Regions of cerebral ischemia located by pyridine nucleotide fluorescence.

The fluorescence of the reduced form of the endogenous pyridine nucleotide nicotinamide adenine dinucleotide was used to map regions of ischemia in cat brain. A remarkably microheterogeneous pattern of increased fluorescence resulted from a critical level of incomplete cerebral ischemia. The fluorescence pattern suggests that ischemia occurs initially in microwatershed zones between penetrating cerebral arteries.

Postischemic hypothermia fails to reduce ischemic injury in gerbil hippocampus.

The objective of this study was to determine whether postischemic hypothermia diminishes ischemic injury in gerbil hippocampus. Cerebral ischemia was produced by occluding both carotid arteries for 5 min, while maintaining the temperature of the cranium and rectum at 38 degrees C. Upon recirculation, the animals were divided into three groups: normothermic (38 degrees C), moderately hypothermic (33 degrees C), and deeply hypothermic (23 degrees C). In the normothermic group, cranial and rectal temperatures were maintained at 38 degrees C for 30 min and 2 h, respectively, prior to the removal of the temperature probes. In the moderately hypothermic group, cranial and rectal temperatures were reduced within 10 min to 33 degrees C for 1 h, and then rewarmed to 38 degrees C. In the deeply hypothermic group, rectal temperature was lowered within 10 min to 23 degrees C for 2 h prior to rewarming to 38 degrees C. After recovery for 1 week, the extent of histologic injury in the hippocampus was assessed in stained sections. Maximal injury was present in the CA1 subfield in all three groups. These results indicate that hippocampal injury was not diminished by postischemic hypothermia during the first 2 h of reperfusion. Thus, pharmacologic studies of postischemic protection in the gerbil model may not be strongly influenced by transient postischemic hypothermia.

Regional alterations of ATP and heat-shock protein-72 mRNA following hypoxia-ischemia in neonatal rat brain.

Neonatal rats, 7 days of age, underwent unilateral carotid artery ligation followed by exposure to hypoxia (8% O2) for 80 min. At the end of the period of hypoxia, and after recovery for 2 or 24 h, regional levels of ATP and heat-shock protein-72 (hsp72) mRNA were measured in adjacent brain sections using ATP-luminescence histochemistry and in situ hybridization, respectively. At the end of hypoxia, ATP levels were decreased in a patchy pattern within the hemisphere ipsilateral to the carotid ligation. In the parietal cortex, the reduction of ATP often occurred in columns oriented perpendicular to the cortical surface. Expression of hsp72 mRNA was not detected prior to recovery, except in the ventricular lining of the ipsilateral hemisphere. However, by 2 h of recovery, hsp72 mRNA was expressed in a diffuse pattern in the ipsilateral hemisphere, even in regions in which the distribution of ATP remained patchy. Although the regional extent of expression varied in different animals, hsp72 mRNA was expressed consistently in the subcortical white matter, which, in some animals, was the only region showing expression. In contrast to the diffuse pattern of expression at 2 h of recovery, expression of hsp72 mRNA at 24 h was highly localized in the superficial layers of cerebral cortex and the pyramidal cell layer of hippocampus. The present results demonstrate that hypoxia-ischemia causes regionally distinct alterations in ATP and hsp72 mRNA that may be related to cell injury in this model.

Correlation between cerebral blood flow and ATP content following tourniquet-induced ischemia in cat brain.

Cerebral ischemia was produced in anesthetized cats using a neck tourniquet, which diminished cortical blood flow to less than 2 ml/100 g/min and depleted levels of ATP throughout the brain. Following a 30-min insult, cortical flow measured with H2 electrodes returned nearly to control, but subsequently decreased to 14-47% of control values. Despite this secondary hypoperfusion, ATP levels adjacent to the H2 electrode were restored to 75% of normal during the 2-h recirculation period. Therefore, this degree of hypoperfusion did not cause a secondary failure of energy metabolism. Following a 60-min insult, impaired reperfusion prevented the regeneration of brain ATP. However, preischemic bilateral craniectomies significantly improved recovery of blood flow and ATP levels following 60 min of ischemia. Therefore, in the present model, insufficient reflow is a primary factor limiting recovery of energy metabolism. Further, surgical decompression prevented the occurrence of "no reflow" caused by 60 min of ischemia.

Spreading depression induces tolerance of cortical neurons to ischemia in rat brain.

Cortical spreading depression (CSD) was induced in male Wistar rats by applying 2 M KCl to the frontal cortex of one hemisphere for 2 h. Saline was applied to the contralateral cortex in the same manner. Following recovery for 24 h, bilateral forebrain ischemia was induced for 6 min, and the animals were permitted to survive for 6 days for assessment of histopathology. The number of necrotic neurons was counted in the cerebral cortex, striatum, and hippocampus of both hemispheres. In separate sets of animals, the effects of KCl application on cortical direct current (DC) potential and regional expression of c-fos mRNA and 72-kDa heat shock protein (hsp72) mRNA were determined. Forebrain ischemia induced selective neuronal necrosis in both hemispheres, but the number of necrotic neurons in the cerebral cortex ipsilateral to the application of KCl was significantly smaller than that in the contralateral cortex (p < 0.02, Wilcoxon signed rank test, n = 7). In the striatum and hippocampus, there were no significant differences in neuronal necrosis between hemispheres. Application of KCl for 2 h induced 11 +/- 2 (mean +/- SD, n = 5) negative deflections of DC potential in the ipsilateral cortex; none were detected in the contralateral cortex. Widespread expression of c-fos mRNA was evident in the ipsilateral cortex, while hsp72 mRNA expression was restricted to the KCl application site. The present results demonstrate that CSD induces tolerance of cortical neurons to ischemia by mechanisms unrelated to hsp72.

Cerebrovascular adaptation after unilateral carotid artery ligation in the rat: preservation of blood flow and ATP during forebrain ischemia.

To investigate long-term adaptations after unilateral carotid artery ligation, the effect of forebrain ischemia on cerebral blood flow and ATP levels was determined at various times after ligation. Unilateral carotid artery ligation was performed in male Wistar rats 0, 3, or 7 days before forebrain ischemia. Laser-Doppler blood flow was monitored bilaterally over the parietal cortex and ATP was measured in the subadjacent cortex of both hemispheres at the end of a 10-minute episode of ischemia. In the 0-day group, forebrain ischemia reduced cortical blood flow to 12% +/- 8% (mean +/- SD) of preischemic values and lowered cortical ATP to 26% +/- 35% of control levels in the ipsilateral hemisphere. Delaying the onset of forebrain ischemia for 3 days after carotid artery ligation significantly improved cortical blood flow (29% +/- 12%, P < 0.05) and ATP levels (92% +/- 11%, P < 0.05) in the ipsilateral hemisphere. Delaying forebrain ischemia for 7 days also significantly improved ipsilateral blood flow (36% +/- 11%, P < 0.05) and ATP levels (81% +/- 29%, P < 0.05) compared with the 0-day group. In the contralateral hemisphere, the reduction in blood flow and ATP levels was not significantly altered by delaying the onset of forebrain ischemia for 3 or 7 days. These results show that unilateral carotid artery ligation induces long-term vascular adaptations that improve the collateral circulation and preserve ATP levels during a subsequent episode of ischemia.

NADH fluorescence and regional energy metabolites during focal ischemia and reperfusion of rat brain.

Transient focal ischemia was produced in rat brain using simultaneous, reversible occlusion of the middle cerebral artery (MCA) and both carotid arteries. NADH tissue fluorescence and regional levels of ATP and lactate were measured after occlusion for 1 or 2.5 h and after reperfusion for 1 or 24 h following a 2.5-h insult. Occlusion for 1 or 2.5 h caused a marked but microheterogenous increase in NADH fluorescence, which was restricted to the MCA territory of the ipsilateral cortex. In this ischemic core, tissue levels of ATP were nearly depleted, while lactate accumulated to 10-13 mmol/kg. Metabolic alterations were less pronounced in regions adjacent to the ischemic core; however, one border region experienced a progressive increase in lactate between 1 and 2.5 h. NADH fluorescence and metabolite levels were not significantly altered in subcortical structures. In animals reperfused after a 2.5-h insult, NADH fluorescence diminished in the ischemic core to abnormally low levels, ATP was restored only to 37-50% of control, and lactate remained elevated. By 24 h, histologic infarction was evident in the regions with metabolic impairment. These results indicate that focal depletion of energy metabolites for 2.5 h caused irreversible impairment of energy metabolism and focal infarction even though lactate accumulation was moderate.

Adaptive preservation of ATP and tolerance to hypoxia following carotid artery ligation in the immature rat.

To investigate the adaptive mechanisms following carotid artery ligation in immature rats, histologic injury and tissue levels of ATP were compared after exposure to identical episodes of hypoxia induced either 3 or 24 h postligation. Histologic injury, assessed in both 9-day- and 23-day-postnatal animals after survival for 1 week, was markedly diminished in animals exposed to hypoxia 24 h postligation compared to that in animals exposed to hypoxia 3 h postligation. In 9-day-postnatal animals, ATP levels in the cerebral cortex ipsilateral to the ligation were depleted during hypoxia to 0.39 +/- 0.49 mmol/kg (mean +/- SD; N = 15) in animals exposed to hypoxia 3 h postligation but were maintained at 2.04 +/- 0.26 mmol/g (N = 18; p < 0.001) in animals exposed to hypoxia 24 h postligation. Thus, preservation of ATP may account for the diminution of cellular injury that results from delaying the onset of hypoxia from 3 to 24 h after carotid artery ligation in immature rats.

Columnar alterations of NADH fluorescence during hypoxia-ischemia in immature rat brain.

Cerebral hypoxia-ischemia was produced in 7-day postnatal rats by unilateral carotid artery ligation combined with systemic hypoxia (8% O2). Levels of high energy phosphates, which were only slightly altered in the contralateral hemisphere, were nearly depleted in the ipsilateral hemisphere during the 3-h hypoxic insult. With hypoxia of between 1 and 3 hours' duration, columnar alterations of cortical NADH fluorescence occurred in the same location and regional pattern as did histologic damage demonstrated previously (Rice et al., 1981). In regions exhibiting columns of NADH fluorescence, there was no evidence of a columnar reduction of high energy phosphates as levels of ATP and phosphocreatine were nearly zero. Recovery from 3 h of hypoxia was accompanied by partial and regionally heterogeneous restoration of ATP within the ipsilateral hemisphere. Columnar variations of NADH fluorescence were not detected in the recovery period; rather, regions with impaired restitution of high energy phosphates exhibited NADH fluorescence that was diminished diffusely compared to the contralateral hemisphere. The correlation between depressed NADH fluorescence and depleted ATP, present as cortical columns during hypoxia and as larger regions during recovery, suggests that decreased formation of NADH may be limiting the resynthesis of high energy phosphates.

Effect of glucose on recovery of energy metabolism following hypoxia-oligemia in mouse brain: dose-dependence and carbohydrate specificity.

Unilateral cerebral hypoxia-oligemia was produced in anesthetized mice using carotid artery occlusion combined with systemic hypoxia (10% O2). In the cerebral cortex ipsilateral to the carotid occlusion, ATP levels were depleted during a 30-min insult, but were restored to 64% of control during 60 min of recovery. Pretreatment of animals with glucose diminished the restoration of ATP in a dose-dependent manner. Thus, when blood glucose levels exceeded 12-13 mM (225 mg/dl), ATP recovery was greatly impaired. Neither galactose nor 3-O-methylglucose mimicked the detrimental effect of glucose. However, pretreatment with mannose, which is readily metabolized by brain, impaired restoration of ATP. The impairment, therefore, appears to be specific for substrates of cerebral metabolism. The ischemic accumulation of lactate in the ipsilateral cortex was augmented by only 30% at blood glucose levels well above the threshold for ATP recovery. Thus, unless recovery of energy metabolism is sensitive to small increments in brain lactate, it is difficult to explain the glucose-induced energy failure on the basis of enhanced lactic acidosis. Ipsilateral cerebral blood flow (CBF), measured with [14C]iodoantipyrine during hypoxia and recovery, was lower in glucose-pretreated than in saline-pretreated animals. However, the poor correlation between CBF and ATP, measured in the same tissue samples at 15 min recovery, failed to substantiate that regeneration of ATP was flow-limited early in recovery.

Mild hypothermia prevents ischemic injury in gerbil hippocampus.

The objective of this study was to define the degree of hypothermia required to diminish ischemic injury to CA1 hippocampal neurons following 5-min bilateral ischemia in the gerbil. The temperature of the body and head was regulated in three groups of animals at 37.5, 35.5, or 32.5 degrees C during 5-min bilateral carotid artery occlusion. Upon recirculation, normothermia was restored in all animals, and recovery was permitted for 1 week. Ischemic injury to CA1 hippocampus was determined using three endpoints: histologic injury, ATP content, and adenylate kinase activity. Reduction of head temperature to 35.5 and 32.5 degrees C during ischemia diminished histologic injury and improved CA1 levels of ATP and adenylate kinase activity in a dose-dependent manner. Indeed, 32.5 degrees C completely abolished ischemic injury to CA1 hippocampus, judging from each of the three endpoints. Reduction of head temperature to 32.5 degrees C delayed but did not prevent the depletion of ATP throughout the hippocampus during the 5-min ischemic insult. These results demonstrate that a decrease in head temperature of only 2 degrees C reduces the degree of CA1 injury in the gerbil model of 5-min bilateral ischemia. Thus, it is imperative to maintain strict normothermia in pharmacologic studies of ischemic protection. Finally, administration of nicardipine to normothermic gerbils failed to diminish ischemic injury in the CA1 hippocampus.

Regional expression of heat shock protein-70 mRNA and c-fos mRNA following focal ischemia in rat brain.

In situ hybridization was used to estimate regional levels of heat shock protein-70 (HSP-70) mRNA and c-fos mRNA in two related models of focal cerebral ischemia. In the first model, permanent occlusion of the distal middle cerebral artery (MCA) alone caused a patchy increase in HSP-70 mRNA by 1 h in the central zone of the MCA territory of the ipsilateral neocortex. Tissue levels of HSP-70 mRNA continued to increase for several hours and remained elevated at 24 h. In contrast to the focal expression of HSP-70, c-fos mRNA was increased throughout the ipsilateral cerebral cortex by 15 min and remained elevated for least 3 h. The wide distribution of c-fos expression suggests it may have been caused by spreading depression. In the second model, severe focal ischemia was produced with a combination of transient (1-h) bilateral carotid artery occlusion and permanent MCA occlusion. Combined occlusion for 1 h without reperfusion caused expression of HSP-70 mRNA only in regions adjacent to the central zone of the MCA territory of the neocortex. However, reperfusion of the carotids for 2 h generated intense expression of HSP-70 mRNA throughout most of the ipsilateral cerebral cortex, white matter, striatum, and hippocampus. The wide-spread increase in HSP-70 mRNA suggests that reperfusion triggered expression in all previously ischemic regions. However, at 24 h of reperfusion, increased levels of HSP-70 mRNA were restricted primarily to the ischemic core of the neocortex. These results suggest that expression of HSP-70 mRNA is prolonged in regions undergoing injury, but is transient in surrounding regions that recover.

Regional induction of c-fos and heat shock protein-72 mRNA following fluid-percussion brain injury in the rat.

To evaluate the cellular response to traumatic brain injury, the expression of mRNA for c-fos and the 72-kDa heat shock protein (hsp72) was determined using in situ hybridization following lateral fluid-percussion injury (2.2-2.4 atm) in rat brain. At 2 h after injury, induction of c-fos mRNA was observed throughout the cortex ipsilateral to the site of injury, while increased expression of hsp72 mRNA was restricted to regions of the cortex surrounding the contusion area. An increase in c-fos mRNA, but not hsp72 mRNA, was observed bilaterally in the CA3 subfield of the hippocampus and the granule cells of the dentate gyrus and in the thalamus ipsilateral to the impact site. By 6 h, increased expression of c-fos mRNA was observed only in the corpus callosum on the impact side; hsp72 mRNA persisted in the deep cortical layers and upper layers of the subcortical white matter below the site of maximal injury. By 24 h, both c-fos and hsp72 mRNA had returned to control levels in all regions of the brain. These results demonstrate that lateral fluid-percussion brain injury triggers regionally and temporally specific expression of c-fos and hsp72 mRNA, which may be suggestive of differential neurochemical alterations in neurons and glia following experimental brain injury.

Induction of ischemic tolerance following brief focal ischemia in rat brain.

The objective of this study was to determine whether brief focal ischemia induces ischemic tolerance in rat brain. Focal ischemia was produced in Wistar rats by occluding the middle cerebral artery (MCA) for 20 min at a distal site. Following recovery for 24 h, the animals were subjected to a 10-min episode of forebrain ischemia using a combination of bilateral carotid artery occlusion and systemic hypotension. Histologic injury, assessed after a survival period of 3-4 days, consisted of selective neuronal necrosis bilaterally in cerebral cortex, striatum, hippocampus, and thalamus superimposed upon a small cortical infarct adjacent to the site of MCA occlusion. However, the intensity of neuronal necrosis in the MCA territory of the neocortex ipsilateral to MCA occlusion was markedly less than that in the contralateral MCA cortex. In contrast, the extent of neuronal necrosis in subcortical structures was similar in both hemispheres. Unexpectedly, animals in which the MCA was manipulated, but not occluded, also exhibited a marked reduction of neuronal necrosis in the ipsilateral MCA neocortex following forebrain ischemia. However, in animals with craniotomy alone, forebrain ischemia caused a similar extent of neuronal necrosis in the MCA neocortex of both hemispheres. Transient occlusion of the MCA induced the focal expression of the 72-kDa heat-shock protein (hsp72) in the MCA territory of the neocortex. Limited expression of hsp72 was also detected following sham occlusion, but not after craniotomy alone. These results demonstrate focal induction of ischemic tolerance in rat neocortex that may be related to expression of heat-shock proteins.

Effect of cortical spreading depression on the levels of mRNA coding for putative neuroprotective proteins in rat brain.

Previous studies have demonstrated that cortical spreading depression (CSD) induces neuronal tolerance to a subsequent episode of ischemia. The objective of the present investigation was to determine whether CSD alters levels of mRNA coding for putative neuroprotective proteins. Unilateral CSD was evoked in male Wistar rats by applying 2 mol/L KCl over the frontal cortex for 2 hours. After recovery for 0, 2, or 24 hours, levels of several mRNA coding for neuroprotective proteins were measured bilaterally in parietal cortex using Northern blot analysis. Levels of c-fos mRNA and brain-derived neurotrophic factor (BDNF) mRNA were markedly elevated at 0 and 2 hours, but not 24 hours after CSD. Tissue plasminogen activator (tPA) mRNA levels were also significantly increased at 0 and 2 hours, but not 24 hours after CSD. Levels of the 72-kDa heat-shock protein (hsp72) mRNA were not significantly increased by CSD, except for a small elevation (20%) at 2 hours recovery. Levels of the 73-kDa heat-shock cognate (hsc73) mRNA were slightly, but significantly, increased at 2 and 24 hours of recovery. Finally, levels of mRNA for protease nexin-1 and glutamine synthetase were not significantly altered by CSD at any time studied. The current results support the hypothesis that neuronal tolerance to ischemia after CSD may be mediated by increased expression of FOS, BDNF, or tPA, but not by increased expression of hsp72, hsc73, nexin-1, or glutamine synthetase.

Regional alterations in glucose consumption and metabolite levels during postischemic recovery in cat brain.

Local CMRgl (LCMRgl) and metabolite levels were measured in the same tissue samples following 4 h of recirculation after 1 h of occlusion of the middle cerebral artery in the cat. The rate of glucose utilization was calculated using direct measurement of tissue deoxyglucose-6-phosphate and using a "lumped" constant corrected in each sample for alterations in tissue glucose. Increased LCMRgl (compared with that in sham-operated animals) occurred in regions with only minor alterations in levels of lactate and phosphocreatine. By contrast, LCMRgl was markedly depressed in regions with major changes in lactate and high-energy phosphates. Interestingly, tissue levels of glucose and unphosphorylated deoxyglucose were abnormally elevated in regions with profound energy failure. These results indicate an inhibition of glucose utilization in regions damaged by ischemia, despite the persistent elevation of tissue lactate. Increased glucose metabolism at 4 h post ischemia was detected only in areas with minor anaerobic alteration of metabolite levels.

Effect of ischemia and reperfusion on lambda of the lumped constant of the [14C]deoxyglucose technique.

We measured the parameter lambda, which is the ratio of the distribution spaces of 2-deoxy-D-glucose (DG) and glucose in the brain, in a model of focal cerebral ischemia in the cat. lambda is the parameter in the lumped constant of the [14C]DG technique most susceptible to changes in ischemia. Cats were subjected to occlusion of the middle cerebral artery for a period of 2 h. During the last 60 min of occlusion, [14C]DG was infused in a programmed fashion so as to obtain a stable arterial blood [14C]DG concentration. The brain was funnel-frozen to preserve tissue metabolites and the frozen brain was sampled regionally (4 to 7-mg samples) for local concentrations of glucose, ATP, phosphocreatine (PCr), and lactate. In a separate series of cats, the infusion of [14C]DG was started after 2 h of occlusion and 3 h of recirculation. In both series, lambda declined slightly for increased levels of tissue glucose and increased appreciably as tissue glucose decreased. A similar relationship was observed between lambda and ATP and PCr, although the correlation was not as clear. Since lambda, and hence the lumped constant, increases in ischemia as well as in postischemic tissue, it is important to measure tissue glucose concentration if quantitative values of local cerebral glucose metabolism are desired in this condition.

Regional expression of c-fos and heat shock protein-70 mRNA following hypoxia-ischemia in immature rat brain.

Cerebral ischemia induces the expression of a number of proteins that may have an important influence on cellular injury. The purpose of this study was to compare the regional effects of hypoxia-ischemia on the expression of the proto-oncogene, c-fos, and the heat shock protein-70 (HSP-70) gene in developing brain. Unilateral hypoxia-ischemia was produced in the brain of immature rats (7, 15, and 23 days after birth) using a combination of carotid artery ligation and systemic hypoxia (8% O2). After recovery for 2 and 24 h, the regional expression of c-fos and HSP-70 mRNA was determined using in situ hybridization. Littermates were permitted to recover for 1 week for assessment of histologic injury. Hypoxia-ischemia increased the expression of both c-fos and HSP-70 mRNA, but the topography of expression varied with the age of the animal as well as the mRNA species. In the 7-day-old group, expression of c-fos at 2 h increased in multiple regions of the ipsilateral hemisphere in nearly one-half of the animals, while HSP-70 mRNA was not expressed until 24 h and, then, predominantly in the hippocampus. In 15- and 23-day-old rats, expression of c-fos was increased at 2 h in the entorhinal cortex and in the dendritic field of the upper blade of the hippocampal dentate gyrus, while HSP-70 mRNA was prominently expressed in neocortex and the cell layers of the hippocampus. Interestingly, the strong expression of HSP-70 mRNA in dentate granule cells did not occur in the innermost layer of cells.(ABSTRACT TRUNCATED AT 250 WORDS)

Preconditioning with cortical spreading depression does not upregulate Cu/Zn-SOD or Mn-SOD in the cerebral cortex of rats.

Previous studies have demonstrated that preconditioning the brain with cortical spreading depression (CSD) induces tolerance to a subsequent episode of ischemia. In other models of preconditioning, induction of ischemic tolerance has been associated with increased expression of the antioxidant enzyme, superoxide dismutase (SOD). The objective of the present study was to determine whether CSD upregulates Cu/Zn-SOD or Mn-SOD. CSD was induced in one hemisphere by applying 2 M KCl to the frontal cortex in Wistar rats. After 2 or 24 h of recovery, Cu/Zn-SOD and Mn-SOD mRNA levels were determined in both hemispheres using Northern blot analysis. In separate rats, Cu/Zn-SOD and Mn-SOD protein levels were determined 24 and 72 h after CSD using Western blot analysis. In addition, total SOD, Cu/Zn-SOD and Mn-SOD enzymatic activities were measured 24 and 72 h after CSD using spectrophotometric and zymographic assays. At the times investigated, no significant differences in mRNA or protein levels for Cu/Zn-SOD or Mn-SOD were observed between the ipsilateral and contralateral cortex. Further, there were no significant differences in Cu/Zn-SOD or Mn-SOD enzymatic activities between the two hemispheres at 24 or 72 h after CSD. In addition, CSD did not alter the activities of Cu/Zn-SOD or Mn-SOD in either hemisphere, relative to those in unoperated animals. Taken together, these results fail to support the hypothesis that CSD-induced tolerance is mediated through the upregulation of Cu/Zn-SOD or Mn-SOD.

Dose-dependent induction of mRNAs encoding brain-derived neurotrophic factor and heat-shock protein-72 after cortical spreading depression in the rat.

Previous studies have demonstrated that cortical spreading depression (CSD) increases the expression of putative neuroprotective proteins. The objective of the present study was to elucidate the relationship between the number of episodes of CSD and steady-state levels of mRNAs encoding brain-derived neurotrophic factor (BDNF), heat-shock protein-72 (hsp72) and c-fos. Wistar rats were administered one, five, or twenty-five episodes of CSD evoked by application of 2 M KCl to the frontal cortex of one hemisphere. Animals were permitted to recover for 30 min, 2 h or 24 h prior to sacrifice. Total RNA was isolated from the parietal cortex of each hemisphere and analyzed using Northern blots. At 30 min recovery, levels of BDNF mRNA were not significantly elevated after 1 episode of CSD, but were increased 4-fold after five episodes of CSD and 11-fold after twenty-five episodes of CSD, relative to levels in the contralateral hemisphere. At 2 h recovery, BDNF mRNA levels increased 2-, 3- and 9-fold, respectively. At 24 h, BDNF mRNA had returned to control levels in all groups. Thus, CSD increased levels of BDNF mRNA in a dose-dependent fashion at the early recovery times. Hsp72 mRNA was below the level of detection after 1 and 5 episodes of CSD. However, after twenty-five episodes of CSD, hsp72 mRNA levels were increased in the ipsilateral hemisphere at 30 min and 2 h recovery. Unlike levels of BDNF and hsp72 mRNA, levels of c-fos mRNA were increased nearly to the same extent at 30 min and 2 h after one, five or twenty-five episodes of CSD before returning to control by 24 h recovery. These results demonstrate that CSD triggers a dose-dependent increase in the expression of genes encoding neuroprotective proteins, which may mediate tolerance to ischemia induced by CSD.