Recent progress in translational research on neurovascular and neurodegenerative disorders

2.1 Noncoding RNAs and neuroprotection

In mammals, the protein-coding ribonucleic acids (RNAs) constitute <2% of the transcriptional output, while noncoding RNAs (ncRNAs) of different classes comprise the remaining 98%. Recent studies showed that experimental stroke alters the expression profiles of various classes of ncRNAs including microRNAs (miRNAs) and long noncoding RNAs (lncRNAs). Hence, the therapeutic potential of ncRNAs in promoting neuroprotection after stroke has been discussed at the meeting.

Rona Giffard (Palo Alto, USA) showed that miRNAs can be manipulated to influence post-stroke outcome. She demonstrated that reducing the levels of miR-181 with either pretreatment or post-treatment with an antagomiR-181 effectively decreased the infarction in mouse models of stroke. Particularly, post-treatment leads to long lasting behavioral improvement (Xu et al., 2015). The miR-181 antagomiR was observed to target molecules that promote apoptosis, oxidative stress and inflammation. Dr. Giffard further showed that manipulation of miR-210 improves mitochondrial function which promotes survival of newborn neurons as well as reduces inflammation. This dual targeting shows promise to improve post stroke recovery (Ouyang et al., 2015).

Raghu Vemuganti (Madison, USA) showed that, miR-29 and miR-7 are major miRNA families that are considered as pro-survival microRNAs and play essential roles in the healthy brain. Following focal ischemia, miR-29c and miR-7a are downregulated significantly and when these miRNAs were replenished using specific premiRs and/or miRNA mimics, post-ischemic brain damage was curtailed with improved functional recovery indicating their essential neuroprotective role (Pandi et al., 2013; Kim et al., 2016). It was further shown that a lncRNA called FosDT induced after stroke mediates ischemic brain damage and treatment with a siRNA cocktail that targets FosDT significantly protected the brain following focal ischemia. Mechanistically, FosDT siRNA-mediated neuroprotection was associated with curtailed scaffolding of the chromatin modifying proteins Sin3A and coREST to the transcription factor REST that is known to promote neuronal death after stroke (Mehta et al., 2015).

Farida Sohrabji (College Station, USA) showed that outcome following manipulation of miRNAs after stroke depends on both age and sex. A case in point is the miRNA Let7f, which negatively regulates the neuroprotective growth factor insulin growth factor (IGF)-1. In a study using adult and middle aged male and female rats, an antagomir to Let7f (anti-Let7f) was delivered intracerebroventricularly 4 hours post stroke (Selvamani et al., 2012). In adult female rats, anti-Let7f significantly reduced infarct volume and improved sensory motor performance. However, this antagomir had no effect on stroke outcomes in males or ovariectomized females. Furthermore, in middle-aged, acyclic females, anti-Let7f, paradoxically, increased infarct volume. These data emphasize the complex interaction between the constitutive endocrine milieu and miRNA mediators, and underscore the importance of considering both sexes in developing stroke interventions.

Kejie Yin (Ann Arbor, USA) reported that the miR-15a/16-1 cluster is involved in the regulation of post-ischemic angiogenesis. Levels of miR-15a/16-1 are increased in the cerebrovasculature in mice at 7 days post stroke. Transgenic overexpression of miR-15a/16-1 leads to reduced cerebral blood vessel formation, increased brain infarction and neurological deficits in stroke mice (Yin et al., 2012). A series of in vitro experiments suggested that the miR-15a/16-1 cluster is a critical mediator in cerebral angiogenesis after ischemic stroke. Pharmacological inhibition may be helpful in developing miR-based neurorestorative drugs to promote post-stroke functional recovery.

2.2 Proteases before, during and after stroke

The plasminogen activation system was initially thought to be solely involved in the control of fibrin degradation by catalyzing the conversion of plasminogen into plasmin via two serine proteases: tissue-type plasminogen activator (tPA) and urokinase-type plasminogen activator (uPA). However, it has been recently discovered by Manuel Yepes (Atlanta, USA) and colleagues that the control of fibrinolysis is not the only role of tPA and uPA, and that instead both PAs are also found in the synapse where they play a central role in the development of synaptic plasticity via mechanisms that not always require their ability to generate plasmin (Wu et al., 2014, 2015). The group used 8–12 weeks old male wild-type (Wt) C57BL/6J mice for their studies.

Dendritic spines (DSs) are specialized dendritic protrusions that receive most of the excitatory input in the brain; and that are replaced by focal areas of swelling known as varicosities following the onset of global cerebral ischemia in adult 2–5 month old mice (Murphy et al., 2008). Data show that during the recovery phase from an ischemic stroke, neurons located in the peri-infarct cortex release uPA, and the interaction of this uPA with its receptor (uPAR) promotes the reorganization of the actin cytoskeleton and the subsequent re-emergence of dendritic spines from dendritic varicosities. This effect, mediated by uPA-uPAR, induced profilin expression and cofilin phosphorylation, leads to the formation of new synaptic contacts and recovery of neurological function after an ischemic injury. Accordingly, treatment with recombinant uPA (ruPA) promotes dendritic spine recovery and improvement in neurological function after ischemic stroke.

Denis Vivien (Caen, France) provided more insights into the multiple physiological and pathophysiological roles of tPA. The enzyme can act on virtually all cell types of the brain, including neurons, endothelial and glial cells. These actions are conducted by enzymatic or growth-factor-like effects on various molecular substrates or receptors. In addition of its fibrinolytic activity, tPA can also influence NMDA receptors signaling and subsequent neuronal fate. Interestingly, the single chain form of tPA (sc-tPA) is the only one capable to promote NMDA receptors-mediated neuronal death. Accordingly, strategies to prevent the tPA-dependent modulation of NMDA receptor signaling lead to an increased therapeutic window for tPA treatment in experimental stroke (Chevilley et al., 2015).

Josune Orbe (Pamplona, Spain) focused on matrix metalloproteinase (MMP)-10. His group investigated the therapeutic and protective effect of combining tPA/MMP-10 as clot-dissolving and neuroprotective agents after experimental model of ischemic stroke in C57BL/6J male mice (4 months old), as well as the molecular pathways involved. tPA/MMP-10 significantly reduced the infarct size in the ischemic stroke model compared with tPA alone. In vitro, MMP-10 reduced tPA-promoted endothelial ionic permeability, preserved claudin-5 expression and decreased ERK1/2 activation. MMP-10 also prevented tPA-mediated neuronal excitotoxicity and Ca²⁺ influx. These findings indicate that MMP-10 improves tPA-thrombolytic therapy by providing additional cytoprotection. All reported results were obtained from experimental models in rodents. Further investigations are needed to confirm the relevance of these data in clinical situations.

2.3 Post-translational modifications and regulation of gene expression in neurodegenerative disorders

It is well-known that neurodegenerative disorders, among those the most prominent, Alzheimer’s disease, are characterized by deposition of post-translationally modified amyloid peptides (Kummer & Heneka, 2014). Recent studies suggest that appearance of modified Aβ correlates with disease progression and cognitive status of patients (Morawski et al., 2014; Mandler et al., 2014). Likewise, clinical programs have been initiated to target those forms with novel therapeutics. Moreover, other factors such as activation of endogenous retroviruses might contribute to neurodegeneration (Emmer et al., 2014).

Stephan von Hörsten (Erlangen, Germany) reported about the putative role of N-terminal modifications of mutant Huntingtin (HTT). HD pathology was analyzed in HTT transgenic rats of different age. He pointed out that multiple processes such as limited proteolysis and pyroglutamate formation might act together to generate toxic HTT aggregates in Huntington’s Disease. Victoria Gröger (Halle, Germany) reviewed the potential role of human endogenous retroviruses for neurodegenerative disorders. Activation of envelope protein expression elicits a strong immune response and subsequent neuronal damage. Jochen Walter reported about phosphorylation of Aβ peptides in Alzheimers’s disease AD and its influence on the aggregation behavior of the amyloid peptide (Rezaei-Ghaleh et al., 2016). Steffen Rossner (Leipzig, Germany) demonstrated the important role of the chemokine CCL2 in Alzheimer’s Disease (AD) and the function of a glutaminyl cyclase (isoQC) for activation of the chemokine. Pathology was studied in Tg2576 mice, a mouse model of Alzheimer’s disease. Finally, Stephan Schilling (Halle, Germany) reported about endoproteolytic events leading to formation of Aβ3-42, which is converted into N-terminally modified pGlu-modified peptides in Alzheimer’s disease.

Summarizing, the reports highlighted that modifications of proteins represent attractive anchor points for future drug development.

3.1 Pressure, penumbra and promising acute stroke therapies

This symposium focused on strong hypothesis driven science and research collaborations with particular potential to translate research from bench to bedside. David Howells (Hobart, Australia) presented data regarding the work of his group using stem cell-derived human neurons as a tool for assessing candidate drugs and therapies (Howells & Macleod, 2013). As a prerequisite for translational attempts, these cells were reproducibly maintained in culture for weeks. Subsequent testing of various putative neuroprotectants in different models of in vitro ischemia produced results largely consistent with in vivo experimental or clinical trial experience, including lack of benefit for NXY-059, and a 4–6 hour window for benefit of hypothermic neuroprotection (Antonic et al., 2014).

Damian McLeod (Callaghan, Australia and Leipzig, Germany) presented in vivo data demonstrating a previously unknown and edema-independent mechanism of intracranial pressure (ICP) elevation 24 hours following stroke. Significant ICP elevation was found even after minor stroke with minimal subsequent edema. Interestingly, this increase was prevented by mild hypothermia in adult male Wistar and Sprague–Dawley rats (Murtha et al., 2014; Murtha et al., 2015). His data demonstrate that ICP elevation after stroke is a potential cause of the so called collateral failure in stroke-in-progression (Fig. 1). Neil Spratt explored potential mechanisms of this intracranial pressure rise based on data obtained from adult rats. Presence of ischemic penumbra in the early phases post stroke was associated with 24 hour ICP elevation, and there appears to be a triggering substance in the cerebrospinal fluid (Beard et al., 2016a, 2016b).

Renée Turner (Adelaide, Australia) presented data exploring the role of the neuropeptide substance P in development of cerebral edema and elevated ICP following large middle cerebral artery (MCA) stroke in adult sheep. A neurokinin 1 antagonist prevented ICP elevation in this model, without the marked midline shift seen in untreated animals or some of those treated with decompressive craniectomy (Gabrielian et al., 2013). This may represent an alternative, less invasive approach to treating the edema associated with large MCA stroke (Corrigan et al., 2016).

Mhairi Macrae (Glasgow, United Kingdom) presented data regarding a promising theranostic based on Glasgow-Oxygen-Level-Dependent (GOLD) technology, which provides MRI-based diagnostic techniques for detection of penumbral tissue alongside improved oxygen delivery to maintain penumbra viability until blood supply is restored. Initial studies using rodent stroke models indicated both diagnostic and treatment potential, and subsequent studies have included testing in healthy volunteers and stroke patients, suggesting potential of these agents to maintain metabolism and prevent lactate elevation in the penumbra as observed in rats (Murtha et al., 2015; Robertson et al., 2015).

3.2 Pathophysiology of vascular cognitive impairment

Chronic cerebrovascular pathology is a leading cause of dementia, and arguably represents the most important modifiable risk factor for Alzheimer’s disease (AD) (Iadecola, 2013). Despite this importance, the mechanisms and pathways contributing to vascular cognitive impairment (VCI) have largely remained elusive.

Gabor Petzold (Bonn, Germany) discussed the role of pro-inflammatory signaling cascades of the neurovascular unit in VCI. He provided evidence that an astrocyte-specific transgenic reduction of white matter inflammation in a mouse model of VCI can ameliorate axonal degeneration, demyelination and gliosis as well as cognitive decline. These data suggest that, similar to AD (Delekate et al., 2014), astrocytes might be centrally involved in the pathophysiology of VCI, and that anti-inflammatory interventions may prove therapeutically beneficial. Data were obtained in adult mice (10–14 weeks of age).

Linked to this work, Karen Horsburgh (Edinburgh, United Kingdom) discussed the contribution of microglial proliferation to axon-glial disruption and impairment of white matter function in a model of VCI. The VCI model is induced in male C57Bl/6J wild-type mice (approx. 4 months old) by permanent bilateral carotid artery stenosis resulting in chronic cerebral hypoperfusion, axon-glial damage and progressive impairment of white matter function over three months. She demonstrated that targeting neuroinflammation dampens microglial proliferation and improves white matter function in 4 month old wild-type mice. In TgAPP (J9) mice (4 months old before the onset of amyloid deposition) after one-month reduced cerebral perfusion, the levels of pro-inflammatory proteins are markedly elevated. Further in another TgAPP model (TgSWDI mice aged 4 months) reduced perfusion over three months culminates in accumulation of vascular amyloid and degenerative changes. This work provides evidence that reduced cerebral hypoperfusion is a common mechanism underlying the pathophysiology of VCI and AD and that neuroinflammation may be a key target for intervention.

Costantino Iadecola (New York, USA) discussed the role of vascular factors in conditions associated with cognitive impairment and dementia. Hypertension, a major cause of VCI, blocks neurovascular coupling and impairs endothelial and blood-brain barrier function through reactive oxygen species (ROS) generated by NADPH oxidase 2 (Nox2). In models of AD, there are major cerebrovascular alterations that precede amyloid deposition in the brain parenchyma (plaques) and blood vessels (amyloid angiopathy). The mechanisms involve binding of Aβ to the scavenger receptor CD36, activation of Nox2 and vascular oxidative stress as shown in a study using 3- to 4-mo-old male C57BL6/J or CD36^0/0, and Tg2576 mice (Park et al., 2011). Thus, deletion of CD36 rescues the cerebrovascular dysfunction in Tg2576 mice (aged 3, 18, and 22 months) and, remarkably, alleviates cerebral amyloid angiopathy without affecting amyloid plaques in the parenchyma (Park et al., 2013). Next, the lab investigated the mechanisms by which the ROS alter cerebrovascular function. It was found that Aβ induces DNA damage and activation of the DNA repair enzyme polyADP ribose polymerase (PARP) in cerebral endothelial cells, leading to formation of polyADP ribose, activation of the ion channel TRPM2 by ADP ribose, massive increases in endothelial Ca²⁺, and endothelial dysfunction (Park et al., 2014). The latter study was performed in male 3- to 4-month-old C57BL/6J mice, and mice lacking PARP-1 or Nox2, as well as in Tg2576 and TRPM2-null mice. Collectively, these observations indicate that vascular alterations are important both in vascular and neurodegenerative dementias and suggest that targeting vascular risk factors and improving cerebrovascular function offers the opportunity to mitigate their devastating impact on human health.

Tim Murphy (Vancouver, Canada) described an adult (2–4 months old) mouse model of small-vessel disease that is characterized by occlusions resulting from endovascular injection of fluorescent microspheres that target ∼12μm diameter penetrating arterioles, which can be localized by histology (Silasi et al., 2015) or in vivo imaging. Advantages of the model are its simplicity with respect to surgery and ability to recover animals for long-term studies. Using Thy1-green fluorescent protein (GFP) transgenic mice, the impact of micro-occlusions on neuronal structure was determined. Micro-occlusions in the hippocampus led to microinfarcts in ∼7% of lodged microspheres cases, while axons within white matter tracts as well as the striatum and thalamus became bleb-like or disrupted. Neocortical layer 5 neurons were relatively resistant to structural damage, with <2% of the lodged microspheres producing obvious neuronal damage by assessment of GFP-labeled dendrites. This small vessel disease model will be used for assessment of regional functional connectivity to understand how small sporadic lesions affect network dynamics. The model can be combined with new mouse home-cage imaging systems where image capture can be automated.

Tracy Farr (Nottingham, United Kingdom) has been using diffusion tensor imaging to explore structural changes in connectivity in the adult hypoperfused C57BL/6J mouse brain. Graph theory is a mathematical modeling technique that can advance our understanding of how the diseased brain changes by quantifying diffusion trajectories and converting them into a series of numbers. Hypoperfusion reduced the number of distinct groups of highly interconnected brain structures, though connectivity increased within these sub-groups. Some network parameters were excellent neuroimaging biomarkers, as they were accurate predictors of cognitive outcome.

3.3 The role of blood in brain injury

While the majority of strokes are of ischemic nature, mortality is comparably higher following hemorrhagic stroke, i.e. up to 50% within the first 30 days after the insult (van Asch et al., 2010). Bleeding can be caused by traumatic brain injury (TBI) or occur spontaneously in hemorrhagic strokes, while intracerebral (ICH) and subarachnoid hemorrhage (SAH) account for 15% and 5% of all strokes, respectively. The hematoma increases intracranial pressure and is associated with oxidative stress, inflammation, and cell death (Mracsko & Veltkamp, 2014; Pluta et al., 2009). However, the direct contribution of the vasculature, blood or its components to disease pathophysiology remains to be elucidated.

Recent studies suggest that early brain injury following SAH is caused by constriction of pial arterioles (Friedrich et al., 2012). Nikolaus Plesnila (Munich, Germany) demonstrated that nitric oxide inhalation improved vasospasm in both microvasculature and large cerebral vessels following SAH in mice (6–8 weeks old). In addition, it also reduced cerebral edema and neuronal cell death, increased survival and improved neurological function in the animals (Terpolilli et al., 2015).

Oliver Kempski (Mainz, Germany) showed that infusion of autologous blood in a rat model of acute subdural hematoma increased lesion growth, microglia and astrocyte activation and worsened behavioral deficits. Intracranial pressure and cerebral blood flow were comparable between both groups (Jussen et al., 2016).

Marietta Zille (New York, USA) systematically investigated cell death mechanisms in an in vitro model of hemorrhagic stroke in primary E15 CD1 mouse cortical neurons. Similar to ischemic stroke, hemorrhagic stroke appears to induce death via regulated necrosis, although preliminary data indicate that cell death is not via parthanatos but rather via necroptosis or ferroptosis, an iron-dependent form of cell death.

Breakdown of hemoglobin further leads to iron overload thought to mediate part of the observed neuronal toxicity following hemorrhage. Preclinical evidence suggests that the iron chelator deferoxamine mesylate (DFO) protects against hemorrhagic stroke in aged (18 months old) rats (Okauchi et al., 2009). Magdy Selim (Boston, USA) discussed the ongoing efforts to develop DFO as a treatment for ICH patients. A phase I clinical trial established that daily infusions of DFO after ICH are feasible and well tolerated without causing serious adverse events (Selim et al., 2011). However, a subsequent phase II clinical trial using high doses of DFO was stopped due to an increased incidence of acute respiratory distress syndrome (ClinicalTrials.gov NCT01662895), and a modified protocol (iDEF) is currently underway (ClinicalTrials.gov NCT02175225). Magdy Selim stressed that transitional research is difficult and imperfect. While initial supporting proof-of-concept animal data is important, ongoing bi-directional communications between basic researchers and clinicians are required at all stages. He also questioned the cost-effectiveness and reliability of phase-I, small, dose-finding, studies and spoke about the need for future evaluation of alternative approaches for drug delivery to minimize systemic toxicity.

The presented studies highlight that blood is a major contributor to cell death and outcome following hemorrhagic stroke. Future studies will further elucidate the underlying mechanisms and allow for development of novel therapeutic options for hemorrhagic stroke patients beyond the currently existing, more mechanistically driven approaches of hematoma evacuation, substitution of coagulation factors and blood pressure regulation.

4.1 The role of large animal models of CNS injury

Animal models are essential in understanding the complex pathophysiology of neurological diseases, with the selection of appropriate animal models being fundamental to clinical success. In recent years, there has been resurgence in the use of large animal models of CNS injury, for example stroke, traumatic brain injury and spinal cord injury. Such models provide an intermediate to validate interventions of outstanding promise identified in rodent models prior to clinical assessment, because such a rigorous step-wise validation is more likely to yield greater clinical success. Recent developments in animal models of CNS injury were presented in the symposium. Whilst the clinical translation of novel therapeutic agents for stroke has been extremely poor to date, Matthew Gounis highlighted that in contrast there has been great success in the translation of medical devices for mechanical thrombectomy. Indeed, combining in vitro (van der Marel et al., 2016; Chueh et al., 2012, 2013) with in vivo models (Gounis et al., 2013; Nogueira et al., 2012; van der Bom et al., 2011) was described as a way to provide evidence to direct device development and implementation.

Renée Turner (Adelaide, Australia) described a number of ovine CNS injury models including transient/permanent stroke (Wells et al., 2015), traumatic brain injury and concussion. The advantages of the sheep as a model for CNS injury were detailed. For example, sheep have a more human-like brain structure (Fig. 2) that includes a gyrencephalic pattern, clear anatomical discrimination between grey and white matter (including dense white matter tracts), strong fibrous dura mater and tentorium cerebelli, and large size compared with the rodent brain, in addition to a remarkably similar cerebrovascular anatomy (with the exception of the rete mirabile).

Anna Leonard (Adelaide, Australia) outlined a porcine model of spinal cord injury and demonstrated the advantages of this model over smaller models. Particularly, the anatomical size of the spinal cord and intrathecal space, in addition to the ability to perform clinically relevant procedures, including imaging and intrathecal pressure monitoring, improve the potential translation to the clinic (Lee et al., 2013; Jones et al., 2012).

In summary, the number and quality of large animal models of CNS injury has increased tremendously over the past few years and may represent one approach to overcome the translational roadblock and improve outcome for patients with CNS injury.

4.2 Current challenges and next steps for endovascular treatments

In 2015, mechanical thrombectomy has robustly shown beneficial effects in large randomized clinical trials, mainly by demonstrating an improved functional outcome but also reduced lesion size and decreased mortality (Campbell et al., 2015). Clinicians are recently challenged by the optimum kind of sedation, since general anesthesia was found to be associated with a worsened outcome (Dhakal et al., 2015), the fastest track to recanalization while considering intra-hospital proceedings (e.g. characterized by the ‘door-to-recanalization’ time), and the increasing need of capacities for post-interventional intensive care. However, on the fringes of the 9th ISN&N, neuroscientists and clinicians discussed about the potential impact of neuroprotective approaches that could be applied simultaneously with mechanical thrombectomy. Considering the lack of positive translations from hopeful pre-clinical neuroprotective strategies, the newly established kind of acute vessel re-opening via the intra-arterial route offers a potent chance to place neuroprotective drugs in close regional association to the initial, ischemia-affected tissue. In this direction, hopeful data originated from Michael E. Maniskas et al. (2016), who noted a significant decrease in lesion volume after intra-arterial application of verapamil subsequent to the recanalization of the middle cerebral artery in mice. In a more general perspective, new concepts for interacting within the complex pathways of neurotoxicity were discussed in the setting of acute vessel recanalization. Interesting data on this topic emerged from José Castillo et al. (2016), who reported on the difficile glutamate ratio between the blood and the brain, which could be modified by a group of pharmacological active substances working as ‘glutamate grabber’, potentially impeding effects emerging from a glutamate-related excitotoxicity.

These approaches might represent tasks for upcoming, translational-oriented studies, which, however, will require the close collaboration of basic neuroscientists, neurologists and neuroradiologists. Pre-clinical studies that started in well-established small animal models, exhibiting excellent standardization of environmental factors as well as lesion location and size, with confirming investigations in large animal studies might represent a hopeful way before strategies will be tested in the clinical setting.

5.1 Immunity in post-stroke neuronal damage and repair

Infiltration of inflammatory leukocytes is an established feature of human stroke. Currently, the activation of the immune system is recognized as a key mechanism in all stages of stroke pathophysiology, including neuronal plasticity and regeneration (Iadecola & Anrather, 2011).

Mathias Gelderblom (Hamburg, Germany) pointed out that innate-like lymphocytes contribute to the tissue damage via local production of IL-17, which leads to a subsequent detrimental infiltration of neutrophils (Gelderblom et al., 2012). All data were obtained in models using adult mice. Kate Lykke Lambertsen demonstrated neuroprotective effects of microglial-derived transmembrane-TNF. She reported that the selective blockade of soluble-TNF is associated with beneficial effects in a mouse model of experimental stroke (Madsen et al., 2015). Arthur Liesz (Munich, Germany) reviewed the role of post-stroke immunomodulation. He emphasized that cerebral ischemia results in a multiphasic dysbalance of the peripheral immune homeostasis. A strong over-activation of peripheral immunity in the acute phase is followed by a low-grade chronic inflammatory state in the chronic phase. Employing multiphoton microscopy, Jens Neumann (Magdeburg, Germany) reported data from rodent studies indicating that microglia are recruited to injured neurons, where they exert neuroprotective effects (Neumann et al., 2008). Finally, Halina Offner (Portland, USA) demonstrated the protective role of a sub-population of B cells, called regulatory B cells (Bregs). Data from her lab showed that IL-10⁺ B-cells elicited protective effects by decreasing the proinflammatory status of the microglia, and by inducing an anti-inflammatory M2 phenotype in microglia in eight-week old female C57BL/6 wild-type mice (Benedek et al., 2016).

Overall, the talks highlighted recent developments in post-stroke immunology. On one side, divergent roles of immune cell subpopulations in inflammatory collateral damage and neuroprotection after stroke were addressed. On the other side, the differing roles of brain-invading leukocytes in the chronic stages after brain injury for neuronal regeneration and repair mechanisms were emphasized.

5.2 The spleen as a novel target for stroke therapies

The spleen session focused on the last 10 years of basic and clinical research supporting a role for the spleen in the pathophysiology of ischemic stroke. Halina Offner initially discussed her seminal discoveries that the spleen contracts after stroke and releases inflammatory cells into the circulation (Pennypacker & Offner, 2015; Offner et al., 2006). These studies along with others in the field have stimulated research on the spleen, an important component of the immune response to stroke.

Dr. Offner (Portland, USA) discussed sex differences in spleen responses that have clear implications for the development of immunological based treatments in stroke patients (Banerjee et al., 2013). Consequently, the spleen became a potential target for new therapies. For example, Robert Mays (Cleveland, USA) discussed how a specific bone marrow-derived adherent stem cell line (MultiStem) may reduce brain injury and promote recovery through mechanisms involving the spleen. Dr. Mays presented the results of a double-blind, placebo-controlled clinical trial in which acute stroke patients were randomized to receive intravenous infusion of MultiStem or placebo within 24 to 48 hours after symptom onset (Hess et al., 2014). Sean Savitz (Houston, USA) then completed the session on clinical studies confirming that the spleen likely reduces in size in patients with acute stroke and that spleen contraction is associated with higher age, prior stroke, and African-American ethnicity. Data were also shown that serum inflammatory cytokines are significantly increased in patients with contracted spleens.

6.1 Roots and routes for stroke recovery

Stroke leads to large-scale dysfunction of neural networks beyond the local site of ischemic brain injury. Reorganization of affected networks can foster functional recovery, which may be furthered with neurostimulation techniques. Consequently, assessment of stroke-induced impairments of brain circuits, post-stroke remodeling of functional and structural connections, and effects of neurorestorative treatments could lead to significant advancement in stroke diagnostics, prognostics and therapeutics (Dijkhuizen et al., 2014) (Fig. 3). Latest insights in the characteristics, impact and mutability of post-stroke neural network alterations were presented and discussed in the session ‘Roots and routes for stroke recovery’.

Willem M. Otte (Utrecht, The Netherlands) explained how different network analysis methods inform on neural network topology from structural and functional magnetic resonance imaging (MRI) data. His studies in animal models and patients illustrated the difficulties of direct comparison between networks in control and pathological brains, which often differ in network density. Alternative strategies, such as minimum spanning tree and generative modeling (Sinke et al., 2016), can overcome these issues and allow more reliable characterization of topological network modifications after brain injury.

Adam Q. Bauer and Jin-Moo Lee (both St. Louis, USA) reported on changes in functional connectivity in the post-stroke mouse brain as measured with a relatively new functional imaging technique, based on optical intrinsic signals (White et al., 2011). Restoration of interhemispheric functional connectivity, which typically coincides with post-stroke sensorimotor recovery (Bauer et al., 2014), was impeded by optogenetic stimulation of the contralesional somatosensory cortex subacutely after photothrombotic stroke in the left somatosensory cortex. In a parallel study, whisker deprivation enhanced functional remapping of the forelimb representation into the whisker barrel cortex, which was accompanied by improved behavioral recovery. Their studies clearly demonstrated how post-stroke cortical remapping and functional recovery may be influenced – negatively or positively – by specific modulation of brain activity in peri- or contralesional areas.

Randolph J. Nudo (Kansas City, USA) showed the potential of the promising concept of closed-loop neuromodulation in a rat model. Microelectrodes were implanted in the premotor cortex, and microdevices that discriminate individual action potentials, were used to process discriminated spikes from multiple input channels, in order to trigger electrical stimulation of the somatosensory cortex (Guggenmos et al., 2013). Continuing spike-triggered cortical stimulation through this closed-loop system effectively improved sensorimotor function after focal injury to the primary motor cortex, emphasizing the ability of neural interface systems to restore functional connectivity and promote recovery after brain damage.

6.2 Neurorehabilitation, human brain plasticity and non-invasive neuromodulation

Yun-Hee Kim (Seoul, Republic of Korea) provided an expert overview on brain stimulation for neural network plasticity after stroke. Based on recent findings that motor training can increase functional recovery by enhancing neuroplasticity in the cortico-spinal tract (Lazaridou et al., 2013), she critically reviewed the clinical impact on invasive and non-invasive stimulation paradigms. For example, high frequency repetitive transcranial magnetic stimulation (Chang et al., 2010) and bilateral transcranial direct current stimulation (tDCS, Sunwoo et al., 2013) can induce functional improvements in human stroke survivors in the subacute and chronic phase of stroke. A possible explanation was given by Maria Rüger (Cologne, Germany), who showed that tDCS can enhance recruitment of endogenous neuroprogenitor cells in the adult male Wistar rat (Rueger et al., 2012), being accompanied by improved functional outcome (Fig. 4).

Complex information processing and sensory functions may also benefit from non-invasive stimulation. Dorothee Saur (Leipzig, Germany) presented recent data on language network reorganization, including underlying dynamics and region-specific roles for recovery. She also introduced a potential rational for increasing language network reorganization using non-invasive brain stimulation based on the results of functional MRI in the sense of aneuroimaging-guided network stimulation. In addition, she presented the results of the study by Meinzer et al. (2016), which demonstrated increasing reorganization in the language network even when stimulation was applied to an ancillary non-damaged network, e.g. the motor network. Bernhard Sabel (Magdeburg, Germany) is an expert in partial blindness and vision restauration (Bola et al., 2014). He presented promising data on noninvasive repetitive transorbital alternating current stimulation, which induced long lasting (6 months and longer) vision improvements in patients suffering from optic neuropathy in clinical trials (Gall et al., 2011, 2016). The trial revealed a more than 20% larger visual field size and 60% better vision in impaired visual field sector.

Ulf Ziemann (Tübingen, Germany) reviewed current concepts for individualized non-invasive brain stimulation for neurorehabilitation. He showed evidence that non-invasive brain stimulation might change activity and excitability of brain networks and highlighted potential supporting effects in stroke rehabilitation. On the other hand, effect sizes using current stimulation paradigms seem limited. Combining stimulation with supportive therapies such as pharmacotherapies might help to further increase effect sizes (Ziemann et al., 2015), but large scale clinical trials still need to be conducted.