Background and Purpose In vivo microdialysis was introduced in 1982 as a technique to study cerebral neurochemistry in awake, freely moving animals. In small animals, bilateral carotid occlusion produces a 7- to 10-fold increase in extracellular glutamate concentrations. This rapidly falls with reperfusion. Increase in extracellular glutamate is currently believed to be a major factor in initiating neuronal injury. Glutamate antagonists are currently undergoing clinical trials in acute stroke. Human data on the extracellular levels of glutamate and other amino acids in the normal or ischemic brain are limited. In this communication we wish to report the extracellular concentrations of glutamate, serine, glutamine, glycine, taurine, alanine, and γ-aminobutyric acid, as monitored by in vivo microdialysis, in the simulated ischemic model of the temporal lobe of the human brain.
Methods Intracerebral microdialysis was carried out in five patients who underwent resection of the temporal lobe for intractable epilepsy. Surgical excision leads to an acute (from partial to total, ie, from incomplete to complete) ischemic state of the resected brain. This was our model to study the changes in human extracellular fluid during acute focal ischemic conditions.
Results Extracellular glutamate concentrations were 15 to 30 μmol/L in the preischemic samples. This increased to 380.69±42.14 μmol/L with partial (incomplete) ischemia and reached a peak of 1781.67±292.34 μmol/L (>100-fold) with total isolation of the temporal pole (complete ischemia). The levels fell to 394.52±72.93 μmol/L 20 minutes after resection. Similar trends were observed with the onset of ischemia in the dialysate levels of serine, glutamine, glycine, alanine, taurine, and γ-aminobutyric acid.
Conclusions Our results show that there is a significant increase in extracellular glutamate and other neurotransmitters with ischemia in the temporal lobe model of the human brain. This increase is of a higher magnitude than that in small animals.
Changes in the chemical microenvironment are important in understanding the mechanisms of cerebral ischemia. It is now generally accepted that after an ischemic insult to the brain, the massive extracellular increases in the concentration of excitatory amino acids are toxic to the neurons.1 Most of the work substantiating this hypothesis is based on either global ischemia (bilateral carotid occlusion) or focal ischemia (carotid artery occlusion plus middle cerebral artery occlusion) in small animals.2 Translating the cause and effect of experiments on these models to the human brain is currently in progress, with several glutamate antagonists undergoing clinical trials.1 However, it is important to study the effects of ischemia on glutamate levels in humans to see whether a situation exists that is similar to that documented in small animals. Therefore, our objective was to develop a focal ischemic model of the human brain wherein neurochemical changes could be monitored as ischemia ensues. Temporal lobe resection for epilepsy surgery leads to an acute (from partial to total, ie, from incomplete to complete) ischemic situation in the resected brain. We used this as our model for acute focal ischemia. Microdialysis has become a widely accepted method for sampling the extracellular fluid of the brain.3 We herein describe a method of in vivo microdialysis on an acute focal ischemic model of the human brain. We measured seven extracellular amino acids and their response to acute (from partial to total, ie, from incomplete to complete) ischemia in the temporal lobes of awake or anesthetized patients subjected to surgery.
The only other published work that deals with a similar situation is the neurometabolic monitoring of the frontal lobe during resection of frontal lobe tumors in five patients.4 The purpose of the study was to monitor energy-related metabolites and amino acid transmitters in “tumor-free” frontal cortical tissue.
Subjects and Methods
Model of Ischemia
All patients scheduled for excision surgery for intractable epilepsy were approached to give their consent. Five patients with temporal lobe resection for intractable epilepsy were included in this study. Two of these patients were awake and had neuroleptic analgesia; the remainder had a general anesthetic with endotracheal intubation. The area of the temporal lobe planned for excision served as our model for in vivo microdialysis. The use of microdialysis in humans was approved by the University Advisory Committee on Ethics in Human Experimentation.
In Vivo Microdialysis
Priming of the Probes
Custom-designed CMA-10 probes with shaft length of 50 mm, diameter of 0.5 mm, and membrane lengths of 4 mm were soaked in a vial containing 70% ethanol in an in vitro CMA-130 stand for 15 minutes to wash out the glycerol. The inlet and the outlet channels were extended by 100 cm with the use of adapters and extra tubing. The inlet of the probe was then connected to a CMA-100 microinjection pump. The microsyringe was loaded with sterile Ringer’s solution (pH ≈6.7), as used in other in vivo studies of the human brain,4 5 6 and the unit was flushed at a rate of 15 μL/min for 10 to 15 minutes to clear the dead space and remove all residual glycerol, ethanol, and air bubbles. This was done concurrently while the neurosurgeons operated on the skull and exposed the temporal lobe.
Intraoperative Insertion of the Probes
The two probes were then handed to the surgeon and placed into the cortex of the temporal lobe at a depth of 0.5 to 1 cm. The inlet and the outlet tubes were loosely clipped to keep them out of the surgeon’s way. The probes rested on the raised muscle flap and were not fixed or anchored because this can easily lead to a shearing trauma to the brain. Surgical manipulations of the temporal pole were kept to a minimum to minimize any probe movements. The pump infused sterile Ringer’s solution at a rate of 2 μL/min.
Stabilization of the Probes
Sample collection was begun after insertion of the probes. As expected, the initial recordings were high (implantation trauma), in the range of 200 to 300 μmol/L. However, it was noted that at the end of 30 minutes, consistent baseline values between 15 and 30 μmol/L were recorded (Fig 1⇓). Therefore, preischemic monitoring began at the end of 30 minutes on stabilization of the baseline. This coincided well with the procedure of intraoperative electroencephalographic monitoring that was undertaken before excision of the temporal lobe.
The outlet tubing leads into small sterile plastic vials kept under ice. In vivo microdialysis was performed for more than 2 hours in all patients, and the samples were collected at 10-minute intervals (20 μL per collection). We were able to collect five to six samples before the commencement of surgical resection of the temporal lobe. Once the electroencephalographic recordings were completed and the area to be excised delineated, the surgeons began their excision by meticulous diathermy of all feeding vessels to the area. The temporal lobe was then excised by a combination of sharp and blunt dissection. Intraoperative in vivo microdialysis recordings were continued pari passu. On average, five collections were obtained from the start of ischemia (partial/incomplete) until the end (total/complete). The resected temporal lobe was moved to the microdialysis table with the probes in situ. Dialysis of the resected ischemic brain was continued for an additional 20 minutes. A total of six preischemic, four partial ischemic, one total ischemic, and two postresection dialysates were collected from each patient. The probes were subsequently removed from the resected brain and in vitro recovery of individual probes in standardized solutions obtained. All the dialysate specimens obtained were then transferred on ice to the laboratory for immediate analysis.
Complete sterility was maintained during all steps of the procedure, in keeping with the operative theater techniques and protocols. The entire microdialysis kit (ie, probes, vials, tubing, adapters, specimen bottles, etc) was gas sterilized. We found no significant differences in probe recovery of solutes in standardized solutions either before or after sterilization of the probes.
High-Performance Liquid Chromatography Analysis of Amino Acids
High-performance liquid chromatography analysis was obtained by a fully automated system with precolumn derivatization of amino acids with o-phthaldehyde 2-mercaptoethanol before electrochemical detection (Waters 460 electrochemical detector with a glassy carbon cell and 30-mm gasket) with a 715 Ultra Wisp Sample Processor and a Waters model 510 liquid pump, as detailed earlier.7 8 9 Standard curves of solutions ranging from 0.625, 1.25, 2.5, 25, 100, 300, to 1200 pmol were run in all experiments.
We calculated the mean±SEM for all data. Student’s t test (two-tailed) was used for further analysis of the glutamate response under ischemic conditions. The mean of three baseline preischemic values served as a control. Values were considered statistically significant at P<.01.
In our laboratory, seven extracellular amino acids are analyzed with precision to the 200-fmol range: glutamate, serine, glutamine, glycine, taurine, alanine, and γ-aminobutyric acid (GABA). For the first time, we have obtained baseline values for these extracellular amino acids in the human temporal lobe. The preischemic baseline values in the temporal lobe of the human brain are shown in the Table⇓.
The response of the amino acids, in particular extracellular glutamate, to an acute (from partial to total, ie, from incomplete to complete) ischemic insult is shown in Fig 1⇑. The dialysate glutamate concentrations began to increase from the baseline (preischemic) values of 20.22±3.39 μmol/L to 380.69±18.73 μmol/L (P<.0001) as the brain was rendered partially ischemic. With total ischemia, the values rose to 1783.47±196.01 μmol/L (P<.001), which is nearly a 100-fold increase in the levels of extracellular glutamate compared with the preischemic control levels. However, 20 minutes after resection the dialysate glutamate levels receded toward the baseline values. We can only speculate regarding the mechanism of this decline. Perhaps this reflects a diffusion/cellular exhaustion phenomenon, or it may be related to variable temperatures, ie, relative cooling of the resected brain in vitro. The observed massive increase during acute total ischemia is in complete agreement with other investigations using microdialysis techniques.2 10 Thus, similar to ischemic insults in small animals, the increase in extracellular glutamate with ischemia may be a major contributor to neuronal injury in humans as well.
A similar trend, albeit of a lesser magnitude, was also noted in the dialysate levels of serine, glycine, taurine, alanine, and GABA (Fig 2⇓) with the onset of ischemia. The glutamine levels rose from a baseline level of 1136.36±55.66 μmol/L to ischemic levels of 2331.76±16.63 to 3045.17±405.27 μmol/L and returned to levels of 633.34±7.7 μmol/L in the postresected collections.
The dialysate values were stable in the collections before the commencement of resection of the temporal lobe in all five patients. The increase in extracellular glutamate was sudden and very high with the onset of acute ischemia. In animals with transient ischemia we noted a 7- to 10-fold increase in extracellular glutamate that returned to baseline levels within minutes of reperfusion.10 11 In this context, the increase in extracellular glutamate in the completely isolated human temporal lobe was 10-fold higher. This may either reflect the complete, ie, total ischemia produced in our model or a true species difference. Whether this increase represents “true” increased glutamate release or decreased glutamate uptake is unknown. This 100-fold increase in extracellular glutamate with ischemia needs to be borne in mind when glutamate antagonists in clinical human trials are designed. The massive release of extracellular amino acids as seen in our model during resection of the temporal lobe is in all probability directly related to the acute ischemic insult delivered to the tissue. The sheer magnitude of the amino acid response probably rules out the possibility of cortical spreading depression leading to the accumulation of extracellular glutamate, as seen in animal studies.12
Human experience with in vivo microdialysis is limited. Initial work evaluated the catecholamine levels in the thalamus in parkinsonian patients.4 The dialysate levels of dopamine and its metabolites, GABA, aspartate, glutamate, and taurine, although initially high, returned to a steady state within 10 to 20 minutes with probe insertion. No experimental manipulations were carried out. Thereafter, neurometabolic monitoring of the tumor-free frontal region of the brain was done in an attempt to study human ischemia.5 Five patients were included in the study. The total duration of microdialysis ranged from 30 to 60 minutes. In patient 1 no data were available after 30 minutes of microdialysis. In patient 2 no data were available because the probes were fixed to the retractor, leading to erroneously high recordings, and on histology they were located in tumor tissue. In patient 3 data represented 40 minutes of microdialysis; there was pronounced edema in the tissue surrounding the probe in this patient. Patient 4 had microdialysis for 60 minutes. In patient 5 the total collection period was 30 minutes, and the tissue surrounding the probe was slightly edematous. The emphasis of this study was on energy-related metabolites, ie, lactate, adenosine, inosine, and hypoxanthine, leading to the conclusion that lactate may be a sensitive indicator of metabolic dysfunction in brain edema. In 1992 Persson and Hillered6 carried out chemical monitoring of neurosurgical intensive care patients using intracerebral microdialysis. The authors considered the lactate-pyruvate ratio to be a sensitive indicator of deranged cerebral metabolism. In 1993 During and Spencer13 investigated the usefulness of microdialysis in patients with epilepsy. Their study showed an increase in extracellular glutamate levels during seizures. In 1994 Scheyer et al14 measured the dialysate concentrations of carbamazepine and carbamazepine epoxide in three patients with intractable epilepsy.
We have measured, for the first time, seven amino acids and their response to an acute ischemic insult leading from a state of incomplete (partial) to complete (total) ischemia in the human temporal lobe. We recognize the limitations of this simulated ischemia model with regard to the variability of blood flow changes during resection that was not measured. It is still a reliable model that reflects the chemical changes in the extracellular fluid during ischemia as surgical resection leads to complete (total) ischemia in all cases.
The ischemic model as described in this communication can effectively be used to study a whole array of neurochemical events in the human brain. Morphological and physiological factors, autoradiography of receptors, probe electrode manipulations, and chemical (drug) monitoring of neurochemistry are but a few subjects that may be evaluated with this model.
In conclusion, our results support the clinical potential of in vivo microdialysis to study the role of excitotoxins and neuromodulators in the simulated ischemic model of the temporal lobe of the human brain.
Reprint requests to Ashfaq Shuaib, MD, FRCPC, Department of Medicine, Saskatchewan Stroke Research Centre, Royal University Hospital, Saskatoon, Saskatchewan, Canada S7N OXO.
- Received September 6, 1994.
- Revision received November 9, 1994.
- Accepted January 31, 1995.
- Copyright © 1995 by American Heart Association