Several EPITARGET partners took part in the Jasper's Basic Mechanisms of the Epilepsies Workshops held in Yosemite National Park, March 4-8, 2019 and presented their work in relation to EPITARGET:
Multimodal profiling of epileptogenic phenotype after TBI: machine-learning approach to discover novel biomarkers
Robert Ciszek, Niina Lapinlampi, Eppu Manninen, Pedro Andrade, Ezrie Samuel, Tomi Paananen, Noora Puhakka, Olli Gröhn, Jussi Tohka, Asla Pitkänen*
A.I.Virtanen Institute for Molecular Sciences, University of Eastern Finland, Kuopio, Finland. *presenting author: email@example.com
BACKGROUND: Traumatic brain injury (TBI) causes 10-20% of structural epilepsies. The risk of epilepsy varies between 4-53%, depending on the type and severity of TBI. Lack of biomarkers hinder the enrichment of TBI cohorts with subjects at the highest risk of epileptogenesis.
OBJECTIVE: To identify plasma molecular, brain histology, behavioral, magnetic resonance imaging, or electro-encephalogram -derived single or combinatory diagnostic biomarkers for posttraumatic epileptogenesis.
HYPOTHESIS: Parameters signaling from altered network activity alone or together with molecular/structural indicators of brain pathology will indicate subjects who will develop epilepsy with a high sensitivity and specificity.
DATASETS:TBI was induced with lateral fluid-percussion injury. Altogether 154 rats (117 TBI, 25 sham-operated controls, 14 naïve) underwent a 6-month post-TBI follow-up. Molecular, histologic, behavioral, imaging and EEG datasets were collected using common data elements and stored in RedCap electronic database. Normality of data was tested with the Shapiro–Wilk test and groups were compared with t-test or Wilcoxon-Mann-Whitney rank-sum test.
MACHINE LEARNING (ML) PIPELINE: ML pipeline consisted of feature selection, classification, and interpretation steps. Candidate features were selected by utilizing a combination of filter, wrapper and embedding methods. The selected features were used to train gradient boosted trees, support vector machines, logistic regression, k-nearest neighbor, and Gaussian naïve Bayes classifiers to differentiate epileptic rats from non-epileptic. Classification was performed using 10fold cross-validation, with classifier hyper-parameter optimization performed via 10-fold nested grid search. Classifier performance was evaluated in terms of area under the receiver operating characteristic curve (AUC), accuracy, specificity and sensitivity. The feature usage of the classifiers was interpreted to retrace the differentiability of epileptic rats with ML to potential biomarkers in the data.
CURRENT STATUS OF INTERIM BIOMARKER ANALYSIS: From TBI animals, 28% (n=33) developed epilepsy. From each animal, we derived > 300 features for subsequent analysis. We will present the interim analysis of data currently being entered to the multimodal database.
Cerebro-spinal fluid and blood microRNAs as biomarkers of epilepsy
University of Ferrara, Italy University Vita-Salute San Raffaele, Milan, Italy
There is a major unmet need for biomarkers of epilepsy. For example, biomarkers would facilitate epilepsy diagnosis and monitoring, ease antiepileptogenic and antiseizure therapy discovery and reduce the cost of clinical trials by enriching subject populations. Biofluids such as blood and cerebro-spinal fluid (CSF) offer a potential source of molecular biomarkers. In particular, biofluid microRNAs (miRNAs) possess attractive properties, including stability and cheap assay techniques. I will summarize and discuss findings from my lab and others, from both animal models and epilepsy patients. Altered miRNA profiles in biofluids have been found in association with epileptogenesis and epilepsy in animal models, and blood levels of several miRNAs were found altered in patients with epilepsy and in patients with drug-resistant compared to drug-responsive seizures. Although often associated with encouraging receiver-operating characteristic (ROC) curve analyses, these results remain preliminary. In addition, available studies display significant pre-analytical and analytical limitations. I will describe some relevant knowledge gaps and propose experiments to close these. There is an urgent need to define a strategic roadmap to facilitate epilepsy biomarker identification, characterization and clinical validation.
Microvascular Pathology in Seizures and Epileptogenesis: A Mechanism-Driven Biomarker and New Target for Treatment?
Alon Friedman1,2 and Daniela Kaufer3
1 Ben-Gurion University of the Negev, Beer-Sheva 2 Dalhousie University, Halifax, NS, 3 Berkeley University, CA
Our research on the role blood-brain barrier (BBB) dysfunction in epileptogenesis, revealed a role for leaky BBB and serum albumin in the activation of astrocytes, excitatory synaptogenesis and pathological plasticity in reducing seizure threshold and ictogenesis. Recent studies further indicate that BBB permeability is increased within minutes after seizures, highlighting a new mechanism for their role in epileptogenesis. To test the potential of BBB imaging as a biomarker for the epileptogenic brain, we developed contrast-enhanced magnetic resonance imaging (CE-MRI) protocols for the quantitative assessment of vascular permeability. In the rodent status-epilepticus model, we found that early BBB pathology in specific brain regions (e.g. pyriform cortex) is a sensitive and specific predictor (AUC=0.96, p<0.0001) for epilepsy. We further report focal BBB pathology (most common in the pyriform cortex) in 37% of dogs with epilepsy (n=46). Post-mortem analysis in 3 dogs confirmed uptake of serum albumin in astrocytes. Finally, in 25 patients with epilepsy, CE-MRI revealed interictal focal BBB pathology in >60% of patients, often co-localized with the suspected epileptic region. We suggest CE- MRI as a diagnostic approach for epilepsy. Clinical trials are expected to confirm the localization and prevalence of microvascular pathology in different epilepsy syndromes, and test the potential of vascular-targeted therapeutics in the treatment of epilepsy.
Characterisation of an infantile rat model of epileptic encephalopathy
Teresa Ravizza, PhD
Department of Neuroscience, Mario Negri Institute for Pharmacological Research, Milano, Italy
Paediatric status epilepticus (SE) results from acquired, metabolic, immune, genetic or unknown causes. We characterized an infantile rat model of de novo SE to study the pathological mechanisms ignited by unremitting seizures in the immature brain that lead to devastating sequelae including cognitive disorders and epilepsy.
Postnatal day (P)13 rats were electrode-implanted for EEG analysis and injected intra-amygdala with 2 µg kainic acid to evoke SE lasting for about 4 h. Immunohistochemical and RT-qPCR analyses showed astrocyte and microglia activation in forebrain and induction of the ictogenic molecules IL-1β and HMGB1 and the oxidative stress marker Nrf2 during one week post-SE. Degenerating neurons were detected.
A cohort of SE-exposed rats was longitudinally video-EEG monitored and exposed to MRI, and behavioral tests were done to evaluate cognition. Epilepsy developed in 60% of P65 rats 1 month post-SE which was similar in onset, severity and duration in all animals. MRI showed progressive brain atrophy before epilepsy onset likely representing a consequence of the SE at P13. Rats displayed impairment in the Morris Water Maze after epilepsy onset thus representing the encephalopathic effects of spontaneous seizures. This rat model can be exploited for mechanistic studies, to test novel antiepiletogenic or disease modifying drugs and for developing biomarkers of disease onset and progression.
Anti-inflammatory strategies for disease-modifications
Department of Neuroscience, Mario Negri Institute for Pharmacological Research, Milano, Italy
Neuroinflammation is a common hallmark of epileptogenic brain region in both human epilepsy and related animal models. Experimental evidence shows that neuroinflammation is involved in the pathogenesis of seizures. Notably, the initial clinical studies are supportive of the experimental evidence in animal models. The neuroinflammatory response in epilepsy has generated cellular and molecular targets for the development of potential new drugs, or for repurposing anti-inflammatory drugs which act on specific pathogenic mechanisms. This mechanistic approach aims at overcoming the mere symptomatic control of seizures while providing disease modification effects. In line with this intent, we demonstrated that a combination of anti-inflammatory drugs targeting the IL-1beta system and the Toll-like receptors (TLR4) prevents disease progression after transient administration either during epileptogenesis or after epilepsy onset. Similar therapeutic outcomes were attained by reducing oxidative stress, which is a phenomenon reinforcing neuroinflammation, with a combination of drugs in medical use, or by enhancing the endogenous resolution response. Finally, anti-inflammatory drugs may offer neuroprotection when timely applied during epileptogenesis and improve the neurological comorbidities. Since neuroinflammatory mediators are chiefly produced in glial cells, specific targeting of microglia and astrocytes is under investigation to elucidate the complexity and dynamics of the neuroinflammatory response in epilepsy.
Multi-targeted drug combinations for antiepileptogenesis
University of Veterinary Medicine Hannover, Department of Pharmacology, Toxicology and Pharmacy
Epilepsy is a complex network phenomenon that, as yet, cannot be prevented or cured. We recently proposed multitargeted, network-based approaches to prevent epileptogenesis by combining clinically available drugs that were rationally chosen to impact diverse epileptogenic processes. In order to test this strategy preclinically, we developed an algorithm for testing such combinations of repurposed drugs in rodents, derived from human clinical drug development phases, including pharmacokinetics and tolerability testing (Phase I and Phase IIa) as well as evaluation of efficacy (Phase IIb) in mouse and rat models of acquired epilepsy. Overall, 18 drugs, including anti-inflammatory, anti-oxidant, neuroprotective, GABA-potentiating, and anti-glutamatergic drugs, were tested in 12 drug combinations in the intrahippocampal kainate mouse model. As yet, the most effective drug combination was to combine a presynaptically acting drug (levetiracetam) with a postsynaptically acting drug (topiramate), but significant antiepileptogenic or disease-modifying effects were also obtained with several other drug combinations. Next, the most promising drug combinations will be tested in a rat model, in which epilepsy develops after traumatic brain injury. The ultimate goal is to translate our findings to clinical epilepsy prevention trials. Overall, our data provide a rich collection of network-based combinatorial therapies as a basis for antiepileptogenic efficacy testing.
to the top
Blood-Brain Barrier Leakage A New Biomarker in Transient Ischemic Attacks
In Stroke. 2019;50:00-00. doi: 10.1161/ STROKEAHA.119.025247
Yonatan Serlin, MD1*; Jonathan Ofer, BSc2*; Gal Ben-Arie, MD2*; Ronel Veksler, MD, BSc2; Gal Ifergane, MD4; Ilan Shelef, MD3; Jeffrey Minuk, MD5; Anat Horev, MD4†; Alon Friedman, MD, PhD2,6†
1 Neurology Residency Training Program (Y.S.), McGill University, Montreal, QC, Canada
2 Department of Physiology and Cell Biology, Ben-Gurion University of the Negev, Beer-Sheva, Israel (J.O., R.V., A.F.), Soroka Medical Center, Beer-Sheva, Israel
3 Department of Medical Imaging (G.B.-A., I.S.), Soroka Medical Center, Beer-Sheva, Israel
4 Department of Neurology (G.I., A.H.), Soroka Medical Center, Beer-Sheva, Israel
5 Department of Neurology and Neurosurgery, Jewish General Hospital (J.M.), McGill University, Montreal, QC, Canada
6 Department of Medical Neuroscience, Dalhousie University, Halifax, NS, Canada (A.F.).
* These authors have contributed equally.
† Drs Horev and Friedman are joint co-senior authors.
Background and Purpose—The diagnosis of transient ischemic attack is challenging. Evidence of acute ischemia on MRI diffusion-weighted imaging is highly variable and confirmed in only about one-third of patients. This study investigated the significance of blood-brain barrier dysfunction (BBBD) mapping in patients with transient neurological deficits, as a diagnostic and prognostic biomarker required for risk stratification and stroke prevention.
Methods—We used dynamic contrast-enhanced MRI to quantitatively map BBBD in a prospective cohort study of 57 patients diagnosed with transient ischemic attack/minor stroke and 50 healthy controls.
Results—Brain volume with BBBD was significantly higher in patients compared with controls (P=0.002). BBBD localization corresponded with the clinical presentation in 41 patients (72%) and was more extensive in patients with acute infarct on diffusion-weighted imaging (P=0.05). Patients who developed new stroke during follow-up had a significantly greater BBBD at the initial presentation (P=0.03) with a risk ratio of 5.35 for recurrent stroke.
Conclusions—This is the first description of the extent and localization of BBBD in patients with transient ischemic attack/ minor stroke. We propose BBBD mapping as a valuable tool for detection of subtle brain ischemia and a promising predictive biomarker required for risk stratification and stroke prevention.
to the top