Skip to main content

European Stroke Research Network

Final Report Summary - EUSTROKE (European Stroke Research Network)

Executive Summary:

A milestone in reinforcing synergies to build a strong base for scientific excellence in stroke research was the formation of the European Stroke Network (ESN) - a unique collaboration and merge of EUSTROKE with the EU-funded consortium ARISE (grant agreement n° 201024). The two consortia became a united force with a combined consortium agreement, shared meetings, joint preclinical and clinical trials, mutual dissemination activities, shared management, and a joint scientific advisory board. The powerful depth of multidisciplinary knowledge, the merging of resources and training activities, the development of mutual research platforms for brain imaging, regeneration after stroke, and in vitro and in vivo modeling served to provide an unsurpassed level of added value European health research.

Thus, the European Stroke Network was a collaborative effort that brought together researchers, government, industry, the non-profit sector, and patient group associations to tackle the devastating burden of stroke. This network put Europe at the forefront of stroke research through its multi-disciplinary research program, high-quality training for European scientists and clinicians, and national and global partnerships.

The ESN was dedicated to decreasing the physical, social, and economic consequences of stroke on the individual and on society. In pursuit of this goal, it integrated European research teams working on cerebrovascular biology, stroke prevention, imaging, and clinical trials to 1) promote excellence in European stroke research, 2) bring new approaches for stroke prevention and treatment to clinical practice, 3) train stroke researchers, 4) establish clinical trial networks, and 5) create European added value through partnerships.

The output of the ESN was remarkable and highly significant results evolved from the network, setting the pace for future stroke research endeavors. Indeed, some aspects of the work were in fact revolutionary and challenged current dogmas on stroke pathophysiology to enable totally new ways of approaching this disease. Great advances were made in understanding how to protect the brain against stroke through new concepts on how different processes are coordinated by various cells of the brain. Much was learned about how the brain protects itself against injuries. Moreover, significant progress was made on harnessing the power of resident neural stem cells in the brain for repairing stroke and bringing these together with other cells that promote formation of new blood vessels. Much effort was undertaken to elucidate how the blood-brain barrier breaks down in stroke, using this knowledge to develop new strategies for treating brain edema. Novel imaging strategies were developed for diagnosis and assessment of stroke recovery. Importantly, the results of the research are now culminating in first clinical stroke trials of the ESN Trial Network, which incorporates over 1350 stroke centers worldwide through our partner SITS. Thus, major contributions with an impact on reducing the burden of stroke have been made by ESN research.

A particular highlight of ESN activities was the formation of a joint transatlantic cooperation with the Canadian Stroke Network involving peer-reviewed partner proposals arising from members of the two networks. The high quality results of this cooperation set the stage for future joint international research projects with the European Union.

Project Context and Objectives:
In the development of new therapies for stroke, most efforts in the past have focused on the cellular and molecular events of neuronal susceptibility to injury, and on developing approaches to protect neurons from the consequences of ischemia and other injurious events. Researchers have strived to discover agents that inhibit cell pathways leading to cell injury or death after ischemic stroke. Indeed, a multitude of such agents were found to decrease neuronal injury in preclinical models, which led to further testing of these “neuroprotectant” agents in prospective clinical trials. Unfortunately, all neuroprotective agents failed to show any beneficial activity in patients with stroke.

EUSTROKE researchers hypothesized that the disappointing results of these neuroprotective therapies may reflect our insufficient understanding of the pathophysiology of ischemic injury in the brain. Ischemia initiates inflammation, increases microvascular permeability (which produces tissue edema), and causes local bleeding, in addition to having direct effects on cells. Ischemic stroke has such effects because stroke is a vascular disorder affecting neuronal function. Because neurons constitute less than 5 percent of the cells in cerebral grey matter, ischemia affects not only neurons but also astrocytes and other glial cells that support the neurons, the axons of neurons that relay their signals to other cells, and the microvessels that supply oxygen and nutrients to them. These observations provided a firm basis for shifting the previous focus from neurons alone to the complex of neurons, microvessels that supply them, and the supportive cells (astrocytes, other glial cells, and resident inflammatory cells). This “neurovascular unit” appeared to be a more realistic target of ischemic injury.

Thus, the primary focus of research in EUSTROKE was upon cerebral blood vessels and their relationships with other brain cells. This focus was justified by a growing body of evidence indicating that neurons, glia, and vascular cells are closely related developmentally, structurally and functionally. Indeed, the term “neurovascular unit” was seen to highlight the intimate functional relationships between these cells and their coordinated pattern of reaction to injury.

Our understanding of the molecular biology of the NVU was still very limited at the onset of the EUSTROKE project. This was due to the fact that research progress regarding the NVU had been hampered by the difficulty in overcoming interdisciplinary barriers of researchers and clinicians in the different fields of vascular biology, neurobiology, physiology, and cell biology. Thus, knowledge on the physiology of the NVU was still reduced to that of a binary-state system – functioning with a closed/controlled BBB or non-functioning with an open/interrupted BBB. However, we realized that this does not reflect the true functionality of the NVU, as barrier responses are more complex involving all cells of the NVU, and BBB breakdown develops over time and manifests with different severities. By interdisciplinary research on the molecular biology of the NVU, the idea was to overcome the logistic barriers that have hampered progress in the field and have prohibited the development of successful therapeutic regimens in stroke aimed at the level of the NVU.

Establishing in vitro models of the NVU

A first objective of EUSTROKE research was to establish novel in vitro models of the neurovascular unit for study of its role in cerebrovascular disease. These models were to be validated and then to be used to help identify the cellular and molecular mechanisms involved in maintenance of the NVU. Such knowledge was seen to provide a basis for understanding how ischemia leads to breakdown of the NVU. The novel models were likewise considered to be essential for developing new treatments for opening occluded brain arteries and for providing a framework to study the formation of new vessels after stroke. It was hoped that these models would likewise enable research on new approaches to prevention of stroke.

Understanding breakdown of the NVU in stroke

Breakdown of the NVU is generally described as a breakdown of the diffusion barrier measured by the diffusion of vascular tracers across the BBB. This simple concept, however, does not fully reflect the complex biology of alterations at the level of the NVU that result in impaired barrier function. Thus, to better characterize gradual or progressive changes leading to breakdown of the NVU we first needed to define the molecular alterations in the NVU in stroke. This objective was to be achieved by performing systematic studies on the modulation/changes of the BBB function in stroke, characterizing molecular changes of injured endothelium or other components of the NVU, including alterations in the underlying extracellular matrix proteins, or factors released from the injured endothelium and/or other cells of the NVU. Moreover, given the recognized intimate functional relationship among cells comprising the neurovascular unit, it was vital that we understand how the ion transporters/channels of these cells work in concert in both health and disease.

Tackling the problem of brain reperfusion injury

In the treatment of acute stroke, restoration of the blood supply can reduce the extent of brain tissue injury by salvaging a reversibly damaged area of tissue called the penumbra. Reperfusion, however, can have serious detrimental effects, including fatal edema or intracranial hemorrhage following thrombolysis. In animal stroke models, reperfusion after a long ischemic period can cause a larger infarct than that associated with permanent vessel occlusion. Thus, while reperfusion may reduce infarct size and improve clinical outcome in some patients, in others it may exacerbate the brain injury and produce a so-called “cerebral reperfusion injury.” At the onset of the EUSTROKE project, it was believed that leukocytes play an important role in the development of cerebral reperfusion injury. Previous studies had shown that during reperfusion, activated leukocytes interact with endothelial cells and plug capillaries, disrupt the BBB through the release of neutrophil-derived oxidants and proteolytic enzymes, extravasate from capillaries and infiltrate brain tissue, releasing cytokines that mediate inflammation. These processes were understood to produce an inflammatory cascade, resulting in the deterioration of the salvageable penumbra.

Thus, an important objective of EUSTROKE was to define leukocyte populations contributing to reperfusion injury following ischemic stroke. Moreover, by elucidating the steps involved in the extravasation processes across CNS vessels, it was thought that new targets could be discovered for effective therapy of reperfusion injury.

Adjunct therapies for thrombolysis – stabilizing the neurovascular unit

In the area of acute stroke therapy, project objectives included research on the interaction of thrombolytic therapies and breakdown of the neurovascular unit. Such knowledge could provide avenues for early so-called adjunct therapies (therapies given in combination) aimed at improving the safety of t-PA administration and at stabilizing or “priming” the neurovascular unit for thrombolytic treatment. Moreover, adjunct therapies could be developed to also support endogenous mechanisms for combating acute vessel occlusion. For example, elevated MMP-9 plasma levels had been reported to be correlated with an increased frequency of hemorrhagic transformation following stroke. From previously published results, it was known that dipyridamole significantly reduces MMP-9 release, particularly in its activated form. Thus, studies were to be undertaken to evaluate the effect of adjunct therapy with dipyridamole upon MMP-9 levels and consequent secondary hemorrhages with rt-PA.

Increasing evidence from in vitro and animal studies had indicated that application of ultrasound as an adjunct to thrombolytic therapy may offer a new approach to improve its effectiveness. First clinical trials in stroke patients by members of the consortium had shown that continuous transcranial ultrasound can augment standard t-PA lysis. Moreover, animal studies had demonstrated that intravenously injected microbubbles may dramatically enhance the effects of ultrasound and may even produce recanalization without fibrinolytic agents. Thus, an important objective of EUSTROKE was to obtain new insights into the molecular effects of ultrasound and microbubbles upon the neurovascular unit to enable optimization of this promising therapy in acute stroke patients. This included studies of the effect upon regulation of the expression and activity of active blood-to-brain transporters and brain-to-blood efflux transporters. Further objectives were to study global changes in the gene expression profile of human brain endothelial cells induced by t-PA and the modifier effect on those genes by adjunct therapies.

Stroke prevention through the neurovascular unit

Endothelial dysfunction is an early causal step in the development of atherosclerosis. Vascular risk factors lead to increased production of reactive oxygen species and decreased bioavailability of endothelial-derived nitric oxide that eventually lead to endothelial dysfunction, endothelial activation, and propagation of vascular disease. For a number of years it was recognized that endothelial dysfunction is an important predictor of cardiac events: For example, individuals with coronary endothelial dysfunction as determined by coronary angiography have a largely increased risk for cardiac events. Moreover, it was also known that coronary endothelial dysfunction also predicts increased risk for cerebrovascular events. However, it was not known what role cerebrovascular endothelial (dys)function or functional integrity of the neurovascular unit plays with regards to stroke risk and susceptibility to damage. Although endothelial progenitor cells (EPC) from the bone marrow can also contribute to vascular repair and improve endothelial function, it was unclear whether EPC levels correlate with cerebrovascular function and/or stroke risk. Also, the potential therapeutic impact for stroke had not been addressed. Objectives in the area of stroke prevention therefore included both experimental and clinical research (from in vitro models, ex vivo preparations, animal models of disease, to imaging and prospective studies in patients) to address the important question as to how cerebrovascular endothelial (dys)function and integrity of the neurovascular unit contributes to stroke pathogenesis. The hypothesis to test was whether endothelial dysfunction not only increases with stroke risk but also indicates worse outcome in those suffering from a stroke (susceptibility for damage). The study objectives included the characterization of the effects of dietary factors, physical activity, aging, progenitor cells, and lipid lowering on the neurovascular unit and the susceptibility for stroke.

New impulses in stroke imaging of the NVU

Recent developments in MRI research opened up unique opportunities for developing in vivo measurements of structural and functional changes at the level of the neurovascular unit. These included velocity-selective arterial spin labeling techniques, methods to quantify absolute cerebral blood flow, and microvessel density measurements for characterizing tumor angiogenesis. Thus, a primary objective of EUSTROKE was to apply these new techniques in stroke models to elucidate the pathophysiology of the NVU. Likewise, new tracer kinetic models were to be applied to allow a more precise analysis of BBB integrity. Further imaging goals were to apply new cellular and molecular MRI strategies to assess in vivo temporal and spatial characteristics of cell distribution (e.g. leukocyte infiltration or stem cell migration) and cell surface receptor expression (e.g. cell adhesion molecule upregulation) following stroke. Mapping of interactions within the neurovascular unit was to be performed with positron emission tomography.

Fostering tissue regeneration after stroke

Despite the potential therapeutic importance, little was known about the mechanisms that induce angiogenesis in the brain following ischemia. Thus, work was undertaken to understand these mechanisms and to use different strategies to improve brain neo-vascularization and perfusion after stroke. Specific objectives were to define the conditions that improve bone marrow-derived progenitor cell localization in the ischemic brain such as factors able to increase their mobilization from the bone marrow or the adhesion molecules and chemokines that increase their homing. A main goal was to improve neural regeneration by creating a favorable vascular niche for neural stem cells.

Establishing an ESN Trials Network

A European Trials Network would offer opportunities to faster and more cost-effective translation of pre-clinical research to clinical experience and facilitate investigator driven trials. Advantages include provision of cost effective monitoring procedures, procedures for submitting trial concepts, and clinical study protocols online, peer review systems, educational and certification modules adapted to GCP and individual trial settings, publication of study protocols on line and data ownership by investigators. An important objective was to adapt the SITS (Safe Implementation of Treatment for Stroke) network and database to support an ESN Trial Network. SITS is the world’s largest stroke treatment network with more than 1350 participating stroke clinics in over 30 countries with the purpose of broad implementation and quality development of evidence-based stroke treatments. Further objectives included clinical review of ongoing preclinical research within the consortium and counsel to basic researchers on matters regarding the clinical relevance and potential challenges or opportunities of translating the mechanism or compound under study into clinical practice.

Project Results:
Modeling the Neurovascular Unit

Scientific interest of neurobiologists has been focused on neurons and on their interactions with glial cells that support their function in the central nervous system (CNS). It is now becoming clear that brain microvessels (exchange surface of 20 m2 and 700 km long in humans) also contribute with neurons and glial cells to the formation of the well-structured neurovascular units (NVU) that are involved in the regulation of the permeability of the blood-brain barrier (BBB). Due to the complexity of the NVU, the investigations of the cellular and the molecular events than can modulate the functions of the BBB would be difficult, or even impossible, to carry out in vivo. Although it is true that in vitro modeling of the BBB is a simplification of the in vivo situation, cell-based BBB assays are powerful and flexible experimental tools to study transport and the dynamic functions of the BBB in normal and pathological states.

In the first phase of the project, current state of the art models of the blood-brain barrier were transferred to other members of the ESN consortium. Two partners in particular had a long-term expertise in establishing and validating in vitro BBB models, which recapitulate most of the unique properties of the brain endothelium. The co-culture model (bovine brain capillary-endothelial cells and rat glial cells) is considered the reference BBB model already used in a number of academic labs and pharmaceutical companies. The murine model has been established more recently on the same concept and is well characterized. Both models were transferred to the members of EUSTROKE and ARISE for in vitro investigations in the various work packages. Twenty-eight researchers from nine different countries visited the University of Artois to learn these models.

The human cell line hCMEC/D3 was developed and characterized. Although it suffers from inadequate barrier properties (low TEER and high sucrose permeability), this model is of human origin and can be used for basic research in signal transduction and receptor studies. These cells are now used in 300 laboratories including those involved in the consortium.

To understand the molecular, membrane, and cellular mechanisms involved in BBB maintenance, we have to work on models where it is possible to apply the techniques of cell biology and physiology at the single cell level. To this end, new in vitro models with additional features of the NVU were set-up. The addition of pericytes, and thereby the role of pericytes in controlling the integrity of the BBB, was investigated in a three culture system, in particular, its role in the induction of the barrier phenotype in the first day of the culture with endothelial cells derived from stem cells. The results showed that the addition of pericytes induced a full BBB phenotype by Wnt3a secretion. A full characterization of the extracellular matrix of the mouse model was carried out to investigate its role in the stabilization of the NVU after an ischemic insult.

Taking advantage of the transgenic mice, in vitro models of the BBB originating from KO mice for adhesion molecules (ICAM1, ICAM2, JAM, ALCAM, or IL-1βR) were set up and transferred to other work packages to determine the role of these molecules in leucocyte transmigration. In addition, we have established pMBMEC cultures from knock-in mice expressing a C-terminal GFP fusion protein of VE-cadherin allowing the visualization of endothelial cell junctions by means of live cell imaging in vitro. This model was successfully used to differentiate between the transcellular and paracellular diapedesis of leucocytes. Using the KO mice approach, the role of PPARα in the fenofibrate-mediated neuroprotection of the BBB was confirmed in in vivo studies. In addition, it was shown in more detail that the brain capillary endothelial cells are the target of this neuroprotective effect. Furthermore, co-culturing of smooth muscle cells and endothelial cells originating from brain arterioles and post-capillary venules were set up and transferred to other work packages in order to study the interaction with leucocytes during an ischemic insult.

The development of a human in vitro model of the BBB that incorporates intraluminal flow was achieved by using hCMEC/D3 cells. In comparison to other flow devices (DIV-BBB, Ibidi system), the “Endoflow” device allows a better differentiation of the cell line in terms of paracellular permeability. Characterization of the other BBB markers (transporters, efflux pumps) is underway. The results support the initial hypothesis and our first conclusion using the Ibidi system that laminar flow is a key parameter to consider for further optimization of in vitro models of the BBB.

A stable and reproducible model of the human BBB for large-scale production was also generated based on stem cells. The cells isolated from cord blood-derived hematopoietic stem cells (CD34+) were initially differentiated into ECs. These cells were then co-cultured with pericytes: only this type of cells of the NVU allows the induction of the BBB properties. These endothelial cells express tight junctions and transporters that are typically observed in brain endothelium, and maintain expression of the in vivo BBB properties for at last 20 days. The model is very reproducible and has already been transferred to several laboratories. Importantly, this model might be used for the screening of newly synthesized molecules to predict the pharmacological effect of a neuroactive drug in humans after systemic administration.

Breakdown of the neurovascular unit

Alterations of basement membranes and cell-matrix receptors within the NVU

Leukocyte extravasation across the BBB into the ischemic parenchyma was considered a key event in exacerbating stroke pathology. To define whether components of the extracellular matrix (ECM) modulate leukocyte-matrix interaction in stroke, mice lacking the endothelial cell laminin 4 (Lama4-/-) or laminin 5 (Lama5-/-) were investigated in an ischemic setting. Loss of laminin 5 appeared to be dispensable in stroke but lack of laminin 4 led to larger lesions and increased vessel permeability. Surprisingly, neither the total number nor the proportions of the different myeloid subsets were affected in the Lama4-/- background. Identical to wildtype mice, CD45+ leukocytes and Ly6G+ neutrophils remained associated with blood vessels and did not transmigrate into the ischemic hemisphere (Sorokin et al., manuscript in preparation).

Deletion of the heparan sulfate proteoglycan agrin from the parenchymal basement membrane (BM) destabilizes barrier integrity of the BBB. Under steady state conditions, agrin elucidates its function by maintaining junctional localization of endothelial adherens junction proteins in vitro and in vivo (Steiner et al. 2013, submitted). These results reflect the finely tuned interplay between components of the ECM and cell-matrix receptors, which is required for an appropriate interaction between peripheral leukocytes, and the BBB. The BBB is composed of a monolayer of specialized endothelial cells interconnected by complex tight junctions, an underlying endothelial basement membrane (BM) and a second parenchymal BM, marking the border to the CNS parenchyma. Together with the associated ensheathing layer of astrocyte endfeet, the parenchymal BM defines the glia limitans. All of the ECM components have the potential to bind to EC receptors and therefore influence the BBB integrity as demonstrated by genetic deletion of the endothelial laminin or the parenchymal agrin. Even though the lack of laminin 4 did not render the BBB permissive for neutrophil extravasation, there are no compensatory mechanisms in place to counteract this deficit as evidenced by exacerbated lesion pathology. In addition, lack of the heparan sulfate glykosaminoglycan agrin in the endothelial basement membrane lead to destabilization of endothelial adherens junctions in vivo. Taken together these observations demonstrate that endothelial cell-matrix interactions directly affect junctional integrity of brain endothelial cells and thus BBB permeability.

Role of pericytes in NVU breakdown and repair in stroke

Pericytes are embedded in the endothelial BM and they pose an integral part in sustaining BBB function. However, previous in vitro models of the NVU largely ignored pericytes but demonstrated a significant number of barrier characteristics such as tight junctions, low paracellular permeability, and efflux transporters in co-culture with astrocytes. Pericyte contact and non-contact models where shown to induce a regular endothelial monolayer of contact-inhibited cells, express tight junction proteins, and become impermeable to small hydrophilic molecules (Vandenhaute et al. 2011, Curr Neurovasc Res). Consequently, the new model was applied to investigate the putative role of pericytes in the porcine stress syndrome (PSS), a model of human malignant hyperthermia (MH). Carrier of a point mutation in the ryr1/hal gene, encoding the ryanodine receptor, display enhanced calcium release from intracellular stores. At the ultrastructural level, brain pericytes did not seem to be affected. However, pericytes isolated from ryr1-mutated pigs (Nn and nn genotype, bearing one or two mutant (n) ryr1/hal alleles, respectively) were not able to maintain low BBB permeability when compared to NN pericytes from wild-type animals and rather increased BBB permeability in vitro (Vandenhaute et al. 2012, Fluids Barriers CNS). The pericyte contact model accounts for EC-pericyte physical interaction and closely resembles the in vivo BBB whereas the non-contact in vitro BBB model might be suitable for analysis of the endothelia monolayer. The newly established models were central in revealing the inability of the mutated pericytes to maintain low BBB permeability and highlight their potential for investigating the contribution of pericytes in various vascular diseases.

Role of astrocytic aquaporin4 in causing vasogenic and cytotoxic edema during stroke

The polarized expression of the water channel protein aquaporin-4 (AQP4) on astroglial endfeet surrounding brain microvessels suggests a role in brain water balance. Loss of astrocyte foot process anchoring to the parenchymal BM accompanied by the loss of polarized localization of AQP4 to astrocyte endfeet has been shown to be associated with vasogenic/extracellular edema in neuroinflammation. In the context of the cytotoxic/intracellular edema, staining for the astrocyte endfoot anchorage protein -dystroglycan (DG) was dramatically reduced, and AQP4 and Kir4.1 showed a loss of polarized localization to astrocytic endfeet. Interestingly, mice deficient for agrin expression in the brain lack polarized localization of DG and AQP4 at astrocytic endfeet and do not develop early cytotoxic/intracellular edema following transient ischemia (Steiner et al., 2012, Glia). These data indicate that the binding of astrocyte DG to agrin embedded in the parenchymal BM promotes polarized localization of AQP4 to astrocyte endfeet. Reduced DG protein levels and redistribution of AQP4 as a consequence of stroke might therefore counteract early edema formation and reflect a beneficial mechanism operating in the brain to minimize damage upon ischemia.

Regulation of tight junction integrity in brain endothelial cells

The transmembrane protein claudin-5 is known as the key tight junction (TJ) protein at the BBB, although the molecular mechanisms by which it regulates TJ tightness are poorly understood. Mass spectrometry revealed the Guanine nucleotide-binding protein subunit αi2 (Gαi2) as binding partner of claudin-5 and that this association is correlated with TJ integrity in a human model of the BBB, the hCMEC/D3 cells. A selective expression of Gαi2 was observed in human brain vasculature cell-cell junctions in situ as well. Moreover, small interfering RNA-mediated depletion of Gαi2 or claudin-5 similarly increases their paracellular permeability and delays TJ recovery after mannitol treatment (Luissint et al. 2012, JCBFM). Altogether, the findings clearly demonstrate that Gαi2 is important for the permeability restriction to small molecules by the brain endothelial cells, which reflects the integrity of their TJ.

Adjunct therapies for thrombolysis

Sonothrombolysis with ultrasound and microbubbles

The only approved therapy for acute thromboembolic stroke remains thrombolysis with recombinant tissue plasminogen activator (rt-PA) within the first 4.5 hours after symptom onset. However, recanalization efficacy is far from optimal, as only 20-40 % of the patients treated with rt-PA achieve partial or full recanalization. On the other hand, rt-PA may provide additional benefits besides clot lysis. Indeed, it was recently shown that patients receiving rt-PA treatment had a better outcome even in cases of persistent MCA occlusion as compared to patients with persistent occlusion not receiving rt-PA. Our studies were aimed to elucidate possible mechanisms of this finding in a new model of thromboembolic stroke closely mimicking human pathophysiology.

The results demonstrate a clear correlation of the functional cerebral blood flow (CBF) data and protein expression of inflammatory mediators, apoptosis and stress genes after rt-PA treatment. Middle cerebral artery occlusion (MCAO) significantly increased the activation of IL-6, TNF-α, hsp70, and Caspase-3. However, this activation of protein expression did not occur in those animals that recanalized with rt-PA treatment as opposed to animals that showed no recanalization after rt-PA treatment. Quite intriguing is our observation that inhibition of the inflammatory factors, stress genes and apoptosis correlates with CBF. Furthermore, the infarct sizes do not correlate with the CBF data. As in the human situation there is a significant decrease in infarct size in all animals treated with rt-PA regardless of whether thrombolysis was successful or not. These data indicate that even if we cannot observe a recanalization after rt-PA treatment it still has an effect. We postulate that this effect is mediated through an improvement in microcirculatory flow. This is in agreement with other data demonstrating that rt-PA treatment both alone and in combination with contrast-enhanced ultrasound provides beneficial effects at the level of the microvasculature. Thus, our results show that rt-PA treatment decreases ischemic lesion volume, which indicates that the success of thrombolysis therapy is not limited to the recanalization of the arterial main stem occlusion. In addition, there is a clear correlation between the protein expression of inflammatory mediators, apoptosis, and stress genes with the recanalization data after rt-PA treatment.

These data provided an encouraging framework for further studies to compliment rt-PA therapy with adjunct therapies that help in opening the cerebral microcirculation after large vessel occlusion. MCA occlusion was performed in C57 black/6J mice as detailed previously. The MCAO mice were subjected to treatment with ultrasound and microbubbles 40 minutes following MCAO and monitored with laser Doppler. Optimized ultrasound parameters consisted of 10 pulses of 50 μs length with a delay of 1 ms between each pulse at a frequency of 1 MHz and 1 MPa pressure. There was a 2-sec pause between each 10-pulse sequence to allow microbubble replenishment. The pulse sequence was applied in 11 of 20 mice undergoing successful MCAO.

Unexpectedly, sonothrombolysis with ultrasound and microbubbles did not lead to MCA recanalization in any of the treated animals. Thus, as opposed to other models of arterial occlusion allowing a large percentage of successful recanalization with ultrasound and microbubbles, the clot arising from thrombin injection into the middle cerebral artery of C57 black/6J mice is highly resistant to lysis with ultrasound and microbubbles. On the other hand, we could observe a significant increase in blood flow adjacent to the infarction. Thus, this approach could be helpful as an adjunct therapy to rt-PA in humans even in cases where the thrombolytic effect of this combination is not substantial, depending upon the thrombus constitution.

Dipyridamole as possible adjunct therapy

Further work was undertaken to assess whether dipyridamole (DP) could be an effective adjunct therapy to rt-PA. In vitro studies with human brain endothelial cells (hCMEC/D3) showed that six hours treatment with r-tPA, at a dose that is translatable to the r-tPA plasma concentration found in acute ischemic stroke patients receiving thrombolytic therapy, was not toxic to cultured cerebral endothelial cells. Higher doses, however, induced a significant cell viability reduction. Oxygen-glucose deprivation (OGD) triggered a higher reduction and the combination of OGD and r-tPA treatment showed a summative reduction of cell viability in an r-tPA dose-dependent manner. We then tested the potential effect of dipyridamole as an adjuvant treatment with r-tPA. Our results showed that DP administrated as a pre-treatment, co-treatment or post-treatment did not protect hCMEC/D3 from cell viability reduction induced by r-tPA under OGD. Our in vivo studies using MCA occlusion in C57 black/6J mice likewise failed to show a positive effect of dipyridamole as an adjuvant treatment for r-tPA.

Effect of ultrasound and microbubbles on BBB efflux transporters

An interesting question arose concerning the effect of ultrasound and microbubbles on the neurovascular unit. It was postulated that some of the positive effects, especially the influence upon drug delivery across the BBB, might be caused through modulation of so-called efflux transporters of the BBB, i.e. transporters that remove drugs from the brain tissue once they have crossed the BBB. We studied mouse endothelial cells were grown in specially sealed OptiCell dishes, which can be immersed in water for treatment with ultrasound and microbubbles. Analyses were performed at various acoustic pressures at 2 h, 4 h, and 6 h following treatment. In the 2 h samples, we discovered a significant decrease of various efflux transporters treated with ultrasound as compared untreated samples. At 6 hours, there was then a significant increase in a dose-dependent (acoustic pressure) manner. To our knowledge, these are the first experiments performed showing that ultrasound and microbubbles can affect BBB efflux transporter functions. The results show that application of ultrasound with microbubbles significantly decreases the function of efflux transporters up to 4 hours after application. This knowledge could be used to selectively deliver drugs that are permeable to the BBB barrier, but quickly pumped out by efflux transporters. Further experiments are required to determine the duration of upregulation of efflux transporters after 6 hours, which could conceivably be useful for increasing efflux transport of undesirable compounds, e.g. β-amyloid from patients with Alzheimer’s disease.

Identifying gene expression that is altered by r-tPA under conditions simulating ischemic stroke

A microarray study was performed in order to carry out a genome-wide comparison of brain endothelial cells treated with r-tPA under OGD. A total of 2500 analyzed transcripts were regulated by 6 h OGD. Among them, 299 genes were differentially expressed only by OGD without r-tPA treatment, and 767 genes were regulated by OGD independently of the r-tPA dosage. However, r-tPA treatment was not sufficient to significantly alter the gene expression profile, either under OGD or control conditions. In order to specifically study the effect of r-tPA under OGD, genes which expression was affected by OGD without r-tPA were excluded from the analysis. Thus, two genes were found to be regulated by a subtoxic dose of r-tPA under OGD, whereas 15 genes were regulated by high dosages of r-tPA under OGD compared to control conditions. A global functional analysis of the genes altered by r-tPA was performed. Genes identified had functions related to cell movement, cardiovascular system development and function, cardiovascular disease and organism injury, demonstrating the complexity of the response to different dosages of r-tPA in an in vitro model of the human blood brain barrier.

From understanding reperfusion injury to challenging dogmas of stroke pathophysiology

When a blood clot that occludes a cerebral vessel is resolved with corresponding reconstitution of blood flow this can lead to what is known as 'reperfusion injury', i.e. vessel damage. In most tissues, such vessel damage is not only the result of the transient hypoxic conditions but also the subsequent influx of immune cells that interact with the endothelial cells lining the inner surface of the vessel and migrate into the underlying tissue, thereby contributing to both damage of the vessels and of the surrounding tissue.

Immune cell influx or inflammation that occurs in autoimmune diseases (e.g. multiple sclerosis) or is caused by infection is associated with upregulation and/or activation of defined molecular complexes on the endothelial cell surface, specifically molecules such as VCAM-1, ICAM-1 and 2, E-selectin, that bind to partner molecules or receptors on the immune cell surface. Interactions between these adhesion complexes mediate binding of the immune cells to the endothelial surface. In addition, molecular complexes that occur at the junctions between adjacent endothelial cells (JAMs, PECAM-1, CD99), together with an underlying extracellular matrix layer (the basement membrane) that is secreted basally by endothelial cells, act collectively to form a tight barrier between the lumen of the vessel and the surrounding tissue that also has to penetrated by the infiltrating immune cells. Cerebral vessels are particularly impermeable to the movement of both soluble molecules and to cells, including immune cells, constituting what is known as the blood-brain barrier (BBB) which results from specialized 'complex tight junctions' and a double basement membrane structure secreted by both endothelial cells and the subjacent astrocytes that articulate with the outer-most blood vessel border.

The molecular complexes involved in the initial tethering of immune cells to the endothelial cells and their subsequent firm adhesion and migration across the endothelial layer and underlying basement membrane have been well characterized in animal models of neuroinflammation. Based on this information, the question was addressed of whether the molecular cues controlling immune cell infiltration into the brain during neuroinflammation are also relevant to reperfusion injury after ischemic stroke and could be exploited for the development of new and more effective stroke treatments. This is particularly important given the lack of success of clinical trials of treatments for limiting brain damage after ischemic stroke. In particular, clinical trials targeting adhesion molecules on endothelium or immune cells after ischemic stroke have shown no benefit.

Early stroke research, which was based largely on biochemical analyses of excised post-stroke samples or involved the use of relatively non-specific immune cell markers (such as myeloperioxidase), suggested a role for so-called polymorphoneutrophil granulocytes (PMNs) which were considered to infiltrate into the brain tissue rapidly after stroke and to interact with neurons, resulting in neuronal death and hence being major contributors to brain damage. ESN work therefore focused on early stages after ischemic stroke and on new, highly specific immune cell markers that permit unequivocal identification of PMNs. In contrast to all other studies to date, the study focused on locating PMNs within the brain using high-resolution immunofluorescence microscopy together with markers of the blood vessel lumens and the border to the brain tissue (via specific basement membrane molecules).

Stroke samples collected from four different EUSTROKE groups and from two independent stroke laboratories outside of the EUSTROKE network revealed the absence of PMN presence within the brain tissue. Rather, PMNs were restricted to the lumens and walls of larger vessels. In collaboration with three independent Neuropathology departments in Germany (at the Universities of Tubingen, Freiburg and Frankfurt), these data were confirmed on a set of rare human samples collected from patients experiencing an ischemic stroke shortly before death. PMN localization in the experimental mouse tissues and the human samples did not correlate strictly with expression of endothelial cell adhesion molecules implicated in immune cell adhesion in neuroinflammation, nor with sites of blood clots (platelet accumulation) or increased vessel permeability. Further studies using a powerful in vitro model of the BBB established within the EUSTROKE network, permitted for the first time investigation of the effects of hypoxia and glucose deprivation on interactions between PMNs and cerebral endothelial cells. While hypoxia and glucose deprivation was shown to be sufficient to upregulate adhesion molecules on the endothelium, it was not sufficient to permit migration of the immune cells across the endothelial layer. The in vivo and in vitro data strongly suggest a new, as yet to be fully characterized mode of PMN interaction with the endothelial cell surface and opens new potential for the development of more effective therapies to minimize brain damage after ischemic stroke.

The data also strongly reinforce the concept of the neurovascular unit, i.e. the functional connection between the vessel lumen and the neurons within the brain: Although PMNs were located within vessel lumens, neuronal defects were apparent indicating that PMNs do not require physical contact with neurons to cause damage.

As a consequence of our unexpected results our publication of these data (Enzmann et al., Acta Neuropathol. 2013;125:395-412) has attracted considerable attention both from the scientific community and from the public at large, including commentaries on our article and several articles in newspapers in both Germany and Switzerland. The findings were highlighted at the 2013 meeting of the American Association for the Advancement of Science in Boston.

The data generated by this work has revolutionized the view of how immune cells impact on stroke severity and will lead to the rewriting of textbooks. It has opened new avenues for the development of therapies to minimize brain damage after ischemic stroke.

Preventing Stroke via the Neurovascular Unit

Effects of lipid lowering

A central observation was that inhibition of cholesterol synthesis with HMG-CoA reductase inhibitors (statins) reduces the size of ischemic strokes. The underlying mechanisms relate to improved bioavailability of nitric oxide (NO), which leads to improved endothelial cell function and protection of the neurovascular unit. These findings have led to several clinical trials and to the implementation in the current treatment guidelines. Importantly, results showed that statins given intravenously also protect from focal brain ischemia for up to 4 hours after an event. This finding opens the way to test this strategy as a novel treatment for patients with acute stroke.

A second (albeit probably less powerful) option to use lipid lowering for stroke protection was explored with regard to the inhibition of the proliferator-activated receptor-α (PPAR-α) employing fenofibrate. We observed that fenofibrate preventively protects the BBB endothelium in vitro by restricting pathologic hyperpermeability.

Effects of functional foods

Food supplementation with plant sterol and plant stanol esters is advocated to reduce serum cholesterol concentrations; however, the effects on vascular function are unknown. Extensive animal studies demonstrated that, compared to normal chow (NC), healthy mice fed with NC supplemented with 2% plant sterol ester show increased plant sterol but equal cholesterol plasma concentrations. Elevated levels of sterols and stanols were associated with endothelial dysfunction and increased depositions in the brain. In contrast to the expectation, sterol supplementation impaired endothelium-dependent vasorelaxation and increased cerebral lesion size after middle cerebral artery occlusion. To relate these data to humans, we examined plasma and aortic valve concentrations of plant sterol in 82 consecutive patients with aortic stenosis. Patients eating PSE-supplemented margarine (n=10) showed increased plasma concentrations and 5-fold higher sterol concentrations in aortic valve tissue. The project showed that food supplementation with plant sterol esters impairs endothelial function, aggravates ischemic brain injury, effects atherogenesis in mice, and leads to increased tissue sterol concentrations in humans. Therefore, prospective studies are warranted that evaluate not only effects on cholesterol reduction, but also on clinical endpoints.

Effects of exercise

EUSTROKE collaborators characterized the effects of a mouse model of voluntary physical exercise on endothelial function, cerebral blood flow, and cerebral ischemia. Physical activity upregulates endothelial nitric oxide synthase (eNOS), improves endothelium function, and protects from vascular disease. We therefore tested whether voluntary running would enhance neovascularization and long-term recovery following mild brain ischemia. Wild-type mice were exposed to 30 minutes of middle-cerebral artery occlusion (MCAo) and reperfusion. Continuous voluntary running on wheels conferred long-term upregulation of eNOS in the vasculature and of endothelial progenitor cells (EPCs) in the spleen and bone marrow (BM). This was associated with higher numbers of circulating EPCs in the blood and enhanced neovascularization. Moreover, engraftment of TIE2/LacZ-positive BM-derived cells was increased in the ischemic brain. Four weeks after the insult, trained animals showed higher numbers of newly generated cells in vascular sites, increased density of perfused microvessels, and sustained augmentation of cerebral blood flow within the ischemic striatum. Moreover, running conferred tissue sparing and improved functional outcome at four weeks. The protective effects of running on angiogenesis and outcome were completely abolished when animals were treated with a NOS inhibitor or the antiangiogenic compound endostatin after brain ischemia, and in animals lacking eNOS expression. Voluntary physical activity improves long-term stroke outcome by eNOS-dependent mechanisms related to improved angiogenesis and cerebral blood flow.

Telomere erosion is a central component of aging and telomere-associated proteins regulate cellular senescence and survival. Studies therefore examined the effects of exercising on vascular telomere biology and endothelial apoptosis in mice, and the effects of long-term endurance training on telomere biology in a clinical study. The data show that exercise up-regulated telomerase activity in the thoracic aorta and in circulating mononuclear cells compared to sedentary controls, increased vascular expression of telomere repeat-binding factor 2, and reduced the expression of vascular apoptosis regulators such as Chk2, p16, and p53. Mice preconditioned by voluntary running exhibited a marked reduction of lipopolysaccharide-induced aortic endothelial apoptosis. Transgenic mouse studies show that endothelial NO synthase (eNOS) and telomerase reverse transcriptase (TERT) synergize to confer endothelial stress-resistance following physical activity.

To test the significance of these data in humans, telomere biology in circulating leukocytes of young and middle-aged track and field athletes was analyzed. Peripheral blood leukocytes isolated from endurance athletes showed increased telomerase activity, expression of telomere-stabilizing proteins, and down-regulation of cell-cycle inhibitors compared to untrained individuals. Long-term endurance training was associated with reduced leukocyte telomere erosion compared untrained controls. In summary, physical activity regulates telomere-stabilizing proteins in mice and in humans and thereby protects from stress-induced vascular apoptosis. These data confirm the importance of exercise for the prevention of ischemic damage to the neurovascular unit.

Data of a prospective training study in humans suggest that a controlled high-intensity interval training for three months leads to an increase in telomerase activity and a reduction in the expression of senescence markers in circulating blood cells in healthy subjects without relevant training experience. In concert with previous studies, these findings set the stage to address the question which training modality, intensity, and duration are optimal to exert vascular “anti-aging effects,” and whether monocyte senescence proteins are useful markers of the preventive potency of a training intervention.

Effects of proliferator-activated receptor-γ (PPAR-γ) agonism

Searching for ways to pharmacologically modify cellular senescence, Work Package 5 found that the PPAR-γ agonist pioglitazone up-regulates telomerase activity and telomere-stabilizing proteins, and reduces senescence markers in the vascular wall and in circulating mononuclear cells. These effects and the reduction of LPS-induced endothelial apoptosis by thiazolidinediones depend on TERT. The data were confirmed in humans. These findings underscore the important role of telomere regulating proteins for vascular cell function and survival. Further studies revealed that the dual PPAR-α/γ agonist aleglitazar augments number, function, and survival of EPC, improves neoangiogenesis, restores endothelial function, and retards atherosclerosis. These results extend the findings of fenofibrate (see above) and suggest that activation PPAR-γ and PPAR-α may be beneficial. Clinically, however, potential side effects of this strategy have to be watched carefully.

Effects of diesel exhaust particles (pollution)

It was also demonstrated that intranasal instillation of diesel exhaust particles reduces the number and function of endothelial progenitor cells in C57Bl/6 and ApoE-/- mice. This finding was confirmed in cultured human EPCs. In ApoE -/- mice, the reduction in EPC was associated with a reduction in neoangiogenesis and an increase of atherosclerotic lesion formation.

New areas of research: mental stress and the NVU

Since some of the work on stroke prevention was achieved ahead of the anticipated time schedule, the ESN resources therefore allowed to address newly emerging important aspects of stroke prevention that were not detailed in the original grant application. An important area relates to the vascular effects of mental stress that are largely unknown but of great potential importance for the prevention of stroke. In addition, heart rate is emerging as a novel important vascular risk factor that can be specifically modified. The importance of heart rate for the effects of stress is not known. We therefore aimed to study effects of chronic stress and heart rate on endothelial function and cerebral ischemia. We demonstrated in a novel rodent model of stress that the increased heart rate could be prevented by the I(f)-channel-inhibitor ivabradine. Current studies utilizing this model are testing the effects of different treatment conditions on vascular reactive oxygen species, signal transduction, endothelial function, and the induction of ischemic stroke by transient middle cerebral artery occlusion.

Angiogenesis and Vascular Stabilization

When vessels are obstructed by a thrombus blood cannot reach the downstream regions of the brain. The nervous tissue, in absence of oxygen and nutrients, is damaged and may undergo necrosis. It is therefore important to be able to reconstitute an efficient new vascular network able to penetrate in the ischemic areas and restore perfusion. This is a difficult task, though, since newly growing vessels are usually fragile and when hemorrhage occurs, it may create profound damage to the nervous tissue. The work performed on the topic of angiogenesis and vascular stabilization had the general aim of improving blood perfusion of the ischemic regions of the brain by creating a new but also stable vascular network.

We first investigated, in a mouse model of stroke, the possibility that, by administering specific vascular growth factors, we could induce new vessel formation and improve ischemic brain perfusion in a stable and lasting way. We identified factors able to stimulate proliferation of the existing vessels within the ischemic brain. Most importantly, we could increase the stability of the new vessels by treatment with particular agents and by increasing the recruitment of specialized cells called pericytes. These cells embrace the new vessels and protect them from hemorrhages. These observations open novel pharmacological possibilities to induce the repair and recovery of the vasculature of damaged areas and eventually to rescue brain functions.

As a second task, we took advantage from the available knowledge on vascular stem cells that are known to circulate in blood and to be recruited in areas of ischemic damage. We defined the agents able to increase the number of vascular stem cells in the circulation and we also identified the adhesive molecules able to improve the localization of vascular progenitors in areas of damage. Further work included the isolation and culture of these vascular stem cells, thus permitting direct evaluation of their role in regeneration after stroke.

However, improving perfusion may not be enough to induce brain regeneration since differentiated nervous cells are usually stable and unable to repair the lesion efficiently. Neural stem cells are present in defined regions of the brain and, in appropriate conditions, can give origin to differentiated neural cells and eventually restore the damaged tissue. Neural stem cells, however, are very few in the brain parenchyma and are usually unable to effectively repair the damaged tissues. It is therefore important to increase their number and differentiation potential. To this end, we recreated an ideal niche for stem cells by growing them in contact with vascular cells as it occurs in brain. The cross talk between nervous and vascular cells and vice versa was found to be important to improve the efficiency of neural stem cells proliferation. Likewise, experiments with stem cell medium, i.e. the cocktail of substances secreted by stem cells, demonstrated that the medium has many of the same signaling effects as stem cells. “Cell-free” therapy with stem cell medium may offer a new approach for stimulating tissue regeneration after stroke.

Overall, these results contribute to the development of pharmacological and cellular therapeutic strategies to improve the rescue of brain damage after stroke.

Functional and molecular imaging of the neurovascular unit

The overall objective of work directed upon ‘Functional and molecular imaging of the neurovascular unit’ was to develop and apply new imaging methods to assess neurovascular changes in the brain after stroke. There was close interaction with the ‘Brain imaging platform’ in ARISE.

During the duration of the EUSTROKE project, MRI protocols were developed that enable quantification of critical vascular factors after stroke. These include blood flow, blood volume, and blood vessel density, which can be measured in a single MRI scan session using an iron oxide-based contrast agent that remains in the blood for prolonged time. With this approach, we have shown in a rat stroke model that blood vessels around a stroke lesion significantly remodel over time – which may include new formation of vessels (i.e. angiogenesis) – which was successfully validated by microscopic vessel counting in the extracted brains. Such vascular remodeling may be crucial for neurorepair mechanisms. As we have shown in parallel rat studies using functional and structural MRI methods, intact neuronal networks undergo significant modifications associated with functional recovery. These findings have demonstrated that reorganization of neuronal networks after stroke occurs at a whole-brain level, which may significantly depend on vascular remodeling to support active endogenous recovery mechanisms. Furthermore, we have shown that diffusion kurtosis imaging, an advanced structural MRI technique, can detect alterations in brain cell populations and neuronal tracts that may not be detectable with other MRI techniques.

A major goal of our studies was to develop and apply MRI methods to assess inflammation in the brain after stroke. Neuroinflammation may contribute to damage as well as recovery after stroke, but the exact role of different inflammatory processes remains largely unknown. To further study this we and our ARISE partners have established protocols to image inflammatory events in the living brain. We have optimized a MRI protocol that allows quantification of the permeability of the blood-brain barrier. The blood-brain barrier normally restricts the passage of compounds from the blood to the brain. However, after stroke the blood-brain barrier may be disrupted which can worsen damage to the brain by enhanced infiltration of inflammatory cells and increased risk of subsequent bleeding. By combining dynamic MRI scanning with a special infusion protocol that establishes stable concentration of a gadolinium-based MR contrast agent in the blood, we showed that changes in blood-brain barrier permeability can be calculated with high sensitivity. We applied this protocol in a number of studies, including an anti-inflammatory therapy study in collaboration with our ARISE partners. In this study, we tested the effect of dexamethasone-loaded long-circulating liposomes in a rat embolic stroke model, and found that blood-brain barrier permeability was reduced in treated animals as compared to controls. Moreover, we have incorporated our MRI-based blood-brain barrier permeability measurements in computer algorithms that predict outcome after stroke based on multiple acute MRI scans. We found that adding information on blood-brain barrier permeability improves the accuracy of estimating risk of bleeding after stroke. These algorithms can also be applied in the clinic, and may in the future contribute to improved diagnosis and treatment monitoring in patients with acute stroke.

Novel cellular and molecular MRI strategies were developed that are based on the use of specific nanoparticles, such as iron oxide particles with conjugated antibodies, to label cells or to bind to molecular targets. We have used these approaches to monitor the invasion of inflammatory cells into the brain, and to detect the expression of inflammatory molecules in the vessel wall. Our MRI data on cellular infiltration showed that the influx of inflammatory cells from the bloodstream is relatively limited after stroke, at least in rodent models, which was reproduced in the labs of other EUSTROKE and ARISE partners. On the other hand, with our ARISE colleagues we have shown in mice that the upregulation of inflammatory molecules in the brain vasculature can be non-invasively detected with MRI, which provides unique opportunities for in vivo molecular imaging of neuroinflammation.

We further studied in rat models mimicking clinical stroke firstly, whether better imaging agents to map cells suffering from lack of oxygen during stroke could be developed; secondly, whether administering pure oxygen during the blockade of the brain artery could reduce the brain damage caused by stroke; and thirdly, whether agents able to block inflammation of the brain tissue following stroke could reduce the death of brain cells (neurons).

We studied three novel PET probes and compared them to the currently widely used agent, which is a very good one but is limited by its slow accumulation in the affected brain tissue and as such is not optimal for diagnostic purposes since acute stroke requires urgent treatment. Unfortunately, none of the three novel imaging probes tested proved better or even equal to the reference, so this topic will need more research in the future.

An important research result was the observation that pure oxygen administered during blockade of the blood vessel did not significantly improve brain oxygenation or reduce brain damage if the blockade was prolonged or permanent, whereas it was able to completely prevent brain damage if the duration of the blockade was short (15 minutes in this rat model). This indicates that when treating patients with acute stroke administering pure oxygen may be helpful in delaying brain damage but that this benefit is materialized only if the blockade of the artery is quickly removed. This finding has direct potential clinical applications.

One particular agent was tested that was supposed to partially block inflammation within and around the area of brain damage following stroke, and hopefully in turn reduce brain damage given that inflammation may kill neurons. We found that this agent was able to reduce inflammation in the late stage (4 weeks after the stroke) but not in the earlier stage (at 2 weeks, as assessed using imaging in the living animal), and it only partly prevented neuron death. This suggests that using specific drugs it should be possible to at least in part control the brain inflammation that follows stroke in man, and that by doing so brain damage may be reduced. This finding may therefore have useful clinical applications. However, what would be needed at this point is drugs that more effectively block inflammation, perhaps at an earlier stage after the stroke, and result in improved neurological function and not just reduced brain damage. More work in this area is therefore required, but the outlook seems promising.

Establishing a ESN Trials Network

The Clinical Trial Platform (a joint WP of EUSTROKE and ARISE) provided a clinical trial network of centers associated with preclinical sites to allow for swift transition into clinical phase II trials, and assisted the coordination of the ongoing clinical study and trial activities of the partners, such as data entry and management and data sharing. It is coordinated by the Karolinska Institute and Heidelberg University. The Trial Platform also helped with the coordination of the design, fund raising, ethical and regulatory issues of phase II studies emerging from the ESN associated researchers and their collaborators (e.g. EUROHYP-1, PREDICT, WAKEUP).

The Trial Platform responded to several new opportunities for enhanced clinical stroke research in Europe. These included:

• Increased opportunities of translating preclinical research to clinical practice/trials. There are several European driven pre-clinical research projects that soon could enter the clinical stage. The Trial Platform offered opportunities for faster and more cost-effective translation of pre-clinical research to clinical experience regardless if initiated from academic institutions or from industry.

• Increased opportunities to run investigator-driven trials. The industry-independent Trial Platform facilitated investigator driven trials. Examples include, but are not limited to: provision of cost effective monitoring procedures in line with GCP, procedures for submitting trial concepts, and clinical study protocols online, peer review systems, educational and certification modules adapted to GCP and individual trial settings, publication of study protocols on line and data ownership by investigators.

• Increased demand to optimize conditions for collaboration with industry and SMEs. The Trial Platform was a strong partner to industry (SYGNIS, PAION, QUICKCOOL, SITS). By having such a Trial Platform, the scientific community took a more active role for study protocol development and transparency, data management and publication. The industry does not have access to clinically well-characterized biobanks, and it is therefore the task of the academic community to investigate this material and provide novel targets for drug development.

• Opportunity for enhanced knowledge sharing among stroke practitioners interested in participating in clinical trials. The Trial platform played an important role in promoting knowledge sharing on front-line stroke research and trial logistics, educational support and support in contacts with ethics committees.

The Trial Platform was based on the SITS Network (Safe Implementation of Treatments for Stroke). The SITS Network is the world’s largest stroke treatment network with more than 1350 participating stroke centers in over 30 countries. Participating clinics are primarily located in Europe but also in Asia, the Middle East, Australia, and South America. The objective of SITS is broad implementation and quality development of evidence-based stroke treatments.

The ESN adapted the SITS network and database to support the Trial Platform. SITS supported pre-clinical trial design and pre-clinical data acquisition to facilitate the transition from pre-clinical results into clinical application. We discussed the results of the preclinical testing of innovative therapies within the consortium and selected promising candidates for clinical trial (e.g. IL-1RA). The Trial Platform helped coordinate ongoing clinical trial activities of the partners, and assisted in coordinating the design and fund raising of II studies emerging from the research.

Potential Impact:
Stimulating and sustaining multidisciplinary basic biomedical research

A central goal of the FP7 Cooperation Work Program Health was to “stimulate and sustain multidisciplinary basic biomedical research.” This goal was clearly addressed in EUSTROKE, which boasted a unique multidisciplinary group of internationally leading and well recognized basic researchers able to address the diverse aspects of the molecular biology of the neurovascular unit and to apply meaningful measures of functional integrity of the NVU. This gathered expertise on the key cellular and acellular components of the NVU is unique worldwide, as exemplified by the expertise on endothelium and endothelial junctions, extracellular matrix constituents of the NVU and in vitro models for the NVU. This provided not only a fundamentally strong basis for the understanding of the functional biology of the NVU, but also the pathological consequences of alterations in components of the NVU that may occur as a consequence of stroke. Expertise in CNS inflammation and breakdown of the barrier function of the NVU further provided the knowledge for strategic analyses of the molecular steps leading to edema and reperfusion injury following stroke.

EUSTROKE linked and supported stroke researchers with other multidisciplinary networks to achieve maximal translation of research results to stroke-specific research priorities. The integration of multidisciplinary research in EUSTROKE enabled a strong interaction between technology and biology, which was maximally exploited through training and educational activities in the European Stroke Network. The network ensured that the economic benefits of European stroke research, partnerships, and training programs were maximized. By working closely with partners and through collaborations with the private sector, EUSTROKE fostered streamlined translation of new European discoveries to clinical stroke applications and facilitated the commercialization of research outcomes. These measures not only stimulate, but also sustain multidisciplinary basic biomedical research.

The European Stroke Network – Scale and Objectives

The large scale of the ESN is manifested in 23 leading multidisciplinary research groups, 7 SMEs, one large industrial partner, and the patient organization group Stroke Alliance for Europe, which unites large stroke patient organizations from 18 European countries. The efforts of the research program were facilitated by the integration of the SITS network consisting of more than 1350 stroke centers throughout the world. The collaborative project followed the initiative by the European Union to strengthen the European Research Area (ERA). The ESN thus reinforced existing synergies to build a strong base for scientific excellence in stroke research. It promoted the ERA not only by a strong research program, which was complemented by integrating and disseminating activities. As such, the project fulfilled functions that go beyond research, providing for the general advancement of knowledge for both science and society. Clearly, the scale and objectives of the European Stroke Network for promoting excellence in European stroke research, bringing new approaches for stroke prevention and treatment to clinical practice, training stroke researchers, establishing clinical trial networks, and creating European added value through partnerships were unique in the field of European stroke research.

Contributing to the development of alternatives to animal testing in biomedical research

Cell culture and in vitro approaches were used throughout our project to replace animal experiments. The consortium displayed a full battery of in vitro models of the neurovascular unit and strictly followed the 3R rule wherever it was possible by reducing, refining, and replacing animal experimentations. In fact, EUSTROKE had an entire work package (WP1 – Modeling the Neurovascular Unit) devoted to developing new in vitro models to answer a considerable number of research questions, thus avoiding the use of animal experiments.

Contribution to Societal Objectives

Through integration of the Stroke Alliance for Europe (SAFE) in EUSTROKE, the project provided a flexible framework for involving key stakeholders throughout Europe for review of the project’s research goals and its achievements. Members of the consortium attended and spoke at numerous meetings of SAFE. This offered opportunities for stakeholders to provide guidance in shaping the ultimate objectives of the program. It also facilitated an open exchange of information and points of view in the area of EU Health Policy and other EU initiatives of interest or concern to patients. This strategy promoted development of common positions on European health policy issues, giving patient organizations a central position in the provision of healthcare in Europe. The Stroke Alliance for Europe, which represents stroke patient organizations from 18 European countries, was a partner in the EUSTROKE consortium with well-defined tasks for supporting public science events, promoting stroke awareness, and assisting the consortium in durability issues.

Improving European Health Care

The ESN addressed the specific cross-border challenge of improving health care. It focused its activities to help solve significant health problems related to death and disability caused by cerebrovascular disease. The burden of stroke to the European Union, due to its size and complexity, is appropriately addressed at the European level. Realization of the goals of the ESN has made a significant impact upon reducing the devastating effects of this disease, a major socio-economic challenge facing the European Union. Community social objectives have been met in our proposal through promotion of health and of the quality of life. The ESN strived to help solve major societal problems related to the high level of morbidity arising from stroke. The further impact of the ESN will be the development of successful avenues to bring emerging therapeutic strategies to clinical trials. The ESN addressed key health problems associated with an ageing population and strived to establish successful strategies for reducing the impact of stroke through delivery of new adjunct and regenerative therapies. Moreover, ESN research placed a strong emphasis upon prevention of stroke with new strategies for clinical application.

Added value in performing the work at the European level

The critical mass of the European Stroke Network in terms of financial and human resources exceeded the means of a single country in the European Union. The resources of the ESN therefore reflect a considerable added value in performing the work at the European level. Moreover, the ESN co-operation was economically meaningful (scale economies) and offered positive effects in terms of stimulating private research. Importantly, the ESN combined complementary national skills, in particular in a highly interdisciplinary fashion to achieve the ambitious goals of developing new approaches to therapy and prevention of stroke via the neurovascular unit. The ESN addressed the specific cross-border challenge of improving health care. The burden of stroke to the European Union, due to its size and complexity, is appropriately addressed at the European level. In this way, the ESN contributed to the program strategy though European added value. The social objectives were met in the project through promotion of health and of the quality of life.

Better understanding of brain dysfunction

ESN research has made major contributions to our current understanding of stroke pathophysiology, which is the basis for the development of novel treatments. For example, through the interdisciplinary collaboration of vascular biologists, preclinical as well as clinical stroke researchers we were able to refute an old dogma regarding the invasion of the brain by granulocytes after stroke. This is highly relevant, as these findings may partially explain why certain anti-inflammatory strategies have failed in clinical testing (e.g. Enlimomab), but also point to alternative strategies to modulate the inflammatory process towards brain protection. Another example concerns the discovery of a master switch of cell death or cell survival (PEA15/HKII) which opens up novel therapeutic avenues for stroke, but also for tumor therapy. Pharmacological interventions in this signaling pathway are currently further explored with the pharmaceutical company SANOFI.

Translation into clinical applications including scientific validation of experimental results

Bench to beside translation was and still is the major bottleneck of stroke research. ESN research was internationally leading the efforts to improve translational success by improving the positive predictive value of preclinical stroke research. These included:

• Establishment and use of predictive stroke models, which included clinically relevant confounders such as comorbidities, gender, and age

• Exploration of clinically relevant issues (bedside to bench), such as infection before and after stroke

• Identification of biomarkers that predict complications of stroke and hence allow preventive treatment, first in animal models, then validated the approach in clinical studies (e.g. PREDICT study)

• Establishment of novel non-invasive imaging approaches to obtain 'fingerprints' of disease processes in experimental models which can now be used in clinical trials for therapeutic stratification, pathophysiological investigation, outcome monitoring, and as surrogate parameters

Increasing European competitiveness

Stroke research in the EU is highly competitive on the international level. ARISE and EUSTROKE were further contributing to this outstanding role of European research in this area. ARISE and EUSTROKE researchers were invited by the National Institutes of Health/National Institutes for Neurological Disorders and Stroke to participate in their quest to define US research priorities in the stroke field for the next decade (SPRG-process).

Competitiveness is the basis for collaboration: Through a collaboration with the Canadian Stroke Network, we were able to pool resources, generate mutual training opportunities and exchange of research expertise. Complex issues can be broken down and distributed in a coordinated fashion between partners. The vast experience of the European and Canadian Stroke Networks suggest that further benefits can be obtained from enhancement of their previously developed avenues of collaboration. We established a unique transatlantic pilot cooperation, the first of its kind in stroke research, to bundle expertise of some of the best researchers across the Atlantic. Six joint projects utilizing expertise from 22 research centers were chosen for this unique cooperation, the results of which could set the stage for further international efforts to overcome the translational roadblock.

Boosting the innovative capacity of European health-related industries and businesses

'High-tech' SMEs were important contributors to the ESN. The innovation and the integration of SOLVO, SITS, SYGNIS, NSGENE, QuickCool, and PAION ensured that new knowledge is disseminated and translated into new therapies and clinical practice. Development of new pharmaceuticals has a high attrition rate. Data collection tools were developed by SITS to fulfill the requirements of randomized controlled trials (RCTs), in particular novel data entry procedures were developed that are suitable for ambulatory use in pre-hospital settings and for acute phase intervention trials. These developments have now been implemented in international clinical trials (SITS Upper Treatment Window Monitoring, SITS Open Artery by Thrombectomy in Acute Occlusive Stroke, SITS Dabigatran Study). SOLVO’s participation in ESN research served to boost the presence of this SME in the biotechnology world, especially in the area of transporter interaction profiling of drugs commonly used in stroke. QuickCool was able to demonstrate the safety and efficacy of its novel, minimally invasive and portable techniques for early and safe whole-body cooling and preferential cooling of the brain. Clinical testing is under way. PAION is clinically developing the novel strategies to improve the safety of thrombolysis that were developed within the consortium. It is remarkable that in a field, which due to massive failures in clinical trials has been largely abandoned by the pharmaceutical industry, ARISE and EUSTROKE were able to involve pharmaceutical companies in joint projects which are currently being scaled up and rolled out towards clinical testing.

Main dissemination activities and exploitation of results

ESN results were disseminated in scientific articles, brochures for laypersons, conferences, workshops, books, TV appearances and public awareness campaigns throughout Europe.

A particular highlight of dissemination was the COMED film on the European Stroke Network, which was organized by the European Commission and filmed by BROADVIEW TV. This documentary film has achieved widespread viewing on Facebook and YouTube. The film is also featured on the website of the Stroke Alliance for Europe.

The results of the ESN research were made available to the scientific community at large via a large number of publications (> 300) in highly esteemed journals, including Nature family journals, PNAS, Brain, Lancet Neurology, J Neurosci., etc. Several of these articles were featured in editorials and press releases with subsequent coverage in news media (e.g. Enzmann et al. 2012). Many more articles describing ESN research are currently in preparation or in submission.

Members of the ESN consortium organized numerous scientific meetings and symposia or took part in such meetings as invited and keynote speakers (>250). Of special note, at the European Stroke Conferences (Nice, Stockholm, Barcelona, Lisbon and London), we organized dedicated and very well attended ESN symposia highlighting the results of our collaboration.

The ESN home page (www.europeanstrokenetwork.eu) served as an important communication platform. The website facilitated regular and programmed dissemination of data and promoted dialogue between investigators. The website has ‘open’ and ‘restricted’ areas (that are accessible only to consortium members). The website also includes regularly updated ‘consensus statements’ that represent the views of the consortium on study design and reporting. The website played an important role in targeting the general public.

An important dissemination activity was the training of researchers, key staff, including managers and industrial executives, even beyond the ESN consortium. ARISE and EUSTROKE organized and advertised these courses in such a way that they were open to all professionals interested even if they were not members of the consortia. External participation included neurologists and neuroscientists, students, but also technicians from academia and industry. A particular focus with regular courses held was on issues of modeling of stroke (techniques, limitations, quality), on brain imaging (in particular harmonization of protocols), brain inflammation, and behavioral assessment in models of stroke. Together with the publishing house Springer, ESN researches produced a widely read book on Modeling Stroke, which is a current standard source of information for basic stroke research laboratories world-wide ('Rodent models of stroke'; http://www.springer.com/biomed/neuroscience/book/978-1-60761-749-5). We also produced instructional videos available via the internet (e.g. J. Vis. Exp. 2011 Jan 6 ;(47), making expert know how of the consortium available to the research community at large.

Numerous activities targeted increasing stroke awareness in the general public: In stroke treatment, 'time is brain', and the ESN was well positioned to participate in the propagation of this knowledge. In addition, since a number of stroke risk factors can be prevented or ameliorated (hypertension, smoking, diabetes, etc.), educating the public about these risk factors and strategies to decrease personal stroke risk are highly important. The ESN improved European stroke awareness by participating in national as well as international health campaigns, such as the 'World Brain Awareness Day', the 'World Stroke Day' etc. ESN scientists gave public lectures, appeared on TV or other news media, and organized local awareness campaigns (e.g. 'Berlin gegen den Schlaganfall', Berlin against stroke). The ESN published and distributed a brochure designed to explain to a lay audience the aims, objectives and accomplishments of the research network. This was put together by our partner Stroke Alliance for Europe. The brochure was recently translated into eight languages and distributed throughout Europe.

Outlook and future research

Towards international stroke research cooperations

The ESN has continued to concentrate upon the feasibility of transatlantic cooperations as a possibility for assuring ESN durability. To this extent, the ESN coordinators have formally recommended a stroke research cooperation to the NINDS. As an indication of the need to integrate European excellence in US science, two members of the ESN were invited to participate in the NINDS Stroke Progress Review Group, which sets the stage for future funding activities of this organization.

The ESN coordinators organized an international symposium on transatlantic/transpacific stroke research collaborations in Potsdam, Germany, which took place on May 5, 2012. Some of the world’s most esteemed researchers attended this meeting to form a consensus statement on such a research strategy. This statement was recently published in the journal Stroke: Dirnagl U, Hakim A, Macleod M, Fisher M, Howells D, Alan SM, Steinberg G, Planas A, Boltze J, Savitz S, Iadecola C, Meairs S: A concerted appeal for international cooperation in preclinical stroke research. Stroke 2013;44:1754-1760. Essential elements of this appeal included the following:

• Stroke research was seen to be an international effort with scientists already collaborating on many individual cerebrovascular research projects. However, scaling up collaborations to the level required to generate the resources and synergies needed requires a structured process. Ideally, such a process can be simultaneously executed by groups of researchers and clinicians (bottom up) and by funding agencies and scientific societies (top down) in a coordinated way.

o An example for a successful 'bottom up' interaction of international scientists was the development of thrombolysis for stroke in the 1980s -2000s, which in its course also led to an improvement in clinical stroke trial expertise.

o A 'top down' example of a highly successful structured international research collaboration in biomedicine is the deciphering of the human genome.

o We may also learn on how to organize and govern multinational research from ongoing international collaborations, such as in genetics. Triggered by the advent of rapid methodological advances in genotyping large numbers of individuals successful approaches for project selection, data deposition and distribution, collaborative analysis, publication and protection of intellectual property claims on a large, international scale were developed recently (e.g. GAIN Collaborative Research Group 2007), many of which can serve as templates for other fields.

o Extreme but instructive examples from outside medicine include research collaborations in physics, which are focused on experiments utilizing expensive equipment (particle accelerators and detectors, e.g. European Organization for Nuclear Research CERN: www.cern.org) or on research on matter under extreme conditions of temperature, pressure or density (e.g. Extreme Matter Institute EMMI: www.gsi.de/emmi. Apparently, the physics community has established a culture of collaboration that makes possible such highly successful large-scale initiatives and pervades their entire research operation, including publication practice.

• Many lessons can be learned from other examples of international scientific cooperation, in particular with respect to organization, scale up, and governance. The consensus statement stressed that a strong foundation for development of international stroke research collaborations has been set by the ESN and CSN that are currently demonstrating the benefits of sharing results, distributing tasks, and bundling expertise in stroke research. The experiences of the ESN and CSN have demonstrated that multidisciplinary expertise can provide high levels of complementarity. Pooling of resources, mutual training opportunities and exchange of research expertise have served to enhance and accelerate the process of translation. Complex issues can be broken down and distributed in a coordinated fashion between partners. The vast experience of these stroke networks suggest that further benefits could be reaped from enhancement of their previously developed avenues of collaboration.

A major breakthrough in international stroke research was a joint research activity between the European Stroke Network and the Canadian Stroke Network. Official proposal guidelines were announced in January 2012. The CSN promoted this collaboration with 1 million dollars allotted to Canadian researchers. The importance of adhering to specific objectives of the ESN within the realms of the current contracts with the European Commission was a prerequisite for the proposals. Thus through this highly focused approach, Canadian researchers complement ESN goals with additional funding from the Canadian Stroke Network. Six ESN-CSN projects received excellent scores from esteemed stroke reviewers from the United States to move forward in this unique effort. There has been extensive interaction between researchers of the two organizations. A CSN-ESN Research Update meeting recently took place in Montreal on October 18, 2013 within the realms of Vascular 2013. The first results of this cooperation were impressive and clearly demonstrated the added value of transatlantic cooperation.

The ESN further consolidated activities of FP7 collaborative stroke research projects. Members of EUROHYP, WAKE-UP, and TargetBrain attended ESN meetings with the purpose of strengthening interactions and to move as one force in the direction of international cooperations. There was a special meeting of these FP7 collaborative stroke projects at the European Stroke Conference in Lisbon, Portugal on May 25, 2012. One avenue for strengthening European Stroke Research may be a European Research Infrastructure Consortium.

In line with the growing international recognition of the European Stroke Network, the ESN was invited to present and discuss the activities of the collaborative project at the 2013 Annual Meeting of the American Association for the Advancement of Science (AAAS), one of the most widely recognized global-science gatherings with broad U.S. and international media coverage. The meeting in Boston highlighted the “unreasonable effectiveness” of the scientific enterprise in creating economic growth, solving societal problems, and satisfying the essential human drive to understand the world in which we live. The vast number of international attendees demonstrated the global nature of the event.

Reflecting ESN’s continuing efforts to extend the collaboration of stroke researchers beyond Europe, the session entitled “Stroke Research: New Concepts and Innovative Solutions” also featured speakers from the U.S. and Canada with strong ties to the consortium. The AAAS stroke session was moderated by the director of the EC’s Health Directorate. The deputy director of the National Institute of Neurological Disorders and Stroke (NINDS) participated in the session as discussant. The European Commission stressed the importance of the transatlantic research cooperation with the Canadian Stroke Network and characterized the symposium as an opportunity to exchange ideas and identify common scientific challenges that need to be solved to speed up the discovery of new treatments.

Improving the predictiveness of preclinical stroke research

The ESN has spearheaded the notion that the decision to move from animal experiments to patients should be based on robust, high quality data. We have proposed that preclinical translational stroke research should learn from the experience of clinical stroke research. The ESN was ideally placed for this approach, as preclinical and clinical researchers were working side by side. In our joint research we have established measures to improve the quality of clinical trials and hence the robustness of our results. This included strategies to minimize bias (randomization, blinding, allocation concealment), a priori Power analysis and other biostatistical advances, careful definition of the primary and secondary endpoints, data monitoring and auditing, internationalization and inclusion of many centers, external steering committees and safety monitoring, rigid publication standards, trial registries, among others. Not only have we applied these measures within our consortium, we have propagated them into the scientific community at conferences and through articles and editorials, as well as in the recent consensus statement on international stroke research cooperation. A future quest will be to continue a process by which preclinical researchers are enabled to understand the limitations of their approaches, to limit bias, and to improve the predictiveness of their research.

List of Websites:

Project website address: www.europeanstrokenetwork.eu

Project coordinator: Prof. Dr. Stephen Meairs
Department of Neurology
University Hospital Mannheim
University of Heidelberg
Tel.: +49 621 383 3550
Fax: +49 621 383 3807
E-mail: meairs@neuro.ma.uni-heidelberg.de

Scientific project manager: Dr. Laszlo Szabo
E-mail: laszlo.szabo@uni-heidelberg.de