Electrophysiological properties of CA1 neurons protected by postischemic hypothermia in gerbils

Hypothermia: Impact on plasticity following brain injury

Therapeutic hypothermia (TH) is a potent neuroprotectant against multiple forms of brain injury, but in some cases, prolonged cooling is needed. Such cooling protocols raise the risk that TH will directly or indirectly impact neuroplasticity, such as after global and focal cerebral ischemia or traumatic brain injury. TH, depending on the depth and duration, has the potential to broadly affect brain plasticity, especially given the spatial, temporal, and mechanistic overlap with the injury processes that cooling is used to treat. Here, we review the current experimental and clinical evidence to evaluate whether application of TH has any adverse or positive effects on postinjury plasticity. The limited available data suggest that mild TH does not appear to have any deleterious effect on neuroplasticity; however, we emphasize the need for additional high-quality preclinical and clinical work in this area.

Mild hypothermia as a treatment for central nervous system injuries: Positive or negative effects

Besides local neuronal damage caused by the primary insult, central nervous system injuries may secondarily cause a progressive cascade of related events including brain edema, ischemia, oxida-tive stress, excitotoxicity, and dysregulation of calcium homeostasis. Hypothermia is a beneficial strategy in a variety of acute central nervous system injuries. Mild hypothermia can treat high intra-cranial pressure following traumatic brain injuries in adults. It is a new treatment that increases sur-vival and quality of life for patients suffering from ischemic insults such as cardiac arrest, stroke, and neurogenic fever following brain trauma. Therapeutic hypothermia decreases free radical produc-tion, inflammation, excitotoxicity and intracranial pressure, and improves cerebral metabolism after traumatic brain injury and cerebral ischemia, thus protecting against central nervous system dam-age. Although a series of pathological and physiological changes as well as potential side effects are observed during hypothermia treatment, it remains a potential therapeutic strategy for central nervous system injuries and deserves further study.

Combination therapy with insulin-like growth factor-1 and hypothermia synergistically improves outcome after transient global brain ischemia in the rat

OBJECTIVES

Hypothermia has a well-established neuroprotective effect and offers a foundation for combination therapy for brain ischemia. The authors evaluated the effect of combination therapy with insulin-like growth factor-1 (IGF-1) and hypothermia on brain structure and function in the setting of global brain ischemia and reperfusion in rats.

METHODS

Male Sprague-Dawley rats were randomly assigned to groups by a registrar. Animals were subjected to 8 minutes of global brain ischemia using bilateral carotid occlusion and systemic hypotension, followed by 7 days (Stage I dose studies) or 28 days (Stage II outcome studies) of reperfusion. Sham controls were subjected to surgery, but not ischemia. Stage II animals were randomized to no treatment, IGF-1 at the dose determined in Stage I, hypothermia (32°C for 4 hours), or a combination of IGF-1 and hypothermia. Stage II animals underwent 21 days of spatial memory testing. At 7 days (Stage I) or 28 days (Stage II), brains were harvested for counting of CA1 neurons. The primary Stage II outcome was a neurologic outcome index computed as the ratio of viable CA1 neurons per 300-μm field to the number of days to reach success criteria on the memory task.

RESULTS

Stage I experiments confirmed the neuroprotective effect of the hypothermia protocol and IGF-1 at a dose of 0.6 U/kg. Stage II studies suggested that early neuroprotection with hypothermia and IGF-1 was not well maintained to 28 days and that combination therapy was more beneficial than either IGF-1 or hypothermia alone. Median and interquartile ranges (IQRs) of viable neurons per 300-μm field were 114 (IQR = 99.5 to 136) for sham, three (IQR = 2 to 4.8) for untreated ischemia, four (IQR = 3 to 70.25) for ischemia treated with IGF-1 alone, 25 (IQR = 3 to 70) for ischemia treated with hypothermia alone, and 78 (IQR 47.3 to 97.5) for ischemia treated with combination therapy. Days to memory success criteria were 13.6 (IQR = 11.5 to 15.5 days) for sham, 23.5 (IQR = 20 to 25.5 days) for untreated ischemia, 17.5 (IQR = 15.5 to 25.5 days) for ischemia treated with IGF-1, 15 (IQR = 14.5 to 21 days) for ischemia treated with hypothermia, and 13.5 (IQR = 12.25 to 18.5 days) for ischemia treated with combination therapy. Neurologic outcome indices were 8.5 (IQR = 7.4 to 9.5) for sham, 0.14 (IQR = 0.08 to 0.2) for untreated ischemia, 0.18 (IQR = 0.17 to 4.6) for ischemia treated with IGF-1, 0.7 (IQR = 0.2 to 4.8) for ischemia treated with hypothermia, and 5.7 (IQR = 3.3 to 6.2) for ischemia treated with combination therapy. Statistically significant differences in neuron counts, days to memory test criteria, and outcome index were found between sham and untreated ischemic animals. Of the three treatment regimens, only combination therapy showed a statistically significant difference from the untreated ischemic group for neuronal salvage (p = 0.02), days to criteria (p = 0.043), and outcome index (p = 0.014).

CONCLUSIONS

Combination therapy with IGF-1 (0.6 U/kg) and therapeutic hypothermia (32°C for 4 hours) at the onset of reperfusion synergistically preserves CA1 structure and function at 28 days after 8 minutes of global brain ischemia in healthy male rats.

Prolonged therapeutic hypothermia does not adversely impact neuroplasticity after global ischemia in rats

Hypothermia improves clinical outcome after cardiac arrest in adults. Animal data show that a day or more of cooling optimally reduces edema and tissue injury after cerebral ischemia, especially after longer intervention delays. Lengthy treatments, however, may inhibit repair processes (e.g., synaptogenesis). Thus, we evaluated whether unilateral brain hypothermia (∼33°C) affects neuroplasticity in the rat 2-vessel occlusion model. In the first experiment, we cooled starting 1 hour after ischemia for 2, 4, or 7 days. Another group was cooled for 2 days starting 48 hours after ischemia. One group remained normothermic throughout. All hypothermia treatments started 1 hour after ischemia equally reduced hippocampal CA1 injury in the cooled hemisphere compared with the normothermic side and the normothermic group. Cooling only on days 3 and 4 was not beneficial. Importantly, no treatment influenced neurogenesis (Ki67/Doublecortin (DCX) staining), synapse formation (synaptophysin), or brain-derived neurotropic factor (BDNF) immunohistochemistry. A second experiment confirmed that BDNF levels (ELISA) were equivalent in normothermic and 7-day cooled rats. Last, we measured zinc (Zn), which is important in plasticity, with X-ray fluorescence imaging in normothermic and 7-day cooled rats. Hypothermia did not alter the postischemic distribution of Zn within the hippocampus. In summary, cooling significantly mitigates injury without compromising neuroplasticity.

Prolonged hypothermia in rat: a safety study using brain-selective and systemic treatments

Hypothermia is an effective neuroprotectant for cardiac arrest and perinatal ischemic injury. Hypothermia also improves outcome after traumatic brain injury and stroke. Although the ideal treatment parameters (duration, delay, and depth) are not fully delineated, prolonged cooling is usually more effective than shorter periods. There is the concern that extended cooling may be hazardous to brain plasticity and cause damage. In order to evaluate this possibility, we assessed the effects of 3 days of systemic hypothermia (32°C) in rats subjected to a sham stroke surgery. There were no detrimental behavioral effects or signs of brain damage. As even longer cooling may be needed in some patients, we cooled (∼32°C) the right hemisphere of rats for 3 or 21 days. Physiological variables, functional outcome, and measures of cell injury were examined. Focal brain cooling for 21 days modestly decreased heart rate, blood pressure, and core temperature. However, focal hypothermia did not affect subsequent behavior (e.g., spontaneous limb usage), cell morphology (e.g., dendritic arborization, ultrastructure), or cause cell death. In conclusion, prolonged mild hypothermia did not harm the brain of normal animals. Further research is now needed to evaluate whether such treatments affect plasticity after brain injury.

The "Refrige-a-RAT-or": an accurate, inexpensive, and clinically relevant small animal model of therapeutic hypothermia

BACKGROUND

Physical and molecular mechanisms for the neuroprotective effect of therapeutic hypothermia are not completely understood, and new therapeutic applications incorporating hypothermia remain to be developed and tested. Clinically relevant animal models of therapeutic hypothermia are not well established or consistent.

OBJECTIVES

The objective was to develop and test an inexpensive small animal therapeutic hypothermia system that models those in widespread clinical use and verify that such a system confers neuroprotection in a rat model of global brain ischemia.

METHODS

A water-cooled extracorporeal system and attendant anesthesia/sedation protocol were developed and tested. In Stage 1, animals were instrumented for brain, temporalis, and rectal temperature monitoring, and the system was tested for its effect on temperature and hemodynamics. In Stage 2, animals were instrumented for rectal temperature only, subjected to global brain ischemia by two-vessel occlusion and hypotension for 8 minutes, and given either sham therapy (37°C) or hypothermia (32°C) for 4 hours. Viable CA1 neurons were counted at 7 days.

RESULTS

The system was well tolerated, provided exquisite control of animal core and brain temperatures, and conferred robust neuroprotection at 7 days. The median and interquartile ranges (IQRs) of viable neurons per 300-μm field were 130 (IQR = 128 to 135) for sham control, 19 (IQR = 15 to 30) for untreated ischemic animals, and 101 (IQR = 94 to 113) for ischemic animals treated with hypothermia (p < 0.05 for comparison between all groups).

CONCLUSIONS

Like human protocols, this model incorporates sedation and analgesia, results in robust neuroprotection, is well tolerated, and offers exquisite temperature control. The system is noninvasive and inexpensive and offers a model that is similar to methods used in clinical practice. This system will be of interest to investigators using small animal models to examine neuroprotective mechanisms of hypothermia and translational strategies that combine hypothermia with targeted pharmacotherapy.

Unilateral brain hypothermia as a method to examine efficacy and mechanisms of neuroprotection against global ischemia

Hypothermia, especially applied during ischemia, is the gold-standard neuroprotectant. When delayed, cooling must often be maintained for a day or more to achieve robust, permanent protection. Most animal and clinical studies use whole-body cooling-an arduous technique that can cause systemic complications. Brain-selective cooling may avoid such problems. Thus, in this rat study, we used a method that cools one hemisphere without affecting the contralateral side or the body. Localized brain hypothermia was achieved by flushing cold water through a metal tube attached to the rats' skull. First, in anesthetized rats we measured temperature in the cooled and contralateral hemisphere to demonstrate selective unilateral cooling. Subsequent telemetry recordings in awake rats confirmed that brain cooling did not cause systemic hypothermia during prolonged treatment. Additionally, we subjected rats to transient global ischemia and after recovering from anesthesia they remained at normothermia or had their right hemisphere cooled for 2 days (∼32°C-33°C). Hypothermia significantly lessened CA1 injury and microglia activation on the right side at 1 and 4 week survival times. Near-complete injury and a strong microglia response occurred in the left (normothermic) hippocampus as occurred in both hippocampi of the untreated group. Thus, this focal cooling method is suitable for evaluating the efficacy and mechanisms of hypothermic neuroprotection in global ischemia models. This method also has advantages over many current systemic cooling protocols in rodents, namely: (1) lower cost, (2) simplicity, (3) safety and suitability for long-term cooling, and (4) an internal control-the normothermic hemisphere.

α-Synuclein levels are elevated in cerebrospinal fluid following traumatic brain injury in infants and children: the effect of therapeutic hypothermia

α-Synuclein is one of the most abundant proteins in presynaptic terminals. Normal expression of α-synuclein is essential for neuronal survival and it prevents the initiation of apoptosis in neurons through covalent cross-linking of cytochrome c released from mitochondria. Exocytosis of α-synuclein occurs with neuronal mitochondrial dysfunction, making its detection in cerebrospinal fluid (CSF) of children after severe traumatic brain injury (TBI) a potentially important marker of injury. Experimental therapeutic hypothermia (TH) improves mitochondrial function and attenuates cell death, and therefore may also affect CSF α-synuclein concentrations. We assessed α-synuclein levels in CSF of 47 infants and children with severe TBI using a commercial ELISA for detection of monomeric protein. 23 patients were randomized to TH based on published protocols where cooling (32-33°C) was initiated within 6-24 h, maintained for 48 h, and then followed by slow rewarming. CSF samples were obtained continuously via an intraventricular catheter for 6 days after TBI. Control CSF (n = 9) was sampled from children receiving lumbar puncture for CSF analysis of infection that was proven negative. Associations of initial Glasgow Coma Scale (GCS) score, age, gender, treatment, mechanism of injury and Glasgow Outcome Scale (GOS) score with CSF α-synuclein were compared by multivariate regression analysis. CSF α-synuclein levels were elevated in TBI patients compared to controls (p = 0.0093), with a temporal profile showing an early, approximately 5-fold increase on days 1-3 followed by a delayed, >10-fold increase on days 4-6 versus control. α-Synuclein levels were higher in patients treated with normothermia versus hypothermia (p = 0.0033), in patients aged <4 years versus ≥4 years (p < 0.0001), in females versus males (p = 0.0007), in nonaccidental TBI versus accidental TBI victims (p = 0.0003), and in patients with global versus focal injury on computed tomography of the brain (p = 0.046). Comparisons of CSF α-synuclein levels with initial GCS and GOS scores were not statistically significant. Further studies are needed to evaluate the conformational status of α-synuclein in CSF, and whether TH affects α-synuclein aggregation.

Protection in animal models of brain and spinal cord injury with mild to moderate hypothermia

For the past 20 years, various laboratories throughout the world have shown that mild to moderate levels of hypothermia lead to neuroprotection and improved functional outcome in various models of brain and spinal cord injury (SCI). Although the potential neuroprotective effects of profound hypothermia during and following central nervous system (CNS) injury have long been recognized, more recent studies have described clinically feasible strategies for protecting the brain and spinal cord using hypothermia following a variety of CNS insults. In some cases, only a one or two degree decrease in brain or core temperature can be effective in protecting the CNS from injury. Alternatively, raising brain temperature only a couple of degrees above normothermia levels worsens outcome in a variety of injury models. Based on these data, resurgence has occurred in the potential use of therapeutic hypothermia in experimental and clinical settings. The study of therapeutic hypothermia is now an international area of investigation with scientists and clinicians from every part of the world contributing to this important, promising therapeutic intervention. This paper reviews the experimental data obtained in animal models of brain and SCI demonstrating the benefits of mild to moderate hypothermia. These studies have provided critical data for the translation of this therapy to the clinical arena. The mechanisms underlying the beneficial effects of mild hypothermia are also summarized.

Therapeutic hypothermia in experimental models of focal and global cerebral ischemia and intracerebral hemorrhage

Experimental evidence shows that therapeutic hypothermia (TH) protects the brain from cerebral injury in multiple ways. In different models of focal and global cerebral ischemia, mild-to-moderate hypothermia reduces mortality and neuronal injury and improves neurological outcome. In models of experimental intracerebral hemorrhage (ICH), TH reduces edema formation but does not show consistent benefi cial effects on functional outcome parameters. However, the number of studies of hypothermia on ICH is still limited. TH is most effective when applied before or during the ischemic event, and its neuroprotective properties vary according to species, strains and the model of ischemia used. Intrinsic changes in body and brain temperature frequently occur in experimental models of focal and global cerebral ischemia, and may have infl uenced studies on other neuroprotectants. This might be one explanation for the failure of a large amount of translational clinical neuroprotective trials. Hypothermia is the only neuroprotective therapeutic agent for cerebral ischemia that has successfully managed the transfer from bench to bedside, and it is an approved therapy for patients after cardiac arrest and children with hypoxic-ischemic encephalopathy. However, the implementation of hypothermia in the treatment of stroke patients is still far from routine clinical practice. In this article, the authors describe the development of TH in different models of focal and global cerebral ischemia, point out why hypothermia is so efficient in experimental cerebral ischemia, explain why temperature regulation is essential for further neuroprotective studies and discuss why TH for acute ischemic stroke still remains a promising but controversial therapeutic option.

Bioenergetics of cerebral ischemia: a cellular perspective

In cerebral ischemia survival of neurons, astrocytes, oligodendrocytes and endothelial cells is threatened during energy deprivation and/or following re-supply of oxygen and glucose. After a brief summary of characteristics of different cells types, emphasizing the dependence of all on oxidative metabolism, the bioenergetics of focal and global ischemia is discussed, distinguishing between events during energy deprivation and subsequent recovery attempt after re-circulation. Gray and white matter ischemia are described separately, and distinctions are made between mature and immature brains. Next comes a description of bioenergetics in individual cell types in culture during oxygen/glucose deprivation or exposure to metabolic inhibitors and following re-establishment of normal aerated conditions. Due to their expression of NMDA and non-NMDA receptors neurons and oligodendrocytes are exquisitely sensitive to excitotoxicity by glutamate, which reaches high extracellular concentrations in ischemic brain for several reasons, including failing astrocytic uptake. Excitotoxicity kills brain cells by energetic exhaustion (due to Na(+) extrusion after channel-mediated entry) combined with mitochondrial Ca(2+)-mediated injury and formation of reactive oxygen species. Many (but not all) astrocytes survive energy deprivation for extended periods, but after return to aerated conditions they are vulnerable to mitochondrial damage by cytoplasmic/mitochondrial Ca(2+) overload and to NAD(+) deficiency. Ca(2+) overload is established by reversal of Na(+)/Ca(2+) exchangers following Na(+) accumulation during Na(+)-K(+)-Cl(-) cotransporter stimulation or pH regulation, compensating for excessive acid production. NAD(+) deficiency inhibits glycolysis and eventually oxidative metabolism, secondary to poly(ADP-ribose)polymerase (PARP) activity following DNA damage. Hyperglycemia can be beneficial for neurons but increases astrocytic death due to enhanced acidosis.

Input-output relations in the entorhinal cortex-dentate-hippocampal system: evidence for a non-linear transfer of signals

In the current study we analyzed the input-output relations in the entorhinal-dentate-hippocampal system, a major network involved in long-term memory. In anesthetized guinea pigs, the system was driven by activation of perforant path neurons in the entorhinal cortex (ENT), via presubicular fibers directly stimulated in the dorsal psalterium. Perforant path neuron discharge activated in parallel the dentate gyrus (DG) and hippocampal field CA2. Whereas the output from the DG activated hippocampal field CA3, the output from the sole field CA2 was sufficient for activation of field CA1. Signals from field CA3 operated in concert with CA2, likely contributing to discharge field CA1. These findings indicate the existence of two in parallel disynaptic systems: an ENT-CA2-CA1 and an ENT-DG-CA3 system. The convergence of the latter with the former gives origin the classical trisynaptic circuit, the ENT-DG-CA3-CA1 system. The input-output relations between the population excitatory postsynaptic potentials (pEPSP) evoked in the DG, CA3, CA2 and CA1 and the population spike (PS) evoked in the structure upstream (the input) were described by smooth sigmoid curves. In contrast, the input-output relations of the PS versus the pEPSP within each structure were described by steep sigmoid curves. The net input-output functions of the DG (ENT-DG system), field CA2 (ENT-CA2 system), field CA3 (ENT-DG-CA3 system) and field CA1 (ENT-CA2-CA1&ENT-DG-CA3-CA1 system) were described by sigmoid curves. While the DG and field CA2 exhibited steep sigmoids, fields CA3 and CA1 had less steep sigmoid functions. The present study demonstrates that all structures downstream to the ENT operate according to sigmoid input-output functions, characterized by specific parameters. These different behaviors may contribute to different memory processes. We additionally demonstrate that field CA1 can be activated by field CA2, independently from field CA3. This functional dissociation between CA3 and CA1 may subserve specific roles of each field in memory encoding/retrieval.

Astrocytic contributions to bioenergetics of cerebral ischemia

Astrocytes are multifunctional cells that interact with neurons and other astrocytes in signaling and metabolic functions, and their resistance to pathophysiological conditions can help restrict loss of tissue after an ischemic event provided adequate nutrients are supplied to support their requirements. Astrocytes have substantial oxidative capacity and mechanisms to upregulate glycolytic capability when respiration is impaired. An astrocytic enzyme that synthesizes a powerful activator of glycolysis is not present in neurons, endowing astrocytes with the ability to sustain ATP production under restrictive conditions. The monocarboxylic acid transporter (MCT) isoforms predominating in astrocytes are optimized to facilitate very large increases in lactate flux as lactate concentration increases within (1-3 mM) and above (>3 mM) the normal range. In sharp contrast, the major neuronal MCT serves as a barrier to increased transmembrane transport as lactate rises above 1 mM, restricting both entry and efflux. Lactate can serve as fuel during recovery from ischemia but direct evidence that lactate is oxidized by neurons (vs. astrocytes) to maintain synaptic function is lacking. Astrocytes have critical roles in regulation of ionic homeostasis and control of extracellular glutamate levels, and spreading depression associated with ischemia places high demands on energy supplies in astrocytes and contributes to metabolic exhaustion and demise. Disruption of Ca2+ homeostasis, generation of oxygen free radicals and nitric oxide, and mitochondrial depolarization contribute to astrocyte death during and after a metabolic insult. Novel pharmaceutical agents targeted to astrocytes and hyperoxic therapy that restores penumbral oxygen level during energy failure might improve postischemic outcome.

Long-term effects of clomethiazole in a model of global ischemia

The failure of neuroprotective drugs in clinical trials has raised questions about the predictive value of animal models. To address this issue we reexamined the efficacy of clomethiazole using functional and histological outcome measures in combination with long-term survival times. Gerbils were exposed to 5 min of global ischemia and received 400 mg/ml clomethiazole (via osmotic minipump) plus a bolus injection (60 mg/kg) 30 min after ischemia. Brain temperature was maintained at approximately 36.5 degrees C during ischemia and for the first 30 min after ischemia, and was monitored in all groups for 24 h. Subgroups of clomethiazole-treated gerbils had their temperatures regulated in the normothermic range while in other animals temperature was not controlled. Open-field habituation tests were conducted 5, 10, 30, and 60 days after occlusion. CA1 cell counts and CA1 slice recordings were done at the conclusion of behavioral testing. Clomethiazole significantly attenuated CA1 cell loss at 10-, 30-, and 60-day survival. A modest reduction in habituation deficits was evident only on Day 10 (P < 0.05). Similarly, field potential amplitude was not maintained in the rostral CA1 region. Clomethiazole produced mild hypothermia that developed over several hours. Based on short-term CA1 cell counts, clomethiazole provided significant histological protection with limited functional preservation. Neuroprotection disappeared when longer survival times (60 day) were employed and temperature confounds eliminated. These data demonstrate the necessity of utilizing more clinically relevant survival times and carefully monitoring/regulating postischemic temperature when assessing potential neuroprotective compounds.

Delayed transient ischemic attacks kill some CA1 neurons previously salvaged with postischemic hypothermia: neuroprotection undone

Delayed hypothermia reduces ischemic hippocampal CA1 injury. However, there are residual structural and functional abnormalities. Therefore, we studied whether these apparently vulnerable rescued neurons are susceptible to secondary insults. All gerbils were subjected to normothermic forebrain ischemia (ISC, 5 min) or SHAM operation. Gerbils were treated with mild hypothermia (HYPO; 33 degrees C for 24 h+35 degrees C for 24 h) beginning 12 h after surgery, or they remained normothermic (NORMO). Then 5 and 6 days following ISC/SHAM operation gerbils received sublethal transient ischemic attacks (TIA, 1.5 min) or sham (SH) surgeries. Behavioral testing was done and animals survived for 30 days for quantification of medial, middle and lateral CA1 sector cell death. The SHAM groups were not significantly different. The ISC+NORMO+SH group lost 87.3% (of SHAM) of medial CA1 neurons, which was not significantly exacerbated in the ISC+NORMO+TIA group (91.1%, P=0.633). However, the ISC+HYPO+TIA group (58.8% loss) had significantly more cell death than the ISC+HYPO+SH group (42.8%; P=0.035), although CA1 protection was still better than in ISC+NORMO groups (P<0.001). Trends were similar in middle and lateral CA1, but the deleterious effects of TIAs were not statistically significant. Behavioral testing did not distinguish groups with or without TIA, but did reveal deficits in ISC+NORMO groups and protection in ISC+HYPO groups. These data, like previous ultrastructural findings, show that while most hypothermia-rescued CA1 neurons are healthy, some are susceptible. Perhaps other neuroprotectants, especially weaker ones, might be undone by delayed insults (e.g. TIA, fever).

Effects of hypothermia on neuronal-vascular function after cerebral ischemia in piglets

We determined whether cerebral arteriolar dilation to N-methyl-d-aspartate (NMDA), a response dependent on stimulation of cortical neurons and inhibited by anoxic stress, would be preserved by hypothermia during and following ischemia. Pial arteriolar diameters in anesthetized piglets were determined via intravital microscopy. Arteriolar responses to NMDA (10, 50, and 100 micromol/l) were measured before and 1 h after 10 min of global ischemia. Piglets were exposed to either total body or selective brain cooling (33-34 degrees C). Arteriolar dilation to lower doses or to 100 micromol/l NMDA was not affected by hypothermia alone (51 +/- 3 vs. 46 +/- 7%, normothermia vs. hypothermia; n = 7) in nonischemic animals. However, arteriolar responses to 100 micromol/l NMDA were clearly attenuated after ischemia despite body cooling during ischemia (53 +/- 3 vs. 32 +/- 6%; n = 8), hypothermia during ischemia and early reperfusion (49 +/- 10 vs. 20 +/- 3%; n = 8), or selective brain cooling (48 +/- 5 vs. 20 +/- 5%; n = 10). In contrast, pretreatment with indomethacin resulted in complete preservation of NMDA-induced vasodilation after ischemia. Thus, hypothermia fails to protect against neuronal dysfunction during ischemia.