An Optogenetic Kindling Model of Neocortical Epilepsy

Abstract

Epileptogenesis is the gradual process by which the healthy brain develops epilepsy. However, the neuronal circuit changes that underlie epileptogenesis are not well understood. Unfortunately, current chemically or electrically induced epilepsy models suffer from lack of cell specificity, so it is seldom known which cells were activated during epileptogenesis. We therefore sought to develop an optogenetic variant of the classical kindling model of epilepsy in which activatable cells are both genetically defined and fluorescently tagged. We briefly optogenetically activated pyramidal cells (PCs) in awake behaving mice every two days and conducted a series of experiments to validate the effectiveness of the model. Although initially inert, brief optogenetic stimuli eventually elicited seizures that increased in number and severity with additional stimulation sessions. Seizures were associated with long-lasting plasticity, but not with tissue damage or astrocyte reactivity. Once optokindled, mice retained an elevated seizure susceptibility for several weeks in the absence of additional stimulation, indicating a form of long-term sensitization. We conclude that optokindling shares many features with classical kindling, with the added benefit that the role of specific neuronal populations in epileptogenesis can be studied. Links between long-term plasticity and epilepsy can thus be elucidated.

Figure 1

Optokindling via simultaneous EEG recording and ChR2 stimulation in awake behaving animals. (A) Coronal M1 section immunostained for EYFP indicated ChR2 expression in L2/3, 5, and 6, though predominantly in L2/3. Inset shows close-up of L2/3 ChR2-expressing PCs. (B) To simultaneously activate ChR2 and acquire EEG, ferrules and recording screws were implanted bilaterally above M1, without penetrating the cortex. Fiber optic cables were air-coupled to 445-nm lasers. EEG signals were processed by an extracellular amplifier, but not pre-amplified. A computer (not shown) TTL-gated the lasers and digitized amplified EEG signals. (C) In each stimulation session, M1 was kindled (during “Induction”) with 15 bouts of 3-second-long 50-Hz bursts of 5-ms 445-nm laser pulses, divided into three sweeps delivered once a minute. Sessions were repeated at least 25 times every two days. In this sample session from a non-naïve animal, a prominent electrographic seizure was evoked in the first induction sweep. EEG responses to 30-Hz paired-pulse laser stimuli were recorded for 10 minutes before and 20 minutes after the kindling induction. Data is represented as mean ± SEM here and throughout the manuscript unless stated otherwise. Inset: Paired-pulse EEG responses before (red) and after (blue) indicated a change in EEG dynamics but not amplitude.

Figure 2

Optokindling and classical kindling share hallmark seizure features. (A) Optokindling required both laser stimulation and ChR2 expression. The number of stimulated ChR2-expressing animals that developed electrographic seizures (9 of 12 animals) was higher than unstimulated ChR2 controls (0 of 4) and stimulated no-ChR2 controls (0 of 5) (Fisher’s exact test, p = 0.001). (B) The behavioral seizure severity, as measured by a modified Racine score, increased over sessions (Spearman’s rank correlation test, rho = 0.957, p < 0.001, n = 12 animals). (C) An increasing number of electrographic seizures developed in stimulated ChR2-expressing animals (Spearman’s rank correlation test, rho = 0.862, p < 0.001, n = 30 seizures from 9 animals). (D) Once seizures arose, seizure duration gradually increased over sessions (r = 0.674, p < 0.001, n = 30 seizures from 9 animals). Open circles represent individual seizures, whereas closed circles are averages over the one session before and after. Linear fits are made to the entire data set. (E) Seizure threshold, measured as time to electrographic seizure onset from start of induction, decreased across sessions (r = −0.478, p = 0.008, n = 30 seizures from 9 animals; symbols as in D). Gray boxes denote the three 50-Hz induction epochs. Linear fits were made to the individual data points, not binned data.

Figure 3

Kindled animals retained a long-term increase in seizure susceptibility. (A) Rekindled mice had more severe behavioral seizures compared to naïve animals (Kruskal-Wallis test, p < 0.001, n = 5 rekindled animals, n = 5 naïve animals). Racine scores from the eight rekindling sessions (blue) were compared with the first eight sessions in naïve animals (red). (B) Rekindled animals (“rek”) had more severe behavioral seizures than naïve animals (“kin”; Student’s paired t test, p = 0.009, n = 5 animals). (C) Electrographic seizures in rekindled animals occurred in earlier sessions than in naïve mice (Mann-Whitney’s U = 183.5, p < 0.001, n = 4 animals). (D) Electrographic seizures in rekindled mice occurred after fewer sessions than in naïve animals (Student’s paired t test, p = 0.003, n = 4 animals). (E) Electrographic seizure duration was indistinguishable between kindled and rekindled animals (Student’s paired t test, p = 0.43, n = 4 animals). (F) The seizure threshold, as measured by time to electrographic seizure onset after start of light stimulation (see Fig. 1C), was not different in kindled and rekindled mice (Student’s paired t test, p = 0.86, n = 4 animals).

Figure 4

Immunohistology revealed no astrocytic reactivity or neuronal loss. (A) Sample coronal slices from an optogenetically kindled animal stained for EYFP to tag ChR2-expressing cells (“ChR2”), GFAP to label for astrocytic reactivity (“GFAP”), and NeuN to assess neuronal cell body counts (“NeuN”). (B) Astrocytic reactivity, as indicated by upregulated GFAP expression, was indistinguishable between animals with evoked seizures (“ChR2 stim”, n = 59 sections) and the two control groups (“no stim ChR2”, n = 42; “stim no ChR2”, n = 19; one-way ANOVA, p = 0.11). (C) Neuronal cell density did not differ between the three animal cohorts (“ChR2 stim”, n = 23, “no stim ChR2”, n = 14 and “stim no ChR2”, n = 13, one-way ANOVA, p = 0.10; compare Fig. 2A).

Figure 5

Evoked EEG responses exhibited long-term plasticity. (A) Example first and second EEG responses due to paired-pulse laser stimulation averaged during baseline periods before and after induction in one session. (B) Ensemble EEG response amplitude averaged across all sessions in one animal showed a within-session potentiation of the second but not the first response. (C) The magnitude of plasticity of the first EEG response remained unaffected across sessions and animals (left, p = 0.32, n = 9 stim ChR2 animals vs. n = 4 no stim ChR2 animals, Friedman test). The pre-induction baseline first response remained at the same amplitude across sessions and animals (right, responses normalized to the first two sessions indicated by vertical dashed lines, p = 0.99, stim ChR2 vs. no stim ChR2, Friedman test). Red: stim ChR2, gray: no stim ChR2. (D) The magnitude of plasticity of the second EEG response remained elevated across sessions and animals (left, p < 0.001, stim ChR2 vs. no stim ChR2, Friedman test), although waned in the first five sessions. The pre-induction baseline second response remained potentiated across sessions and animals (right, normalized as in C, p < 0.001, stim ChR2 vs. no stim ChR2, Friedman test), although seemed to saturate, perhaps as plasticity waned (left). Blue: stim ChR2, gray: no stim ChR2.

Figure 6

High-frequency oscillations peak before low-frequency oscillations. (A) Sample EEG trace illustrating an optogenetically evoked electrographic seizure, with automatically detected seizure start and end indicated by vertical dashed lines (see Methods and Supplementary Fig. S3). Note that seizure begins with rapid oscillations and ends with slow oscillations. Laser-light stimulation bouts are indicated in blue (Fig. 1C and Supplementary Fig. S3). (B) Wigner transform of electrographic seizure in (A) showing graded decay in high-frequency components as well as a gradual increase in low-frequency power. (C) FFT power z-score traces for delta (red, 4–8 Hz) and ripple bands (blue, 80–250 Hz) derived from the EEG trace in (A) shows how high-frequency oscillations peak (blue arrow) before their low-frequency counterparts (red arrow). (D) Peak power of high frequencies occurred earlier compared with that of low frequencies (6.4 ± 1 sec, n = 51 seizures from 9 animals, p < 0.001, t test for difference of mean compared to time zero). Data points indicate the difference between the low and high frequency peak times for individual seizures. Box plot shows first quartile, median, and third quartile with whiskers denoting one standard deviation from the mean.