Abstract
Electrocorticographic Dynamics as a Novel Biomarker in Five Models of Epileptogenesis
Milikovsky DZ, Weissberg I, Kamintsky L, Lippmann K, Schefenbauer O, Frigerio F, Rizzi M, Sheintuch L, Zelig D, Ofer J, Vezzani A, Friedman A. J Neurosci 2017;37:4450–4461.
Postinjury epilepsy (PIE) is a devastating sequela of various brain insults. While recent studies offer novel insights into the mechanisms underlying epileptogenesis and discover potential preventive treatments, the lack of PIE biomarkers hinders the clinical implementation of such treatments. Here we explored the biomarker potential of different electrographic features in five models of PIE. Electrocorticographic or intrahippocampal recordings of epileptogenesis (from the insult to the first spontaneous seizure) from two laboratories were analyzed in three mouse and two rat PIE models. Time, frequency, and fractal and nonlinear properties of the signals were examined, in addition to the daily rate of epileptiform spikes, the relative power of five frequency bands (theta, alpha, beta, low gamma, and high gamma) and the dynamics of these features over time. During the latent pre-seizure period, epileptiform spikes were more frequent in epileptic compared with nonepileptic rodents; however, this feature showed limited predictive power due to high inter- and intra-animal variability. While nondynamic rhythmic representation failed to predict epilepsy, the dynamics of the theta band were found to predict PIE with a sensitivity and specificity of >90%. Moreover, theta dynamics were found to be inversely correlated with the latency period (and thus predict the onset of seizures) and with the power change of the high-gamma rhythm. In addition, changes in theta band power during epileptogenesis were associated with altered locomotor activity and distorted circadian rhythm. These results suggest that changes in theta band during the epileptogenic period may serve as a diagnostic biomarker for epileptogenesis, able to predict the future onset of spontaneous seizures.
Commentary
Before the concept of epileptogenesis as a process leading to epilepsy had been established in the late eighties/early nineties there was epilepsy as a disease (i.e., a final product) and search for its biomarkers. The search was focused on such biomarkers that could clearly distinguish between those patients who had epilepsy and those who did not. As this was easiest to do in a setting with similar genetic background (i.e., in families with epilepsy patients), these studies were predominantly genetic (1). In humans, they were focusing on discovery of “epilepsy genes” (2) and in animal models on discovery of translatable biomarker genes or MRI findings (3–5). Additional studies centered on the prediction of epilepsy refractoriness (6). Later on, biomarkers of epileptogenesis were sought for early detection of the process. The ultimate goal was and still is to distinguish those patients in whom the initial (precipitating) event would eventually lead to development of epilepsy from those without any future epilepsy. The knowledge of ongoing epileptogenesis could make it possible to intervene with disease-modifying or even antiepileptogenic treatments (once we have at least one, of course). Covering all epilepsies with biomarkers for epileptogenesis is still a futuristic view as the precipitating events may be diverse, not identified or even not identifiable. It is therefore reasonable to focus on those epileptic processes following a clearly defined injurious (precipitating) event, such as stroke or brain trauma, which provides a fixed point for the onset of the process of epileptogenesis (7).
Post-injury epilepsy (covering both post-stroke and post-traumatic brain injury [TBI] epilepsy) is difficult to predict and similarly difficult to treat. Only a small proportion of patients after TBI develop the disease and the outcome is also dependent on the severity of brain injury: at 1 year after the trauma, epilepsy will develop in approximately 6% of patients with severe brain injury (characterized by brain contusion or intracranial hematoma or loss of consciousness/amnesia longer than 24 hours). This number will eventually rise to 13% in 10 years and to 17% in 20 years after the TBI (8). On the other hand, the number of patients with TBI due to various reasons is significantly increasing in the population (9). Development of post-stroke epilepsy is similarly difficult to estimate. Proportion of patients with epilepsy at 1 year after stroke (mostly ischemic in nature in the adult population) is anywhere between 2.5 and 15 percent. Obviously, not every stroke will lead to precipitation of epilepsy as the strokes are not the same: 10 years after a stroke in anterior circulation more than 40% of patients develop post-stroke epilepsy. On the other hand, 10 years since the event strokes in posterior circulation will result in only 5% occurrence of epilepsy (10). Numbers illustrate that these two types of epilepsy summarized as post-injury, are relatively frequent to be worrisome, but not so common to warrant blanket preventative approaches with doubtful efficacy and some side effects (7, 11). This picture may change with current and future findings that modulation of vascular pathology and inflammatory signaling after injury may significantly improve the outcome (12); hence, the search for biomarkers that may provide hints as to which patients from those affected by either stroke or TBI will eventually develop epilepsy. The importance of discovery of reliable biomarkers for epileptogenesis cannot be emphasized enough as our current understanding of the epileptogenic process is that subconvulsive seizures (further solidifying epilepsy) may be present well before the first clinical seizure occurs (13).
Milikovsky et al. therefore decided to analyze electrocorticograms (ECoG) in five models of epileptogenesis relevant to stroke or TBI. All models had a clearly defined initial precipitating event. Three of the models were induced in mice, specifically by intracerebroventricular administration of albumin (resulting in epilepsy in 12/14 subjects), intra-cerebroventricular transforming growth factor-β (3/3 having epilepsy), and intracerebroventricular interleukin-6 (3/5). For further confirmation and species independence, two models were induced in rats: photothrombosis accomplished by skull illumination after intravenous administration of rose bengal and electrically-induced status epilepticus (EISE) in the ventral hippocampus (90-minute, 50-Hz constant current electrical stimulation). The authors looked at many EEG parameters (total of 22) and confirmed some previous negative findings (e.g., the number of interictal spikes has no predictive value for development of epilepsy). As theta rhythm has been implicated in epilepsy and cognitive decline in an epilepsy model (14, 15), they further concentrated on changes in the theta rhythm during the time course of epileptogenesis. Spectral analysis of ECoG focused on the power of theta frequency revealed that there was a temporal decline in the relative theta-band power in those subject that eventually developed epilepsy compared with those who (from the induction groups) did not or compared with controls with no precipitating event. This decline was also expressed as increased theta-power slope (over time). There was also a positive correlation between the absolute value of this theta slope parameter and time to the first seizure after the initial impact. Finally in epileptic mice, there was breakdown of circadian cyclicity of theta rhythm compared with nonepileptic mice. Findings in rat models were consistent with those in mice. In the robust EISE model, in which all subjects eventually would develop seizures, comparison of theta slopes had to be done between subjects developing early seizures versus late seizures. Interestingly, in both rat models there was increase in gamma (60–100 Hz) activity with negative theta-higher gamma correlation (cf. (16)) for high frequency EEG activity in human and animal epilepsy).
The findings of the study are quite important as first, a biomarker for the progression of epileptogenesis was sorely missing and second, any approach providing information as to whether the initial impact is associated with epilepsy development is extremely valuable. Interestingly, findings of increased gamma activity with epileptogenesis progression correlate with recent findings showing high frequency oscillations in the cortical TBI core and adjacent areas recorded in 60% of rats after TBI (17). These high-frequency oscillations may also serve as predictors of ongoing epileptogenesis. The authors correctly accentuate the fact that persistence pays off: recording theta activity at a single time point did not provide any meaningful data; only reconnaissance over time was fruitful. On the other hand, some of their small group sizes are concerning. A bonus point worth emphasizing is that greater theta slopes are associated with shorter latency to the onset of spontaneous seizures, indicating that this biomarker may provide additional information about the dynamics of epileptogenesis. As the authors focus on theta activity, they need to reconcile the fact that in their rodent models, theta is the leading EEG (ECoG) rhythm., whereas in humans this role belongs to the alpha rhythm. Indeed in rodents signals are recorded directly from the surface of the cortex while in humans, there are scalp EEG electrodes. So the questions are – Can a similar outcome be observed with the development of alpha rhythm in humans? Or alternatively, is this feature something the human theta rhythm will support as well, so the spectral power regress is frequency-linked and independent of the functional relevance of individual EEG rhythms? Is it possible to compare easily surface ECoG and scalp EEG? Prospective and repetitive EEG recordings in clinical trials in patients after TBI or stroke may help resolve these issues in human epilepsy.