An emerging role of mast cells in cerebral ischemia and hemorrhage

Abstract

Mast cells (MCs) are perivascularly located resident cells of hematopoietic origin, recognized as effectors in inflammation and immunity. Their subendothelial location at the boundary between the intravascular and extravascular milieus, and their ability to rapidly respond to blood- and tissue-borne stimuli via release of potent vasodilatatory, proteolytic, fibrinolytic, and proinflammatory mediators, render MCs with a unique status to act in the first-line defense in various pathologies. We review experimental evidence suggesting a role for MCs in the pathophysiology of brain ischemia and hemorrhage. In new-born rats, MCs contributed to brain damage in hypoxic-ischemic insults. In experimental cerebral ischemia/reperfusion, MCs regulated permeability of the blood-brain barrier, brain edema formation, and the intensity of local neutrophil infiltration. MCs were reported to play a role in the tissue plasminogen activator-mediated cerebral hemorrhages after experimental ischemic stroke, and to be involved in the expansion of hematoma and edema following intracerebral hemorrhage. Importantly, the MC-stabilizing drug cromoglycate inhibited MC-mediated adverse effects on brain pathology and improved survival of experimental animals. This brings us to a position to consider MC stabilization as a novel initial adjuvant therapy in the prevention of brain injuries in hypoxia-ischemia in new-borns, as well as in ischemic stroke and intracerebral hemorrhage in adults.

• Blood-brain barrier
• cerebral hemorrhage
• cerebral ischemia
• ischemic brain edema
• mast cells
• reperfusion injury

Introduction

Ischemic and hemorrhagic strokes are among the foremost causes of death and disability worldwide. Massive brain edema is the leading cause of death in large hemispheric strokes [1] and is only modestly alleviated by the limited conservative treatments available today. The two main detrimental effects of brain edema are: 1) mass effect with compression of normal brain structures (herniation), and 2) decrease in cerebral blood flow. Both effects damage the penumbra's (tissue-at-risk) metabolism and facilitate its transformation into an infarction.

Aimed at fast restoration of circulation, thrombolysis with tissue plasminogen activator (tPA) is the only approved therapy in acute ischemic stroke within 4.5 hours after symptom onset [2]. Fear of tPA-mediated hemorrhage is often a reason for withholding this otherwise beneficial treatment. In addition, recanalization of the occluded artery (spontaneously or with thrombolysis) may cause reperfusion injury (depending on severity and duration of the preceding ischemia), a dominant cause of which seems to be disruption of the blood-brain barrier (BBB). This leads to extravasation of plasma and erythrocytes, i.e. to formation of edema and of hemorrhage, respectively, and to inflammatory cell infiltration, as well. Moreover, tPA-treated patients with acute myocardial infarction have had a 1.1% [3] and with pulmonary embolism a 3% [4] risk for cerebral hemorrhage, so adding to the overall burden of this often fatal condition.

In contrast to ischemic stroke, no widely approved clinical therapy [5], [6] exists for intracerebral hemorrhage (ICH), which is associated with poor outcome with only 40%–50% of the patients surviving through the first year [7], [8], and many of them with chronic disability. The dismal outcome derives from early brain damage involving direct tissue damage, evolving mass effect of hematoma and edema, and also from delayed brain damage. The latter is mediated by toxins associated with blood break-down products, thrombin, inflammation [9–11], excitotoxicity [12], and disruption of the BBB [13].

Key messages

• Mast cell mediators are capable of disrupting the blood-brain barrier.

• Mast cells aggravate brain edema and inflammatory cell infiltration in ischemic stroke.

• Mast cells increase brain damage in hypoxic-ischemic brain damage.

• Mast cells play a role in the pathophysiology of hemorrhage formation following thrombolysis of acute ischemic stroke, as well as of spontaneous intracerebral hemorrhage.

Hypoxic-ischemic brain damage is associated with 20%–50% mortality of asphyxiated new-borns during the perinatal period, and up to 25% of the survivors suffer from permanent disability [14].

Here, we review the recent experimental evidence of mast cell involvement in the above-described pathologies of cerebral ischemia and hemorrhage.

Mast cells—basic facts

Mast cells (MCs) are effector cells of inflammation and immunity [15], [16]. They contain granules that store preformed mediators, which are released upon activation [17]. Derived from pluripotential stem cells of the bone-marrow [18], [19], MCs enter the circulation as committed immature precursor cells [20]. Thereafter, they home in to specific tissues, where they undergo differentiation under the specific microenvironmental conditions of the homing tissue [18], [21], MC differentiation being principally stem-cell-factor-dependent [22]. In the rodent brain, however, other factors are likely to be necessary for MC differentiation and survival including interleukin (IL)-3 and nerve growth factor [23–25]. The latter was found to be synthesized and stored by MCs themselves, suggesting an effective autoregulatory mechanism [26].

Abbreviations

Table

BBB	blood-brain barrier
CNS	central nervous system
ICH	intracerebral hemorrhage
IL	interleukin
MC	mast cell
MMP	matrix metalloproteinase
TIMP	tissue inhibitor of metalloproteinase
TNF	tumor necrosis factor
tPA	tissue plasminogen activator

Since MC differentiation occurs within the homing tissue under local microenvironmental conditions, MCs have a noteworthy heterogeneity allowing them to express different functions in health and disease [18], [19]. Based on the preformed mediator content, two types of mature MCs are recognized in rodents: serosal and connective tissue-type MCs in the pleura, the peritoneum, and the skin, and mucosal MCs in the nasal and gastrointestinal mucosa [27]. It is not yet known whether the brain MCs represent a third phenotype [27–30]. However, the various subpopulations of brain MCs—dural, perivascular, and parenchymal—may represent different phenotypes which express, to a variable degree, some features of either main phenotype [31–34]. Some differences exist in the mediator palette between rodents and humans (Table I).

Table I.  Main differences in the preformed mediators between rodents and humans.

Species	Rodents		Humans
Mast cell type	mucosal	serosal	MCT	MCTC
Biogenic amine	histamine	histamine, 5-HT	histamine	histamine
Neutral protease	chymase, cathepsin G	tryptase ^a, chymase, cathepsin G, carboxypeptidase	tryptase	tryptase, chymase, cathepsin G, carboxypeptidase
Proteoglycan	chondroitin sulphate, heparin?	heparin	chondroitin sulphate, heparin	chondroitin sulphate, heparin

^aIn some subpopulations only.

5-HT = serotonin.

T = tryptase, TC = tryptase, chymase

Mast cells—activation and mediators

Activation of MCs is a response to a wide range of various biological, chemical, and physical stimuli [16], [35], [36]. Such activation leads to the release of a vast array of mediators. Observation of pH-dependent activation of MCs is relevant for conditions of local lactate acidosis after cerebral ischemia [16]. Mediator release happens in two phases with the first fast-response phase involving release of preformed mediators stored in the granules of MCs and typically including histamine, serotonin, heparin, chymase, tryptase, and cathepsin G. Other types of mediators such as prostaglandins, leukotrienes, thromboxanes, and cytokines are synthesized de novo and secreted by the activated MCs. Excellent reviews on this topic have been published recently [16], [22], [37]. Here, we only discuss possible roles of individual mediators in the context of the pathophysiology of brain hemorrhage and ischemia.

It is important to note that MCs react to activation not only stereotypically with the explosive (anaphylactic) degranulation leading to acute release of mediators into the extracellular space, but also in a more controlled fashion resulting in differential release of mediators, i.e. the so-called piecemeal degranulation [37–40]. This allows stimulus-specific release of individual mediators, which may also vary among the different populations of MCs [35], [36]. However, whether or not piecemeal degranulation of mast cell products exists is still under debate. Nevertheless, different mast cell triggers, e.g. IL-1, IL-33, CD30, lipopolysaccharide, can induce a specific set of mediators, independent of exocytosis [37].

Mast cell location and abundance in the brain

The presence of MCs within the nervous system was noticed more than 100 years ago [41], and since then the presence of MCs in the vertebrate brain has been studied in various experimental settings [18], [27], [28], [36], [42]. MCs are found in the rat pia mater already in late fetal life, and they also enter the central nervous system (CNS) postnatally during days 7–8 [43]. MC progenitors enter the CNS during development via penetrating blood vessels, with which they remain associated (Figure 1) [31], [40], [43]. The known association of MCs with the vascular bed (preferentially at branching points) during development is dependent on contact of the blood vessel with astroglial processes [44]. In the mammalian brain, MCs are typically found in the dura mater, leptomeninges, choroid plexus, thalamus, hypothalamus, olfactory bulb, and the mesencephalon [27], [28], [32], [33], [42], [45]. They are also present in the cerebral cortex [27], [28], [46], often at the branching points of cortical penetrating arterioles, localized deep in the basal lamina nesting between glial processes [47].

Figure 1.  Tryptase-stained section of human brain showing mast cells (arrow-heads) in typical perivascular position.

Role of mast cells in brain ischemia and hemorrhage—cell cultures, animal models, and potential clinical applications

The first evidence suggesting MC involvement in the pathophysiology of ischemic stroke came from our laboratory [48], [49], and later also from other laboratories [50]. To pharmacologically activate or stabilize MCs in the rats, we systematically used compound 48/80 and sodium cromoglycate, respectively.

Cell cultures with mast cells

In an in vitro model of ischemia involving oxygen-glucose deprivation, activation of MCs could be induced [51]. Acute ischemic stroke in selected patients can currently be managed with tPA thrombolysis, this mode of pharmaceutical treatment being the only currently approved mode of therapy. Thus, it is of importance to note that tPA can activate MCs in vitro in a dose-dependent manner [52]. When co-cultured with hippocampal neurons, activated MCs potentiated glutamate excitotoxicity, presumably by histamine release, suggesting a role for MCs under conditions of cerebral ischemia with enhanced glutamatergic neurotransmission [53].

Cerebral palsy and cerebral ischemia in neonates

In a retrospective study of children with cerebral palsy, the investigators detected higher levels of several perinatal circulating cytokines, including IL-9 [54]. In addition, in a new-born mouse model of cerebral palsy, IL-9 pretreatment of the mice exacerbated brain damage [55], an effect which was later shown to be MC-dependent [56]. IL-9, in turn, has been shown to be a growth factor of human mast cells [57]. Histological changes in cerebral palsy include ischemic-like and hypoxic-ischemic cortical damage [58], [59]. In the immature rat, hypoxic-ischemic brain damage led to a rapid increase of the cerebral population of MCs and their activation, the activated MCs being present in the pia, parenchyma, and, importantly, in the regions showing neuronal loss [60]. Importantly, MC stabilization reduced MC count in the latter region and limited brain damage more than 50% compared with controls. In this study, it was noted that only hypoxia-ischemia—but not hypoxia alone—led to the brain damage, confirming stimulus-specific reactions of MCs. A similar study by a different group [61] showed contribution of MC-derived histamine to ischemia-induced neuronal death in the postnatal day 7 new-born rats. Observation of ipsilaterally upregulated expression of MC-associated genes after hypoxic-ischemic brain damage [62] provides another body of evidence supporting the involvement of MCs in this particular condition.

Focal transient cerebral ischemia—brain edema, hemorrhage formation, and inflammation

Vasogenic brain edema, hemorrhage formation, and inflammatory response are all associated with disruption of the BBB. Before we address the possible interactions between MCs and the BBB at the molecular level, we provide the available experimental data showing MC involvement in these various brain ischemia-dependent processes.

Brain edema is a major determinant of acute stroke mortality in man [1]. In the rat, we showed MC involvement in ischemic brain edema early after focal cerebral ischemia onset [63]. Specifically, we found almost 90% increase after treatment with the MC-degranulating agent compound 48/80, and almost 40% reduction of ischemic brain edema after treatment with the MC-stabilizing agent sodium cromoglycate. The role of MCs as a major determinant of brain edema in this animal model was confirmed in genetically MC-deficient rats, which had 60% less brain edema compared with their wild-type MC-competent litter-mates [63]. Moreover, a role for histamine in the development of brain edema has been observed in a rat model of bilateral common carotid artery occlusion, which represents global rather than focal cerebral ischemia [64], [65]. Since histamine is the major MC mediator during acute stimulation of MCs, the above finding implies MC involvement also in this scenario.

Thrombolysis-mediated hemorrhage formation can devastate clinical prognosis following successful stroke thrombolysis and may occur even in patients with acute myocardial infarction or with pulmonary embolism, i.e. in patients treated with thrombolytic agents without any on-going cerebral ischemia [3], [4]. We reported a robust tPA-mediated activation of MCs in vitro[52], a finding not reported for any other fibrinolytic drug so far. Importantly, in in vivo experiments in rats, which underwent focal cerebral ischemia-reperfusion and postischemic tPA administration, we found a 70- to 100-fold increase in hemorrhage formation. Pharmacological MC stabilization with cromoglycate led to significant reduction in the tPA-mediated hemorrhage formation at 3 (97%), 6 (76%), and 24 hours (96%) compared with controls. Moreover, MC-deficient rats showed robust reduction of tPA-mediated hemorrhage compared with their wild-type MC-competent litter-mates. Importantly, these reductions in hemorrhage formation translated into better neurological outcomes and reduced the mortality. Specifically, in the control group 30% and in the tPA-alone-treated group 65% of animals died, while no deaths occurred in the tPA-treated group if the animals had been treated with the MC-stabilizing drug sodium cromoglycate. Again, a similar beneficial effect was observed in the MC-deficient rats.

MCs contribute to CNS inflammation [36], [66], [67], particularly by recruiting leukocytes to sites of vascular inflammation [68–70]. Recently, we reported MCs to play a significant role in neutrophil infiltration following experimental cerebral ischemia-reperfusion injury in the rat [63]. Thus, MC stabilization reduced neutrophil infiltration by 40%, and the reduction was even stronger (almost 50%) in MC-deficient rats, while MC degranulation enhanced neutrophil infiltration. In a different study [52], an intravenous administration of tPA was accompanied by a 2- to 3-fold increase in neutrophil infiltration in ischemic hemispheres and by an up to 6-fold increase in the intact hemisphere, suggesting an independent proinflammatory effect of tPA. Moreover, the numbers of brain tissue neutrophils in MC-deficient rats subjected to focal cerebral ischemia and subsequent tPA treatment were less than 40% of those in their wild-type litter-mates.

Blood-brain barrier as the main target of mast cells

Disruption of the BBB in the above-mentioned scenarios is a consequence of ischemia-reperfusion injury. Outside the CNS, MCs have been shown to be involved in the pathophysiology of elements of ischemia-reperfusion injury in the myocardium [71–75], kidney [76], intestine [77–81], skeletal muscle [82–84], and the lungs [83], [85], [86].

Differentiation or migration of intraparenchymal MCs in the developing brain occurs during the formation of the BBB [87], and the time-course of MC migration from the choroid fissure to the thalamus coincides with the formation of the vascular network [31]. Since MCs not only show characteristic perivascular location in the brain, but many of their mediators are profoundly vasoactive [28], [31], [32], [34], it is reasonable to hypothesize that MCs play a role in the regulation of the BBB. MCs were shown to be involved in the regulation of the BBB in healthy doves [88], and they also regulated the BBB under conditions of acute stress challenge in the rat [89]. In that latter study, the authors noticed that a relatively limited population of MCs may be sufficient to mount an effective alteration of the BBB permeability. Furthermore, MC-mediated BBB disruption has been documented to precede the onset of clinical symptoms in multiple sclerosis [90], [91].

Our group addressed the role of MCs in the BBB disruption following focal cerebral ischemia [63]. In the rat model of focal cerebral ischemia, we observed a 50% increase in the BBB disruption after triggering of MC degranulation, while pharmacological stabilization of MCs reduced it by 50%. The role of MCs in the regulation of the BBB disruption was confirmed by demonstrating 50% less BBB disruption in genetically MC-deficient rats compared with their wild-type MC-competent litter-mates. Similar experiments with additional administration of tPA supported the involvement of MCs in mediating the BBB disruption [52]. Most probably, MCs with their armamentarium of bioactive mediators influence all the described consequences of ischemia-reperfusion injury by targeting the BBB and the basal lamina (Figure 2), both of which normally secure the integrity of the cerebral microvasculature and the extravascular environment of the CNS.

Figure 2.  Schematic illustration of the neurovascular unit (middle) and the components of the blood-brain barrier (bottom left) and the basal lamina (bottom right) (ECM = extracellular matrix; ZO = zona occludens; AF6 = afadin; 7H6 = tight junction associated phosphoprotein).

The BBB is composed of tight interendothelial junctions in capillaries and in postcapillary venules (Figure 2). Furthermore, several integral transmembrane proteins contribute to the BBB integrity (e.g. claudin, occludin, and junctional adhesion molecules). The second barrier—the basal lamina—is a specialized part of the extracellular matrix that connects the endothelial cells to the adjoining cell layers and further to the smooth muscle layer of the media (Figure 2). The basal lamina loses its integrity very early after the onset of ischemia, as demonstrated by rapid disappearances of the three main constituents of the basal lamina, i.e. laminin, fibronectin, and collagen type IV [92], [93].

Some evidence shows involvement of proteases, e.g. matrix metalloproteinases (MMPs), plasminogen activators, and cathepsins, in the disruption of the matrix of the basal lamina. The MMPs comprise a large family of proteolytic proenzymes, which require activation. Once activated, they can degrade most of the protein constituents of the neurovascular matrix, i.e. collagen, elastin, fibronectin, vitronectin, and gelatine [94]. Involvement of the gelatinases MMP-2 and MMP-9 in the degradation of the components of the basal lamina, with ensuing BBB disruption and development of edema and hemorrhage, has been demonstrated experimentally in several pieces of work [92], [95–100]. Furthermore, MMP-mediated disruption of the tight junction proteins occludin and claudin (Figure 2) was reported in the rat [100], [101]. Finally, clinical data have revealed a positive correlation between the degree of brain edema and the level of MMP-9 in the cerebrospinal fluid [102], suggesting the involvement of this MMP also in human CNS pathology involving BBB damage.

In view of the processes described above, it is of relevance that peripheral MCs are able to migrate across tissues. Thus MCs in the bovine gall-bladder migrated from the adjacent connective tissue across the basal membrane to enter the epithelium and so to reach the luminal surface of the gall-bladder [103]. MCs are also able to enhance the antibody-mediated injury of the skin basement membrane in mice [104]. Moreover, in vitro data show that both resting and activated MCs can attach to and move across a variety of extracellular matrix proteins, including fibronectin and laminin, for which MCs have receptors [105–107]. The involvement of MCs in the proteolytic degradation of endothelial basement membrane with ensuing detachment of the endothelial cell, i.e. endothelial erosions, has been extensively studied in the context of atherosclerosis and its thrombotic complications [108], [109].

How can MCs disrupt the BBB and degrade the components of the basal lamina (Figure 2)? There are several possible mechanisms involved, in which the vasoactive and the matrix-degrading components of MCs act in concert. MCs produce a potent vasodilatatory mediator, histamine, which contributes up to 90% of the histamine content in the thalamus and up to 50% of that in the whole brain [110]. It is of fundamental importance for the pathogenic mechanisms of acute ischemic brain conditions that, of the resident cells in the brain, only MCs are able to acutely respond by releasing massive amounts of preformed vasodilatatory mediators, such as histamine. Under experimental conditions in vitro, histamine binding to its specific receptors on the endothelial surface leads to increased phosphorylation of adherens junction and loosening of vascular-endothelial-cadherin-mediated adhesion of adjacent endothelial cells [111], [112]. The same effect can be induced by MC-derived protease tryptase [113], a process which can be aggravated by other MC-neutral proteases—chymase, and cathepsin G [114]. Again, all these three MC proteases are acutely released from activated MCs along with histamine. Moreover, cerebral MC-derived chymase [31] is a potent protease which cleaves fibronectin and occludin and activates procollagenases [115–117] even in the presence of the MMP-inhibiting protein TIMP (tissue inhibitor of metalloproteinase)-1 [118]. Indeed, MC-derived chymase can degrade TIMP-1 [119], and a key role for MC chymase in the activation of pro-MMP-2 and pro-MMP-9 was demonstrated recently by Pejler and co-workers [120]. Similarly, MC tryptase is also able to activate MMPs [121] and to degrade TIMP-1 [119]. Moreover, MCs themselves can release gelatinase A (MMP-2) and gelatinase B (MMP-9) [122]. Regarding cathepsin G, this chymotrypsin-like neutral protease present in some MCs is able to cleave many components of the extracellular and pericellular matrix, including fibronectin and vitronectin [109], [123]. The role of cathepsins in microvascular matrix degradation following focal cerebral ischemia has been noticed [124]. Finally, apart from proteases, MCs can release tumor necrosis factor (TNF)-α and IL-1, two cytokines shown to be involved in the modulation of BBB permeability and consequently in ischemic brain edema formation [125–127]. Indeed, MCs do possess a potent armamentarium to target the components of the BBB and basal lamina shortly after their activation, whereas de novo production of mediators reactivates and maintains the process.

The mechanism of postthrombolytic hemorrhage formation still awaits detailed understanding. It is of interest to note that it occurs with any of the fibrinolytic substances in clinical use, i.e. tPA, streptokinase, prourokinase (can be activated by human MC tryptase [128]), and reteplase [129], [130]. They all are potent serine proteinases, which, in addition to their local plasminogen-activating effect to dissolve fibrin locally, also induce systemic plasminemia [131]. Plasmin, again, has an independent proinflammatory activity that causes direct brain tissue damage [132], [133], and in addition it degrades a range of extracellular matrix proteins by activating MMPs [134]. Interestingly, inhibition of MMPs reduced tPA-mediated mortality in experimental ischemia-reperfusion [135]. Perhaps even more important in the setting of thrombolysis of an acute ischemic stroke is the observation that the development of hemorrhage may be further promoted by heparin, also acutely released by activated MCs, which are the only source of this strong anticoagulant [136], [137]. Heparin released locally from perivascularly positioned MCs may prevent the formation of hemostatic plugs, which otherwise would patch any BBB breaches. Thus, the endogenously released heparin may aggravate erythrocyte extravasation and contribute to hemorrhage formation. In dogs bearing MC tumors, intense activation of the fibrinolytic system has been observed after injection of minute amounts the MC-degranulating agent 48/80 [138]. Superimposed on the fact that (in addition to heparin) tPA itself belongs to the palette of MC-derived mediators [139], this confirms the antithrombotic and fibrinolytic potential of MCs.

Infiltration of neutrophilic granulocytes is held to compromise microcirculation and BBB integrity and to contribute to the release of free radicals. One of the most potent mediators possibly involved in the recruitment of granulocytes is TNF-α, which represents a distinct type of MC mediators, since after MC stimulation it is both acutely released from preformed pools stored in the cytoplasmic secretory granules, and then continuously synthesized de novo and secreted [22]. Being probably the only cell type containing preformed stores of TNF-α [140], MCs are likely to represent a critical initial cellular source of TNF-α during inflammation [16]. Later during the inflammatory response, TNF-α is also produced by neutrophils, eosinophils, T cells and B cells, and macrophages [141]. The role of MC-derived TNF-α in inflammatory cell recruitment was reported in various non-cerebral tissues [141–143]. However, as much as 50% of the recruitment of neutrophils during MC-dependent cutaneous late phase reactions in the mouse was independent of MC-derived TNF-α, suggesting that other MC-associated mediators also contributed to this response [142]. Indeed, MCs are a source of other mediators that could also promote leukocytic infiltration, including IL-1, members of the macrophage inflammatory protein-1 family [141], [144], [145], and other potent chemotactic substances, such as platelet-activating factor, IL-4, IL-5, IL-8, and neutrophil, eosinophil, and macrophage chemotactic factors. Together, these chemoattractants are likely to contribute to the early steep chemotactic gradient, after which neutrophils start to transmigrate. Since MC stabilization reduced both the number of transmigrated neutrophils and of those still adherent to the endothelial wall [63], MC blocking not only reduced passive neutrophil trafficking through BBB breaches, but probably also attenuated the perivascular chemotactic gradient up which the neutrophils migrate. The proposed proinflammatory effect of tPA might be mediated not only by the described MC activation and the released of MC-derived mediators, but also directly by plasmin, which is able to induce the synthesis of platelet-activating factor, activate the terminal complement cascade [146], and to recruit neutrophils when injected into the brain [133].

The fact that clinical trials based on inhibition of neutrophils have failed to have therapeutic success [147], [148] may be explained by the neutrophils arriving at the scene too late to be able to influence the early damage [149]. In contrast, the resident MCs are already present in the brain tissue at the onset of ischemia, and may thus offer an alternative target for early antichemotactic intervention before the involvement of any circulating inflammatory cells.

Brain hemorrhage

MC involvement in the pathophysiology of ICH was shown in a rat model in our laboratory [150]. In that study, MC inhibition led to significant reductions in the growth of ICH volume during a 24-hour follow-up. Furthermore, depending on the route of administration of the MC-stabilizing agent sodium cromoglycate, brain edema was reduced by 50%–75%. MC degranulation, again, led to opposite effects. The specific role for MCs in this setting was supported by similar findings in the genetically MC-deficient rats. Importantly, the degree of brain edema corresponded with mortality of the animals. Thus, all rats treated with the MC-stabilizing agent survived and showed good outcome, whereas those treated with saline or with the MC-degranulating agent had mortality rates of 45% and 55%, respectively. This striking result could be observed again, when we repeated the experiment using the genetically MC-deficient rats: all MC-deficient rats stayed alive, while 25% of their wild-type litter-mates died.

The pathophysiological basis of MC-dependent brain edema during ICH is uncertain. We propose that the release of MC-derived vasoactive substances (histamine) and proteolytic enzymes (tryptase, chymase, and cathepsin G) increases vascular permeability and disrupts the basal lamina of the vasculature along with the surrounding extracellular tissue matrix, leading to secondary aggravation of extravasated blood cells and further extravascular spreading of serum proteins. Furthermore, mast cell heparin markedly accelerates the generation of another potent vasoactive substance, bradykinin [151]. Moreover, several MC-mediated cytokines (e.g. TNF-α and various ILs) are likely to be involved in this scenario. Indeed, some human studies show a correlation between plasma concentrations of TNF-α and IL-6 and magnitude of perihematomal brain edema and hematoma growth [152], [153]. Although the cellular sources of these cytokines was not known in the human studies, hematoma growth has been attributed not only to continued bleeding from the primary source [154] but also to a mechanical disruption of the surrounding vessels in the peripheral zone around the ICH [154], [155], which is a possible stimulus for enhanced local MC activation. As discussed above, upon activation MCs acutely release heparin, an endogenous anticoagulant, which may promote impaired hemostasis and prolonged extravasation of erythrocytes and the occurrence of secondary hemorrhage. This hypothesis is in line with our recent observations in a rat intracerebral hemorrhage model (injection of autologous blood into the basal ganglia), in which the significant hematoma growth observed in saline- and compound 48/80-treated groups could be significantly inhibited by pharmacological stabilization of MCs or by using MC-deficient rats in the experiment [150].

Conclusions

To sum up the current evidence about the emerging role of MCs in cerebral ischemia and hemorrhage, it has become evident only in recent experimental studies that MCs, which represent a stationary, resident cell population in the mammalian brain and meninges, participate in several clinically important sequelae of the progression of ischemic and hemorrhagic stroke, namely in BBB damage, brain edema, hemorrhage formation, and in the inflammatory response to such events. The observed effects are well in line with the biphasic phlogistic responses of MCs described in type-I hypersensitivity reactions and the ensuing late phase reactions. MCs reside in the pivotal perivascular position and contain an abundance of vasoactive, proinflammatory, anticoagulant, and proteolytic mediators released immediately after activation, and many of these mediators are released also during the late phase response. This provides MCs with a multitude of strategic opportunities to interact at the level of the neurovascular unit. Hence, MC stabilization may represent an efficient strategy by blocking simultaneously several molecular cascades that participate in the progressive tissue damage associated with ischemic and hemorrhagic stroke. This contrasts with the prevailing approach of blocking just one single pathway, which has repeatedly failed in clinical trials employing neuroprotective agents in ischemic stroke.

In addition to a role in ischemic stroke, MCs seem to participate also in hemorrhagic cerebral events due to enhancement of vascular injury and aggravated bleeding. Indeed, all described antihemostatic effects of MCs superimposed on their vasculopathic proteolytic potential afford to MCs a fibrinolytic capacity in the microvascular milieu. Then, the activated MCs clearly have the potential to exert unwanted side-effects as an acute response to therapeutic administration of thrombolytics after ischemic stroke. Stabilization of MCs, again, is a potential novel adjuvant therapy to prevent these occasionally devastating complications of thrombolysis. This bold proposal is mainly based on the experimental data in rats, which can differ from the human situation. This issue clearly requires further studies to establish its potential role as an adjunct mode of the acute therapy of cerebral ischemia and hemorrhage in man. Fortunately, MC-stabilizing pharmacological agents are safe in clinical use, and so we are ready to explore novel indications for this old drug, such as acute ischemic stroke and intracerebral hemorrhage.

Acknowledgements

Declaration of interest: The authors report no conflicts of interest. The authors alone are responsible for the content and writing of the paper.