Amelioration of brain damage after focal ischemia in the rat by a novel inhibitor of lipid peroxidation

Rifampicin attenuates the MPTP-induced neurotoxicity in mouse brain

Rifampicin, an antibacterial drug, is highly effective in the treatment of tuberculosis and leprosy. Recently, it has been reported to have neuroprotective effects in in vitro and in vivo models. This study was designed to elucidate its neuroprotective effects against 1-methyl-4-phenyl-1,2,3,6-tetrahydropyridine (MPTP)-induced neurotoxicity (known as an in vivo mouse model of Parkinson's disease). Mice were injected intraperitoneally (i.p.) with MPTP (10 mg/kg) four times at 1-h intervals, and brains were analyzed 3 or 7 days later. Rifampicin at 20 mg/kg (i.p., twice) had protective effects against MPTP-induced neuronal damage (immunohistochemical changes in tyrosine hydroxylase) in both the substantia nigra and striatum. Rifampicin also protected against the MPTP-induced depletions of dopamine, 3,4-dihydroxyphenylacetic acid (DOPAC), and homovanillic acid (HVA) in the striatum. The maximal concentrations of rifampicin between 30 and 240 min after a single rifampicin injection (20 mg/kg, i.p.) were 2.6 microM (at 30 min) in plasma and 0.77 microM (at 60 min) in striatum. Next, the effects of rifampicin on oxidative stress [lipid peroxidation in mouse brain homogenates and free radical-scavenging activity against diphenyl-p-picrylhydrazyl (DPPH)] were evaluated to clarify the underlying mechanism. At 1 microM or more, rifampicin significantly inhibited both lipid peroxidation in the striatum and free radical production. These findings suggest that in mice, rifampicin can reach brain tissues at concentrations sufficient to attenuate MPTP-induced neurodegeneration in the nigrostriatal dopaminergic neuronal pathway, and that an inhibitory effect against oxidative stress may be partly responsible for its observed neuroprotective effects.

Nitrone derivatives of trolox as neuroprotective agents

Synthesis of nitrone derivatives of trolox is described. Their biological evaluation was performed in vitro for scavenging different free radicals, inhibiting Fe(2+)-induced lipid peroxidation, and in vivo in a permanent middle cerebral artery occlusion model in mice. New compounds exert pharmacological activities comparable to or better than those of trolox or nitrone-type reference compounds.

Minocycline inhibits oxidative stress and decreases in vitro and in vivo ischemic neuronal damage

The neuroprotective effects of minocycline-which is broadly protective in neurologic-disease models featuring cell death and is being evaluated in clinical trials-were investigated both in vitro and in vivo. For the in vivo study, focal cerebral ischemia was induced by permanent middle cerebral artery occlusion in mice. Minocycline at 90 mg/kg intraperitoneally administered 60 min before or 30 min after (but not 4 h after) the occlusion reduced infarction, brain swelling, and neurologic deficits at 24 h after the occlusion. For the in vitro studies, we used cortical-neuron cultures from rat fetuses in which neurotoxicity was induced by 24-h exposure to 500 microM glutamate. Furthermore, the effects of minocycline on oxidative stress [such as lipid peroxidation in mouse forebrain homogenates and free radical-scavenging activity against diphenyl-p-picrylhydrazyl (DPPH)] were evaluated to clarify the underlying mechanism. Minocycline significantly inhibited glutamate-induced cell death at 2 microM and lipid peroxidation and free radical scavenging at 0.2 and 2 microM, respectively. These findings indicate that minocycline has neuroprotective effects in vivo against permanent focal cerebral ischemia and in vitro against glutamate-induced cell death and that an inhibition of oxidative stress by minocycline may be partly responsible for these effects.

Antioxidant LY231617 enhances electrophysiologic recovery after global cerebral ischemia in dogs

OBJECTIVE

The potent antioxidant LY231617 (2,6-bis(1,1-dimethylethyl)-4-[[(1-ethyl)amino]methyl]phenol hydrochloride) is cytoprotective in models of focal and global cerebral ischemia. We tested the hypothesis that administration of LY231617, before the insult, would improve recovery of cerebral electrical activity and metabolic function after transient global cerebral ischemia by improving cerebral blood flow (CBF) during the reperfusion period.

DESIGN

Randomized, controlled, prospective study.

SETTING

Research laboratory at a university teaching hospital.

SUBJECTS

Twenty-four male beagle dogs.

INTERVENTIONS

All experiments were performed under pentobarbital anesthesia and controlled conditions of normoxia, normocarbia, and normothermia. Twelve control dogs received 20 mL/kg saline (vehicle) bolus into the right atrium and 0.01 mL/kg/min i.v., beginning 20 mins before 13 mins of global cerebral ischemia (by aortic occlusion). The dogs in the drug-treated group received LY231617 as a 10-mg/kg bolus 20 mins before ischemia and 5 mg/kg/hr throughout reperfusion (n = 12). CBF was measured using radiolabeled microspheres.

MEASUREMENTS AND MAIN RESULTS

Total CBF, cerebral oxygen consumption, and somatosensory evoked potentials (SEP) were measured during 240 mins of reperfusion. CBF was similar in both vehicle- and LY231617-treated animals at baseline and throughout the experimental period. In all animals, SEP became isoelectric between 60 and 100 secs after cross-clamping of the ascending aorta. SEP amplitude recovery was significantly higher in drug-treated animals compared with controls (73%+/-15% vs. 39%+/-14% [mean+/-SEM] from baseline at 120 mins [p<.05] and 86%+/-12% vs. 49%+/-14% from baseline at 240 mins [p< .05]).

CONCLUSIONS

LY231617 improves recovery of cerebral electrical function after complete transient global ischemia via mechanisms unrelated to cerebral circulatory effects.

Usual and unusual methods for detection of lipid peroxides as indicators of tissue injury in cerebral ischemia: what is appropriate and useful?

1. Free radical-dependent lipid peroxidation processes have long been thought to contribute to brain damage following stroke or cerebral ischemia/reperfusion. 2. The preponderance of evidence for this belief has been derived indirectly, through diminution of tissue injury indices (e.g., brain infarct volume) facilitated by application of free radical scavenger substances. 3. Direct, unequivocal evidence for lipid peroxidation in terms of classical assays (detection of conjugated diene absorbance or thiobarbituric acid-reactive substances) is considerably less common, and its validity can be questioned. 4. Correlations of treatment-induced diminishment of brain injury indices with reductions in lipid peroxidation level are rarer still. 5. Reasons underlying the disparity between the belief that lipid peroxidation contributes to ischemic brain injury and direct evidence for this contribution (at least acutely) are proposed, along with evidence that new methods are being developed which should provide the basis for obtaining a definitive answer.

Sensorimotor impairments in rats with cerebral infarction, induced by unilateral occlusion of the left middle cerebral artery: strain differences and effects of the occlusion site

Enormous differences exist between rat strains with respect to the infarct volume induced by unilateral middle cerebral artery (MCA) occlusion. We performed three experiments to address the following questions. Firstly, whether the pattern of MCA-occlusion (MCA-O) induced sensorimotor impairments in rats are strain dependent; secondly, whether proximal (i.e., close to its origin) and distal occlusions (above the lenticulostriate branch) of the MCA affect infarct volume and the behavioral impairments to a different extent; and thirdly, whether there is a relationship between the infarct volume and behavioral deficits. We found that the pattern of sensorimotor malfunctions induced by proximal unilateral MCA-O were highly strain dependent. Of the eight strains tested, Winkelmann-Wistar rats, Spontaneously Hypertensive Stroke-Prone rats, and Wistar-Kyoto rats were most severely affected. By contrast, Brown-Norway rats showed only mild behavioral deficits after the MCA-O. The second experiment confirmed that proximal occlusions induced slightly more behavioral malfunctions than distal occlusions did. Histological evaluation of the brain damage caused by proximal and distal MCA-O, confirmed that distal MCA-O damaged nearly exclusively cortical areas, and spared the caudate/putamen. An exploratory analysis of the relationship between infarct volume and behavioral deficits did not indicate that the severity of sensorimotor malfunctions can be predicted from the size of the infarct.

Effects of LY231617 and angiotensin IV on ischemia-induced deficits in circular water maze and passive avoidance performance in rats

The antioxidant LY231617 has previously been shown to offer significant protection against postischemic cell death in the hippocampus and corpus striatum of rats. The present results extend this observation by demonstrating a concomitant protection against the spatial memory deficits that accompany damage to the hippocampus, as measured by the circular water maze task. These animals were further tested for changes in associative memory by employing a passive avoidance conditioning task. No deficits in passive avoidance conditioning were measured among the 4-vessel occlusion animals treated with LY231617 or vehicle. However, the intracerebroventricular injection of angiotensin IV (Ang IV) immediately prior to foot-shock conditioning improved retention of the conditioned response during the subsequent 2-day period. These results suggest that LY231617 can offer considerable protection against global ischemia-induced cell death in the hippocampus with resulting preservation of spatial memory abilities. In addition, untreated animals that suffered cell losses in the hippocampus remained capable of responding to the facilitory effect of centrally administered Ang IV on a non-spatial memory task. The hypothesized mechanisms of the protection characteristics of LY231617, and the nootropic effect of Ang IV, are discussed.

Sensorimotor impairments in Wistar Kyoto rats with cerebral infarction, induced by unilateral occlusion of the middle cerebral artery: recovery of function

Wistar Kyoto (WKY) rats with cerebral infarction induced by permanent unilateral occlusion of the middle cerebral artery (MCA) and sham-operated rats were tested in a series of simple behavioral test 2, 16 and 37 days after surgery. In addition, the motility of the animals was measured over a period of 62 h, after the third test series. A subset of the tests appeared to be suitable to assess the effects of cerebral infarction, namely, grasping reflex of contralateral hindpaw, circling behavior, forelimb flexion, hindlimb flexion, and latency to fall off a square bridge. Except for the impaired grasping reflex of the contralateral hindpaw, there was spontaneous complete recovery of function by the third test session, 37 days after surgery. Some of the other tests might not have been sensitive enough to detect the effects of the unilateral MCA-occlusion (MCA-O) on behavior. Moreover, the WKY rats were very inactive in some of the tests, so that reliable scoring of the effects was not always possible. A rat strain other than the WKY strain might be more suitable to study the behavioral consequences of MCA-O.

Suppressive effects of tranilast on pulmonary fibrosis and activation of alveolar macrophages in mice treated with bleomycin: role of alveolar macrophages in the fibrosis

We have reported that tranilast, an anti-allergic drug that inhibits chemical mediator release from mast cells, suppresses bleomycin (BLM)-induced pulmonary fibrosis in mice through mechanisms other than inhibiting chemical mediator release from mast cells. The purpose of this paper is to examine the effect of tranilast on alveolar macrophage (AM) activation and on the development of fibrosis in ICR mice instilled with BLM intratracheally. Twenty eight days after the BLM instillation (0.01 mg/mouse), AM often migrated into alveolar spaces surrounding the fibrotic areas. Flow cytometry analysis for the size and density of AM (MAC-1 positive cells) suggested that AM were activated not only in the earlier acute inflammatory phase, but also in the later chronic phase. The p.o. administration of tranilast suppressed an increase of AM activity to produce reactive oxygen species in BLM-instilled mice, and it inhibited the subsequent development of pulmonary fibrosis. In vitro treatment with tranilast suppressed the reactive oxygen species production from murine peritoneal macrophages. However, several different anti-oxidants failed to inhibit the development of fibrosis. These results suggest that the activation of AM plays an important role in the development of fibrosis, and it is likely that tranilast suppresses fibrosis by inhibiting AM activation but not by scavenging reactive oxygen species.

The glycine antagonist and free radical scavenger 7-Cl-thio-kynurenate reduces CA1 ischemic damage in the gerbil

We examined whether 7-Cl-thio-kynurenate, a potent antagonist at the glycine site of the N-methyl-D-aspartate receptor which also inhibits lipid peroxidation, protected CA1 pyramidal cells following transient forebrain ischemia. Global ischemia was produced in anesthetized gerbils by 5 min bilateral carotid artery occlusion; hippocampal injury was assessed seven days later. 7-Cl-thio-kynurenate (100 mg/kg, i.p. x 5) dramatically attenuated ischemia-induced CA1 cell loss (from 95 +/- 1 to 7 +/- 3%): the protection was associated with a delayed and marked reduction in the animals' temperature. However, when the gerbils were maintained normothermic for at least 360 min, 7-Cl-thio-kynurenate still provided partial (54 +/- 11%) but significant protection. No protection was observed when a reduction in temperature with a time course similar to that caused by 7-Cl-thio-kynurenate was experimentally induced in saline-treated ischemic animals. In situ hybridization revealed that expression of NMDA-R1, a subunit of the N-methyl-D-aspartate receptor, was selectively reduced in CA1 seven days following global ischemia. In ischemic gerbils treated with 7-Cl-thio-kynurenate, protected CA1 cells were still able to express normal amounts of NMDA-R1 messenger RNA. Our results demonstrate that 7-Cl-thio-kynurenate, a glutamate receptor blocker possessing radical scavenger properties, is effective in reducing CA1 hippocampal damage following global ischemia in the gerbil. Since there is growing evidence that a positive feedback interaction between activation of glutamate receptors and free radical formation may be responsible for the generation of ischemic brain damage, drugs capable of interfering with both pathogenic mechanisms may be useful in preventing post-ischemic neuronal death.

Neuronal protection with superoxide dismutase in repetitive forebrain ischemia in gerbils

The underlying mechanism for severe damage with repetitive ischemia is not fully understood. Because of prolonged periods of reperfusions between the brief insults, we speculated that the severe damage may be secondary to excessive generation of oxygen free radicals. In this study we tested the efficacy of peg-superoxide dismutase (SOD) in a model of repeated ischemia in gerbils. Superoxide dismutase (SOD) or vehicle (saline) was delivered through osmotic pumps into the lateral ventricles continuously from the onset of the insult until the gerbils were sacrificed 6 days later. Three doses of SOD were used in the experiments (110, 150, and 190 units per microliter). Damage was assessed using a 0-4 point scoring system and statistical comparisons were done using the Mann-Whitney U-test. There was significant protection in the hippocampus (p < 0.05), striatum (p < 0.001), and substantia nigra reticulata (p < 0.05) in the lowest dose SOD-treated group (110 units per microliter). Animals treated with 150 units showed lesser (but significant) protection in the thalamus, medial geniculate nucleus, and striatum. In the animals treated with the higher dose of SOD (190 units per microliter), the extent of damage was no different than vehicle-treated controls in the cortex, striatum, and hippocampus. Compared to controls, neuronal damage was, however, significantly more severe in the medial geniculate nucleus and the thalamus in the high-dose SOD-treated animals (p < 0.05). Our experiments suggest that the SOD may have a small therapeutic window. Higher doses may either have no neuroprotective effects or may be harmful.

Effect of transient reduction of cerebral blood flow on membrane anisotropy and lipid peroxidation in different rat brain areas

Light-microscopical studies revealed that oligemic hypoxia for 24 and 60 min as produced by bilateral clamping of the carotid arteries (BCCA) in normotension does not produce neuronal cell necrosis in the vast majority of rat brain. Less than 5% of cases showed a pattern of mild selective neuronal necrosis as would be expected in ischemia. However, significant changes in both lipid peroxidation (as measured by MDA formation) and membrane anisotropy (measured by DPH or TMA-DPH, respectively, as a fluorescence probe) in cortical and striatal, but not in hippocampal, membrane fractions could be measured in ex vivo studies. Twenty-four and 60 min of BCCA without reperfusion decreased lipid peroxidation in the cerebral cortex but not in the striatum. BCCA, either for 24 or 60 min, and 60 min of reperfusion produced no changes in lipid peroxidation in either structure. However, 24 and 60 min of BCCA followed by 14 days of reperfusion led to a significant increase in MDA formation in the striatum, while lipid peroxidation in the cortex was only increased after 60 min of BCCA. Cortical as well as striatal membrane anisotropy increased significantly 14 days later in rats submitted to BCCA for 24 or 60 min. The study shows an increased lipid peroxidation 2 weeks after a transient reduction in cerebral blood flow although no neuronal necrosis could be observed in general.

Superoxide dismutase, catalase, and U78517F attenuate neuronal damage in gerbils with repeated brief ischemic insults

Repeated ischemic insults at one hour intervals result in more severe neuronal damage than a single similar duration insult. The mechanism for the more severe damage with repetitive ischemia is not fully understood. We hypothesized that the prolonged reperfusion periods between the relatively short ischemic insults may result in a pronounced generation of oxygen free radicals (OFRs). In this study, we tested the protective effects of superoxide dismutase (SOD) and catalase (alone or in combination), and U78517F in a gerbil model of repetitive ischemia. Three episodes (two min each) of bilateral carotid occlusion were used at one hour intervals to produce repetitive ischemia. Superoxide dismutase and catalase were infused via osmotic pumps into the lateral ventricles. Two doses of U78517F were given three times per animal, one half hour prior to each occlusion. Neuronal damage was assessed 7 days later in several brain regions using the silver staining technique. The Mann-Whitney U test was used for statistical comparison. Superoxide dismutase showed significant protection in the hippocampus (CA4), striatum, thalamus and the medial geniculate nucleus (MGN). Catalase showed significant protection in the striatum, hippocampus, thalamus, and MGN and the substantia nigra reticulata. Combination of the two resulted in additional protection in the cerebral cortex. Compared to the controls, there was little protection in a dose of 3 mg/kg of U78517F. There was significant protection with a dose of 10 mg/kg in the hippocampus (CA4), striatum, thalamus, medial geniculate nucleus and the substantia nigra reticulata.(ABSTRACT TRUNCATED AT 250 WORDS)

Brain angiotensin receptor subtypes in the control of physiological and behavioral responses

This review summarizes emerging evidence that supports the notion of a separate brain renin-angiotensin system (RAS) complete with the necessary precursors and enzymes for the formation and degradation of biologically active forms of angiotensins, and several binding subtypes that may mediate their diverse functions. Of these subtypes the most is known about the AT1 site which preferentially binds angiotensin II (AII) and angiotensin III (AIII). The AT1 site appears to mediate the classic angiotensin responses concerned with body water balance and the maintenance of blood pressure. Less is known about the AT2 site which also binds AII and AIII and may play a role in vascular growth. Recently, an AT3 site was discovered in cultured neoblastoma cells, and an AT4 site which preferentially binds AII(3-8), a fragment of AII now referred to as angiotensin IV (AIV). The AT4 site has been implicated in memory acquisition and retrieval, and the regulation of blood flow. In addition to the more well-studied functions of the brain RAS, we review additional less well investigated responses including regulation of cellular function, the modulation of sensory and motor systems, long term potentiation, and stress related mechanisms. Although the receptor subtypes responsible for mediating these physiologies and behaviors have not been definitively identified research efforts are ongoing. We also suggest potential contributions by the RAS to clinically relevant syndromes such as dysfunctions in the regulation of blood flow and ischemia, changes in cognitive affect and memory in clinical depressed and Alzheimer's patients, and angiotensin's contribution to alcohol consumption.

CGS-19755 is neuroprotective during repetitive ischemia: this effect is significantly enhanced when combined with hypothermia

In small animals the damaging effects of repetitive ischemia are more severe than a single insult of similar duration. Prolonged release of glutamate may correlate with the degree of damage. We report the protective effects of CGS-19755 (an N-methyl-D-aspartate receptor blocker), hypothermia or CGS-19755 in combination with mild hypothermia, in a gerbil model of repetitive ischemia. We used 3 min of forebrain ischemia and repeated it for a total of three times as 1-h intervals. Damage was assessed seven days after the insult. In the group where only CGS-19755 was used, significant neuronal protection was evident in the hippocampus (CA1 and CA3), striatum, and medial geniculate nucleus. With hypothermia significantly less damage was seen in the cerebral cortex, hippocampus (CA1 and CA4), and substantia nigra reticulata. When CGS-19755 was combined with mild hypothermia the effects of repetitive ischemia were completely abolished in all but one gerbil. Compared to hypothermia alone, significant protection was seen in the cerebral cortex, hippocampus (sibiculum, CA1 and CA4), striatum, medial geniculate nucleus, thalamus, and substantia nigra reticulata. The use of N-methyl-D-aspartate receptor blockers may protect the brain in repetitive ischemia. Combination with hypothermia may further enhance this protection.

The antioxidant LY231617 reduces global ischemic neuronal injury in rats

BACKGROUND AND PURPOSE

In the rat four-vessel occlusion model with 30 minutes of ischemia most agents have failed to be of benefit when given after ischemia. Because postischemia administration is more clinically relevant, we evaluated the antioxidant LY231617 (2,6-bis(1,1-dimethylethyl)-4-[[(1-ethyl)amino]methyl]phenol hydrochloride]) when administered after 30 minutes of four-vessel occlusion.

METHODS

Male Wistar rats were subjected to 30 minutes of four-vessel occlusion. LY231617 was either given orally 30 minutes before ischemia or intravenously beginning at 30 minutes after the onset of ischemia. Hippocampal CA1 layer and striatal damage were rated on a scale of 0-3 (0, no damage; 3, > 90% cell loss). We also evaluated the ability of LY231617 to prevent iron-dependent lipid peroxidation and to prevent hydrogen peroxide-induced neuronal death of hippocampal neurons in primary culture by exposing cultures to a 50-microM concentration of hydrogen peroxide for 15 minutes in the presence of LY231617.

RESULTS

Oral administration of LY231617 reduced both striatal and hippocampal CA1 damage by > 75% (p < 0.0001). In two separate experiments in which LY231617 was given intravenously beginning 30 minutes after occlusion, hippocampal and striatal damage were reduced by approximately 50% (p < 0.03) in the first experiment and by approximately 41% (p < 0.02) in the second experiment. Addition of 5 microM of LY231617 to primary hippocampal neuronal cultures antagonized the lethal effect of hydrogen peroxide (p < 0.05). Iron-dependent lipid peroxidation was also inhibited in a dose-related fashion.

CONCLUSIONS

The significant reduction of ischemia-induced or hydrogen peroxide-induced neuronal damage and inhibition of lipid peroxidation by LY231617 observed in this study suggest that reactive oxygen intermediates play an important role in the events leading to neuronal death after global ischemia/reperfusion.

Effects of baicalein and alpha-tocopherol on lipid peroxidation, free radical scavenging activity and 12-O-tetradecanoylphorbol acetate-induced ear edema

The effects of baicalein, a flavonoid, and alpha-tocopherol (vitamin E) on lipid peroxidation in rat forebrain homogenates, on free radical scavenging action against diphenyl-p-picrylhydrazyl (DPPH), and on 12-O-tetradecanoylphorbol acetate (TPA)-induced ear edema in mice were studied. Baicalein inhibited lipid peroxidation in forebrain homogenates, DPPH-induced free radical and TPA-induced ear edema as potently as did quercetin and nordihydroguaiaretic acid (NDGA), a lipoxygenase inhibitor, and more potently than BW755C, a mixed cyclooxygenase and lipoxygenase inhibitor. Lipid peroxidation in forebrain homogenates, DPPH-induced free radical and TPA-induced ear edema were also inhibited by alpha-tocopherol. Flavone showed no reaction. These results suggest that lipid peroxidation may play an important role in the pathogenesis of TPA-induced ear edema in mice.

A phenothiazine derivative reduces rat brain damage after global or focal ischemia

BACKGROUND AND PURPOSE

We previously reported that 2-(10H-phenothiazin-2-yloxy)-N,N-dimethylethanamine hydrochloride is a potent inhibitor of iron-dependent lipid peroxidation in vitro and can protect primary cultures of rat hippocampal neurons from hydrogen peroxide-induced toxicity. Because oxidants may play an important role in mediating postischemic tissue injury, we evaluated this agent in two rat models of transient cerebral ischemia.

METHODS

In a model of global forebrain ischemia, 23 male Wistar rats were subjected to 10 minutes of four-vessel occlusion followed by 72 hours of reperfusion. The rats received three intraperitoneal injections of either vehicle (2% aqueous acacia) or test agent (40 mg/kg). In a model of focal stroke, 19 spontaneously hypertensive rats were subjected to 2 hours of tandem middle cerebral and ipsilateral common carotid artery occlusion followed by 24 hours of reperfusion. The rats received three intraperitoneal injections of either vehicle (2% aqueous acacia) or test agent (40 mg/kg).

RESULTS

In the global model, the phenothiazine significantly protected the CA1 layer of the hippocampus, with a reduction in mean damage score from 2.1 +/- 0.3 for control rats to 1.0 +/- 0.4 for treated rats (p less than 0.05). In the transient focal stroke model, the compound reduced cortical infarct volume from 130.1 +/- 10.3 mm3 for control rats to 95.2 +/- 24.5 mm3 for treated rats (p less than 0.02).

CONCLUSIONS

Although the primary mechanism responsible for the protective effect is unclear at the present time, our study is consistent with the hypothesis that oxidant-mediated lipid peroxidation may be involved in the pathophysiology of postischemic brain injury.