Final Report Summary - EPITARGET (Targets and biomarkers for antiepileptogenesis)
Executive Summary:
EPITARGET designed a unique set of first in its kind preclinical epilepsy-related common data elements (CDEs) and case report forms (CRFs) for different animal models and experimental procedures (www.epitarget.eu). These data were utilised through electronic RedCap database creatred at 01-ULUND server, where information on animals and performed procedures are stored and shared, including granting access of whole research community in the field. This includes data on plasma protein biomarkers, microRNAs, imaging MRI and PET biomarkers. Moreover, EPITARGET developed organic transistors that can be used to both stimulate neurons and record their activity. These devices are now being adapted to be tested in the operating theatres as MRI compatible electrodes to aid combinatorial biomarker identification. Moreover, these electrodes were shown to even stop seizures.
EPITARGET identified new targets for treatment, include the immunoproteasome, metalloproteinases, specialised pro-resolving mediators promoting resolution of inflammation, glycogen synthase kinase 3β, the chemokine CX3CR1, the adhesion molecule N-cadherin and oxidative stress-related molecules. These were also validated in human specimens from patients with drug-resistant epilepsies. Pharmacological targeting of such molecules lead to a >50% reduction in the frequency of spontaneous seizures, improvement in neuronal cell loss and rescue of cognitive deficits in rodents. EPITARGET also identified five disease-modifying treatments based on targeting neuroinflammation and oxidative stress. Combined anti-inflammatory or anti-oxidant drugs at onset of epileptogenesis significantly improved long-term disease outcomes, i.e. the frequency and/or duration of spontaneous seizures, neuronal cell loss and memory deficits. These therapeutic effects persisted long after drug treatment indicating disease-modification effects. These findings have high translational value since some of these drugs are medically used for other neurological or inflammatory diseases.
For more effective drug discovery and increased translational impact, EPITARGET designed “medium-throughput drug screening” model with intrahippocampal kainate in mice, followed by the refined intrahippocampal kainate model in rats. For the same purpose, EPITARGET developed the Theiler’s viral encephalitis epilepsy model in mice, and a model of BBB dysfunction induced by intracerebral albumin infusion in rodents. These model are a significant contribution to the community to facilitate trslational therapeutic value of preclinical studies in epilepsy.
EPITARGET designed and developed an extensive portfolio of Herpes Simplex Virus (HSV)-based vectors that are suitable for combinatorial gene therapy of epilepsy. This portfolio includes amplicon- and replication defective-vectors, all safe and capable of short-term or long-term transgene expression, i.e. prospectively applicable for use during epileptogenesis, to prevent development or progression of epilepsy, or in drug-resistant epilepsy, to control seizures. Selected combinations of therapeutic genes have been already engineered in these vectors and are ready for testing in epilepsy models.
EPITARGET has built up an extensive biosample resource comprising sera, CSF as well as biopsy brain tissue from patients undergoing epilepsy surgery for seizure control. The biobank is complemented by harmonized clinical data, altogether allowing for the extensively analyses, e.g. based on bioinformatics and systems biology. This biobank was used to identify substantially increased expression levels for MTF1 and CaV3.2 in human pharmacoresistant TLE patients. This finding may facilitate the development of new therapies focusing on this ‘MTF1-CaV3.2-neuronal bursting’ mechanism.
EPITARGET partners carried out comprehensive gene-regulatory network analyses in human brain tissue samples of temporal lobe epilepsy (TLE) patients, employing systems genetics approaches to characterize the genetic regulation of pathophysiological pathways in TLE. Using surgically acquired hippocampi from pharmacoresistant TLE patients included in the EPITARGET database, gene-regulatory network was identified containing a specialized, highly expressed transcriptional module encoding proconvulsive cytokines and Toll-like receptor signalling genes. Complementary RNA sequencing analysis in a mouse model of TLE demonstrated the pro-convulsive module specific to the epileptic hippocampus. In the respective set of TLE patients from the EPITARGET database, revealed that sestrin-3 (SESN3) positively regulates the module in macrophages, microglia and neurons, and SENS3 knockdown in a zebrafish epilepsy model attenuated chemically induced behavioral seizures in vivo.
EPITARGET has demonstrated that dysregulation of the (immuno)proteasome system during epileptogenesis occurs in epileptic rat model and human brain and may serve as biomarker as well as therapy target in the future. The consortium also demonstrated that TBI animal models and correspondingly human epileptic hippocampi provide new pathogenetic insights in how albumin extravasation that occurs upon BBB dysfunction in various brain injuries and ways to predispose neural circuitry to the development of chronic inhibition deficits and suggest new tissue biomarkers and a novel rationale for therapy targets in TBI-induced epileptogenesis.
EPITARGET established a systems genetics approach in animal models for identifying cell membrane receptors with a direction-specified influence over disease-related mRNA profiles. The conservation of several mRNA modules across human and mouse TLE models was established and provides an independent line of evidence for the validity of these modules to assess seizure risks and points to innovative treatment options. In particular, microglia-related combinatorial mRNA module was shown to correlate strongly with seizure activity. The analysis and bioinformatic characterization of potential therapeutic targets predicted the tyrosine kinase receptor Csf1R in complementary mouse model and human epileptic hippocampi. The released corresponding mRNA signature of module 18 has a strong potential in this regard.
Project Context and Objectives:
EPITARGET is a large European consortium of universities, research institutions and SMEs dedicated to understanding complex mechanisms and identify biomarkers of epileptogenesis to improve treatment and diagnosis of epilepsy. The overall scientific objectives of EPITARGET are:
Using multidisciplinary strategies of basic, preclinical and clinical research to:
1. Identify novel biomarkers and their combinations that will define the different stages of epileptogenesis and predict/diagnose early and late stages of the evolution of the disease. This will pave the way to improved diagnostics and better patient stratification, as well as development of patient-specific preventive strategies. In this objective EPITARGET will identify at least 2 biomarkers that in combination will predict whether epileptogenesis is triggered in the post-insult period and will stratify those individual subjects that are at risk of developing epilepsy.
2. Unravel the complex patho-physiology of epileptogenesis and design new, disease-modifying combinatorial treatment strategies specifically targeted to the different stages of epileptogenesis. These novel treatment strategies are expected to be capable of preventing the development of epilepsy in at-risk patients and stop its progression after the onset of the disease. Different epileptogenesis-associated processes, such as brain damage and structural/functional reorganization, neurogenesis, brain-blood barrier (BBB) dysfunction, inflammation, free radical formation, genetic and epigenetic alterations will be individually addressed and specific treatments directly or indirectly targeting these mechanisms will be combined. The choice of the best combinatorial therapeutic strategy will be assisted by systems biology approaches. This strategy will help to identify key aetiological factors, thus limiting redundancy in the choice of multiple drug targets and facilitating a rational drug discovery process. In this objective EPITARGET will identify at least 2 patho-physiological mechanisms of epileptogenesis that will be targeted in combination in animals at risk of developing epilepsy (post-insult period).
3. Translate the knowledge obtained in experimental models to patients in order to improve diagnosis, achieve better patient stratification, and develop new antiepileptogenic treatments and means to predict their efficacy. The objective is to verify and validate biomarkers in blood and brain tissue samples obtained from patients after potentially epileptogenic brain insults. These tissues include traumatic brain injury (TBI) and post-mortem brain tissue from patients that had status epilepticus (SE) or TBI (early-stage epileptogenesis), as well as from those with chronic epilepsy. These specimens are the best closest match to the experimental brain tissue of post-insult epileptogenesis in animal models. In this objective EPITARGET will perform clinical validation of the combinatorial biomarker approach (objective 1), and thereby obtain data from patients to prove the concept for future clinical applications.
To achieve these objectives and maximise outcome, EPITARGET tackles the complexity of epileptogenesis by adopting concerted and complementary actions of participating partners, attacking various aspects of epileptogenesis at different levels. EPITARGET combines a powerful arsenal of both established and innovative multidisciplinary research strategies, tools and platforms. Analysis of data is assisted by creating an animal and human database, and a bioinformatics approach within the consortium.
Project Results:
SEE REPORT ATTACHED
1.3.1. Identifying novel biomarkers and their combinations in animal models
Globally, an estimated 2.4 million people are diagnosed with epilepsy each year. Thus, a new person is diagnosed with epilepsy every 13 seconds (http://www.who.int/mediacentre/factsheets/fs999/en). In 60% of those affected, epileptogenesis is initiated by structural causes such as traumatic brain injury (TBI) or stroke. Over 40 hypothesis-driven monotherapy approaches have demonstrated some disease-modifying effects in animal models of epileptogenesis. Currently, however, no clinical treatments are available to stop or alleviate epileptogenesis in at-risk patients or to alleviate the course of epilepsy after its diagnosis.
One reason for the stalled progression of compounds showing proof-of-concept evidence in animal models to clinical antiepileptogenesis trials is the lack of diagnostic biomarkers for epileptogenesis that could be used to stratify patient populations for antiepileptogenesis trials and reduce study costs. Also, little effort has been aimed at identifying and using response biomarkers that could inform about target engagement in the early treatment development phase and provide an early go/no-go signal for further development.
One main objective of EPITARGET was to identify single or combinatorial biomarkers for epileptogenesis, that is: (a) prognostic biomarkers for the development of epilepsy after epileptogenic brain insults, (b) diagnostic biomarkers of ongoing epileptogenesis in one or several epileptogenic etiologies. The modalities investigated were plasma protein and microRNA biomarkers and brain MRI and PET imaging biomarkers. All these modalities are relatively noninvasive and widely available. In addition, WP1 developed novel tools for biomarker analysis.
EPITARGET partners 01-ULUND, 04-UEF, 05-IRFMN, 06-TiHo, 07-UCL, 09-NENCKI, 10-AMU, 11-IC, 12-UNIFER, 13-AMC, 17-MHH have worked as a team, generating multiple animal models and state-of-art novel tools for biomarker analysis, and complicated data analyses. There have been significant collaborations with WP1, 2, 3 and 5 in implementation and dissemination of data obtained in WP1.
1.3.1.1. Generation of an Epilepsy Preclinical Biomarker Bank (EPBB)
In spirit to perform a statistically powered preclinical analysis of the sensitivity and specificity of plasma and imaging biomarkers for epileptogenesis, EPITARGET decided to harmonize the methodologies and data analysis. This was a pioneering effort considering the preclinical research in Europe.
EPITARGET (a) generated the first preclinical epilepsy-related common data elements (CDEs) and case report forms (CRFs) for different animal models and experimental procedures (www.epitarget.eu) (b) designed electronic RedCap database on 01-ULUND server to store information on animals and procedures performed. Research data is stored at each study site, (c) provided access to RedCap for Epitarget users, (d) trained Epitarget users to enter and retrieve data in RedCap
1.3.1.2. Plasma protein biomarkers
Our initial goal was to multiplex 11 plasma proteins: SBDP150 and HMGB1 (acute necrosis), SBDP120 (delayed apoptosis), S100β (brain injury), GFAP (astrogliosis), UCH-L1 and NSE (neuronal cell body injury), MAP2 (dendritic injury), MBP (myelin injury), collagen VI (extracellular matrix), and EMAP-II/III-6 (microglia), in order to perform analysis on multiple markers in small plasma volumes available from rat tail vein. However, development of the multiplex protein analysis platform by Bioviron did not work out, and EPITARGET partners decided to continue with a hypothesis-driven approach, assessing one plasma biomarker candidate at a time by using ELISA or Simoa. After a wide range of technical difficulties (no antibody/assay available; assay described in literature but not showing protein expression at EPITARGET sample time points), we reviewed new information in the literature and took advantage of the mechanistic work of EPITARGET partners in order to identify novel plasma protein biomarkers and optimize their analysis. These included Collagen VI, HMGB1, clusterin, Sushi-repeat protein X-linked 2, neurofilament light. Of the candidates tested in rodent/human plasma samples and rat/human brain tissue, HMGB1 is at the most advanced stage, showing promise to predict epileptogenesis after status epilepticus in animal models.
1.3.1.3. Plasma microRNAs
In the frame of EPITARGET, we designed and conducted an unbiased, multi-centre, longitudinal study on circulating miRNAs during epileptogenesis, using four different refined models of epilepsy (lithium-pilocarpine, perforant path stimulation, amygdala stimulation and lateral fluid-percussion injury), in which only a subset of animals develops spontaneous seizures. Altogether 157 eligible samples were sent to GenomeScan (Leiden, the Netherlands) for small RNA-seq using an Illumina platform. No common miRNA alterations were observed across all 4 models. Only a few were common to 2 models. Therefore, it was decided to run a meta-analysis of the data. As there was a low variation between the effect sizes, a fixed effect model (FEM) was considered appropriate for analysis. FEM meta-analysis identified 5 miRNAs that were significantly different between animals that would subsequently become epileptic and those that would not become epileptic at 2 d post insult. Three of these were already known (namely miR-129-5p, miR-138-5p and miR-3085) and two were not (ENSRNOG00000054305 and ENSRNOG00000054389). Interestingly, miR-129-5p inhibition has been associated with synaptic downscaling in vitro and reduced epileptic seizure severity in vivo, and miR-129-5p is upregulated in different epileptic models. In addition, miR-138-5p is a potential regulator of memory performance, a candidate regulator of P53 (involved in cell cycle, proliferation and apoptosis), and it is down-regulated in several phases of epilepsy.
Based on these data, we performed a technical verification on the 3 known miRNAs using digital droplet PCR (ddPCR). Residual aliquots of the samples employed for RNA-seq were RNA retro-transcribed and ddPCR analysed. Due to delays caused by unforeseen difficulties in animal modelling and miRNA assays, technical validation and analyses in independent cohorts of animals will be being done after the EPITARGET funding period using the additional funds the EPITARGET partners have been able to get. If these further experiments will provide positive results, some of these miRNAs may become eligible for testing in human epilepsy.
1.3.1.4. MRI biomarkers
Multiparametric MRI data was collected in several timepoints along the evolution of epileptogenesis. The analysis of the data turned out to be more challenging as anticipated and therefore a part of resource originally allocated to Task 5.1 (Development of novel MRI tools) were directed to advanced analysis of rich longitudinal MRI data obtained in biomarker study in Task 2, instead of doing a separate longitudinal biomarker study for new techniques.
At the first stage, analysis was focused to T2 values in the perilesional cortex. Because of variable lesion size lesions and progression conventional co-registration-based image analysis pipelines were non-applicable. For visualization purposes, unfolded cortical maps of T2 relaxation were computed. For these maps, the T2 relaxation coefficient was measured along the curvature of the cortex and over the imaging slices, while being averaged over the cortical thickness. Consequently, T2 relaxation was observed on the flattened cortex, allowing convenient visualization.
The differently classified pathological volumes computed from T2 MRI measurements 2 d and 7 d after TBI were used to fit a multiple linear regression model to explain necrotic lesion volume 21 d after TBI The result implies that an early (< 1 week after TBI) T2 MRI measurement is highly indicative of later necrotic tissue volume. The result implies that cortical lesion volume is not a sufficient metric to explain posttraumatic epileptogenesis.
In the second part, analysis was targeted to the thalamus. T2 relaxation time on days 2, 7, and 21 and DTI parameters on days 7 and 21 after TBI were utilized to model epileptogenesis using various logistic regression models. The different parameter maps at different time points were registered to a common template for a region-of-interest (ROI) type analysis. The approaches included using a single MRI parameter map at a single time point to model epileptogenesis and using multiple parameter maps at multiple time points to model epileptogenesis.
Based on the findings, analysis of multiple MRI parameters in the thalamus may be combined to form a model to distinguish between individuals with and without epileptogenesis (Figure 1). However, the approach used had relatively low sensitivity and requires further development or must be combined with other biomarkers to achieve clinical relevance. The manuscript of the results is under preparation.
Figure 1. Modelling epileptogenesis by using the mean and standard deviation of T2 parameter values on days 2, 7, and 21 as well as of diffusion parameters CL and CP on days 7 and 21 in a logistic regression model produced a statistically significant regression model. (A) The model produced a c-statistic (area under a receiver operating characteristic curve, AUC) of 0.89 with a 95% confidence interval (CI) of 0.81-0.94. To assess the possibility of overfitting, leave-one-out (LOO) and K-fold cross-validation (CV) methods were used. (B) LOO CV produced a c-statistic of 0.75 (95% CI 0.65-0.84). (C) K-fold CV produced a c-statistic of 0.70 (95% CI 0.58-0.79).
1.3.1.5. PET biomarkers
Identification of prognostic and predictive biomarkers for epileptogenesis and treatment response.
EPITARGET investigated various positron emission tomography (PET), single photon emission computed tomography (SPECT) and magnetic resonance imaging (MRI) protocols throughout experimental epileptogenesis (e.g. TSPO ligands for [11C]PK11195 PET and [18F]GE180 PET (microglia activation), [18F]FDG PET (glucose metabolism), [18F]FET (amino acid turnover), [68Ga]DTPA PET, [99mTc]DTPA SPECT, [18F]FMZ (GABAA receptor expression), gadolinium-DTPA-enhanced T1 MRI (blood-brain barrier (BBB) integrity) and T2 MRI (e.g. cell edema)). We could acquire distinct time profiles of various epileptogenesis-associated changes (exemplarily illustrated in Figure 2 for the pilocarpine rat model).
Figure 2. Imaging-based analysis of time profiles of epileptogenesis-associated changes. To facilitate a better overview, data were normalized to baseline.
Furthermore, the imaging protocols were investigated in different post status epilepticus animal models of epileptogenesis including the systemic pilocarpine rat post-SE model as well as the focal intra-hippocampal kainate post-SE mouse model. As neuro-inflammation and BBB integrity are promising candidates for biomarkers, but also for treatments, the time profiles were successfully adapted to treatment protocols in order to quantify acute treatment effects of anti-epileptogenic drug candidates, including dexamethasone, minocycline, fingolimod (FTY720) or curcumin.
17-MHH also performed a multi-tracer study investigating the suitability of various imaging and behavioural markers regarding their predictive biomarker properties. The applied study design was adapted to the 04-UEF design for TBI-induced epileptogenesis. A modified pilocarpine rat model, developed by 06-TiHo within EPITARGET recently, in which only part of the animals develop chronic epilepsy, was chosen for this study. Imaging was performed at 17-MHH, whereas behavioural testing and video-EEG monitoring was carried out at 06-TiHo. Both glucose hypometabolism and neuroinflammation (Figure 3) were able to detect differences between seizing and non-seizing animals using voxel-based analysis (statistical parametric mapping (SPM)). In addition, HMGB1 blood values at day 8 post SE (Figure 3A) differentiated well between seizing and non-seizing animals using the ROC analysis (Figure 3C) AUC = 0.86 p = 0.02) resulting in a cut-off value of 4 ng/ml. This result was even improved when combining HMGB1 blood levels with TSPO PET (AUC = 0.91 p = 0.02). This approach provided a promising candidate protocol for translation to the clinic.
Figure 3. Analysis of high mobility group box 1 protein (HMGB1) blood levels and TSPO PET (neuroinflammation). (A) Blood levels were analysed by ELISA at baseline, 1 day and 8 days post SE (upper graph). Data is presented as mean ± SD, asterisks indicate significant differences (p < 0.05 one-way ANOVA and Tukey’s multiple comparison post hoc test for comparison between testing days and unpaired Student‘s t-test for comparison BL with following measurements in one group). (B) Imaging results are presented as statistical parametric maps. Voxel-wise comparisons (Student’s t test) of TSPO binding on days 12 to 14 post SE (volume of distribution Vt, middle box) are shown and only clusters with significantly different voxels are visualized (p < 0.05 minimum cluster size of 100 voxels). Hot and cold scale represents the p value for each voxel. (C) Receiver operator characteristic (ROC, lower graph) curves comparing seizing and non-seizing rats for thalamic TSPO PET (18F-GE180 Vt), HMGB1 levels at day 8 post SE, or combination of both.
A validated platform for in vitro autoradiography based on existing ([18F]GE180, [18F]FMZ ) and novel PET radiotracers was developed by EPITARGET ([18F]FPEB (glutamatergic neuro-receptor MGluR5), [18F]deprenyl (MAO-B, astrocyte activation)) was established and applied to brain slices of rat and mouse models of epileptogenesis provided, but also to the P21 rat model. Here, [18F]deprenyl could differentiate between HMGB1-high animals (developing epilepsy) and HMGB1-low animals (not developing epilepsy).
1.3.1.6. Novel MRI and PET tools and novel in vivo technologies for biomarker identification
Novel MRI tools.
In this subtask EPITARGET focused on developing novel MRI methods for detecting structural abnormalities after potentially epileptogenic brain injuries. Especially, the focus was in white matter abnormalities. We were able to develop two novel approaches, based on phase imaging or magnetization transfer with zero-echo time detection (SWIFT), and relaxation in the rotating frame. The techniques were tested for detection of white matter damage in traumatic brain injury and/or in demyelination lesions caused by lysophospatidyl choline injection. We also tested novel diffusion MRI method (double pulsed field gradient diffusion MRI), which turned out to not to be feasible for biomarker studies because extensively long acquisition time required for high -quality data.
In the first part of the work we examined the phase contrast in zero-echo time pulse sequence SWIFT. Phase contrast can be used sensitively and specifically detected hemorrhage, calsifications and myelin abnormalities. The majority of previous MRI phase imaging has been based on gradient recalled echo (GRE) sequences. EPITARGET studied phase contrast behavior due to small off-resonance frequency offsets in brain using SWIFT, a FID-based sequence with nearly zero acquisition delay. 1D simulations and a phantom study were conducted to describe the behavior of phase accumulation in SWIFT. Imaging experiments of known brain phase contrast properties were conducted in a perfused rat brain comparing GRE and SWIFT.
Additionally, a human brain sample was imaged. EPITARGET demonstrated how SWIFT phase is orientation-dependent and correlates well with GRE, linking SWIFT phase to similar off-resonance sources as GRE. The acquisition time was shown to be analogous to TE for phase accumulation time. Using experiments with and without a magnetization transfer preparation, the likely effect of myelin water pool contribution was seen as a phase increase for all acquisition times. Due to the phase accumulation during acquisition, SWIFT phase contrast can be sensitized to small frequency differences between white and gray matter using low acquisition bandwidths. This opens up new possibilities for more specific detection of white matter abnormalities potentially contributing to abnormal connectivity and networks during epileptogenesis.
In the second part, EPITARGET compared the sensitivity of SWIFT MRI imaging with that of magnetization transfer contrast and conventional diffusion tensor imaging (DTI) to detect mild to moderate damage in cortico-thalamic pathway after traumatic brain injury. MRI data from 6 TBI and 5 sham operated animals was collected and MRI findings were compared with histological stainings for thionin, myelin and iron. Magnetization transfer ratio (MTR) was decreased throughout cortical layers in somatosensory cortex indicating demyelination verified qualitatively by histology, while DTI changes were more variable. All analyzed white matter areas displayed decreased MTR with the largest change in the caudal internal capsule. Histology verified highest demyelination in the same area, although also substantial gliosis was detected. Also, in this area, DTI parameters were variable, although statistically significant changes were detected also with DTI. Our results show how MTR using SWIFT can be a valuable tool in assessing subtle changes both in cortex, and in white matter, after TBI in the corticothalamic pathway, which has been shown to play role in development of epilepsy.
In the third part, demyelination was assessed relaxation along fictitious field (RAFF) MRI contrast. This novel MRI contrast is able to probe slow molecular motion regime and is therefore proposed to be specific to myelin bound water fraction and or rotational correlation time of methyl groups in myelin. The contrast was tested in the intracranial LPC injection model, which causes demyelination without significand cell death and only mild gliosis. Demyelination was induced in two different areas in corpus callosum (cc) and dorsal tegmental tract (dtt). Corpus callosum represents an area with highly organized axonal bundles while dorsal tegmental tract was chosen to exemplify more complex axonal architecture. The data collected 3 days after LPC injection showed that both RAFF and DTI could detect deymelination (Figure 4, left panel) and were correlated with myelin content while, in DTT, only RAFF could detect demylination (Figure 4, right panel) and correlated with myelin content. This shows that RAFF can be used to detect demyelination regardless of underlying microstructure and provides a novel specific tool for epilepsy biomarker search.
Figure 4. Quantitative MRI parameters measured in LPC injection induced lesion in corpus callosum (left) and dorsal tegmental tract (DTT, right). (A) RAFF4 and magnetization transfer derived parameters (B) T1sat and (C) MTR showed lesion in both areas while DTI derived parameters (D) fractional anisotropy (FA), (E) mean diffusivity (MD), (F) axial diffusivity (AD), and (G) radial diffusivity failed to show demyelination in DTT.
We also implemented orientationally invariant acquisition scheme for double pulsed field gradient (dPFG) MRI. Analysis of rats at 5 months after traumatic brain injury the results indicated that when macroscopic level anisotropy is removed in orientationally invariant dPFG approach, the remaining contrast provides unique information of microscopic level anisotropy. In quantitative ROI analysis, apparent eccentricity derived from dPFG data showed significant changes in perilesional cortex at group level if compared to contralateral cortex. However, after careful analysis it became evident that dPFG images may have too low contrast to noise ratio and requires too long acquisition to be practical in search for biomarkers for epileptogenesis.
Development of novel PET radiotracers.
Within EPITARGET, starting with numerous candidates, novel nuclear imaging probes were developed and optimized by 07-UCL and 17-MHH, and made available for preclinical testing in EPITARGET (including transfer of radio-synthesis methods). This included 123I-labeled VCAM-1 antibody coated micron-sized particles of iron oxide (MPIOs, targeting neuro-inflammation), a peptidic iodine-labeled VCAM-1 tracer, a dually labeled human serum albumin tracer bearing radio-iodine and a fluorescent tag (targeting blood-brain barrier disturbance), [18F]FPEB (targeting glutamatergic neuro-receptors), [18F]deprenyl (MAO-B, astrocyte activation), and novel innovative synthesis methods (e.g. the use of sulfonium salts labelling precursors). These tracers could be partially evaluated during the EPITARGET funding period and are available for further preclinical and clinical testing.
Novel in vivo technologies for biomarker identification.
Our goal was to develop electrodes able to record electrical and molecular (GABA/glucose) signals and make them MRI compatible. Moreover, 10-AMU produced new in vivo technologies, which may have a large scientific and societal impact. In addition, EPITARGET being a highly dynamic consortium, naturally emerging new collaborations (initially not planned in the project) within the consortium, led to important new results.
GABA sensors.
While testing the GABA sensor devices, we realized that it is at present not possible to stabilize the enzymes close to the recording device. This raises complex chemistry and physics (at the nanoscale) questions well beyond the state of current knowledge. We decided to discontinue this part and rather focus on metabolic sensors.
We discovered that the modus operandi of such sensors is fraught with caveats. The glucose oxidase used to measure glucose concentration produces high levels of H2O2, which damage cells. The situation is even worse with commercial sensors, which questions the validity and interpretation of all papers making use of them. We had to go beyond the state-of-the-art, designing several technological solutions, and finally choosing one, which is presently tested in vivo. It took us several years to solve this technical difficulty. The solution found by EPITARGET is likely to revolutionize the technology of molecular sensors in the brain. The decision was made to shift towards adapting the organic technology for therapeutic purposes in epilepsy (instead of biomarkers).
Controlling brain activity with stimulation and the release of neuroactive compounds.
We demonstrated that organic transistors can be used to both stimulate neurons and record their activity (Williamson et al., 2015a, Adv Mat, IF 25.8). We demonstrated an organic electronic pump able to record activity and deliver drug at the same site, achieving exquisite control of brain activity, in particular stopping seizures (Williamson et al., 2015b, Adv Mat; Jonsson, 2016, PNAS). Press releases were issued, largely picked up by media (TV, radio, newspapers, blogs etc.). The pump is part of a traveling exhibition made by Inserm for the general public. These devices are being adapted to be tested in the operating room. Based on these results, and their further development, Adam Williamson obtained an ERC starting grant and a permanent research position at Inserm. We wrote a book chapter was written on these technological solutions (https://www.worldscientific.com/doi/abs/10.1142/9789813148611_0011).
MRI compatible electrodes.
This turned out to be a very challenging scientific question, as we wished to change the highly conducting gold lines of the electrodes, which generate artefacts in the MRI, into highly conducting carbon-based wires. Our chemist collaborators invented a new material satisfying our requirements. We made the electrodes and demonstrated their MRI compatibility and their performance to record electrophysiological signals. We could not perform the final test with implanted animals in the MRI due to the failure of 03-MVT to provide an interface. However, we intend to finish the project and publish the results.
1.3.2. Unravelling the complex pathophysiology of epileptogenesis for preclinical development of new disease-modifying treatments
Targets.
Novel molecules and cellular pathways involved in key etiopathogenic mechanisms of epileptogenesis have been identified and characterized in electrical and chemical status epilepticus (SE) rodent models. These new targets include the immunoproteasome, metalloproteinases, specialised pro-resolving mediators promoting resolution of inflammation, glycogen synthase kinase 3β, the chemokine CX3CR1, the adhesion molecule N-cadherin and oxidative stress-related molecules. The changes in the expression of some of these molecules during epileptogenesis in rodent brain tissues were validated in human specimens from patients with drug-resistant structural epilepsies or in patients who died after SE. Pharmacological targeting of such molecules lead to a >50% reduction in the frequency of spontaneous seizures, improvement in neuronal cell loss and rescue of cognitive deficits.
Treatments.
We identified five disease-modifying treatments based on the combination of novel mechanism-targeted drugs (i.e. neuroinflammation and oxidative stress). We found that the administration of combined anti-inflammatory or combined anti-oxidant drugs after SE or at disease onset significantly improved long-term disease outcomes, i.e. the frequency and/or duration of spontaneous seizures, neuronal cell loss and memory deficits. These therapeutic effects persisted after drug discontinuation therefore indicating disease-modification effects. Importantly, these findings have high translational value since some of these drugs are medically used for other neurological or inflammatory diseases.
Model refinement.
A two-stage in vivo approach was established for facilitating 1. preclinical screening of antiepileptogenic drugs and 2. finding new biomarkers.
1. For the first stage, which should allow “medium-throughput drug screening”, the intrahippocampal kainate model in mice was chosen. This model is characterised by a latent period of several days followed by a high frequency of spontaneous electrographic seizures, thus allowing short EEG recording periods. For stage 2, to test the most interesting drug combinations coming out of stage 1, the refined intrahippocampal kainate model in rats was chosen.
Stage 1 in mice were used to determine: (a) the tolerability of novel combination of clinically used drugs; (b) the antiepileptogenic efficacy of tolerable combinations of clinically used drugs. Eighteen rationally chosen drug from different mechanistic categories (i.e. anti-inflammatory, neuroprotective, and classical antiepileptic drugs) were tested in 12 combinations. Four effective drug combinations were identified, with levetiracetam+topiramate being the most effective combination.
2. For the preclinical identification of biomarkers of epileptogenesis, it is mandatory to use experimental models with a similar severity of primary injury but providing animals with and without spontaneous seizures. In this context, several established post-SE models of acquired epilepsy were refined. By using these refined models, novel phenotypic, blood, MRI and EEG biomarkers were characterized and validated in WP1.
Novel models of epileptogenesis.
They were developed to support preclinical studies by the scientific community: the Theiler’s viral encephalitis model in mice, an immature rat model to study the encephalopathic effects of epilepsy, and a model of BBB dysfunction induced by intracerebral albumin infusion in rodents.
1.3.3. Developing polymeric nanotechnology-based formulations for combinatorial and larger biologic compound delivery into the brain
The delivery of disease-modifying treatments to the brain was advanced using gene therapy based on HSV-based amplicon vectors and polymeric nanotechnology-based formulations. A portfolio of amplicon vectors was constructed to evaluate the most suitable technologies and promoters. Specific strategies for short-term or long-term gene therapy were explored, including multi-cistronic constructs for the inclusion of several therapeutic genes in amplicon vector, as well as replication-defective HSV vectors gene combinations aiming at modulation of the NPY signal with promising results.
Nanotechnology-based formulations of two potential therapeutics were formulated and successfully delivered via the nose-to-brain administration route using the MET platform developed by 19-NM. We were able to exemplify the feasibility of this approach using the large peptide anakinra (17.3 kD) which has anti-inflammatory and potentially anti-epileptic properties blocking the IL1 receptor. Pharmacokinetic studies (07-UCL, 19-NM) indicated that anakinra reached the brain and hippocampus but concentrations were probably insufficient to allow effective blocking of epileptogenesis specifically in the acute kainate model (05-IRFMN). The approach was also demonstrated for a candidate small molecule with positive pharmacokinetics data suggesting further evaluation in preclinical models (19-NM).
1.3.4. Structured human epilepsy tissue banking as basis for translational characterization of biomarkers and therapy targets for antiepileptogenesis
From patients undergoing epilepsy surgery for seizure control, 08-UBMC has built up an extensive biosample resource comprising sera, CSF as well as biopsy brain tissue. The bank is complemented by harmonized clinical data and we have extensively used these resources for analyses. In this context, 131 human temporal lobe epilepsy fresh frozen hippocampi (HS n=90, lesion associated temporal lobe epilepsy n=41) were used for experimental approaches scrutinizing a low voltage-dependent calcium channel as biomarker of hyperexcitability conversion in epileptogenesis.
08-UBMC together with 02-L&B had previously shown in an animal model, that transcriptional upregulation of the voltage dependent calcium channel CaV3.2 leads to an increased propensity for burst discharges of hippocampal neurons in epileptic hippocampi. Current data from animal models further suggested a role for the metal-responsive transcription factor 1 (MTF1) in promoter activation of CaV3.2. To translate our findings into the human disease context and investigate whether the MTF1-CaV3.2 pathway is also involved in human TLE, we analyzed patients with hippocampal sclerosis (HS; Figure 5) for their MTF1 and CaV3.2 expression levels. Fresh frozen human non-epileptic control tissue is not available for obvious reasons.
Therefore, we used ‘lesion-associated’ TLE hippocampi as epileptic controls. In such patients, lesions including low-grade tumors as well as dysplasias are typically in the close vicinity of the hippocampal formation, which is resected in the course of lesionectomy. The sequential pattern of epileptogenesis after a brain insult is not present in ‘lesion-associated’ TLE clinical histories. Intriguingly, patients with HS show a higher MTF1 and CaV3.2 expression compared to lesion-associated controls (Figure 5; HS: n = 79; Lesion associated: n = 35; t-test: ***P ≤ 0.001). In addition, we observed a strong positive correlation between MTF1 and CaV3.2 expression levels (Figure 5), indicating that also in human TLE, the MTF1 dependent CaV3.2 upregulation is active – even in patients, which are regarded pharmacoresistant considering nowadays available drug treatment regimens.
Improved therapies for TLE are urgently needed, since nowadays still 30% of respective patients are refractory to pharmacotherapy. Intriguingly, for human TLE patients with the most frequent pathology pattern, i.e. HS, we here observed evidence for increased expression of CaV3.2 and MTF1 and a strong correlation between their expression levels. Many HS patients, in contrast to lesion-associated TLE have a clinical history of insult-induced epileptogenesis.
Figure 5. MTF1 and CaV3.2 expression levels co-segregate and are increased in hippocampal tissue of patients with hippocampal sclerosis (HS). (a) Epileptogenesis in a human individual without any previous neurological symptoms. The patient initially manifested clinically with SE. Coronal T2-weighted fast spin echo (A-C) and axial Diffusion-weighted spin echo EPI images show the rapid development of a right-sided hippocampal sclerosis in the clinical course. Initially, there is hippocampal swelling (A: arrow) associated with cytotoxic edema of the CA1 sector (D: arrow). Two weeks later, swelling and cytotoxic edema are somewhat regredient but still present (B, E). Only 8 weeks later, cytotoxic edema has disappeared (F) and hippocampal atrophy, i.e. HS has manifested (C: arrow). (b) Quantitative determination of MTF1 and CaV3.2 mRNA. MTF1 and CaV3.2 are significantly more abundantly expressed in hippocampal tissue of TLE patients with HS versus hippocampi from patients with lesion-associated TLE, i.e. in which seizures are explained by lesions such as low-grade neoplasms and/or focal dysplasia in the immediate vicinity or even including the hippocampal formation (HS: n = 79; Lesion associated: n = 35; t-test: ***P ≤ 0.001 with synaptophysin as reference gene). (c) Regression analyses of CaV3.2 mRNA versus MTF1 mRNA expression. A strong positive correlation between the expression of this TF and the dependent Ca2+ channel mRNA is even present in the heterogeneous group of human HS hippocampi.
Furthermore, the finding that MTF1 and CaV3.2 expression levels are substantially increased in human TLE patients currently regarded as pharmacoresistant, from which hippocampi were obtained in the present study, may serve as a strong argument for the development of therapy options in humans focusing on this ‘MTF1-CaV3.2-neuronal bursting’ mechanism as potential new treatment option (van Loo et al., Nat Commun 2015). Thus, the characterization of the MTF1-CaV3.2 axis have provided new biomarkers for epileptogenesis by complementary analyses in TLE mouse models and human epileptic brain tissue and thereby represents successful translational analyses by partners 08-UBMC and 02-L&B in the framework of EPITARGET.
Impaired memory performance is a major comorbidity of chronic progressive epilepsy. Molecular biomarkers for this aspect have not been systematically characterized. In a joint study, partners 02-L&B together with 08-UBMC have used a combinatorial approach starting from large scale gene expression from biopsy hippocampi and corresponding genetic data from a large series (n=79) pharmacoresistant TLE patients (Figure 6; Bungenberg et al., Neurobiol Dis 2016).
These patients were stratified according to presurgical memory impairment using an established four-tired grading system ranging from ‘average’ to ‘very severely’. Multimodal cluster analyses revealed molecules specifically associated with synaptic function and abundantly expressed in TLE patients with substantially impaired memory performance. In a subsequent promoter analysis, the single nucleotide polymorphism rs744373 C-allele was found to be associated with high mRNA levels of bridging integrator 1 (BIN1)/Amphiphysin 2, i.e. a major component of the endocytotic machinery and located in a crucial genetic AD risk locus.
Figure 6. Functional gene clusters with low expression levels in figural memory in the low performance vs. average performance group (A, B). ‘Gene ontology (GO) enrichment analysis (GOEA) was performed using the Biological Networks Gene Ontology tool (BiNGO, 3.0.3) and Cytoscape (3.0). The enrichment map represents the GO terms as nodes with enrichment significance encoded by color intensity and number of genes involved by the node size of the respective GO term. GOEA analysis of genes with a negative diffscore <-2.5 (n=383), i.e. genes for which the low performance group has a lower expression signal compared to the average performance group was carried out. Using a high p-value (0.00005) the categories with the highest enrichment score (17.28) have a common functional background, being mainly related to the neurological system, such as the synapse, neuron projection and synaptic transmission (modified from Bungenberg et al., Neurobiol Dis 2016).
Using in vitro luciferase transfection assays, the BIN1 promoter activation was observed to be genotype dependent and strongly increased by reduced binding of the transcriptional repressor TGIF. These data indicate that poor memory performance in patients with TLE strongly corresponds to distinctly altered neuronal transcript signatures, which - as demonstrated for BIN1 - can correlate with a particular allelic promoter variant (Bungenberg et al., Neurobiol Dis 2016). These promoter variants characterized in a joint effort of 02-L&B together with 08-UBMC can be directly analysed presurgically and may be of future value as biomarkers for tailor made clinical management with respect to major comorbidities of chronic progressive epilepsies.
1.3.5. Systems-biological approach based characterization of Sestrin-3 as epileptogenesis biomarker
11-IC jointly with 08-UBMC have used these recently established infrastructural preconditions in order to provide an integrated systems-biological approach to characterize biomarkers of epileptogenesis after SE complementarily in animal models and human TLE brain tissue samples. Therefore, the partners initially carried out comprehensive gene-regulatory network analyses in human brain tissue samples of temporal lobe epilepsy (TLE). We employed systems genetics approaches to characterize the genetic regulation of pathophysiological pathways in TLE (Figure 7). Using surgically acquired hippocampi from 129 pharmacoresistant TLE patients included in the EPITARGET virtual database, we identified a gene-regulatory network genetically associated with TLE that contains a specialized, highly expressed transcriptional module encoding proconvulsive cytokines and Toll-like receptor signaling genes (Johnson et al., Nat Commun 2015).
Figure 7. Identification of the TLE-network and functionally specialized transcriptional modules in human epileptic hippocampus. (a) Gene co-expression network identified in the hippocampus of TLE patients (TLE-network). Nodes represent genes and edges represent significant partial correlations between their expression profiles (FDR<5%). Node colour indicates the best GWAS P-value of association with focal epilepsy for SNPs within 100 kb of each gene. Boxes mark two transcriptional modules within the network. (b) Kyoto Encyclopedia of Genes and Genome (KEGG) pathways significantly enriched in the TLE-network (FDR<5%). The fold enrichment for each KEGG pathway is reported on the side of each bar. (c) Module-1 and Module-2 details. The size of each node is proportional to its degree of interconnectivity within each module. Light blue indicates genes showing nominal association with susceptibility to focal epilepsy by GWAS. Numbers in parenthesis indicate multiple microarray probes representing the same gene. (d) KEGG pathways significantly enriched in Module-1 (top) and Module-2 (bottom; FDR<5%). (e) Module-1 is significantly highly expressed in the hippocampus of TLE patients. mRNA expression of Module-1 (n=80 probes, representing 69 unique annotated genes) as compared with Module-2 (n=60 probes, representing 54 unique annotated genes), other network genes (n=371 probes, representing 319 unique annotated genes) and all other probes represented on the microarray (n=48,256). **P=3.8x10-4; ***P<10-10, Mann–Whitney test, two-tailed (adapted from Johnson et al., Nat Commun 2015).
Complementary RNA sequencing analysis in a mouse model of TLE using 100 epileptic and 100 control hippocampi demonstrated the pro-convulsive module to be preserved across species, specific to the epileptic hippocampus and abundant in expression also in the chronic stage of recurrent epileptic seizures (Figure 8). In the respective set of TLE patients from the EPITARGET database, we mapped the trans-acting genetic control of this proconvulsive module to Sestrin 3 (SESN3), and demonstrated that SESN3 intriguingly positively regulates the module in macrophages, microglia and neurons (Johnson et al., Nat Comm 2015).
Figure 8. TLE-network conservation in mouse epileptic hippocampus. (a) Human TLE-network genes that are conserved and co-expressed (84%) in the mouse hippocampus. Each node in the network represents a transcript that had significant partial correlation with at least another transcript in the network (FDR<5%). Conserved Module-1 and Module-2 genes are indicated in blue and green, respectively. (b) Distribution of significant partial correlations (FDR<5%) between pairs of transcripts from 10,000 bootstrap permutation samples in epileptic (top) and control (bottom) mouse hippocampus. In each case, the red line indicates the actual number of significant partial correlations (FDR<5%) between all genes in the network. The number of significant partial correlations observed in control hippocampus was no different from chance expectation (P=0.659). In contrast, the number of significant partial correlations detected in epileptic hippocampus was significantly higher than expected by chance (P=0.001). (c) Differential expression of Module-1 genes between control and epileptic mouse hippocampus shows specific enrichment for TLR-signaling and cytokine genes among the upregulated genes. Stars denote significant fold changes between epileptic and control mouse hippocampus (FDR<5%); blue bars indicate TLR-signaling and cytokine genes (from Johnson et al., Nat Comm 2015).
Finally, Morpholino-mediated Sesn3 knockdown in a recent zebrafish epilepsy model confirmed the regulation of the transcriptional module, and intriguingly attenuated chemically induced behavioral seizures in vivo. These results underline that the recent launching of the ‘virtual human brain tissue database’ represents a powerful repository for the integrated and harmonized experimental use and optimal exploitation of the rarely available epileptogenesis human brain tissue and corresponding biofluid specimens in the framework of EPITARGET and has already substantially contributed to externally highly visible research activities. The outlined results were only possible by the active mutual interactions of numerous partners of the consortium (01-ULUND, 02-L&B, 04-UEF, 07-UCL, 08-UBMC, 09-NENCKI, 11-IC, 12-UNIFER, 14-BGU) and EPITARGET provides a unique structure for extrapolating these results into even more refined biomarker molecules for epilepsy also after the termination of EPITARGET. The data represent an integrated approach to define a biomarker mechanisms and potential therapy target based on integrated analyses in human brain tissue and corresponding animal models and further demonstrates the experimental power of the EPITARGET epilepsy brain tissue biobank.
1.3.6. Structured human status epilepticus (SE) tissue banking and analyses of immune-related mechanisms as biomarkers of early epileptogenesis.
10-AMC has achieved the structured biobanking of post mortem human brain tissue samples of patients after status epilepticus (SE) and traumatic brain injury (TBI), of surgical brain tissue and biofluids (CSF and blood) of patients after TBI, and of fresh frozen hippocampal biopsy specimen of pharmacoresistant epilepsy patients as well as of blood from a subset of patients. These biospecimens have been extensively used for analyses of the use of molecules involved in early epileptogenesis as biomarkers.
The proteasome is a multisubunit enzyme complex critical for protein degradation. 05-IRFMN revealed a dysregulation of the (immuno)proteasome pathway in a rat model of chronic epilepsy after SE (Mishto et al., Brain Behav Immun 2015). During inflammation, the constitutive subunits are replaced by their inducible counterparts, which results in the formation of the immunoproteasome. In post-SE rats, the expression of the β1i and β5 subunits was higher in neurons within the hippocampus and piriform cortex during the acute phase (one day post-SE) compared to controls. At this time point, β5 expression was also higher in astrocytes within the hippocampus and piriform cortex. During the latent (one week post-SE) and chronic phase (three months post-SE), the expression of β1, β1i, β5 and β5i was higher in neurons and in most cases also in astrocytes within the hippocampus and piriform cortex compared to controls. Interestingly, in rats that developed a progressive form of epilepsy the expression of (immuno)proteasome subunits was higher than in rats with a non-progressive form.
10-AMC with 05-IRFMN have translated these data from rodent models into the human disease context by addressing the expression pattern of constitutive (β1, β5) and immunoproteasome (β1i, β5i) subunits by immunohistochemistry in surgical specimens from patients with malformations of cortical development and in patients with temporal lobe epilepsy (TLE) and hippocampal sclerosis, as well as in the hippocampus and cortex of patients that died after status epilepticus (SE) or traumatic brain injury (TBI) and during epileptogenesis in the post-SE epilepticus rat model (van Scheppingen et al., J Neuroinf 2016). Thereby, increased expression was observed in both, focal cortical dysplasia, FCD IIa and b; cortical tubers from patients with Tuberous Sclerosis Complex (TSC). Intriguingly, β1, β1i, β5 and β5i were detected (within cytosol and nucleus) in dysmorphic neurons, balloon/giant cells and reactive astrocytes. In human TLE, the expression of β1, β1i, β5 and β5i was higher in hippocampal neurons and astrocytes compared to controls and was most pronounced in patients with hippocampal sclerosis. In patients that died after SE, the expression of subunits β5 and β1i was higher in neurons and astrocytes, respectively, compared to controls. Increased expression was particularly detected in the cortex of TBI patients.
Subsequently, in vivo experiments showed that (immuno)proteasome protein expression and the number of spontaneous seizures was lower in rats that were treated with immunomodulator drug rapamycin (inhibitor of the mammalian target of rapamycin, mTOR pathway) compared to vehicle-treated rats. In vitro studies using cultured human astrocytes showed that gene expression of all subunits increased after IL-1β stimulation, which could be attenuated by the immunomodulators rapamycin or curcumin. Dysregulation of the (immuno)proteasome system during epileptogenesis occurs in epileptic rat model and human brain and may serve as biomarker as well as therapy target in the future.
Distinct miRNAs including miR-9a-3p have been suggested as putative biomarkers of epileptogenesis after SE in rats (12-UNIFER & 11-IC; Roncon et al., Sci Rep 2015). Conversely, post mortem human brain tissue has been used in several studies addressing the cellular expression of microRNAs deregulated during epileptogenesis in experimental models (13-AMC ; van Scheppingen et al., Glia, 2016). Increasing evidence supports the involvement of microRNAs in the regulation of astrocyte-mediated inflammatory response. To study the role of inflammation-related microRNAs in epilepsy associated pathologies, real-time PCR and in situ hybridization were employed to characterize the expression of miR21, miR146a and miR155. In addition, cultured human astrocytes were used to study the regulation of the expression of these miRNAs in response to the proinflammatory cytokine IL-1β and to evaluate the effects of overexpression or knockdown of miR21, miR146a and miR155 on inflammatory signaling. IL-1β stimulation of cultured glial cells strongly induced intracellular miR21, miR146a and miR155 expression, as well as miR146a extracellular release. IL-1β signaling was differentially modulated by overexpression of miR155 or miR146a, which resulted in pro- or anti-inflammatory effects respectively. These data provide supportive evidence that inflammation-related microRNAs play in focal epilepsies. In particular, miR146a and miR155 appear to be key players in the regulation of astrocyte-mediated inflammatory response, with miR146a as most interesting anti-inflammatory therapeutic candidate. This specific miRNA has been further evaluated in collaboration with 05-IRFMN in WP2.
Resolution of inflammation is a highly coordinated and active process that is controlled by endogenous pro-resolving lipid or peptide mediators and it is instrumental to switch off acute inflammation. If this mechanism fails, inflammation might perpetuate resulting in varying degree of tissue injury or dysfunction. 05-IRFMN has investigated whether the brain immune response triggered by epileptogenic injuries is inefficiently controlled by pro-resolving endogenous molecules thus resulting in chronic inflammation using rodent models of epileptogenesis. Increased gene expression of pro-resolving G-protein coupled receptors (GPCRs; ALXR and ChemR23) was observed in resected hippocampus of patients with TLE and hippocampal sclerosis, as well as in the hippocampus of patients that died within 5 days or more than 7 days after SE, i.e. reflecting the stage of putative early epileptogenesis in humans. In control, ALXR and ChemR23 were expressed predominantly in neurons, while after SE and in epileptic tissue we observed an increase in the expression of the two receptors in glial cells (Figure 9). These data agree with the increase in glial expression of the two receptors observed in mice after SE and in the chronic epileptic phase.
Figure 9. Expression of ALXR and Chem23 in control human brain tissue as well as at early stages in post mortem tissue after SE and in chronic mesial (m)TLE.
14-BGU demonstrated that recurrent seizures and the associated excessive glutamate release lead to increased vascular permeability in the rat cerebral cortex, through activation of NMDA receptors (Vazana et al., J Neurosci 2016). After blood-brain-barrier (BBB) disruption, albumin extravasation induces astrocytic activation and potentially seizures (Kim et al., J Neurosci Res. 2016). Astrocytes immediately respond to brain damage and play a critical role in the development of postinjury epilepsy. Astrocytes were shown to regulate extracellular matrix (ECM) remodeling, which can affect plasticity and stability of synapses and, in turn, contribute to the epileptogenic process.
Complementarily, post mortem brain tissue of patients that died after TBI has been used to support the occurrence of the BBB dysfunction and brain inflammation in perilesional cortex and thalamus in lateral fluid percussion injury rat model (04-UEF; manuscript in preparation). Evaluation of both cortex and thalamus of patients that died a few months after TBI confirmed the association between BBB dysfunction and activation of astrocytes, microglia, as well as deposition of iron and calcifications (Figure 10, Figure 11).
Figure 10. Blood-brain barrier dysfunction in perilesional human cortex and thalamus 6 months after TBI. A, B: iron deposits (blue); C, D: calcifications; E, F: albumin.
Figure 11. Brain inflammation in perilesional human cortex and thalamus 6 months after TBI. A, B: CD68 positive macrophages/microglia/monocytes; C ,D: vimentin positive astrocytes; E, F: HLA-DP, DQ,DR positive microglia.
TBI triggers a cascade of molecular and cellular changes leading to development of comorbidities such as progressive neuronal loss and hippocampal memory decline. 04-UEF analyzed microRNAs (miRNAs) related to chronic alterations in hippocampal molecular networks using Affymetrix microarray dataset obtained from rat dentate gyrus at 3-months after lateral fluid-percussion injury. Microarray data was used for Ingenuity Pathway Analysis (IPA), which revealed 18 up-regulated targets for miR-124, suggesting a down-regulation of miR-124 (z-score=-4.176 p<0.01). The deregulation of this specific miRNA has been confirmed in different brain injury models. Evaluation by in situ hybridization performed in rat and TBI human tissue confirmed the predominant neuronal expression of this miRNA within both hippocampus and cortex. Analysis of data is in progress and extended to the post-SE human material. Intriguingly, 04-UEF has in a parallel approach demonstrated that a TBI transcript signature (TBI-sig) consisting of 4964 regulated genes in the perilesional cortex and 1966 in the thalamus (FDR < 0.05) of the TBI rat model insinuates susceptibility to distinct drug treatment (Lipponen et al., Sci Rep 2016). 04-UEF observed by LINCS analysis 11 compounds that showed a strong connectivity with the TBI-sig in neuronal cell lines. Of these, celecoxib and sirolimus have been demonstrated to have a disease-modifying effect in in vivo animal models of epilepsy. These results reveal several molecules with pre-clinical evidence as biomarkers for epileptogenesis in TBI human brain tissue and corresponding animal models and provide arguments in favor of certain therapy regimens.
The present results in the framework of EPITARGET clearly demonstrate the applicability of human brain tissue samples of patients that died after SE as well as TBI. The biobank relies on highly selected and very rare human brain tumour samples and represents a unique scientific asset based on the multilateral activities within the network structure of EPITARGET.
1.3.7. Characterization of 'early epileptogenesis' biomarkers identified in TBI animal models in human biopsy brain tissue of patients with epilepsy.
14-BGU together with 08-UBMC have focused on alterations of TGFß signaling related to expression of inflammatory genes and degradation of perineuronal nets around inhibitory interneurons particularly at early stages after distinct epileptogenic insults to the brain (Kim et al., Sci Rep 2017). Brain damage following stroke or brain trauma frequently results in epileptogenesis. Such transient insults commonly involve lasting dysfunction of the blood-brain barrier (BBB). Previous data from 14-BGU had shown that chemically-induced disruption of the BBB induces pathological hyperexcitability manifested by epileptiform activity and recurring seizures. Exposure of the brain environment to generally serum-derived albumin is sufficient to induce epileptiform activity and recurring seizures, and this pathological cascade depends on activation of TGFβ signalling.
14-BGU observed albumin extravasation in brain regions with hyperexcitable network activity in the injury models used in the respective study – peri-infarct hippocampus following photothrombotic cortical stroke and partially isolated undercut cortex (Kim et al., Sci Rep 2017). Comparative transcriptome analyses in respective models predicted remodelling of extracellular matrix (ECM) as a common response to different types of injuries. ECM-related transcriptional changes were induced by the serum protein albumin via TGFβ signalling in primary astrocytes. 14-BGU studied, whether the predicted ECM alteration may be associated or even causatively involved in chronic deficits of inhibitory signalling and pathological hyperexcitability observed following injury. Particularly, a specialized ECM structure named perineuronal nets (PNNs) is tightly associated with parvalbumin (PV) expressing fast-spiking inhibitory interneurons. Co-staining of wisteria floribunda agglutinin (WFA), a broad marker labelling PNNs, with immunostaining for parvalbumin (PV), a marker for fast-spiking interneurons revealed significant main effects revealed that perineuronal nets around PV(+) interneurons were degraded following traumatic brain injury induced by the undercut model in the present study only seven days after induction of the trauma.
14-BGU together with 08-UBMC next studied whether albumin-induced TGFβ signalling activation and subsequent PNN degradation occur in the hippocampus of patients with focal epilepsies, i.e. temporal lobe epilepsy (TLE) who frequently report transient brain insults to have occurred months to years before the onset of chronic recurrent seizures. 14-BGU quantified phosphorylation of Smad2 (pSmad2) protein as a marker of TGFβ receptor-mediated signalling in hippocampal tissues resected from TLE patients and autopsy controls from 08-UBMC. The number of GFAP (+) astrocytes co-localized with pSmad2 protein was significantly increased in hippocampal sections of TLE patients compared to controls (Figure 12). Consistent with the findings in the rodent brains the percentage of PV(+)/WFA(+) interneurons was significantly decreased in the hippocampus resected from TLE patients in compared with controls (Figure 12). These results corroborate the potential role of TGFβ-regulated ECM remodelling in the pathology of chronic hyperexcitability.
Figure 12. Increased expression of astrocytic pSmad2 and reduced association of PNNs around PV(+) cells in human epileptic hippocampi. (a) Representative micrographs of human hippocampi stained for GFAP and phosphorylated Smad2, a downstream effecter of TGFβ signalling. The hippocampi were resected from temporal lobe epilepsy (TLE) patients or age-matched autopsy controls. Representative GFAP(+) astrocytes expressing pSmad2(+) are indicated by white arrowheads. Scale bar=50μm. (b) Co-localization of pSmad2 with GFAP was increased in TLE patients (n=5) compared to controls (n=3). (c) Representative confocal micrographs of human hippocampal tissues stained for GFAP, PV, and PNNs (WFA) resected from TLE patients or controls. Scale bar=50μm. White and blue arrowheads indicate representative PV(+)/WFA(+) and PV(+)/WFA(−) cells, respectively. (d) The percentage of PV(+)/WFA(+) cells was decreased in the hippocampus of TLE patients (n=4) compared to controls (n=4). Mann-Whitney test, two-tailed, was used for statistical analyses. *p<0.05. (e) A working model. Several points for feed-forward loops (i-iii) can lead to chronic hyperexcitability and the development of epilepsy. The dysfunction of blood-brain barrier (BBB) and the ensuing entry of albumin into brain parenchyma activate TGFβ signaling. Comparative transcriptome analyses predicted the activation of the common core signalling transduction including MAPK pathway, Stat3, NFκB, AP-1, and ETS1. These signalling pathways can elicit the reciprocal activation of inflammation and extracellular matrix (ECM) remodelling (i). ECM remodelling can trigger the transformation of a latent from of TGFβ to its active form75 as well as exacerbate BBB dysfunction58 (ii). The degradation of perineuronal nets (PNNs) around fast-spiking interneurons and aberrant excitatory synaptogenesis occur in the course of ECM remodelling, presumably leading to functional alterations in inhibition and abnormality in synaptic plasticity that may contribute to excitation/inhibition (E/I) imbalance and ultimately the occurrence of seizures. Finally, seizures per se cause BBB dysfunction, inflammation, and the upregulation of MMPs activity (from Kim et al., Sci Rep 2017).
These results integrating data from TBI animal models and correspondingly human epileptic hippocampi provide new pathogenetic insights in how albumin extravasation that occurs upon BBB dysfunction in various brain injuries and ways to predispose neural circuitry to the development of chronic inhibition deficits and suggest new tissue biomarkers and a novel rationale for therapy targets in TBI-induced epileptogenesis.
1.3.8. Use of biosample collections comprising blood, CSF and brain tissue from patients with chronic progressive epilepsy for translational characterization of biomarkers and therapy targets
Individual and combinatorial biomarkers have been characterized and released with strong impact for assessment of the risk and potential gain for certain parameters. This includes the establishment of a systems genetics approach starting from gene expression data from hippocampi of the pilocarpine-induced status epilepticus model for identifying cell membrane receptors with a direction-specified influence over disease-related mRNA profiles (11-IC, 02-L&B, 05-IRFMN, 08-UBMC, 13-AMC). 11-IC has designed a very complex novel systems neurobiological approach in order to characterize biomarkers for chronic progressive epilepsy and derive new therapy concepts (Figure 13).
Figure 13. The study plan was based on 100 mice with epilepsy (pilocarpine post status epilepticus model of temporal lobe epilepsy) and 100 control (pilocarpine-naïve) matched littermate mice. At 4 weeks post status epilepticus, each mouse was continuously monitored using 3D accelerometery and video monitoring for 14 days to record seizure frequency and severity. High-throughput mRNA sequencing (RNA-seq) was generated using RNA from snap-frozen whole hippocampus samples from the mice and gene expression profiles were used to generate co-expression modules. Coexpression modules with a potential relationship to epilepsy were prioritized using the following criteria: (i) differential co-expression between epileptic and healthy hippocampus (mouse and human TLE), (ii) correlation of module expression with seizure frequency (mouse), and (iii) conservation in the human epileptic hippocampus. Modules meeting these criteria were considered candidate modules for epilepsy, and subjected to ‘Causal reasoning analytical framework for target discovery’ (CRAFT) analysis to identify membrane receptors predicted to restore disease module expression toward health (modified from Srivastava et al., Nature Commun 2018).
Analysis of the biological terms and canonical pathways enriched among the 28 co-expression modules in the mouse epileptic hippocampus revealed that the modules were generally enriched for specific functions. Among the modules with overlapping functions, modules 5, 16, and 18 were enriched for “immune response” processes (Benjamini–Hochberg (BH)-corrected P = 2.1 × 10−11, P = 1.4 × 10−6, and P = 1.3 × 10−33, respectively), and modules 10, 14, 26, and 29 were enriched for neuronal functions including “synaptic transmission” (BH P = 4.4 × 10−11, P = 0.02 P = 4.0 × 10−3, and P = 4.0 × 10−4, respectively). To provide insights into the cell-type expression of the modules, cell-type marker genes derived from singlecell RNA-seq analysis of the mouse hippocampus were used. In this context, “immune response” modules 16 and 18 were enriched for microglia marker genes, whilst “synaptic transmission” modules 10, 14, 26, and 29 were specific for neuronal cell types. To further prioritize the modules in terms of their relationship to epilepsy, the correlation between each module’s expression and seizure frequency was explored (Figure 14).
Figure 14. Correlation of module expression with seizures. (a) For each co-expression module from the epileptic mouse hippocampus, we plotted the significance (−Log10 FDR) of the Spearman’s correlation between the module’s eigengene and seizure frequency (bar plot), and the percentage of variance in seizure frequency explained by the module’s eigengene (R2, dotted line). Modules marked with a blue arrow are the modules differentially co-expressed between the epileptic mouse hippocampus and the control mouse hippocampus. Modules highlighted in grey (bar plot) are significantly (FDR <0.05) correlated with seizure frequency. (b) Volcano plot of average (Spearman’s) correlation of a module’s genes with seizure frequency (X-axis) versus the significance of the module’s enrichment for genes individually correlated with seizure frequency (QTT genes) (Y-axis) for the nine modules differentially co-expressed in epilepsy and correlated with seizures by module eigenegene (modified from Srivastava et al., Nature Commun 2018).
For the nine modules differentially co-expressed in epilepsy and correlated with seizures (i.e. modules 2, 5, 8, 10, 16, 18, 21, 22, and 24), it was then assessed whether the module was conserved in the human epileptic hippocampus. Using human orthologs of mouse module genes and genome-wide gene expression data from 122 human epileptic hippocampus samples surgically ascertained from TLE patients, we found that all nine modules were conserved in the human epileptic hippocampus (FDR<0.05) (Figure 15).
Figure 15. Correlation of module expression with seizures. Grey – not significant at FDR 5%. * = modules differentially co-expressed in mouse and human TLE + correlated with seizures + conserved in human TLE (modified from Srivastava et al., Nature Commun 2018).
The conservation of these nine modules across human and mouse TLE provides an independent line of evidence for the validity of these modules, and further supports the relevance of the pilocarpine post SE mouse model of TLE to human TLE. As a final assessment of the relationship of these nine mouse TLE modules to human epilepsy, we tested whether each module was also differentially co-expressed in human TLE. In this analysis, for each module, we compared intra-module correlations in the human epileptic hippocampus with that in the non-diseased human hippocampus using post-mortem hippocampal samples ascertained from people with no history of psychiatric or neurological disease. Among the nine mouse modules differentially co-expressed in epilepsy and correlated with seizures, seven (5, 10, 16, 18, 21, 22, and 24) were also differentially co-expressed in human TLE (Figure 15). These seven modules were selected for further analysis. Specifically, we hypothesized that focusing on these seven modules (and by extension their enriched functional pathways) would provide a starting point for the development of new therapy approaches for epilepsy. Microglia-related combinatorial mRNA modules (module 18) correlate strongly with seizure activity. The further bioinformatic characterization of potential therapeutic targets predicted the tyrosine kinase receptor Csf1R in complementary mouse model and human epileptic hippocampi (Srivastava et al., Nature Commun 2018). The predicted effect of Csf1R blockade in attenuating seizures was then validated in pre-clinical epilepsy models. The released corresponding mRNA signature of module 18 has a strong potential to assess seizure risks and points to innovative treatment options.
Potential Impact:
SEE ATTACHED REPORT
1.4.1. Potential impact
1.4.1.1. Identifying novel biomarkers and their combinations in animal models for further preclinical studies towards more precise diagnoses and treatments
Novel biomarkers.
EPITARGET has provided evidence of plasma HMGB1 with or without combination to TSPO PET as a biomarker for epileptogenesis after SE. EPITARGET has also identified several miRNAs that will be further investigated as prognostic and predictive biomarkers for epileptogenesis. In addition, EPITARGET MRI studies have revealed imaging biomarkers for experimental post-traumatic epilepsy and progression of brain damage. These data will impact diagnosis and stratification of patients in clinical practice, and pave the way towards more affordable preclinical and clinical treatment trial to combat acquired epileptogenesis in humans.
05-IRFMN, 04-UEF and 17-MHH groups also generated promising data concerning the use of [18F]deprenyl as a novel prognostic biomarker. [18F]Deprenyl autoradiography could differentiate between HMGB1-high animals (developing epilepsy) and HMGB1-low animals (not developing epilepsy).
Social isolation and epilepsy.
The Debski et al study made us realize that social isolation may constitute an important confounding factor in epilepsy research. We performed a dedicated study, which questions the interpretation of the results obtained when animals are singly housed (tens of thousands), as commonly done in epilepsy (Manouze et al., 2019, eNeuro). This study was highlighted in Science, discussed in several commentaries and editorial in the October issue of Epilepsia; and generated a record high number of tweets in the scientific community. The impact of this study on future experimental and preclinical work based on epilepsy models is going to be significant as animal housing conditions, particularly social isolation, can be a major confounding factor for interpretation and comparisons of molecular, behavioural, and imaging data obtained in different laboratories in trials aiming at treatment and biomarker discovery.
Information processing in the “epileptic” brain.
The Frigerio et al. paper led us to ask whether information processing may be drastically modified in experimental epilepsy. As a first step, we characterize information processing in the “normal” brain. We have unravelled the basic properties of the neuronal language during sleep (Clawson et al., 2019, Science Adv). We have now finished, the second part, demonstrating how the neuronal language is modified in epilepsy (Clawson et al., in preparation). This study will have impact on how information processing should be considered in further studies on epilepsy models and taken in account as a significant comorbidity to be accounted for in treatment paradigms.
Epilepsy Preclinical Biomarker Bank (EPBB).
EPITARGET RedCap database (see section 1.3.1.1) was the first systematic preclinical database in epilepsy for multicentre use. It will allow different study sites to link the animal specific information with the data collected (e.g. miRNA-seq data). It will remain as a part of 01-ULUND’s informatics infrastructure. If new funding opportunities arise, investments will be made to maintain it as long as possible. This database will have impact on epilepsy research community in terms of defining CDEs for collaborative and multicentre studies, as well as providing access to already established database for future datamining and networking. The experience gained paves the way in preclinical laboratories towards large multicentre biomarker and therapy studies, which is expected to increase the reproducibility and translational value of preclinical data to clinic, and thus, speed up the clinical biomarker and therapy development.
EPITARGET RedCap and EPBB have already significantly helped the efforts of harmonization of preclinical epilepsy studies. For example, it has served as a template for ILAE Working Group of Preclinical Harmonization, which includes many EPITARGET partners. It was also used to help in design of the NIH Centers-without-Walls funded EpiBioS4Rx preclinical RedCap database at LONI in USC.
1.4.1.2. Unravelling the complex pathophysiology of epileptogenesis for preclinical development of new disease modifying treatments
This project enabled a better understanding of the mechanisms leading to epilepsy development in structural forms of epilepsy with an acquired aetiology. This novel information also set the basis for investigating the same mechanisms in other forms of epilepsy therefore searching for common mechanisms across the epilepsies. By highlighting novel targets for antiepileptogenesis interventions we developed combinatorial drugs treatments for taking into consideration the multiple contributing mechanisms that in concert are responsible for epileptogenesis and to attain more significant disease modifying effects vs single treatments. We identified novel disease-modifying combinatorial treatments using anti-inflammatory and anti-oxidant drugs which are in medical use, and also novel investigational drugs. These medically used drugs with a good tolerability and safety profile could be tested by repurposing these drugs in clinical trials in drug-resistant epilepsies or in prospective studies for the prevention of epilepsy in patients who are at risk of developing epilepsy. In view of the enhanced pathological excitability underlying also non-epileptic disorders, such as migraine, neuropathic pain and bipolar disorders, the novel antiepileptogenic treatments may have additional therapeutic indications other than epilepsy. Therefore this research line opened novel opportunities and created impact on better epilepsy treatment, and particularly, hypothesis/mechanism driven combinatorial approach in this regard.
1.4.1.3. Developing polymeric nanotechnology-based formulations for combinatorial and larger biologic compound delivery into the brain
The delivery of biologics and small molecule drugs to the brain remains a significant barrier to the successful pharmacotherapy not only of epilepsy but also for many other diseases such as e.g. neurodegenerative conditions. The viral vectors generated and characterized in the frame of EPITARGET are promising tools for the gene therapy of epilepsy as well as other neurological disorders. Work under EPITARGET with the MET delivery system towards using the nose-to-brain route has added important evidence of the feasibility of this route. Specifically, this has increased our understanding of the range of molecules this might be applicable to but also restrictions with respect to achievable dose/ concentration.
1.4.1.4. Translation of biomarkers and therapy targets into the human disease context
With respect to the translation of biomarkers and therapy targets into the human disease context (WP04), the potential impact of EPITARGET is particularly significant due to the fact that highly selected human epilepsy brain tissue sample based data is integrated with results from a wide range of animal model studies. This holds true for animal model data derived from blood samples, peripheral organs, brain tissue, electrophysiological data, behavioural data, and imaging data from different models of epileptogenesis. Furthermore, brain and blood levels of antinflammatory mediators and the cellular expression of their cognate receptors in SE and TBI models have been made available for translational analyses in human epilepsy brain tissue and corresponding biofluids. As we provided examples for under ‘Main Results’, successful translation of biomarkers of distinct excitability mechanisms emergent during epileptogenesis as well as inflammatory dynamics as surrogate markers of altered seizure threshold and biomarkers in serum/CSF was accomplished. Furthermore, several biomarker candidates reflect mechanisms that can serve as therapy targets. These data can have significant impact by further translation from the preclinical human disease in the clinical context. The human brain tissue repositories generated in the framework of EPITARGET will be further used in the future and even extended by ongoing introduction of human epilepsy brain tissue samples. Given the integration of these data with exact clinical information, this unique clinic-scientific asset will outreach EPITARGET and provide one of the worldwide leading tissue repositories on human epilepsy for translational study approaches with high impact.
1.4.1.5. New collaborations and jobs
EPITARGET created the opportunity for its partners to engage in new research collaborations that may have not been possible otherwise. For instance, 05-IRFMN and 04-UEF’s teams together looked at the consequence of inflammation on signal integration in the dendrites (Frigerio et al., 2018, Mol Neurobiol). 09-NENCKI, 06-TiHo, 08-UBMC and 04-UEF demonstrated the circadian cycling of the molecular architecture of the hippocampus and its alteration in epilepsy (Debski et al., Science Adv, in revision – on BioRxiv). A spin off paper is in preparation on the circadian regulation of metabolism. The partners involved in the identification of miRNA biomarker (Task 3.1) namely 04-UEF, 09-NENCKI, 11-IC, 12-UNIFER and 10-AMC, generated a common paper (Srivastava et al., eNeuro 2017), designed CDE for the multi-center study (van Vliet et al., Epilepsia 2017) and are currently continuing their collaboration to finalize the data jointly collected. Lastly, EPITARGET generated several position papers to which many partners contributed (e.g. Lapinlampi et al., 2017, Epi Res; Klein et al., 2018, Epilepsia).
These collaborations have a significant impact on European networking and future collaborations on European arena. The human biospecimen repository partially created on the basis of EPITARGET on a comprehensive European scale, the experimental workflows in translating biomarker candidates from experimental animal research into the human disease context that have been established through the integrated efforts of EPITARGET consortium members and particularly the knowledge generated on these individual aspects as well as the downstream steps for e.g. patenting of key results have been able to harness individual activities in national Epilepsy Centres for leveraging their quality onto a European collaborative stage with quality levels to compete worldwide as documented by the seminal output of EPITARGET. Several research initiatives on a European level have been initiated, in which one or more EPITARGET partners have prominent roles as PIs or coordinators (e.g. UBMC, AMC).
Another important outcome is the new collaborative network Epi-CLUSTER which has received financial support from EBRA (see section 3.3.2 in D6.03). This collaborative network is mainly based on EPITARGET and other epilepsy consortia and their beneficiaries, and is designed to reach consensus between various epilepsy constellations how to work together, how to coordinate the efforts together with other initiatives (e.g. Epilepsy Forum 2020), agreeing on main objectives and speaking with a single voice at a European level in alignment with EBRA’s objectives and activities. Other collaborative applications for grants are being developed building on the epilepsy consortia meeting epiXchange held in Brussels in May 2018, which also was based on networking between EPITARGET and other EU funded epilepsy projects. Such efforts and collaborations would not have been possible without EPITARGET, as well as other epilepsy consortia.
In addition, the work carried out in EPITARGET had a significant impact on the career of several young researchers involved in the project, by creating new jobs and giving opportunities for career development. Thus, the paper published on the Debski et.al study (see section 1.3.1.1.) played an instrumental role in the permanent research position obtained by Pascale Quilichini in the 10-AMC group. That same group published a general review on the pilocarpine model of epilepsy (Lévesque et al., 2016, J Neurosc Met, 100 citations), and two book chapters on models and a general theory of vulnerability to depression, in particular in epilepsy (Bernard, 2016, Cold Spring Harbor; Ivanov & Bernard, 2017, Models of Seizures and Epilepsy). This part has been extremely successful with several high-profile publications (2xAdv Mat, Science Adv, PNAS – two others in revision), largely covered by the media, with translational potential (under way). The results allowed two scientists from the group to obtain very highly competitive permanent positions at Inserm (8 per year for all neuroscience) and an ERC starting grant; ensuring continuity for the projects.
1.4.2. Dissemination activities
1.4.2.1. Publications
Over the course of the project, the partners published a total of 177 peer-reviewed articles in 72 different journals (see section 2. Use and Dissemination of Foreground). These include several publications in high impact journals, such as The Lancet Neurology (Impact factor 2018: 28.755) Advanced Materials (IF 2018: 25.809) and Acta Neuropathologica (IF 2018: 18.174) (Table 1).
Most articles had less than 15 citations, however 19 papers had between 50 and 100 citations and 7 papers had more than 100 citations (including 2 papers with more than 200 citations; see Table 1).
Table 1. EPITARGET Publications figures
A major effort in dissemination was the generation of a seminal review paper on the topic of the consortium by EPITARGET PIs. The review, entitled ‘Advances in the Development of Biomarkers for Epilepsy’ was published in the Lancet Neurology, one of the premiere journals in Neurology (Pitkänen et al., Lancet Neurology, 2016). This review paper has generated great interest in the field and was largely cited . Due to the rapid developments in the fields, EPITARGET partners are working towards generating two additional authoritative reviews with a similar impact: a first review paper will address novel developments in the biomarker field (led by A. Pitkänen of UEF); and the second will address combinatorial treatments (led by A. Vezzani of IRFMN, and W. Löscher of TiHo).
The EPITARGET consortium also contributed (coordinated by A. Pitkänen, UEF), within the epiXchange cluster (see section 1.4.2.2. Events) to an article published in Epilepsia Journal, presenting major achievements by the cluster in the areas of biomarkers, genetics, therapeutics, comorbidities and biobanks as well as recommendations for future research to develop and bring novel solutions to the patients (Advancing research toward faster diagnosis, better treatment, and end of stigma in epilepsy , Pitkänen et al., 2019).
1.4.2.2. Events
In addition to publications, the partners engaged in more than 350 dissemination activities see section 2. Use and Dissemination of Foreground). These included over 200 oral presentations and 100 posters communicating project results at major scientific events such as the European Congresses of Epileptology (ECE), the International Epilepsy Congresses (IEC) and Workshops on Neurobiology of Epilepsy (WONOEP) , or the Annual Meetings of the American Epilepsy Society .
The consortium also held four Young Researchers’ Symposia back-to-back with the project’s General Assembly meetings, during which PhD students and Post-doc researchers were able to present their work in the project. To accompany these symposia, abstract booklets were compiled by ARTTIC including abstracts of all oral presentations and posters. Awards in the form of a conference participation fee of a maximum amount of 400€ were distributed to the best oral presentation and best poster. The Executive Board members and SAB members constituted the jury.
EPITARGET was also counted among the seven currently ongoing and former collaborative projects funded by the EC (Desire, EpimiRNA, EPIPGX, Epistop, EPITARGET, EpiCare and Epixchange) which co-organised the epiXchange conference to disseminate project results and discuss the future of epilepsy research in the EU and worldwide. The conference took place on 23 May 2018 in Brussels.
The first day of the conference was open for the public and highlighted the achievements from the participation projects so far, their impact and main challenges and next steps. It included keynote talks from renowned scientists in the field and also provided a platform for scientists to present their work during a poster session. The event gathered 170 participants (28 from EPITARGET) from 18 European and 5 non-European countries, in addition to 250 watching the live-stream on Facebook . Officials from the EC, the European Brain Council, EMA, patient organisations (e.g. Epilepsy Foundation and Federazione Italiana Epilessie) and pharma (strong presence of UCB pharma) also attended the conference were 28 talks and 38 posters divided into 5 main themes were presented. epiXchange was promoted by 800 tweets and retweets from October 2017 to August 2018. Presentations and videos from the conference can be found on www.epixchange2018.eu.
In connection to conference, a workshop was held on 24 May at the DG Research and Innovation to discuss policy and strategy on future epilepsy research with a global approach. The main focus was to identify gaps and needs, impact enhancement of results, like pathways (including regulatory) to clinical guidelines, product development and policy recommendations. This policy day aimed to define and analyse bottlenecks and give suggestions for additional efforts need to be made to overcome those. This discussion day addressed regulatory pathways for treatments or diagnosis and highlight what kind of infrastructure is needed for the future.
1.4.2.3. Website and social media
Publications, events and project news were regularly communicated about on the project’s website (https://www.epitarget.eu/) and Facebook page (Figure 20) as well as in the project newsletters.
Figure 20. Screenshot of EPITARGET's Facebook page
The EPITARGET logo has been developed in period 1 by partner 15-GABO:mi. All dissemination material was based on the colour scheme established in the logo to generate a consistent corporate project identity.
The EPITARGET website went online in the 1st period and was updated continuously since then with regular news items as well as update on the project content.
Since the launch on 19.05.2014 the EPITARGET website had 20,911 visits with an average session duration of 00:01:29 and on average 2.03 pages opened per session. These numbers include all page visits.
The website is accessed from all over the world, also from outside of Europe (Figure 21): the top three countries of origin of visitors are the US, Germany and France. The website has totalized 41,902 page views.
The website was maintained by 15-GABO:mi until 30 June 2016 and by 20-ART since 1 July 2016.
Figure 21. Geographical distribution of all website sessions longer than 2 seconds between 19 May 2014 and 31 October 2019.
The latest news on the EPITARGET project were circulated with a biannual newsletter which was implemented as part of the EPITARGET website . The newsletter included updates on EPITARGET meetings and awards as well as information on the latest EPITARGET publications and posters.
1.4.3. Exploitation of results
EPITARGET produced Deliverable 6.3 “Final plan on dissemination, exploitation and transfer of final results” which has been actively prepared in the last two years of the project under the lead of 02 L&B and with contributions from all partners.
In the beginning of the project (Month 6), the Consortium agreed on and implemented a plan for dissemination (see D6.02) carrying out numerous communication and dissemination activities to make the project results known. Firstly, the partners published a total of 177 peer-reviewed articles in 72 different journals (see section 1.4.2.1.). In addition, the partners engaged in more than 350 dissemination activities (see section 1.4.2.2.) including oral presentations (>200) and posters (>100) at major scientific conferences. EPITARGET also organized four Young Researchers’ Symposia, as well as co-hosting the epiXchange conference (May 2018, Brussels). Communication of publications, events and project news was achieved via the project website, Facebook and in a regular newsletter, following the plan for dissemination.
With regards to exploitation, EPITARGET has been successful in generating a large range of exploitable results in different categories. Firstly, several new drug combinations with antiepileptogenic/disease-modifying potential have been tested during the project with some promising result. An example for such a promising combination is the combination of an NMDA receptor antagonist with an AMPA receptor antagonist, or the administration of combinations of levetiracetam with topiramate. Secondly, EPITARGET has established a number of unique biosample banks and databases that will continue to drive research and translation, including tissue banks and databases about experiment types performed in the consortium. Third, EPITARGET has successfully developed preclinical data on a number of promising prognostic biomarkers, or novel technologies to enable biomarker measurements. Examples for this achievements are the use of the biomarker HMGB1 for predicting the development of posttraumatic epilepsy in a rat model of neurotrauma. Fourth, EPITARGET has also been active in developing technologies aimed at measuring electrophysiological activity biomarkers. For example, EPITARGET PIs have developed novel in-vivo techniques, primarily organic electrode systems to detect activity biomarkers of epileptogenesis with unprecedented accuracy.
Most of these results require further research to reach all the way to the “end-user” (patients or physicians). Therefore, the EPITARGET consortium has discussed and devised different ways of collaborating and obtaining funding for future research, including collaborations under EBRA, under ERANET and with NIH.
List of Websites:
https://www.epitarget.eu/
EPITARGET designed a unique set of first in its kind preclinical epilepsy-related common data elements (CDEs) and case report forms (CRFs) for different animal models and experimental procedures (www.epitarget.eu). These data were utilised through electronic RedCap database creatred at 01-ULUND server, where information on animals and performed procedures are stored and shared, including granting access of whole research community in the field. This includes data on plasma protein biomarkers, microRNAs, imaging MRI and PET biomarkers. Moreover, EPITARGET developed organic transistors that can be used to both stimulate neurons and record their activity. These devices are now being adapted to be tested in the operating theatres as MRI compatible electrodes to aid combinatorial biomarker identification. Moreover, these electrodes were shown to even stop seizures.
EPITARGET identified new targets for treatment, include the immunoproteasome, metalloproteinases, specialised pro-resolving mediators promoting resolution of inflammation, glycogen synthase kinase 3β, the chemokine CX3CR1, the adhesion molecule N-cadherin and oxidative stress-related molecules. These were also validated in human specimens from patients with drug-resistant epilepsies. Pharmacological targeting of such molecules lead to a >50% reduction in the frequency of spontaneous seizures, improvement in neuronal cell loss and rescue of cognitive deficits in rodents. EPITARGET also identified five disease-modifying treatments based on targeting neuroinflammation and oxidative stress. Combined anti-inflammatory or anti-oxidant drugs at onset of epileptogenesis significantly improved long-term disease outcomes, i.e. the frequency and/or duration of spontaneous seizures, neuronal cell loss and memory deficits. These therapeutic effects persisted long after drug treatment indicating disease-modification effects. These findings have high translational value since some of these drugs are medically used for other neurological or inflammatory diseases.
For more effective drug discovery and increased translational impact, EPITARGET designed “medium-throughput drug screening” model with intrahippocampal kainate in mice, followed by the refined intrahippocampal kainate model in rats. For the same purpose, EPITARGET developed the Theiler’s viral encephalitis epilepsy model in mice, and a model of BBB dysfunction induced by intracerebral albumin infusion in rodents. These model are a significant contribution to the community to facilitate trslational therapeutic value of preclinical studies in epilepsy.
EPITARGET designed and developed an extensive portfolio of Herpes Simplex Virus (HSV)-based vectors that are suitable for combinatorial gene therapy of epilepsy. This portfolio includes amplicon- and replication defective-vectors, all safe and capable of short-term or long-term transgene expression, i.e. prospectively applicable for use during epileptogenesis, to prevent development or progression of epilepsy, or in drug-resistant epilepsy, to control seizures. Selected combinations of therapeutic genes have been already engineered in these vectors and are ready for testing in epilepsy models.
EPITARGET has built up an extensive biosample resource comprising sera, CSF as well as biopsy brain tissue from patients undergoing epilepsy surgery for seizure control. The biobank is complemented by harmonized clinical data, altogether allowing for the extensively analyses, e.g. based on bioinformatics and systems biology. This biobank was used to identify substantially increased expression levels for MTF1 and CaV3.2 in human pharmacoresistant TLE patients. This finding may facilitate the development of new therapies focusing on this ‘MTF1-CaV3.2-neuronal bursting’ mechanism.
EPITARGET partners carried out comprehensive gene-regulatory network analyses in human brain tissue samples of temporal lobe epilepsy (TLE) patients, employing systems genetics approaches to characterize the genetic regulation of pathophysiological pathways in TLE. Using surgically acquired hippocampi from pharmacoresistant TLE patients included in the EPITARGET database, gene-regulatory network was identified containing a specialized, highly expressed transcriptional module encoding proconvulsive cytokines and Toll-like receptor signalling genes. Complementary RNA sequencing analysis in a mouse model of TLE demonstrated the pro-convulsive module specific to the epileptic hippocampus. In the respective set of TLE patients from the EPITARGET database, revealed that sestrin-3 (SESN3) positively regulates the module in macrophages, microglia and neurons, and SENS3 knockdown in a zebrafish epilepsy model attenuated chemically induced behavioral seizures in vivo.
EPITARGET has demonstrated that dysregulation of the (immuno)proteasome system during epileptogenesis occurs in epileptic rat model and human brain and may serve as biomarker as well as therapy target in the future. The consortium also demonstrated that TBI animal models and correspondingly human epileptic hippocampi provide new pathogenetic insights in how albumin extravasation that occurs upon BBB dysfunction in various brain injuries and ways to predispose neural circuitry to the development of chronic inhibition deficits and suggest new tissue biomarkers and a novel rationale for therapy targets in TBI-induced epileptogenesis.
EPITARGET established a systems genetics approach in animal models for identifying cell membrane receptors with a direction-specified influence over disease-related mRNA profiles. The conservation of several mRNA modules across human and mouse TLE models was established and provides an independent line of evidence for the validity of these modules to assess seizure risks and points to innovative treatment options. In particular, microglia-related combinatorial mRNA module was shown to correlate strongly with seizure activity. The analysis and bioinformatic characterization of potential therapeutic targets predicted the tyrosine kinase receptor Csf1R in complementary mouse model and human epileptic hippocampi. The released corresponding mRNA signature of module 18 has a strong potential in this regard.
Project Context and Objectives:
EPITARGET is a large European consortium of universities, research institutions and SMEs dedicated to understanding complex mechanisms and identify biomarkers of epileptogenesis to improve treatment and diagnosis of epilepsy. The overall scientific objectives of EPITARGET are:
Using multidisciplinary strategies of basic, preclinical and clinical research to:
1. Identify novel biomarkers and their combinations that will define the different stages of epileptogenesis and predict/diagnose early and late stages of the evolution of the disease. This will pave the way to improved diagnostics and better patient stratification, as well as development of patient-specific preventive strategies. In this objective EPITARGET will identify at least 2 biomarkers that in combination will predict whether epileptogenesis is triggered in the post-insult period and will stratify those individual subjects that are at risk of developing epilepsy.
2. Unravel the complex patho-physiology of epileptogenesis and design new, disease-modifying combinatorial treatment strategies specifically targeted to the different stages of epileptogenesis. These novel treatment strategies are expected to be capable of preventing the development of epilepsy in at-risk patients and stop its progression after the onset of the disease. Different epileptogenesis-associated processes, such as brain damage and structural/functional reorganization, neurogenesis, brain-blood barrier (BBB) dysfunction, inflammation, free radical formation, genetic and epigenetic alterations will be individually addressed and specific treatments directly or indirectly targeting these mechanisms will be combined. The choice of the best combinatorial therapeutic strategy will be assisted by systems biology approaches. This strategy will help to identify key aetiological factors, thus limiting redundancy in the choice of multiple drug targets and facilitating a rational drug discovery process. In this objective EPITARGET will identify at least 2 patho-physiological mechanisms of epileptogenesis that will be targeted in combination in animals at risk of developing epilepsy (post-insult period).
3. Translate the knowledge obtained in experimental models to patients in order to improve diagnosis, achieve better patient stratification, and develop new antiepileptogenic treatments and means to predict their efficacy. The objective is to verify and validate biomarkers in blood and brain tissue samples obtained from patients after potentially epileptogenic brain insults. These tissues include traumatic brain injury (TBI) and post-mortem brain tissue from patients that had status epilepticus (SE) or TBI (early-stage epileptogenesis), as well as from those with chronic epilepsy. These specimens are the best closest match to the experimental brain tissue of post-insult epileptogenesis in animal models. In this objective EPITARGET will perform clinical validation of the combinatorial biomarker approach (objective 1), and thereby obtain data from patients to prove the concept for future clinical applications.
To achieve these objectives and maximise outcome, EPITARGET tackles the complexity of epileptogenesis by adopting concerted and complementary actions of participating partners, attacking various aspects of epileptogenesis at different levels. EPITARGET combines a powerful arsenal of both established and innovative multidisciplinary research strategies, tools and platforms. Analysis of data is assisted by creating an animal and human database, and a bioinformatics approach within the consortium.
Project Results:
SEE REPORT ATTACHED
1.3.1. Identifying novel biomarkers and their combinations in animal models
Globally, an estimated 2.4 million people are diagnosed with epilepsy each year. Thus, a new person is diagnosed with epilepsy every 13 seconds (http://www.who.int/mediacentre/factsheets/fs999/en). In 60% of those affected, epileptogenesis is initiated by structural causes such as traumatic brain injury (TBI) or stroke. Over 40 hypothesis-driven monotherapy approaches have demonstrated some disease-modifying effects in animal models of epileptogenesis. Currently, however, no clinical treatments are available to stop or alleviate epileptogenesis in at-risk patients or to alleviate the course of epilepsy after its diagnosis.
One reason for the stalled progression of compounds showing proof-of-concept evidence in animal models to clinical antiepileptogenesis trials is the lack of diagnostic biomarkers for epileptogenesis that could be used to stratify patient populations for antiepileptogenesis trials and reduce study costs. Also, little effort has been aimed at identifying and using response biomarkers that could inform about target engagement in the early treatment development phase and provide an early go/no-go signal for further development.
One main objective of EPITARGET was to identify single or combinatorial biomarkers for epileptogenesis, that is: (a) prognostic biomarkers for the development of epilepsy after epileptogenic brain insults, (b) diagnostic biomarkers of ongoing epileptogenesis in one or several epileptogenic etiologies. The modalities investigated were plasma protein and microRNA biomarkers and brain MRI and PET imaging biomarkers. All these modalities are relatively noninvasive and widely available. In addition, WP1 developed novel tools for biomarker analysis.
EPITARGET partners 01-ULUND, 04-UEF, 05-IRFMN, 06-TiHo, 07-UCL, 09-NENCKI, 10-AMU, 11-IC, 12-UNIFER, 13-AMC, 17-MHH have worked as a team, generating multiple animal models and state-of-art novel tools for biomarker analysis, and complicated data analyses. There have been significant collaborations with WP1, 2, 3 and 5 in implementation and dissemination of data obtained in WP1.
1.3.1.1. Generation of an Epilepsy Preclinical Biomarker Bank (EPBB)
In spirit to perform a statistically powered preclinical analysis of the sensitivity and specificity of plasma and imaging biomarkers for epileptogenesis, EPITARGET decided to harmonize the methodologies and data analysis. This was a pioneering effort considering the preclinical research in Europe.
EPITARGET (a) generated the first preclinical epilepsy-related common data elements (CDEs) and case report forms (CRFs) for different animal models and experimental procedures (www.epitarget.eu) (b) designed electronic RedCap database on 01-ULUND server to store information on animals and procedures performed. Research data is stored at each study site, (c) provided access to RedCap for Epitarget users, (d) trained Epitarget users to enter and retrieve data in RedCap
1.3.1.2. Plasma protein biomarkers
Our initial goal was to multiplex 11 plasma proteins: SBDP150 and HMGB1 (acute necrosis), SBDP120 (delayed apoptosis), S100β (brain injury), GFAP (astrogliosis), UCH-L1 and NSE (neuronal cell body injury), MAP2 (dendritic injury), MBP (myelin injury), collagen VI (extracellular matrix), and EMAP-II/III-6 (microglia), in order to perform analysis on multiple markers in small plasma volumes available from rat tail vein. However, development of the multiplex protein analysis platform by Bioviron did not work out, and EPITARGET partners decided to continue with a hypothesis-driven approach, assessing one plasma biomarker candidate at a time by using ELISA or Simoa. After a wide range of technical difficulties (no antibody/assay available; assay described in literature but not showing protein expression at EPITARGET sample time points), we reviewed new information in the literature and took advantage of the mechanistic work of EPITARGET partners in order to identify novel plasma protein biomarkers and optimize their analysis. These included Collagen VI, HMGB1, clusterin, Sushi-repeat protein X-linked 2, neurofilament light. Of the candidates tested in rodent/human plasma samples and rat/human brain tissue, HMGB1 is at the most advanced stage, showing promise to predict epileptogenesis after status epilepticus in animal models.
1.3.1.3. Plasma microRNAs
In the frame of EPITARGET, we designed and conducted an unbiased, multi-centre, longitudinal study on circulating miRNAs during epileptogenesis, using four different refined models of epilepsy (lithium-pilocarpine, perforant path stimulation, amygdala stimulation and lateral fluid-percussion injury), in which only a subset of animals develops spontaneous seizures. Altogether 157 eligible samples were sent to GenomeScan (Leiden, the Netherlands) for small RNA-seq using an Illumina platform. No common miRNA alterations were observed across all 4 models. Only a few were common to 2 models. Therefore, it was decided to run a meta-analysis of the data. As there was a low variation between the effect sizes, a fixed effect model (FEM) was considered appropriate for analysis. FEM meta-analysis identified 5 miRNAs that were significantly different between animals that would subsequently become epileptic and those that would not become epileptic at 2 d post insult. Three of these were already known (namely miR-129-5p, miR-138-5p and miR-3085) and two were not (ENSRNOG00000054305 and ENSRNOG00000054389). Interestingly, miR-129-5p inhibition has been associated with synaptic downscaling in vitro and reduced epileptic seizure severity in vivo, and miR-129-5p is upregulated in different epileptic models. In addition, miR-138-5p is a potential regulator of memory performance, a candidate regulator of P53 (involved in cell cycle, proliferation and apoptosis), and it is down-regulated in several phases of epilepsy.
Based on these data, we performed a technical verification on the 3 known miRNAs using digital droplet PCR (ddPCR). Residual aliquots of the samples employed for RNA-seq were RNA retro-transcribed and ddPCR analysed. Due to delays caused by unforeseen difficulties in animal modelling and miRNA assays, technical validation and analyses in independent cohorts of animals will be being done after the EPITARGET funding period using the additional funds the EPITARGET partners have been able to get. If these further experiments will provide positive results, some of these miRNAs may become eligible for testing in human epilepsy.
1.3.1.4. MRI biomarkers
Multiparametric MRI data was collected in several timepoints along the evolution of epileptogenesis. The analysis of the data turned out to be more challenging as anticipated and therefore a part of resource originally allocated to Task 5.1 (Development of novel MRI tools) were directed to advanced analysis of rich longitudinal MRI data obtained in biomarker study in Task 2, instead of doing a separate longitudinal biomarker study for new techniques.
At the first stage, analysis was focused to T2 values in the perilesional cortex. Because of variable lesion size lesions and progression conventional co-registration-based image analysis pipelines were non-applicable. For visualization purposes, unfolded cortical maps of T2 relaxation were computed. For these maps, the T2 relaxation coefficient was measured along the curvature of the cortex and over the imaging slices, while being averaged over the cortical thickness. Consequently, T2 relaxation was observed on the flattened cortex, allowing convenient visualization.
The differently classified pathological volumes computed from T2 MRI measurements 2 d and 7 d after TBI were used to fit a multiple linear regression model to explain necrotic lesion volume 21 d after TBI The result implies that an early (< 1 week after TBI) T2 MRI measurement is highly indicative of later necrotic tissue volume. The result implies that cortical lesion volume is not a sufficient metric to explain posttraumatic epileptogenesis.
In the second part, analysis was targeted to the thalamus. T2 relaxation time on days 2, 7, and 21 and DTI parameters on days 7 and 21 after TBI were utilized to model epileptogenesis using various logistic regression models. The different parameter maps at different time points were registered to a common template for a region-of-interest (ROI) type analysis. The approaches included using a single MRI parameter map at a single time point to model epileptogenesis and using multiple parameter maps at multiple time points to model epileptogenesis.
Based on the findings, analysis of multiple MRI parameters in the thalamus may be combined to form a model to distinguish between individuals with and without epileptogenesis (Figure 1). However, the approach used had relatively low sensitivity and requires further development or must be combined with other biomarkers to achieve clinical relevance. The manuscript of the results is under preparation.
Figure 1. Modelling epileptogenesis by using the mean and standard deviation of T2 parameter values on days 2, 7, and 21 as well as of diffusion parameters CL and CP on days 7 and 21 in a logistic regression model produced a statistically significant regression model. (A) The model produced a c-statistic (area under a receiver operating characteristic curve, AUC) of 0.89 with a 95% confidence interval (CI) of 0.81-0.94. To assess the possibility of overfitting, leave-one-out (LOO) and K-fold cross-validation (CV) methods were used. (B) LOO CV produced a c-statistic of 0.75 (95% CI 0.65-0.84). (C) K-fold CV produced a c-statistic of 0.70 (95% CI 0.58-0.79).
1.3.1.5. PET biomarkers
Identification of prognostic and predictive biomarkers for epileptogenesis and treatment response.
EPITARGET investigated various positron emission tomography (PET), single photon emission computed tomography (SPECT) and magnetic resonance imaging (MRI) protocols throughout experimental epileptogenesis (e.g. TSPO ligands for [11C]PK11195 PET and [18F]GE180 PET (microglia activation), [18F]FDG PET (glucose metabolism), [18F]FET (amino acid turnover), [68Ga]DTPA PET, [99mTc]DTPA SPECT, [18F]FMZ (GABAA receptor expression), gadolinium-DTPA-enhanced T1 MRI (blood-brain barrier (BBB) integrity) and T2 MRI (e.g. cell edema)). We could acquire distinct time profiles of various epileptogenesis-associated changes (exemplarily illustrated in Figure 2 for the pilocarpine rat model).
Figure 2. Imaging-based analysis of time profiles of epileptogenesis-associated changes. To facilitate a better overview, data were normalized to baseline.
Furthermore, the imaging protocols were investigated in different post status epilepticus animal models of epileptogenesis including the systemic pilocarpine rat post-SE model as well as the focal intra-hippocampal kainate post-SE mouse model. As neuro-inflammation and BBB integrity are promising candidates for biomarkers, but also for treatments, the time profiles were successfully adapted to treatment protocols in order to quantify acute treatment effects of anti-epileptogenic drug candidates, including dexamethasone, minocycline, fingolimod (FTY720) or curcumin.
17-MHH also performed a multi-tracer study investigating the suitability of various imaging and behavioural markers regarding their predictive biomarker properties. The applied study design was adapted to the 04-UEF design for TBI-induced epileptogenesis. A modified pilocarpine rat model, developed by 06-TiHo within EPITARGET recently, in which only part of the animals develop chronic epilepsy, was chosen for this study. Imaging was performed at 17-MHH, whereas behavioural testing and video-EEG monitoring was carried out at 06-TiHo. Both glucose hypometabolism and neuroinflammation (Figure 3) were able to detect differences between seizing and non-seizing animals using voxel-based analysis (statistical parametric mapping (SPM)). In addition, HMGB1 blood values at day 8 post SE (Figure 3A) differentiated well between seizing and non-seizing animals using the ROC analysis (Figure 3C) AUC = 0.86 p = 0.02) resulting in a cut-off value of 4 ng/ml. This result was even improved when combining HMGB1 blood levels with TSPO PET (AUC = 0.91 p = 0.02). This approach provided a promising candidate protocol for translation to the clinic.
Figure 3. Analysis of high mobility group box 1 protein (HMGB1) blood levels and TSPO PET (neuroinflammation). (A) Blood levels were analysed by ELISA at baseline, 1 day and 8 days post SE (upper graph). Data is presented as mean ± SD, asterisks indicate significant differences (p < 0.05 one-way ANOVA and Tukey’s multiple comparison post hoc test for comparison between testing days and unpaired Student‘s t-test for comparison BL with following measurements in one group). (B) Imaging results are presented as statistical parametric maps. Voxel-wise comparisons (Student’s t test) of TSPO binding on days 12 to 14 post SE (volume of distribution Vt, middle box) are shown and only clusters with significantly different voxels are visualized (p < 0.05 minimum cluster size of 100 voxels). Hot and cold scale represents the p value for each voxel. (C) Receiver operator characteristic (ROC, lower graph) curves comparing seizing and non-seizing rats for thalamic TSPO PET (18F-GE180 Vt), HMGB1 levels at day 8 post SE, or combination of both.
A validated platform for in vitro autoradiography based on existing ([18F]GE180, [18F]FMZ ) and novel PET radiotracers was developed by EPITARGET ([18F]FPEB (glutamatergic neuro-receptor MGluR5), [18F]deprenyl (MAO-B, astrocyte activation)) was established and applied to brain slices of rat and mouse models of epileptogenesis provided, but also to the P21 rat model. Here, [18F]deprenyl could differentiate between HMGB1-high animals (developing epilepsy) and HMGB1-low animals (not developing epilepsy).
1.3.1.6. Novel MRI and PET tools and novel in vivo technologies for biomarker identification
Novel MRI tools.
In this subtask EPITARGET focused on developing novel MRI methods for detecting structural abnormalities after potentially epileptogenic brain injuries. Especially, the focus was in white matter abnormalities. We were able to develop two novel approaches, based on phase imaging or magnetization transfer with zero-echo time detection (SWIFT), and relaxation in the rotating frame. The techniques were tested for detection of white matter damage in traumatic brain injury and/or in demyelination lesions caused by lysophospatidyl choline injection. We also tested novel diffusion MRI method (double pulsed field gradient diffusion MRI), which turned out to not to be feasible for biomarker studies because extensively long acquisition time required for high -quality data.
In the first part of the work we examined the phase contrast in zero-echo time pulse sequence SWIFT. Phase contrast can be used sensitively and specifically detected hemorrhage, calsifications and myelin abnormalities. The majority of previous MRI phase imaging has been based on gradient recalled echo (GRE) sequences. EPITARGET studied phase contrast behavior due to small off-resonance frequency offsets in brain using SWIFT, a FID-based sequence with nearly zero acquisition delay. 1D simulations and a phantom study were conducted to describe the behavior of phase accumulation in SWIFT. Imaging experiments of known brain phase contrast properties were conducted in a perfused rat brain comparing GRE and SWIFT.
Additionally, a human brain sample was imaged. EPITARGET demonstrated how SWIFT phase is orientation-dependent and correlates well with GRE, linking SWIFT phase to similar off-resonance sources as GRE. The acquisition time was shown to be analogous to TE for phase accumulation time. Using experiments with and without a magnetization transfer preparation, the likely effect of myelin water pool contribution was seen as a phase increase for all acquisition times. Due to the phase accumulation during acquisition, SWIFT phase contrast can be sensitized to small frequency differences between white and gray matter using low acquisition bandwidths. This opens up new possibilities for more specific detection of white matter abnormalities potentially contributing to abnormal connectivity and networks during epileptogenesis.
In the second part, EPITARGET compared the sensitivity of SWIFT MRI imaging with that of magnetization transfer contrast and conventional diffusion tensor imaging (DTI) to detect mild to moderate damage in cortico-thalamic pathway after traumatic brain injury. MRI data from 6 TBI and 5 sham operated animals was collected and MRI findings were compared with histological stainings for thionin, myelin and iron. Magnetization transfer ratio (MTR) was decreased throughout cortical layers in somatosensory cortex indicating demyelination verified qualitatively by histology, while DTI changes were more variable. All analyzed white matter areas displayed decreased MTR with the largest change in the caudal internal capsule. Histology verified highest demyelination in the same area, although also substantial gliosis was detected. Also, in this area, DTI parameters were variable, although statistically significant changes were detected also with DTI. Our results show how MTR using SWIFT can be a valuable tool in assessing subtle changes both in cortex, and in white matter, after TBI in the corticothalamic pathway, which has been shown to play role in development of epilepsy.
In the third part, demyelination was assessed relaxation along fictitious field (RAFF) MRI contrast. This novel MRI contrast is able to probe slow molecular motion regime and is therefore proposed to be specific to myelin bound water fraction and or rotational correlation time of methyl groups in myelin. The contrast was tested in the intracranial LPC injection model, which causes demyelination without significand cell death and only mild gliosis. Demyelination was induced in two different areas in corpus callosum (cc) and dorsal tegmental tract (dtt). Corpus callosum represents an area with highly organized axonal bundles while dorsal tegmental tract was chosen to exemplify more complex axonal architecture. The data collected 3 days after LPC injection showed that both RAFF and DTI could detect deymelination (Figure 4, left panel) and were correlated with myelin content while, in DTT, only RAFF could detect demylination (Figure 4, right panel) and correlated with myelin content. This shows that RAFF can be used to detect demyelination regardless of underlying microstructure and provides a novel specific tool for epilepsy biomarker search.
Figure 4. Quantitative MRI parameters measured in LPC injection induced lesion in corpus callosum (left) and dorsal tegmental tract (DTT, right). (A) RAFF4 and magnetization transfer derived parameters (B) T1sat and (C) MTR showed lesion in both areas while DTI derived parameters (D) fractional anisotropy (FA), (E) mean diffusivity (MD), (F) axial diffusivity (AD), and (G) radial diffusivity failed to show demyelination in DTT.
We also implemented orientationally invariant acquisition scheme for double pulsed field gradient (dPFG) MRI. Analysis of rats at 5 months after traumatic brain injury the results indicated that when macroscopic level anisotropy is removed in orientationally invariant dPFG approach, the remaining contrast provides unique information of microscopic level anisotropy. In quantitative ROI analysis, apparent eccentricity derived from dPFG data showed significant changes in perilesional cortex at group level if compared to contralateral cortex. However, after careful analysis it became evident that dPFG images may have too low contrast to noise ratio and requires too long acquisition to be practical in search for biomarkers for epileptogenesis.
Development of novel PET radiotracers.
Within EPITARGET, starting with numerous candidates, novel nuclear imaging probes were developed and optimized by 07-UCL and 17-MHH, and made available for preclinical testing in EPITARGET (including transfer of radio-synthesis methods). This included 123I-labeled VCAM-1 antibody coated micron-sized particles of iron oxide (MPIOs, targeting neuro-inflammation), a peptidic iodine-labeled VCAM-1 tracer, a dually labeled human serum albumin tracer bearing radio-iodine and a fluorescent tag (targeting blood-brain barrier disturbance), [18F]FPEB (targeting glutamatergic neuro-receptors), [18F]deprenyl (MAO-B, astrocyte activation), and novel innovative synthesis methods (e.g. the use of sulfonium salts labelling precursors). These tracers could be partially evaluated during the EPITARGET funding period and are available for further preclinical and clinical testing.
Novel in vivo technologies for biomarker identification.
Our goal was to develop electrodes able to record electrical and molecular (GABA/glucose) signals and make them MRI compatible. Moreover, 10-AMU produced new in vivo technologies, which may have a large scientific and societal impact. In addition, EPITARGET being a highly dynamic consortium, naturally emerging new collaborations (initially not planned in the project) within the consortium, led to important new results.
GABA sensors.
While testing the GABA sensor devices, we realized that it is at present not possible to stabilize the enzymes close to the recording device. This raises complex chemistry and physics (at the nanoscale) questions well beyond the state of current knowledge. We decided to discontinue this part and rather focus on metabolic sensors.
We discovered that the modus operandi of such sensors is fraught with caveats. The glucose oxidase used to measure glucose concentration produces high levels of H2O2, which damage cells. The situation is even worse with commercial sensors, which questions the validity and interpretation of all papers making use of them. We had to go beyond the state-of-the-art, designing several technological solutions, and finally choosing one, which is presently tested in vivo. It took us several years to solve this technical difficulty. The solution found by EPITARGET is likely to revolutionize the technology of molecular sensors in the brain. The decision was made to shift towards adapting the organic technology for therapeutic purposes in epilepsy (instead of biomarkers).
Controlling brain activity with stimulation and the release of neuroactive compounds.
We demonstrated that organic transistors can be used to both stimulate neurons and record their activity (Williamson et al., 2015a, Adv Mat, IF 25.8). We demonstrated an organic electronic pump able to record activity and deliver drug at the same site, achieving exquisite control of brain activity, in particular stopping seizures (Williamson et al., 2015b, Adv Mat; Jonsson, 2016, PNAS). Press releases were issued, largely picked up by media (TV, radio, newspapers, blogs etc.). The pump is part of a traveling exhibition made by Inserm for the general public. These devices are being adapted to be tested in the operating room. Based on these results, and their further development, Adam Williamson obtained an ERC starting grant and a permanent research position at Inserm. We wrote a book chapter was written on these technological solutions (https://www.worldscientific.com/doi/abs/10.1142/9789813148611_0011).
MRI compatible electrodes.
This turned out to be a very challenging scientific question, as we wished to change the highly conducting gold lines of the electrodes, which generate artefacts in the MRI, into highly conducting carbon-based wires. Our chemist collaborators invented a new material satisfying our requirements. We made the electrodes and demonstrated their MRI compatibility and their performance to record electrophysiological signals. We could not perform the final test with implanted animals in the MRI due to the failure of 03-MVT to provide an interface. However, we intend to finish the project and publish the results.
1.3.2. Unravelling the complex pathophysiology of epileptogenesis for preclinical development of new disease-modifying treatments
Targets.
Novel molecules and cellular pathways involved in key etiopathogenic mechanisms of epileptogenesis have been identified and characterized in electrical and chemical status epilepticus (SE) rodent models. These new targets include the immunoproteasome, metalloproteinases, specialised pro-resolving mediators promoting resolution of inflammation, glycogen synthase kinase 3β, the chemokine CX3CR1, the adhesion molecule N-cadherin and oxidative stress-related molecules. The changes in the expression of some of these molecules during epileptogenesis in rodent brain tissues were validated in human specimens from patients with drug-resistant structural epilepsies or in patients who died after SE. Pharmacological targeting of such molecules lead to a >50% reduction in the frequency of spontaneous seizures, improvement in neuronal cell loss and rescue of cognitive deficits.
Treatments.
We identified five disease-modifying treatments based on the combination of novel mechanism-targeted drugs (i.e. neuroinflammation and oxidative stress). We found that the administration of combined anti-inflammatory or combined anti-oxidant drugs after SE or at disease onset significantly improved long-term disease outcomes, i.e. the frequency and/or duration of spontaneous seizures, neuronal cell loss and memory deficits. These therapeutic effects persisted after drug discontinuation therefore indicating disease-modification effects. Importantly, these findings have high translational value since some of these drugs are medically used for other neurological or inflammatory diseases.
Model refinement.
A two-stage in vivo approach was established for facilitating 1. preclinical screening of antiepileptogenic drugs and 2. finding new biomarkers.
1. For the first stage, which should allow “medium-throughput drug screening”, the intrahippocampal kainate model in mice was chosen. This model is characterised by a latent period of several days followed by a high frequency of spontaneous electrographic seizures, thus allowing short EEG recording periods. For stage 2, to test the most interesting drug combinations coming out of stage 1, the refined intrahippocampal kainate model in rats was chosen.
Stage 1 in mice were used to determine: (a) the tolerability of novel combination of clinically used drugs; (b) the antiepileptogenic efficacy of tolerable combinations of clinically used drugs. Eighteen rationally chosen drug from different mechanistic categories (i.e. anti-inflammatory, neuroprotective, and classical antiepileptic drugs) were tested in 12 combinations. Four effective drug combinations were identified, with levetiracetam+topiramate being the most effective combination.
2. For the preclinical identification of biomarkers of epileptogenesis, it is mandatory to use experimental models with a similar severity of primary injury but providing animals with and without spontaneous seizures. In this context, several established post-SE models of acquired epilepsy were refined. By using these refined models, novel phenotypic, blood, MRI and EEG biomarkers were characterized and validated in WP1.
Novel models of epileptogenesis.
They were developed to support preclinical studies by the scientific community: the Theiler’s viral encephalitis model in mice, an immature rat model to study the encephalopathic effects of epilepsy, and a model of BBB dysfunction induced by intracerebral albumin infusion in rodents.
1.3.3. Developing polymeric nanotechnology-based formulations for combinatorial and larger biologic compound delivery into the brain
The delivery of disease-modifying treatments to the brain was advanced using gene therapy based on HSV-based amplicon vectors and polymeric nanotechnology-based formulations. A portfolio of amplicon vectors was constructed to evaluate the most suitable technologies and promoters. Specific strategies for short-term or long-term gene therapy were explored, including multi-cistronic constructs for the inclusion of several therapeutic genes in amplicon vector, as well as replication-defective HSV vectors gene combinations aiming at modulation of the NPY signal with promising results.
Nanotechnology-based formulations of two potential therapeutics were formulated and successfully delivered via the nose-to-brain administration route using the MET platform developed by 19-NM. We were able to exemplify the feasibility of this approach using the large peptide anakinra (17.3 kD) which has anti-inflammatory and potentially anti-epileptic properties blocking the IL1 receptor. Pharmacokinetic studies (07-UCL, 19-NM) indicated that anakinra reached the brain and hippocampus but concentrations were probably insufficient to allow effective blocking of epileptogenesis specifically in the acute kainate model (05-IRFMN). The approach was also demonstrated for a candidate small molecule with positive pharmacokinetics data suggesting further evaluation in preclinical models (19-NM).
1.3.4. Structured human epilepsy tissue banking as basis for translational characterization of biomarkers and therapy targets for antiepileptogenesis
From patients undergoing epilepsy surgery for seizure control, 08-UBMC has built up an extensive biosample resource comprising sera, CSF as well as biopsy brain tissue. The bank is complemented by harmonized clinical data and we have extensively used these resources for analyses. In this context, 131 human temporal lobe epilepsy fresh frozen hippocampi (HS n=90, lesion associated temporal lobe epilepsy n=41) were used for experimental approaches scrutinizing a low voltage-dependent calcium channel as biomarker of hyperexcitability conversion in epileptogenesis.
08-UBMC together with 02-L&B had previously shown in an animal model, that transcriptional upregulation of the voltage dependent calcium channel CaV3.2 leads to an increased propensity for burst discharges of hippocampal neurons in epileptic hippocampi. Current data from animal models further suggested a role for the metal-responsive transcription factor 1 (MTF1) in promoter activation of CaV3.2. To translate our findings into the human disease context and investigate whether the MTF1-CaV3.2 pathway is also involved in human TLE, we analyzed patients with hippocampal sclerosis (HS; Figure 5) for their MTF1 and CaV3.2 expression levels. Fresh frozen human non-epileptic control tissue is not available for obvious reasons.
Therefore, we used ‘lesion-associated’ TLE hippocampi as epileptic controls. In such patients, lesions including low-grade tumors as well as dysplasias are typically in the close vicinity of the hippocampal formation, which is resected in the course of lesionectomy. The sequential pattern of epileptogenesis after a brain insult is not present in ‘lesion-associated’ TLE clinical histories. Intriguingly, patients with HS show a higher MTF1 and CaV3.2 expression compared to lesion-associated controls (Figure 5; HS: n = 79; Lesion associated: n = 35; t-test: ***P ≤ 0.001). In addition, we observed a strong positive correlation between MTF1 and CaV3.2 expression levels (Figure 5), indicating that also in human TLE, the MTF1 dependent CaV3.2 upregulation is active – even in patients, which are regarded pharmacoresistant considering nowadays available drug treatment regimens.
Improved therapies for TLE are urgently needed, since nowadays still 30% of respective patients are refractory to pharmacotherapy. Intriguingly, for human TLE patients with the most frequent pathology pattern, i.e. HS, we here observed evidence for increased expression of CaV3.2 and MTF1 and a strong correlation between their expression levels. Many HS patients, in contrast to lesion-associated TLE have a clinical history of insult-induced epileptogenesis.
Figure 5. MTF1 and CaV3.2 expression levels co-segregate and are increased in hippocampal tissue of patients with hippocampal sclerosis (HS). (a) Epileptogenesis in a human individual without any previous neurological symptoms. The patient initially manifested clinically with SE. Coronal T2-weighted fast spin echo (A-C) and axial Diffusion-weighted spin echo EPI images show the rapid development of a right-sided hippocampal sclerosis in the clinical course. Initially, there is hippocampal swelling (A: arrow) associated with cytotoxic edema of the CA1 sector (D: arrow). Two weeks later, swelling and cytotoxic edema are somewhat regredient but still present (B, E). Only 8 weeks later, cytotoxic edema has disappeared (F) and hippocampal atrophy, i.e. HS has manifested (C: arrow). (b) Quantitative determination of MTF1 and CaV3.2 mRNA. MTF1 and CaV3.2 are significantly more abundantly expressed in hippocampal tissue of TLE patients with HS versus hippocampi from patients with lesion-associated TLE, i.e. in which seizures are explained by lesions such as low-grade neoplasms and/or focal dysplasia in the immediate vicinity or even including the hippocampal formation (HS: n = 79; Lesion associated: n = 35; t-test: ***P ≤ 0.001 with synaptophysin as reference gene). (c) Regression analyses of CaV3.2 mRNA versus MTF1 mRNA expression. A strong positive correlation between the expression of this TF and the dependent Ca2+ channel mRNA is even present in the heterogeneous group of human HS hippocampi.
Furthermore, the finding that MTF1 and CaV3.2 expression levels are substantially increased in human TLE patients currently regarded as pharmacoresistant, from which hippocampi were obtained in the present study, may serve as a strong argument for the development of therapy options in humans focusing on this ‘MTF1-CaV3.2-neuronal bursting’ mechanism as potential new treatment option (van Loo et al., Nat Commun 2015). Thus, the characterization of the MTF1-CaV3.2 axis have provided new biomarkers for epileptogenesis by complementary analyses in TLE mouse models and human epileptic brain tissue and thereby represents successful translational analyses by partners 08-UBMC and 02-L&B in the framework of EPITARGET.
Impaired memory performance is a major comorbidity of chronic progressive epilepsy. Molecular biomarkers for this aspect have not been systematically characterized. In a joint study, partners 02-L&B together with 08-UBMC have used a combinatorial approach starting from large scale gene expression from biopsy hippocampi and corresponding genetic data from a large series (n=79) pharmacoresistant TLE patients (Figure 6; Bungenberg et al., Neurobiol Dis 2016).
These patients were stratified according to presurgical memory impairment using an established four-tired grading system ranging from ‘average’ to ‘very severely’. Multimodal cluster analyses revealed molecules specifically associated with synaptic function and abundantly expressed in TLE patients with substantially impaired memory performance. In a subsequent promoter analysis, the single nucleotide polymorphism rs744373 C-allele was found to be associated with high mRNA levels of bridging integrator 1 (BIN1)/Amphiphysin 2, i.e. a major component of the endocytotic machinery and located in a crucial genetic AD risk locus.
Figure 6. Functional gene clusters with low expression levels in figural memory in the low performance vs. average performance group (A, B). ‘Gene ontology (GO) enrichment analysis (GOEA) was performed using the Biological Networks Gene Ontology tool (BiNGO, 3.0.3) and Cytoscape (3.0). The enrichment map represents the GO terms as nodes with enrichment significance encoded by color intensity and number of genes involved by the node size of the respective GO term. GOEA analysis of genes with a negative diffscore <-2.5 (n=383), i.e. genes for which the low performance group has a lower expression signal compared to the average performance group was carried out. Using a high p-value (0.00005) the categories with the highest enrichment score (17.28) have a common functional background, being mainly related to the neurological system, such as the synapse, neuron projection and synaptic transmission (modified from Bungenberg et al., Neurobiol Dis 2016).
Using in vitro luciferase transfection assays, the BIN1 promoter activation was observed to be genotype dependent and strongly increased by reduced binding of the transcriptional repressor TGIF. These data indicate that poor memory performance in patients with TLE strongly corresponds to distinctly altered neuronal transcript signatures, which - as demonstrated for BIN1 - can correlate with a particular allelic promoter variant (Bungenberg et al., Neurobiol Dis 2016). These promoter variants characterized in a joint effort of 02-L&B together with 08-UBMC can be directly analysed presurgically and may be of future value as biomarkers for tailor made clinical management with respect to major comorbidities of chronic progressive epilepsies.
1.3.5. Systems-biological approach based characterization of Sestrin-3 as epileptogenesis biomarker
11-IC jointly with 08-UBMC have used these recently established infrastructural preconditions in order to provide an integrated systems-biological approach to characterize biomarkers of epileptogenesis after SE complementarily in animal models and human TLE brain tissue samples. Therefore, the partners initially carried out comprehensive gene-regulatory network analyses in human brain tissue samples of temporal lobe epilepsy (TLE). We employed systems genetics approaches to characterize the genetic regulation of pathophysiological pathways in TLE (Figure 7). Using surgically acquired hippocampi from 129 pharmacoresistant TLE patients included in the EPITARGET virtual database, we identified a gene-regulatory network genetically associated with TLE that contains a specialized, highly expressed transcriptional module encoding proconvulsive cytokines and Toll-like receptor signaling genes (Johnson et al., Nat Commun 2015).
Figure 7. Identification of the TLE-network and functionally specialized transcriptional modules in human epileptic hippocampus. (a) Gene co-expression network identified in the hippocampus of TLE patients (TLE-network). Nodes represent genes and edges represent significant partial correlations between their expression profiles (FDR<5%). Node colour indicates the best GWAS P-value of association with focal epilepsy for SNPs within 100 kb of each gene. Boxes mark two transcriptional modules within the network. (b) Kyoto Encyclopedia of Genes and Genome (KEGG) pathways significantly enriched in the TLE-network (FDR<5%). The fold enrichment for each KEGG pathway is reported on the side of each bar. (c) Module-1 and Module-2 details. The size of each node is proportional to its degree of interconnectivity within each module. Light blue indicates genes showing nominal association with susceptibility to focal epilepsy by GWAS. Numbers in parenthesis indicate multiple microarray probes representing the same gene. (d) KEGG pathways significantly enriched in Module-1 (top) and Module-2 (bottom; FDR<5%). (e) Module-1 is significantly highly expressed in the hippocampus of TLE patients. mRNA expression of Module-1 (n=80 probes, representing 69 unique annotated genes) as compared with Module-2 (n=60 probes, representing 54 unique annotated genes), other network genes (n=371 probes, representing 319 unique annotated genes) and all other probes represented on the microarray (n=48,256). **P=3.8x10-4; ***P<10-10, Mann–Whitney test, two-tailed (adapted from Johnson et al., Nat Commun 2015).
Complementary RNA sequencing analysis in a mouse model of TLE using 100 epileptic and 100 control hippocampi demonstrated the pro-convulsive module to be preserved across species, specific to the epileptic hippocampus and abundant in expression also in the chronic stage of recurrent epileptic seizures (Figure 8). In the respective set of TLE patients from the EPITARGET database, we mapped the trans-acting genetic control of this proconvulsive module to Sestrin 3 (SESN3), and demonstrated that SESN3 intriguingly positively regulates the module in macrophages, microglia and neurons (Johnson et al., Nat Comm 2015).
Figure 8. TLE-network conservation in mouse epileptic hippocampus. (a) Human TLE-network genes that are conserved and co-expressed (84%) in the mouse hippocampus. Each node in the network represents a transcript that had significant partial correlation with at least another transcript in the network (FDR<5%). Conserved Module-1 and Module-2 genes are indicated in blue and green, respectively. (b) Distribution of significant partial correlations (FDR<5%) between pairs of transcripts from 10,000 bootstrap permutation samples in epileptic (top) and control (bottom) mouse hippocampus. In each case, the red line indicates the actual number of significant partial correlations (FDR<5%) between all genes in the network. The number of significant partial correlations observed in control hippocampus was no different from chance expectation (P=0.659). In contrast, the number of significant partial correlations detected in epileptic hippocampus was significantly higher than expected by chance (P=0.001). (c) Differential expression of Module-1 genes between control and epileptic mouse hippocampus shows specific enrichment for TLR-signaling and cytokine genes among the upregulated genes. Stars denote significant fold changes between epileptic and control mouse hippocampus (FDR<5%); blue bars indicate TLR-signaling and cytokine genes (from Johnson et al., Nat Comm 2015).
Finally, Morpholino-mediated Sesn3 knockdown in a recent zebrafish epilepsy model confirmed the regulation of the transcriptional module, and intriguingly attenuated chemically induced behavioral seizures in vivo. These results underline that the recent launching of the ‘virtual human brain tissue database’ represents a powerful repository for the integrated and harmonized experimental use and optimal exploitation of the rarely available epileptogenesis human brain tissue and corresponding biofluid specimens in the framework of EPITARGET and has already substantially contributed to externally highly visible research activities. The outlined results were only possible by the active mutual interactions of numerous partners of the consortium (01-ULUND, 02-L&B, 04-UEF, 07-UCL, 08-UBMC, 09-NENCKI, 11-IC, 12-UNIFER, 14-BGU) and EPITARGET provides a unique structure for extrapolating these results into even more refined biomarker molecules for epilepsy also after the termination of EPITARGET. The data represent an integrated approach to define a biomarker mechanisms and potential therapy target based on integrated analyses in human brain tissue and corresponding animal models and further demonstrates the experimental power of the EPITARGET epilepsy brain tissue biobank.
1.3.6. Structured human status epilepticus (SE) tissue banking and analyses of immune-related mechanisms as biomarkers of early epileptogenesis.
10-AMC has achieved the structured biobanking of post mortem human brain tissue samples of patients after status epilepticus (SE) and traumatic brain injury (TBI), of surgical brain tissue and biofluids (CSF and blood) of patients after TBI, and of fresh frozen hippocampal biopsy specimen of pharmacoresistant epilepsy patients as well as of blood from a subset of patients. These biospecimens have been extensively used for analyses of the use of molecules involved in early epileptogenesis as biomarkers.
The proteasome is a multisubunit enzyme complex critical for protein degradation. 05-IRFMN revealed a dysregulation of the (immuno)proteasome pathway in a rat model of chronic epilepsy after SE (Mishto et al., Brain Behav Immun 2015). During inflammation, the constitutive subunits are replaced by their inducible counterparts, which results in the formation of the immunoproteasome. In post-SE rats, the expression of the β1i and β5 subunits was higher in neurons within the hippocampus and piriform cortex during the acute phase (one day post-SE) compared to controls. At this time point, β5 expression was also higher in astrocytes within the hippocampus and piriform cortex. During the latent (one week post-SE) and chronic phase (three months post-SE), the expression of β1, β1i, β5 and β5i was higher in neurons and in most cases also in astrocytes within the hippocampus and piriform cortex compared to controls. Interestingly, in rats that developed a progressive form of epilepsy the expression of (immuno)proteasome subunits was higher than in rats with a non-progressive form.
10-AMC with 05-IRFMN have translated these data from rodent models into the human disease context by addressing the expression pattern of constitutive (β1, β5) and immunoproteasome (β1i, β5i) subunits by immunohistochemistry in surgical specimens from patients with malformations of cortical development and in patients with temporal lobe epilepsy (TLE) and hippocampal sclerosis, as well as in the hippocampus and cortex of patients that died after status epilepticus (SE) or traumatic brain injury (TBI) and during epileptogenesis in the post-SE epilepticus rat model (van Scheppingen et al., J Neuroinf 2016). Thereby, increased expression was observed in both, focal cortical dysplasia, FCD IIa and b; cortical tubers from patients with Tuberous Sclerosis Complex (TSC). Intriguingly, β1, β1i, β5 and β5i were detected (within cytosol and nucleus) in dysmorphic neurons, balloon/giant cells and reactive astrocytes. In human TLE, the expression of β1, β1i, β5 and β5i was higher in hippocampal neurons and astrocytes compared to controls and was most pronounced in patients with hippocampal sclerosis. In patients that died after SE, the expression of subunits β5 and β1i was higher in neurons and astrocytes, respectively, compared to controls. Increased expression was particularly detected in the cortex of TBI patients.
Subsequently, in vivo experiments showed that (immuno)proteasome protein expression and the number of spontaneous seizures was lower in rats that were treated with immunomodulator drug rapamycin (inhibitor of the mammalian target of rapamycin, mTOR pathway) compared to vehicle-treated rats. In vitro studies using cultured human astrocytes showed that gene expression of all subunits increased after IL-1β stimulation, which could be attenuated by the immunomodulators rapamycin or curcumin. Dysregulation of the (immuno)proteasome system during epileptogenesis occurs in epileptic rat model and human brain and may serve as biomarker as well as therapy target in the future.
Distinct miRNAs including miR-9a-3p have been suggested as putative biomarkers of epileptogenesis after SE in rats (12-UNIFER & 11-IC; Roncon et al., Sci Rep 2015). Conversely, post mortem human brain tissue has been used in several studies addressing the cellular expression of microRNAs deregulated during epileptogenesis in experimental models (13-AMC ; van Scheppingen et al., Glia, 2016). Increasing evidence supports the involvement of microRNAs in the regulation of astrocyte-mediated inflammatory response. To study the role of inflammation-related microRNAs in epilepsy associated pathologies, real-time PCR and in situ hybridization were employed to characterize the expression of miR21, miR146a and miR155. In addition, cultured human astrocytes were used to study the regulation of the expression of these miRNAs in response to the proinflammatory cytokine IL-1β and to evaluate the effects of overexpression or knockdown of miR21, miR146a and miR155 on inflammatory signaling. IL-1β stimulation of cultured glial cells strongly induced intracellular miR21, miR146a and miR155 expression, as well as miR146a extracellular release. IL-1β signaling was differentially modulated by overexpression of miR155 or miR146a, which resulted in pro- or anti-inflammatory effects respectively. These data provide supportive evidence that inflammation-related microRNAs play in focal epilepsies. In particular, miR146a and miR155 appear to be key players in the regulation of astrocyte-mediated inflammatory response, with miR146a as most interesting anti-inflammatory therapeutic candidate. This specific miRNA has been further evaluated in collaboration with 05-IRFMN in WP2.
Resolution of inflammation is a highly coordinated and active process that is controlled by endogenous pro-resolving lipid or peptide mediators and it is instrumental to switch off acute inflammation. If this mechanism fails, inflammation might perpetuate resulting in varying degree of tissue injury or dysfunction. 05-IRFMN has investigated whether the brain immune response triggered by epileptogenic injuries is inefficiently controlled by pro-resolving endogenous molecules thus resulting in chronic inflammation using rodent models of epileptogenesis. Increased gene expression of pro-resolving G-protein coupled receptors (GPCRs; ALXR and ChemR23) was observed in resected hippocampus of patients with TLE and hippocampal sclerosis, as well as in the hippocampus of patients that died within 5 days or more than 7 days after SE, i.e. reflecting the stage of putative early epileptogenesis in humans. In control, ALXR and ChemR23 were expressed predominantly in neurons, while after SE and in epileptic tissue we observed an increase in the expression of the two receptors in glial cells (Figure 9). These data agree with the increase in glial expression of the two receptors observed in mice after SE and in the chronic epileptic phase.
Figure 9. Expression of ALXR and Chem23 in control human brain tissue as well as at early stages in post mortem tissue after SE and in chronic mesial (m)TLE.
14-BGU demonstrated that recurrent seizures and the associated excessive glutamate release lead to increased vascular permeability in the rat cerebral cortex, through activation of NMDA receptors (Vazana et al., J Neurosci 2016). After blood-brain-barrier (BBB) disruption, albumin extravasation induces astrocytic activation and potentially seizures (Kim et al., J Neurosci Res. 2016). Astrocytes immediately respond to brain damage and play a critical role in the development of postinjury epilepsy. Astrocytes were shown to regulate extracellular matrix (ECM) remodeling, which can affect plasticity and stability of synapses and, in turn, contribute to the epileptogenic process.
Complementarily, post mortem brain tissue of patients that died after TBI has been used to support the occurrence of the BBB dysfunction and brain inflammation in perilesional cortex and thalamus in lateral fluid percussion injury rat model (04-UEF; manuscript in preparation). Evaluation of both cortex and thalamus of patients that died a few months after TBI confirmed the association between BBB dysfunction and activation of astrocytes, microglia, as well as deposition of iron and calcifications (Figure 10, Figure 11).
Figure 10. Blood-brain barrier dysfunction in perilesional human cortex and thalamus 6 months after TBI. A, B: iron deposits (blue); C, D: calcifications; E, F: albumin.
Figure 11. Brain inflammation in perilesional human cortex and thalamus 6 months after TBI. A, B: CD68 positive macrophages/microglia/monocytes; C ,D: vimentin positive astrocytes; E, F: HLA-DP, DQ,DR positive microglia.
TBI triggers a cascade of molecular and cellular changes leading to development of comorbidities such as progressive neuronal loss and hippocampal memory decline. 04-UEF analyzed microRNAs (miRNAs) related to chronic alterations in hippocampal molecular networks using Affymetrix microarray dataset obtained from rat dentate gyrus at 3-months after lateral fluid-percussion injury. Microarray data was used for Ingenuity Pathway Analysis (IPA), which revealed 18 up-regulated targets for miR-124, suggesting a down-regulation of miR-124 (z-score=-4.176 p<0.01). The deregulation of this specific miRNA has been confirmed in different brain injury models. Evaluation by in situ hybridization performed in rat and TBI human tissue confirmed the predominant neuronal expression of this miRNA within both hippocampus and cortex. Analysis of data is in progress and extended to the post-SE human material. Intriguingly, 04-UEF has in a parallel approach demonstrated that a TBI transcript signature (TBI-sig) consisting of 4964 regulated genes in the perilesional cortex and 1966 in the thalamus (FDR < 0.05) of the TBI rat model insinuates susceptibility to distinct drug treatment (Lipponen et al., Sci Rep 2016). 04-UEF observed by LINCS analysis 11 compounds that showed a strong connectivity with the TBI-sig in neuronal cell lines. Of these, celecoxib and sirolimus have been demonstrated to have a disease-modifying effect in in vivo animal models of epilepsy. These results reveal several molecules with pre-clinical evidence as biomarkers for epileptogenesis in TBI human brain tissue and corresponding animal models and provide arguments in favor of certain therapy regimens.
The present results in the framework of EPITARGET clearly demonstrate the applicability of human brain tissue samples of patients that died after SE as well as TBI. The biobank relies on highly selected and very rare human brain tumour samples and represents a unique scientific asset based on the multilateral activities within the network structure of EPITARGET.
1.3.7. Characterization of 'early epileptogenesis' biomarkers identified in TBI animal models in human biopsy brain tissue of patients with epilepsy.
14-BGU together with 08-UBMC have focused on alterations of TGFß signaling related to expression of inflammatory genes and degradation of perineuronal nets around inhibitory interneurons particularly at early stages after distinct epileptogenic insults to the brain (Kim et al., Sci Rep 2017). Brain damage following stroke or brain trauma frequently results in epileptogenesis. Such transient insults commonly involve lasting dysfunction of the blood-brain barrier (BBB). Previous data from 14-BGU had shown that chemically-induced disruption of the BBB induces pathological hyperexcitability manifested by epileptiform activity and recurring seizures. Exposure of the brain environment to generally serum-derived albumin is sufficient to induce epileptiform activity and recurring seizures, and this pathological cascade depends on activation of TGFβ signalling.
14-BGU observed albumin extravasation in brain regions with hyperexcitable network activity in the injury models used in the respective study – peri-infarct hippocampus following photothrombotic cortical stroke and partially isolated undercut cortex (Kim et al., Sci Rep 2017). Comparative transcriptome analyses in respective models predicted remodelling of extracellular matrix (ECM) as a common response to different types of injuries. ECM-related transcriptional changes were induced by the serum protein albumin via TGFβ signalling in primary astrocytes. 14-BGU studied, whether the predicted ECM alteration may be associated or even causatively involved in chronic deficits of inhibitory signalling and pathological hyperexcitability observed following injury. Particularly, a specialized ECM structure named perineuronal nets (PNNs) is tightly associated with parvalbumin (PV) expressing fast-spiking inhibitory interneurons. Co-staining of wisteria floribunda agglutinin (WFA), a broad marker labelling PNNs, with immunostaining for parvalbumin (PV), a marker for fast-spiking interneurons revealed significant main effects revealed that perineuronal nets around PV(+) interneurons were degraded following traumatic brain injury induced by the undercut model in the present study only seven days after induction of the trauma.
14-BGU together with 08-UBMC next studied whether albumin-induced TGFβ signalling activation and subsequent PNN degradation occur in the hippocampus of patients with focal epilepsies, i.e. temporal lobe epilepsy (TLE) who frequently report transient brain insults to have occurred months to years before the onset of chronic recurrent seizures. 14-BGU quantified phosphorylation of Smad2 (pSmad2) protein as a marker of TGFβ receptor-mediated signalling in hippocampal tissues resected from TLE patients and autopsy controls from 08-UBMC. The number of GFAP (+) astrocytes co-localized with pSmad2 protein was significantly increased in hippocampal sections of TLE patients compared to controls (Figure 12). Consistent with the findings in the rodent brains the percentage of PV(+)/WFA(+) interneurons was significantly decreased in the hippocampus resected from TLE patients in compared with controls (Figure 12). These results corroborate the potential role of TGFβ-regulated ECM remodelling in the pathology of chronic hyperexcitability.
Figure 12. Increased expression of astrocytic pSmad2 and reduced association of PNNs around PV(+) cells in human epileptic hippocampi. (a) Representative micrographs of human hippocampi stained for GFAP and phosphorylated Smad2, a downstream effecter of TGFβ signalling. The hippocampi were resected from temporal lobe epilepsy (TLE) patients or age-matched autopsy controls. Representative GFAP(+) astrocytes expressing pSmad2(+) are indicated by white arrowheads. Scale bar=50μm. (b) Co-localization of pSmad2 with GFAP was increased in TLE patients (n=5) compared to controls (n=3). (c) Representative confocal micrographs of human hippocampal tissues stained for GFAP, PV, and PNNs (WFA) resected from TLE patients or controls. Scale bar=50μm. White and blue arrowheads indicate representative PV(+)/WFA(+) and PV(+)/WFA(−) cells, respectively. (d) The percentage of PV(+)/WFA(+) cells was decreased in the hippocampus of TLE patients (n=4) compared to controls (n=4). Mann-Whitney test, two-tailed, was used for statistical analyses. *p<0.05. (e) A working model. Several points for feed-forward loops (i-iii) can lead to chronic hyperexcitability and the development of epilepsy. The dysfunction of blood-brain barrier (BBB) and the ensuing entry of albumin into brain parenchyma activate TGFβ signaling. Comparative transcriptome analyses predicted the activation of the common core signalling transduction including MAPK pathway, Stat3, NFκB, AP-1, and ETS1. These signalling pathways can elicit the reciprocal activation of inflammation and extracellular matrix (ECM) remodelling (i). ECM remodelling can trigger the transformation of a latent from of TGFβ to its active form75 as well as exacerbate BBB dysfunction58 (ii). The degradation of perineuronal nets (PNNs) around fast-spiking interneurons and aberrant excitatory synaptogenesis occur in the course of ECM remodelling, presumably leading to functional alterations in inhibition and abnormality in synaptic plasticity that may contribute to excitation/inhibition (E/I) imbalance and ultimately the occurrence of seizures. Finally, seizures per se cause BBB dysfunction, inflammation, and the upregulation of MMPs activity (from Kim et al., Sci Rep 2017).
These results integrating data from TBI animal models and correspondingly human epileptic hippocampi provide new pathogenetic insights in how albumin extravasation that occurs upon BBB dysfunction in various brain injuries and ways to predispose neural circuitry to the development of chronic inhibition deficits and suggest new tissue biomarkers and a novel rationale for therapy targets in TBI-induced epileptogenesis.
1.3.8. Use of biosample collections comprising blood, CSF and brain tissue from patients with chronic progressive epilepsy for translational characterization of biomarkers and therapy targets
Individual and combinatorial biomarkers have been characterized and released with strong impact for assessment of the risk and potential gain for certain parameters. This includes the establishment of a systems genetics approach starting from gene expression data from hippocampi of the pilocarpine-induced status epilepticus model for identifying cell membrane receptors with a direction-specified influence over disease-related mRNA profiles (11-IC, 02-L&B, 05-IRFMN, 08-UBMC, 13-AMC). 11-IC has designed a very complex novel systems neurobiological approach in order to characterize biomarkers for chronic progressive epilepsy and derive new therapy concepts (Figure 13).
Figure 13. The study plan was based on 100 mice with epilepsy (pilocarpine post status epilepticus model of temporal lobe epilepsy) and 100 control (pilocarpine-naïve) matched littermate mice. At 4 weeks post status epilepticus, each mouse was continuously monitored using 3D accelerometery and video monitoring for 14 days to record seizure frequency and severity. High-throughput mRNA sequencing (RNA-seq) was generated using RNA from snap-frozen whole hippocampus samples from the mice and gene expression profiles were used to generate co-expression modules. Coexpression modules with a potential relationship to epilepsy were prioritized using the following criteria: (i) differential co-expression between epileptic and healthy hippocampus (mouse and human TLE), (ii) correlation of module expression with seizure frequency (mouse), and (iii) conservation in the human epileptic hippocampus. Modules meeting these criteria were considered candidate modules for epilepsy, and subjected to ‘Causal reasoning analytical framework for target discovery’ (CRAFT) analysis to identify membrane receptors predicted to restore disease module expression toward health (modified from Srivastava et al., Nature Commun 2018).
Analysis of the biological terms and canonical pathways enriched among the 28 co-expression modules in the mouse epileptic hippocampus revealed that the modules were generally enriched for specific functions. Among the modules with overlapping functions, modules 5, 16, and 18 were enriched for “immune response” processes (Benjamini–Hochberg (BH)-corrected P = 2.1 × 10−11, P = 1.4 × 10−6, and P = 1.3 × 10−33, respectively), and modules 10, 14, 26, and 29 were enriched for neuronal functions including “synaptic transmission” (BH P = 4.4 × 10−11, P = 0.02 P = 4.0 × 10−3, and P = 4.0 × 10−4, respectively). To provide insights into the cell-type expression of the modules, cell-type marker genes derived from singlecell RNA-seq analysis of the mouse hippocampus were used. In this context, “immune response” modules 16 and 18 were enriched for microglia marker genes, whilst “synaptic transmission” modules 10, 14, 26, and 29 were specific for neuronal cell types. To further prioritize the modules in terms of their relationship to epilepsy, the correlation between each module’s expression and seizure frequency was explored (Figure 14).
Figure 14. Correlation of module expression with seizures. (a) For each co-expression module from the epileptic mouse hippocampus, we plotted the significance (−Log10 FDR) of the Spearman’s correlation between the module’s eigengene and seizure frequency (bar plot), and the percentage of variance in seizure frequency explained by the module’s eigengene (R2, dotted line). Modules marked with a blue arrow are the modules differentially co-expressed between the epileptic mouse hippocampus and the control mouse hippocampus. Modules highlighted in grey (bar plot) are significantly (FDR <0.05) correlated with seizure frequency. (b) Volcano plot of average (Spearman’s) correlation of a module’s genes with seizure frequency (X-axis) versus the significance of the module’s enrichment for genes individually correlated with seizure frequency (QTT genes) (Y-axis) for the nine modules differentially co-expressed in epilepsy and correlated with seizures by module eigenegene (modified from Srivastava et al., Nature Commun 2018).
For the nine modules differentially co-expressed in epilepsy and correlated with seizures (i.e. modules 2, 5, 8, 10, 16, 18, 21, 22, and 24), it was then assessed whether the module was conserved in the human epileptic hippocampus. Using human orthologs of mouse module genes and genome-wide gene expression data from 122 human epileptic hippocampus samples surgically ascertained from TLE patients, we found that all nine modules were conserved in the human epileptic hippocampus (FDR<0.05) (Figure 15).
Figure 15. Correlation of module expression with seizures. Grey – not significant at FDR 5%. * = modules differentially co-expressed in mouse and human TLE + correlated with seizures + conserved in human TLE (modified from Srivastava et al., Nature Commun 2018).
The conservation of these nine modules across human and mouse TLE provides an independent line of evidence for the validity of these modules, and further supports the relevance of the pilocarpine post SE mouse model of TLE to human TLE. As a final assessment of the relationship of these nine mouse TLE modules to human epilepsy, we tested whether each module was also differentially co-expressed in human TLE. In this analysis, for each module, we compared intra-module correlations in the human epileptic hippocampus with that in the non-diseased human hippocampus using post-mortem hippocampal samples ascertained from people with no history of psychiatric or neurological disease. Among the nine mouse modules differentially co-expressed in epilepsy and correlated with seizures, seven (5, 10, 16, 18, 21, 22, and 24) were also differentially co-expressed in human TLE (Figure 15). These seven modules were selected for further analysis. Specifically, we hypothesized that focusing on these seven modules (and by extension their enriched functional pathways) would provide a starting point for the development of new therapy approaches for epilepsy. Microglia-related combinatorial mRNA modules (module 18) correlate strongly with seizure activity. The further bioinformatic characterization of potential therapeutic targets predicted the tyrosine kinase receptor Csf1R in complementary mouse model and human epileptic hippocampi (Srivastava et al., Nature Commun 2018). The predicted effect of Csf1R blockade in attenuating seizures was then validated in pre-clinical epilepsy models. The released corresponding mRNA signature of module 18 has a strong potential to assess seizure risks and points to innovative treatment options.
Potential Impact:
SEE ATTACHED REPORT
1.4.1. Potential impact
1.4.1.1. Identifying novel biomarkers and their combinations in animal models for further preclinical studies towards more precise diagnoses and treatments
Novel biomarkers.
EPITARGET has provided evidence of plasma HMGB1 with or without combination to TSPO PET as a biomarker for epileptogenesis after SE. EPITARGET has also identified several miRNAs that will be further investigated as prognostic and predictive biomarkers for epileptogenesis. In addition, EPITARGET MRI studies have revealed imaging biomarkers for experimental post-traumatic epilepsy and progression of brain damage. These data will impact diagnosis and stratification of patients in clinical practice, and pave the way towards more affordable preclinical and clinical treatment trial to combat acquired epileptogenesis in humans.
05-IRFMN, 04-UEF and 17-MHH groups also generated promising data concerning the use of [18F]deprenyl as a novel prognostic biomarker. [18F]Deprenyl autoradiography could differentiate between HMGB1-high animals (developing epilepsy) and HMGB1-low animals (not developing epilepsy).
Social isolation and epilepsy.
The Debski et al study made us realize that social isolation may constitute an important confounding factor in epilepsy research. We performed a dedicated study, which questions the interpretation of the results obtained when animals are singly housed (tens of thousands), as commonly done in epilepsy (Manouze et al., 2019, eNeuro). This study was highlighted in Science, discussed in several commentaries and editorial in the October issue of Epilepsia; and generated a record high number of tweets in the scientific community. The impact of this study on future experimental and preclinical work based on epilepsy models is going to be significant as animal housing conditions, particularly social isolation, can be a major confounding factor for interpretation and comparisons of molecular, behavioural, and imaging data obtained in different laboratories in trials aiming at treatment and biomarker discovery.
Information processing in the “epileptic” brain.
The Frigerio et al. paper led us to ask whether information processing may be drastically modified in experimental epilepsy. As a first step, we characterize information processing in the “normal” brain. We have unravelled the basic properties of the neuronal language during sleep (Clawson et al., 2019, Science Adv). We have now finished, the second part, demonstrating how the neuronal language is modified in epilepsy (Clawson et al., in preparation). This study will have impact on how information processing should be considered in further studies on epilepsy models and taken in account as a significant comorbidity to be accounted for in treatment paradigms.
Epilepsy Preclinical Biomarker Bank (EPBB).
EPITARGET RedCap database (see section 1.3.1.1) was the first systematic preclinical database in epilepsy for multicentre use. It will allow different study sites to link the animal specific information with the data collected (e.g. miRNA-seq data). It will remain as a part of 01-ULUND’s informatics infrastructure. If new funding opportunities arise, investments will be made to maintain it as long as possible. This database will have impact on epilepsy research community in terms of defining CDEs for collaborative and multicentre studies, as well as providing access to already established database for future datamining and networking. The experience gained paves the way in preclinical laboratories towards large multicentre biomarker and therapy studies, which is expected to increase the reproducibility and translational value of preclinical data to clinic, and thus, speed up the clinical biomarker and therapy development.
EPITARGET RedCap and EPBB have already significantly helped the efforts of harmonization of preclinical epilepsy studies. For example, it has served as a template for ILAE Working Group of Preclinical Harmonization, which includes many EPITARGET partners. It was also used to help in design of the NIH Centers-without-Walls funded EpiBioS4Rx preclinical RedCap database at LONI in USC.
1.4.1.2. Unravelling the complex pathophysiology of epileptogenesis for preclinical development of new disease modifying treatments
This project enabled a better understanding of the mechanisms leading to epilepsy development in structural forms of epilepsy with an acquired aetiology. This novel information also set the basis for investigating the same mechanisms in other forms of epilepsy therefore searching for common mechanisms across the epilepsies. By highlighting novel targets for antiepileptogenesis interventions we developed combinatorial drugs treatments for taking into consideration the multiple contributing mechanisms that in concert are responsible for epileptogenesis and to attain more significant disease modifying effects vs single treatments. We identified novel disease-modifying combinatorial treatments using anti-inflammatory and anti-oxidant drugs which are in medical use, and also novel investigational drugs. These medically used drugs with a good tolerability and safety profile could be tested by repurposing these drugs in clinical trials in drug-resistant epilepsies or in prospective studies for the prevention of epilepsy in patients who are at risk of developing epilepsy. In view of the enhanced pathological excitability underlying also non-epileptic disorders, such as migraine, neuropathic pain and bipolar disorders, the novel antiepileptogenic treatments may have additional therapeutic indications other than epilepsy. Therefore this research line opened novel opportunities and created impact on better epilepsy treatment, and particularly, hypothesis/mechanism driven combinatorial approach in this regard.
1.4.1.3. Developing polymeric nanotechnology-based formulations for combinatorial and larger biologic compound delivery into the brain
The delivery of biologics and small molecule drugs to the brain remains a significant barrier to the successful pharmacotherapy not only of epilepsy but also for many other diseases such as e.g. neurodegenerative conditions. The viral vectors generated and characterized in the frame of EPITARGET are promising tools for the gene therapy of epilepsy as well as other neurological disorders. Work under EPITARGET with the MET delivery system towards using the nose-to-brain route has added important evidence of the feasibility of this route. Specifically, this has increased our understanding of the range of molecules this might be applicable to but also restrictions with respect to achievable dose/ concentration.
1.4.1.4. Translation of biomarkers and therapy targets into the human disease context
With respect to the translation of biomarkers and therapy targets into the human disease context (WP04), the potential impact of EPITARGET is particularly significant due to the fact that highly selected human epilepsy brain tissue sample based data is integrated with results from a wide range of animal model studies. This holds true for animal model data derived from blood samples, peripheral organs, brain tissue, electrophysiological data, behavioural data, and imaging data from different models of epileptogenesis. Furthermore, brain and blood levels of antinflammatory mediators and the cellular expression of their cognate receptors in SE and TBI models have been made available for translational analyses in human epilepsy brain tissue and corresponding biofluids. As we provided examples for under ‘Main Results’, successful translation of biomarkers of distinct excitability mechanisms emergent during epileptogenesis as well as inflammatory dynamics as surrogate markers of altered seizure threshold and biomarkers in serum/CSF was accomplished. Furthermore, several biomarker candidates reflect mechanisms that can serve as therapy targets. These data can have significant impact by further translation from the preclinical human disease in the clinical context. The human brain tissue repositories generated in the framework of EPITARGET will be further used in the future and even extended by ongoing introduction of human epilepsy brain tissue samples. Given the integration of these data with exact clinical information, this unique clinic-scientific asset will outreach EPITARGET and provide one of the worldwide leading tissue repositories on human epilepsy for translational study approaches with high impact.
1.4.1.5. New collaborations and jobs
EPITARGET created the opportunity for its partners to engage in new research collaborations that may have not been possible otherwise. For instance, 05-IRFMN and 04-UEF’s teams together looked at the consequence of inflammation on signal integration in the dendrites (Frigerio et al., 2018, Mol Neurobiol). 09-NENCKI, 06-TiHo, 08-UBMC and 04-UEF demonstrated the circadian cycling of the molecular architecture of the hippocampus and its alteration in epilepsy (Debski et al., Science Adv, in revision – on BioRxiv). A spin off paper is in preparation on the circadian regulation of metabolism. The partners involved in the identification of miRNA biomarker (Task 3.1) namely 04-UEF, 09-NENCKI, 11-IC, 12-UNIFER and 10-AMC, generated a common paper (Srivastava et al., eNeuro 2017), designed CDE for the multi-center study (van Vliet et al., Epilepsia 2017) and are currently continuing their collaboration to finalize the data jointly collected. Lastly, EPITARGET generated several position papers to which many partners contributed (e.g. Lapinlampi et al., 2017, Epi Res; Klein et al., 2018, Epilepsia).
These collaborations have a significant impact on European networking and future collaborations on European arena. The human biospecimen repository partially created on the basis of EPITARGET on a comprehensive European scale, the experimental workflows in translating biomarker candidates from experimental animal research into the human disease context that have been established through the integrated efforts of EPITARGET consortium members and particularly the knowledge generated on these individual aspects as well as the downstream steps for e.g. patenting of key results have been able to harness individual activities in national Epilepsy Centres for leveraging their quality onto a European collaborative stage with quality levels to compete worldwide as documented by the seminal output of EPITARGET. Several research initiatives on a European level have been initiated, in which one or more EPITARGET partners have prominent roles as PIs or coordinators (e.g. UBMC, AMC).
Another important outcome is the new collaborative network Epi-CLUSTER which has received financial support from EBRA (see section 3.3.2 in D6.03). This collaborative network is mainly based on EPITARGET and other epilepsy consortia and their beneficiaries, and is designed to reach consensus between various epilepsy constellations how to work together, how to coordinate the efforts together with other initiatives (e.g. Epilepsy Forum 2020), agreeing on main objectives and speaking with a single voice at a European level in alignment with EBRA’s objectives and activities. Other collaborative applications for grants are being developed building on the epilepsy consortia meeting epiXchange held in Brussels in May 2018, which also was based on networking between EPITARGET and other EU funded epilepsy projects. Such efforts and collaborations would not have been possible without EPITARGET, as well as other epilepsy consortia.
In addition, the work carried out in EPITARGET had a significant impact on the career of several young researchers involved in the project, by creating new jobs and giving opportunities for career development. Thus, the paper published on the Debski et.al study (see section 1.3.1.1.) played an instrumental role in the permanent research position obtained by Pascale Quilichini in the 10-AMC group. That same group published a general review on the pilocarpine model of epilepsy (Lévesque et al., 2016, J Neurosc Met, 100 citations), and two book chapters on models and a general theory of vulnerability to depression, in particular in epilepsy (Bernard, 2016, Cold Spring Harbor; Ivanov & Bernard, 2017, Models of Seizures and Epilepsy). This part has been extremely successful with several high-profile publications (2xAdv Mat, Science Adv, PNAS – two others in revision), largely covered by the media, with translational potential (under way). The results allowed two scientists from the group to obtain very highly competitive permanent positions at Inserm (8 per year for all neuroscience) and an ERC starting grant; ensuring continuity for the projects.
1.4.2. Dissemination activities
1.4.2.1. Publications
Over the course of the project, the partners published a total of 177 peer-reviewed articles in 72 different journals (see section 2. Use and Dissemination of Foreground). These include several publications in high impact journals, such as The Lancet Neurology (Impact factor 2018: 28.755) Advanced Materials (IF 2018: 25.809) and Acta Neuropathologica (IF 2018: 18.174) (Table 1).
Most articles had less than 15 citations, however 19 papers had between 50 and 100 citations and 7 papers had more than 100 citations (including 2 papers with more than 200 citations; see Table 1).
Table 1. EPITARGET Publications figures
A major effort in dissemination was the generation of a seminal review paper on the topic of the consortium by EPITARGET PIs. The review, entitled ‘Advances in the Development of Biomarkers for Epilepsy’ was published in the Lancet Neurology, one of the premiere journals in Neurology (Pitkänen et al., Lancet Neurology, 2016). This review paper has generated great interest in the field and was largely cited . Due to the rapid developments in the fields, EPITARGET partners are working towards generating two additional authoritative reviews with a similar impact: a first review paper will address novel developments in the biomarker field (led by A. Pitkänen of UEF); and the second will address combinatorial treatments (led by A. Vezzani of IRFMN, and W. Löscher of TiHo).
The EPITARGET consortium also contributed (coordinated by A. Pitkänen, UEF), within the epiXchange cluster (see section 1.4.2.2. Events) to an article published in Epilepsia Journal, presenting major achievements by the cluster in the areas of biomarkers, genetics, therapeutics, comorbidities and biobanks as well as recommendations for future research to develop and bring novel solutions to the patients (Advancing research toward faster diagnosis, better treatment, and end of stigma in epilepsy , Pitkänen et al., 2019).
1.4.2.2. Events
In addition to publications, the partners engaged in more than 350 dissemination activities see section 2. Use and Dissemination of Foreground). These included over 200 oral presentations and 100 posters communicating project results at major scientific events such as the European Congresses of Epileptology (ECE), the International Epilepsy Congresses (IEC) and Workshops on Neurobiology of Epilepsy (WONOEP) , or the Annual Meetings of the American Epilepsy Society .
The consortium also held four Young Researchers’ Symposia back-to-back with the project’s General Assembly meetings, during which PhD students and Post-doc researchers were able to present their work in the project. To accompany these symposia, abstract booklets were compiled by ARTTIC including abstracts of all oral presentations and posters. Awards in the form of a conference participation fee of a maximum amount of 400€ were distributed to the best oral presentation and best poster. The Executive Board members and SAB members constituted the jury.
EPITARGET was also counted among the seven currently ongoing and former collaborative projects funded by the EC (Desire, EpimiRNA, EPIPGX, Epistop, EPITARGET, EpiCare and Epixchange) which co-organised the epiXchange conference to disseminate project results and discuss the future of epilepsy research in the EU and worldwide. The conference took place on 23 May 2018 in Brussels.
The first day of the conference was open for the public and highlighted the achievements from the participation projects so far, their impact and main challenges and next steps. It included keynote talks from renowned scientists in the field and also provided a platform for scientists to present their work during a poster session. The event gathered 170 participants (28 from EPITARGET) from 18 European and 5 non-European countries, in addition to 250 watching the live-stream on Facebook . Officials from the EC, the European Brain Council, EMA, patient organisations (e.g. Epilepsy Foundation and Federazione Italiana Epilessie) and pharma (strong presence of UCB pharma) also attended the conference were 28 talks and 38 posters divided into 5 main themes were presented. epiXchange was promoted by 800 tweets and retweets from October 2017 to August 2018. Presentations and videos from the conference can be found on www.epixchange2018.eu.
In connection to conference, a workshop was held on 24 May at the DG Research and Innovation to discuss policy and strategy on future epilepsy research with a global approach. The main focus was to identify gaps and needs, impact enhancement of results, like pathways (including regulatory) to clinical guidelines, product development and policy recommendations. This policy day aimed to define and analyse bottlenecks and give suggestions for additional efforts need to be made to overcome those. This discussion day addressed regulatory pathways for treatments or diagnosis and highlight what kind of infrastructure is needed for the future.
1.4.2.3. Website and social media
Publications, events and project news were regularly communicated about on the project’s website (https://www.epitarget.eu/) and Facebook page (Figure 20) as well as in the project newsletters.
Figure 20. Screenshot of EPITARGET's Facebook page
The EPITARGET logo has been developed in period 1 by partner 15-GABO:mi. All dissemination material was based on the colour scheme established in the logo to generate a consistent corporate project identity.
The EPITARGET website went online in the 1st period and was updated continuously since then with regular news items as well as update on the project content.
Since the launch on 19.05.2014 the EPITARGET website had 20,911 visits with an average session duration of 00:01:29 and on average 2.03 pages opened per session. These numbers include all page visits.
The website is accessed from all over the world, also from outside of Europe (Figure 21): the top three countries of origin of visitors are the US, Germany and France. The website has totalized 41,902 page views.
The website was maintained by 15-GABO:mi until 30 June 2016 and by 20-ART since 1 July 2016.
Figure 21. Geographical distribution of all website sessions longer than 2 seconds between 19 May 2014 and 31 October 2019.
The latest news on the EPITARGET project were circulated with a biannual newsletter which was implemented as part of the EPITARGET website . The newsletter included updates on EPITARGET meetings and awards as well as information on the latest EPITARGET publications and posters.
1.4.3. Exploitation of results
EPITARGET produced Deliverable 6.3 “Final plan on dissemination, exploitation and transfer of final results” which has been actively prepared in the last two years of the project under the lead of 02 L&B and with contributions from all partners.
In the beginning of the project (Month 6), the Consortium agreed on and implemented a plan for dissemination (see D6.02) carrying out numerous communication and dissemination activities to make the project results known. Firstly, the partners published a total of 177 peer-reviewed articles in 72 different journals (see section 1.4.2.1.). In addition, the partners engaged in more than 350 dissemination activities (see section 1.4.2.2.) including oral presentations (>200) and posters (>100) at major scientific conferences. EPITARGET also organized four Young Researchers’ Symposia, as well as co-hosting the epiXchange conference (May 2018, Brussels). Communication of publications, events and project news was achieved via the project website, Facebook and in a regular newsletter, following the plan for dissemination.
With regards to exploitation, EPITARGET has been successful in generating a large range of exploitable results in different categories. Firstly, several new drug combinations with antiepileptogenic/disease-modifying potential have been tested during the project with some promising result. An example for such a promising combination is the combination of an NMDA receptor antagonist with an AMPA receptor antagonist, or the administration of combinations of levetiracetam with topiramate. Secondly, EPITARGET has established a number of unique biosample banks and databases that will continue to drive research and translation, including tissue banks and databases about experiment types performed in the consortium. Third, EPITARGET has successfully developed preclinical data on a number of promising prognostic biomarkers, or novel technologies to enable biomarker measurements. Examples for this achievements are the use of the biomarker HMGB1 for predicting the development of posttraumatic epilepsy in a rat model of neurotrauma. Fourth, EPITARGET has also been active in developing technologies aimed at measuring electrophysiological activity biomarkers. For example, EPITARGET PIs have developed novel in-vivo techniques, primarily organic electrode systems to detect activity biomarkers of epileptogenesis with unprecedented accuracy.
Most of these results require further research to reach all the way to the “end-user” (patients or physicians). Therefore, the EPITARGET consortium has discussed and devised different ways of collaborating and obtaining funding for future research, including collaborations under EBRA, under ERANET and with NIH.
List of Websites:
https://www.epitarget.eu/