Plastic biochannel hybridization devices: a new concept for microfluidic DNA arrays

A xanthine oxidase hydroxyl radical (.OH)-generating system was created for sustained in vitro production of *OH. This assay was coupled with microdialysis sampling to elucidate the factors that influence microdialysis calibration during radical trapping. A *OH trapping agent, 4-hydroxybenzoic acid, was included either in the microdialysis perfusion fluid or in the medium external to the microdialysis probe. Xanthine oxidase enzymatic activity was reproducible and had an average activity measured by UV absorbance of produced uric acid of 0.037 +/- 0.005 deltaAU/min (n = 5). A considerable amount of variance in the rate and amount of the product, 3,4-dihydroxybenzoic acid (3,4-DHBA), was observed when one microdialysis probe was placed in the reaction mixture. When two microdialysis probes were placed in the reaction mixture, a greater rate and amount of 3,4-DHBA was observed. Different concentrations of 3,4-DHBA were obtained between quiescent and stirred systems.     

Flow dependent changes in the effective surface area of microdialysis probes

The relative efficiencies of microdialysis probes were determined both in vitro and in vivo using tritiated water. Tritiated water (THO) freely distributes throughout the fluid spaces of an experimental animal and, at equilibrium, the brain extracellular concentration of THO is the same as the plasma concentration. Microdialysis probes were inserted into the right caudoputamen of anesthetized rats. The rats were injected with THO and after one hour microdialysis samples were collected at flow rates between 0.2 and 10.0 ul/min. The in vitro relative efficiency for THO was computed as the ratio of the THO concentration in the dialysate to that of the solution the probe was immersed in. The in vivo relative efficiency was computed as the ratio of the concentration of THO in the brain dialysate to that measured in the plasma of the rat. Both the in vitro and in vivo relative efficiencies for THO decrease with increasing flow rates, but they differ from each other except at very low flow rates (less than 0.25 ul/min). The in vitro relative efficiency at a given probe flow is the maximum efficiency that can be attained in vivo at that flow. The surface of effective exchange (Se) is the fraction of that maximum which is attained in vivo. This study also demonstrates how the effective surface area can be computed at any probe flow rate and how it can be used as a correction factor.     

Simultaneous measurement of blood and brain microdialysates of granisetron in rat by high-performance liquid chromatography with fluorescence detection

Simultaneous microdialysis probes in the blood and brain and sensitive high-performance liquid chromatography with fluorescence detection were used to examine the granisetron concentration in the jugular vein and frontal cortex of rats after drug administration. Two microdialysis probes were inserted into the right jugular vein and frontal cortex of male Sprague–Dawley rats to which granisetron (6 mg/kg, i.v.) had been administered. Dialysates were automatically collected using a microfraction collector. Samples were eluted with a mobile phase containing 25 mM acetate buffer (pH 4.8)–acetonitrile (72:28, v/v). Excitation and emission wavelengths were set at 305 and 360 nm, respectively, on a scanning fluorescence detector. The limit of quantification for granisetron was 0.5 ng/ml. The in vitro recovery of granisetron was 29.7±1.2% (n=6) for the jugular vein microdialysis probe and 6.1±0.5% (n=6) for the frontal cortex microdialysis probe. The increasing brain/blood concentration ratio of granisetron suggests that granisetron penetrates the blood–brain barrier.     

Evaluation of extracellular lipid peroxidation in brain cortex of anaesthetized rats by microdialysis perfusion and high-performance liquid chromatography with fluorimetric detection

A method for in vivo evaluation of lipid peroxidation in the extracellular space of anaesthetized rat brain cortex was developed. This method involved the use of microdialysis perfusion and high-performance liquid chromatography. The microdialysates, eluted from implanted probes, were reacted with thiobarbituric acid (TBA) prior to analysis by an HPLC system equipped with a fluorescence detector (excitation and emission wavelengths were 515 and 550 nm, respectively). Lipid peroxidation in the extracellular space was evaluated as the concentration of malondialdehyde, a lipid peroxidation end product which reacts with TBA to form a fluorescent conjugate. Significantly increased production of malondialdehyde following hydrogen peroxide perfusion (0.03%, 0.3% at a flow-rate of 1 μl/min) was observed in the brain cortex of anaesthetized rats.     

On-line derivatization for continuous and automatic monitoring of brain extracellular glutamate levels in anesthetized rats: a microdialysis study

Glutamate is an important excitatory amino acid in central nervous system. We developed a method for in vivo, continuous and automatic monitoring of brain extracellular glutamate, as well as other amino acids in anesthetized rat. This method involves the use of microdialysis perfusion technique and a high-performance liquid chromatography system equipped with a fluorescence detector. The microdialysate (perfused at a flow-rate of 1 μl/min) was on-line derivatized with o-phthaldehyde (perfused at 2 μl/min) through a mixing tee prior to the injection onto the HPLC column. The efficiency of this on-line derivatization was equivalent to that performed with an off-line manner. The effect of cerebral ischemia (2 h) and reperfusion (2 h) in brain cortex of anesthetized rats was monitored using this method. In addition to glutamate, extracellular concentrations of other amino acids, such as aspartate, glutamine, glycine, taurine and γ-aminobutyric acid, were also simultaneously monitored with this on-line method. Since monitoring of extracellular amino acids by microdialysis perfusion is intensively used in neuroscience investigations, this simple and convenient method would be useful in the future applications.     

Determination of baicalin in rat cerebrospinal fluid and blood using microdialysis coupled with ultra-performance liquid chromatography-tandem mass spectrometry

An in vivo microdialysis sampling method coupled with ultra-performance liquid chromatography-tandem mass spectrometry (UPLC-MS/MS) was employed for continuous simultaneous monitoring of unbound baicalin in rat blood and brain. Microdialysis probes were inserted into the jugular vein and brain cerebrospinal fluid (CSF) of Sprague-Dawley rats then, following administration of baicalin at doses of 24mg/kg via the candal vein, samples were collected every 20min and injected directly into the UPLC-MS/MS system. In vitro recoveries of the probes were 19.26% and 18.38%, while in vivo recoveries of the probes were 15.0% and 17.52% for blood and brain, respectively. This improved method offers a rapid quantitative procedure for the determination of baicalin with a retention time of only 1.6min. The lower limit of quantification (LLOQ) and the lower limit of detection (LLOD) based on a signal-to-noise ratio of 5 were 2.37 and 0.1ng/ml for anticoagulant citrate dextrose (ACD) solution, and 1.185 and 0.3ng/ml for artificial cerebrospinal fluid (aCSF), respectively. The pharmacokinetics results indicated that baicalin could pass through the blood-brain barrier (BBB) and was detectable in brain dialysate. These in vivo microdialysis-based measurements provide a technique for simple sampling and rapid sensitive analysis of unbound baicalin in rat blood and CSF and for further application in pharmacokinetic studies.     

Hemorrhage-induced regional brain thyrotropin-releasing hormone release in conscious rats measured by microdialysis

The septum, nucleus accumbens and preoptic area in the brains of conscious, freely moving rats were perfused using microdialysis probes. The TRH concentration significantly increased in the septum after withdrawal of 30% of the total blood volume but remained at constant levels in the other brain areas. Also, high potassium dose-dependently stimulated TRH release in vivo. These results suggest that blood loss stimulates septal TRH release, probably by membrane depolarization of TRH-containing nerve terminals.     

Determination of Kainic Acid-Induced Release of Nitric Oxide Using a Novel Hemoglobin Trapping Technique with Microdialysis

We have demonstrated the usefulness of a novel hemoglobin-trapping technique to quantify nitric oxide (NO) concentrations in vivo. Concentric microdialysis probes were implanted into the hippocampus of rats under urethane anesthesia and perfused with 1 μM oxyhemoglobin in artificial CSF to sequester NO in extracellular fluid. The concentration of methemoglobin was then determined spectrophotometrically. The basal level of NO in hippocampus was 2.2 ± 0.5 nM(in vitro sensitivity of the probe was 0.2 nM). Administration of 13 mg/kg, i.p., of kainic acid (KA) produced a maximal 5.3-fold increase at 100 min in NO levels (11.8 ± 0.2 nM). This response was significantly attenuated by pretreatment with the NO synthase inhibitor N-monomethyl-L-arginine (50 mg/kg, 30 min before KA). These results demonstrate that a microdialysis probe using a novel hemoglobin-trapping technique possesses adequate sensitivity to determine the basal levels of NO and document the ability of KA to increase these levels via a NO synthase-mediated mechanism.     

Hydroxyl radicals detected via brain microdialysis in rats breathing air and during hyperbaric oxygen convulsions

Microdialysis was done on 300-400 g, awake, male rats with microdialysis probes inserted through guide cannulas into the striatum (Bregma co-ordinates A 0.5, L 2.9, D -4.0 for guide cannulas implanted 5 days previously). Rats were exposed to hyperbaric oxygen (HBO; 6 atm absolute, 5 atm gauge pressure of oxygen with carbon dioxide absorbed by soda lime). Artificial cerebrospinal fluid (CSF) containing 5 mM sodium salicylate was perfused at 1 microl/min and collected over sequential 10 min intervals with rats breathing air, then HBO, and after decompression. Times to convulsions and duration and severity of convulsions were observed and recorded. CSF samples were analyzed for 2,3- and 2,5-dihydroxybenzoic acid (DHBA), reaction products of hydroxyl radicals with salicylate, by HPLC and compared to authentic standards. Recovery of DHBAs was 48% from fluid surrounding microdialysis probes, based on in vitro tests. The average time to the first convulsion was 21 min and rats convulsed an average of 4 times during 40 min in HBO. There were no significant differences in hydroxyl radical production by this protocol during any of the 10 min collection periods in air or HBO (average in pmoles for 10 microl of all samples: 2,3-DHBA = 7.0 +/- 2.5 and 2,5-DHBA = 11.3 +/- 4.1). The failure to detect an increase in hydroxyl radicals in HBO prior to or during convulsions appears valid since each rat served as its own control.     

Effects of α-Phenyl-tert-Butylnitrone and Selegiline on Hydroxyl Free Radicals in Rat Striatum Produced by Local Application of Glutamate

Abstract: The hydroxyl radical is a very reactive oxygen species that damages biomolecules in the brain and in other tissues. The possible pharmacological intervention to prevent hydroxyl radical formation was studied in vivo using the microdialysis technique in brains of nonanesthetized rats. Hydroxyl radicals form stable adducts [mainly 2,3-dihydroxybenzoic acid (2,3-DHBA) and 2,5-DHBA)] via an aromatic hydroxylation reaction with salicylic acid. 2,3-DHBA was separated and quantified by HPLC and electrochemical detection. Microdialysis probes were implanted into the striatum 1 day before measurement of levels of hydroxyl radicals. The next day, the probes were first perfused for 120 min with a modified Ringer's solution containing 5 m M salicylic acid, to obtain stable baselines. Afterward, the perfusion solution was switched to another solution that in addition contained 50 m M glutamate, to stimulate radical formation. Twenty minutes later, α-phenyl- tert -butylnitrone (PBN; 100 mg/kg), selegiline (10 mg/kg), or saline was administered intraperitoneally. The glutamate perfusion produced marked two- to 2.5-fold increases in 2,3-DHBA content. Treatment with PBN significantly antagonized the rise of 2,3-DHBA level, indicating that PBN is a direct radical scavenger not only in vitro but also in vivo. Acute treatment with selegiline failed to reduce significantly the glutamate-induced radical formation. The acute experiments presented here do not support the suggestion that the neuroprotective effects of selegiline described in the literature are due to a potential hydroxyl radical scavenging property of the drug.     

Effects of Different Semipermeable Membranes on In Vitro and In Vivo Performance of Microdialysis Probes

The in vitro and in vivo performance of three different semipermeable microdialysis membranes was compared: a proprietary polycarbonate-ether membrane made by Carnegie Medecin; cuprophan, a regenerated cellulose membrane; and polyacrylonitrile. When microdialysis probes were tested in a stirred in vitro solution, large and statistically significant differences among the three membranes in extraction of acid metabolites (3,4-dihydroxyphenylacetic acid, 5-hydroxyindoleacetic acid, and homovanillic acid) and acetaminophen were found. Polyacrylonitrile had the highest extractions in vitro. In contrast, when microdialysis probes were implanted in vivo (in rat striatum), extraction of acid metabolites and acetaminophen did not differ significantly among the different membranes. These results are consistent with predictions made by a mathematical model of microdialysis and can be explained by the fact that in vitro the main factor limiting extraction is membrane resistance to diffusion, whereas tissue resistance to diffusion plays a more dominant role in vivo. These findings suggest that (aside from differences in surface area), the choice of semipermeable membrane will generally have little effect on in vivo microdialysis results. Furthermore, in vitro measurements of microdialysis probe extractions are not a reliable way of calibrating in vivo performance.     

Amino acid concentrations in hypothalamic and caudate nuclei during microwave-induced thermal stress: Analysis by microdialysis

Exposure to radiofrequency radiation (RFR) may produce thermal responses. Extracellular amino acid concentrations in the hypothalamus (Hyp) and caudate nucleus (CN) were measured by using in vivo microdialysis before and during exposure to RFR. Under urethane anesthetic, each rat was implanted stereotaxically with a nonmetallic microdialysis probe and temperature probe guides and then placed in the exposure chamber. The rat laid on its right side with its head and neck placed directly under the wave guide. Temperature probes were placed in the left brain, right brain, face (subcutaneously), left tympanum, and rectum. Each microdialysis sample was collected over a 20 min period. The microdialysis probe was perfused for 2 h before the rat was exposed to 5.02 GHz radiation (10 μs pulse width, 1000 pulses/s). The right and left sides of the brain were maintained at approximately 41.2 and 41.7 °C, respectively, throughout a 40 min exposure period. Initially when the brain was being heated to these temperatures, the time-averaged specific absorption rates (SARs) for the right and left sides of the brain were 29 and 40 W/kg, respectively. Concentrations of aspartic acid, glutamic acid, serine, glutamine, and glycine in dialysate were determined by using high-pressure liquid chromatography with electrochemical detection. In the Hyp and CN, the concentrations of aspartic acid, serine, and glycine increased significantly during RFR exposure (P < .05). These results indicate that RFR-induced thermal stress produces a general change in the amino acid concentrations that is not restricted to thermoregulatory centers. Changes in the concentrations of glutamic acid (Hyp, P = .16; CN, P = .34) and glutamine (Hyp, P = .13; CN, P = .10) were not statistically significant. Altered amino acid concentrations may reveal which brain regions are susceptible to damage in response to RFR-induced thermal stress. Bioelectromagnetics 18:277–283, 1997. © 1997 Wiley-Liss, Inc.     

Lack of antilipolytic effect of lactate in subcutaneous abdominal adipose tissue during exercise.

The purpose of our study was to evaluate the potential inhibition of adipose tissue mobilization by lactate. Eight male subjects (age, 26. 25 +/- 1.75 yr) in good physical condition (maximal oxygen uptake, 59.87 +/- 2.77 ml. kg-1. min-1; %body fat, 10.15 +/- 0.89%) participated in this study. For each subject, two microdialysis probes were inserted into abdominal subcutaneous tissue. Lactate (16 mM) was perfused via one of the probes while physiological saline only was perfused via the other, both at a flow rate of 2.5 microl/min. In both probes, ethanol was also perfused for adipose tissue blood flow estimation. Dialysates were collected every 10 min during rest (30 min), exercise at 50% maximal oxygen consumption (120 min), and recovery (30 min) for the measurement of glycerol concentration. During exercise, glycerol increased significantly in both probes. However, no differences in glycerol level and ethanol extraction were observed between the lactate and control probes. These findings suggest that lactate does not impair subcutaneous abdominal adipose tissue mobilization during exercise.     

Redistribution of fetal circulation during repeated asphyxia in sheep: effects on skin blood flow, transcutaneous PO2, and plasma catecholamines

To improve the understanding of fetal responses to labour, we have ascertained whether reduced fetal skin blood flow after asphyxia reflects redistribution of the circulation, and if so, whether this can be detected by transcutaneous PO2 monitoring. We also studied the relation between plasma concentrations of catecholamines and organ blood flow. Eight experiments were conducted on 8 acutely-prepared fetal sheep in utero between 125 and 135 days of gestation. In each fetus 11 episodes of asphyxia were induced within 33 min by intermittent arrest of uterine blood flow for 90 s. The distribution of blood flow was measured before and after asphyxia (at 35.5 min) by the isotope-labelled microsphere method. Blood samples were drawn at 0, 33 (i.e. after 90 s recovery), and 40 min to determine blood gases, acid-base balance, and catecholamine concentrations. Fetal transcutaneous PO2, heart rate, arterial blood pressure, and arterial O2 saturation were recorded continuously. Repeated fetal asphyxia increased plasma catecholamine concentrations and caused a circulatory redistribution to the brain (181% change), adrenals (116% change), and lungs (105% change) at the expense of many peripheral organs, particularly of the skin (-61% change). The pattern of these changes was different from that observed by others in persistent hypoxia or asphyxia. The decrease in skin blood flow, which depressed transcutaneous PO2 and increased the arterial-transcutaneous PO2 difference, correlated with the decrease in blood flow to other peripheral organs and with an increase in blood flow to the brain stem. We conclude that reduced blood flow to the fetal skin after repeated episodes of asphyxia indicates circulatory redistribution, which can be detected by transcutaneous PO2 measurements. We suggest that monitoring of variables that depend on skin blood flow may improve fetal surveillance during complicated labour.     

Spin label electron paramagnetic resonance study in thylakoid membranes from a new herbicide-resistant D1 mutant from soybean cell cultures deficient in fatty acid desaturation

We examined the effect of fluvastatin, a 3-hydroxy-3-methylglutaryl coenzyme A (HMG-CoA) reductase inhibitor, on the production of hydroxyl radical (*OH) generation via nitric oxide synthase (NOS) activation by an in vivo microdialysis technique. The microdialysis probe was implanted in the left ventricular myocardium of anesthetized rats and tissue was perfused with Ringer's solution through the microdialysis probe at a rate of 1 microl/min. Sodium salicylate in Ringer's solution (0.5 nmol/microl/min) was infused directly through a microdialysis probe to detect the generation of *OH. Induction of [K(+)](o) (70 mM) or tyramine (1 mM), significantly increased the formation of *OH trapped as 2,3-dihydroxybenzoic acid (DHBA). The application of N(G)-nitro-L-arginine methyl ester (L-NAME), a NOS inhibitor, significantly decreased the K(+) depolarization-induced *OH formation, but the effect of tyramine significantly increased the level of 2,3-DHBA. When fluvastatin (100 microM), an inhibitor of low-density lipoprotein (LDL) oxidation, was administered to L-NAME-pretreated animals, both KCl and tyramine failed to increase the level of 2,3-DHBA formation. The effect of fluvastatin may be unrelated to K(+) depolarization-induced *OH generation. To examine the effect of fluvastatin on ischemic/reperfused rat myocardium, the heart was subjected to myocardial ischemia for 15 min by occlusion of the left anterior descending coronary artery (LAD). When the heart was reperfused, a marked elevation of the level of 2,3-DHBA was observed. However, in the presence of fluvastatin (100 microM), the elevation of 2,3-DHBA was not observed in ischemia/reperfused rat heart. Fluvastatin, orally at a dose of 3 mg/kg/day for 4 weeks, significantly blunted the rise of serum creatine phosphokinase and improved the electrocardiogram 2 h after coronary occlusion. These results suggest that fluvastatin is associated with a cardioprotective effect due to the suppression of noradrenaline-induced *OH generation by inhibiting LDL oxidation in the heart.     

Adenosine deaminase activity in rat intestine: assay with a microdialysis technique

Using microdialysis, we measured adenosine deaminase activity in rat intestine by detecting inosine, a breakdown product of adenosine. The dialysis probe consisted of a 3 x 0.22 mm dialysis fiber with a 50,000 mol wt cut off. When the probe was perfused at 1 microl/min in vitro, the average relative recovery rate of inosine was 22.1+/-0.9%). The dialysis probe was implanted in the intestinal mucosa and perfused with Tyrode solution containing adenosine at 1 microl/min. The dialysate samples were analyzed for inosine by high-performance liquid chromatography with ultraviolet (HPLC-UV) detection at 260 nm. When adenosine (100-1000 microM) was perfused, the level of inosine increased dose-dependently and was saturatable at about 1 mM adenosine. The ED50 of adenosine was 192.6 microM, with a maximum attainable inosine concentration of 59.7 microM. In the presence of aminoguanidine, a adenosine deaminase inhibitor (10 mM or 10 n mol/microl/min), the elevation of inosine was not observed. The dialysis technique makes it possible to measure adenosine deaminase activity in intestinal mucosa.     

A small, removable microdialysis probe

A miniaturized, concentric, microdialysis probe is described. It is constructed from 36 gauge stainless steel tubing inside of 26 gauge tubing, with a cellulose hollow fiber tip 0.2 mm in diameter and 2 mm long. It has a 6000 molecular weight cut off that excludes enzymes but collects monoamines, their metabolites, and other small neurochemicals. In vitro tests show relative recovery rates of 5-10%. Absolute recovery measured in picograms was independent of the perfusate flow rate inside the probe. Tests in awake rats with probes in the nucleus accumbens showed stable amounts of catecholamines and metabolites collected during repeated 20 min samples. After ip amphetamine, release of dopamine in the accumbens increased from 20 to 40 pg per sample while DOPAC and HVA decreased from about 1500 to 500 pg. Tests of multiple site sampling succeeded in obtaining norepinephrine and dopamine plus three metabolites (DOPAC, HVA and 5HIAA) from four probes simultaneously in four different brain sites in each rat. Five day continuous samples or monthly intermittent samples can be obtained with this microdialysis probe.     

6-hydroxydopamine increases the hydroxylation and nitration of phenylalanine in vivo: implication of peroxynitrite formation

In the present study, we investigated the effect of the dopaminergic neurotoxin 6-hydroxydopamine (6-OHDA) on hydroxyl free radical and peroxynitrite formation in vivo using D-phenylalanine as a novel mechanistic probe. In vivo microdialysis was carried out in the striatum of freely moving male Wistar rats. The microdialysis probes were perfused with artificial cerebrospinal fluid containing 5 mM D-phenylalanine (flow rate 2 microL/min). After obtaining a stable baseline 6-OHDA was delivered into the striatum via reverse microdialysis for 60 min. HPLC measurements of the effluent were performed using photodiode array detection for determination of phenylalanine derived o-tyrosine and m-tyrosine (as hydroxylation markers) as well as of nitrotyrosine and nitrophenylalanine (as nitration markers). The basal levels of the hydroxylation derived products of phenylalanine were approximately 100-fold higher than those of the nitration derived products. 6-OHDA (0.1, 1, 10 mM) significantly increased o- and m-tyrosine up to nine- and 13-fold, respectively, whereas levels of 3-nitrotyrosine and 4-nitrophenylalanine were significantly increased up to 422- and 358-fold, respectively. The results demonstrate that phenylalanine is a sensitive in vivo marker for 6-OHDA-induced hydroxylation and nitration reactions which are clearly concentration dependent. We conclude that peroxynitrite formation is involved in 6-OHDA-induced neurochemical effects.     

Fluoxetine Partly Exerts its Actions Through GABA: A Neurochemical Evidence

Fluoxetine, as a serotonin re-uptake inhibitor augments serotonin concentration within the synapse by inhibiting the serotonin transporter. The contribution of amino acids has also been shown in depression. We hypothesized that fluoxetine exerts its actions at least in part by intervening brain signaling operated by amino acid transmitters. Therefore the aim of this study is to supply neurochemical evidence that fluoxetine produces changes in amino acids in cerebrospinal fluid of rats. Sprague-Dawley rats were anesthetized and concentric microdialysis probes were implanted stereotaxically into the right lateral ventricle. Intraperitoneal fluoxetine (2.5 or 5 mg/kg) or physiological saline was administered and the probes were perfused with artificial cerebrospinal fluid at a rate of 1 μl/min. In the chronic fluoxetine group, the rats were treated daily with oral fluoxetine solution or inert syrup for 3 weeks. The microdialysis probes were placed on the 21st day and perfused the next day. Fluoxetine was ineffective in changing the cerebrospinal fluid GABA levels at the dose of 2.5 mg/kg but produced a significant increase in the perfusates following injection of 5 mg/kg of fluoxetine (P < 0.05). Oral fluoxetine administration (5 mg/kg) for 21 days also elevated the CSF GABA levels by approximately 2-fold (P < 0.05). l-glutamic acid levels were not affected in all groups. These neurochemical findings show that fluoxetine, a selective serotonin re-uptake inhibitor affects brain GABA levels indirectly, and our results suggest that acute or chronic effects may be involved in beneficial and/or adverse effects of the drug.     

Hydroxyl radicals detected via brain microdialysis in rats breathing air and during hyperbaric oxygen convulsions

Abstract

Microdialysis was done on 300–400 g, awake, male rats with microdialysis probes inserted through guide cannulas into the striatum (Bregma co-ordinates A 0.5, L 2.9, D –4.0 for guide cannulas implanted 5 days previously). Rats were exposed to hyperbaric oxygen (HBO; 6 atm absolute, 5 atm gauge pressure of oxygen with carbon dioxide absorbed by soda lime). Artificial cerebrospinal fluid (CSF) containing 5 mM sodium salicylate was perfused at 1 µl/min and collected over sequential 10 min intervals with rats breathing air, then HBO, and after decompression. Times to convulsions and duration and severity of convulsions were observed and recorded. CSF samples were analyzed for 2,3- and 2,5-dihydroxybenzoic acid (DHBA), reaction products of hydroxyl radicals with salicylate, by HPLC and compared to authentic standards. Recovery of DHBAs was 48% from fluid surrounding microdialysis probes, based on in vitro tests. The average time to the first convulsion was 21 min and rats convulsed an average of 4 times during 40 min in HBO. There were no significant differences in hydroxyl radical production by this protocol during any of the 10 min collection periods in air or HBO (average in pmoles for 10 µl of all samples: 2,3-DHBA = 7.0 ± 2.5 and 2,5-DHBA = 11.3 ± 4.1). The failure to detect an increase in hydroxyl radicals in HBO prior to or during convulsions appears valid since each rat served as its own control.