Regional brain energy metabolism after complete versus incomplete ischemia in the rat in the absence of severe lactic acidosis

Intracellular activation of full-length human TREK-1 channel by hypoxia, high lactate, and low pH denotes polymodal integration by ischemic factors

TREK-1, a two-pore domain potassium channel, responds to ischemic levels of intracellular lactate and acidic pH to provide neuroprotection. There are two splice variants of hTREK1: the shorter splice variant having a shorter N-terminus compared with the full-length hTREK1 with similar C-terminus sequence that is widely expressed in the brain. The shorter variant was reported to be irresponsive to hypoxia-a condition attributed to ischemia, which has put the neuroprotective role of hTREK-1 channel into question. Since interaction between N- and C-terminus of different ion channels shapes their gating, we re-examined the sensitivity of the full-length as well as the shorter hTREK-1 channel to intracellular hypoxia along with lactate. Single-channel data obtained from the excised inside-out patches of the full-length channel expressed in HEK293 cells indicated an increase in activity as opposed to a decrease in activity in the shorter isoform. However, both the isoforms showed an increase in activity under combined hypoxia, 20mM lactate, and low pH 6 condition, albeit with subtle differences in their individual actions, confirming the neuroprotective role played by hTREK-1 irrespective of the differences in the N-terminus among the splice variants. Furthermore, E321A mutant that disrupts the interaction of the C-terminus with the membrane showed a decrease in activity with hypoxia indicating the importance of the C-terminus in the hypoxic response of the full-length hTREK-1. We propose an increase in activity of both the splice variants of hTREK-1 in combined hypoxia, high lactate, and low pH conditions typically associated with ischemia provides neuroprotection.

l-Lactate mediates neuroprotection against ischaemia by increasing TREK1 channel expression in rat hippocampal astrocytes in vitro

Brain ischaemia is a highly debilitating condition where shortage of oxygen and glucose leads to profuse cell death. Lactate is a neuroprotective metabolite whose concentrations increase up to 15-30 mmol/L during ischaemia and TREK1 is a neuroprotective potassium channel which is upregulated during ischaemia. The aim of this study was to investigate the effect of l-lactate on TREK1 expression and to evaluate the role of l-lactate-TREK1 interaction in conferring neuroprotection in ischaemia-prone hippocampus. We show that 15-30 mmol/L l-lactate increases functional TREK1 protein expression by 1.5-3-fold in hippocampal astrocytes using immunostaining and electrophysiology. Studies with transcription blocker actinomycin-D and quantitative PCR indicate that the increase in TREK1 expression is due to enhanced TREK1 mRNA transcription. We further report that l-lactate-mediated increase in TREK1 expression is via protein kinase A (PKA)-dependent pathway. This is the first report of an ischaemic metabolite affecting functional expression of an ion channel. Our studies in an in vitro model of ischaemia using oxygen glucose deprivation show that 30 mmol/L l-lactate fails to reduce cell death in rat hippocampal slices treated with TREK1 blockers, PKA inhibitors and gliotoxin. The above effects were specific to l-lactate as pyruvate failed to increase TREK1 expression and reduce cell death. l-Lactate-induced TREK1 upregulation is a novel finding of physiological significance as TREK1 channels contribute to neuroprotection by enhancing potassium buffering and glutamate clearance capacity of astrocytes. We propose that l-lactate promotes neuronal survival in hippocampus by increasing TREK1 channel expression via PKA pathway in astrocytes during ischaemia. Insufficient blood supply to the brain leads to cerebral ischaemia and increase in extracellular lactate concentrations. We incubated hippocampal astrocytes in lactate and observed increase in TREK1 channel expression via protein kinase A (PKA). Inhibition of TREK1, PKA and metabolic impairment of astrocytes prevented lactate from reducing cell death in ischaemic hippocampus. This pathway serves as an alternate mechanism of neuroprotection. Cover image for this issue: doi: 10.1111/jnc.13326.

Lactate modulates the intracellular pH sensitivity of human TREK1 channels

Tissue acidosis and high lactate concentrations are associated with cerebral ischaemia. The degree of acidosis is dependent on circulating glucose concentration, hyperglycaemia being associated with increased acidosis. Among other agents, lactate and protons have been shown to activate the leak potassium channel; TREK1 (TWIK related potassium channel 1) from the intracellular side and its increased activity is implicated in tolerance towards ischaemic cell damage. In the present study, we show that ischaemic concentrations of lactate (30 mM) at pH 7.0 and 6.5, commonly observed during ischemia, cause robust potentiation of human TREK1 (hTREK1) activity at single-channel level in cell-free inside-out membrane patches, while 30 mM lactate at pH 6.0 to 5.5, commonly observed during hyperglycaemic ischemia, reduces hTREK1 channel activity significantly. The biphasic effect of 30 mM lactate (ischaemic concentrations) on modulation of hTREK1 by varying pH conditions is specific since basal concentrations of lactate (3 mM) and 30 mM pyruvate at pH 7.0 and 5.5 failed to show similar effect as lactate. Experiments with deletion and point mutants of hTREK1 channel suggest that lactate changes the pH modulation of hTREK1 by interacting differently with the histidine residue at 328th position (H328) above and below its pKa (∼6.0) in the intracellular carboxyl-terminal domain of TREK1. This lactate-induced pH modulation of hTREK1 is absent in C-terminal deletion mutant, CTDΔ100, and is similar in E321A-hTREK1 mutant as in wild-type hTREK1 suggesting that it is independent of pH-sensitive glutamate residue at 321st position. Such a differential pH-dependent effect of lactate on an ion channel function has not been reported earlier and has important implications in different stages of ischaemia.

Anoxia-evoked intracellular pH and Ca2+ concentration changes in cultured postnatal rat hippocampal neurons

The ratiometric indicators 2',7'-bis-(2-carboxyethyl)-5-(and-6)-carboxyfluorescein and Fura-2 were employed to examine, respectively, intracellular pH (pHi) and calcium ([Ca2+]i) changes evoked by anoxia in cultured postnatal rat hippocampal neurons at 37 degrees C. Under both HCO3-/CO2- and HEPES-buffered conditions, 3-, 5- or 10-min anoxia induced a triphasic change in pHi consisting of an initial fall in pHi, a subsequent rise in pHi in the continued absence of O2 and, finally, a further rise in pHi upon the return to normoxia, which recovered towards preanoxic steady-state pHi values if the duration of the anoxic insult was < or = 5 min. In parallel experiments performed on sister cultures, anoxia of 3, 5 or 10 min duration evoked rises in [Ca2+]i which, in all cases, commenced after the start of the fall in pHi, reached a peak at or just following the return to normoxia and then declined towards preanoxic resting levels. Removal of external Ca2+ markedly attenuated increases in [Ca2+]i, but failed to affect the pHi changes evoked by 5 min anoxia. The latency from the start of anoxia to the start of the increase in pHi observed during anoxia was increased by perfusion with media containing either 2 mM Na+, 20 mM glucose or 1 microM tetrodotoxin. Because each of these manoeuvres is known to delay the onset and/or attenuate the magnitude of anoxic depolarization, the results suggest that the rise in pHi observed during anoxia may be consequent upon membrane depolarization. This possibility was also suggested by the findings that Zn2+ and Cd2+, known blockers of voltage-dependent proton conductances, reduced the magnitude of the rise in pHi observed during anoxia. Under HCO3-/CO2-free conditions, reduction of external Na+ by substitution with N-methyl-D-glucamine (but not Li+) attenuated the magnitude of the postanoxic alkalinization, suggesting that increased Na+/H+ exchange activity contributes to the postanoxic rise in pHi. In support, rates of pHi recovery from internal acid loads imposed following anoxia were increased compared to control values established prior to anoxia in the same neurons. In contrast, rates of pHi recovery from acid loads imposed during anoxia were reduced, suggesting the possibility that Na+/H+ exchange is inhibited during anoxia. We conclude that the steady-state pHi response of cultured rat hippocampal neurons to transient anoxia is independent of changes in [Ca2+]i and is characterized by three phases which are determined, at least in part, by alterations in Na+/H- exchange activity and, possibly, by a proton conductance which is activated during membrane depolarization.

Global ischemia-induced inhibition of the coupling ratio of calcium uptake and ATP hydrolysis by rat whole brain microsomal Mg(2+)/Ca(2+) ATPase

Ischemia is associated with a loss of cytosolic calcium homeostasis. Intracellular stores, particularly in endoplasmic reticulum, are critical for the maintenance of calcium homeostasis. Recent studies have shown that ischemia significantly inhibited microsomal calcium uptake mediated by Mg(2+)/Ca(2+) ATPase, the major mechanism of endoplasmic reticulum calcium sequestration. This study was initiated to determine whether the decreased calcium uptake caused by ischemia was the result of inhibition of Mg(2+)/Ca(2+) ATPase activity or an uncoupling of calcium uptake from ATP hydrolysis. The microsomal Mg(2+)/Ca(2+) ATPase specific inhibitor thapsigargin partially inhibited ATPase activity and completely inhibited calcium uptake. ATPase inhibited by thapsigargin was considered microsomal Mg(2+)/Ca(2+) ATPase. Ischemia from 5 to 60 min had no significant effect on thapsigargin sensitive ATPase activity. However, under identical conditions, increasing ischemia from 5 to 60 min significantly inhibited microsomal calcium uptake. Comparing calcium uptake to ATP hydrolysis as ischemia increased from 5 to 60 min revealed that the coupling ratio of calcium molecules sequestered to ATP molecules hydrolyzed became significantly decreased. The results demonstrated that the effect of ischemia on microsomal calcium uptake was mediated by an uncoupling of calcium transport from Mg(2+)/Ca(2+) ATPase activity.

Mechanisms of neuronal cell death in Wernicke's encephalopathy

Wernicke's Encephalopathy (WE) is a serious neurological disorder resulting from thiamine deficiency, encountered in chronic alcoholics and in patients with grossly impaired nutritional status. Neuropathologic studies as well as Magnetic Resonance Imaging reveal selective diencephalic and brainstem lesions in patients with WE. The last decade has witnessed major advances in the understanding of pathophysiologic mechanisms linking thiamine deficiency to the selective brain lesions characteristic of WE. Activities of the thiamine-dependent enzyme alpha-ketoglutarate dehydrogenase, a rate-limiting tricarboxylic acid cycle enzyme are significantly reduced in autopsied brain tissue from patients with WE and from rats treated with the central thiamine antagonist, pyrithiamine. In the animal studies, evidence suggests that such enzyme deficits result in focal lactic acidosis, cerebral energy impairment and depolarization resulting from increased release of glutamate in vulnerable brain structures. It has been proposed that this depolarization may result in N-Methyl-D-Aspartate receptor-mediated excitotoxicity as well as increased expression of immediate early genes such as c-fos and c-jun resulting in apoptotic cell death. Other mechanisms involved in thiamine deficiency-induced cell loss may involve free radicals and alterations of the blood-brain barrier. Additional studies are still required to identify the site of the initial cellular insult and to explain the predilection of diencephalic and brainstem structures due to thiamine deficiency.

Real-time measurement of ischemia-evoked glutamate release in the cerebral cortex of four and eleven vessel rat occlusion models

Interstitial levels of the neurotransmitter glutamate and cerebral blood flow changes were compared in two models of rat forebrain ischemia using the dialysis electrode technique and laser doppler flowmetry with brain temperature controlled. Ten-minute periods of cerebral ischemia were elicited by the four and an eleven vessel occlusion and compared to carotid artery transection. Elapsed time from the onset of ischemia to the ischemic plateau was 76.8+/-57.9 s in 4VO vs. 14.8+/-1.3 s in 11VO animals. Percent residual cerebral blood flow (CBF) was 13.5+/-8.8% during 4VO as opposed to 4.5+/-2.9% during 11VO. Concomitantly, cerebral glutamate levels rose to 255. 7+/-72.8 micromol l-1 in the 4VO animals in comparison with levels of 138.5+/-78.7 and 135.7+/-40.2 micromol l-1 in the 11VO and carotid transection animals. During the first 89.6+/-47.4 s of reperfusion, glutamate levels rose to a second higher peak of 315. 1+/-179.2 micromol l-1 in 7 of 12 animals. Following reperfusion, glutamate levels in the 4VO and 11VO animals returned towards basal levels. This study demonstrates that 11VO causes a rapid drop in CBF to near zero levels, better mimicking complete forebrain ischemia than the traditional 4VO technique. Moreover, the 'low flow' state of cerebral ischemia, produced by traditional 4VO, results in a higher interstitial level of glutamate than a 'no flow' state, as exhibited by the 11VO technique. The dialysis electrode, used simultaneously with laser doppler flowmetry, real-time data acquisition, and continuous brain temperature control, in this new rat model, provides real-time evidence that glutamate levels in the interstitial space are enhanced during a low flow state of cerebral ischemia. Furthermore, not before demonstrated, glutamate transients are seen to occur during the first 90 s of reperfusion, and, to the best of our knowledge, the glutamate levels recorded by this technique are the highest in the literature.

Brain lactate, not glucose, fuels the recovery of synaptic function from hypoxia upon reoxygenation: an in vitro study

Lactate has been considered for many years to be a useless, and frequently, harmful end-product of anaerobic glycolysis. In the present in vitro study, lactate-supplied rat hippocampal slices showed a significantly higher degree of recovery of synaptic function after a short hypoxic period than slices supplied with an equicaloric amount of glucose. More importantly, all slices in which anaerobic lactate production was enhanced by pre-hypoxia glucose overload exhibited functional recovery after a prolonged hypoxia. An 80% recovery of synaptic function was observed even when glucose utilization was blocked with 2-deoxy-D-glucose during the later part of the hypoxic period and during reoxygenation. In contrast, slices in which anaerobic lactate production was blocked during the initial stages of hypoxia did not recover their synaptic function upon reoxygenation despite the abundance of glucose and the removal of 2-deoxy-D-glucose. Thus, for brain tissue to show functional recovery after prolonged period of hypoxia, the aerobic utilization of lactate as an energy substrate is mandatory.

Relative phosphocreatine and nucleoside triphosphate concentrations in cerebral gray and white matter measured in vivo by 31P nuclear magnetic resonance

Rates of ATP metabolism generally are higher in cerebral gray matter compared to white matter. In order to study the physiology of this regional difference in vivo, the 1-dimensional chemical shift imaging technique (1D-CSI) was used to acquire 31P nuclear magnetic resonance spectra from 2.5 mm slices of 4-week old piglet brains. Spectra from predominantly gray matter slices (estimated 76% gray matter, 7 mm below the scalp) were compared to predominantly white matter slices (56% estimated white matter, 13 mm below the scalp) as assessed by magnetic resonance images. The 1D-CSI technique introduced no systematic changes in the ratio of signals from a single chamber phantom containing a phosphocreatine (PCr) and ATP solution. Gray matter slices showed a PCr/NTP ratio of 0.93 +/- 0.11 (mean +/- S.D.) using a 2 s interpulse interval, a value very close to the ratio in surface coil localized spectra. The predominantly white matter slices showed a PCr/NTP ratio of 1.32 +/- 0.18 (P < 0.02 for gray versus white matter). Using the estimated percentages of gray and white matter in the two slices and calculated concentrations from fully relaxed spectra, the gray matter PCr/NTP ratio is approximately 0.77, while the ratio in white matter is approximately 2.18. The difference in PCr/NTP measured in vivo suggests that either the total NTP concentration is higher or the steady state PCr concentration is lower in gray matter than in white matter in the piglet brain.

Global incomplete cerebral ischemia produces predominantly cortical neuronal injury

BACKGROUND AND PURPOSE

We determined the neuropathologic damage in a canine model of global incomplete ischemia commonly used in a variety of physiological experiments.

METHODS

We induced 20 minutes of incomplete ischemia in dogs (n = 9) by increasing intracranial pressure via intraventricular infusion of artificial cerebrospinal fluid to maintain a cerebral perfusion pressure of 10 mm Hg while keeping body temperature at 38 degrees C during and immediately after ischemia. After a 7-day recovery period, animals were perfusion-fixed for neuropathology. In hematoxylin and eosin preparations, ischemic neuronal injury was assessed, neurons were counted, and percentage of cell damage was determined.

RESULTS

No focal neurological deficits or overt seizures were observed during the 7-day recovery period. In superior temporal gyrus, 49 +/- 11% and 70 +/- 10% damage (mean +/- SEM) was observed in layer III pyramidal cells in the crown and sulcus, respectively. All neocortical regions examined showed neuronal damage in layers III and/or V. In hippocampus, 59 +/- 11% damage of pyramidal neurons occurred in CA1, with dorsal (septal) hippocampus showing more injury than ventral (temporal) portions. The caudate nucleus (head) exhibited 27 +/- 7% neuronal injury. In cerebellar cortex (anterior lobule), 70 +/- 7% damage of Purkinje cells occurred, but different folia of cerebellum showed varying degrees of injury. Brain stem and thalamus were minimally affected despite reduced blood flow. Inflammatory changes (leukocytic infiltration and neuronal incrustations) were observed, but only when neuronal degeneration was severe. Pancellular necrosis and infarction did not occur.

CONCLUSIONS

This animal model of ischemia causes reproducible neuronal injury primarily in cortical regions without pancellular necrosis and infarction. Damage to subcortical areas is less severe than to cortical areas, despite comparable reductions in regional cerebral blood flow. Therefore, in the presence of regionally uniform but incomplete cerebral ischemia, neocortical and hippocampal pyramidal neurons and cerebellar cortical Purkinje cells are more likely than subcortical neurons to degenerate; alternatively, pyramidal and Purkinje neurons degenerate before neostriatal neurons in this model. This neuronal degeneration may represent an intrinsic cellular mechanism without major contribution of cytotoxic pathways associated with inflammation.

Intracellular alkalosis during hypoxia in newborn mouse brain in the presence of systemic acidosis: a phosphorus magnetic resonance spectroscopic study

We investigated the in vivo changes in cerebral energy metabolism and pHi in newborn mice noninvasively during 8 h of hypoxia with FiO2 = 5%, using phosphorus magnetic resonance spectroscopy continuously. The intracellular brain pH (pHi) increased from 7.20 +/- 0.03 to 7.36 +/- 0.03 (P < 0.05) at 1 h of hypoxia and then decreased gradually. On the other hand, the mixed arterial and venous blood pH decreased gradually during hypoxia, reaching a minimum value of 7.16 +/- 0.01 at the end of the hypoxia. There was no significant difference in PCO2 between control (47.4 +/- 0.8 mm Hg) and 1-h hypoxic (49.0 +/- 1.1 mm Hg) mice. The blood glucose concentration was significantly increased at 1 h of hypoxia. These results indicate that the alkaline shift in pHi during hypoxia was caused neither by systemic alkalosis due to hypocapnia nor hypoglycemia.

Hypoglycemia prevents increase in lactic acidosis during reperfusion after temporary cerebral ischemia in rats

Sequential 31P and 1H MRS was used to measure cerebral phosphate metabolites, intracellular pH, and lactate in normoglycemic and hypoglycemic rats during 30 min of complete cerebral ischemia and 5.5 h of reperfusion. These results were correlated with brain levels of free fatty acids (FFAs), excitatory amino acids, cations, and water content at death. The lactate/N-acetyl aspartate ratio was not significantly different between groups before or during occlusion. During reperfusion, the ratio was higher in normoglycemic rats from 3 to 85 min (p < or = 0.05), and recovery time was faster in hypoglycemic rats (29 vs 45 min; p = 0.04), suggesting reduced lactate production and faster recovery of aerobic metabolism. During occlusion, significant but comparable decrease of intracellular pH occurred in each group. Intracellular pH was higher in hypoglycemic rats at 140 min and 260 min of reperfusion. Water content, Na and K+ concentrations, and FFA and excitatory amino acid levels were not significantly different between groups, but hypoglycemic rats had less depletion of levels of Mg2+ (p = 0.011). These results show that hypoglycemia has a limited but potentially beneficial effect on postischemic lactic acidosis.

Acute brain swelling in cerebral embolization model of rats. II: Biochemical study

To investigate the mechanism of acute brain swelling after cerebral embolism in rats, the authors evaluated the regional changes in tissue pH, ATP, blood-brain barrier function, and water content. The deep cerebrum--thalamus, hypothalamus, and hippocampus--had alkaline change in tissue pH accompanied with the disruption of the blood-brain barrier, decrease of ATP, and increase of water content. The cortex and the caudate putamen, on the other hand, did not show edema though acidic change and ATP decrease were seen. These results suggest that an early vasogenic edema in the deep cerebrum is a main factor for evolving acute brain swelling in this embolization model of rats.

Simultaneous 31P NMR spectroscopy and laser Doppler flowmetry of rat brain during global ischemia and reperfusion

The relationship between blood flow and metabolism was studied in halothane-anaesthetized, normothermic rats submitted to 30 min global ischemia by four-vessel occlusion. Phosphocreatine (PCr), ATP, intracellular pH and intracellular magnesium (pMg) were measured by 31P NMR spectroscopy, and blood flow by laser Doppler flowmetry. Prior to ischemia the PCr/ATP ratio of fully relaxed spectra was 2.4 +/- 0.3, intracellular pH was 7.26 +/- 0.15 and pMg was 3.26 +/- 0.13. Vascular occlusion led to complete cessation of blood flow in four out of eight rats, and to incomplete ischaemia (< 10% of control) in the other four animals. During vascular occlusion EEG flattened and energy metabolism broke down in all but one animal with a residual blood flow of 8% of control. pH declined to 6.70 +/- 0.08. The speed of electrophysiological and metabolic recovery after 30 min ischemia varied considerably from animal to animal. Variability depended mainly on the recirculation delay (i.e., the interval from vascular release to normalization of blood flow) but was independent of residual blood flow during ischemia, pre-ischemic glucose, ischemic or post-ischemic acidosis, or the degree of post-ischemic hypoperfusion. After 3 h recirculation PCr and intracellular pH returned to normal but pMg was slightly increased, and ATP was reduced by up to 50% in all animals except the rat with incomplete breakdown of energy metabolism during ischemia. The dissociation between PCr and ATP is attributed to a loss of total adenylate, the severity of which depends on the quality of post-ischemic recirculation.(ABSTRACT TRUNCATED AT 250 WORDS)

Post-ischemic brain tissue alkalosis suppressed by U74006F

We monitored chronically (for 1 week) the effect of the 21-aminosteroid U74006F, a potent lipid peroxidation inhibitor, on the pH profile of the rat brain following transient forebrain ischemia. Eight rats were treated initially with 3 mg/kg i.v. of U74006F 1 min after reperfusion. A second dose of 1.5 mg/kg i.v. was given 60 min after reperfusion. A vehicle group (n = 9) was treated in the same manner, using the same volume of the vehicle solution, 20 mM citric acid, 3 mM sodium citrate, and 8 mM NaCl. Statistically significant interaction between group and time (P = 0.003) was detected for pH. Brain pH of the vehicle treated animals were significantly higher than the U74006F treated group at 24 h (P = 0.009) and 48 h (P = 0.009) of reperfusion. Chronic post-ischemic brain tissue alkalosis at 24 h (pH 7.22 +/- 0.12) and 48 h (pH 7.25 +/- 0.11) post-ischemia, observed among the vehicle treated animals (and untreated animals), was suppressed by treatment with U74006F. These results suggest a coupling between post-ischemic brain tissue alkalosis and free radical induced lipid peroxidation.

A new method for producing temporary complete cerebral ischemia in rats

A new model of temporary complete cerebral ischemia was developed and tested in 64 rats. With use of microsurgical techniques, both pterygopalatine and external carotid arteries were occluded and the basilar artery was coagulated to reduce potential collateral CBF during ischemia. After this preliminary five-vessel occlusion, temporary global ischemia was induced by occluding the common carotid arteries (CCAs) with microclips. To validate the method, CBF was measured autoradiographically in 24 anatomical regions at death after 5 min of ischemia or after 15 min of ischemia followed by 5 min of reperfusion. Mean arterial blood pressure and arterial blood gases remained stable under controlled endotracheal ventilation and anesthesia (halothane, 70% N2O, and 30% O2) throughout the CBF experiments, except for a 10-15% increase in mean arterial blood pressure for 1-5 min after bilateral CCA occlusion. After the initial five-vessel occlusion, the EEG did not change, and local CBF levels were comparable to those in anesthetized non-surgical controls. When the CCAs were occluded, the EEG flattened rapidly; after 5 min of ischemia, autoradiography showed no detectable blood flow in the forebrain and cerebellum. The local CBF levels measured after 15 min of temporary global ischemia and 5 min of reperfusion demonstrated relatively homogeneous postischemic hyperperfusion; only two of eight rats had several 1- to 3-mm areas of no-reflow. Survival studies showed increasing motor impairment after 10, 15, 30, and 60 min of temporary CCA occlusion. Ischemic neuronal damage was observed histologically in the hippocampus and basal ganglia 24 h after 10 min of temporary ischemia.(ABSTRACT TRUNCATED AT 250 WORDS)

Age-related differences in recovery of blood flow and metabolism after cerebral ischemia in swine

We tested two hypotheses: 1) that cerebral blood flow, oxygen consumption, and evoked potentials recover to preischemic values at 120 minutes of reperfusion more completely in 1-2-week-old piglets than in 6-10-month-old pigs after complete ischemia; and 2) that recovery of cerebral blood flow, oxygen consumption, and electrical function in piglets and pigs at 120 minutes of reperfusion is better after incomplete than after complete ischemia. During 30 minutes of ischemia produced by intracranial pressure elevation, cerebral blood flow determined by the microspheres technique was decreased to 0-1 ml/min/100 g with complete ischemia, to 1-10 ml/min/100 g with severe incomplete ischemia, or to 10-20 ml/min/100 g with moderate incomplete ischemia. During reperfusion after complete ischemia, both piglets and pigs demonstrated hyperemia but delayed hypoperfusion occurred in more brain regions in pigs, oxygen consumption returned to preischemic values in piglets but not in pigs (70 +/- 10% of preischemic values), and evoked potentials recovered better in piglets than in pigs (24 +/- 4% and 9 +/- 4% of preischemic values, respectively). Both piglets and pigs had fewer brain areas with hyperemia and hypoperfusion and improved oxygen consumption and electrical function during recovery from incomplete than from complete ischemia. We speculate that piglets tolerate complete ischemia better than pigs because of decreased reperfusion injury and that both groups recover better from incomplete than complete ischemia because of improved substrate supply during ischemia.

Nuclear magnetic resonance spectroscopy study of human brain after cardiac resuscitation

We used 31P nuclear magnetic resonance spectroscopy to study the cerebral metabolic function of eight patients with severe postischemic anoxic encephalopathy secondary to cardiac arrest. Spectroscopy was performed at 18 +/- 13 and 64 +/- 20 hours after resuscitation. Glasgow Coma Scale scores at the time of initial and repeat spectroscopy were 3.6 +/- 1.2 and 3.5 +/- 1.2, respectively. In those patients whose spectra were of adequate quality to monitor pH, all demonstrated tissue alkalosis in at least one brain region. The mean brain pH at initial spectroscopy was 7.14 +/- 0.09 and was significantly alkalotic when compared with age- and sex-matched normal controls (pH = 6.98 +/- 0.04, p less than 0.0001). Five of the eight patients showed at least one region of persistent alkalosis at repeat spectroscopy, whereas one patient demonstrated severe acidosis with a pH of 6.42. Spectra demonstrated marked metabolic heterogeneity, ranging from normal in appearance to complete obliteration of all high-energy phosphates with only inorganic phosphate remaining.

Neurological protection by dichloroacetate depending on the severity of injury in the paraplegic rat

Hyperglycemia has been shown to exacerbate neurological deficit associated with central nervous system ischemia. Iodoacetate or dichloroacetate was administered intraperitoneally to rats in a study to examine the role of glycolysis in hyperglycemic exacerbation of neurological deficit. Sprague-Dawley rats were injected with saline, iodoacetate, or dichloroacetate and then made paraplegic by temporary occlusion for 10, 12, 13, or 15 minutes of the right and left subclavian arteries and the aorta distal to the left subclavian artery. Glycolytic blockage by iodoacetate was lethal in doses of 15 mg/kg or more, whereas rats receiving 10 mg/kg survived but showed no significant neurological improvement compared to the saline-treated control group. Dichloroacetate, 500 mg/kg, protected neurological function, which suggests a possible detrimental role for lactate accumulation and the benefit of maintaining tricarboxylic acid cycle activity by stimulating pyruvate dehydrogenase. The protection seen with dichloroacetate depended on the severity of ischemic injury. Dichloroacetate administration had a minimal effect on neurological outcome with occlusion periods of 13 and 15 minutes, mild improvement with 12 minutes of occlusion, and a significant protective effect with a 10-minute occlusion period. The dose-response nature of ischemic injury and neurological outcome in this rat model of paraplegia therefore appears to play an important role in determining the effect observed with a specific intervention.

Effect of plasma glucose concentration on cerebral metabolism during partial ischemia in neonatal piglets

We used neonatal piglets to determine the influence of plasma glucose concentration on cerebral energy metabolism during and immediately after partial ischemia. We assessed cerebral metabolism using in vivo phosphorus-31 magnetic resonance spectroscopy. Arterial plasma glucose concentration was increased in four piglets by systemic infusions of dextrose in water for comparison with infusions of saline in four controls or decreased in eight piglets by fasting for 24-48 hours for comparison with four fed piglets. Plasma glucose concentration showed a significant linear correlation with intracellular pH (r = -0.7, p less than 0.05). Piglets that developed hypoglycemia during partial ischemia had a smaller reduction in intracellular pH and a larger increase in inorganic phosphate content than piglets that were normoglycemic or hyperglycemic during ischemia. Similar differences persisted during the first 5 minutes of postischemic reperfusion. Subsequently, the cerebral concentrations of phosphorylated compounds returned to normal in all piglets. Our results demonstrate that 1) arterial plasma glucose concentration influences cerebral energy metabolism and intracellular pH during ischemia, 2) neonatal piglets can develop profound brain acidosis, and 3) brain acidosis during ischemia does not influence the restoration of cerebral phosphorylated compounds to control levels during the first 90 minutes after ischemia.

Chronic cerebral intracellular alkalosis following forebrain ischemic insult in rats

We measured cerebral intracellular pH using in vivo phosphorus-31 nuclear magnetic resonance spectroscopy during 1 week after forebrain ischemia or sham operation in eight and seven rats, respectively. Mean maximum pH was significantly higher (p less than 0.003) in the ischemic group than in the sham-operated group (7.34 +/- 0.03 and 7.19 +/- 0.02, respectively). The difference between mean maximum pH and baseline pH (7.08 +/- 0.01 in each group) was significantly greater (p less than 0.02) in the ischemic group than in the sham-operated group. In the ischemic group, alkalosis occurred primarily after 48-72 hours of recirculation. We speculate that brain tissue alkalosis occurring chronically after ischemia is associated with delayed ischemic neuronal death.

Rodent models of cerebral ischemia

The use of physiologically regulated, reproducible animal models is crucial to the study of ischemic brain injury--both the mechanisms governing its occurrence and potential therapeutic strategies. Several laboratory rodent species (notably rats and gerbils), which are readily available at relatively low cost, are highly suitable for the investigation of cerebral ischemia and have been widely employed for this purpose. We critically examine and summarize several rodent models of transient global ischemia, resulting in selective neuronal injury within vulnerable brain regions, and focal ischemia, typically giving rise to localized brain infarction. We explore the utility of individual models and emphasize the necessity for meticulous experimental control of those variables that modulate the severity of ischemic brain injury.

Evaluation of the 11CO2 positron emission tomographic method for measuring brain pH. II. Quantitative pH mapping in patients with ischemic cerebrovascular diseases

A practical method has been developed that, using 11CO2 and positron emission tomography (PET), computes and maps (a) "effective pH" (pHt), a weighted average of intra- and extracellular pH, and (b) "clearance" (K1), product of blood flow and 11CO2 extraction. This method, together with measurements of cerebral blood flow (CBF) and oxygen extraction fraction (OEF), was applied to 12 patients with cerebral ischemia or stroke. The regional K1 was positively correlated with CBF (n = +0.78). The k1/CBF ratio, representing the extraction fraction ratio of 11CO2 to H2 15O, was negatively correlated with CBF (r = -0.54), suggesting that 11CO2 extraction decreases as flow increases. In five acute stroke patients within 2 days of onset, the injured cortex had lower CBF (20.6 ml/min/100 g), higher OEF (78.1%), and lower pHt (6.96) than the contralateral cortex (CBF = 41.4 ml/min/100 g, OEF = 53.3%, pHt = 7.00), suggesting intracellular acidosis with intact cell membranes. In three stroke patients 5-8 days after onset, the injured cortex had higher CBF (60.9 ml/min/100 g), lower OEF (32.0%), and higher pHt (7.12) than the contralateral cortex (CBF = 45.3 ml/min/100 g, OEF = 58.0%, pHt = 7.06), which suggested an increase in extracellular volume compartment reflecting loss of cell membrane integrity. This method provides information on the regional tissue acid-base status and cell membrane integrity, which may be prognostic of tissue viability.

NMR spectroscopic investigation of the recovery of energy and acid-base homeostasis in the cat brain after prolonged ischemia

The effects of 1 h of complete global ischemia on the recovery of high-energy phosphates, intracellular pH (pHi), and lactate in the cat brain in vivo was investigated by 31P and 1H NMR spectroscopy. Ischemia led to a decrease in creatine phosphate (CrP), nucleoside triphosphates (NTP), and pHi, while inorganic phosphate and lactate increased. Intracellular pH decreased from a control value of 7.07 +/- 0.04 to 6.17 +/- 0.12 after 1 h of ischemia (N = 7). The degree of metabolic recovery after recirculation was variable. In three animals CrP and NTP were detected within 4 min and NTP increased to greater than or equal to 90% of control within 1 h; these levels were maintained for the 3 h of observation. In four other animals, CrP and NTP reached only 20 to 80% of control; however, high-energy phosphates decreased and lactate increased spontaneously between 1 and 2.5 h. Immediately following recirculation, pHi decreased further by an average of 0.3 units. The rate of recovery of cerebral pHi was slower than that of PCr and NTP for the majority of animals. Recovery of pHi was not detected for an average of 32 min after recirculation--by this time, NTP had attained 80 +/- 10% of their preischemic level. Recovery of pHi (and lactate) was not observed in two animals where PCr and NTP recovered transiently to only 30-43% of the preischemic level. Recovery of cerebral pHi was markedly heterogeneous in one animal, since two Pi peaks were detected shortly after recirculation.(ABSTRACT TRUNCATED AT 250 WORDS)

The role of glucose uptake and metabolism in hyperglycemic exacerbation of neurological deficit in the paraplegic rat

Previous studies indicate that hyperglycemia, particularly that induced by exogenous glucose administration, exacerbates neurological deficits in the rat spinal cord ischemic model. The effect of inhibition of glucose uptake (glucose transporter) and initial metabolism (hexokinase) on neurological outcome was evaluated in the present investigation using the competitive inhibitors 2-deoxyglucose (2-DG) and 3-O-methylglucose (3-OMG). Sprague-Dawley rats, weighing 200 to 300 gm each, received either 0.25, 1, or 2 gm/kg 2-DG; 2 gm/kg 3-OMG; 2 gm/kg glucose; or an equivalent volume of 0.9% saline intraperitoneally. Rats were intubated and ventilated with 1% to 1.5% halothane. The aortic arch was exposed and snares were placed on the right and left subclavian arteries and the aorta distal to the left subclavian artery. The three vessels were occluded for 10, 11, 12, or 13 minutes. Lower-extremity neurological deficits were evaluated at 1, 4, 18, and 24 hours postocclusion based on a 15-point scale (normal = 0, severe deficit = 15). Lower-extremity neurological deficits were significantly less severe in the groups treated with 2-DG (0.25 and 1 gm/kg) at 18 and 24 hours postocclusion (p less than 0.05 for 0.25 gm/kg and p less than 0.005 for 1 gm/kg, Student's t-test with Bonferroni correction). The lower 2-DG dose of 0.25 gm/kg did not significantly increase the plasma glucose level, suggesting that the glucose transporter was not markedly inhibited, and that the improved neurological outcome was more likely due to inhibition of hexokinase. The higher 2-DG dose of 1 gm/kg afforded protection despite significantly increasing the plasma glucose level, implying a strong inhibition of both the glucose transporter and hexokinase. Administration of 3-OMG, which only inhibits glucose uptake and not hexokinase, actually worsened the neurological deficit in a manner similar to that observed in rats treated with glucose. The authors conclude that the activity of the glucose transporter by itself does not significantly contribute to hyperglycemic exacerbation of neurological deficits. In contrast, the hexokinase step, at least in combination with the transporter and possibly alone, plays a significant role in hyperglycemic exacerbation of the lower-extremity neurological deficit in the paraplegic rat.

Lactic acid-induced swelling in C6 glial cells via Na+/H+ exchange

One of the primary consequences of ischemia is tissue acidification due to anaerobic production of lactic acid. Upon reperfusion and recovery of pH, cytotoxic edema often ensues. Na+/H+ exchange, a mechanism involved in the regulation of intracellular pH (pHi), is activated by low intracellular pH, is dependent on extracellular Na+, and is inhibited by low extracellular pH (pH less than 6) or by amiloride. In this study we explore the role of Na+/H+ exchange in cell swelling following cytoplasmic acidification of C6 glioma cells. Postischemic intracellular acidification was simulated in vitro by exposure of cells in suspension to: (1) 20 or 140 mM lactic acid; or (2) 10 microM oligomycin. pHi was monitored fluorimetrically using the intracellularly trapped pH-sensitive dye bis(carboxyethyl)carboxyfluorescein. Cell volume was measured electronically with a Coulter Counter/Channelyzer. Both simulations of ischemia caused intracellular acidification followed by recovery. pHi recovery was mediated by Na+/H+ exchange, since it was amiloride-sensitive and Na+-dependent. This pHi reversal following lactic acid-induced acidification was also inhibited at pHo less than 6. Volume measurements showed that cells suspended in 140 mM Na-lactate/lactic acid swelled by 19% over 15 min. This swelling was Na+-dependent, and inhibited by amiloride and pHo less than 6. These results suggest that Na+/H+ exchange may be involved in cell swelling following cytoplasmic acidification, and thus may be involved in postischemic cytotoxic brain edema.

In vivo 31P NMR spectroscopy of energy rich phosphates in the brain of the hyperammonemic rat

Hyperammonemia is a major contributing factor to the neurological abnormalities observed in hepatic encephalopathy and in congenital defects of ammonia detoxication. In rats variable changes in labile energy rich phosphates in the brain have been observed in hyperammonemia using biochemical methods. Using 31P-NMR spectroscopy however no significant changes of the relative concentrations of the energy rich phosphates alpha, beta and gamma-ATP, phosphocreatine, inorganic phosphate and the pH were found in the fronto parietal cortex of the urease treated hyperammonemic rat. Alterations in the metabolites of these compounds do not appear to be a major pathomechanism of ammonia toxicity in this brain area.

Cerebral carbohydrate metabolism during severe ischemia in fetal sheep

The effect of cephalic hypotension on brain metabolism was studied in 10 unanesthetized, normoxic (PaO2 greater than 17 mm Hg), late-gestation fetal lambs. Perfusion pressure (cephalic arterial minus sagittal venous pressure) was 40 +/- 1 mm Hg (SEM) during control and was reduced to 10 +/- 1 by occlusion of the Grachio-cephalic artery. Cerebral blood flow was measured with microspheres, and arterial and sagittal vein blood samples were analyzed for oxygen content, glucose, and lactate. During the occlusion, oxygen consumption decreased from 125 +/- 8 to 95 +/- 4 (p less than 0.05) (all values mumol 100 g-1 min-1), and glucose uptake increased from 20 +/- 3 to 25 +/- 1 (p less than 0.05). During the control period, there was no net lactate flux; during the occlusion, lactate excretion was 5.7 +/- 1.4 (p less than 0.005). The control glucose and oxygen uptakes demonstrated a normal 6:1 molar ratio; however, during the occlusion, 9.4 mumol 100 g-1 glucose min-1 were taken up in excess of expected aerobic glucose metabolism. If all of this glucose were anaerobically metabolized to lactate, three times the measured efflux would be produced. The transport properties of the fetal blood-brain barrier may be important factors in perinatal brain injury.

Acidosis and ischemic brain damage

It is now widely accepted that acidosis is an important component of the pathogenetic events that lead to ischemic brain damage. The objective with this article is to recall the evolution of the concept, to describe the conditions under which tissue acidosis arises and causes enhanced brain damage, and to probe into the cellular and molecular mechanisms involved.

Insulin administration protects from paraplegia in the rat aortic occlusion model

The effect of insulin induced hypoglycemia was evaluated in a rat aortic occlusion model of ischemic paraplegia. One hour before aortic occlusion, 200-250 g Sprague-Dawley rats received either 1 cc of saline or 0.5 units regular insulin in 1 cc saline. Rats were then anesthetized, intubated, and ventilated with halothane (1-1.5%). The aortic arch was exposed and snares were placed on the right and left subclavian arteries and the aorta distal to the left subclavian. The three vessels were occluded for 10 min. Lower extremity neurologic deficit was evaluated at 1, 4, 18, and 24 hr postocclusion based on a 15-point scale (0 = normal, 15 = severe deficit). Lower extremity neurologic deficit was significantly decreased in the insulin-treated group at 18 and 24 hr postocclusion (P = 0.005 and 0.006, respectively, Student's test). Blood glucose concentration was significantly lower at the time of occlusion in the insulin-treated group when compared to the saline-treated group (P = 0.001). We conclude that in this rat model of paraplegia, insulin induced hypoglycemia is associated with a reduction in lower extremity neurologic deficit produced by temporary aortic occlusion.

31P NMR characterization of graded traumatic brain injury in rats

Irreversible tissue injury following central nervous system trauma is believed to result from both mechanical disruption at the time of primary insult, and more delayed "autodestructive" processes. These delayed events are associated with various biochemical changes, including alterations in phosphate energy metabolism and intracellular pH. Using 31P NMR, we have monitored the changes in phosphorus energy metabolism and intracellular pH in a single hemisphere of the rat brain over an 8-h period following graded, traumatic, fluid percussion-induced brain injury. Following trauma the ratio of phosphocreatine to inorganic phosphate (PCr/Pi) declined in each injury group. This decline was transitory with low injury (1.0 +/- 0.5 atm), biphasic with moderate (2.1 +/- 0.4 atm) and high (3.9 +/- 0.9 atm) injury, and sustained following severe injury (5.9 +/- 0.7 atm). The initial PCr/Pi decline in the moderate and high injury groups was associated with intracellular acidosis; however, the second decline occurred in the absence of any pH changes. Alterations in ATP occurred only in severely injured animals and such changes were associated with marked acidosis and 100% mortality rate. After 4h, the posttraumatic PCr/Pi ratio correlated linearly with the severity of injury. We suggest that a reduced posttraumatic PCr/Pi ratio may be indicative of altered mitochondrial energy production and may predict a reduced capacity of the cell to recover from traumatic injury.

Traumatic brain injury in the rat: alterations in brain lactate and pH as characterized by 1H and 31P nuclear magnetic resonance

Application of both phosphorus (31P) and proton (1H) magnetic resonance spectroscopy (MRS) to the study of brain metabolism permits the noninvasive measurement of intracellular pH and brain lactate level. We have used water-suppression 1H MRS with novel lactate-editing techniques, together with 31P MRS, to characterize sequential changes in brain lactate level and pH in vivo over an 8-h period following fluid-percussion brain injury of graded severity in the rat. A transient fall in intracellular pH (from 7.09 +/- 0.07 at baseline to 6.88 +/- 0.09 at 40 min postinjury) occurred in animals subjected to moderate- (1.5-2.2 atm) and high- (2.5-3.3 atm) but not low-level (0.1-1.2 atm) injury; intracellular pH returned to baseline by 90 min postinjury. Transient elevations in brain lactate level were observed that temporally paralleled and were significantly correlated with the pH changes for all injury levels (r = 0.93, p less than 0.001). Postinjury alterations in intracellular brain pH and lactate level were identical in magnitude in animals subjected to either moderate or high-level injury. However, animals subjected to moderate injury had a moderate chronic neurological deficit that persisted up to 4 weeks postinjury, whereas animals subjected to a high level of injury showed greater histopathological damage and a more severe chronic neurological deficit. These data suggest that the extent of posttraumatic intracellular cerebral acidosis in our model of experimental head injury is not directly related to the severity of functional neurological deficit.

Effects of traumatic brain injury on cerebral high-energy phosphates and pH: a 31P magnetic resonance spectroscopy study

Traumatic injuries to the CNS produce tissue damage both through mechanical disruption and through more delayed autodestructive processes. Delayed events include various biochemical changes whose nature and time course remain to be fully elucidated. Magnetic resonance spectroscopy (MRS) techniques permit repeated, noninvasive measurement of biochemical changes in the same animal. Using phosphorus MRS, we have examined certain biochemical responses of rats over an 8-h period following lateralized brain injury (1.5-2.5 atmospheres) using a standardized fluid-percussion model recently developed in our laboratory. Following injury, the ratio of phosphocreatine to inorganic phosphate (PCr/Pi) showed a biphasic decline: The first decline reached its nadir (4.8 +/- 0.4 to 2.8 +/- 0.7) by 40 min post-trauma with recovery by 100 min, followed by a second decline by 2 h that persisted for the remaining 6-h observation period (mean 2.5 +/- 0.5). The first, but not the second, decrease in PCr/Pi was associated with tissue acidosis (pH 7.10 +/- 0.03 to 6.86 +/- 0.11). No changes in ATP occurred at any time during the injury observation period. Such changes may be indicative of altered mitochondrial energy production following brain injury, which may account for the reduced capacity of the cell to recover from traumatic injury.

Simultaneous 31P- and 1H-nuclear magnetic resonance studies of hypoxia and ischemia in the cat brain

The objective of this study was to evaluate simultaneous 31P/1H nuclear magnetic resonance (NMR) spectroscopy as a technique for monitoring and correlating changes in brain energy metabolism during hypoxia and ischemia. Five cats were studied with a protocol that involved 20 min of hypoxia (PaO2 20 mm), 60 min of recovery, 10 min of hypoxia with relative ischemia (bilateral carotid occlusion, PaO2 20 mm), and 60 min of recovery. Bifrontal and biparietal electrocorticograms (ECoG) were monitored continuously during the entire protocol. The results demonstrate that the degree of metabolic response is different in individual cats, but a number of quantitative relationships between metabolic parameters are consistently observed for all cats. First, there is agreement between increases in lactate and changes in intracellular pH; the observed relationship corresponds to an in vivo cerebral buffer capacity of 29 mumol/g/pH unit. Second, the delayed recovery of PCr is due to the effect of metabolic acidosis on the creatine kinase equilibrium and not to a delayed recovery of the ATP/ADP ratio. Third, the observed rate of lactate clearance from the cell is zero-order (k = 0.36 mumol/g/min) for lactate levels greater than 5 microns/g and may be composed of both lactate efflux from the cell and lactate oxidation.

Histopathological and hemodynamic consequences of complete versus incomplete ischemia in the rat

The primary objective of this study was to compare the histopathological consequences of complete versus incomplete ischemia under experimental conditions that limit lactate accumulation. Fasted rats underwent 1 h of either complete or incomplete ischemia by a procedure combining bilateral common carotid artery occlusion, halothane-induced systemic hypotension, and CSF pressure elevation. Histopathological outcome was evaluated 4 h later and was graded on a 4-point scale. Incomplete ischemia resulted in ischemic neuronal damage within selectively vulnerable brain regions. In contrast, complete ischemia, in addition to diffuse neuronal damage, resulted in focal sites of parenchymal necrosis with vascular stasis. Perfusion defects were detected by carbon black infusion within cortical and subcortical regions following only 25 min of complete, but not incomplete, ischemia. Ultrastructural abnormalities at the same duration of complete ischemia included a high frequency of endothelial microvilli and compressed lumina with severe perivascular astrocytic swelling. When recirculation was instituted for 1 h following 1 h of complete ischemia, regions of nonperfusion were detected autoradiographically. Thus, when the degree of lactic acidosis is controlled, prolonged periods of complete ischemia result in a more severe pathological outcome compared to incomplete ischemia. Focally impaired postischemic cerebral perfusion appears to be an important factor in infarct formation under the present experimental conditions.

Vascular aspects and hemodynamic consequences of central nervous system injury

We document here microvascular alterations occurring in models of mild, moderate, and severe cerebral ischemic injury. The relationship of the vascular abnormalities to the generation of hemodynamic alterations was also evaluated. Following periods of severe incomplete ischemia, scanning electron microscopic analysis of cerebral microvessels revealed the widespread production of cerebral endothelial microvilli. These microvilli increased in frequency as the ischemic insult was prolonged and remained prominent during periods of recirculation. Although these luminal projections would not be expected to inhibit reperfusion completely, they might increase microvascular resistance, leading to moderate hemodynamic impediments extending into the post-ischemic period. Similar periods of complete ischemia resulted in more severe microvascular alterations. Light and electron microscopic studies revealed a high frequency of compressed capillary lumina with vascular stasis. These compressed vessels were consistently surrounded by swollen astrocytic foot processes. When recirculation was instituted for 1 hr following 1 hr of complete ischemia, regions of non-perfusion were detected autoradiographically within brain regions destined to undergo ischemic infarction. Finally, in an attempt to determine the consequences of a primary microvascular insult on brain structure and function, the endothelial layer of microvessels in the cerebral cortex was injured using a noninvasive photochemical method. Endothelial damage led to platelet aggregation in both pial and intraparenchymal vessels. Occlusive thrombi were frequently detected with perivascular edema associated with vascular compression and severe focal ischemia. Ultrastructural blood-brain barrier studies using the horseradish peroxidase tracer demonstrated that protein leakage at the site of primary vascular injury resulted in tracer material in brain regions remote from the pathological lesion. The widespread leakage of protein tracer was associated with decreased blood flow in remote brain regions at several postirradiation periods. These data emphasize the importance of injury-induced microvascular dysfunction in the generation of brain lesions and hemodynamic abnormalities.