Neuroprotective and Neurotoxic Properties of General Anesthetics: A Double-Edged Sword

This is an open-access article, published by Evidence Based Communications (EBC). This work is licensed under the Creative Commons Attribution 4.0 International License, which permits unrestricted use, distribution, and reproduction in any medium or format for any lawful purpose.To view a copy of this license, visit http://creativecommons.org/licenses/by/4.0/. From Department of Anaesthesiology and Critical Care, University of Pennsylvania, Philadelphia, USA.

G eneral anesthetics, especially volatile anesthetics, are commonly used to maintain general anesthesia for surgeries in patients across a spectrum of ages. Preconditioning and postconditioning with general anesthetics have been shown to provide neuroprotection against reperfusion injury and hypoxia (1,2). However, there is growing apprehension that general anesthetics may cause permanent cognitive dysfunction following peri-and post-operative periods. Particularly, the neonatal and elderly populations seem most vulnerable to the neurotoxic risk of anes-thetics. Results from our laboratory and others suggested that the dual effects of general anesthetics depend on the concentration and the duration of anesthetic exposure (3)(4)(5).
Determining the neurotoxicity of general anesthetics in humans from the results of cellular and animal studies poses several challenges. Firstly, the timing of neuro-developmental stages varies among different animal species. For example, while periods of rapid human brain development occur from the third trimester through age two, neuro-anatomical and neuro-physiologic evidence suggests that similar development occurs in rodents from birth to postnatal day 14 (6,7). Hence, it is difficult to infer specific intervals of patient susceptibility. Secondly, the lack of surgical stimuli, inflammatory responses, and comparable diseases within many animal models complicates their adaptation to humans. Because most preclinical studies have focused on cellular and animal studies, research evaluating general anesthetic neurotoxicity in humans is sparse. Knowing how much the exposure time of anesthetics should be adjusted from animals to humans is one of the questions of animal-human translational research. Therefore, an understanding of the mechanisms behind general anesthetics'dual effects can only help to optimize neuroprotection and minimize neurotoxicity in anesthetized patients.

General Anesthetic-Induced Neuroprotection
Pre-and postconditioning with anesthetics are utilized for neuroprotection. A brief subclinical application of anesthetics just before or after neurotoxic stimuli increases tolerance and decreases neuronal cell damage. General anesthetics have been shown to be protective against cell death induced by glutamate, hypoxia, ischemia, and oxygen-glucose deprivation (OGD) in cell culture and animal studies. For example, use of isoflurane, sevoflurane, and halothane in clinicallyused concentrations reduces lactate dehydrogenase (LDH) release, a marker for tissue damage (8). Short-term administration of isoflurane also decreases apoptosis in OGD -exposed rat hippocampal slices (9) and increases the survival rate of Purkinje neurons exposed to glutamate (10). The possible mechanisms of general anestheticinduced neuroprotection are discussed below.

Role of Calcium and Cell Survival Signaling Pathways
Previous research indicated that intracellular calcium (Ca 2 + ) concentration is a defining factor for the neuroprotective properties of general anesthetics. As an important second messenger in cell growth and differentiation, calcium is maintained at 100 nM between the intra-and extracellular space under normal conditions (11)(12)(13)(14). Sarcoplasmic reticulum Ca 2 + -ATPase (SERCA) pumps, inositol-1,4,5-triphosphate receptors (In-sP3R), ryanodine receptors (RYRs), and Ca 2 +binding proteins all regulate Ca 2 + within the endoplasmic reticulum (ER). When RYRs and In-sP3, a component of the phospholipase C pathway (15), activate InsP3Rs, Ca 2 + is released from the ER.
Besides the ER, the mitochondria are another significant organelle involved in the regulation of intracellular Ca 2 + concentration. The mitochondrial permeability transition pores (mPTP) and the mitochondrial ATP-sensitive K (mito-KATP) channels both play crucial roles in Ca 2 + regulation and apoptosis. Overall activation of mPTP increases mitochondrial permeability inside the mitochondria and results in mitochondrial swelling, rupture, and the release of cytochrome c (CytC), an activator of pro-apoptotic factors. To achieve mitochondrial protection, the mitoKATP channel stays open during ischemia and reperfusion to conserve the mitochondrial ATP production rate and to mitigate Ca 2 + overrelease (16).
A moderate increase in Ca 2 + concentration is thought to be the mechanism of hypoxia and isoflurane preconditioning-induced cytoprotection. Gray et al. (17) reported that slight increases in Ca 2 + concentration in the presence of isoflurane stimulated the mitogen-activated protein kinase (MAPK) signaling cascade, a cell-survival pathway. Likewise, Bickler and Fahlman (18) compared the expression of signal transduction genes following hypoxic or isoflurane preconditioning in rat hippocampal slices. In both cases, preconditioning initiated an increase in intracellular Ca 2 + levels and established similar cell survival rates. When hypoxic and isoflurane preconditionings were applied together, the neuroprotective effect was higher than when they were administered alone and neuronal death decreased following the combination pretreatment (19).
Researchers have also examined the effect of postconditioning with general anesthetics. Zhou et al. (20) examined the effect of isoflurane postconditioning on neonatal hypoxic-ischemic brains in rats. These authors reported reduced infarct size at 48 hours and attenuated brain atrophy at four weeks. Pretreating the rats with the sphingosine-1-phosphate (S1P) antagonist before the postconditioning treatment antago- nized isoflurane-induced neuroprotection. Therefore, the data indicate that isoflurane-induced neuroprotection occurs via activation of S1P/PI3K/Akt through modest increases in cytosolic Ca 2+ concentration.
Overall, these studies demonstrated that preand postconditioning with general anesthetics moderate the Ca 2 + level within cells and activate cell survival signal pathways in anesthetic-induced neuroprotection.

Role of Receptor Interactions
Volatile anesthetics mainly act on ligand-gated ion channels and two-pore domain K + channels. The three ligand-gated ion channels typically targeted by most anesthetics are the γ-aminobutyric acid type A (GABAA) receptor, the inhibitory glycine receptor, and the N-methyl-D -aspartate (NMDA) receptor (21)(22)(23)(24). With the exception of xenon and cyclopropane, almost all anesthetics potentiate GABA-induced Clcurrents and serve as agonists on GABAA receptors (21). The antagonistic properties of general anesthetics on NMDA receptors are thought to be responsible for the analgesic, neuroprotective, and immobilizing effects of anesthetics (8,10,25).
The activation of inhibitory synaptic K + current by anesthetics has also been well researched. The three types of K + channels are the voltage-gated (Kv), inwardly rectifying (Kir), and two-pore (K2p) domain channels (26)(27)(28)(29). KATP is one of the members of Kir and, as previously mentioned, plays an essential role in Ca 2 + regulation and cell death. The K + channels TREK-1 and TASK are two of the most important members of the K2p family. While chloroform, diethyl ether, halothane and isoflurane activate TREK-1, only halothane and isoflurane stimulate TASK (30). The activation of GABA receptors and inhibitory K + channels and the inhibition of NMDA receptors by anesthetics cause hyperpolarization and reduced glutamate release. As a result, these receptor interactions have an absolute effect on decreasing excitotoxicity during ischemia, hypoxia, and reperfusion.
Another possible mechanism for the neuroprotective effects of volatile anesthetics is the enhancement or preservation of excitatory amino acid transporters (EAAT), especially that of Type 3 (EAAT3). EAATs are responsible for glutamate re-uptake from the extracellular space into cells. The major glutamate transporter in neurons is EAAT3. Previous studies have shown that volatile anesthetics enhance EAAT3 activity and gene expression in cells via protein kinase C (PKC) (31,32). EAAT3 also transports cysteine, a rate-limiting substrate for the synthesis of a major intracellular antioxidant, glutathione. Lee et al. (33) reported that volatile anesthetics attenuate the tert-Butyl hydroperoxide-induced decrease in EAAT3 activity, thus reducing the transport L-cysteine and L-glutamate in rat EAAT3 expressed in xenopus oocytes. These authors concluded that volatile anesthetics increase EAAT3 activity and the neuroprotective effect of volatile anesthetics may stem from the preservation of EAAT3 function under oxidative stress.
Both NMDA and GABA receptors take part in memory and learning formation in the hippocampus. Compared to a control group, mice anesthetized with isoflurane presented better cognitive function (34). Up-regulation of NR2B, a subunit of the NMDA receptor, was found in the hippocampus of mice, and pretreatment with NR2B selective antagonists mitigated the improved cognitive function induced by isoflurane. Also, isoflurane enhanced long-term potentiation in the CA1 region of the hippocampus (34).

Role of Nitric Oxide Synthase and Anti-Inflammatory Properties of General Anesthetics
Nitric oxide (NO) is produced endogenously from the amino acid L-arginine by nitric oxide synthase (NOS). Four different isoforms of NOS have been classified: endothelial (eNOS), neuronal (nNOS), inducible (iNOS) and mitochondrial (mtNOS) (35)(36)(37). All forms of NOS are expressed in the brain (38,39). The authors concluded that eNOS, nNOS, and mtNOS activities are constitutive and calcium-and calmodulin-dependent. However, iNOS activity is independent of calcium or calmodulin; instead, cytokines, infection, inflammation, or cardiac stress induce its response.
The effects of NO are highly dependent upon concentration. At low levels, NO can act as an intracellular and intercellular regulatory messenger. At high levels, NO can be cytotoxic by producing free oxygen radicals, activating post-translational peroxynitrite and inducing apoptosis (14,39). The isoform of NOS that produces NO is functionally important. For example, both overproduction of NO by iNOS and hypoproduction of NO by mtNOS are involved in the pathogenesis of aging (39). Different sites of activity may be important in determining whether NO effects are neuroprotective or neurotoxic.
Isoflurane or halothane preconditioning provided prolonged neuroprotection against ischemia in rats induced by middle cerebral artery occlusion (MCAO) and in primary cortical neurons induced by OGD (40). The preconditioning increased iNOS protein levels and the administration of an iNOS inhibitor prevented anesthetic-induced infarct volume reduction in rats. In OGD -exposed hippocampal cells, preconditioning with isoflurane increased iNOS mRNA levels (41). Isoflurane preconditioning in neonatal rats also induced neuroprotection in a perinatal stroke model (42). Isoflurane increased iNOS levels and the pretreatment with an iNOS inhibitor antagonized isoflurane-induced neuroprotection. Thirty minutes of inhalation of 2.8% isoflurane 24 hours before spinal cord ischemia in rats significantly increased iNOS expression in the spinal cord motor neurons (43). Also, isoflurane induced ischemic preconditioning in ischemic rat hearts by increasing expression and activity of iNOS (44).
Inflammation is common in central nervous system (CNS) diseases involving brain trauma, stroke, brain infection, and neurodegeneration. Microglial cells play a critical role in the brain inflammatory response. During neuroinflammation, microglial cells are activated and pro-inflammatory cytokines and reactive oxygen species are produced. Xu et al. (45) tested how isoflurane preconditioning affects microglial function in mouse microglial cultures. One hour exposure of isoflurane before application of lipopolysaccharide (LPS) and interferonγ (IFNγ) to microglial cells decreased cell death, iNOS expression, and glutamate release. These results indicate that isoflurane reduces microglial activity induced by LPS and IFNγ and consequently inhibits the iNOS-NO-glutamate pathway activated by microglia. Anti-inflammatory activity was not only reported with isoflurane but also with desflurane (46) and sevoflurane (47).

Differential Effects of Inhalational Anesthetics in Neuroprotection
Inhalational anesthetics protect the brain during ischemia and hypoxia by reducing cerebral metabolic rate, cortical electrical activity, cerebral oxygen consumption, and lactate accumulation through sustaining ATP stores and phosphocreatine depletion (48,49). Positron emission tomography studies in humans demonstrate that inhalational anesthetics globally and regionally diminish cerebral metabolism (50,51). Among the inhalational anesthetics, isoflurane provides better cerebral protection than halothane and enflurane. Michenfelder et al. (52) investigated the records of 2,223 patients who underwent carotid endarterectomy at the Mayo Clinic between 1972 and 1985 to compare the effects of isoflurane, halothane, and enflurane on cerebral blood flow (CBF), incidence of electro encephalographic (EEG) ischemic changes, and neurologic outcome. CBF was lower in the isoflurane group, compared to the halothane and enflurane groups. There was no difference among groups in neurologic outcome, but ischemic EEG changes were significantly less in the isoflurane group than those in the halothane and enflurane groups. Another study reported that desflurane lowered plasma endothelin levels after anesthesia in patients undergoing intracranial aneurysm clipping (54). This effect of desflurane may contribute to the prevention of unwanted acute cerebral vasospasm.

General Anesthetic-Induced Neurotoxicity
Since most anesthetics are NMDA antagonists and GABA A agonists, both of these receptors have been extensively studied in neurotoxic research. It has been shown that the blockage of NMDA receptors and stimulation of GABAA receptors caused apoptotic neurodegeneration in animals when given in the early neonatal period (6,(54)(55)(56)(57). Ethanol, an NMDA receptor antagonist and a GABAA receptor agonist, induced apoptotic neurodegeneration in animals (6,56). Fetal alcohol syndrome, a consequence of intrauterine exposure of the fetus to ethanol in humans, causes neurobehavioral abnormalities ranging from learning disabilities to psychosis, mirroring the effect of NMDA and GABAA re- ceptors. Other factors involved in anesthetic-induced neurotoxicity are discussed below.

Role of Calcium
Abnormal calcium release from the ER via either InsP3R and/or RYR plays important role in anesthesia-mediated neurotoxicity. Wei et al. (58) used PC 12 pheochromocytoma cells and primary cortical neurons to compare the cytotoxic effects of equipotent concentrations of isoflurane and sevoflurane. The application of isoflurane (2.4% ) for 24 hours increased LDH release but the application of sevoflurane (4% ) for the same timeframe had no effect on both cell lines. Isoflurane increased the pro-apoptotic activities of caspase 3 and 9, but sevoflurane affected neither of the signaling proteins. The authors hypothesized that isoflurane induces apoptosis by increasing calcium release from the ER through the activation of RYRs. Cells that had a pretreatment of dantrolene, an antagonist of RYR, exhibited significantly decreased isoflurane-induced apoptosis. Eckenhoff et al. (59) demonstrated that both halothane and isoflurane enhance amyloid β (Aβ) oligomerization and Aβ-induced cell toxicity in rat pheochromocytoma cells. Aβ is produced from amyloid β precursor protein (APP) by aspartyl protease β-site APP-cleaving enzyme (BACE) or β-secretase and γ-secretase (60). Similar studies showed that BACE and γ-secretase levels were increased in primary cortical and hippocampal neurons and in H4 human neuroglioma cells following treatments of isoflurane and sevoflurane (61)(62)(63)(64). In these studies, when the isoflurane-induced increase in cytosolic calcium level was reduced using calcium chelators, the isoflurane-induced cytotoxic effects were attenuated. For example, memantine, a NMDA partial antagonist and USA Food and Drug Administration (FDA)-approved Alzheimer's disease (AD) medication, prevented isoflurane-elicited caspase-3 activity and apoptosis in cell lines (64).
While all inhalational anesthetics have the capacity to induce cell damage, the three commonly used inhalational anesthetics do not appear to have the same potency. Our results showed that isoflurane is more potent than sevoflurane or desflurane in causing cell damage in the rodent developing brain (65). Other studies confirmed that isoflurane seems to be more potent in causing calcium release from the ER and cell death, as well as huntingtin protein aggregation, than equipotent concentrations of sevoflurane or desflurane (61,(65)(66)(67)(68). It should be noted that isoflurane is also more potent than sevoflurane in providing neuroprotection via preconditioning mechanisms.

Effect of Age on Neuroprotection and Neurotoxicity
Various studies showed that neurologic outcome after inhalational anesthetic exposure depends on age of the subject. In one study, maternal exposure of isoflurane during gestational day 21 of rats had no marked effect on fetal brains (4). Apoptosis was similar in both exposed and control fetal brains and spatial memory and learning were not impaired. However, when isoflurane concentration was increased to 3% (1 hour) in gestational day 21, caspase-3 positive cells in both the hippocampal CA1 region and the retrosplenial cortex of neonatal rats were significantly increased (5). In another study (69), rats were exposed to isoflurane during gestational day 14 and studied behaviorally at the age of 8 weeks. Deficits in spatial memory and anxiety were found in rats receiving isoflurane. We conclude that isoflurane may inhibit or promote apoptosis in fetal brains, depending not only on the time of exposure but also the concentration used.
Combined midazolam, nitrous oxide and isoflurane exposure to 7-day-old rats led to apoptotic neurodegeneration, synaptic dysfunction, and long-term impairment in spatial learning and memory (70). In 14-day-old rats, similar results were not found, suggesting that the developing brain is more vulnerable during the peak time of synaptogenesis (71). Exposure of 7-dayold mice to only isoflurane did not influence locomotion, learning, or memory at 2 months. Yet, isoflurane (0.75% ) combined with nitrous oxide (6 hours) increased caspase-3 activity in spinal cords of 7-day-old rats (72). Likewise, a subclinical concentration of sevoflurane increased caspase-3 activity in 7-day-old mice (73,74). When the concentration is increased, sevoflurane not only increased apoptosis but also caused persistent learning and social interaction  (76) found that apoptotic and neuronal cell death activities of equipotent concentrations of isoflurane, sevoflurane, and desflurane were similar in 7-day-old mice. Low concentrations of isoflurane combined with nitrous oxide elicited apoptosis (70)(71)(72). On the other hand, higher concentrations of isoflurane were needed when given with just oxygen in the 7-8 days old rodents (5,76). It seems that the combination of administered anesthetics influences neurotoxicity. Again, the peak of synaptogenesis seems to be the most vulnerable time for anesthetics exposure.
Stratmann et al. (77) investigated the effects of 1 minimum alveolar concentration (MAC) isoflurane exposure on neurogenesis and neurocognitive function in 7-day-and 60-day-old rats. In 60-day-old rats, isoflurane increased neuronal differentiation and progenitor proliferation and improved neurocognitive functions. However, in newborn rats, isoflurane decreased progenitor proliferation and caused persistent and progressive neurocognitive deficit. Thus, the authors suggested that anesthetic-induced cognitive dysfunction seems to occur during specific periods of brain development, as the 60-day old rats, although young, displayed limited impairment. Similarly, repeated isoflurane exposure (1.7% solution applied for 35 minutes over 4 consecutive days) was studied in PN14 rats and mice as well as PN60 rats (78). Persistent and progressive learning and memory deficits were detected in the younger group but not the older group. Undifferentiated hippocampal neural stem cells were reduced in the PN14 rats, indicating that young brains are more vulnerable to isoflurane. In aged rats, isoflurane-nitrous oxide anesthesia has been shown to cause long-term spatial memory impairment (79)(80)(81). These results showed a trend of newborn and aged rats exhibiting the greatest effects of neurotoxicity.
Despite evidence of neurodegeneration induced by anesthetics in cell culture and neonatal animal brain slices, human studies remain limited. The lack of established biomarkers for human neurodegeneration complicates proper analysis. Further, human studies would require decades of follow-up in very large cohorts of patients to evaluate behavioral outcome after in utero or neona-tal anesthesia exposure. Kalkman et al. (82) calculated that a sample size of 2,268 children would be needed to perform a large-scale, retrospective study to investigate disturbed neurobehavioral development in children exposed to anesthetics at ages 0-6 years and reach a statistically valid conclusion. Although it was not statistically significant, their pilot study showed that children undergoing surgery before month 24 had more behavioral disturbance than children having surgery after age two. Learning disabilities were detected in children receiving multiple anesthetic exposures prior to age four, but not in children exposed to a single anesthetic exposure (83).
In elderly humans, long-term cognitive problems can occur following anesthesia and previous surgeries may increase the risk of neurodegenerative diseases (84,85). As a result, postoperative delirium and cognitive decline (POCD) are important issues in the geriatric population. POCD is characterized by disturbances in memory, concentration, language comprehension, and social interactions. It can persist for weeks and months after surgery, and may even become permanent. Besides advanced age, other risk factors for POCD are limited education, diabetes, atherosclerotic heart disease, hypotension, hypoxia, and cerebral emboli during surgery (86)(87)(88). Delirium mostly occurs on postoperative days one and three and can be associated with disorientation, temporary memory dysfunction, and sleepwake cycle alterations. POCD occurs in 10-26% of patients over 65 years of age and significantly increases morbidity and mortality. The interaction of anesthetics with the central muscarinic cholinergic system is thought to be the underlying mechanism of POCD since the system's reduced activity has been found in the pathogenesis of diseases involving mood and memory (89,90).
Cognitive function decline in patients examined at 7 days and at 3 months postoperatively were similar after both regional and general anesthesia (91). For patients at day 3 only, POCD incidence was higher in the general anesthesia group than in the regional anesthesia group (92). Following coronary artery bypass graft surgery, neurocognitive function was better in patients who received isoflurane compared to patients who received desflurane and sevoflurane. The incidence of POCD was similar in elderly patients  The cytosolic concentration of calcium can be increased either by calcium influx through the voltage-dependent calcium channel (VDCC) or the N-methyl-D-aspartate (NMDA) receptor or by calcium release from the endoplasmic reticulum (ER) through the 1,4,5-trisphosphate receptor (InsP3R) or ryanodine receptor (RYR). Short-term exposure to low-concentrations of a general anesthetic like isoflurane induces a moderate cytosolic calcium increase. An adequate calcium elevation acts as a pre-or postconditioning mechanism and promotes pro-survival pathways like ATP production, mitogen-activated protein kinase (MAPK) signaling cascade activation, and neurogenesis. However, extended exposure to high-concentrations of general anesthesia disrupts calcium homeostasis and increases cytosolic calcium to toxic levels, provoking intrinsic pathway apoptosis, activating pro-apoptotic molecules calpains and inhibiting neurogenesis.
receiving sevoflurane or desflurane for non-cardiac surgery (93). Two years after non-cardiac surgery, POCD incidence among elderly patients was 1% (94). After minor surgery, POCD incidence was 6.8% at 7 days and 6.6% at 3 months (95). Significant risk factors for POCD were determined to be an age greater than 70 years and being an in-patient rather than an out-patient, suggesting that long-term hospitalization should be avoided in elderly patients. The nature and severity of cognitive decline were evaluated at discharge and 3 months after non-cardiac surgery (96). At discharge, 56% of patients had cognitive decline with equal type and severity. But at 3 months, 13% of patients had only memory dysfunction, 8% had only executive function decline, and 9% of patients had both. POCD also occurred in middle-aged patients, but it was resolved in a shorter time period (97). Long-term consequences of POCD were studied for 8.5 years following non-cardiac surgery (98). Mortality rates were higher in patients who had POCD at 3 months than in patients who had POCD at 7 days.
An important question in neurotoxic studies concerns whether the degree of anesthetic-induced neurodegeneration in the developing brain is correlated to the level of anesthetic-induced memory and learning impairments. Although one previous study suggested an association between anesthetic-mediated neurodegeneration and memory and learning disabilities in rodent developing brains (70), our recent studies did not support this statement (61,74), consistent with recent findings from other labs (99). Recent studies suggested that some additional mechanisms other than anesthesia-mediated neurodegeneration may underlie the anesthesia-mediated cognitive dysfunction, such as the reduction of synapses in the developing brain (99). However, further studies are needed to understand the effects of general anesthetics on synapse structures and their contribution to anesthesia-mediated cognitive dysfunction.

Conclusions
Preclinical studies suggested that general anesthetics may be both neuroprotective and neurotoxic, depending on concentration and duration ( Figure 1). Anesthetics used at low concentration for a short time may provide neuroprotection via induction of moderate elevation of cytosolic calcium concentration. However, anesthetics used at a high concentration for prolonged times may cause cell damage by excessive and abnormal elevation of cytosolic calcium concentration. Compounding the problem, specific subsets of the human population are more vulnerable to the effects of general anesthetics than others.
Although human clinical practice typically involves a combination of anesthetics, intubation, and controlled respiration, most animal studies fail to replicate these clinical conditions. Many experimenters use just one type of anesthetic combined with oxygen and permit spontaneous respiration. In cellular studies, it is uncertain if the response of cells treated directly with anesthetics mirrors that of human cells during general anesthesia. Human cells, with circulation and physiological barriers, may need different concentrations of anesthetics than what has been used with cultured cells. Finally, research restrictions imposed on human studies provide yet another problem for data translation. The necessity for long-term follow-up of subjects to study brain development and the difficulty of establishing a homogenous study group make human studies costly. From the studies of both neuroprotection and neurotoxicity in this review, we suggest that general anesthetics act as stress factors, such as ischemia. Short cerebral ischemia or adequate exposure of general anesthetics at low concentrations precondition different neurons of the CNS and provide neuroprotection by inducing endogenous neuroprotective mechanisms. However, prolonged cerebral ischemia or excessive exposure to general anesthetics will cause neuronal damage. Future clinical studies are needed to determine the dividing point of the general anesthetic exposure level at which the effects of general anesthetics will be converted from neuroprotection to neurotoxicity.