunting and the novel optical kindling method proposed by Khoshkhoo et al. provides an exciting tool to break down the unpredictability of seizures and could be valuable in providing mechanistic insights. However, as with existing seizure models, one should carefully consider the unique features of the optical-kindling and focal experimental manipulations, as they may influence the outcomes. Thus, whether the nature of interneuron involvement identified in optically kindled seizures is model specific or informs mechanisms of spontaneous epileptic seizures remains to be examined.