Effects of Glucose and PaO 2 Modulation on Cortical Intracellular Acidosis, NADH Redox State, and Infarction in the Ischemic Penumbra

Transcription

1 Effects of Glucose and PaO 2 Modulation on Cortical Intracellular Acidosis, NADH Redox State, and Infarction in the Ischemic Penumbra Robert E. Anderson, BS; William K. Tan, PhD; Heidi S. Martin; Fredric B. Meyer, MD Background and Purpose During focal cerebral ischemia, the ischemic penumbra or border-zone regions of moderate cortical blood flow reductions have a heterogeneous development of intracellular cortical acidosis. This experiment tested the hypotheses that (1) this acidosis is secondary to glucose utilization and (2) this intracellular acidosis leads to recruitment of potentially salvageable tissue into infarction. Methods Brain ph i, regional cortical blood flow, and NADH redox state were measured by in vivo fluorescent imaging, and infarct volume was assessed by triphenyltetrazolium chloride histology. Thirty fasted rabbits divided into 6 groups of 5 each were subjected to 4 hours of permanent focal ischemia in the presence of hypoglycemia ( 2.8 mmol/l), moderate hyperglycemia ( 11 mmol/l), and severe hyperglycemia ( 28 mmol/l) under either normoxia or moderate hypoxia (PaO 2 50 mm Hg). Results Preischemic hyperglycemia led to a more pronounced intracellular acidosis and retardation of NADH regeneration than in the hypoglycemia groups under both normoxia and moderate hypoxia in the ischemic penumbra. For example, 4 hours after ischemia, brain ph i in the severe hyperglycemia/normoxia group measured 6.46, compared with 6.84 in the hypoglycemia/normoxia group (P 0.01), and NADH fluorescence measured 173% compared with 114%. Infarct volume in the severe hyperglycemia/normoxia group measured % of total hemispheric volume, compared with % in the hypoglycemia/normoxia group (P 0.01). Conclusions Hyperglycemia significantly worsened both cortical intracellular brain acidosis and mitochondrial function in the ischemic penumbra. This supports the hypothesis that the evolution of acidosis in the ischemic penumbra is related to glucose utilization. Furthermore, the observation that hypoglycemia significantly decreased infarct size supports the postulate that cortical acidosis leads to recruitment of ischemic penumbra into infarction. (Stroke. 1999;30: ) Key Words: acidosis redox, NADH cerebral infarction glucose cerebral ischemia, focal rabbits When the cortical surface is imaged during focal cerebral ischemia, there is variation in the development of intracellular acidosis. 1 3 In regions of severe cerebral blood flow reductions or evolving infarction, the decline in intracellular ph (ph i ) is relatively uniform and profound. In the border-zone region of moderate ischemia surrounding the evolving infarction, however, there is a heterogeneous distribution of intracellular acidosis. The significance of this intracellular acidosis in the ischemic penumbra is unclear. The mechanisms by which acidosis contributes to neuronal injury may include facilitating free radical formation, activating ph-dependent endonucleases with DNA fragmentation, or altering intracellular Ca 2 regulation Despite the probable deleterious effects of acidosis, the effects of hyperglycemia in focal cerebral ischemia remain controversial For example, it has been published that hyperglycemia reduces, does not alter, or increases damage after transient See Editorial Comment, page 170 focal cerebral ischemia It has also been published that hypoglycemia decreases the degree of pan-necrosis. 13 It is possible that varying effects of hyperglycemia may be due to differences in collateral blood flow and therefore the degree of lactic acid production. 15 The effects of hyperglycemia may also be dependent on reperfusion. 16 Adding to the controversy are in vitro observations that acidosis may ameliorate neuronal injury caused by glutamate and anoxia. 17,18 This experiment tested the hypotheses that (1) the development of cortical intracellular acidosis in the ischemic penumbra is a result of glucose utilization and (2) this acidosis leads to recruitment of potentially salvageable tissue into infarction. To test this hypothesis, in vivo fluorescence imaging was used to measure brain ph i, regional cortical blood flow (rcbf), and the NADH redox state in the New Received July 30, 1998; final revision received September 16, 1998; accepted October 20, From the Thoralf M. Sundt, Jr, Neurosurgical Research Laboratory, Mayo Clinic and Mayo Graduate School of Medicine, Rochester, Minn. Reprint requests to R.E. Anderson, Department of Neurosurgery, Mayo Clinic, 200 First St SW, Rochester, MN anderson.robert@mayo.edu 1999 American Heart Association, Inc. Stroke is available at 160

2 Anderson et al January Zealand White rabbit. 19 Histological assessment of infarction was performed in the acute setting with tetrazolium staining. Materials and Methods Animal Preparation After review and approval by the Institutional Animal Care and Use Committee, 30 New Zealand White rabbits weighing between 3.5 and 4.5 kg that had been fasted overnight were induced with sodium pentothal 40 mg/kg, operated on under 2.5% halothane anesthesia, and studied under 1.5% halothane anesthesia, respectively. A tracheotomy was performed, and the animals were placed on a Harvard respirator (Harvard Apparatus). The animals were given 0.15 mg/kg pancuronium bromide (Pavulon, Organon Inc) to abolish respiratory efforts. The animals were kept normoxic and normocarbic during the surgical procedure and through the experimental procedure with supplemental CO 2 and O 2. A craniectomy was performed with a high-speed air drill (Hall Surgical, Division of Zimmer) with the aid of an Olympus operating microscope. The majority of the frontal and parietal cortex was exposed for imaging. The dura was removed and carefully cauterized at the margins of the craniectomy, then covered with Saran Wrap to prevent surface oxygenation and to keep the brain moist. Blood loss for the surgical preparation did not exceed 5 ml. After the surgical preparation, the animal was moved from the operating table and placed on an intravital-type microscope stand. The microscope was focused on an area centered about the suprasylvian gyrus, with 1.5 cm 2 of cortex imaged for brain ph i, rcbf, and NADH fluorescence measurements. Arterial blood pressure was measured by a Statham strain gauge attached to the femoral artery catheter and recorded on a Grass model 78 polygraph. The animals were kept normothermic ( C) by use of a heating blanket (K-Pad, Gorman-Rupp), and core body temperature was monitored with a rectal digital thermometer. PaCO 2,PaO 2, and ph a (arterial) measurements were performed on a London Radiometer blood gas analyzer (PHM-73). Serum glucose and lactate were measured on a YSI 2300 Stat glucose/l-lactate analyzer. Brain temperature was not measured in this study; however, it was maintained at core body temperature with an air heater (Air-Therm, WPI). This device blows temperature-regulated air at a low constant velocity toward the surface of the brain. The temperature at the surface of the brain is regulated to within 0.2 C of core body temperature. Severe focal cerebral ischemia of the parietal cortex was induced by cauterization of the right middle cerebral artery (MCA). To reduce the potential for collateral blood flow, the right vertebral artery at the third cervical vertebra, the contralateral common carotid artery, and the anterior segment of the MCA were cauterized before the beginning of the experimental protocol. Thirty New Zealand White rabbits, divided into 6 groups of 5 each, were subjected to 4 hours of focal ischemia without reperfusion in the presence of the following preexisting conditions: (1) moderate hyperglycemia/normoxia (glucose, 11 mmol/l; PaO 2, 150 mm Hg), (2) severe hyperglycemia/normoxia (glucose, 28 mmol/l; PaO 2, 150 mm Hg), (3) hypoglycemia/normoxia (glucose, 2.8 mmol/l; PaO 2, 150 mm Hg), (4) moderate hyperglycemia/moderate hypoxia (glucose, 11 mmol/l; PaO 2, 50 mm Hg), (5) severe hyperglycemia/moderate hypoxia (glucose, 28 mmol/l; PaO 2, 50 mm Hg), and (6) hypoglycemia/moderate hypoxia (glucose, 2.8 mmol/l; PaO 2, 50 mm Hg). Hypoglycemia was induced with an intermediate-action (1- to 1.5-hour onset), long-duration ( 12 hours) isophane insulin suspension (NPH Iletin I, Lilly) at 20 U/kg SC 2 hours before ischemia. Severe hyperglycemia was induced with a bolus infusion of 10 ml (0.5 g/ml IV) dextrose (Abbot Laboratories) 1 hour before ischemia, supplemented by 10 ml/h IV continuous perfusion during the ischemic period. Brain ph i, rcbf, and intrinsic NADH were measured with in vivo fluorescence imaging. 20 Five additional animals were used as sham nonischemic controls to assess the stability of the preparation. Brain ph i, NADH redox state, rcbf, systemic parameters including serum glucose and lactate, and core body temperature were measured at 30-minute intervals throughout the animal preparation and ischemic period. In Vivo Video Fluorescent Instrumentation Instrumentation was designed to perform serial panoramic video imaging of cortical brain ph i and rcbf with umbelliferone fluorescence. 20 The optical characteristics were such that the majority or a portion of the exposed hemisphere, ph i, and rcbf could be studied simultaneously through a large craniectomy by varying the degree of magnification. The use of umbelliferone as a noninvasive in vivo technique for measuring brain ph i and CBF has been described previously. 20 Umbelliferone is nontoxic, fat soluble, and freely diffusible across the blood-brain barrier, and it rapidly equilibrates across cell membranes and is distributed through the cytoplasm as an uncharged molecule. 20 Umbelliferone was prepared for injection by dissolving 0.2 g of indicator in 200 ml of 5% glucose-saline solution at 90 C for 30 minutes. The solution was then filtered through a 0.22-mesh filter before injection. For each measurement, 1.0 ml of umbelliferone was injected retrogradely through a catheter placed in the right lingual artery. The measurements were separated by 30-minute intervals to allow for sufficient clearance of the indicator out of the brain tissue. The ph-sensitive indicator umbelliferone has 2 fluorophors, anionic and isobestic. The anionic and isobestic forms are excited at 370 and 340 nm, respectively, and have a common emission at 450 nm. The fluorescence of the anion varies directly with ph, whereas the fluorescence of the isobestic form varies directly only with the indicator concentration. Therefore, it is possible to create a nomogram from the ratio of 340- to 370-nm excitations to determine brain ph i. NADH fluorescent images were acquired before umbelliferone was injected into the lingual artery for correction of background fluorescence. Intrinsic NADH fluorescence images excited at 370 nm were stored for later analysis of mitochondrial function. The scale factor for the percent change in NADH fluorescence from baseline is set so that at 100%, the level represents the level of NADH fluorescence in normal brain, whereas an increase to 300% represents brain death. The scale factor is confirmed by random measurement of NADH fluorescence levels at death and comparing it with baseline nonischemic values in the same animal. It is important to note that a primary source of artifact in the measurement of NADH fluorescence is hemoglobin interference. It has been shown previously that use of bright-field illumination, as used in this optical system to reduce scatter, minimizes the effect of this type of interference. 21 Sundt et al, 22 in a monkey model, correlated biopsy analysis of NADH and NADPH with that of NADH fluorescence measurements in normal brain and brain at death. Use of both techniques showed that there was a close relationship in the degree of change between normal brain and brain at death. The images from the 340-nm excitation were processed to compute rcbf from the 1-minute initial slope index with a partition coefficient of unity for umbelliferone. 20 The rcbf image was then displayed and stored on tape for final analysis. For processing of the ph i image, ratios of the paired images from the 340- and 370-nm excitations were taken, and the resultant ph i image was then displayed and stored on tape for final analysis. rcbf and ph i as measured by umbelliferone are imaged in those areas that are relatively avascular in surface conducting vessels and therefore contain primarily arterioles and capillary beds. 20 The imaging system facilitates the measurement of rcbf and brain ph i by allowing the investigator to outline cortical areas of interest that are devoid of major surface conducting vessels. Histological Analysis It was determined by the investigators and consulting veterinarian that permitting the animals to survive for several days to allow for infarct maturation would not be acceptable under current institutional animal care guidelines, given analgesic expectations. Recognizing these limitations, the animals were killed at the end of the ischemic period for acute histological assessment, which did not allow for infarct maturation. At the end of each experiment, the brain was removed. To increase the firmness of the brain for sectioning, it was immersed in saline and chilled for 30 minutes. The brain was then sliced into 4-mm coronal sections, yielding slices (Figure 1) that represented 2 anterior locations, 2 central locations, and 2 posterior

3 162 Glucose and Brain ph i Figure 1. Typical photographs of acute histological assessment with TTC staining after 4 hours of severe focal cerebral ischemia without reperfusion in the New Zealand White rabbit. A, Normoxic/hypoglycemic animal; B, normoxic/moderately hyperglycemic animal; C, normoxic/severely hyperglycemic animal; D, moderately hypoxic/hypoglycemic animal; E, moderately hypoxic/moderately hyperglycemic animal; and F, moderately hypoxic/severely hyperglycemic animal. Numbers 1 through 4 in photographs refer to locations of the sections 2, 4, 6, and 8 mm from the frontal pole, respectively. locations. The sections were then immersed in a 37 C solution of 2% 2,3,5-triphenyltetrazolium chloride (TTC) in saline. To enhance TTC penetration, the sections were suspended within the TTC solution for 30 minutes in a shaker bath maintained at 37 C. Sections were removed from the TTC solution, placed flat on one another in separate fixation cassettes, and stored in 10% buffered formalin. Photographic slides of each section were taken 1 week later. For assessment of the amount of tissue damage, each section was photographed, and the area of infarction was identified, traced, and digitized. The total area of the hemisphere was also determined. Definition of Ischemic Penumbra The purpose of this experiment was to determine the effects of glucose and oxygen manipulations on brain ph i and NADH redox state in the ischemic penumbra. Determination of the location of the ischemic penumbra was made as follows. Thirty minutes after the onset of ischemia, in the immediate proximal distribution of the MCA along the sylvian fissure, there is a relatively small region of cortex in which rcbf declines to 12 ml 100 g 1 min 1,oran 80% reduction in rcbf values compared with preischemic measurements. Under normoglycemic conditions, this cortex has ph i reductions to This region has previously been demonstrated to be an evolving infarction with necrosis under light microscopic examination. 19 Distal to this zone of severe rcbf reductions is parietal cortex exposed by the craniectomy that has initial rcbf reductions of 20 ml 100 g 1 min 1. This cortex was defined as the ischemic penumbra in this experiment, and the tissue was analyzed by fluorescence imaging and histology. For example, Figures 2 and 3 are composite images of this cortex, which is distal to the smaller zone of early evolving infarction. Within this ischemic penumbra, there is a heterogeneous development of acidosis, which is the issue of interest for this experiment. Statistical Analysis Separate analyses were carried out for each of the 4 variables under study: ph i, rcbf, NADH fluorescence, and infarct volume. ANOVA was used to test the statistical significance of differences between groups. Results were considered statistically significant at a value of P Data are presented as mean SEM. All analysis was conducted with CSS (Statsoft) statistical software. Results Systemic Parameters, Nonischemic Control Group, and Video Acquisition There were no significant differences over time between animals studied in each of the 6 groups in the measurements of PaCO 2, PaO 2, ph a, mean arterial blood pressure, body temperature, glucose, and hematocrit (Table). The control sham-operated nonischemic animals had stable ph i, rcbf, and NADH fluorescence measurements throughout the experiment, thus confirming the stability of the preparation. The blood serum levels of PaO 2, %ScO 2, and lactate in the moderate hypoxia groups were significantly different from those of the normoxia groups (P 0.01). Figures 2 and 3 show composite video pictures of 2

4 Anderson et al January Figure 2. Composite photograph is an example of a hypoglycemia/moderate hypoxia experiment. Before ischemia, brain ph i is uniform across the cortical surface, measuring 7.03 in this experiment. Note heterogeneity in focal CBF before ischemia. Yellow pixels are in areas of the larger surface conducting vessels. Over the next 4 hours of ischemia after occlusion of the MCA, although there is a marked reduction in CBF to 10 ml 100 g 1 min 1, brain ph i is virtually unchanged, measuring NADH redox state increased by an average of 40% over this time period. typical experiments: a hypoglycemic/moderately hypoxic animal and a hyperglycemic/normoxic animal, respectively. Experimental Groups (Figure 4) Normoxic Groups Brain ph i. Baseline preischemic brain ph i was uniform over the exposed cortex in all groups, measuring Within the ischemic penumbra, there was a heterogeneous development of acidosis, as illustrated in an example experiment, Figure 3. After 2 hours of ischemia, overall brain ph i in the ischemic penumbra declined to and in the moderate and severe hyperglycemia groups, respectively (P 0.01 compared with preischemic values). In the hypoglycemia group, brain ph i declined to , which was not significantly different from preischemic values. After 4 hours of ischemia, brain ph i declined further, to in both the moderate and severe hyperglycemia groups (P 0.01). In the hypoglycemia group, brain ph i was , not significantly different from preischemic values. Figure 5 depicts a series of brain ph i histograms of a hyperglycemic/normoxic animal (Figure 3), demonstrating that before occlusion, brain ph i was relatively homogeneous and then became markedly heterogeneous after the onset of focal ischemia. Regional Cortical Blood Flow. rcbf in all groups was ml 100 g 1 min 1 before ischemia. After 2 hours of ischemia, rcbf in the ischemic penumbra fell significantly (P 0.01) in all groups studied, to 15 ml 100 g 1 min 1. After 4 hours of ischemia, rcbf declined further to 10 ml 100 g 1 min 1 in all groups (P 0.01). Figure 5 depicts a series of rcbf histograms of a hyperglycemic/normoxic animal (Figure 3) demonstrating that before occlusion, rcbf was relatively heterogeneous and then became more heterogeneous after MCA occlusion. NADH Fluorescence. NADH fluorescence was uniform over the exposed cortex in all groups, measuring % before ischemia. After 2 hours of ischemia, NADH fluorescence levels in the ischemic penumbra increased to 152% in both the moderate and severe hyperglycemia groups (P 0.01). In the hypoglycemia group, NADH fluorescence increased to % (P 0.01). After 4 hours of ischemia, NADH fluorescence increased further, to % and % in the moderate and severe hyperglycemia groups, respectively (P 0.01 compared with preischemic values). However, in the hypoglycemia group, NADH redox state improved, measuring %, which was not significantly different from preischemic values. Figure 5 depicts a series of NADH redox state histograms of a hyperglycemic/ normoxic animal (Figure 3) demonstrating increased heterogeneity during the period of occlusion. Areas of Infarction as Measured With TTC. Infarct volume in the moderate and severe hyperglycemia groups was % and % of total hemisphere volume. The hypoglycemia group showed a significantly (P 0.01) smaller infarct volume ( %) than the other 2 groups. Analysis showed that

5 164 Glucose and Brain ph i Figure 3. This composite photograph is an example of a hyperglycemia/normoxia experiment. Before ischemia, brain ph i is uniform across the cortical surface, measuring 7.02 in this experiment. Note that there is heterogeneity in focal CBF before ischemia. Yellow pixels are in areas of the larger surface conducting vessels. Over the next 4 hours of ischemia after occlusion of the MCA, although there is a marked reduction in CBF, brain ph i becomes progressively acidotic, and NADH redox state increased by an average of 87% over this time period. hemispheric volumes in all normoxic study groups were not significantly different, indicating that there was no significant early edema formation, which might alter infarct volumes. Moderately Hypoxic Groups Brain ph i. Brain ph i was uniform over the entire exposed cortex in all groups, measuring before ischemia. After 2 hours, brain ph i in the ischemic penumbra fell significantly in the moderate and severe hyperglycemia groups, to 6.50 (P 0.01 compared with preischemic values). In the hypoglycemia group, brain ph i declined to , which was not statistically different from preischemic values. After 4 hours of ischemia, brain ph i declined further, to and in the moderate and severe hyperglycemia groups, respectively (P 0.01 compared with preischemic values). In the hypoglycemia group, brain ph i measured , which was not statistically different from preischemic values. Figure 6 depicts a series of histograms of a hypoglycemic/moderately hypoxic animal (Figure 2) demonstrating homogeneity of ph i before occlusion and throughout the ischemic period. Regional Cortical Blood Flow. rcbf in all groups was ml 100 g 1 min 1 before ischemia. After 2 hours of ischemia, rcbf fell significantly in the ischemic penumbra (P 0.01) in all groups studied, to 16 ml 100 g 1 min 1. After 4 hours of ischemia, rcbf further declined to 11 ml 100 g 1 min 1 in all groups (P 0.01). Figure 6 depicts a series of rcbf histograms of a hypoglycemic/moderately hypoxic animal (Figure 2) demonstrating that before occlusion, rcbf was relatively heterogeneous and then became more heterogeneous with reductions in rcbf after MCA occlusion. NADH Fluorescence. NADH fluorescence was uniform over the exposed cortex in all groups, measuring % before ischemia. After 2 hours of ischemia, NADH fluorescence levels in the ischemic penumbra increased to 173% in both the moderate and severe hyperglycemia groups (P 0.01 compared with preischemic values). In the hypoglycemia group, NADH fluorescence increased to %. After 4 hours of ischemia, NADH fluorescence increased further, to % and % in the moderate and severe hyperglycemia groups, respectively (P 0.01 compared with preischemic values). In the hypoglycemia group, NADH redox state increased slightly, to an overall increase of %, which was not statistically different from preischemic values. Figure 6 depicts a series of NADH redox state histograms of a hypoglycemic/moderately hypoxic animal (Figure 2) demonstrating very little change in the histograms during ischemia. Areas of Infarction as Measured With TTC. Infarct volume in the moderate and severe hyperglycemia groups measured % and % of total hemisphere volume. The hypoglycemia group showed a significantly smaller infarct volume ( %) compared with the other 2 groups

6 Anderson et al January Systemic Parameters at 4 Hours of Occlusion MABP, PaCO 2, mm Hg ph a mm Hg PaO 2, mm Hg ScO 2, % Lactate, mmol/l Glucose, mmol/l HCT, % Rectal Temperature, C Normoxia groups Hypoglycemia Moderate hyperglycemia Severe hyperglycemia Moderate hypoxia groups Hypoglycemia Moderate hyperglycemia Severe hyperglycemia MABP indicates mean arterial blood pressure. Values are mean SEM; n 5 per group. (P 0.05). Analysis of hemispheric volumes in all moderate hypoxia study groups were not significantly different, indicating that there was no significant early edema formation, which might alter infarct volumes. Discussion The significant findings of the study are as follows. First, in both the normoxia and moderate hypoxia groups, hyperglycemia facilitates development of cortical acidosis in the ischemic penumbra. Second, both moderate and severe hyperglycemia were associated with worsening mitochondrial function in the ischemic penumbra, as assessed by an increase in NADH fluorescence. Alternatively, hypoglycemia led to improvement, with a reduction in NADH redox state. Third, hypoglycemia resulted in a significant reduction in the size of early infarcts as measured by TTC staining. There was not a significant difference in size of infarct between the moderate and severe hyperglycemia groups, because the maximum degree of cortical infarction was realized in both groups. These findings were observed in severe focal cerebral ischemia of 4 hours duration without reperfusion. Together, these results support the hypotheses that the development of cortical intracellular acidosis in the ischemic penumbra is dependent on glucose utilization and that this acidosis leads to increases in cortical infarction size. Hyperglycemia Hyperglycemia has been known to worsen neurological damage in different models of cerebral ischemia. In clinical studies, diabetic patients who have experienced a stroke appear to have a significantly worse outcome than nondiabetic patients Mortality rates have been shown to be increased in acute stroke patients who have fasting levels of glucose 6.1 mmol/l. 26 In a more recent study, Toni et al 27 suggested that when serum glucose levels were lowered in diabetic stroke patients, a better outcome was achieved. A number of investigators have studied the effects of hyperglycemia in models of focal cerebral ischemia. These studies have shown conflicting results, either worsening of ischemic damage 11,15,28 32 or amelioration of ischemic damage. 13,33 37 The discrepancy in these studies could be explained in part by the degree of collateral blood supply. In models of focal incomplete cerebral ischemia, in which there was some collateral flow ( trickle flow phenomenon), hyperglycemia appeared to exacerbate ischemic damage. 15,31,38 However, in other studies of focal cerebral ischemia, in which there was a greater amount of collateral blood flow, the effects of hyperglycemia on ischemic damage were less. 13,34,35 In our study of the ischemic penumbra, increasing the serum glucose levels before the onset of focal ischemia resulted in greater brain intracellular acidosis, increased the heterogeneity of ph i, adversely effected NADH redox state, and significantly increased infarct volume compared with the hypoglycemia group. The increased heterogeneity of ph i with increasing plasma glucose concentration observed in this study is in agreement with that of LaManna et al. 2 In a model of cardiac arrest and using neutral red histophotometry, they showed that when the plasma glucose concentration was increased, there was greater acidosis and increased heterogeneity because of greater accumulation of tissue lactate. They also demonstrated that during ischemia, ph i and tissue lactate accumulation are linearly correlated. Griffith et al 1 also showed increased heterogeneity during ischemia in a model of global ischemia. Hypoglycemia Preischemic moderate hypoglycemia in models of focal cerebral ischemia has not been studied as extensively as in models of global and forebrain ischemia. It is known that hypoglycemic coma without occlusion of the MCA results in neuronal damage in selective areas of the brain Kristián et al 41 showed that during hypoglycemic coma, brain ph i did not become acidotic when the animals were normocapnic, although there were selective areas of neuronal damage. When the animals were made hypercapnic, however, brain ph i became more acidotic, with greater regions of neuronal damage and pan-necrosis. In a model of forebrain ischemia, Smith et al 12 showed that during ischemia, brain ph i was less acidotic in animals with hypoglycemia than in animals with normoglycemia, versus In this study, the plasma glucose level during hypoglycemia was mmol/l. Conversely, in a model of global ischemia using 4-vessel occlusion, Nagai et al 42 noted that ph i, which became acidotic during ischemia, was not different in either the normoglycemic or hypoglycemic setting. The plasma

7 166 Glucose and Brain ph i Figure 4. A, Brain ph i in the ischemic penumbra after 4 hours of normoxia or hypoxia under hypoglycemic, moderately hyperglycemic, and severely hyperglycemic conditions. Dashed line is mean brain ph i before ischemia. Despite marked cerebral blood flow reductions, hypoglycemia significantly attenuated the development of cortical acidosis. B, NADH redox state 4 hours after hypoglycemia, moderate hyperglycemia, or severe hyperglycemia during either normoxia or moderate hypoxia in the ischemic penumbra. Hypoglycemia attenuated the change in NADH redox state despite significant cerebral blood flow reduction. C, rcbf in the ischemic penumbra under various glucose conditions during both normoxia and moderate hypoxia. Dashed line represents mean cerebral blood flow before ischemia, which measured ml 100 g 1 min 1 in all groups. The degree of cerebral blood flow reduction was similar in all groups studied. D, Cortical infarct volume as a percentage of hemispheric volume after 4 hours of ischemia during hypoglycemia, moderate hyperglycemia, or severe hyperglycemia under normoxic or hypoxic conditions. Hyperglycemia in both the normoxia and moderate hypoxia groups increased the size of cortical gray mantle infarction compared with the moderate and severe hyperglycemia groups. Open bars indicate normoxia; cross-hatched bars, moderate hypoxia. Data are expressed as mean SEM. *Statistically different from preischemic values, P **Statistically different from moderate hyperglycemia/normoxia, P Statistically different from moderate hyperglycemia/normoxia group, P glucose level was not specified in that study. There have been several reports in which insulin reduced ischemic damage Hamilton et al 45 showed that by reducing the blood glucose to minimum values but within the physiological range (moderate hypoglycemia), infarct volume could be significantly reduced. In this study, untreated animals had a cortical infarct volume of mm 3, whereas the insulin-treated animals had reductions in infarct volume to mm 3 (43.5%). This is in close agreement with our study, in which we showed a reduction in infarct volume by 47.7%. Furthermore, this experiment demonstrates that hypoglycemia decreases ph i heterogeneity changes during ischemia. This supports the suggestion that moderate hypoglycemia or the avoidance of hyperglycemia may be beneficial for neurosurgical procedures in preventing ischemic damage as a result of temporary arterial occlusion. 46 Moderate Hypoxia/Ischemia Many studies have investigated the effects of hypoxia on the brain Some of these studies were done in animals made hypoglycemic or hyperglycemic. 50,51 Other studies used the Levine rat preparation, which is a model of global ischemia with hypoxia as a method to further exacerbate tissue damage. 54 To the best of our knowledge, no studies of moderate hypoxia in focal cerebral ischemia have been performed. In our study, we noted 3 distinct findings in animals studied with moderate hypoxia: (1) there was no difference in infarct volume, ph i,or NADH redox state between the normoxic and moderately hypoxic animals during moderate hyperglycemia; (2) moderate hypoxia exacerbated brain intracellular acidosis compared with the normoxic animals in the severe hyperglycemia group; and (3) in hypoglycemia groups, although brain ph i was slightly but not significantly alkalotic in the moderate hypoxia group, infarct volume was 180% greater than in the normoxia group. It has been documented that during moderate hypoxia (PaO 2 45 to 55 mm Hg), there are biochemical alterations in brain tissue levels of energy metabolites, NADH redox state, and ph i. ATP, ADP, and AMP do not change unless the PaO 2 is 25 mm Hg, whereas phosphocreatine, NADH redox state, and ph i change at 35 mm Hg. 52,55 However, the lactate and pyruvate levels begin

8 Anderson et al January Figure 5. Composite histogram of the experiment in Figure 3. Before ischemia, the distribution of rcbf is quite heterogeneous compared with brain ph i or NADH redox state. Over the next 2 hours after MCA occlusion, rcbf becomes less heterogeneous, then at hours 3 and 4 becomes more heterogeneous, but with rcbf values significantly less than preischemic values. After occlusion of the MCA, brain ph i becomes very acidotic and heterogeneous over the 4-hour period. The distribution of NADH redox state becomes more heterogeneous over this same time frame. to increase when the PaO 2 is 50 mm Hg. This would explain in part why the severely hyperglycemic/moderately hypoxic animals were markedly more acidotic than the severely hyperglycemic/normoxic animals. It is interesting to note that the increase in NADH fluorescence between the hypoglycemia and hyperglycemia groups was not enhanced with the addition of moderate hypoxia. The results of this study suggest that one adverse effect of acidosis is increased damage to mitochondria. It has been shown previously that during complete and incomplete ischemia, lactic acidosis prevents normalization of mitochondrial respiration. 56,57 Wagner et al 58 demonstrated that a combination of transient anoxia in hyperglycemic cats caused altered mitochondrial respiration. Our study would support the above findings and conclusion that acidosis does adversely alter mitochondrial function during the acute ischemic insult. In fact, the observation that changes in systemic O 2 did not significantly alter NADH fluorescence supports the contention that the observed effects were not due to a failure of substrate delivery for aerobic metabolism but rather a direct effect of hyperglycemia on the mitochondria. The mechanisms by which acidosis might adversely affect mitochondria have been expertly discussed elsewhere. 9 The observation that the apparent adverse

9 168 Glucose and Brain ph i Figure 6. Composite histogram of experiment in Figure 2. Before ischemia, the distribution of rcbf is quite heterogeneous compared with brain ph i or NADH redox state. Over the next 2 hours after MCA occlusion, rcbf becomes less heterogeneous, then at hours 3 and 4 becomes more heterogeneous, but with rcbf values significantly less than preischemic values. The distribution of ph i is unchanged during the 4-hour period of occlusion. The distribution of NADH redox state became less homogeneous over the 4-hour period of occlusion. effects of hyperglycemia on NADH redox state occurred acutely in our experiment would suggest that free radical formation was not the cause but rather that other mechanisms, such as mitochondrial calcium overload, were at play. TTC Staining Technique Tetrazolium salts have been used to determine the area and degree of infarction in myocardial tissue obtained from patients. This technique has been extended for use in experimental animals in the study of cerebral injury as a result of unilateral temporary or permanent MCA occlusion. TTC is a water-soluble salt that is reduced to formazon by the enzyme succinate dehydrogenase in mitochondrial tissue. This in turn stains a deep red color in normal tissue. In ischemic tissue in which mitochondria have been damaged, however, there would be a lack of staining, ie, tissue will be a white or pale color. The reliability of TTC staining as an early marker ( 24 hours after MCA occlusion) of ischemic damage has been controversial.

10 Anderson et al January Infarction can be detected as early as within 1 to 2 hours; however, the color differences between red and white are subtle, making it difficult to delineate the extent of infarction. At infarct times of 3 hours, the infarcted tissue becomes distinctly delineated even before development of histological evidence of infarction. Hatfield et al 59 showed that 5 to 20 minutes after MCA occlusion, the area of damage as assessed by hematoxylin-eosin staining was significantly smaller than that assessed by TTC. At 3 to 4 hours after MCA occlusion, however, there was good correlation between hematoxylin-eosin and TTC staining. At 24 hours, infarct size was not significantly different from the 3 to 4 hours post MCA occlusion group. It can be postulated that when the animals are killed 5 to 20 minutes after MCA occlusion, the cerebral metabolic rate of oxygen is reduced immediately after occlusion. TTC can overestimate infarct size because, although mitochondrial function is compromised, it may potentially recover. Peri-infarct edema can also affect assessment of infarct volume. Other studies have also shown similar results Acknowledgments This work was supported by NIH grant RO The authors are indebted to Mary Soper for manuscript preparation. References 1. Griffith JK, Cordisco BR, Lin C-W, LaManna JC. Distribution of intracellular ph in the brain cortex after global ischemia as measured by color film histophotometry of neutral red. Brain Res. 1992;573: LaManna JC, Griffith JK, Cordisco BR, Lin C-W, Lust WD. Intracellular ph in rat brain in vivo and in brain slices. Can J Physiol Pharmacol. 1992;70:S269 S Tomlinson FH, Anderson RE, Meyer FB. Acidic foci within the ischemic penumbra of the New Zealand White rabbit. Stroke. 1993;24: Barber AA, Bernheim F. Lipid peroxidation: its measurement, occurrence, and significance in animal tissues. Adv Gerontol Res. 1967;2: Rehncrona S, Hauge HN, Siesjö BK. Enhancement of iron catalyzed free radical formation by acidosis in brain homogenates: differences in effect by lactic acid and CO 2. J Cereb Blood Flow Metab. 1989;9: Siesjö BK, Bendek G, Koide T, Westerberg E, Wieloch T. Influence of acidosis on lipid peroxidation in brain tissues in vitro. J Cereb Blood Flow Metab. 1985;5: Combs DJ, Dempsey RJ, Donaldson D, Kindy MS. Hyperglycemia suppresses c-fos mrna expression following transient cerebral ischemia in gerbils. J Cereb Blood Flow Metab. 1992;12: Barry MA, Reynolds JE, Eastman A. Etoposide-induced apoptosis in human HL-60 cells is associated with intracellular acidification. Cancer Res. 1993;53: Siesjö BK, Katsura K, Kristian T. Acidosis related damage. Adv Neurol. 1996;71: Regli L, Held MC, Anderson RE, Meyer FB. Nitric oxide synthase inhibition by L-NAME prevents brain acidosis during focal cerebral ischemia in rabbits. J Cereb Blood Flow Metab. 1996;16: Pulsinelli WA, Waldman S, Rawlinson D, Plum F. Moderate hyperglycemia augments ischemic brain damage: a neuropathologic study in the rat. Neurology. 1982;32: Smith ML, von Hanwehr R, Siesjö BK. Changes in extra- and intracellular ph in the brain during and following ischemia in hyperglycemic and in moderately hypoglycemic rats. J Cereb Blood Flow Metab. 1986;6: Nedergaard M, Diemer NH. Focal ischemia in the rat brain, with special reference to the influence of plasma glucose concentration. Acta Neuropathol. 1987;73: Nedergaard M. Transient focal ischemia in hyperglycemic rats is associated with increased cerebral infarction. Brain Res. 1987;408: Prado R, Ginsberg MD, Dietrich WD, Watson BD, Busto R. Hyperglycemia increases infarct size in collaterally perfused but not end-arterial vascular territories. J Cereb Blood Flow Metab. 1988;8: Yip PK, He YY, Hsu CY, Garg N, Marangos P, Hogen EL. Effect of plasma glucose on infarct size in focal cerebral ischemia-reperfusion. Neurology. 1991;41: Tombaugh GC, Sapolsky RM. Mild acidosis protects hippocampal neurons from injury induced by oxygen and glucose deprivation. Brain Res. 1990;506: Tombaugh GC, Sapolsky RM. Evolving concepts about the role of acidosis in ischemic neuropathology. J Neurochem. 1993;61: Meyer FB, Anderson RE, Sundt TM Jr, Yaksh TL. Intracellular brain ph, indicator tissue perfusion, electroencephalography, and histology in severe and moderate focal cortical ischemia in the rabbit. J Cereb Blood Flow Metab. 1986;6: Anderson RE, Meyer FB, Tomlinson FH. Focal cortical distribution of blood flow and brain ph i determined by in vivo fluorescent imaging. Am J Physiol. 1992;263:H565 H Anderson RE, Sundt TM Jr. Comparison of dark-field and bright-field incident illumination for in-vivo measurements of reduced pyridine nucleotides. Anal Biochem. 1978;91: Sundt TM Jr, Anderson RE, Sharbrough RW. Effect of hypocapnia, hypercapnia, and blood pressure on NADH fluorescence, electrical activity, and blood flow in normal and partially ischemic monkey cortex. J Neurochem. 1976;27: Pulsinelli WA, Levy DE, Sigsbee B, Schere P, Plum F. Increased damage after ischemic stroke in patients with hyperglycemia with or without established diabetes mellitus. Am J Med. 1983;74: Berger L, Hakim AM. The association of hyperglycemia with cerebral edema in stroke. Stroke. 1986;17: Murros K, Fogelholm R, Kettunen S, Vuorela AL, Valva J. Blood glucose, glycosylated haemoglobin and outcome of ischemic brain infarction. J Neurol Sci. 1992;111: Candelise L, Landi G, Orazio EN, Boccardi E. Prognostic significance of hyperglycemia in acute stroke. Arch Neurol. 1985;42: Toni D, De Michele M, Fiorelli M, Bastianello S, Camerlingo M, Sacchetti ML, Argentino C, Fieschi C. Influence of hyperglycemia on infarct size and clinical outcome of acute ischemic stroke patients with intracranial arterial occlusion. J Neurol Sci. 1994;123: Siemkowicz E, Hansen AJ. Clinical restitution following cerebral ischemia in hypo-, normo-, and hyperglycemic rats. Acta Neurol Scand. 1978;58: Ginsberg MD, Welsh FA, Budd WW. Deleterious effect of glucose pretreatment on recovery from diffuse cerebral ischemia in the cat, I: local cerebral blood flow and glucose utilization. Stroke. 1980;11: Venables GS, Miller SA, Gibson G, Hardy JA, Strong AJ. The effects of hyperglycaemia on changes during reperfusion following focal cerebral ischaemia in the cat. J Neurol Neurosurg Psychiatry. 1985;48: Marsh WR, Anderson RE, Sundt TM Jr. Effect of hyperglycemia on brain ph levels in areas of focal incomplete cerebral ischemia in monkeys. J Neurosurg. 1986;65: De Courten-Meyers G, Meyers RE, Schoolfield L. Hyperglycemia enlarges infarct size in cerebrovascular occlusion in cats. Stroke. 1988;19: Jernigan J, Evans OB, Kirshner HS. Hyperglycemia and diabetes improve outcome in a rat model of anoxia/ischemia. Neurology. 1984;34(suppl 1):262. Abstract. 34. Nedergaard M, Astrup J. Infarct rim: effect of hyperglycemia on direct current potential and [ 14 C]2-deoxyglucose phosphorylation. J Cereb Blood Flow Metab. 1986;6: Ginsberg MD, Prado R, Dietrich WD, Busto R, Watson BD. Hyperglycemia reduces the extent of cerebral infarction in rats. Stroke. 1987;18: Zasslow MA, Peral RG, Shuer LM, Steinberg GK, Lieberson RE, Larson CP Jr. Hyperglycemia decreases acute neuronal ischemic changes after middle cerebral artery occlusion in cats. Stroke. 1989;20: Kraft SA, Larson CP Jr, Shuer LM, Steinberg GK, Benson GV, Pearl RG. Effect of hyperglycemia on neuronal changes in a rabbit model of focal cerebral ischemia. Stroke. 1990;21: Rehncrona S, Rosen I, Siesjö BK. Excessive cellular acidosis: an important mechanism of neuronal damage in the brain? Acta Physiol Scand. 1980;110: Auer RN, Olsson Y, Siesjö BK. Hypoglycemic brain injury in the rat: correlation of density of brain damage with the EEG isoelectric time: a quantitative study. Diabetes. 1984;33: Auer RN, Siesjö BK. Biological differences between ischemia, hypoglycemia and epilepsy. Ann Neurol. 1988;24: Kristián T, Gidö G, Siesjö BK. The influence of acidosis on hypoglycemic brain damage. J Cereb Blood Flow Metab. 1995;15: Nagai Y, Naruse S, Weiner MW. Effect of hypoglycemia on changes of brain lactic acid and intracellular ph produced by ischemia. NMR Biomed. 1993;6: Lemay DR, Gehua L, Zelenock GB. Insulin administration protects neurologic function in cerebral ischemia in rats. Stroke. 1988;19:

11 170 Glucose and Brain ph i 44. Voll CL, Auer RN. The effect of postischemic blood glucose levels on ischemic brain damage in the rat. Ann Neurol. 1988;24: Hamilton MG, Tramner BI, Auer RN. Insulin reduction of cerebral infarction due to transient focal ischemia. J Neurosurg. 1995;82: Siesjö BK, Jóhannsson H, Norberg K, Salford LG. Brain function, metabolism and blood flow in moderate and severe arterial hypoxia. In: Ingvar DH, Lassen NA, eds. Brain Work. Alfred Benzon Symposium VIII. Copenhagen, Denmark: Munksgaard; 1975: Norberg K, Siesjö BK. Cerebral metabolism in hypoxic hypoxia, I: pattern of activation of glycolysis: a re-evaluation. Brain Res. 1975;86: Kogure K, Scheinberg P, Utsunomiya Y, Kishikawa H, Busto R. Sequential cerebral biochemical and physiological events in controlled hypoxemia. Ann Neurol. 1977;2: Emoto SE, Kinter D, Feyzi JM, Gilboe DD. Relating cerebral ischemia and hypoxia to insult intensity. J Neurochem. 1988;50: Silver IA, Erecińska M. Extracellular glucose concentration in mammalian brain: continuous monitoring of changes during increased neuronal activity and upon limitation in oxygen supply in normo-, hypo-, and hyperglycemic animals. J Neurosci. 1994;14: Rosenberg GA, White J, Gasparovic C, Crisostomo EA, Griffey RH. Effect of hypoxia on cerebral metabolites measured by proton nuclear magnetic resonance spectroscopy in rats. Stroke. 1991;22: Siesjö BK, Nilsson L. The influence of arterial hypoxemia upon labile phosphates and upon extracellular and intracellular lactate and pyruvate concentrations in the rat brain. Scand J Clin Lab Invest. 1971;27: Coutice FC. The effect of oxygen lack on the cerebral circulation. J Physiol. 1941;100: Levine S. Anoxic-ischemic encephalopathy in rats. Am J Pathol. 1960; 36: Nioka S, Smith DS, Chance B, Subramanian HV, Butler S, Katzenberg M. Oxidative phosphorylation system during steady-state hypoxia in the dog brain. J Appl Physiol. 1990;68: Rehncrona S, Mela L, Siesjö BK. Recovery of brain mitochondrial function in the rat after complete and incomplete ischemia. Stroke. 1979;10: Hillered L, Ernster L, Siesjö BK. Influence of in vitro lactic acidosis and hypercapnia on respiratory activity of isolated rat brain mitochondria. J Cereb Blood Flow Metab. 1984;4: Wagner KR, Kleinholz M, Myers RE. Delayed onset of neurologic deterioration following anoxia/ischemia coincides with appearance of impaired brain mitochondrial respiration and decreased cytochrome oxidase activity. J Cereb Blood Flow Metab. 1990;10: Hatfield RH, Mendelow AD, Perry RH, Alvarez LM, Modha OP. Triphenyltetrazolium chloride (TTC) as a marker for ischemic changes in rat brain following middle cerebral artery occlusion. Neuropathol Appl Neurobiol. 1991;17: Liszczak TM, Hedley-Whyte ET, Adams JF, Han DH, Kolluri VS, Vacanti FX, Heros RC, Zervas NT. Limitations of tetrazolium salts in delineating infarcted brain. Acta Neuropathol (Berl). 1984;65: Han DH, Zervas N, Geyer RP, Adams JS, Ropper AH, Kennedy J, Heros RC, Barsos BG, Borges L, Hedley-Whyte T. Can perfluorochemicals reduce cerebral ischemia? In: Reivich M, Hurtig HR, eds. Intracranial Vascular Disease. New York, NY: Raven Press; 1983: Bederson JB, Pitts LH, Germano SM, Nishimura MC, Davis RL, Bartkowski HM. Evaluation of 2,3,5-triphenyltetrazolium chloride as a stain for detection and quantification of experimental cerebral infarction in rats. Stroke. 1986;17: Park CK, Mendelow AD, Graham DI, McCulloch J, Teasdale GM. Correlation of triphenyltetrazolium chloride perfusion staining with conventional neurohistology in the detection of early brain ischemia. Neuropathol Appl Neurobiol. 1988;14: Isayama K, Pitts LH, Nishimura MC. Evaluation of 2,3,5- triphenyltetrazolium chloride staining to delineate rat brain infarcts. Stroke. 1991;22: Intraischemic hyperglycemia has been widely demonstrated to exacerbate brain injury in a variety of animal models by enhancement of intracellular acidosis, consequent loss of ion homeostasis, and through bioenergetic failure. A host of potential cellular mechanisms have been identified, with particularly good evidence for the involvement of iron-catalyzed production of reactive oxygen species. 1 3 It is increasingly clear that lactic acidosis amplifies multiple subcellular, and potentially mitochondrial, injury cascades during cerebral ischemia. However, results in focal cerebral ischemia have not completely agreed on the deleterious effect of high glucose, in part because glycemic state may alter collateral blood flow and may therefore infuence the fate of penumbral tissues. 4 The preceding article uses sophisticated, panoramic video imaging to evaluate ischemic intracellular ph, regional cortical blood flow, and NADH fluorescence in a carefully controlled rabbit model of permanent focal cerebral ischemia. The data show that preischemic hyperglycemia worsens tissue infarction when compared with the hypoglycemic condition, and this is associated with greater tissue acidosis and increased NADH fluorescence. It is worth noting that the relationship between tissue acidosis, redox state, and tissue infarction may or may not be causal. Further, because the animal model is quite severe and not reversible, the investigators are limited to study of acute pathophysiology without reperfusion. This is important, because hyperglycemic ischemia Editorial Comment has delayed effects that are of clinical interest, such as exaggerated cerebral edema, seizure development, and transformation of selective neuronal injury to pannecrosis. Nevertheless, the temporal and spatial measurement resolution of the experimental approach allows for excellent description and correlation of important, physiological variables during vascular occlusion. The data provide remarkable spatial detail and fresh quantification of tissue ph/perfusion heterogeneity that has previously been speculated to develop during hyperglycemic stroke. Although no specific cellular mechanism is elucidated, the study strongly supports the prevailing concept that acidosis recruits potentially salvageable brain into a state of no return. Patricia D. Hurn, PhD, Guest Editor Anesthesiology/Critical Care Medicine Johns Hopkins Medical Institutions Baltimore, Maryland References 1. Li PA, Siesjo BK. Role of hyperglycaemia-related acidosis in ischaemic brain damage. Acta Physiol Scand. 1997;161: Siesjo BK, Katsura K, Kristian R. Acidosis-related damage. Adv Neurol. 1996;71: Hurn PD, Traystman RJ. ph-mediated brain injury in cerebral ischemia and circulatory arrest. J Intensive Care Med. 1996;11: Prado R, Ginsberg MD, Dietrich WD, Watson BD, Busto R. Hyperglycemia increases infarct size in collaterally perfused but not end-arterial vascular territories. J Cereb Blood Flow Metab. 1988;8: