When perfused neonatal brain slices are studied ex vivo with nuclear magnetic resonance (NMR) spectroscopy, it is possible to use 31P detection to monitor levels of intracellular adenosine triphosphate (ATP), cytosolic pH, and other high-energy phosphates and 1H detection to monitor lactate and glutamate. Adult brain slices of high metabolic integrity are more difficult to obtain for such studies, because the adult cranium is thicker, and postdecapitation revival time is shorter. A common clinical anesthesia phenomenon--loss of temperature regulation during anesthesia, with surface cooling and deep hypothermia, was used to obtain high-quality adult rat cerebrocortical slices for NMR studies.


Spontaneously breathing adult rats (350 g), anesthetized with isoflurane in a chamber, were packed in ice and cooled until rectal temperatures decreased to approximately 30 degrees C. An intraaortic injection of heparinized saline at 4 degrees C further cooled the brain to approximately 18 degrees C. Slices were obtained and then recovered at 37 degrees C in oxygenated medium. Interleaved 31P/1H NMR spectra were acquired continually before, during, and after 20 min of no-flow hypoxia (PO2 approximately 0 mmHg). Histologic (Nissl stain) measurements were made from random slices removed at different times in the protocol. Three types of pretreatment were compared in no-flow hypoxia studies. The treatments were: (1) hyperoxia; (2) hypercapnia (50% CO2); and (3) hypoxia, which was accomplished by washing the slices with perfusate equilibrated with 100% N2 and maintaining a 100% N2 gas flow in the air space above the perfusate.


During hyperoxia, 31P NMR metabolite ratios were identical to those seen in vivo in adult brains, except that, in vitro, the Pi peak was slightly larger than in vivo. A lactate peak was seen in in vitro 1H spectra of slices after metabolic recovery from decapitation, although lactate is barely detectable in vivo in healthy brains. The in vitro lactate peak was attributed to a small population of metabolically impaired cells in an injury layer at the cut edge. NMR spectral resolution from the solenoidal coil exceeded that obtained in vivo in surface coil experiments. Phosphocreatine and ATP became undetectable during oxygen deprivation, which also caused a three- to sixfold increase in the ratio of lactate to N-acetyl-aspartate. Within experimental error, all metabolite concentrations except pHi recovered to control values within 2 h after oxygen restoration. Nissl-stained sections suggested that pretreatment with hypercapnia protected neurons from cell swelling during the brief period of no-flow oxygen deprivation.


Perfused, respiring adult brain slices having intact metabolic function can be obtained for NMR spectroscopy studies. Such studies have higher spectral resolution than can be obtained in vivo. During such NMR experiments, one can deliver drugs or molecular probes to brain cells and obtain brain tissue specimens for histologic and immunochemical measures of injury. Important ex vivo NMR spectroscopy studies that are difficult or impossible to perform in vivo are feasible in this model.

BRAIN edema and intracellular energy failure occur within seconds after the onset of cerebral oxygen deprivation. They persist throughout the course of cerebral oxygen deprivation and are targets of therapy-oriented research. Early energy failure in respiring animal brain slices has been studied with repeated, non-invasive NMR spectroscopy during oxygen deprivation and glutamate toxicity. [1–3]The former provides a model of stroke and hypoxia, whereas the latter approximates penumbral exposure to excitotoxic amino acids. There are inherent limits to NMR studies of brain slices. For example, brain slices are deafferentiated, and they can neither survive indefinitely nor recover neurologic function. In addition, it is possible that the decapitation ischemia associated with tissue isolation might result in tissue alkalosis during reoxygenation. However, metabolic studies with perfused slices avoid confounding variables found in vivo: localization of NMR spectra to a particular region of the brain, cerebral blood flow changes, uncontrollable extracellular concentrations of metabolites or agents, blood-brain barrier limits to drug delivery, hepatic metabolism of drugs, dosage limits to anesthesia, or requirements that anesthesia be used. Additionally, certain sophisticated NMR studies can be performed easily in slices but not so easily in vivo. These include various types of gradient-enhanced, diffusion-weighted spectroscopy and multinuclear, multidimensional NMR experiments. Finally, NMR slice experiments are convenient for studying gene expression. In such studies, NMR spectroscopy provides a means of intracellular physiologic monitoring, and slices are immediately accessible for removal and rapid freezing or fixation. In situ hybridization, immunohistochemistry, and drug alteration of gene expression are convenient in this system.

Compared to neonatal animals, it is considerably more difficult to obtain high-integrity brain slices from older animals, because the brain has a shorter revival time and a longer isolation time. Although hypoxia is an important issue in pediatric and neonatal medicine, perioperative hypoxia/ischemia and stroke are more common in adult humans. For this reason, we investigate the possibility of studying adult brain slices from older animals. We describe an adult rat brain-slice preparation developed for NMR studies. Hypothermic cerebral metabolic protection, [4,5]a well studied phenomenon often used for obtaining slices for electrophysiology [6]studies, was combined with in vivo surface cooling during isoflurane anesthesia during spontaneous ventilation.

Slice Preparation

Animal protocols were approved by the University of California, San Francisco, Committee on Animal Research. For each NMR experiment, 40 cerebrocortical slices, each 350 micro meter thick, were obtained from ten young adult male Wistar rats (250+/-50 g, Simonsen, CA) and studied in a single 20-mm diameter NMR tube. Approximately 25 min was required to anesthetize and cool a single adult rat and then isolate four brain slices. The procedure was as follows. The animal was first placed in a transparent Plexiglas chamber, where anesthesia was induced by spontaneous ventilation with [nearly equal] 2%-4% isoflurane in oxygen.

Anesthetized animals were supine with a head-down tilt of [nearly equal] 15 degrees. When the respiratory rate was observed to decrease to [nearly equal] 1 per second, the chamber was briefly opened and a bed of crushed, packaged ice was placed around the animal, i.e., underneath it and over its legs and lower abdomen, but not over its upper chest. The chamber was closed, anesthesia flow in 100% Oxygen2was restored, and spontaneous respiratory excursions were visually monitored through the Plexiglas. The inspired isoflurane concentration was gradually decreased to [nearly equal] 0.5–1.5% over [nearly equal] 10 min as the animal cooled and as respiratory frequency and amplitude decreased. The rationale for the decrease in anesthetic concentration was that isoflurane solubility and potency increase as body temperature is decreased.* The monitored physiologic endpoint for adjustment of anesthesia concentration was decreased respiratory rate with adequate chest excursions. These were visualized as rectal temperature decreased to [nearly equal] 30 degrees Celsius. In initial efforts during development of the protocol, failure to decrease the inspired anesthetic concentration during cooling resulted in brain slices having substantial ischemic injury and decreased levels of high-energy phosphate metabolites observed by NMR spectroscopy. After achieving a rectal temperature of [nearly equal] 30 degrees C (which took [nearly equal] 10 min), the chamber was opened, and the hypothermic, anesthetized rat was transferred to a surgical table. The chest was opened via a midline thoracotomy with a sharp scissors and the heart exposed. After making left and right ventriculostomy incisions, a syringe containing cold heparinized saline (60 ml, 0.9%, 4+/-2 degrees Celsius) was removed from an ice bucket. The syringe was tipped with a gavage needle, which was placed through the left ventriculostomy, across the left ventricle, into the proximal ascending aorta. Cold saline was injected during a period that lasted [nearly equal] 10 s, with the gavage needle tip being palpated between the index and third fingers. Immediately on completion of the injection, the aorta was pinched closed, and the heart was severed from the body. This was done to prevent subsequent cardiac pumping of warm blood to the head after cold perfusion.

The brain was isolated after immediate subsequent decapitation with a guillotine (Harvard Instruments). Blunt-tipped sharp scissors were used to remove the scalp and divide the calvarium in the dorsal midline, caudal to rostral, starting at the site of decapitation. The right and left calvaria were lifted off the brain, exposing the in situ brain. The olfactory bulb and optic nerves were severed from the brain with one cut of the scissors. A cold spatula, wet with saline, was used to gently pry the brain from the skull.

Rectal temperatures at the time of the intraaortic saline injection averaged 30+/-2 degrees C, providing a good estimate of brain temperature during the 2-min interval between animal removal from the chamber and intraaortic injection. After decapitation and brain isolation, the average cerebrocortical temperature was 17+/- 2 degrees C, as measured with a Grass temperature probe (Omega Engineering, Stamford, CT). This value was taken as a good approximation of brain temperature during the period after intraaortic injection, until slice placement in oxygenated media. Table 1gives average times for various steps in the procedure. From our data in Table 1and an estimate of 2.0 for Q10(the difference in cerebral metabolic rates for two temperatures differing by 10 degrees C), [7]we conclude that 5 min of impaired brain oxygenation during our hypothermia protocol corresponds to [nearly equal] 1.8 min [=(2 min)/(0.9 x 2)+(3 min)/(2.2 x 2)] of impaired brain oxygenation at normothermia (39 degrees C). In addition to hypothermic protection, brain slices in our protocol might have had anesthetic protection. Isoflurane is known to cause electroencephalographic suppression and result in a lower critical value for the ischemic threshold of cerebral blood flow. [8,9]Isoflurane, a vapor anesthetic, was the only agent administered before cutting slices, and it was allowed to wash out after slice isolation. However, we call attention to the fact that, with our technique, it is possible to administer other pharmacologic pretreatments during brain perfusion by the intraaortic injection.

Four slices, each 350 micro meter thick, were cut from every rat brain, using a modified McIlwain technique that was described previously. [10–12]In this technique, a glass slide with shim strips at the edges rests on the brain surface, and a microtome blade slides along the shims, producing slices of uniform thickness. Slices came from frontal and parietal cortex. Each slice, after being cut from the brain, was first placed in fresh cold (4 degrees C) artificial cerebrospinal fluid (ACSF). After completing the isolation of four slices from one animal, they were individually transferred from the cold solution to ACSF at 21 degrees C in a standard 20-mm diameter glass NMR tube (Wilmad, Buena, NJ) that served as the tissue perfusion chamber. Within the NMR tube, the tissue slices rest together on a Teflon mesh located 1 cm above the bottom. The mass of slices spans a height within the tube of approximately 1 cm, and the radiofrequency solenoidal NMR coil that fits snugly about the tube has the slices centered in the solenoid. A schematic diagram of the arrangement is shown in Figure 1. Teflon tubing (1.7 mm OD) was used for inflow and outflow. The inlet tubing descended beyond the Teflon mesh toward the bottom of the tube, so that perfusate flow to the slices came upward. The outlet tubing was situated in the perfusate at a level [nearly equal] 1 cm above the slices. Inflow and outflow of perfusate were driven separately by two pumps (Cole Parmer Masterflex Console Drive), with the rates adjusted to produce a stable fluid level in the NMR tube. Stability occurred when the outlet pump was set to approximately twice the speed of the inflow pump. Further details are given in the schematic diagrams in Figure 1.

Our ACSF, a modified Krebs balanced salt solution, contained 124 mM NaCl, 5 mmol/l KCl, 1.2 mmol/l KH2PO41.2 mmol/l MgSO sub 4, 1.2 mmol/l CaCl2, 26 mmol/l NaHCO3, and 10 mmol/l glucose. The bicarbonate pH buffer for fresh ACSF was maintained by continuous bubbling with a gas mixture of 5% CO2/95% Oxygen2. Extracellular pH of fresh ACSF was constantly at [nearly equal] 7.4. The nonrecirculating flow of fresh ACSF through the NMR tube was 15–20 ml/min. The ACSF temperature was maintained at 21+/-1 degree C for 1 h after the isolation of the last slice, and then increased gradually over 60 min to 37+/-1 degree C.

Ex Vivo Oxygen Deprivation

NMR studies of slice ensembles began after31Phosphorus and1Hydrogen NMR spectra indicated there was complete recovery from decapitation oxygen deprivation. After obtaining 30 min of control data, oxygen deprivation was induced by halting the flow of perfusate for 20 min. Three sets of oxygen deprivation experiments were performed for our adult slices. In the first group, slices were kept normocapnic and hyperoxic before oxygen deprivation (n = 3). In the second group, slices were made hypercapnic for 20 min before oxygen deprivation (n = 3). Hypercapnia was accomplished as in earlier studies [10]by switching to ACSF that had been equilibrated with a gas mixture that was 50% CO2/50% O2. In a third set of experiments, normocapnic slices first underwent 5 min of hypoxia and then 20 min of oxygen deprivation (n = 2). In this third group ("hypoxia pretreatment"), the air space in the NMR tube was flushed with 100% N2, and the perfusate was switched to "degassed" ACSF, i.e., ACSF that had been equilibrated with a gas mixture that was 5% CO2/95% N2. A 120-min reperfusion recovery period with "normal" ACSF (oxygenated with 95% O2/5% CO2) followed each of the oxygen deprivation protocols. Aliquots of perfusate were taken from the slice chamber for PO2determinations, which were measured with a blood gas analyzer (Radiometer ABL 30, Copenhagen, Denmark). Table 2shows resulting levels of PO2, PCO2, and pH in the perfusate for different stages in the protocol.


Random slices, i.e., not simply slices at the top of the tube, were taken from the perfusion chamber at different predetermined time points: before oxygen deprivation, at the end of oxygen deprivation, and at the end of reperfusion recovery. These slices were transferred to tubes containing 10% formalin in phosphate-buffered 0.9% saline solution (pH [nearly equal] 7). After 24–72 h, the tissues were washed, dehydrated, and embedded in paraffin. Adjacent sections, 10 micro meter thick, were cut in the plane of the slices, stained with cresyl violet (Nissl stain), and examined using phase contrast light microscopy (with a Nikon Diaphot TMD inverted microscope). [12]Representative fields of 50–140 adjacent cells, corresponding to areas of [nearly equal] 104micro meter2, were photographed after counting shrunken and dead neurons.

NMR Experiments

Interleaved31Phosphorous/sup 1 Hydrogen spectra were obtained on a Nalorac Quest 4400 4.7 Tesla NMR instrument, operating at 81 and 200 MHz, respectively. NMR methods were derived from those established for studies of neonatal slices. [10–12]Magnetic field homogeneity was optimized by adjusting room-temperature shim currents until the water proton linewidth from the perfusate was less than 0.06 ppm. The 20-mm NMR tube, containing 40 tissue slices, was positioned inside a custom-made, double-tuned, four-turn, 23 x 15-mm solenoidal coil. Total acquisition time for interleaved31Phosphorous/sup 1 Hydrogen spectra [13,14]was 10 min. Each acquisition consisted of 2,048 complex data points for both31Phosphorous and1Hydrogen. Time-sharing was such that it took 0.84 s for a single31Phosphorous acquisition and 1.27 s for a single1Hydrogen acquisition. Individual spectra for31Phosphorous and1Hydrogen were generated after 120 acquisitions in quadrature phase detection mode. The spectral width was+/-4,000 Hz. Typically, for31Phosphorous, a single-pulse experiment was used, and the duration of the radiofrequency excitation pulse was [nearly equal] 30 micro s for a 45 degrees -tip angle, with a pulse interval of 2.11 s. The broad hump from phospholipids in the31Phosphorous spectra was removed by a convolution difference method using an exponential filter of 500 Hz.

Spin-echo1Hydrogen NMR spectroscopy experiments were initiated with a 100-ms low-power presaturation pulse centered on the water resonance. The spin-echo delay for refocusing pulses was 136 ms. Having a long spin-echo permitted substantial discrimination against more rapidly relaxing lipid signals near the lactate peak at 1.32 ppm. Duration of the1Hydrogen radiofrequency excitation pulse was [nearly equal] 105 micro s for a 90 degrees nutation. Each1Hydrogen spectrum was processed to achieve Lorenzian-to-Gaussian transformation by - 12 Hz exponential and 7 Hz Gaussian multiplication.

NMR signal intensities for each metabolite were determined by numeric integration of optimal computer fits to corresponding NMR resonance peaks in the spectra (Nalorac Quest 4400 Curve Fitting Program, Martinez, CA, and MacFID Curve Fitting Program, Tecmag, Bellaire, TX).31Phosphorous metabolite concentrations were measured relative to corresponding signal intensities in the control run. Relative adenosine triphosphate (ATP) levels were determined from the beta-ATP peak at 16.3 ppm. Fully relaxed spectra (20-s interpulse delay) were obtained in special studies to obtain relaxation time corrections for different metabolites. As in previous studies, [10,11]pH computations were calculated from the chemical shift difference between Phosphocreatine (PCr) and Pi. In spectra where extra- and intracellular Piresonances were resolvable, verification of separate Pipeak assignments was based on blood gas pH measurements of extracellular fluid taken from the NMR tube housing the slices. Relative changes in intracellular lactate levels were determined from1Hydrogen spectra, using the N-acetyl-aspartate (NAA) peak as an internal reference. [15].

Statistical Analysis

Data are reported as mean+/-SD. Statistical comparisons were made of average relative NMR metabolite values for different times in the protocol. First, a repeated-measures analysis of variance was used to determine whether relative metabolite values appeared unchanged throughout time, i.e., consistent with the null hypothesis. [16]If the null hypothesis was rejected, multiple comparisons were made using Bonferroni corrected t-tests [17]among the following a priori chosen intervals (i.e., intervals chosen before taking data):(1) control = 0 min, (2) at 20 min (i.e., at the completion of oxygen deprivation), (3) at 60 min (middle of reperfusion recovery period), and (4) at 120 min (end of reperfusion recovery period). Details of our statistical methods have been described previously. [10].

(Figure 2) shows a comparison between a typical in vivo31Phosphorous rat brain spectrum (from an adult male Wistar rat) and a control spectrum of our ex vivo brain-slice preparation. The high quality of the ex vivo spectrum is apparent. The ex vivo high-energy phosphorus metabolite profile closely matches that of the in vivo spectrum, except that it contains two overlapping larger Pipeaks representing intracellular Piand extracellular Pifrom the perfusate. Each spectrum was obtained after [nearly equal] 10 min of data accumulation. The in vivo spectrum was obtained using a 1-cm, two-turn surface coil. The preischemic PCr/ATP ratio measured for the brain slices was 1.54 +/-0.02 (n = 6), whereas that obtained from in vivo studies is [nearly equal] 1.5–1.7. [18].

Typical31Phosphorous spectra from two representative NMR hypoxia experiments are shown in Figure 3, demonstrating typical spectral changes that are subsequently quantitated. Figure 3(B)(hypercapnic pretreatment) shows better recovery of Picompared to Figure 3(A)(hypoxic pretreatment).

Average metabolite data from the three pretreatment studies of no-flow oxygen deprivation are shown in Figure 4. Plots are presented for relative changes in PCr, ATP, pHi, and lactate/NAA. Relative lactate/NAA ratios increased approximately threefold, fourfold, and sixfold, respectively, at the end of the insult period for hyperoxic pretreatment, hypercapnic pretreatment, and hypoxic pretreatment. For all types of pretreatment, PCr and ATP decreased after 30 min of hypoxia to levels that were not detectable by NMR spectroscopy. Recovery of PCr was maximal and close to control (0.98+/-0.05, P < 0.05) for slices pretreated with hypercapnia. In contrast, PCr concentrations recovered only to [nearly equal] 0.80 in slices pretreated either with hypoxia or hyperoxia. Statistically significant differences between the hypercapnia group and the other two groups (oxygen deprivation and hypoxia-oxygen deprivation) were observed for PCr concentrations after 1 and 2 h of postischemic reperfusion (P < 0.05). A slower rate of recovery of ATP concentrations was observed for slices that had the hypoxia pretreatment. The ATP recovery rate was the same for the groups with hyperoxia pretreatment and hypercapnia pretreatment. For all groups, no statistically significant differences were found between ATP concentrations at the end of 1 h of reperfusion recovery. Figure 4(C) shows that intracellular pH at 1 and 2 h after hypoxia was significantly higher in the hypercapnia pretreatment group. For 1 and 2 h after hypoxia, the values were:(1) 7.13+/-0.05 and 7.12+/-0.08 for hypercapnia pretreatment, (2) 6.69+/-0.03 and 6.74+/- 0.06 for hypoxia pretreatment, and (3) 6.67+/-0.03 and 6.78 +/-0.06 for hyperoxia pretreatment.

(Figure 5) shows morphologic changes in neurons studied in phase-contrast micrographs of 10-micro meter Nissl-stained sections. Slices taken immediately after hypothermic brain isolation (Figure 5(A)) and 120 min after decapitation recovery (Figure 5(B)) showed regular, round neuronal nuclei. Slices exposed to either hypoxia pretreatment or hyperoxia pretreatment showed, after 60 min of reperfusion recovery, shrunken and deformed neuronal nuclei inside swollen neurons (Figure 5(D and E)). In contrast, similarly taken slices from the group made hypercapnic before oxygen deprivation showed remarkably preserved morphology (Figure 5(C)). Qualitatively, the group that showed better values of NMR metabolites was also the group that showed better histology. Although the histologic changes in Figure 5(D and E) indicate permanent injury if they persist for more than [nearly equal] 4 h after the onset of oxygen deprivation, these changes do not indicate permanent injury if seen only 2 h after an insult. Such histologic changes are reversible in studies in which insults are mild and appropriate resuscitative measures are initiated rapidly. The primary point is that, in our perfused brain-slice model of hypoxia, NMR metabolic impairment and histology seem to correlate with each other.

NMR metabolic changes in perfused adult brain slices appear to closely resemble those found in vivo. A major difference between ex vivo and in vivo NMR spectra (Figure 2(A and B)) is the ex vivo presence of a large extracellular Pipeak. As shown in Figure 2, substantial augmentation to Picomes from inorganic phosphate in the perfusate. However, in previous studies with neonatal brains, we found that use of 1.2 mM phosphate in the perfusate results in metabolically and histologically healthy brain slices that are viable for > 10 h. Of course, in vivo extracellular Piis very small in healthy brain. Although spectral overlap of extracellular and intracellular Piis an inconvenience in this model, there is nevertheless an advantage to having extracellular Pi. Its presence can provide an extracellular spectroscopic chemical shift reference location, a pH calibration value, and a reference for quantitation. However, reaping all of these benefits requires more sophisticated NMR methods than we have used. [19]With such methods, IOPxit is possible to cleanly separate intra- and extracellular contributions.

It should be noted that, in slice studies, perfusate concentrations might not be identical to interstitial concentrations, just as cerebrospinal fluid concentrations in vivo are not identical to brain interstitial concentrations that are measured with microdialysis. In principle, NMR spectroscopy in our slice model can be used to separately determine interstitial, perfusate, and intracellular concentrations by changing perfusate flow and using more sophisticated NMR techniques, such as diffusion-weighted spectroscopy. [19].

A recent review of previous studies found that low pH during hypoxia is associated with adverse changes in metabolism, histology, and neurologic outcome. [20]The protective effect of hypercapnia pretreatment in our studies does not contradict this finding. Our ex vivo oxygen deprivation period might have been too short to fully exhaust tissue oxygen and astrocyte glycogen, [21–23]and further studies would be required to fully examine such questions. In earlier studies, we found that short periods of mild hypercapnia are tolerable when ATP is maintained. [10]Additionally, some studies have found that low extracellular pH (pHo[nearly equal] 6.5) induces closure of ligand-gated N-methyl-D-aspartate channels, [20,24]leading to protection during hypoxia. Further decreases in pHoto 6.1 have been found to reduce kainate-induced currents. Protective effects of induced brain acidosis via hypercarbic ventilation in focal cerebral ischemia, with maximal protection at pH values of [nearly equal] 6.8 and most protection lost at pH values below 6.5. [25]Acidotic protection during hypoxia was similarly described by Kristian et al. [26]In our stopped-flow hypoxia studies that had hypercapnic pretreatment, the decrease in ATP concentration might have been due to downregulation of ATP synthesis, which has been reported to occur as pHidecreases to 6.3. [27]On the other hand, slices might have been slightly hypoxic during hypercapnic pretreatment, as lactate is seen to increase (Figure 4(D)) before the period of nitrogen hypoxia. It is possible that, when oxygen concentrations in the perfusate descend to near 50%, there is insufficient oxygen for aerobic metabolism in all slices. A similar effect was seen in our earlier neonatal brain-slice study of hypercapnia. [10]It should be remembered that ATP becomes undetectable with NMR spectroscopy at concentrations below [nearly equal] 0.4 mM/l. Therefore, small but protective concentrations of ATP might persist at NMR-invisible levels. Finally, although Nissl stains for the hyperoxic and hypoxic pretreatment groups revealed that all neurons were seen to be swollen 2 h after the insult, such does not indicate that irreversible injury has occurred. Several hours is needed before such changes in conventional stains can be deemed irreversible.

We believe that the high quality of brain tissue under study in this model will accommodate additional, interesting histologic techniques, such as in situ hybridization and immunohistology. These, in turn, will make it possible to distinguish cell-specific molecular events in the context of overall, nonspecific knowledge regarding the energy state. With NMR spectroscopy, one cannot distinguish between the following two situations:(1) when all cells are at 50% energy and (2) when 50% are perfectly healthy and remaining cells are at zero energy. Double- and triple-labeled immunostain studies of protective, structural, and injury proteins, such as hsp70 expression or neurofilament integrity, would help resolve cell-specific ambiguities.

In summary, we conclude that the tolerance of our adult brain-slice preparation to no-flow hypoxia paradigms demonstrates intact metabolic functions. Additionally, it is feasible in this model to conduct studies that combine NMR spectroscopy with invasive histologic techniques. Such investigations have high potential for correlating metabolic and histologic drug effects in adult brain tissue during acute hypoxia and ischemia.

*Eger EI: Isoflurane: A compendium and reference. Madison, Anaquest, 1985.

Badar-Goffer RS, Thatcher NM, Morris PG, Bachelard HS: Neither moderate hypoxia nor mild hypoglycemia alone causes any significant increase in cerebral [Calcium sup 2+] sub i: only a combination of the two insults has this effect. A sup 31 Phosphorus and sup 19 Fluorine NMR study. J Neurochem 1993; 61:2207-14.
Jacquin T, Gillet B, Fortin G, Pasquier C, Beloeil JC, Champagnat J: Metabolic action of N-methyl-D-aspartate in newborn rat brain ex vivo: sup 31 Phosphorus magnetic resonance spectroscopy. Brain Res 1989; 497:296-304
Kauppinen RA, Williams SR: Nondestructive detection of glutamate by sup 1 Hydrogen nuclear magnetic resonance spectroscopy in cortical brain slices from the guinea pig: evidence for changes in detectability during severe anoxic insults. J Neurochem 1991; 57:1136-44.
Chen H, Chopp M, Jiang Q, Garcia JH: Neuronal damage, glial response and cerebral metabolism after hypothermic forebrain oxygen/glucose deprivation in the rat. Acta Neuropathol (Berl) 1992; 84:184-9.
Zhang RL, Chopp M, Chen H, Garcia JH, Zhang ZG: Postischemic (1 hour) hypothermia significantly reduces ischemic cell damage in rats subjected to 2 hours of middle cerebral artery occlusion. Stroke 1993; 24:1235-40.
Pakhotin PI, Belousov AB, Otmakhov NA: Functional stability of the brain slices of ground squirrels, Citellus undulatus, kept in conditions of prolonged deep periodic hypothermia: electrophysiological criteria. Neuroscience 1990; 38:591-8.
Michenfelder JD, Milde JH: The effect of profound levels of bypothermia (below 14 degrees C) on canine cerebral metabolism. J Cereb Blood Flow Metab 1992; 12:877-80.
Messick JM Jr, Casement B, Sharbrough FW, Milde LN, Michenfelder JD, Sundt TM Jr: Correlation of regional cerebral blood flow (rCBF) with EEG changes during isoflurane anesthesia for carotid endarterectomy: critical rCBF. ANESTHESIOLOGY 1987; 66:344-9.
Michenfelder JD, Sundt TM, Fode N, Sharbrough FW: Isoflurane when compared to enflurane and halothane decreases the frequency of cerebral ischemia during carotid endarterectomy. ANESTHESIOLOGY 1987; 67:336-40.
Espanol MT, Litt L, Yang GY, Chang LH, Chan PH, James TL, Weinstein PR: Tolerance of low intracellular pH during hypercapnia by rat cortical brain slices: A sup 31 Phosphorous/sup 1 Hydrogen NMR study. J Neurochem 1992; 59:1820-8.
Espanol MT, Xu Y, Litt L, Yang GY, Chang LH, James TL, Weinstein P, Chan PH: Modulation of glutamate-induced intracellular energy failure in neonatal cerebral cortical slices by kynurenic acid, dizocilpine, and NBQX. J Cereb Blood Flow Metab 1994; 14:269-78.
Espanol MT, Xu Y, Litt L, Chang L-H, James TL, Weinstein PR, Chan PH: sup 19 Fluorine NMR calcium changes, edema, and histology in neonatal rat brain slices during glutamate toxicity. Brain Res 1994; 647:172-6.
Chang LH, Shirane R, Weinstein PR, James TL: Cerebral metabolite dynamics during temporary complete oxygen/glucose deprivation in rats monitored by time-shared sup 1 Hydrogen and sup 31 Phosphorus NMR spectroscopy. Magn Reson Med 1990; 13:6-13.
Xu Y, Cohen Y, Litt L, Chang LH, James TL: Tolerance of low cerebral intracellular pH in rats during hyperbaric hypercapnia. Stroke 1991; 22:1303-8.
Chang LH, Pereira BM, Weinstein PR, Keniry MA, Murphy BJ, Litt L, James TL: Comparison of lactate concentration determinations in ischemic and hypoxic rat brains by in vivo and in vitro 1H NMR spectroscopy. Magn Reson Med 1987; 4:575-81.
Winer BJ: Statistical Principles in Experimental Design. New York, McGraw-Hill, 1971.
Glantz S: Primer of Biostatistics, 3rd edition. New York, McGraw Hill, 1992.
Chang L-H, James TL: NMR Studies of Brain Oxygen/glucose deprivation, Biological Magnetic Resonance. New York, Plenum Press, 1992, pp 135-158.
Van Zijl PC, Moonen CT, Faustino P, Pekar J, Kaplan O, Cohen JS: Complete separation of intracellular and extracellular information in NMR spectra of perfused cells by diffusion-weighted spectroscopy. Proc Natl Acad Sci USA 1991; 88:3228-32.
Katsura K, Kristian T, Smith ML, Siesjo BK: Acidosis induced by hypercapnia exaggerates ischemic brain damage. J Cereb Blood Flow Metab 1994; 14:243-50.
Eyre JA, Stuart AG, Forsyth RJ, Heaviside D, Bartlett K: Glucose export from the brain in man: evidence for a role for astrocytic glycogen as a reservoir of glucose for neural metabolism. Brain Res 1994; 635:349-52.
Lomako J, Lomako WM, Whelan WJ, Dombro RS, Neary JT, Norenberg MD: Glycogen synthesis in the astrocyte: from glycogenin to proglycogen to glycogen. FASEB J 1993; 7:1386-93.
Dringen R, Gebhardt R, Hamprecht B: Glycogen in astrocytes: possible function as lactate supply for neighboring cells. Brain Res 1993; 623:208-14.
Giffard R, Monyer H, Christine C, Choi D: Acidosis reduces NMDA receptor activation, glutamate neurotoxicity, and oxygen-glucose deprivation neuronal injury in cortical culture. Brain Res 1990; 506:339-42.
Simon RP, Niro M, Gwinn R: Brain acidosis induced by hypercarbic ventilation attenuates focal ischemic injury. J Pharmacol Exper Ther 1993; 267:1428-31.
Kristian T, Katsura K, Gido G, Siesjo BK: The influence of pH on cellular calcium influx during oxygen/glucose deprivation. Brain Res 1994; 641:295-302.
Hillered L, Ernster L, Siesjo BK: The influence of in vitro lactic acidosis and hypercapnia on respiratory activity of isolated rat brain mitochondria. J Cereb Blood Flow Metab 1984; 4:430-7.