The Effect of Hyperglycemia on Intracellular Calcium in Stroke

Journal of Cerebral Blood Flow and Metabolism 12:469-476 1992 The International Society of Cerebral Blood Flow and Metabolism Published by Raven Press, Ltd., New York The Effect of Hyperglycemia on Intracellular Calcium in Stroke Nobuo Araki, Joel H. Greenberg, John T. Sladky, Daisuke Uematsu, Andrea Karp, and Martin Reivich Cerebrovascular Research Center, Department of Neurology, University of Pennsylvania, Philadelphia, Pennsylvania, U.S.A. Summary: The effect of hyperglycemia on cytosolic free calcium ([Ca2+l) during temporary focal cerebral ischemia was investigated in cats using a fluorometric technique. The middle cerebral artery (MCA) was occluded for a period of 1 h, after which the clip was removed. In seven animals, plasma glucose was raised to 5(}"7 mg/ dl by infusion of a 5% glucose solution starting 3 min after MCA occlusion, while eight animals were kept normoglycemic during and following occlusion. MCA occlusion induced a significant, but identical, elevation of the [Ca2+]j signal ratio (4/56 nm) in both the normoglycemic group (from 1.4 to 1.97 ±.34, p <.1) and in the hyperglycemic group (from 1.4 to 2. ±.53, p <.1) at the end of the occlusion. Between 1 and 3 min after reopening, the [Ca2+]j signal ratio decreased to control levels in the normoglycemic group (1.4 ±.11 and 1.36 ±.8 at 1 and 3 min after reopening, respectively), but remained elevated in the hyperglycemic group (1.69 ±.18 and 1.65 ±.21 at 1 and 3 min after reopening, respectively). There was a statistically significant difference between the two groups (p <.1). These data suggest that hyperglycemia may be harmful to calcium recovery during the early recirculation period following focal cerebral ischemia. Key Words: Hyperglycemia Cerebral ischemia-calcium-cats. The effect of hyperglycemia on the neuronal injury caused by cerebral ischemia and reperfusion remains controversial. A number of experimental studies have demonstrated that hyperglycemia exacerbates ischemic cell damage (Myers and Yamaguchi, 1977; Ginsberg et ai., 198; Welsh et ai., 198; Kalimo et ai., 1981; Rehncrona et ai., 1981; Pulsinelli et ai., 1982; Paljarvi et ai., 1983; Welsh et ai., 1983; Nedergaard, 1987), whereas other studies (Nedergaard and Astrup, 1986; Ibayashi et ai., Received July 1, 1991; final revision received December 24; accepted December 27, 1991. The present address of Dr. N. Araki is Department of Neurology, Keio University, 35 Shinanomachi, Shinjuku-ku, Tokyo 16, Japan. The present address of Dr. D. Uematsu is Department of Neurology, Urawa City Hospital, 246 Mimuro, Urawacity, Saitama 336, Japan. Address correspondence and reprint requests to Dr. J. H. Greenberg at Room 429, Johnson Pavilion, 36th & Hamilton Walk, University of Pennsylvania, Philadelphia, PA 1914, USA. Abbreviations used: ACSF, artificial cerebrospinal fluid; AM, acetoxymethyl ester; lcbf, local cortical blood flow; lcbv, local cerebral blood volume; MCA, middle cerebral artery; MTT, mean transit time; PDHC, pyruvate dehydrogenase complex. 1986; Ginsberg et ai., 1987; Zaaslow et ai., 1989) suggest that it either decreases or has no effect on the extent of cerebral injury. The possible role of calcium in irreversible cell damage during cerebral ischemia and reperfusion has been a focus of attention over the past decade (Siesj6, 1981; Siesj6 and Bengtsson, 1989). Using an in vivo fluorometric technique with indo-1, a fluorescent Ca2+ indicator, we recently demonstrated in cats that cytosolic free calcium ([Ca2+]j) in the cortex increased in severe ischemia caused by middle cerebral artery (MCA) occlusion (Uematsu et ai., 1988). The purpose of the present study was to examine the effects of glucose on cerebral ischemia and reperfusion from the viewpoint of calcium metabolism. MATERIALS AND METHODS The details of the surgical preparation have been described previously (Uematsu et ai., 1988). Adult male cats (2.(}"3.2 kg) fasted 18 h before surgery were anesthetized by halothane inhalation; induction and maintenance doses were 5. and 1.%, respectively. Following trache- 469

47 N. ARAKI ET AL. ostomy, the respiration was maintained with a mechanical ventilator (Model 622, Harvard Apparatus, Millis, MA, U.S.A.). Both femoral arteries and one femoral vein were catheterized for monitoring arterial blood pressure, to obtain arterial blood for analysis of blood gas parameters, blood glucose concentration, and plasma osmolality, and for the administration of drugs. The left lingual artery was also cannulated so that a.5 ml saline bolus could be intermittently injected to obtain cerebrovascular hemodilution curves. The rectal temperature was maintained at 37 C using a thermostat-controlled heating pad. After fixing the animal's head in a stereotaxic frame, a burr hole was made over the left middle ectosylvian gyrus, and the dura membrane was incised. A quartz cranial window equipped with silver EEG electrodes and inletoutlet tubing was placed into the burr hole and the space beneath the cranial window was filled and superfused continuously with artificial cerebrospinal fluid (ACSF) of the following composition (in mm): Na, 14; K, 2.5; Ca, 1.5; Mg, 1.2; CI, 13; HC3, 14.5; glucose, 3. 3; N-2-hydroxyethylpiperazine-N' -2-ethanesulfonic acid (HEPES), 1; ph 7.35 at 37 C. ACSF containing 7 f.1m indo-l acetoxymethyl ester (Indo-I-AM, Molecular Probe Inc., Eugene, OR, U.S.A.) at a temperature of 37 C was superfused over the cortical surface for 2 h. The membrane-permeant indo-i-am diffuses into the neuronal cells and is hydrolyzed by the action of cytoplasmic esterase to the membraneimpermeant indo-i, which is trapped in the cells as a specific indicator for Ca2+. Following this loading, the remaining extracellular dye was washed out by superfusion for 3 min with ACSF without indo-i-am. A 1 W mercury arc lamp (Osram, Munich, Germany) with a stabilized direct current power supply (Ludl Electric Products, Ltd., Scarsdale, NY, U.S.A.) was utilized as a light source. Ultraviolet excitation light (34112 nm) was focused on a small cortical area, and indo-l-ca2+ fluorescence (4 and 56 nm), NADH fluorescence (464 nm, an isosbestic point of the indo-l spectra), and ultraviolet reflection (34 nm) were selectively transmitted through a four-way quartz fiber bundle to four photomultipliers (RI14 Hamamatsu, Hamamatsu City, Japan) with appropriate barrier filters. Measurements were made from the middle ectosylvian gyrus, considered to be in the central MCA territory (Ginsberg et ai., 1976). The ultraviolet reflection is inversely related to changes in local cortical blood volume (lcbv) (Harbig et ai., 1976). Lingual artery flushes with saline permitted the measurement of the vascular mean transit time, which, along with changes in lcbv, enabled changes in local cortical blood flow (lcbf) to be calculated. The indo-l-ca2+ signals and the NADH signals were corrected for changes in ICBV based on the cerebrocortical hemodilution curves as described previously (Harbig et ai., 1976). The indo-l-ca2 + signals were corrected for the changes due to NADH fluorescence by determining the response of the indo-l signal to a brief period of anoxia prior to indo-i-am loading (Uematsu et ai., 1988). The basal indo-l-ca2+ fluorescence level in the cortical area was determined by subtracting the preloading autofluorescence level from the fluorescence level reached after indo-l loading. Following the correction for hemodynamic and NADH changes, a corrected signal ratio (4 nm/56 nm) was calculated and used as a measure of intracellular free calcium. Each channel was calibrated with an electronically generated constant signal. The four optical signals, the blood pressure, and the electroencephalogram (EEG) from each hemisphere were recorded on a seven-channel polygraph (Model 7D, Grass Instruments Co., Quincy, MA, U.S.A.). The mean transit time (MTT) was determined from an analysis of the cerebrocortical hemodilution curve produced by small bolus injections of saline into the lingual artery. The area under this dilution curve as determined by integrating from the time in which the curve deviated from the control baseline level was divided by the height of the curve to determine the MTT (Meier and Zieder, 1954). The bolus was administered over 1 s and the time duration of the bolus was the same for all injections. Changes in ICBV were measured from the changes in the reflected (34 nm) light during the study with % CBV defined as the reflectance when the blood was washed out from the region of interest, and 1% as the reflectance signal during the control (baseline period). Changes in cerebral blood flow (CBF) were determined by dividing the changes in CBV by MTT. Following the indo-l loading procedure, the exposed left MCA trunk was occluded with a miniature Mayfield clip. After 1 h of MCA occlusion, the clip was removed and the fluorescence was monitored over the first 6 min of reperfusion. Due to the potential for photobleaching of indo-l-ca2 +, the optical signals were measured intermittently for 5 s every 5 to 1 min. Thirty minutes after occlusion, an infusion of a 5% glucose solution was initiated in seven cats whose EEG amplitude was depressed and remained between 1 and 2% of the preocclusion level from 5 to 3 min after MCA occlusion. During the first 5 min, glucose was given at the rate of. 2 g/kg/min, and then changed to a rate of.25 g/kg/min until the end of the study (after 3 h of reperfusion). Eight animals whose EEG showed the same change following occlusion as the hyperglycemic group were used as controls. Three hours after the start of reperfusion, the animal was killed and the brain removed and fixed in 1% formalin for 2 weeks. Histopathologic evaluation was performed in the hematoxylin-eosin stained slides by light microscopy. Coronal sections were made at the level of the cortical window 2-2.5 mm apart in the region of the fluorometric measurements. All of the tissue sections were analyzed by one of the authors (J.T.S.), who was blinded as to whether they were from hyperglycemic or control animals. The region with histological evidence of neuronal injury was outlined on the slides. To insure that artifactual changes in the appearance of neurons were excluded from the area of injury, the ischemic hemisphere was compared with homologous regions of the contralateral hemisphere. In addition to neuronal changes, vacuolation of the neuropil was also required for a region to be classified as having histopathological evidence of ischemic injury. Consequently, some regions with minor histological changes, which may have been the consequence of local ischemia, were not classified as injured. Typically, the differences in the histological appearance between areas of injury showing shrunken pyknotic neurons with vacuolated neuropil and normal regions (or less severely damaged regions that were classified as normal) were quite striking. The outline of the region of histologic damage was marked on the stained sections, and the area of this region was determined using an image analyzer with a video camera. The two groups were compared using analysis of vari- J Cereb Blood Flow Metab, Vol. 12, No.3, 1992

HYPERGLYCEMIA AND INTRACELLULAR CALCIUM 471 ance. If significance was attained, probing tests at individual times were undertaken using Student's two-tailed t test or Scheffe's test. The results were considered statistically significant if the probability value was <.5. All values are expressed as mean ± SD. RESULTS Data on mean arterial blood pressure, arterial blood gas analyses, and plasma glucose concentration are summarized in Table 1. In the hyperglycemic group, arterial blood pressure increased by 1 mm Hg, but it recovered in 1 min and remained almost constant throughout the experiment. There were no significant differences in arterial blood gas values between the two groups. In the hyperglycemia group, plasma glucose increased to 52 ± 121 mg/dl within 1 min after the start of glucose infusion and remained elevated at 622 ± 82 and 651 ± 66 mg/dl at 3 and 6 min after starting the glucose infusion, respectively (Fig. I). Plasma osmolality was checked only in the hyperglycemic group. Before the glucose infusion, the plasma osmolality was 316.1 ± 4.8 mosm/l, but it increased to 33.9 ± 4.2 at 3 min into the glucose infusion, 334.1 ± 4.2 mosm/l at 6 min into the infusion (p <.1), and 338. ± 5.9 mosm/l at 12 min into the infusion. All postinfusion osmolality values were significantly greater than the control level (p <.1). The EEG amplitude of the occluded side was markedly reduced following MCA occlusion and remained between 1 and 2% in both groups prior to the start of glucose infusion (Fig. 2). Twenty minutes after MCA occlusion, the mean EEG amplitude was 14 ± 3% of control in the hyperglycemic group and 15 ± 8% of control in the normoglycemic group. By the end of 6 min of occlusion (3 min after the start of the glucose infusion), the mean EEG amplitude was 16 ± 13% of control in the hyperglycemic group and 14 ± 6% in the control group. During reperfusion, the mean EEG amplitudes of both groups were very similar. In both groups, the [Ca2+]j signal ratio started to increase very soon after MCA occlusion and remained elevated throughout the ischemic period (Fig. 3). Within 1 min of occlusion, the [Ca2+t signal ratio of the hyperglycemic group was 1.86 ±.25 while in the normoglycemic group this ratio was 1.78 ±.23. There was very little further increase in [Ca2+]j throughout the remainder of the occlusion, even after the start of the glucose infusion. Immediately prior to the start of reperfusion, the [Ca2+]j signal ratio was almost identical in both groups (1.97 ±.32 and 2. ±.53 in the hyperglycemic and normoglycemic animals, respectively). After reperfusion, the [Ca2+t signal ratio decreased to the control level in the normoglycemic group, while in the hyperglycemic group the [Ca2+]j decreased slightly but still remained elevated during the reperfusion period. Ten minutes after reopening, this ratio was 1.69 ±.18, and 3 min after reopening it was 1.65 ±. 21. The difference between the normoglycemic group and hyperglycemic group was statistically significant (p <.1) for the first 5 min of reperfusion. By 1 h after release of the occlusion, however, there was no significant difference between the normoglycemic and the hyperglycemic groups. The NADH/NAD redox state became reduced in both groups after MCA occlusion but slowly returned to the control level during occlusion (Fig. 4). During reperfusion, the normoglycemic group showed a rapid oxidization, while the hyperglycemic group remained at the control level, although there were no statistically significant differences between the two groups. In both the normoglycemic and the hyperglycemic animals, the lcbv decreased gradually during MCA occlusion, and recovered during reperfusion (Fig. 5). By the end of the occlusion, the lcbv in the hyperglycemic group was 88.2 ± 12.9% of control, while in the normoglycemic group the lcbv was 89.4 ± 14.9% of control. Thirty minutes into TABLE 1. Physiological data fo r cats subjected to middle cerebral artery occlusion and reperfusion Mean arterial Plasma glucose blood pressure Pa2 PaC2 Arterial concentration Group (mm Hg) (mm Hg) (mm Hg) ph (mg/dl) Before occlusion Hyperglycemic 75 ± 1 122 ± 1 33 ± 3 7.35 ±.5 157 ± 26 N ormoglycemic 74 ± 11 121 ± 8 33 ± 6 7.33 ±.6 15 ± 42 After 3 min of occlusion Hyperglycemic 72±13 127 ± 13 31 ± 3 7.35 ±.5 152 ± 32 Normoglycemic 73 ± 11 12 ± 8 34 ± 6 7.34 ±.6 16 ± 51 After 3 min of reperfusion Hyperglycemic 76 ± 16 135 ± 13 33 ± 3 7.34 ±.5 651 ± 66 Normoglycemic 73 ± 11 128 ± 14 34 ± 3 7.33 ±.6 151 ± 45 Values are mean ± SD, n = 7 for hyperglycemia and n = 8 for normoglycemia. J Cereb Blood Flow Metab. Vol. 12, No.3, 1992

472 N. ARAKI ET AL. I MeA Occlusion I Glucose Infusion Reperfusion 9 8 7 6 5 4 3 2 1 a Cl r:: C') Ul CD (") ::J C') CD ::J., =: ::J 3 (Q a. '-" FIG. 1. Plasma glucose concentration in the normoglycemic (, n = 8) and hyperglycemic (e, n = 7) animals during temporary occlusion of the middle cerebral artery (MCA) and during the reperfusion phase. The infusion of glucose 3 min into the MCA occlusion elevated plasma glucose in the hyperglycemic animals to above 5 mg/dl within 1 min after the start of infusion. a 6 12 18 24 reperfusion, the ICBV was 12.8 ± 13.2% of control in the hyperglycemic group and 16.2 ± 7.3% in the normoglycemic group, neither value significantly different from control. The MTT in the steady state, prior to occlusion, was 2.39 ±.53 sen = 15), there being no difference between the treated and untreated animals at this time. In both groups, the ICBF decreased dramatically after MCA occlusion, with no significant differences seen throughout the ischemic period (Fig. 5). During the first 3 min of reperfusion, the ICBF in the hyperglycemic group was lower than in the normoglycemic group, but by 1 h of reperfusion the ICBF in the two groups was virtually identical. At no time during the study was there significant differences in flow between the two groups. The area of damage was normalized by the total area of the left hemisphere in two coronal sections through the tissue under the window. The damaged area was 29.6 ± 1.8% in the hyperglycemic group and 33.9 ± 12.8% in the normoglycemic group, with no significant difference between the two groups. 1 '- r:: CJ 8 6 '-" CD " 4 ::J ;!: a. E 2 <{ w w a I t.4ca Occlusion I Hyperglycemia a 3 6 9 12 15 18 21 FIG. 2. The amplitude of electroencephalogram (EEG) from the ischemic hemisphere during middle cerebral artery (MCA) occlusion and reperfusion in cats. The EEG amplitude in both hyperglycemic (e, n = 7) and normoglycemic (, n = 8) groups were significantly depressed immediately after MCA occlusion, and remained between 1 and 2% throughout occlusion. During reperfusion, the EEG amplitude gradually recovered to about 25% in both groups 3 min after reperfusion. Values are mean ± SO. o :;: C :: c r:: m Vl 2.5 2. +r---. 1.5 '" C U '--' 1. a MeA Occlusion Hyperglycemia 2 4 6 8 1 12 FIG. 3. The cytosolic free calcium [Ca2+]i signal ratio during middle cerebral artery (MCA) occlusion and reperfusion in cats. The [Ca2+]i signal ratio started to increase after MCA occlusion and remained elevated throughout the ischemia, with no significant differences between the hyperglycemic (e) and normoglycemic () groups. After reperfusion, the [Ca2+] signal ratio decreased in both groups, but the hyperglycemic group remained at a relatively high level. Significant differences ('p <.1) were found between the two groups throughout the first 3 min of reperfusion. Values are mean ± SO. J Cereb Blood Flow Metab, Vol. 12, No.3, 1992

.. HYPERGLYCEMIA AND INTRACELLULAR CALCIUM 473 ::::> 5 < '-' 4 2 3 c +- 2 VI X 1 ""C T Q) - c::: 1 Cl < Z I Cl < Z -1-2 -3-4 MCA Occlusion Hyperglycemia o 2 4 6 8 1 12 FIG. 4. Time course of the changes in NADH/NAD redox state during middle cerebral artery (MCA) occlusion and reperfusion in cats. In both the hyperglycemic (e) and the normoglycemic () groups, the NADH/NAD redox state was reduced after MCA occlusion. This reduced state was maintained in the normoglycemic animals during occlusion, but in the hyperglycemic animals, the NADH/NAD redox state returned to the control level after the start of glucose infusion. During reperfusion, the normoglycemic group showed a rapid oxidization, while the hyperglycemic group remained at the control level. These differences, however, were not statistically significant. Values are mean ± SD. DISCUSSION In the fluorometric technique for the in vivo measurement of changes in cytosolic free calcium, there is significant cross-channel contamination that must be taken into account. The channels primarily sensitive to changes in the fluorescence of the calcium indicator also receive signals due to autofluorescence, NADH fluorescence, and changes in CBV. The methods by which these corrections are made are described in detail in a previous publication (Uematsu et ai., 1988). The hemodynamic correction factors that we use are identical to those described by Harbig et ai. (1976) and Dora (1984), and is based on frequent small flushes with saline to produce artificial hemodilution. Since this correction factor may change during the course of the study, it is important to determine the correction factor frequently. The four wavelengths used in our studies are all close to the isosbestic point of the oxygenated and reduced hemoglobin; therefore, the degree of hemoglobin oxygenation should not contribute appreciably to the results. The corrections for NADH fluorescence were obtained by inducing a short anoxic insult prior to loading of indo-i, when any observed changes in the calcium channels will be due to NADH fluorescence and not indo-l fluorescence. If this correction factor is made carefully, the corrected signal should represent only indo-l fluorescence, with little (or no) NADH component. These studies indicate that recovery of cytosolic free calcium during the reperfusion period following focal cerebral ischemia is delayed more in hyperglycemic animals than in the normoglycemic animals. We have previously shown that only severely stroked cats whose EEG amplitude is depressed to less than 2% of control exhibit an increase in cytosolic free calcium during ischemia (U ematsu et ai., 1988). There also appears to be a good correlation between changes in the EEG following occlusion and histopathological ischemic changes (Tanaka et ai., 1986). Therefore, we have included only animals whose EEG amplitude was depressed to between 1 and 2% of the pre occlusion level 12 L. +- 11 C u 1-9 '-' 8 > CD U -l,- 7 125 o f 1 L. +- C u 75-5 ll. CD 25 U -l Hyperglycemia I MCA Occluslonl I I I o 3 6 9 12 15 18 21 FIG. 5. Changes in local cortical blood volume (ICBV) (upper panel) and local cerebral blood flow (ICBF) (lower panel) during middle cerebral artery (MCA) occlusion and reperfusion. In both the hyperglycemic (e) and the normoglycemic () groups, the ICBV decreased gradually during MCA occlusion and then recovered during reperfusion with no significant differences observed between the two groups. The ICBF decreased dramatically after occlusion in both hyperglycemia and hypoglycemia groups, with no significant differences observed throughout the ischemia. During the first 3 min of reperfusion, the ICBF in the hyperglycemia group was lower than the normoglycemic group, but 1 h after the start of reperfusion, the ICBF in the hyperglycemia group recovered to the same level as the normoglycemic group. Values are mean ± SD. J Cereb Blood Flow Metab, Vol. 12, No.3, 1992

474 N. ARAKI ET AL. between 1 and 3 mm after the MCA was occluded. There are very little data in the literature relating to the changes in brain calcium levels during hyperglycemia. Siemkowictz and Hansen (1981) examined the effects of plasma glucose level on extracellular K +, Ca2 +, and H + following 1 min of global cerebral ischemia in the rat. They found that brain ph was lower in the hyperglycemic animals during ischemia, but that the changes in K + and Ca2 + were not significantly altered by the elevation in glucose. Warner et al. (1987) measured total tissue calcium in the cortex of rats subjected to 1 min of forebrain ischemia followed by 1.5 to 24 h of recirculation. They observed a significant increase in Ca2 + only in hyperglycemic animals that developed seizures; all other animals, both normoglycemic and hyperglycemic, did not show any significant changes in tissue calcium level. At least two significant differences exist between these experiments and the present study. Both the studies of Siemkowictz and Hensen (1981) and that of Warner et al. (1987) employed global ischemia models; the degree of ischemia induced in their models may differ significantly from that in our focal model. In addition, their measurements of either extracellular calcium (Siemkowictz and Hansen, 1981) or total tissue calcium (Warner et ai., 1987) are not equivalent to the measurement of changes in intracellular calcium used in the present study due to shifts of calcium between the extra- and intracellular spaces, as well as the sequestration of calcium by intracellular organelles (Siesjo and Bengtsson, 1989). Following 2 h of MCA occlusion in the cat, Venables et al. (1985) found that pial surface potassium activity remained elevated in the penumbra zone during reperfusion in hyperglycemic cats but returned to normal in normoglycemic animals. Although changes in potassium may not directly relate to changes in calcium, the time course of these potassium changes are very similar to the changes in [Ca2+]j observed in our study. Both phenomena may be due to damage to cellular membranes. Recently, Lundgren et al. (199) examined the postischemic alterations of the pyruvate dehydrogenase complex (PDHC) in normoglycemic and hyperglycemic rats subjected to 15 min of total forebrain ischemia, and found a relative increase in the fraction of activated PDHC in the early recovery phase in hyperglycemic animals compared to normoglycemic animals. They speculate that this is due to either increased intramitochondrial calcium levels or to persistent increases in the NADH/NAD and/or the ADP/ATP ratios. Our observation of a small change in NADH/NAD redox state in hyperglyce- mic animals suggests that the increase in PDHC in the early reperfusion phase may reflect an increase in intramitochondrial calcium levels. Mitochondria will sequester Ca2+ when the cytosolic free calcium is abnormally increased (Dux et ai., 1987) so that a Ca2+ increase in the mitochondria reflects an increase in cytosolic free calcium. During the early reperfusion period following temporary focal ischemia, the normoglycemic animals show a rapid recovery of calcium homeostasis, whereas the hyperglycemic animals exhibit a prolonged elevation in [Ca2+]i' Ginsberg et al. (198) showed that flow and metabolism during ischemia in normoglycemia and hyperglycemia do not differ greatly, but that there was a striking failure of hemodynamic and metabolic recovery in the recirculation period in the hyperglycemic animals. Kozuka et al. (1989) observed that preischemic hyperglycemia markedly accentuated the postischemic depression of local cerebral glucose metabolism, and Welsh et al. (1983) found that both CBF and ATP were significantly lower in animals pretreated with glucose following 6 min of reperfusion. During global ischemia in the cat, Chopp et al. (1987) found that intracellular ph obtained by magnetic resonance spectroscopy fell to the same level in normoglycemia and hyperglycemia during ischemia, whereas there was a decline in intracellular ph in the hyperglycemic group during the initial reperfusion that was greater than that which occurred in the normoglycemic group. These studies indicate the importance of the early phase of reperfusion in the metabolic derangement during hyperglycemic conditions. During both ischemia and reperfusion, hyperglycemic animals exhibit a more severe lactic acidosis than do normoglycemic animals (Welsh et ai., 198; Rehncrona et ai., 198]; Smith et ai., 1986). In other studies, hyperglycemic animals showed a more pronounced drop in intracellular ph only during the initial reperfusion period (Smith et ai., 1986; Chopp et ai., ]987, 1988), or after 3 h of reperfusion (Marsh et ai., 1986). This intracellular acidosis is thought to be closely related to the changes in cytosolic free calcium. Abercrombie and Hart (1986) observed an increase in cytosolic Ca2 + when the ph was decreased, indicating that acidosis can induce a calcium release from intracellular binding sites leading to a reduction in the capacity of the cell to buffer a calcium load (Siesjo, 1988). The persistent elevation in [Ca2+]i observed in the hyperglycemic group during the early reperfusion period in our study may be closely related to this decrease in intracellular ph. Even though we observed a difference in the [Ca2+]i between the hyperglycemic and the normo- J Cereb Blood Flow Metab, Vol. 12, No. 3, 1992

HYPERGLYCEMIA AND INTRACELLULAR CALCIUM 475 glycemic group during the early reperfusion phase, no histopathological differences were found after 3 h of reperfusion. Previous investigations of the effect of hyperglycemia on histological damage in cerebral ischemia have been equivocal. Nedergaard (1987) observed a significantly larger infarct volume in hyperglycemic rats than in normoglycemic animals following 1 or 15 min of MeA occlusion. Tissue with extensive collaterals, such as the neocortex in the rat, appear to be more vulnerable to the deleterious effects of hyperglycemia during MeA occlusion than are regions with a nonanastomosing vascular supply, such as the striatum (Prado et al., 1988). In a photochemically induced thrombotic ischemia model in the rat, Ginsberg et al. (1987) found that larger infarct volumes were correlated with lower plasma glucose levels. Recently, in a permanent model of MeA occlusion in the cat, Zasslow et al. (1989) also observed an inverse relationship between the blood glucose concentration and the area of cortical damage. Although they did not monitor osmolality, they did discuss the possibility that the beneficial effect of hyperglycemia may be due not to glucose per se but to the hyperosmotic effect of glucose administration. Although the osmolality of the control group was not measured, we have previously measured blood osmolality in normoglycemic animals during MeA occlusion and reperfusion and found it to be stable. Since in this study glucose was infused continuously until the end of the study, the osmolality in the hyperglycemic animals increased gradually, becoming significantly higher than the preischemic level and presumably also significantly higher than the normoglycemic animals. Our observation of a difference in calcium recovery during the early reperfusion period, but no difference in histopathological damage, may possibly be due to the effect of hyperosmolality induced by continuous glucose infusion that may protect the brain from edema. In conclusion, our data indicate that hyperglycemia may be harmful to calcium recovery during the early recirculation period following focal cerebral ischemia. The hyperosmolality induced by infusion of glucose may have attenuated cerebral edema and prevented a worsening of the neuronal damage sometimes seen in hyperglycemic animals. Acknowledgment: This work was supported by a grant from the National Institutes of Health, Grant Number NSI939-1S. We wish to thank Ms. Doreen Shikitino and Ms. Emily Daniel for their help in preparing the manuscript. REFERENCES Abercrombie RF, Hart CE (1986) Calcium and proton buffering and diffusion in isolated cytoplasm from myxicola axons. Am J Physiol 25:C391-C45 Chopp M, Frinak S, Walton DR, Smith MB, Welch KMA (1987) Intracellular acidosis during and after cerebral ischemia: In vivo nuclear magnetic resonance study of hyperglycemia in cats. Stroke 18:919-923 Chopp M, Welch KMA, Tidwell CD, Helpem JA (1988) Global cerebral ischemia and intracellular ph during hyperglycemia and hypoglycemia in cats. Stroke 19:1383-1387 Dora E (1984) A simple cranial window technique for optical monitoring of cerebrocortical microcirculation and NADI NADH redox state. Effect of mitochondrial electron transport inhibitors and anoxic anoxia. J Neurochem 42: 11-18 Dux E, Mies G, Hossman KA, Siklos L (1987) Calcium in the mitochondria following brief ischemia of gerbil brain. Neurosci Lett 78:295-3 Ginsberg MD, Reivich M, Frinak S, Harbig K (1976) Pyridine nucleotide redox state and blood flow of the cerebral cortex following middle cerebral artery occlusion in the cat. Stroke 7:125-131 Ginsberg MD, Welsh FA, Budd WW (198) Deleterious effect of glucose pretreatment on recovery from diffuse cerebral ischemia in the cat. I. Local cerebral blood flow and glucose utilization. Stroke 11 :347-354 Ginsberg MD, Prado R, Dietrich WD, Busto R, Watson BD (1987) Hyperglycemia reduces the extent of cerebral infarction in rats. Stroke 18:57-574 Harbig K, Chance B, Kovach AGB, Reivich M (1976) In vivo measurement of pyridine nucleotide fluorescence from cat brain cortex. J Appl PhysioI41:48-488 Ibayashi S, Fujishima M, Sadoshima S, Yoshida F, Shiokawa, Ogata J, Omae T (1986) Cerebral blood flow and tissue metabolism in experimental cerebral ischemia of spontaneously hypertensive rats with hyper-, normo-, and hypoglycemia. Stroke 17:261-266 Kalimo H, Rehncrona S, S6derfeldt B, Olsson Y, Siesj6 BK (1981) Brain lactic acidosis and ischemic cell damage: 2. Histopathology. J Cereb Blood Flow Metab 1:313-327 Kozuka M, Smith MJ, Siesj6 BK (1989) Preischemic hyperglycemia enhances postischemic depression of cerebral metabolic rate. J Cereb Blood Flow Metab 9:478-49 Lundgren J, Cardell M, Wieloch T, Siesj6 BK (199) Preischemic hyperglycemia and postischemic alteration of rat brain pyruvate dehydrogenase activity. J Cereb Blood Flow Metab 1:536-541 Marsh WR, Anderson RE, Sundt TM (1986) Effect of hyperglycemia on brain ph levels in areas of focal incomplete cerebral ischemia in monkeys. J Neurosurg 65:693-696 Meier P, Zierler KL (1954) On the theory of the indicatordilution method for measurement of blood flow and blood volume. J Appl PhysioI6:731-744 Myers RE, Yamaguchi S (1977) Nervous system effects of cardiac arrest in monkeys. Arch Neural 34:65-74 Nedergaard M (1987) Transient focal ischemia in hyperglycemic rats is associated with increased cerebral infarction. Brain Res 48:79-85 Nedergaard M, Astrup J (1986) Infarct rim: Effect of hyperglycemia on direct current potential and e4cj2-deoxyglucose phosphorylation. J Cereb Blood Flow Metab 6:67-615 Paljarvi L, Rehncrona S, S6derfeldt B, Olsson Y, Kalimo H (1983) Brain lactic acidosis and ischemic cell damage: Quantitative ultrastructural changes in capillaries of rat cerebral cortex. Acta Neuropathol (Ber/) 6:232-24 Prado R, Ginsberg MD, Dietrich WD, Watson BD, Busto R (1988) Hyperglycemia increases infarct size in collaterally perfused but not end-arterial vascular territories. J Cereb Blood Flow Metab 8: 186-192 Pulsinelli WA, Waldman S, Rawlinson D, Plum F (1982) Moderate hyperglycemia augments ischemic brain damage: A neuropathologic study in the rat. Neurology 32: 1239-1246 Rehncrona S, Rosen I, Siesj6 BK (1981) Brain lactic acidosis and ischemic cell damage: 1. Biochemistry and neurophysiology. J Cereb Blood Flow Metab 1:297-311 Siemkowicz E, Hansen AJ (1981) Brain extracellular ion com- J Cereb Blood Flow Me/ab. Vol. 12, No. 3, 1992

476 N. ARAKI ET AL. position and EEG activity following 1 minutes ischemia in normo- and hyperglycemic rats. Stroke 12:236-24 Siesjii BK (1981) Cell damage in the brain: A speculative synthesis. J Cereb Blood Flow Metab 1:155-185 Siesjii BK.(1988) Calcium, ischemia, and death of brain cells. Ann NY Acad Sci 522:638-661 Siesjii BK, Bengtsson F (1989) Calcium fluxes, calcium antagonists, and calcium-related pathology in brain ischemia, hypoglycemia, and spreading depression: A unifying hypothesis. J Cereb Blood Flow Metab 9: 127-14 Smith ML, von Hanwehr R, Siesjii BK (1986) Changes in extraand intracellular ph in the brain during and following ischemia in hyperglycemic and in moderately hypoglycemic rats. J Cereb Blood Flow Metab 6:574-583 Tanaka K, Dora E, Urbanics R, Greenberg JH, Toffano G, Reivich M (1986) Effect of the ganglioside GMI on cerebral metabolism, microcirculation, recovery kinetics of ECoG and histology during the recovery period following focal ischemia in cats. Stroke 17: 117-1178 Uematsu D, Greenberg JH, Reivich M, Karp A (1988) In vivo measurement of cytosolic free calcium during cerebral ischemia and reperfusion. Ann Neurol 24:42-428 Venables GS, Miller SA, Hardy JA, Strong AJ (1985) The effects of hyperglycemia on changes during reperfusion following focal cerebral ischaemia in the cat. J Neurol Neurosurg Psychiatry 48:663-669 Warner OS, Smith M, Siesjii BK (1987) Ischemia in normo- and hyperglycemic rats: Effects on brain water and electrolytes. Stroke 18:464-471 Welsh FA, Ginsberg MD, Rieder W, Budd WW (198) Deleterious effect of glucose pretreatment on recovery from diffuse cerebral ischemia in the cat. II. Regional metabolite levels. Stroke 11 :355-363 Welsh FA, Sims RE, McKee AE (1983) Effect of glucose on recovery of energy metabolism following hypoxia-oligemia in mouse brain: Dose-dependence and carbohydrate specificity. J Cereb Blood Flow Metab 3:486-492 Zasslow MA, Pearl RG, Shuer LM, Steinberg GK, Lieberson RE, Larson CP (1989) Hyperglycemia decreases acute neuronal ischemic changes after middle cerebral artery occlusion in cats. Stroke 2:519-523 J Cereb Blood Flow Metab, Vol. 12, No.3, 1992