Nitric Oxide-Dependent Pathways as Critical Factors in the Consequences and Recovery after Brain Ischemic Hypoxia

The obstruction of blood flow by the clot dramatically reduces glucose and oxygen supply in ischemic brain region and triggers ischemic cascades, which include accumulation of lactate and malfunction of ion pumps (Na⁺/K⁺-ATPase and Ca²⁺/H-ATPase), subsequently inducing membrane depolarization and calcium ion (Ca²⁺) overload. Membrane depolarization causes the release of excitotoxic amino acid glutamate and its translocation into the extracellular compartment. The reversal of the activity of glutamate transport proteins prevents glutamate reuptake which results in robust increase in glutamate level and activation of extrasynaptic NMDA receptors [4,16,30,31].

In vivo studies show, that in the MCAO model in rats, immediate early genes (c-fos, c-jun) are activated by protein kinases and other second messengers shortly after the onset of ischemia.

The synthesis of NO^• through nNOS is mainly related to calcium overload induced by glutamate in ischemic neurons [32]. The exacerbated activation of nNOS and the excess NO^• production due to uncoupled state, contributes to neurotoxicity and neurodegeneration by free radical formation [33,34,35].

Neurons are particularly sensitive to stress caused by ROS/RNS overproduction because of the relatively low levels of antioxidants compared with other cells. Physiologically, basal ROS/RNS production in neurons is constitutively generated by mitochondrial metabolism, which is higher than that in other cells due to the necessity to maintain neuronal circuit activity and synaptic transmission. Excessive nNOS activation under hyperactivation of NMDA receptors results in massive and uncontrolled generation of NO^• followed by the downstream formation of ROS/RNS, resulting in neuroinflammation and nitro oxidative burst. The prolonged destabilization of nNOS contributes to neurotoxicity and the potentiation of ischemic damage (Figure 2) [36,37]. Genetically engineered mice that overexpress the free radical scavenger, superoxide dismutase, have less edema formation than wild-type littermates confirming the detrimental role of free radicals in ischemia pathology [38,39,40].

A second glutamate-independent rise in Ca²⁺ is observed 2–3 h after an ischemic episode, further contributing to the activation of Ca²⁺-dependent enzymes (except NOS also proteases, phospholipases, cyclooxygenases, endonucleases) and leading to irreversible changes that promote the apoptotic and necrotic death of neurons [41,42,43,44,45].

In animal studies with the use of nNOS^−/− knockout (KO) mice, in the model of cerebral ischemia-reperfusion, the smaller infarct size comparing to their wild-type littermates was observed. Relative cerebral blood flow after reperfusion was also higher and the levels of nitrates was significantly decreased in nNOS^−/− mice [46,47]. No 3-nitrotyrosine immunoreactivity, a marker for ONOO⁻ formation associated with cell death, was reported in mutant mice deficient in nNOS activity, which were subjected to reversible MCAO [16].

In contrast to nNOS, NO^• generated by eNOS expressed in the endothelial cells has been suggested to have beneficial effects [17]. Here, eNOS not only promotes vascular dilation but also increases vascular smooth muscle cell proliferation and migration, and thereby enhances arteriogenesis after stroke [48].

In this case, eNOS deficient mice suffered from more severe ischemia-reperfusion injury in MCAO model, significantly reduced cerebral blood flow which subsequently resulted in greater infarct size [46,47,49].

The results of in vivo studies clearly suggest that the generation of NO^• and its neurotoxicity after reperfusion is closely related with the activity of nNOS. Additionally, nNOS-derived NO^• may play a major role in early blood-brain barrier BBB disruption following transient focal cerebral ischemia [50].

Immediate early gene products activate transcription sites on the cytokine genes such as tumor necrosis factor-α (TNF-α) or interleukin-1β (IL-1β) that appear within hours after stroke. Appearance of cytokines attracts inflammatory cells such as neutrophils, that appear 4 to 6 h after a stroke [51,52]. P-selectin, E-selectin, vascular cell adhesion molecule (VCAM) and intercellular adhesion molecule (ICAM) appear on the vessels with different time courses, but the ultimate effect is that neutrophils rolling along the vessel attach and cross into the brain. At later times macrophages become the predominate cell type in the injured site [15,53].

The role of glia cells (astrocytes and microglia) is pivotal in progressing cerebral ischemia. Astrocytes play critical role BBB integrity and the maintenance of extracellular ion homeostasis by buffering excitatory transmitters released by neurons and producing trophic factors that support neuronal growth and survival. In response to pathological situations in surrounding tissue such as ischemic injury astrocytes undergo morphological, molecular and functional changes and become reactive astrocytes [54] The changes associated with the reactive state can directly impact synaptic transmission and neuronal circuit activity, thereby potentially contributing to the pathological changes observed after ischemia [55].

On the other hand, resting microglia in the mature brain is a powerful weapon under pathological activation. Following ischemia, as a result of rapid neuronal death, microglia migrates to damaged tissue to exert a neuroprotective effect by clearing dead tissue, inhibiting cytotoxic neuronal damage and releasing neuroprotective growth factors. However, microglial activation triggers iNOS production, which, together with the parallel appearance of reactive astrocytes, results in massive iNOS expression observed relatively late after an ischemic episode (after approximately 24 h) and generation of ROS/RNS [51,55,56,57,58]. In some brain areas, astrocytes also express nNOS, which may be tonically active after chain reactions initiated by ischemia. However, some contrary reports indicate the minor role of microglial iNOS in mediating brain injury after stroke [59].

In contrast to nNOS, the induction of iNOS expression begins 12 h after induction of ischemia, increases progressively over time and reaches maximal levels after approximately 24–48 h as shown in transient MCAO model in animals [60]. The other in vivo studies confirm that NO derived from iNOS appears to contribute to neurotoxicity after ischemic stroke. The infarct and the motor deficits produced by MCAO were smaller in iNOS knockouts than in wild-type mice confirming that iNOS-derived NO^• is one of the factors contributing to the expansion of the brain damage that occurs in the late stage of post-ischemic period [61]. Furthermore, the reduction of infarct size and improvement of neurological deficits was not observed up to 24 h after MCAO, indicating that iNOS does not participate in the initiation of ischemic brain damage [61].

Pathological glial activation escapes endogenous control and turns into an autoaggressive pathomechanism that contributes to secondary neuronal damage occurring after brain ischemia [58]. Leukocytes accumulating at the site of neuronal necrosis act as an additional barrier to this cascade of destructive events by secreting cytokines, including interferon-γ and interleukins (Figure 3).

Overall increased activity of nNOS/iNOS isoforms results in the subsequent production of NO^• in neurons, glia, neutrophils and the rest inflammatory cells inducing the release of inflammatory factors that promote cytotoxicity, including the increased activation of cell adhesion molecules, cytokines, TNFα and matrix metalloproteinases (MMPs) [62]. TNFα, caspases and MMPs are particularly important in triggering apoptotic signals in cells and contribute to BBB damage, neurotoxic substance release, free radical generation, oxidative stress and brain edema [63,64]. As a result, after hours and days the infarct size expands due to excitatory amino acid release, loss of ion homeostasis, decreased pH, inflammation and edema, causing additional damage and apoptosis in the surrounding areas [65]. The strategy for post-ischemic neuroprotective therapies is to target the peri-infarct or penumbra region to prevent or rescue the spreading damage of the initial infarct.

Reassuming, ischemic brain injury involves a complex interaction between leukocytes, glia, neurons and the endothelium, which form a vicious chain of events that results in massive and continuous production of iNOS resulting is subsequent generation of ROS/RNS [59,60]. The pathological events associated with ischemic brain injury involve energy failure, oxidative stress, acidosis, disruption of ion homeostasis, neuronal cell excitotoxicity or inflammation and evolve progressively over time. The early and late stages of ischemia described above concern not only sequential NO^• production but also immunological responses (which are in part related to NO^• generation) described elsewhere in more details (for review see: [66,67]).

The NO^•-mediated S-nitrosylation process is a redox-based posttranslational modification that modulates protein function and activity. S-nitrosylation is the chemical reaction of an NO^• moiety with the sulfhydryl groups of target proteins, which leads to the formation of S-nitrosothiols (R-SNO), producing S-nitrosylated proteins (SNO-proteins) [68].

S-nitrosylation can occur both extracellularly and intracellularly [69]. The S-nitrosylation of the regulatory binding partners of transcription factors (TFs) (for example, HIF-1α) may impose an extranuclear influence on their activation, stability and nuclear targeting [69,70]. On the other hand, the S-nitrosylation of critical redox-sensitive Cys residues in the DNA-binding or allosteric sites of TFs invokes alterations in gene expression [71,72]. Finally, S-nitrosylation may regulate protein function by the covalent addition of an NO group to a cysteine thiol/sulfhydryl group (RSH or, more properly, thiolate anion, RS⁻) to form S-nitrosothiol derivatives (RS-NO), which is the most important aspect of the S-nitrosylation process [70]. The reaction is mediated by NO^•-related species including NO^•, NO⁻ (nitroxyl anion, which is NO^• with one additional electron) and NO⁺ (nitrosonium ion, which has one fewer electron than NO^•) [70].

In addition to the process of S-nitrosylation, functional equivalents of NO⁺ can be transferred from one nitrosothiol to another in a process called transnitrosylation, whereby an NO^• moiety is transferred from a SNO-protein to a free thiol on another protein, which occurs when two proteins interact directly and possess appropriate redox potentials to allow electron transfer [73].

Aberrant S-nitrosylation occurs as a consequence of exacerbated nitrosative stress via the excessive production of NO^•, which nitrosylates cysteine thiols with only partial SNO motifs or located more distant from the NO^• source [74]. These aberrantly S-nitrosylated proteins may contribute to pathological changes by triggering protein misfolding, mitochondrial dysfunction, transcriptional dysregulation, synaptic damage and neuronal injury [68]. In contrast, some SNO-proteins lose NO^• groups from their Cys thiols and undergo denitrosylation, which contributes to the regulation of the SNO signaling cascade [73]. Therefore, both S-nitrosylation and denitrosylation regulate protein activity and may be involved in pathological processes.

Under physiological conditions, nNOS is S-nitrosylated by NO^•. The studies were performed both in vitro in HEK223 cell lines, cultured primary cortical neurons treated with OGD/reoxygenation and in vivo in rat hippocampus during cerebral ischemia-reperfusion. In all experimental schedules the enzyme is S-nitrosylated in resting or physiological state and undergoes significant denitrosylation under oxygen deprivation, which is coupled with its increased activity [75]. The subsequent increase in NO^• production mediates the S-nitrosylation of proteins that initiate apoptotic signals in neurons, such as GluR6, c-Jun N-terminal kinase 3, phosphatases or tensin homolog [76].

In contrast to nNOS, the process of S-nitrosylation inhibits the activity of eNOS by inducing dimer disruption, dephosphorylation and changes in the subcellular targeting status as shown in vitro [77,78].

Two of the most important proteins that undergo aberrant S-nitrosylation in response to ischemic injury are GAPDH [79,80] and matrix metalloproteinase 9 (MMP9) [81]. GAPDH has been implicated in neurotoxicity and neurodegeneration and regulates transcriptional activation, apoptosis initiation, ER to Golgi vesicle shuttling and fast axonal or axoplasmic transport [82]. Following an ischemic episode, GAPDH accumulates rapidly both in the ischemic core and in penumbral apoptotic neurons as shown in vivo using MCAO model in rats [83]. Aberrant GAPDH s-nitrosylation, translation to nucleus, concomitant neuronal death occur during the early stages of reperfusion as shown, e.g., in the rat four-vessel occlusion ischemic model [80].

MMP9 belongs to the zinc metalloproteinase family, which is involved in the degradation of the extracellular matrix. The enzyme is acutely activated during ischemia and is selectively S-nitrosylated by NO^• during cerebral ischemia in vivo [81]. In in vitro studies s-nitrosylation of the cysteine switch at the active site of the enzyme together with ROS-mediated oxidation of MMP9 to sulfinic or sulfonic acid derivatives triggered the apoptotic form of cell death [81], suggesting a potential extracellular proteolysis pathway to neuronal cell death in which s-nitrosylation activates MM9.

HIF-1α is the primary mammalian transcription factor specifically regulated by hypoxia and plays an essential role in cellular and systemic O₂ homeostasis by regulating the expression of genes important in tissue survival, that regulate glycolysis, erythropoiesis, angiogenesis or catecholamine metabolism [84].

HIF is primarily regulated by changes in protein stability and transcriptional activity in oxygen-dependent manner. Under physiological conditions HIF-1α is rapidly degraded and α subunit is hydroxylated by asparagine and proline residues by the family of prolyl-4-hydroxylase domain (PHD) proteins and factor inhibiting HIF (FIH), whose activity is dependent on molecular oxygen, ferrous iron, 2-oxoglutarate [85,86,87]. In response to low tissue oxygenation during ischemic stroke, the activity of PHD and FIH declines resulting in stabilization of HIF-1α, its accumulation within the cell and translocation into the nucleus. It results in the transcriptional activation of several dozen hypoxia-responsive genes through binding the hypoxia-responsive element (HRE) in their promoter region [88,89,90,91].

The activity of HIF-1α is regulated by NO via the mechanism of S-nitrosylation. NO metabolites such as S-nitrosoglutathione (GSNO) and peroxynitrite stabilizes or destabilizes HIF-1α, respectively [92,93,94].

The impact of HIF-1α on neuronal survival upon stroke is controversial, as it mediates both anti- and pro-survival genes [85]. Up regulation of vascular endothelial growth factor (VEGF) or erythropoietin (EPO) promotes adaptation to hypoxic/ischemic stress [95] and results in reduced infarct size after cerebral ischemia in MCAO model in rodents [58,85]. The HIF-1α dependent activation of eNOS transcription and subsequent NO release in the endothelium also contribute to reduction of ischemia infarct volume also shown in MCAO model [96]. Exogenously administered NO metabolite S-nitrosoglutathione (GSNO) was found not only to stabilize HIF-1α and to induce HIF-1α-dependent genes but also to stimulate the regeneration process and to aid in functional recovery in traumatic brain injury animal model [93].

On the other hand HIF-1α regulates iNOS transcription preserving iNOS translation during ischemia, thus contributing to progressive ischemia-induced inflammation [97,98].

Tissue-specific knockouts have also given conflicting results on the role of HIF-1α during ischemia. The studies of Helton et al. showed that the brains from neuron-specific HIF-1α deficient mice were protected from hypoxia-induced cell death, when subjected to normobaric chamber or bilateral carotid artery occlusion (BCAO) model, suggesting that decreasing HIF-1α level can be neuroprotective [84]. Additionally, in HIF-1α KO mice genes involved in apoptotic pathway were uniquely downregulated when compared with WT animals. Endothelial-specific HIF-1α knock-out reduced BBB permeability and brain infarction in diabetic mice subjected to MCAO procedure [99]. In in vitro studies, in primary cortical neuron cultures, silencing HIF-1α attenuated the accumulation of iNOS [100].

On the other hand in the studies of Baranova et al. neuron-specific inactivation of HIF-1α increased brain injury in mice MCAO model [101]. Additionally, neuron-specific PHD inactivation which resulted in up-regulation of HIF-1α lead to smaller infarct size and reduced edema formation in transient MCAO model in mice [102].

In the models of permanent MCAO or acute phase of ischemic stroke in mice the expression of HIF-1α was enhanced in the place of injury, causing the massive production of iNOS [103], indicating that activation of HIF-1α might be involved in the mechanisms through which iNOS promotes cell death or survival constituting a critical factor in widespread inflammation and subsequent pathological events. Another studies showed that intermittent hypoxic conditions after brain ischemia displayed a neuroprotective effect, and despite relatively high expression of HIF-1α, lower expression of iNOS in the border between infarcts and normal tissue was observed, suggesting that overactivation of HIF-1α may suppress the activation of microglia in ischemic mice [104]. Enhanced HIF-1α activation was responsible for triggering the transcription of HIF-regulated genes (VEGF, erythropoietin, eNOS), reduced infarct size and caspase-3 activation in MCAO or common carotid arteries occlusion models in mice [96]. Increase HIF-1α was accompanied by increased iNOS expression as shown in MCAO-reperfusion injury model in rats [105].

In in vitro studies the upregulation of HIF-1α signaling was shown to improve cGMP production following ischemia through the maintenance of cGMP protein kinase activity [106], thus preventing the NO-mediated production of ROS/RNS instead of cGMP.

Overall, the role of HIF-1α in ischemia remains inconclusive, but despite controversial results the role of HIF-1α was shown to be of special importance in ischemia preconditioning and may either promote or prevent neuronal survival [107]. Partially the pro-survival and pro-inflammatory roles in the ischemic brain might depend on the injury model, time point or cell type assessed [101,108].

The contradictory results indicate that phenotype and transcriptional response to hypoxia in vivo is much more complex that would have been supposed. The brain has multiple ways of inducing HIF-1α-dependent genes involved in the response to hypoxia (that promote erythropoiesis, angiogenesis or vasodilation), and one of the alternatives is HIF-2α [109,110,111,112]. HIF-1α and HIF-2α combined KO mice exhibited reduced expression of the anti-survival genes in MCAO model in mice [85]. Even though the mice initially performed better, became more impaired 72 h after reperfusion, accompanied by increased apoptosis and reduced angiogenesis. HIF-1α and HIF-2α can partially compensate for each other, although specific target genes are differentially regulated after ischemia. Combined loss of neuronal HIF-1α and HIF-2α impairs functional recovery after cerebral ischemia, which may be beneficial predominantly in the early phase after stroke, indicating a timely regulated activation-inhibition of hypoxia-regulated cytoprotective and damaging factors which may be important for the functional outcome after stroke [85].

Another example of mutuality that may promote proapoptotic genes in response to hypoxia is p53/HIF-1α interaction [113,114,115]. Specific disruption of this interplay (attributable to the lack of HIF-1α) leads to downregulation and loss of expression of genes that promote cell death [84,98].

Reperfusion of ischemic areas can exacerbate ischemic brain damage through the generation of ROS/RNS (e.g., O₂^•−, hydroxyl radicals and ONOO⁻) by excessive production of both NO^• and O₂^•− by nNOS and iNOS. It has been reported that excess of NO^• immediately reacts with O₂^•− to form ONOO⁻ [120,121], which is responsible for the nitration of both free and protein bound tyrosine residues which are known to disrupt cell signaling cascades leading to tissue injury [120,121]. Therefore, enhanced degradation of ROS/RNS with pharmacological agents has been found to limit the extent of brain damage following stroke-induced excess of NO^• generation. Elimination of free radicals attenuates cytokine formation that drives up regulation of inflammatory proteins including iNOS.

On the other hand ROS provides a redox signal for hypoxic HIF-1α activation [122]. Assuming, that under specific pathological conditions, increased activity of HIF-1α is responsible for ischemia-induced detrimental effects, application of anti-oxidants eliminates ROS and consequently reduces HIF-1α levels by its destabilization and loss of transcriptional activity [123].

Under normal conditions, endogenous protective enzyme systems, such as superoxide dismutase (SOD) and reduced glutathione (GSH), limit the overproduction of free radicals. However, their capacities may be overwhelmed by pathological changes when free radical generation is uncontrolled. Removal of pathologically produced free radicals may be regarded as a viable approach to neuroprotection and may be achieved by scavenging or trapping free radicals. Among the compounds that possess free radical scavenging properties are nitrones, thiols, iron chelators, phenols and catechols. Tirilazad, ebselen and edaravone are examples of compounds with scavenging activity, and compounds with free radical trapping properties include NXY-059 and NSP-116.

Edaravone (Radicut) is a free radical scavenger marketed in Japan to treat acute ischemic stroke [124]. Edaravone scavenged •HO, NO^• and ONOO⁻ in a concentration-dependent manner [125,126,127]. The efficacy of the drug ranges from large clinical improvements to modest effects measured with standard stroke scales when administered up to 72 h following ischemic stroke [128,129].

Tirilazad mesylate, a lazaroid, has been investigated as a neuroprotective agent in patients after acute ischemic stroke. The compound was ineffective in treating acute ischemic stroke [130,131]. Another seleno-organic compound, ebselen, was shown to reduce delayed ischemic neurological deficits after SAH [132] and to improve outcomes after stroke [133,134].

L-arginine, a substrate for NO^• biosynthesis, is an obvious candidate of choice to improve NO bioavailability. The administration of L-arginine increased regional blood flow and prevented tissue damage in the rat MCAO model [135,136,137]. However, contradictory results with L-arginine administration were obtained that showed a beneficial effect [135,136,137], no effect [138] and the potentiation of pathological changes [139,140]. These effects may result from the fact that L-arginine enhances NO^• synthesis via all three isoforms [141]; therefore, synthesis of potentially detrimental NO^• from iNOS or nNOS may counteract the neuroprotective effects of eNOS activation. The ability of L-arginine to stimulate the release of hormones such as insulin [142], glucagon [143], growth hormone [144] or catecholamines [145] and its ability to transform into toxic byproducts such as polyamines or agmatine [146,147,148,149] are also limitations of its use.

NO^• donors, such as organic nitrates, sodium nitroprusside (SNP), sydnonimines, S-nitrosothiols, NONOates and hybrid donors potentiate NO^• production and thus may potentially be regarded as treatment strategies.

Administration of DETA/NONOate [150], SNP [151,152], 3-morpholinosydnonimine [153,154], ZJM-289 [155] or LA-419 [156,157] significantly increased cell proliferation and/or reversed ischemia-induced tissue damage in the selected structures of the rat brain. Stimulation of eNOS activity with concomitant suppression of nNOS and iNOS function was observed [155,156,157]. Beyond this point, neurovascular toxicity instead of vasodilation or infarct size reduction was observed [152].

These results from preclinical studies were confirmed in clinical observations in which SNP reduced the mean arterial blood pressure and led to improvements in CBF in patients after acute stroke [158,159].

The mechanism of the neuroprotective action of NO^• donors includes their abilities to reduce oxidative stress both in the brain and blood, inhibit the expression and activities of MMPs, scavenge NO^• or quench ROS, reduce inflammation, exert antiplatelet effects and attenuate I/R-mediated increases in ICAM-1 and E-selectin mRNA expression [160,161].

Suppression of the hydrolysis of L-arginine into ornithine and urea by arginases increased the production of NO^• and prevented the development of endothelial dysfunction [162], similar to the effects of eliminating asymmetric dimethylarginine (ADMA)—endogenous NOS inhibitor, through the stimulation of ADMA-hydrolyzing enzyme (DDAH II) [163].

The available data concerning the efficacy of the enhancement of NO^• production-based therapies are collected in Table 1.

Selective inhibition of nNOS and iNOS activities is thought to counteract the aftermath of stroke.

The neuroprotective abilities of N(ω)-nitro-L-arginine methyl ester (L-NAME) and 7-nitroindazole (7-NI), widely used NOS/nNOS inhibitors, are time-dependent and are observed in the early stages of ischemic insult [164]. The choice of dose is important, as the compounds prevent at low doses but potentiate ischemia-induced neurodegeneration at high doses [165]. Long-term inhibition of NOS might be too risky because of the off-target effects on eNOS, particularly in patients with cardiovascular risk or metabolic diseases.

Direct or indirect inhibition of iNOS through selective inhibitors, such as aminoguanidine (AG) or the BH₄ rate-limiting enzyme GTP cyclohydrolase I, attenuated cerebral infarction, ischemia-induced pathologies and prevented the progression of cerebral aneurysms [166,167,168].

The available data concerning the efficacy of NOS inhibitor-based therapies are collected in Table 1. For review please see also: [164].

HIF-1α is an essential component in changing the transcriptional repertoire of tissues during oxygen deprivation and plays a pivotal role in the regulation of iNOS activity. Thus, HIF-1α and the genes regulated by it have been the center of intense research. A growing number of pre-clinical studies in rodents suggests that the activation of HIF-1α signaling pathway prior or shortly after ischemic stroke reduces tissue damage and increases functional recovery from ischemic stroke [87,96,186,187,188,189]. One example of an agent that stabilizes the transcriptional activator HIF-1α and activates target genes involved in compensation for ischemia are inhibitors of HIF-1α prolyl hydrolases (PHD1, PHD2 and PHD3) [190]. The beneficial effects of PHD inhibition after ischemia require the activity of HIF-1α as shown in in vitro in oxygen glucose deprivation model of ischemia and in vivo in MCAO model in mice [96]. Dimethyloxalylglycine (DMOG) enhanced the activation of HIF-1α and enhanced transcription of the HIF-regulated genes. In vivo the infarct size, activation of pro-apoptotic proteins and behavioral deficits after stroke were reduced. The effect of DMOG was decreased after inhibition of HIF-1α with digoxin [96]. Similar effects were observed with small molecule hypoxia mimics, such as deferoxamine, cobalt chloride or GSNO. Additionally, these agents increased the expression of HIF-1α target genes [92,93,191]. Cocaine, andrographolide or vitamin E activated the HIF-VEGF pathway, thus increasing microvascular density, restoring local blood flow and protecting the brain from ischemic insults [192,193,194]. However, available data indicate that sustained and prolonged activation of the HIF-1α pathway may lead to a transition from neuroprotection to neurodegeneration, reflecting the dual features of the HIF system [195,196,197], which should be taken into consideration when considering therapy to modify HIF-1α. Some contradictory results indicated that inhibition of HIF-1α improved brain function in ischemia-reperfusion brain injury-related disorders [105,198]. Inhibition of free radical formation, followed by inhibition of HIF-1α activation, apoptosis formation, neutrophil activation and iNOS expression resulted in reduction in the infarct volume in ischemia-reperfusion brain injury (MCAO model) in rats [105]. In rat model of transient MCAO the early inhibition of HIF-1α by application of inhibitors or small interfering RNA reduced infarct size and BBB hyperpermeability, decreased mortality and improved neurological deficits through inhibition of HIF-1α activity [198,199]. The studies of Hsiao et al. showed that the administration of PMC [105].

The therapeutic strategies might lead to pleiotropic activation of HIF signaling through all cell types in CNS, making it difficult to draw conclusions about the significance of the HIF signaling pathway in the treatment of ischemia-reperfusion injury. However, the stimulation of the non-ischemic penumbra regions to initiate HIF-1α (e.g., by inhibition of PHD) and subsequent downstream induction of HIF-1α mediated antiapoptotic, vascular and glycolytic metabolic changes before the area is enveloped by the spreading ischemia may be one of the proposed therapeutic HIF-dependent effect. Additionally, inhibition of endothelial HIF-1α warrants further investigation as a therapeutic target for the treatment of stroke patients with diabetes [99].

The available data concerning the efficacy of HIF-1α-based therapies are collected in Table 2.

At present, it seems that combination therapies are more effective for treating stroke than monotherapies. Scavenging reactive oxygen species and concomitantly inhibiting NO^• synthesis by administering statins with resveratrol, an approved antioxidant or with nifedipine, a calcium channel blocker, is just one example of combination therapy for stroke [226]. Administration of resveratrol upregulated antioxidant enzyme activities, decreased O₂^•− production by downregulating NADPH oxidase activity and attenuated oxidative stress-mediated eNOS uncoupling. Some hope may be placed in combination therapy, which could allow minimization of the dose of NMDA antagonists by administering a compound enhancing its effect, such as caffeine [227] or cordycepin (an adenosine analog) [228]. The use of different agents at the same time enables avoidance of detrimental consequences induced by undesired side effects and reinforcement of beneficial impacts by mutual action. The examples of effective combined treatment therapies in clinical trials are collected in Table 3.