Mechanisms in blood-brain barrier opening and metabolism-challenged cerebrovascular ischemia with emphasis on ischemic stroke

Abstract

Stroke is the leading cause of disability among adults as well as the 2nd leading cause of death globally. Ischemic stroke accounts for about 85% of strokes, and currently, tissue plasminogen activator (tPA), whose therapeutic window is limited to up to 4.5 hours for the appropriate population, is the only FDA approved drug in practice and medicine. After a stroke, a cascade of pathophysiological events results in the opening of the blood-brain barrier (BBB) through which further complications, disabilities, and mortality are likely to threaten the patient’s health. Strikingly, tPA administration in eligible patients might cause hemorrhagic transformation and sustained damage to BBB integrity. One must, therefore, delineate upon stroke onset which cellular and molecular factors mediate BBB permeability as well as what key roles BBB rupture plays in the pathophysiology of stroke. In this review article, given our past findings of mechanisms underlying BBB opening in stroke animal models, we elucidate cellular, subcellular, and molecular factors involved in BBB permeability after ischemic stroke. The contribution of each factor to stroke severity and outcome is further discussed. Determinant factors in BBB permeability and stroke include mitochondria, miRNAs, matrix metalloproteinases (MMPs), immune cells, cytokines, chemokines, and adhesion proteins. Once these factors are interrogated and their roles in the pathophysiology of stroke are determined, novel targets for drug discovery and development can be uncovered in addition to novel therapeutic avenues for human stroke management.

Introduction

Ischemic stroke, the second leading cause of death worldwide, is the result of permanent or transient occlusion in cerebral blood vessels (1, 2). The only FDA-approved drug for ischemic stroke is tissue plasminogen activator (tPA), and its therapeutic time window is limited to 4.5 hours (Emberson et al. 2014). TPA is a thrombolytic agent that degrades fibrin clots via the conversion of plasminogen to the active plasmin. The administration of tPA may prompt a ten-fold increase in risk of intracranial hemorrhage (ICH) (Whiteley et al. 2016). Because tPA administration leads to poor outcomes in ischemic stroke patients whose large arteries were occluded, mechanical thrombectomy due to faster reperfusion\recanalization has been exploited to manage occlusions of large vessels with a therapeutic time window of 24 hours (Elgendy et al. 2016). For select Large Vessel Occlusion (LVO) patients (those admitted into the Stroke Care Unit within therapeutic time window and those who are eligible to be treated with both methods), a combination of both approaches is recommended where the mechanical thrombectomy precedes the tPA administration (Kim et al. 2011).

Over 90% of stroke patients do not receive tPA due to the short time window. Moreover, there is an increased risk of intracerebral hemorrhage when tPA is applied out of this time window (Vishnu and Padma Srivastava 2019). The leaky BBB reduces success rates of outcomes with tPA treatment, and conversely, tPA administration can increase BBB permeability through elevation of matrix metalloproteinase 9 expression and phosphorylation of gap junction protein connexin 43 (Jiang et al. 2017). Not only is this driven by the tPA treatment whose receivers constitute a minority of patients compared to the rest of those missing the 4.5hour therapeutic time window, but it’s also due to the ischemic stroke itself. Some pathophysiological manifestations include BBB rupture, and this disruption is found to contribute to further complications. Such complications include the facilitation of neuroinflammatory factors and vasogenic edema in addition to the risk of hemorrhagic transformation which increases morbidity and mortality after stroke (Cheng et al. 2014; Khanna et al. 2014; Berger et al. 2001).

Previous studies emphasized MMP’s role in mediating BBB disruption via degradation of tight junctions and basal laminas, but owing to the observed biphasic behavior of the BBB breakdown upon ischemic stroke, it is clear that BBB opening is a complex process. Within 4–6 hours of the onset of stroke, there is reversible hyperpermeability in BBB, and after 2–3 days, irreversible BBB opening occurs (Krueger et al. 2013). Before tight junction degradation (4.5 hours from stroke onset), vesicular trafficking is observed and is associated with primary BBB permeability (Krueger et al. 2013).

Following a stroke, bioenergetics of neurons fail resulting in reduction of adenosine triphosphate (ATP), elevated glutamate efflux, oxidative stress induction, and lactic acidosis (Kulik et al. 2008). The coagulation cascade through alterations in shear stress and reactive oxygen species (ROS) production is activated, and platelet and endothelial cells are recruited to the occlusion location (Peerschke et al. 2010). Fibrin forms so that it traps platelets and leukocytes and endangers the BBB integrity (Eltzschig and Carmeliet 2011). Endothelial membrane disruption, triggered by the micro-architectural changes in their cytoskeleton, is the next stage followed by tight junction degradation (Krueger et al. 2015). The tight junction degradation results from induced matrix metalloproteinase (MMP) expression. MMP activity may perpetuate BBB permeability and the detachment of endothelial cells from the extracellular matrix (ECM) (Grossmann 2002). Further, increased vesicular trafficking after stroke is documented to assist in BBB rupture (Knowland et al. 2014; Nahirney et al. 2016).

Due to the complexity of BBB opening after stroke as well as its detrimental roles in outcome and severity of the stroke, a mechanistic-driven drug discovery that holistically focuses on targeting ongoing pathophysiological events on the BBB demands a consolidated map of whole pathological mediators through which the BBB disintegrates at subcellular and molecular levels after stroke. Unraveling these factors can pave the way for novel therapeutic approaches that limit brain damage from stroke and physical disability emergence following stroke onset.

The Blood-Brain Barrier

The BBB, which regulates exchanges between blood and CNS, was initially observed by Goldman in 1909 (Goldmann 1909). During his pharmacological experiments, he failed to stain brain cells via intravenous administration of trypan blue dye. He concluded that a barrier should exist between blood and the CNS. Half of century later, after a series of controversial debates among physiologists about the existence and features of BBB permeability (Hawkins and Davis 2005), anatomical investigations that were empowered by electron microscopy shed more light on the impermeable BBB and its role in active and passive substance transport (Zlokovic 2008).

Subsequently, it has been demonstrated that a major component of the BBB is endothelial cells, which display tight intercellular junctions (TJ) with no fenestration, a quality not visualized in other non-brain endothelial cells(Brightman and Kadota 1993). They can also be distinguished from other non-brain endothelial cells via their low level of non-specific transcytosis (pinocytosis) and high number of mitochondria (Brightman and Kadota 1993). The endothelial cells connect at junctions that are classified into two groups- tight junctions (TJ) and adherence junctions (AJ). Tight junctions are reduced paracellular transport and possess a ―zip-locked‖ structure made of integral membrane proteins that include occludins, claudins family proteins, and membrane-associated guanylate kinases. These proteins are linked to the actin cytoskeleton by the adaptor proteins ZO1, ZO2, and ZO3 (Hawkins and Davis 2005; Zlokovic 2008). The AJ structure is merged with TJ structure and contains PECAM-1, cadherins proteins, and the junctional adhesion molecules (JAMs) JAMA, JAMB, and JAMC (Zhao et al. 2015). In the BBB, a single layer of endothelium cells is surrounded by the basal lamina, while the abluminal surface is attached to pericytes. Additionally, astrocyte end-feet membranes cover the overall capillary (Hawkins and Davis 2005). These cells, with their extracellular matrixes, constitute neurovascular units, and they maintain the integrity and function of the cerebrovascular endothelial cells (Weiss et al. 2009). Brain endothelial cells are different from non-brain endothelial cells in several respects. During the vascularization and developmental process, astrocytes stimulate the formation of a different type of endothelial cell from other non-brain endothelial cells. This conclusion was supported by the experiment wherein injected neonatal astrocytes into the chamber of the eye rapidly led to vascularization of new vessels, displaying impermeability to Evan’s blue dye, similar to BBB (Janzer and Raff 1987). In vitro culture of endothelial cells with astrocytes is demonstrated to increase BBB formation (Neuhaus et al. 1991). Pericytes were found to produce contractile proteins and angiopoietin, adjust capillary blood flow, and induce endothelial expression of occludin (Hori et al. 2004; Bandopadhyay et al. 2001).

Around twenty percent of the body’s oxygen and cardiac output are delivered to the brain, despite the brain representing only 2 percent of total body mass (Kisler et al. 2017; Iadecola 2013). The critical and undisrupted delivery of oxygen and glucose to the brain is reflected in the extensive vascularization of the brain, with capillaries in the human brain being 400 miles in length (Begley and Brightman 2003). Neurons within the brain do energy-expensive computation by electrical and chemical signals and are dependent on potassium, sodium, and calcium ions fluxing across their plasma membranes (Rolfe and Brown 1997). Due to evolutionary pressure to maintain homeostasis in the microenvironment of neuronal synapses, the central nervous system (CNS) was separated from the circulating blood by three borders - the BBB, blood– CSF barrier, and the arachnoid barrier (Abbott 1992).

Transcellular transport across the BBB is generally categorized into five mechanisms that include the following: ion transport, receptor-mediated transport, carrier-mediated transport, active efflux transport, and caveolae-mediated transport. ATP-binding cassette (ABC) proteins, expressed on the surface of endothelial plasma that faces with blood, are ATP-driven efflux pumps responsible for active transportation out of the brain of xenobiotics and endogenous metabolites. For carrier-mediated transportation, more than 300 genes encode membrane-bound proteins that assist the import of various molecules such as vitamins, carbohydrates, amino acids, hormones, fatty acids, and nucleotides. Receptor-mediated transport facilitates the transportation of peptides and proteins that possess receptors like growth hormones, transferrin, insulin, and amyloid-beta (Aβ) (Zhao et al. 2015).

The integrity of the BBB is essential for health, and increased BBB permeability is linked to the pathophysiology of many neurological disorders including stroke, Alzheimer’s diseases (AD), multiple sclerosis (MS), and traumatic brain injury. Ischemic stroke, with the sudden occlusion of a cerebral blood vessel, initiates microvascular injury and induces BBB dysfunction (Derakhshankhah et al. 2020; Ortiz et al. 2014). BBB dysfunction is well documented in stroke patients and experimental stroke animal models. In patients who received tPA treatments, BBB disruption has been visualized through magnetic resonance imaging (MRI) 2 h post tPA treatment and has been associated with brain edema evolution (Hjort et al. 2008). Girad et al. reported that BBB disruption occurs during the first 3 h after stroke onset and argued that delayed BBB alterations are associated with stroke severity and vasogenic edema (Giraud et al. 2015). Recently, we have demonstrated that BBB opens biphasically following in the transient MCAO in mice. We observed a significant increase of Evan’s blue and fluorescein in the ipsilateral hemisphere at 6 and 72 h of reperfusion following tMCAO; however, neither Evan’s blue nor fluorescein was observed at 24 or 48 h reperfusion (Fig. 1) (Hone et al. 2018).

Figure 1. Dynamic BBB opening following tMCAO in mice.

(A) MCA was occluded for 1 h and Evan’s blue was injected into the mice (i.v.). Transcardiac perfusion was then performed and brain images acquired as shown by representative coronal brain sections. Evan’s blue extravasation is denoted by blue arrows. Quantification of Evan’s blue extravasation in the left and right hemispheres. Data are expressed as mean ± S.D.; n=4/group; One-way ANOVA followed by post-hoc Tukey’s test, ***p<0.001; ****p <0.0001. (B) TTC–staining was performed on brain sections from the Evan’s blue extravasation assay using an additional cohort of mice (n=4/group). Evan’s blue extravasation is denoted by blue arrows. (C) Fluorescent dyes (Texas Red and rhodamine-123) were injected into tMCAO mice at different time points. Transcardiac perfusion was performed and whole brains were sectioned (20 μm). Fluorescence images were acquired as shown by representative coronal brain sections followed by staining with H&E or cresyl violet. Red arrows denote Texas Red infiltration. Yellow arrows denote rhodamine-123 extravasation. White arrows denote infarction by cresyl violet staining. Pink arrows denote pathological changes based on H&E staining (Hone et al. 2018).

BBB responses to ischemic events are a focus of attention to understand how to halt the progression of the BBB leakiness. Stroke-induced extravasation of circulating molecules into the CNS was found to occur through paracellular or transcellular pathways. This is initiated about 6 hours after stroke onset. Due to upregulation in the number of endocytotic caveolae forming from the apical surface and being directed the to the basal surface, the BBB become leaky during this phase, tight junctions undergo transient micro-architectural change and there is no evidence for complete degradation and sustained paracellular breach in this time. However, at about 48 hours after the onset of stroke, tight junction proteins undergo profound degradation and sustained damage, and the paracellular pathway become the main cause of BBB permeability (Knowland et al. 2014). Fig. 2 depicts these two pathophysiological breaching pathways.

Figure 2: Cascade of events in stroke pathophysiological opening of the BBB.

After stroke, due to bioenergetics failure in insulted area, the impaired mitochondria in cerebrovascular endothelial cells leads to BBB rupture. Concurrently, upregulated expression of MMPs, exacerbate the BBB integrity by degrading the tight junction proteins. Granulocytes with assistance of adhesion proteins, cytokine and chemokine adhere to endothelial cells and then, infiltrate into the CNS and initiate inflammation. Overexpression of some miRNAs, encapsulated in exosome silenced specific genes, mediating BBB permeability and finally modulating the stroke outcome. All these cascades of events manifest their impact on the BBB permeability spatiotemporally through paracellular or transcellular breaching pathways. The transcytosis precedes the tight junction degradation so that the transcytosis occurs at 6 hours after ischemic stroke while the tight junction degradation perpetuating the BBB occurs at 40–48 hours after stroke.

Given that the increased risk of hemorrhagic transformation (HT) restrains the broad utilization of tPA for acute treatment of ischemic stroke, questions are raised about the therapeutic time window criteria and what tools can enable us to predict which patients tend more to develop HT upon tPA treatment other than clinical CT imaging. The answer may be MRI imaging, with which perfusion-sensitive models can genuinely reflect post-stroke physiological damages in BBB. Moreover, given the MRI imaging accuracy, we have a precise discrimination of ischemic penumbra zone damage in the periphery from the core zone damage harboring already dead cells (Kassner et al. 2005; Scalzo et al. 2013). It is noteworthy to mention that employing the MRI in routine stroke care seems less feasible and accessible. In contrary to MRI, CT perfusion measurements can be accomplished with all modern CT scanners. In particular, based on two recent clinical trials’ outcomes where CT perfusion measurements were employed for risk prediction of HT, it was concluded that the CT perfusion assessment is a very practical tool for stroke physicians and scientists. Thus, CT perfusion assessment enables them to rapidly evaluate the risk of HT in regard to the tPA therapeutic time window extension up to one day. This, in turn, can benefit significantly more stroke patients (Wing and Markus 2019).

The mechanisms underlying stroke-induced BBB disruptions are dynamically interconnected. Interestingly, this complexity can offer combinatory therapies which may synergistically restore the BBB function after stroke. Cellular and molecular factors including mitochondrial bioenergetics, matrix metalloproteinases (MMPs), miRNAs, immune cells, cytokines, chemokines, and adhesion proteins, which are associated with stroke pathophysiology as well as BBB rupture, are depicted in Fig. 2. Each of these factors, in part, influences BBB permeability and exhibit significant roles in the pathophysiological interplay between BBB permeability and stroke.

Mitochondrial Bioenergetics

The brain mitochondria are the cell’s powerhouse and generate ATP to supply energy for several functions (Chandel 2017). ATP generation begins with the initial steps of glycolysis in the cytoplasm, during which pyruvate is produced and transported into the mitochondria. It is then oxidized to CO₂ yielding energy in the form of ATP (Nunnari and Suomalainen 2012).

Based on evolutionary theories, mitochondria were believed to be introduced as an alpha-proteobacterium into the cell through endosymbiosis, which means that following engulfment of the mitochondrion by the ancestor host cell, the mitochondria eventually linked to fundamental cellular compartments and interplays (Evans 2016). Mitochondrial dysfunction reduces longevity through several signaling pathways that regulate cellular senescence and age-dependent decreases in stem cell activity (Sun et al. 2016). Mitochondria play a vital role in the regulation of and responses to both intrinsic and extrinsic stress which can activate adaptive responses or execute cell death (Galluzzi et al. 2012; Oka et al. 2012).

Extremely metabolically active cells are more vulnerable to mitochondrial dysfunctions, especially when it comes to oxidative stress induced by ischemia/reperfusion injury or hypoxia (Fraser 2011; Ren and Simpkins 2015). Cerebrovascular endothelial cells were reported to possess enriched mitochondria due to intense energy-driven transportations of ions and substances between the blood and extracellular fluids (Oldendorf et al. 1977). BBB dysfunction might be associated with mitochondrial abnormalities (Pun et al. 2009).

An in vitro model of oxygen-glucose deprivation and reoxygenation (OGD-R) and ischemia/reperfusion (IR) injury showed that ROS formation and mitochondrial cytochrome c release results in caspase-3 mediated degradation of BBB tight junctions (Alluri et al. 2014). In an in vitro model, it has been shown that LPS-induced inhibition of mitochondria in endothelial cells compromised BBB integrity and worsened stroke outcomes (Doll et al. 2015). Mitochondrial carbonic anhydrase inhibitor (mCAi) reduced the production of bicarbonate as a ROS within mitochondria, as well as limited BBB damages (Salameh et al. 2016). In an in vitro study where oxygen-glucose deprivation was simulated in a co-culture of human brain microvascular endothelial cells and astrocytes, the role of Sirt1-Sirt3 axis in regulating BBB was investigated. Sirt-3, which mainly resides in mitochondria and previously was found to protect neurons against oxidative stress (Dai et al. 2014), exhibited positive effects in reducing BBB permeability against OGD by promoting mitochondrial function (Chen et al. 2018).

Mitochondria also play a crucial role in determining body temperature. Even though induced hypothermia in stroke patients is believed to correlate with better outcomes, it has been shown that if the hypothermia stems from brain endothelial cell mitochondrial impairment, outcomes are worsened (Hu et al. 2016). Moreover, exposure to FCCP 30 minutes prior to tMCAO in murine animals was shown to significantly increase infarct volume in the cortex, striatum, and total hemisphere compared to the control group (Grasmick et al., 2018).

In one of our previous studies, the bacterial infection mimic, LPS, was found to cause the mitochondria-dependent ischemic challenge on BBB permeability, associating with worsen stroke outcome. To validate the mechanism through which compromising the BBB permeability is governed by direct mediation of the mitochondrial respiratory, three pharmacological inhibitors of mitochondrial-respiratory complexes, including Rotenone, FCCP, and oligomycin were added to the cerebrovascular endothelial cells (cCVECs) cultures. In contrast, the degree of FITC-dextran 70 permeability mirrors the BBB disintegration following the small molecule inhibition of the mitochondrial respiratory chains I-II. To further strengthen our speculations and to see whether we are able to generalize this finding to the in vivo model, rotenone was adopted to investigate the BBB permeability as well as infarct size employing epidural application (EA) model and tMCAO model respectively. Taken together, both experiments highlight the fact that rotenone causes mitochondrial-dependent BBB disintegration and increases infarct size. Since the LPs were shown to inhibit mitochondrial function, it seems very reasonable to ascribe the LPS-induced observations in tMCAO model to the increased BBB permeability and infarct volume. (Doll et al. 2015). Moreover, exposure to FCCP 30 minutes prior to tMCAO in murine animals was shown to significantly increase infarct volume in the cortex, striatum, and total hemisphere compared to the control group (Grasmick et al. 2018).

A recent study suggested that astrocytes release functional extracellular mitochondrial particles and then transfer them into the neighbor neurons after stroke (Hayakawa et al. 2016). This finding, besides other pieces of evidence indicating mitochondria exchange as a mode of cell to cell communications between neurons, raises the possibility of mitochondria exchange between BBB component cells. Hence, one should investigate this important question. The therapeutic application of the mitochondria transfer is demonstrated by the in-vitro transfection of endothelial progenitor cells (EPCs)-derived extracellular mitochondria into the endothelial cells. Such a procedure protects BBB integrity as well as augments ATP level and angiogenesis (Hayakawa et al. 2018). The same promising results within the oxygen-glucose deprivation (OGD) stroke model was observed (Hayakawa et al. 2018). This platform, which translates mitochondrial repair in light of stem cell therapies, was investigated in MCAO model where co-localization of EPCs’ mitochondria in stroke rats’ neurons restored the BBB integrity and CNS hemostasis (Borlongan et al. 2019). Nevertheless, in order to translate the mitochondria-based regenerative medicine into the clinical settings, more reliable inquiries on the safety and efficacy of the treatment should be addressed.

Additionally, mitophagy as a catabolic mechanism whereby the dysfunctional or excessive mitochondria producing reactive oxygen species (ROS) are eliminated through specific autophagy, has been documented to contribute to stroke pathophysiology (Guan et al. 2018). However, no study has determined the role of mitophagy in stroke-induced BBB disruption. Future studies ought to shed more light on the statue of the mitochondrial biogenesis/degradation in the stroke-induced BBB permeability. As mitochondrial dysfunction associates with metabolic disorders accompanying/preceding stroke onset, the capitalization of the mitochondria in the stroke translational research needs more attention. Thereby, more investigations into the dynamic behavior of mitochondria, its metabolite fingerprinting, and turnover in the BBB constituent cells under the influence of stroke and metabolic disorders should be undertaken.

MicroRNAs

Exosomes, which perform intercellular communications among cells, are small membrane vesicles, with a diameter ranging from 30 to 100 nanometers. They are generated by cells and are released into biological fluids such as blood and cerebrospinal fluids. Besides, they encapsulate proteins and genetic materials like DNA, mRNA, and microRNAs (Lai and Breakefield 2012; György et al. 2015).

Strikingly, only 1.5% of DNA sequences in the human genome are found to encode for proteins, while approximately 98% of the genome is reported to be transcribed into diverse non-protein-coding RNAs (Taft et al. 2010; Bartel 2004). MicroRNAs (miRNAs) are single-stranded short non-coding RNAs (~20 nt) that can inhibit the translation of mRNA through complementary binding or enhanced degradation of mRNA (Beermann et al. 2016). A single miRNA can inhibit hundreds of mRNAs, and one mRNA can complement with hundreds of miRNAs (Afonso-Grunz and Muller 2015).

miRNAs have been documented to be involved in early stroke-induced stress (Bihl et al. 2018; Sørensen et al. 2014; Ren et al. 2018). Upregulation of miR-15a is seen with reduced oxygen-glucose deprivation in ischemic vascular injury, and its suppression ameliorates cerebrovascular endothelial cell death and keeps the BBB integrity intact (Yin et al. 2010a). Neuro-2a cells, in which miR-200b and miR-200c were transfected then exposed to the oxygen/glucose deprivation, survived cellular death (Bihl et al. 2018). The miR-497 antagomir exploitation in vivo model of middle cerebral artery occlusion (MCAO) reduced subsequent infarct volume (Yin et al. 2010b). Intravenous injection of miR-15a/16–1 antagomir lowered cerebral infarct volume and brain water content as well as improved neurobehavioral assessments [66]. These changes were associated with the upregulation of anti-apoptotic proteins, Bcl-2 and Bcl-w, with suppression of pro-inflammatory cytokines (Yang et al. 2017).

In mice with experimental autoimmune encephalomyelitis (EAE), miR-155 is overexpressed at the surface of the neurovascular unit upon cytokine stimuli. The miR-155 was demonstrated to mediate BBB permeability through targeting annexin-2, claudin-1, DOCK-1, and syntenin-1(Lopez-Ramirez et al. 2014). Moreover, it was shown that miR-155 knockout mice protect the BBB integrity amid MCAO through negative modulation of notch signaling pathway and neuronal apoptosis (Jiang et al. 2019). Another group has also demonstrated miR-155 knockout protects against brain damage by cerebral ischemia and decreased hemorrhagic burden (Suofu et al. 2020). The MCAO-challenged miR-132 knockout mice have been shown to have less stroke-induced BBB opening. The study justified this observation through reduced MMP-9 mRNA levels by miR-132 which could potentially rescue the tight junction proteins from degradation (Zuo et al. 2019).

The miR-181c, encapsulated in cancer-derived extracellular vehicles (EVs), was found to exert destructive effects on the BBB through degradation of phosphoinositide-dependent protein kinase-1 (PDPK-1) in a mouse model of brain metastasis. Since PDPK1 in cerebrovascular endothelial cells modulates the localization of N-cadherin and actin, silencing its gene via miR-181c compromises BBB integrity (Tominaga et al. 2015). In cystathionine-β-synthase deficient mice hyperhomocysteinemia, where the level of homocysteine exceeds normal state, BBB permeability is increased. miR-29b was shown to mediate BBB permeability through modulating DNMT3b and consequently regulating the expression of matrix metalloproteinases (Kalani et al. 2014).

BBB permeability and integrity were reported to be impacted by alterations in miRNA concentrations. For example, when brain endothelial cells were treated with homocysteine, the BBB was disrupted. The BBB leakiness was ascribed to a mechanism wherein increased expression levels of miR-29b upon treatment with homocysteine leads to DNA-methyltransferase 3 suppression. This event causes upregulation of matrix metallopeptidase 9 and eventually, BBB disruption (Kalani et al. 2014). In an animal model of MCAO, overexpression of miR-150 impaired BBB by suppressing brain endothelial tyrosine-protein kinase receptor TIE-2 expression as well as decreasing claudin-5 expression. A miR-150 antagonist for post-stroke treatment resulted in BBB protection, decreased infarct volume, and attenuated neurologic deficits (Fang et al. 2016). Another study showed perturbations in BBB permeability during HIV-1 infection is due to the upregulation of brain endothelial miR-101, which suppresses VE-cadherin protein expression (Mishra and Singh 2013). During traumatic brain injury, both VEGF and angiopoietin-1/TIE-2 expression is increased, a consequence of elevated levels of miR-21. These signaling pathways were suggested to protect BBB permeability by augmenting angiogenesis and cerebrovascular integrity (Ge et al. 2015; Ge et al. 2014).

In a study conducted by our lab, miR-34a knockout mice revealed less neuronal deficits and reduced infarct volume when subjected to tMCAO (Ren et al. 2019; Hu et al. 2020). Based on the fact that the miR-34a had been found to increase in brain ischemic hemispheres at 6 hour and 24 hour after stroke onset, we investigated the role of miR-34a in mediating in vitro phenotypic leakiness through compromising mitochondrial activity as well as tight junction staining (Bukeirat et al. 2016). We have also reported miR-34a KO mice show less intense BBB opening which was ascribed to the notion that the miR-34a targets mitochondrial cytochrome C (CYC) and impairs the oxidative phosphorylation energy metabolism in ipsilateral hemispheres tissues and also in the primary CECs (Hu et al. 2020).

Overexpression of miRNA-125a-5p is associated with increased levels of ZO-1 and VE-cadherin proteins in tight junctions (Almutairi et al. 2016). Another study reported that during neuroinflammation, miRNA-98 overexpression protected BBB integrity and functionality by decreasing leukocyte adhesions (Almutairi et al. 2016). miR-21 was documented to play a neuroprotective role against BBB breakdown during ischemia-reperfusion in rats induced by MCAO. The underlying molecular mechanism of this miR-21 effect appears to be inhibition of MAP2K3 (Yao et al. 2018). miR-130a, expressed by brain microvascular endothelial cells, was shown to increase BBB permeability (Wang et al. 2017). Administration of antagomir-130a in MCAO animal model reduced BBB rupture, lowered the infarct volume, and improved neurological function via inhibition of Homeobox A5 expression (Wang et al. 2017). Table.1 is a summary of the in vivo stroke model experiments in which miRNAs affected BBB opening.

Table 1:

miRNAs involved in BBB permeability in stroke (in vivo experiments)

miRNAs	BBB permeability	Animal type, sex and age.	Effects	Technical features	References
miR-21	↓	Specific pathogen-free (SPF) Sprague-Dawley (SD) rats, male	Beneficial	I/R injury, occlusion for 90 minutes, reperfusion for 24 hours	(Yao, Wang et al. 2018)
miR-150	↑	Sprague Dawley rats, male, 6–8 weeks old	Detrimental	Permanent pMCAO	(Fang et al., 2016)
miR-130a	↑	Sprague Dawley rats, male	Detrimental	tMCAO, ischemia for 90 minutes, reperfusion for 30 minutes	(Y. Wang et al., 2017).
miR-15a	↓	C57/B6 mice, male	Beneficial	tMCAO, ischemia for 90 minutes, reperfusion for 24 hours	(Yin, Deng, Hamblin, et al., 2010)
miR-34a	↑	Mice, male, 3~6 month-old	Detrimental	tMCAO, 1 h ischemia followed by reperfusion	(Hu et al., 2020).
miR-132	↓	Mice, male, 8 week-old	Beneficial	tMCAO, 1 hour ischemia pursued by reperfusion	(Zuo et al., 2019)
miR-155	↓	Mice, male, 8 month-old	Beneficial	tMCAO, 90 min ischemia followed by reperfusion	T. Jiang et al., 2019)

The capacity of miRNAs, which simultaneously affect several targets, are involved in the pathogenesis of ischemic stroke and, as such, are novel targets for ischemic stroke therapy. Since miRNAs reside in biological fluids, like blood and cerebrospinal fluid, they are more likely to represent the physiological states of the body. Therefore, one can claim that the miRNAs may be biomarkers for BBB disruption during a stroke.

Matrix Metalloproteinases (MMPs)

Matrix metalloproteinases, which are also known as matrixins, are a family of 23 zinc-dependent protein cleaving enzymes which are classified based on their substrate specificity or their domain structure (Klein and Bischoff 2011). MMPs demonstrated cleavage of the extracellular matrix proteins collagen (MMPs-1, −8, and −13), gelatin (MMPs-2 and −9), and matrilysin (MMPs-7 and −26) as well as degradation of other substrates such as contractile proteins and pro-forms of signaling molecules. They play significant roles not only in normal physiological processes like cell migration, angiogenesis and tissue morphogenesis, but they also play a role in pathophysiological states, such as neuroinflammation, neurodegeneration, and cancer metastasis (Nagase et al. 2006; Paiva and Granjeiro 2014; Sternlicht and Werb 2001; Lee and Pienaar 2014; Hannocks et al. 2017).

MMP-2 and MMP-9 have diverse substrates, and both play important roles in BBB permeability (Candelario-Jalil et al. 2011). MMP-2 is involved in the initial phase of BBB opening (maximum 3 hours) (Candelario-Jalil et al. 2011). MMP-9 is a key player in the delayed opening of the BBB after ischemic stroke (Adibhatla and Hatcher 2008). After a stroke, furin is produced by hypoxia-inducible factor-alpha (HIF-1α). Following furin production, MMP-14 becomes activated, and the activated MMP-14 eventually activates MMP-2, which usually presents in a latent form. Activation of MMP-2 leads to the degradation of the basal lamina and tight junction proteins (Yang and Rosenberg 2015).

In sepsis-induced Wistar rats, the activation of both MMP-2 and MMP-9 is a significant step in BBB disintegration (Dal-Pizzol et al. 2013). Also, MMP2 and MMP9 expression were higher in hemorrhagic areas around medium-sized brain vessels and capillaries of AD brains (Hernandez-Guillamon et al. 2012). In a genetically obese mouse model, the overexpression of MMP-9 in the microvasculature was reported to potentiate BBB leakiness upon focal ischemia with MCAO, independently of the leptin deficiency or glycemic physiological circumstances (McColl et al. 2010). In addition, a five-fold increase in levels of the neutrophil-derived MMP-9 residing in the brains of IL-1β-challenged mice was found to be the important driving force in the higher rates of tight junction degradation after the MCAO (McColl et al. 2008). Both MMP-2,9 are deemed to cause BBB damage in HIV encephalopathy (Louboutin et al. 2011).

In rat and mice model of stroke, 3–8 h after stroke, plasma levels of MMP-2 and MMP-9 are increased, and BBB permeability is increased (Lu et al. 2009). In a rodent model of stroke, elevated MMP2 levels are associated with degradation of claudin-5 and, consequently, BBB breakdown (Rosenberg et al. 1998; Yang et al. 2007). When MMP-2 is directly injected into the brain, BBB breaks down, and hemorrhage is observed (Rosenberg et al. 1992). In addition to MMP-2 and MMP-9, MMP-1 was also found to degrade claudin and occludin proteins in metastatic brain cells (Wu et al. 2015b). In a MCAO model, it was shown that increased MMP-2 activity coincides with dextran leakage from the BBB in the ischemic ventromedial striatum. However, decreased dextran leakage was observed in MCAO animals from which an MMP-2 inhibitor was also administered (Liu et al. 2012).

MMP activity is regulated by endogenous tissue inhibitor of metalloproteinases (TIMPs) through binding to specific and alternative sites of the activated MMPs (Malemud 2006). TIMP-1 and TIMP-2 are two key players that mediate the inhibition of MMPs after experimental stroke.

TIMP-1 knockout mice, subjected to the 30 minutes of MCAO, displayed elevated MMP-9 activity and more BBB rupture compared to wild-type mice. Strikingly, the deletion of the TIMP-2 gene in mice model of MCAO was accompanied by BBB opening (Fujimoto et al. 2008).

Pharmacological studies have targeted the inhibition of MMPs to shed more light on mechanisms by which MMPs cause BBB breakdown. Hydrogen in a hypertensive stroke model was discovered to alleviate BBB leakiness in the hippocampus through suppression of MMP-9 activity (Geng et al. 2015). Similar observations were documented for the acid-ascorbic in the MCAO of male rats that eventually suggested the extension of the therapeutic window of tPA administration. Ascorbic acid attenuates the BBB rupture by suppressing MMP-9 activity while MMP-9 overexpression followed by delayed administration of recombinant tPA in the control group without ascorbic acid, resulting in more neuronal deficit, increased infarct size as well as brain edema (Allahtavakoli et al. 2015). Another study in hypoxia-induced rats showed the positive effects of administration of angiotensin 1–7 on the downregulation of MMP-9. Expression of tight junction proteins, including claudin-5 and ZO-1, were recovered, and BBB permeability reduced (Wu et al. 2015a). One-hour treatment with normobaric hyperoxia (NBO) in a MCAO rat model for up to 5 minutes after stroke onset showed reduced degradation of tight junction proteins by MMP-2 and MMP-9 as well as a decline in BBB disruption (Jin et al. 2013).

MMP-3, which is activated during transient MCAO (Solé et al. 2004), was reported to mediate BBB opening during inflammation. After injection of lipopolysaccharide (LPS) in mice whose MMP-3 gene was knocked out, tight junction proteins including claudin-5 and occludin were degraded minimally, and diminished neutrophil infiltration was observed (Gurney et al. 2006). Both intravenous and intra-arterial delivery of MMP-12 shRNA-expressing plasmid to rats subjected to MCAO was shown to reduce BBB permeability and infarct size (Chelluboina et al. 2015).

Although the MMPs are well-known for their ability to degrade extracellular proteins, there is some evidence indicating that their substrates also include intracellular proteins. Proteolysis of nucleus proteins results in DNA fragmentation and apoptosis (Amantea et al. 2008; Kwan et al. 2004). 15–30 minutes after MCAO in rat brains, the increased intranuclear activity of MMP-2, 9, and 13 were reported. Thus, one of the contributing factors in oxidative DNA damage, which is seen as an early event after cerebral ischemia, was ascribed to the intracellular activation of MMPs (Cuadrado et al. 2009; Gasche et al. 2001).

In a MCAO mice animal model, where the stroke was accompanied with a systematic inflammation induced by peripheral interleukin-1 beta, there was a five-fold increase in neutrophil-derived MMP-9 in the control group as well as a transformation of the transient BBB rupture to sustained BBB injury. However, inhibition of MMP-9 alleviated detrimental effects of systemic inflammation on edema, neurological deficit, brain damage. Therefore, neutrophils were assumed to be the major source of elevated MMP-9 activity (McColl et al. 2008).

In stroke patients undergoing a hemorrhagic transformation after tPA administration, a marked increase in MMP-9 activity was reported (Chaturvedi and Kaczmarek 2014). Therefore, a proposed therapy to avoid fatal complications is tPA combination therapy accompanied by MMP-9 inhibitors. Bio-therapeutic inhibitors of MMP-9, using monoclonal antibodies, siRNA, or shRNA demonstrated an effective decrease in infarct size, as well as a reduction in life-threatening complication stemmed from BBB disruption (Chaturvedi and Kaczmarek 2014). Administration of the antibiotic, minocycline, inhibited not only MMP-9 activity but also displayed neuroprotective properties in ischemic stroke models (Yang et al. 2015). Another small molecule that can be repurposed for MMP-9 inhibition is Atorvastatin, which belongs to the statin class of drugs. Historically, Atorvastatin was formerly commercialized by Pfizer with the trade name of ―Lipitor‖, in order to lower the blood cholesterol. Administration of Atorvastatin in embolic middle cerebral artery occlusion rat model has extended the therapeutic time window of tPA to 6 hours without raising the risk for hemorrhagic transformation (Zhang et al. 2009).

Although the inhibition of MMPs is an effective therapeutic strategy combined with tPA administration, late administration of the MMP inhibitors has negative impacts on stroke patients. This limited-time window is due to MMPs possessing beneficial roles, including angiogenesis and neurogenesis, during the recovery phase of acute ischemic stroke. Therefore, to protect and restore the BBB integrity through inhibition of MMPs, more research on the exact mechanisms related to different isoforms of MMPs should be done.

Immune cells, cytokines and chemokines

Once an ischemic stroke occurs, affected cells produce signal alarm molecules, including high-mobility group box 1 (HMGB1) and brain-derived antigens that activate peripheral immune cells. Upon activation through Toll-Like Receptors (TLRs) and purinergic receptors, immune cells migrate to the site of injury with the assistance of cytokines and chemokines. After a stroke, brain-derived antigens like myelin basic protein (MBP), proteolipid protein (PLP), and myelin oligodendrocyte glycoprotein (MOG) leak from the CNS to the periphery system and activate immune responses. Following these events, antigen-specific immune cells migrate to and aggregate in the ischemic area of the brain. They can directly worsen stroke outcome by induction of cell death signaling. Additionally, they can indirectly exacerbate the stroke outcome by increasing BBB permeability for extravasation of more immune cells (Ren et al. 2012).

The role of cytokines and chemokines in BBB breakdown could be both protective and disruptive (Zhao et al. 2017). During the initial days after ischemic stroke, polymorphonuclear neutrophils (PMNs) are abundant and compromise BBB integrity by a cascade of biological events like producing MMPs, proteases, elastases, reactive oxygen species, and disorganizing junctional proteins (Gelderblom et al. 2009; Perez-de-Puig et al. 2015). On the other hand, PMNs release anti-inflammatory molecules, including annexin-1, resolvins, and lipoxin A4. Together, they serve to ameliorate the post-stroke inflammatory responses, thereby exerting protective effects on BBB integrity (Kolaczkowska and Kubes 2013).

The same scenario was uncovered for macrophage/microglia cells, in which both show dualistic roles in BBB opening after stroke. Macrophages/microglia that differentiate to the pro-inflammatory phenotype exacerbate BBB injury, whereas those macrophages/microglia differentiating to anti-inflammatory phenotypes play notable roles in BBB repair and protection (Fan et al. 2016; Gliem et al. 2012; Xiong et al. 2016).

Inflammation is described as an interplay between pro-inflammatory cytokines and anti-inflammatory cytokines. On the one hand, the pro-inflammatory cytokines, including interleukin-1 (IL-1), IL-12, IL-18, granulocyte-macrophage-colony stimulating factor, tumor necrosis factor-α (TNF-α), and gamma-interferon (IFN-gamma) worsen outcomes through producing fever, inflammation, and even sometimes shock and death. Whereas anti-inflammatory cytokines, like transforming growth factor-beta, IL-4, IL-10, and IL-13, reduce inflammation symptoms (111–114).

In experimental models of multiple sclerosis, BBB permeability is associated with early CNS inflammation and chronic release of cytokines such as TNF-α and IL-6 (Minagar and Alexander 2003; de Vries et al. 1996). Pro-inflammatory cytokines, if chronically produced, can cause brain pathology (Clark et al. 2010). In vitro studies of cerebrovascular endothelial cells from multiple species demonstrate that TNF-α can breakdown the BBB barrier by delocalizing and reducing expression of tight junction proteins like claudin-5, occludin, and ZO-1(Nishioku et al. 2010; Fiala et al. 1997; Lopez-Ramirez et al. 2012; Förster et al. 2008). It has also been shown that TNF-α affects BBB permeability through inducing cyclooxygenase-2 (COX2) release in brain microvascular endothelial cells (BMECs) (Mark and Miller 1999). Additionally, studies on brain microvascular endothelial cells have shown that expression of chemokines and cell adhesion molecules, which play a pivotal role in adhesion and transmigration of leukocytes across BBB, are upregulated with TNF-α treatment (Larochelle et al. 2011).

IL-1β is a pro-inflammatory cytokine that mediates some pathophysiology related to the stroke. In rats after permanent MCAO, peripheral administration of recombinant human interleukin-1 receptor antagonist (rhIL-1ra) reduces infarct volume and BBB permeability (Relton et al. 1996). These preclinical outcomes led to an assessment of the safety and tolerability of the IL-1 antagonist receptor in human phase 2 trial studies (Ridker 2018; Sobowale et al. 2016).

IL-6 was found to increase paracellular permeability in brain microvascular endothelial cells via an increase in ROS and downregulation of tight junction proteins (Almutairi et al. 2016). IL-6 plays a role in vascular inflammation and pathological responses to stress. An ex vivo study using ovine cerebral microvessels showed that IL-6 reduced levels of claudin-5 and 15 occludin. However, other studies did not find a correlation between IL-6 and paracellular permeability (Chaudhuri et al. 2008; Rochfort and Cummins 2015). Mice infected with the West Nile virus, whose IL-28 receptor protein was knocked out, exhibited greater neuroinvasion and severity compared to those normal mice infected with the West Nile virus. The antiviral effects of IL-28 were ascribed to IL-28 signaling, contributing to the tightening of BBB and decreasing its permeability against the invasion of pathogens into the brain via elevation of transendothelial electrical resistance (Lazear et al. 2015). IL-10 is an anti-inflammatory cytokine that inhibits IL-1 and TNF-α. IL-10 was reported to attenuate lesion volume in experimental stroke by either administration of the protein or transfection of its gene into animals (Spera et al. 1998).

Macrophage migration inhibitory factor (MIF), a constitutively expressed cytokine in several cell lines with multifunctional physiological roles, has been found to worsen the MCAO-challenged injury via distorting tight junctions and increasing infarct volume. This preclinical observation gains more significance when the clinical measurement of cytokine biomarkers in stroke patients’ blood samples exhibits MIF upregulation in ischemic stroke patients. This MIF upregulation might precede and trigger upregulation of other pro-inflammatory cytokines and chemokines (Liu et al. 2018).

The significance of IL-9 in BBB integrity in a tMCAO model on rats was uncovered through two distinct pathways, including astrocyte-mediated VEGF-A downstream pathway and astrocyte-independent endothelial nitric oxide synthase pathway. Both pathways can exacerbate stroke and undermine BBB integrity. The implication is that anti-IL-9 antibodies capable of passing through partially opened BBB could improve stroke outcome. However, further genetic deletion studies to determine the contribution of VEGF-A activation on the BBB permeability are warranted (Tan et al. 2019).

Chemokines are secondary pro-inflammatory chemotactic cytokines, that after induction by primary pro-inflammatory mediators, such as interleukin-1 (IL-1) or TNF-α, selectively regulate migration, trafficking, and homing of leukocytes to the nearby tissue. Their biological and medical significance results from their specificity (Graves and Jiang 1995) in that their mechanism of action was shown to rely on the G-protein-coupled receptors on the surface of the outer membrane of leukocytes (Ransohoff 2002). Since some receptors may be activated by several chemokines, and some might act upon several receptors, the underlying mechanism behind their promiscuous interaction is poorly understood (Yao and Tsirka 2014). Fractalkine and macrophage inflammatory protein-1α (MIP-1α) are harmful chemokines observed in stroke injury (Wang et al. 2007). Monocyte chemoattractant protein‑1 (MCP-1) is a member of the CC subfamily that is transiently expressed by parenchymal neurons during inflammation. In multiple sclerosis and ischemic stroke, it recruits monocytes, macrophages, and activated lymphocytes into the brain (Mennicken et al. 1999).

The absence of the MCP1 receptor (CCL2) in rats whose ischemic stroke was generated by tMCAO was found to act as a protecting factor against BBB breakdown following reperfusion injury. The diminishing CCL2/CCR2 axis activity by inactivating CCR2 receptors exerted influences on the cytokine expression profile so that it turns in the direction of an anti-inflammatory phenotype (Dimitrijevic et al. 2007). In the ischemic stroke animal model, MCP-1 is reported to be elevated in cerebrospinal fluid (CSF) one day after the onset of symptoms (Losy and Zaremba 2001). Mounting evidence shows that MCP-1 compromises BBB integrity not only through indirect effects (chemotactic recruitment of leukocytes to infiltrate from the periphery into the CNS) but also through direct actions. These include downregulation and redistribution of tight junction proteins, as well as a perturbation in cytoplasmic actin cytoskeleton in brain microvasculature endothelial cells (Song and Pachter 2004; Stamatovic et al. 2005). Redistributing tight junction proteins was reported to be mediated through their phosphorylation, and further studies have demonstrated binding of MCP-1 to CCR2 receptor activates Protein Kinase C (PKC) and Rho kinase (Stamatovic et al. 2006).

Chemokine receptor type 5 (CCR5) is an integral G-protein coupled receptor on the surface of white blood cells, and based on a large cohort study on stroke patients, the mutation causing loss of function in CCR5 was found to result in more significant recovery and better cognitive functions. This finding was in accordant with the results where antagonizing CCR5 by the administration of Maraviroc, an FDA approved drug for HIV, exhibited similar positive effects on the motor recovery in a post-stroke model(Joy et al. 2019).

Table.2 summarizes in vivo investigations in which the inflammatory-related factors were shown to engage in stroke-induced BBB disruption.

Table 2:

Inflammatory-related factors involved in blood-brain barrier permeability in stroke (in vivo experiments)

Inflammatory-related factors		BBB permeability	Animal type, sex and age	Effects	Technical features	References
Cytokines	IL-1β	↑	Sprague–Dawley rats, male	Detrimental	pMCAO	(B. W. McColl, Rothwell, & Allan, 2008)
	TNF-α	↓	Sprague–Dawley rats, male, 3month-old	Beneficial	tMCAO, occlusion for 2 h and reperfusion for 48 h	(Clark, Alleva, & Vissel, 2010)
	MIF	↑	Male Wistar rats, 13–15week-old	Detrimental	tMCAO occlusion for 50 min followed by reperfusion	(Y.-C. Liu et al., 2018)
	IL-9	↑	Sprague–Dawley rats, male	Detrimental	tMCAO, occlusion for 1.5 h	(Tan et al., 2019)
Chemokines	MCP-1	↑	Mice, male	Detrimental	tMCAO, occlusion for 30 min, reperfusion for 1–5 days	(Dimitrijevi c, Stamatovic, Keep, & Andjelkovic, 2007)

To date, several therapeutic targets through regulation of inflammatory responses after stroke onset have been investigated, while depending on the types and phases of the stroke insults, different elements of the inflammation cascade were addressed. For example, different types of immune cells exist, and each one possesses diverse subtypes. Further investigations should clarify the characteristics of each subtype after ischemic stroke onset as well as make an inquiry on their effects on BBB permeability. This will eventually help us translate them into the clinical settings. Additionally, with respect to progress made by immunomodulation (immunotherapy), which holds promises for stroke therapy and autoimmune diseases (Xia et al. 2016), further studying on how regulatory T-cell (Treg) therapy protects BBB integrity in ischemic stroke patients may set the stage for stroke treatment with a wide therapeutic time window.

Adhesion proteins

Cells that comprise the BBB interact directly and indirectly (through surrounding extracellular matrix) with matrix adhesion proteins on their surface (Berrier and Yamada 2007). Intercellular adhesion molecules (ICAMs) are widely distributed on the surface of the brain microvascular endothelial cells. They mediate intercellular interactions through acting as pivotal ligands for integrin proteins. ICAM-1 was found to play a key role in neuroinflammatory responses, especially when it is stimulated by pro-inflammatory cytokines (Dietrich 2002).

Fast circulating leukocytes with the assistance of ICAM-1 adhere firmly to the endothelial cells and can transmigrate across BBB into the CNS. This stepwise event results in BBB disruption (Brown 1997). Antisense oligonucleotides to ICAM-1 mRNA, as well as the antibody of ICAM-1, were observed to attenuate stroke severity and improve experimental outcomes (Kanemoto et al. 2002; Vemuganti et al. 2004). Integrins are a family of matrix receptor glycoproteins that possess a transmembrane heterodimer structure, consisting of an alpha and beta subunit. Upon binding to ECM ligands, they induce an adaptation resulting in alterations in response to surrounding microenvironments by activating critical cellular signaling pathways, including MAP kinase, Rho, Rac, and Cdc42 (Miyamoto et al. 1996). Their expression on the surface of endothelial cells, which is induced by chemokines and cytokines, precedes the leukocyte adhesion to activated endothelial cells(Smith 1993).

The anti-integrin drugs, including humanized CD11/CD18 antibody and recombinant neutrophil inhibiting factor (rNIF) peptide, were shown to be effective in treating stroke in rodents; however, they were not effective in the clinical trials (Becker 2002; Krams et al. 2003). Integrin activation results in impaired VE-cadherin localization at adherent junctions between endothelial cells, which may play a role in increasing BBB permeability (Alghisi et al. 2009). Inhibition of the integrin α˅βΙΙΙ with cyclo-RGDfV in male/female mice that underwent tMCAO was shown to improve neurological deficits and decrease brain edema and BBB permeability through inhibition of VEGF-mediated vascular breakdown (Shimamura et al. 2006).

Another protein notably expressed on the surface of endothelial cells is vascular cell adhesion molecule-1 (VCAM-1). When this adhesion protein binds to its cognate ligand, VLA-4, it mediates inter-cellular recognition between endothelial cells and leukocytes and subsequently gives the ability of vascular transmigration to the VLA-4- overexpressing leukocytes (Richard et al. 2015). Associating with the pathophysiology of stroke, the VCAM-1/ VLA-4 axis constitutes a reliable therapeutic target for halting the extravasation of leukocytes and consequently protecting the BBB against stroke (Langhauser et al. 2014). However, the preclinical efficacy of the VCAM-1/VLA-4 blockade in a robust multi-centralized and randomized mice trial was found to vary depending on the severity of infarct size and timing of the IV administration related to the tMCAO. This outcome discrepancy led to two recent human trials where collectively more than 450 stroke patients were treated with placebo and natalizumab. The natalizumab is a recombinant immunoglobulin G4 monoclonal antibody potentiated to tether to the VLA-4 and thereby blocking the VCAM-1/VLA-4 axis in the interface of leukocytes and endothelial cells. Both clinical trials outcomes were unable to verify the clinical efficacy of this strategy for combating stroke-induced complication including but not limited to BBB rupture (Ramiro et al. 2018).

Prospective

BBB rupture is the cause of hemorrhagic transformation and increased mortality after tPA treatment in stroke, especially when tPA treatment is delayed. The aforementioned studies indicate that BBB disruption plays a pivotal role in stroke outcomes. Following the stroke, several pathophysiological downstream pathways converge through BBB permeability that leads to increased mortality. This damage happens early after stroke, indicating that BBB disintegration can influence thrombolytic therapy.

The multifactorial pathophysiology of stroke, which encompasses several cells including neurons, astrocytes, and endothelial cells, makes the rational drug design difficult yet provides multiple targets for drug therapy. The unmet clinical needs should consider not only properly inhibiting detrimental pathways but also boosting innate neuroprotective pathways against stroke complications.

One of the important causative factors in the failure to translate preclinical animal studies into clinical studies stems from the fact that the stroke usually occurs in older adults where comorbidities, including diabetics, hypertension, and cardiovascular disease are more likely to affect the therapeutic efficacy and safety. Another reason for translational failure is that most conventional animal models like MCAO do not entirely mimic cellular and molecular mechanisms underpinning of ischemic stroke in human subjects.

As mitochondrial dysfunction associates with metabolic disorders accompanying/preceding stroke onset, the capitalization of the mitochondria in the context of the stroke translational research needs more attention. Therefore, interrogating mitochondria’s dynamic behavior, its metabolite fingerprinting, and its homeostasis in each BBB constituent cell under the influence of stroke can give us precious clues around novel therapeutic interventions.

From a metabolism standpoint where organs such as liver and pancreas are holistically intertwined against metabolic challenges like hypoxia and glucose deprivation, it is not surprising to observe meaningful association between ischemic stroke severity and liver physiological response. The pathophysiological responses to the metabolism-challenged brain is not limited to the brain and likely involves diverse liver metabolic responses. Given the evidence that the administration of Liver X receptors (LXR) agonists rescued BBB damage through reducing the MMP-2 and MMP-9 levels (ElAli and Hermann 2012), other potent and safe pharmacological treatments may be available. Moreover, cumulative evidence suggests the BBB is the bridge between the gut microbiome and the brain (Logsdon et al. 2018), whereby intestinal enzymes and receptor proteins contribute to BBB damages. This may represent another target for pharmacological interventions, curbing the metabolism-challenged BBB damage.

Another key factor in the context of metabolism and energy homeostasis is hypoxic induced factor alpha (HIF-1α) nuclear receptor, which is responsible for governing energy metabolism and neurogenesis through activating transcription of several pro-surviving genes in the response to the hypoxia. Given their uniquely strong sensitivity to the oxygen concentration gradians during ischemic events, investigating their role in mediating different spatiotemporal cellular transcriptomes that are associated with BBB damage phenotype is warranted.

Due to the multi-factorial etiology of the stroke-induced BBB disruption, two pharmacological approaches, polypharmacy and multi-valent drug design, might be considered. Polypharmacy is defined as the simultaneous administration of multiple drugs where each of them has a different mechanism of action. For example, with tPA, exploiting combinatorial therapies, potentiating the inhibition of the MMP-2,9 activity or expression, ROS accumulation, pro-inflammatory cytokine and chemokine production might confer promising translatable outcomes. Beside the polypharmacy where the possibility of adverse drug-drug interactions is elevated, designing multi-valent small molecules could be of clinical use and lay a groundwork for further studies. This is primarily because the latter can offer the ability to chemically modify and synthesize small molecules that have been taken to address possible adverse effects.

Clarification of the mechanisms of BBB permeability after stroke could provide novel therapeutic targets to combat not only stroke but also other neurological disorders where BBB opening directly or indirectly contributes to their progress and elevated mortality.

Figure 3: The interconnected factors governing the BBB opening through transcellular or paracellular permeability after ischemic stroke.

The pharmacological inhibition of mitochondrial respiratory chain reactions results in the BBB opening through cellular death (in this case apoptosis) and consequent BBB opening. The extravasation of the FITC-dextran-70 through cerebrovascular tight junction stained with ZO-1 red antibodies as well as extravasation of Evans blue dye in the epidural application (EA) model were the methods measuring BBB opening. A super-antigen, like LPS, was found to lead to BBB disintegration via mitochondrial impairment similarly to pharmacological inhibition of the mitochondrial functions (Doll et al. 2015). Mitochondrial homeostasis via mitophagy or mitochondrial biogenesis, accompanied with inter-neuronal mitochondrial transfer are increased in ischemic stroke and other neurodegenerative diseases. Some classes of MMP-2, MMP-9, and small molecule inhibitors such as minocycline were clinically documented to be effective in extension of the tPA therapeutic time window by protecting BBB (Yang et al. 2015). BBB opening is followed by the brain edema, neuronal death, and worsened physiological and behavioral outcomes.