Authors :
Presenting Author: Andrew Knox, MD, M.S. – University of Wisconsin School of Medicine and Public Health
Christopher Thompson, Ph.D. – Research Assistant Professor, Pharmacology, Northwestern University Feinberg School of Medicine; Bethany Stieve, Ph.D. – Research Associate, Neurology, University of Wisconsin, Madison; Dillon Scott, B.S. – Research Specialist, Neurology, University of Wisconsin, Madison; Alfred George, M.D. – Professor and Chair, Pharmacology, Northwestern University Feinberg School of Medicine
Rationale:
Pathogenic SCN1A variants are a common cause of genetic epilepsy, and in vitro studies can determine the functional effects of variants. Descriptions of channel function are usually dichotomized as “gain of function” or “loss of function,” implying two different pathways to ictogenesis. We hypothesize that a lowered threshold for depolarization block may be a common pathway by which both variant types lead to seizures. We used an in-silico interneuron model to test the effects of gain and loss of function variants on neuronal firing.
Methods:
In vitro experimental data from three variant SCN1A channels (KO - complete loss of function, I1356M – partial loss of function, and T226M – gain of function) were incorporated into a detailed multicompartment interneuron model with 11 channels, simulated Ca dynamics, and a complex dendritic tree. Excitatory Post-Synaptic Potentials (EPSPs) of increasing frequency, both alone and with Inhibitory Post-Synaptic Potentials (IPSPs) of increasing frequency at a ratio of 3:1 at the neuron soma were used to stimulate neuronal firing in 200 simulations. In a second set of 100 simulations, EPSPs and IPSPs of increasing frequency at a ratio of 3:1 stimulated 12 different randomly selected dendritic compartments.
Results:
Stimulus frequency sufficient to produce a single spike did not vary significantly between WT and variant neurons (Fig 1, Kruskal-Wallis ANOVA, p=.132 for EPSPs only, p = 0.131 for EPSPs and IPSPs). Firing frequency did not substantially differ between neuron types up to the point of depolarization block (Fig 2). In contrast, depolarization block occurred at significantly lower stimulus frequency for all variants (Fig 1, Kruskal-Wallis ANOVA, p = 6.37*10-76 for EPSPs alone, p=2.80*10-58 for EPSPs and IPSPs). Similar results were observed in the second set of simulations (Fig 2), with the additional finding that simulations in which dendritic compartments receiving IPSPs were more proximal to the soma relative to those receiving EPSPs required higher stimulus frequency to achieve initial spiking and depolarization block.
Conclusions:
We discovered that a lowered threshold for depolarization block occurs among SCN1A variants across a variety of stimulation paradigms, regardless of location of stimulation within the dendritic tree. This finding suggests that depolarization block may be the common pathway by which both gain of function and loss of function variants initiate seizures. It is plausible that selective interneuron depolarization block blunts network inhibition, which in turns leads to additional EPSPs and worsening depolarization block in interneurons, establishing a positive feedback loop leading to ictogenesis. In vivo or in vitro studies that show selective depolarization block in interneurons alone could confirm this hypothesis and provide a specific target for new therapies; in-silico studies could help optimize dosing of therapies meant to prevent ictogenesis by modulating sodium channel function.
Funding:
This work was supported by the Lily’s Grace Grant Fund and an NINDS Center Without Walls Grant (U54-HL108874).