On-Demand Seizures Facilitate Rapid Screening of Therapeutics for Epilepsy

Reviewer #1 (Public review):

Summary:

This important study by Takano et. al. describes a novel approach for optogenetically evoking seizures in an etiologically relevant mouse model of epilepsy. The authors developed a model that can trigger seizures "on demand" using optogenetic stimulation of CA1 principal cells in mice rendered epileptic by an intra-hippocampal kainate (IHK) injection into CA3. The authors discuss their model in the context of the limitations of current animal models used in epilepsy drug development. In particular, their model addresses concerns regarding existing models where testing typically involves inducing acute seizures in healthy animals or waiting on infrequent, spontaneous seizures in epileptic animals.

Strengths:

A strength of this manuscript is that this approach may facilitate the evaluation of novel therapeutics since these evoked seizures are demonstrated as being sufficiently similar to spontaneous seizures in these same mice which are more laborious to analyze. The data demonstrating the commonality of pharmacology and EEG features between evoked seizures and spontaneous seizures in epileptic mice, while also being different from evoked seizures in naïve mice, are convincing despite concerns regarding the biological significance of the differences in effect sizes of these features. The structural, functional, and behavioral differences between a seizure-naïve and epileptic mouse are complex and important issues. This study positively impacts the wider epilepsy research community by investigating seizure semiology and pharmaceutical responses in these populations.

Weaknesses:

While the data generally supports the authors' conclusions, a weakness of this manuscript lies in their analytical approach where EEG feature-space comparisons used the number of spontaneous or evoked seizures as their replicates as opposed to the number of IHK mice; these large data sets tend to identify relatively small effects of uncertain biological significance as being highly statistically significant. Furthermore, the clinical relevance of similarly small differences in EEG feature space measurements between seizure-naïve and epileptic mice is also uncertain. Finally, the multiple surgeries and long timetable to generate these mice may limit the value compared to existing models in drug-testing paradigms.

Reviewer #2 (Public review):

Summary:

The authors have attempted to modify and adapt the IH-KA model in mice to provide an improved approach to screening for new ASDs by partially mitigating the problem of randomly occurring seizures and relatively low seizure frequency in the IH-KA model. The authors used KA micro-injections to selectively kill the hippocampal CA3 area as a way to induce temporal lobe "epileptogenesis" (TLE), and then used optogenetics to activate CA1 pyramidal cells specifically. This approach allowed the authors to trigger generalized seizures where the tonic-clonic pattern of electrical activity was reminiscent of actual tonic-clonic behavioral convulsions. Administration of levitracetam (LEV) and diazepam (DZP), two widely used ASDs with different mechanisms, reduced the probability of optogenetically activated epileptic seizures in IH-KA mice, thus seeming to provide evidence for a new approach to screen ASDs. A variety of problems and issues with the approach and the results lead to confounds that raise serious concerns about the conclusions.

Major strengths and weaknesses of the Methods and Results:

Strengths:

The authors have designed a method for triggering seizures, and the figures show bona fide electrographic seizures with concomitant convulsive behavioral components. The optogenetically evoked seizures in IH-KA mice had the electrical properties of actual seizures and the tonic-clonic components were readily apparent. These seizures appeared different from seizures evoked in naïve mice, and the authors attribute this difference to the epileptogenic process, but this may not be correct.

The ASDs (i.e., LEV and DZP) reduced the success rate of the optogenetically evoked seizures in IH-KA mice, thus suggesting the potential usefulness of the model for testing ASDs. The paper discusses whether the Epilepsy Therapy Screening Program (ETSP) will be able to use this modification of the IH-KA model in place of (1) ASD screening with acute seizures in naïve animals, where the brain has not undergone "epileptogenesis", (2) testing ASDs on hippocampal paroxysmal discharges (HPDs) in the IH-KA model, which has undergone epileptogenesis, or (3) spontaneous epileptic seizures in animal models of TLE based on systemic treatments that lead to acute convulsive status epilepticus that have later undergone epileptogenesis. This proposed version of the IH-KA model aims to address the former problem (#1, above) by using a mouse model of TLE, and to address the latter problems (#2 and #3, above) of the seemingly random occurrence of epileptic seizures and the low seizure frequency by using optogenetically "triggered" seizures.

Weaknesses

Although the figures provide excellent examples of individual electrographic seizures and compare induced seizures in epileptic and naïve animals, it is unclear which criteria were used to identify an actual seizure induced by the optogenetic stimulus, versus a hippocampal paroxysmal discharge (HPD), an "afterdischarge", an "electrophysiological epileptiform event" (EEE, Ref #36, D'Ambrosio et al., 2010 Epilepsy Currents), or a so-called "spike-wave-discharge" (SWD). Were HPDs or these other non-seizure events ever induced using stimulation in animals with IH-KA? A critical issue is that these other electrical events are not actual seizures, and it is unclear whether they were included in the column showing data on "electrographic afterdischarges" in Figure 5 for the studies on ASDs. This seems to be a problem in other areas of the paper, also.

The differences between the optogenetically evoked seizures in IH-KA vs naïve mice are interpreted to be due to the "epileptogenesis" that had occurred, but the lesion from the KA-induced injury would be expected to cause differences in the electrically and behaviorally recorded seizures - even if epileptogenesis had not occurred. This is not adequately addressed.

The authors did not test whether an apparent "kindling" effect, apparently seen in naïve controls, also occurred in animals micro-injected with kainic acid (KA). This effect could cause model instability that might result in variability in response to ASDs. It is not clear whether the number of optogenetically induced seizures in epileptic animals would affect the response to drugs. It is also unclear how much of an improvement the animal model in the present work is over other similar models of TLE, where electrically triggered seizures could simply be applied to one of them.

The authors offer little mention of other research using animal models of TLE to screen ASDs, of which there are many published studies - many of them with other strengths and/or weaknesses. For example, although Grabenstatter and Dudek (2019, Epilepsia) used a version of the systemic KA model to obtain dose-response data on the effects of carbamazepine on spontaneous seizures, that work required use of KA-treated rats selected to have very high rates of spontaneous seizures, which requires careful and tedious selection of animals. The ETSP has published studies with an intra-amygdala kainic acid (IA-KA) model (West et al., 2022, Exp Neurol), where the authors claim that they can use spontaneous seizures to identify ASDs for DRE; however, their lack of a drug effect of carbamazepine may have been a false negative secondary to low seizure rates. The approach described in this paper may help with confounds caused by low or variable seizure rates. These types of issues should be discussed, along with others.

While the paper may be relevant for the ETSP and contract research organizations (CROs), the paper was not written to attract the interest of biological scientists, even those in this specific area of epilepsy research. It may be of low interest to other neuroscientists.

The outcome measure for testing LEV and DZP on seizures was essentially the fraction of unsuccessful or successful activations of seizures, where high ASD efficacy is based on showing that the optogenetic stimulation causes fewer seizures when the drug is present. The final outcome measure is thus a percentage, which would still lead to a large number of tests to be assured of adequate statistical power. Thus, there is a concern about whether this proposed approach will have high enough resolution to be more useful than conventional screening methods so that one can obtain actual dose-response data on ASDs.

The key issue the authors aim to address is the 30-40% of patients with DRE, but the real problem with DRE patients is not that these people have seizures with no effect of the ASDs; rather, although ASD may reduce seizure burden, these patients continue to have some remaining seizures even after high doses of ASDs, which often leads to adverse effects from the particular ASDs.

In several sections of the paper, the authors argue that two different groups are similar on the basis that no statistical difference was found between the two groups (i.e., p > 0.05); however, the failure to find a statistically significant difference, particularly with relatively small sample sizes, is not rigorous evidence that the two groups are actually similar - they are just "not significantly different."

It remains unclear that the optogenetically induced seizures in this model are better than similarly induced seizures in a naïve animal, and there is no evidence that the model will be useful for finding new ASDs to treat DRE.

Do the results support the conclusions?

Although the Results show examples of clear tonic-clonic seizures, it is not at all clear whether this approach is a significant improvement over previous methods used on animal models of TLE. The presented data from this method shows it provides an ability to detect the effect of widely used ASDs, but not that it will have the resolution to find better ASDs. The outcome measure of successful vs failed seizure inductions does not necessarily translate to a pathway for finding new ASDs for DRE, which often is a function of the side effects of the proposed new ASD. Although the recorded seizures in IH-KA rats differ in waveform from the ones in naïve mice, this could be due to the pattern of damage resulting from the micro-injection of KA or the density of expressed Chr2, which could be affected by sclerosis.

Impact and utility of methods and data.

The authors state that this approach should be used to test for and discover new ASDs for DRE, and also used for various open/closed loop protocols with deep-brain stimulation; however, the paper does not actually discuss rigorously or critically the background literature on other published studies in these areas or how this approach will improve future research for a broader audience than the ETSP and CROs. Thus, it is not clear whether the utility will apply more widely and how extensive a readership will be attracted to this work.

Final Conclusions:

Although this is an Interesting if not elegant new model for testing ASDs, it could be seen as a version of kindling (plus brain damage) in a rodent model, where some of the pathology of TLE is induced through focal injection of KA in the CA3 area of the hippocampus. Unfortunately, no evidence was presented that it will be any better than other models, although it could be faster and maybe easier than models based on spontaneous seizures. Although it has some similarities to the pathology of human TLE, the ablating part of the hippocampus does not account for the more widespread pathology that usually occurs elsewhere in the brain, as studied with imaging and with anatomy in surgical specimens from patients with DRE.

Although this approach with seizure induction via an optogenetic approach adds specificity to the type of cell that is stimulated (i.e., CA1 pyramidal cells), it is not apparent why this provides a better or more effective tool than simple electrical induction of seizures in any TLE model. Most important, it remains unclear how this addresses any aspect of drug resistance. To improve the ASD discovery process, an important new model must make a significant reduction in seizure burden, and would ideally improve the percentage of patients that become seizure-free. It is not clear how this model will do that.

In the end, the authors have created a model with some of the pathology of TLE, where they can then induce actual seizures via specific optogenetic stimulation. So, although it is potentially elegant work, it remains unclear what new information this model will tell us about epilepsy, and most importantly DRE - or how it will improve treatment outcomes.

Reviewer #3 (Public review):

Summary:

Chen et al. develop and characterize a new approach for screening drugs for epilepsy. The idea is to increase the ability to study seizures in animals with epilepsy because most animal models have rare seizures. Thus, the authors use the existing intrahippocampal kainic acid (IHKA) mouse model, which can have very unpredictable seizures with long periods of time between seizures. The authors employ an additional method to trigger seizures in the IHKA model. This method is closed-loop optogenetic stimulation of area CA1. There are several assumptions: area CA1 is the best location, triggered seizures are the same as spontaneous seizures, and this method will be useful despite requiring a great deal of effort. Regarding the latter, using a mouse model with numerous seizures (such as the pilocarpine model) might be more efficient than using a modified IHKA protocol that requires viral injection for optogenetics, fiber insertion requiring additional surgery, and accurate targeting to reliably trigger seizures on-demand. Aside from these caveats, the authors do succeed in studying seizures more readily in a mouse model of rare seizures. However, the seizures are evoked, not spontaneous. As currently presented, it is not clear how the triggered seizures can be used to investigate if antiseizure medication can reduce seizure burden as measured by seizure severity and seizures per day.

The authors modified the IHKA model to inject KA into CA3 instead of CA1 in order to preserve the CA1 pyramidal cells that they will later stimulate. To express the excitatory opsin channelrhodopsin (ChR2) in area CA1, they use a virus that expresses ChR2 in cells that express the Thy-1 promoter. The authors demonstrate that CA3 delivery of KA can induce a very similar chronic epilepsy phenotype to the injection of KA in CA1 and show that optical excitation of CA1 can reliably induce seizures. These are the strengths of the study.

While the authors show that electrophysiological signatures of induced vs spontaneous seizures are similar in many ways, the authors also show several differences and it is not clear if these differences are meaningful. Notably, the induced seizures are robustly inhibited by the antiseizure medication levetiracetam and variably but significantly inhibited by diazepam, similar to many mouse models with chronic recurrent seizure activity. I agree with the authors that this modified IHKA model will be of most value for higher throughput screening of potential antiseizure therapies, but with the caveat that the data may not generalize to other epilepsy models or humans. The authors evaluate the impact of repeated stimulation on the reliability of seizure induction and show that seizures can be reliably induced by CA1 stimulation for as long as 16 days, but the utility of the model would be better demonstrated if seizures could be shown to be inducible over the range of weeks to months.

Strengths:

(1) The authors show that the IHKA model of chronic epilepsy can be modified to preserve CA1 pyramidal cells (but at a cost of CA3 cells), allowing on-demand optogenetic stimulation of CA1 that appears to lower seizure threshold and thus trigger a seizure event.

(2) The authors show that repeated reactivation of CA1 even in untreated mice can promote kindling and induction of seizure activity, indeed generating two mouse models in total.

(3) Many electrophysiological signatures are similar between the induced and spontaneous seizures, and induced seizures reliably respond to treatment with antiseizure medications.

(4) Given that more seizures can be observed per mouse using on-demand optogenetics, this model enhances the utility of each individual mouse.

Weaknesses:

(1) Evaluation of seizure similarity using the SVM modeling and clustering is not sufficiently explained to show if there are meaningful differences between induced and spontaneous seizures. SVM modeling did not include analysis to assess the overfitting of each classifier since mice were modeled individually for classification.

(2) The difference between seizures and epileptiform discharges or trains of spikes (which are not seizures) is not made clear.

(3) The utility of increasing the number of seizures for enhancing statistical power is limited unless the sample size under evaluation is the number of seizures. However, the standard practice is for the sample size to be the number of mice.

(4) Seizure burden is not easily tested.

(5) It is unlikely that long-term adaptation to CA1-stimulated seizure induction is absent in these mice. A duration of evaluation longer than 16 days is warranted in light of the downward slope at days 13-16 for induced seizures in Figure 4C.

(6) Human epilepsy is extensively heterogeneous in both etiology and individual phenotype, and it may be hard to generalize the approach.

(7) No mention or assessment of mouse sex as a biological variable.

Author response:

In this initial response to the public review, we outline our plan to address the major concerns raised. Below, we provide a general categorization of the suggestions and our corresponding responses

Weakness #1: Statistical Concerns - using the number of seizures (rather than the number of animals) may identify small effects that could be insignificant. Effect size should be taken into consideration.

Reviewer 1:

“While the data generally supports the authors' conclusions, a weakness of this manuscript lies in their analytical approach where EEG feature-space comparisons used the number of spontaneous or evoked seizures as their replicates as opposed to the number of IHK mice; these large data sets tend to identify relatively small effects of uncertain biological significance as being highly statistically significant.”

Reviewer 2:

“In several sections of the paper, the authors argue that two different groups are similar on the basis that no statistical difference was found between the two groups (i.e., p > 0.05); however, the failure to find a statistically significant difference, particularly with relatively small sample sizes, is not rigorous evidence that the two groups are actually similar - they are just "not significantly different.”

Reviewer 3:

“(3) The utility of increasing the number of seizures for enhancing statistical power is limited unless the sample size under evaluation is the number of seizures. However, the standard practice is for the sample size to be the number of mice.”

Reviewer 3:

“(1) Evaluation of seizure similarity using the SVM modeling and clustering is not sufficiently explained to show if there are meaningful differences between induced and spontaneous seizures. SVM modeling did not include analysis to assess the overfitting of each classifier since mice were modeled individually for classification.”

We understand the reviewers’ concerns. In this work, we used linear mixed effect model to address two levels of variability –between animals and within animals. The interactive linear mixed effect model shows that most (~90%) of the variability in our data comes from within animals (Residual), the random effect that the model accounts for, rather than between animals. Since variability between animals are low, the model identifies common changes in seizure propagation across animals, while accounting for the variability in seizures within each animal. Therefore, the results we find are of changes that happen across animals, not of individual seizures. We will make text edits to enhance understanding of the linear mixed effect model.

To address the point raised about similarity, we will explain how the SVM classifier was trained. The purpose of the SVM is not to identify meaningful differences between induced and spontaneous seizures. Rather, it is to classify EEG sections as “seizures” or non-seizures, demonstrating the gross similarity between induced and spontaneous seizures despite minor differences. We will make text clarifications for the SVM model.

Weakness #2: Clinical and biological significance is unclear.

Reviewer 1:

“Furthermore, the clinical relevance of similarly small differences in EEG feature space measurements between seizure-naïve and epileptic mice is also uncertain.”

Reviewer 2:

“While the paper may be relevant for the ETSP and contract research organizations (CROs), the paper was not written to attract the interest of biological scientists, even those in this specific area of epilepsy research. It may be of low interest to other neuroscientists… The key issue the authors aim to address is the 30-40% of patients with DRE, but the real problem with DRE patients is not that these people have seizures with no effect of the ASDs; rather, although ASD may reduce seizure burden, these patients continue to have some remaining seizures even after high doses of ASDs, which often leads to adverse effects from the particular ASDs… It remains unclear that the optogenetically induced seizures in this model are better than similarly induced seizures in a naïve animal, and there is no evidence that the model will be useful for finding new ASDs to treat DRE.”

Reviewer 3:

“(6) Human epilepsy is extensively heterogeneous in both etiology and individual phenotype, and it may be hard to generalize the approach.”

Reviewer 2:

“The authors state that this approach should be used to test for and discover new ASDs for DRE, and also used for various open/closed loop protocols with deep-brain stimulation; however, the paper does not actually discuss rigorously or critically the background literature on other published studies in these areas or how this approach will improve future research for a broader audience than the ETSP and CROs. Thus, it is not clear whether the utility will apply more widely and how extensive a readership will be attracted to this work.”

We appreciate the reviewer’s concerns. We will revise the manuscript to better emphasize the potential significance of our approach. The on-demand seizure model can be applied to address biologically and clinically relevant questions beyond its utility in drug screening. For example, crossing the Thy1-ChR2 mouse line with genetic epilepsy models, such as Scn1a mutants, could reveal how optogenetic stimulation differentially induces seizures in mutant versus non-mutant mice, providing insights into seizure generation and propagation in Dravet Syndrome. Due to the cellular specificity of optogenetics, we also envision this approach being used to study circuit-specific mechanisms of seizure generation and propagation. Regarding drug-resistant epilepsy (DRE) and anti-seizure drug (ASD) screening, we agree with the reviewer that probing new classes of ASDs for DRE represents the critical goal. However, we believe a full exploration of additional ASD classes and/or modeling DRE lies outside the scope of this manuscript.

Weakness #3: Definition of Seizure is unclear

Reviewer 2:

“Although the figures provide excellent examples of individual electrographic seizures and compare induced seizures in epileptic and naïve animals, it is unclear which criteria were used to identify an actual seizure induced by the optogenetic stimulus, versus a hippocampal paroxysmal discharge (HPD), an "afterdischarge", an "electrophysiological epileptiform event" (EEE, Ref #36, D'Ambrosio et al., 2010 Epilepsy Currents), or a so-called "spike-wave-discharge" (SWD). Were HPDs or these other non-seizure events ever induced using stimulation in animals with IH-KA? A critical issue is that these other electrical events are not actual seizures, and it is unclear whether they were included in the column showing data on "electrographic afterdischarges" in Figure 5 for the studies on ASDs”

Reviewer 3:

“(2) The difference between seizures and epileptiform discharges or trains of spikes (which are not seizures) is not made clear.”

Reviewer 2:

“The differences between the optogenetically evoked seizures in IH-KA vs naïve mice are interpreted to be due to the "epileptogenesis" that had occurred, but the lesion from the KA-induced injury would be expected to cause differences in the electrically and behaviorally recorded seizures - even if epileptogenesis had not occurred. This is not adequately addressed.”

Thank you for pointing out the unclear definition of the seizures analyzed. We agree and will revise the text to clarify this issue. In this manuscript, we focused on tonic-clonic seizures. We analyzed animal behavior during evoked events, and a high percentage of induced electrographic events were accompanied by behavioral seizures with a Racine scale of three or above. Regarding epileptogenesis, our model is based on the IHK model, in which spontaneous tonic-clonic seizures occur a few to several days after KA injection. These mice are, by definition, epileptogenic. We will further clarify this methodology in the text.

Weakness #4: Similarity/Difference with Kindling Not Clear

Reviewer 2:

“The authors did not test whether an apparent "kindling" effect, apparently seen in naïve controls, also occurred in animals micro-injected with kainic acid (KA). This effect could cause model instability that might result in variability in response to ASDs. It is not clear whether the number of optogenetically induced seizures in epileptic animals would affect the response to drugs. It is also unclear how much of an improvement the animal model in the present work is over other similar models of TLE, where electrically triggered seizures could simply be applied to one of them.”

Reviewer 3:

“(5) It is unlikely that long-term adaptation to CA1-stimulated seizure induction is absent in these mice. A duration of evaluation longer than 16 days is warranted in light of the downward slope at days 13-16 for induced seizures in Figure 4C.”

We appreciate the reviewer’s comments regarding the “kindling effect” as well as its similarity to the kindling model. We will carefully assess the data and address this in the revised manuscript. In electrical kindling, the activated cellular population is non-specific, including both excitatory and inhibitory neurons. In our model, we specifically activate predominantly excitatory neurons (Thy1-positive neurons), which we observed to participate in convulsant-induced seizures (as demonstrated in Thy1-GCaMP experiments). We consider this specificity an improvement over the kindling model, making our approach more biologically relevant.

Weakness #5: Time needed to generate model is significant. Unclear if animals were pre-selected

Reviewer 1:

“Finally, the multiple surgeries and long timetable to generate these mice may limit the value compared to existing models in drug-testing paradigms.

Reviewer 2:

“The authors offer little mention of other research using animal models of TLE to screen ASDs, of which there are many published studies - many of them with other strengths and/or weaknesses. For example, although Grabenstatter and Dudek (2019, Epilepsia) used a version of the systemic KA model to obtain dose-response data on the effects of carbamazepine on spontaneous seizures, that work required use of KA-treated rats selected to have very high rates of spontaneous seizures, which requires careful and tedious selection of animals. The ETSP has published studies with an intra-amygdala kainic acid (IA-KA) model (West et al., 2022, Exp Neurol), where the authors claim that they can use spontaneous seizures to identify ASDs for DRE; however, their lack of a drug effect of carbamazepine may have been a false negative secondary to low seizure rates. The approach described in this paper may help with confounds caused by low or variable seizure rates. These types of issues should be discussed, along with others.”

We appreciate the reviewer’s insights. In an existing model investigating spontaneous tonic-clonic seizures (such as the intra-amygdala kainate injection model), the time investment is back-loaded, requiring two to three weeks per condition while counting spontaneous seizures, which may occur only once a day. In contrast, our model requires a front-loaded time investment. Once the animals are set up, we can test multiple drugs within a few weeks, providing significant time savings. Additionally, we did not pre-screen animals in our study. Existing models often pre-select mice with high rates of spontaneous seizures, whereas in our model, seizures can be induced even in animals with few spontaneous seizures. We believe that bypassing the need for pre-screening is a key advantage of our induced seizure model.

Reviewer 3:

“(7) No mention or assessment of mouse sex as a biological variable.”

Thank you for pointing this out. Both female and male animals were included in this study: Epileptic cohort: 7 males, 3 females; Naïve cohort: 3 males, 4 females