Neonatal Desflurane Exposure Induces More Robust Neuroapoptosis than Do Isoflurane and Sevoflurane and Impairs Working Memory

• In animal models, neonatal exposure to volatile anesthetics induces neuroapoptosis, leading to memory deficits in adulthood. However, comparative effects of neonatal exposure to desflurane are largely unknown.

• In a murine model, desflurane was more neurotoxic than sevoflurane or isoflurane in the developing brain.

EXPOSURE of neonatal animals to several classes of drugs induces widespread neuronal death and long-term cognitive function impairments. Although mechanisms are largely unknown, several reports during the past few years have demonstrated that a single administration of isoflurane or sevoflurane at a clinically relevant concentration induces neurodegeneration in the developing brain of a variety of animals ranging from rodents1,–,3to rhesus monkeys.4Little is known, however, about the in vivo  neurotoxicity of the most recently introduced volatile anesthetic, desflurane. Because these three ethers—isoflurane, sevoflurane, and desflurane—are commonly used as inhalation anesthetics in obstetric and pediatric medicine, concerns about the neurotoxicity of these drugs in the developing brain require urgent resolution.

There is a hypothesis that neurodegeneration might be induced through inhibition of N -methyl-D-aspartate (NMDA) receptors and/or activation of γ-aminobutyric acid type A receptors in the developing brain,5,–,7although this view has not been established definitively. Furthermore, it was reported that drugs with activity in both NMDA and γ-aminobutyric acid type A receptors might induce more neurotoxicity than drugs that activate a single receptor type; however, the underlying mechanism is not fully understood.8It is therefore noteworthy that halogenated ethers are reported to exert their effects at both NMDA and γ-aminobutyric acid type A receptors.9,–,12It was reported that activation of γ-aminobutyric acid type A receptors by isoflurane, sevoflurane, and desflurane is qualitatively and quantitatively similar among the three drugs.10However, these halogenated ethers may differ in their effects on NMDA receptors.13Although it is unclear if a drug's ability to inhibit NMDA receptor activity correlates with risk of neuronal apoptosis in the developing brain, it is obviously important to investigate the differences in neurodegenerative potencies of these anesthetics.

In animal models, neonatal exposure to volatile anesthetics reportedly leads to cognitive deficits in adulthood.1,6,14However, whether halogenated ethers produce different effects in developing brains on learning ability in adulthood, and whether these effects correlate with the severity of apoptotic neurodegeneration, are not known. In the current article, we studied the potential risks of a clinically relevant concentration of desflurane compared with other halogenated ethers, isoflurane and sevoflurane, on neurodegeneration in the developing mouse brain. We also studied the comparative effects of these drugs on cognitive function in these mice as adults. Our results suggest that desflurane induces relatively greater neurodegeneration and cognitive deficits than do sevoflurane and isoflurane, and provide the basis for the differential use of halogenated ethers in obstetric and pediatric medicine.

All experiments were conducted according to the institutional ethical guidelines for animal experiments of the National Defense Medical College and were approved by the Committee for Animal Research at National Defense Medical College (Tokorozawa, Saitama, Japan). The C57BL/6 mice used in this study were maintained on a 12-h light–dark cycle (lights on from 07:00 to 19:00) with room temperature at 21°± 1°C. Mice had ad libitum  access to water and food.

On postnatal day 6 (P6), mice were placed in a humid chamber with manipulating gloves and exposed to anesthetics. The total gas flow was 2 l/min, using 25% O²as a carrier. We measured the oxygen and anesthetic agent fractions using a gas analysis system (Capnomac Ultima, GE Healthcare, Tokyo, Japan). During exposure to the anesthetic, the mice were kept warm on a mat heated to 38°± 1°C.

For the caspase-3 experiment, neonatal littermate mice (n = 80) were randomly assigned to receive the following treatments: 25% oxygen for 3 h;, 4% desflurane for 3 h, 8% desflurane for 3 h, 3% sevoflurane for 3 h, 2% isoflurane for 3 h, 25% oxygen for 6 h, 4% desflurane for 6 h, 8% desflurane for 6 h, 3% sevoflurane for 6 h, or 2% isoflurane for 6 h in 25% oxygen. For the terminal deoxynucleotidyl transferase-mediated dUTP nick end-labeling (TUNEL), a separate group of neonatal littermate mice (n = 25) were randomly assigned to the last five treatments. In Western blot analysis, a separate group of neonatal littermate mice (n = 60) were randomly assigned to all treatments. In the behavioral experiment, a separate group of neonatal littermate mice (n = 72) were randomly assigned to 25% oxygen for 6 h and the last three treatments. Mortality during anesthetic exposure for 6 h was 15%, 13%, 18%, and 0% for desflurane, sevoflurane, isoflurane, and 25% oxygen, respectively.

Considering that minimum alveolar concentration (MAC) values in neonate rodents might decrease as a function of anesthesia duration,15we measured MAC values at each time point after induction of anesthetics (1–6 h). Measurements were carried out under the same conditions as the anesthesia treatment study described above. The generally used up-and-down method16with tail-clamp takes more than 15 min as each intertrial interval time to allow for equilibration of brain tissues and inspired gas concentrations, and several concentration trials are needed to determine MAC. We considered that it might be difficult to determine MAC values at a specific time point using this method, because MAC values might be altered as a function of anesthesia duration; therefore, we kept an independent group of mice on fixed, constant concentrations. At each time point (from 1–6 h), supramaximal pain stimuli were generated by application of alligator clamps to the mice's tail. The stimulation was continued for 30 s or until mice moved. Immobility was defined as lack of any purposeful movement except breathing. Measurements were carried out for several concentrations with different groups of mice. After measurements, data were integrated for dose-response curves at specific time points and MAC values were determined. In this method, intertrial interval times for equilibration were not needed and many mice could be examined at once.

Percutaneous oxygen saturation (SpO²) was measured using a SpO²monitoring system (Tosca 500; IMI Co, Saitama, Japan). A monitoring sensor was attached to the tail of each mouse.

Arterial blood sampling from the left cardiac ventricle was performed immediately after removal from the maternal cage (0 h) or at the end of anesthesia (6 h). Samples were analyzed immediately after blood collection using a blood gas analyzer (ABL800; Radiometer, Copenhagen, Denmark).

Apoptosis was evaluated by immunohistochemical staining for activated caspase-3 and by TUNEL as previously described.1Mice were transcardially perfused with 0.1 M phosphate buffer containing 4% paraformaldehyde immediately or 6 h after anesthesia, for activated caspase-3 staining or for TUNEL, respectively. The brains were histologically analyzed using paraffin-embedded sections (5 μm thick). Before analysis, sections were dewaxed in xylene and hydrated using a graded ethanol series. Antigen retrieval was performed by immersing mounted tissue sections in 0.01 mM sodium citrate (pH 6.0) and heating in an autoclave (121°C) for 5 min. Deparaffinized sections were blocked for endogenous peroxidase activity as described previously, followed by blocking with a nonspecific staining blocking reagent (Dako, Glostrup, Denmark) for 1 h to reduce background staining.1The sections were then incubated with antiactive caspase-3 antiserum (Cell Signaling Technology, Beverly, MA) at dilutions of 1:100 in antibody diluent (Dako) overnight in a humidified chamber at 4°C. Subsequently, sections were incubated with peroxidase-conjugated secondary antibody (Dako EnVision+ system; Dako). Immunoreactivity was revealed using 3,3-diaminobenzine-tetrachloride (DAB, Vector Laboratories, Burlingame, CA) according to the manufacturer's instructions. Finally, the sections were counterstained with hematoxylin. An investigator blinded to the treatment conditions counted the number of activated caspase-3 positive (caspase-3⁺) cells.

The TUNEL was performed using an in situ  apoptosis detection kit (ApopTag fluorescein; Chemicon, Temecula, CA) according to the manufacturer's protocol; DAB was used to reveal reactivity. Sections were counterstained with hematoxylin. An investigator blinded to the treatment conditions counted the number of positive cells.

The mouse forebrain was quickly removed and homogenized in four volumes of homogenization buffer containing 50 mM Tris-HCl, pH 7.4, 150 mM NaCl, 1% NP-40, 0.5% sodium deoxycholate, 0.1% sodium dodecyl sulfate, protease inhibitor cocktail (Complete, Roche Diagnostics, Penzberg, Germany), and phosphatase inhibitors (20 mM glycerophosphate, 1 mM Na³VO⁴, 2 mM NaF). After homogenization, a portion of each sample was immediately frozen at −80°C. The remaining homogenate was centrifuged at 15,000g  for 30 min at 4°C. The supernatant solutions were separated and stored at −80°C until use. The amount of protein concentration in each sample was measured using a bicinchoninic acid protein assay kit (Pierce, Rockford, IL).

NIH3T3 mouse fibroblast cells were grown in Dulbecco Modified Eagles Medium supplied with 10% bovine calf serum. Cells were treated with 1 μM staurosporine (Cayman Chemical, Ann Arbor, MI) for 5 h, as an apoptosis-triggering reagent.

The homogenate proteins were subjected to sodium dodecyl sulfate polyacrylamide gel electrophoresis. Then, proteins were transferred onto polyvinylidene difluoride membranes (Immobilon-P; Millipore, Bedford, MA). The blots were immunoreacted with anticleaved poly-(adenosine diphosphate-ribose) polymerase (PARP; 1:1,000, rabbit polyclonal; Cell Signaling) or anti-β-actin (1:5,000, mouse monoclonal; Sigma Chemical Company, St. Louis, MO) antibodies, and the protein bands were visualized by chemiluminescence detection system (SuperSignal West Pico; Pierce).

Some sets of male mice for behavioral studies were exposed to anesthetics at P6, as described previously.1,17At 3 weeks of age, these mice were weaned and housed in groups of four animals per cage. These mice were further allowed to mature. At the indicated ages, they were subjected to behavioral tests such as open-field, elevated plus-maze, Y-maze, and fear conditioning tests. The movement of each mouse was monitored and analyzed using a computer-operated video tracking system (SMART, Barcelona, Spain). In tasks using an apparatus with arms, arm entry was counted when all four legs of the animal entered each arm. The apparatus was cleaned after each trial. All apparatuses used in this study were made by O'Hara & Co., Ltd. (Tokyo, Japan).

Emotional responses to a novel environment were measured by an open-field test using 6-week-old mice, by a previously described method.17Activity was measured as the total distance traveled (meters) in 10 min.

The elevated plus-maze test was performed as previously described.17This test is used to evaluate anxiety-related behavior in rodents. The elevated plus–maze consisted of two open arms (25 × 5 cm) and two enclosed arms, with all arms elevated to a height of 50 cm above the floor. Mouse behavior was recorded during a 10-min test period. The percentage of time spent in the open arms was used as an index of anxiety-like behavior. Mice used for the test were 6 weeks of age.

This study was performed as previously described17to assess spatial working memory. The symmetrical Y-maze made of acrylic consisted of three arms (25 × 5 cm) separated by 120° with 15-cm-high transparent walls. Each mouse was placed in the center of the Y-maze, and the mouse was allowed to freely explore the maze for 8 min. The sequence and total number of arms entered were recorded. The percentage of alternations was the number of triads containing entries into all three arms divided by the maximum possible number of alternations (total number of arm entries minus 2) × 100. Mice used for the test were 6 weeks of age.

This is a simple and sensitive test of hippocampal-dependent and hippocampal-independent learning, as described previously.1Briefly, the conditioning trial for contextual and cued fear conditioning comprised a 5-min exploration period followed by three conditioned stimulus–unconditioned stimulus pairings separated by 1 min each: unconditioned stimulus, 1 mA foot shock with a 1-s duration; conditioned stimulus, 80-db white noise with a 20-s duration; the unconditioned stimulus was delivered during the last seconds of conditioned stimulus presentation. A contextual test was performed in the conditioning chamber for 5 min in the absence of white noise 24 h after conditioning. A cued test (for the same set of mice) was performed by presenting a cue (80-db white noise, 3-min duration) in an alternate context with distinct visual and tactile cues. The freezing response rate (absence of movement in any part of the body during 1 s) was scored automatically and used as a measure of fear memory. Mice used for the test were 7 weeks of age. Mice subjected to the test were not used for any further testing.

Statistical analysis was performed using GraphPad Prism 5.0 (GraphPad Software Inc, La Jolla, CA). To obtain MAC value at each time point, nonlinear fit-log (agonist) versus  normalized response – variable slope was performed. Comparisons of the means of each group were performed using a one-way ANOVA followed by Neuman-Keuls multiple comparison post hoc  test. In the Y-maze task, comparisons of group performance relative to random performance were carried out using a two-tailed one-sample Student t  test. Two-way ANOVA for repeated measures was performed to examine differences of the percutaneous oxygen saturation among anesthetics (factor 1) and among time (factor 2) (a measure repeated over time). P  values less than 0.05 were considered statistically significant. Values are presented as mean ± SEM.

It was suggested that inspired concentrations needed to produce 1 MAC of anesthetics might decrease progressively over time during anesthesia in neonatal rodents.15Thus, MAC values for desflurane, sevoflurane, and isoflurane in P6 mice were measured as a function of anesthesia duration (fig. 1A). The MAC values at 1 h anesthesia were determined to be 11.5% for desflurane, 3.8% for sevoflurane, and 2.7% for isoflurane, respectively; MAC values at 3 h anesthesia were determined as 8.5% for desflurane, 3.3% for sevoflurane, and 1.8% for isoflurane; and MAC values at 6 h anesthesia were determined as 7.0% for desflurane, 2.5% for sevoflurane, and 1.5% for isoflurane (fig. 1B). Our data suggest that MAC values in 6-day-old mice for these anesthetics decrease similarly as a function of anesthesia duration.

Six-day-old mice were exposed to 8% desflurane, 3% sevoflurane, or 2% isoflurane. At 1 h, these concentrations are estimated within the range of 0.70 to 0.79 MAC multiples (desflurane: 0.70 MAC multiple; sevoflurane: 0.79 MAC multiple; isoflurane: 0.74 MAC multiple) in P6 mice. At 3 h, these concentrations are estimated within the range of 0.90 to 1.12 MAC multiples (desflurane: 0.94 MAC multiple; sevoflurane: 0.90 MAC multiple; isoflurane: 1.12 MAC multiple) in P6 mice. At 6 h, these concentrations are estimated within the range of 1.14 to 1.33 MAC multiples (desflurane: 1.14 MAC multiple; sevoflurane: 1.20 MAC multiple; isoflurane: 1.33 MAC multiple) in P6 mice.

To assess oxygenation levels, we examined percutaneous oxygen saturation (SpO²) in mice every hour during anesthetic treatment. Control samples were obtained from pups exposed to 25% oxygen during the same period. SpO²did not differ significantly among groups (table 1; two-way repeated measures ANOVA, drugs: F = 0.59, P = 0.67, time: F = 0.81, P = 0.57, interaction between drugs and time: F = 1.41, P = 0.12, n = 5 mice for each), suggesting that neonatal exposure to desflurane, sevoflurane, and isoflurane did not induce significant disturbances in oxygenation. Analysis of arterial blood before and after 6-h anesthesia revealed no significant differences in pH; partial pressure of arterial oxygen (PaO²) was observed among all groups (table 2, one way ANOVA, pH: F = 1.85, P = 0.16, PaO²: F = 2.38, P = 0.09, n = 5 mice for each group). Partial pressure of carbon dioxide (PaCO²) was increased in all anesthesia groups compared with 6-h control groups without significant difference among the anesthesia groups (table 2, one-way ANOVA with Neuman-Keuls multiple comparison post hoc  test, F = 15.58, P < 0.0001).

To investigate neurotoxic effects on the developing brain by desflurane, sevoflurane, or isoflurane, we performed a histologic evaluation of activated caspase-3⁺cells (fig. 2). Neonatal exposure to 4% or 8% desflurane for 6 h induced dramatic increases in the number of activated caspase-3⁺cells compared with time-matched controls in the regions including layer II of the parietal cortex, layer IV of the sensory cortex, the dorsal hippocampal commissure, and the retrosplenial cortex (figs. 2B, D, and J and fig. 3). The numbers of activated caspase-3⁺cells with 4% desflurane for 6 h were significantly less than those induced by 8% desflurane with the same treatment duration (fig. 3). Figure 4show high power views of layer IV in the sensory cortex on neonatal brains after anesthetic treatments. Note that relatively well-preserved activated caspase-3⁺neurons were seen occasionally. These neurons may be in an early stage of apoptosis.

To verify that the immunohistochemically detectable reactivity represented authentic apoptosis, and to quantify the apoptotic response, we examined the expression of cleaved PARP (another biomarker for confirming apoptotic cell death) in forebrain extracts by Western blot analysis (fig. 5). Compared with a time-matched control group, exposure to 4% or 8% desflurane for 6 h induced significantly more expression of cleaved PARP (fig. 5, B and C). Exposure to 4% desflurane for 6 h induced significantly less expression of PARP than that in mice exposed to 8% desflurane with the same treatment duration (fig. 5, B and C), suggesting that desflurane has concentration-dependent effects on neuroapoptosis. The specificity of the antibody for cleaved PARP was confirmed by comparing Western blots of extracts from NIH/3T3 cells treated with staurosporine (1 μM) versus  untreated cells (fig. 5A). The 89 kDa band was present only in staurosporine-treated NIH/3T3 cells and corresponded to the band seen in extracts from brains of anesthesia-treated pups (fig. 5A).

We also performed TUNEL as an independent measure of apoptotic cell death. Because the number of TUNEL⁺cells in the cortex was greatly increased at 6 h after anesthesia compared with immediately after anesthesia, we observed TUNEL staining 6 h after anesthesia. The pattern of TUNEL staining at 6 h after anesthesia was similar to that of activated caspase-3 staining (figs. 6and 7).

We further investigated whether a shorter treatment time (3 h) with the same concentration (8%) of desflurane also causes neuroapoptosis. Western blot analysis showed that differences between expression of cleaved PARP induced by treatment with 8% desflurane for 3 h and those of time-matched control mice were not statistically significant (fig. 5, B and E); however, anesthesia with 8% desflurane for 6 h induced significant increase of cleaved PARP expression compared with time-matched controls as described above, suggesting that desflurane has time-dependent effects on neuroapoptosis.

Previous studies reported that sevoflurane1,18and isoflurane12,16cause neurodegeneration in the developing brains of neonatal mice. We examined effects of neonatal exposure to sevoflurane or isoflurane at almost equivalent doses. The numbers of activated caspase-3⁺cells in mice with 8% desflurane (fig. 2C), 3% sevoflurane (fig. 2E), 2% isoflurane (fig. 2G), 4% desflurane (fig. 2I), or control (fig. 2A) for 3 h did not differ significantly in all brain regions examined (table 3, one-way ANOVA, all P  values more than 0.05). Similarly, Western blot analysis showed that these exposures did not induce significant differences in expression of the cleaved PARP compared with time-matched control mice (fig. 5E).

However, neonatal exposure to 3% sevoflurane or 2% isoflurane for 6 h significantly increased activated caspase-3⁺cells, with a pattern similar to that of 8% desflurane (fig. 2B, D, F, and H). Moreover, fewer activated caspase-3⁺cells in mice treated with sevoflurane or isoflurane were seen than in those with equivalent exposure to desflurane for 6 h in some regions (fig. 2D, F, H, and 8). Figure 4shows high-power views of layer IV in the neonatal sensory cortex after anesthetic treatments, indicating similar anatomic patterns among these anesthetics. Western blot analysis showed immunoreactivities for cleaved PARP in mice treated with sevoflurane or isoflurane for 6 h to be significantly greater than those in time-matched controls (fig. 5B), suggesting time-dependent effects on neuroapoptosis similar to desflurane. However, reactivity in pups treated with 8% desflurane for 6 h was significantly greater than in pups exposed to other agents (fig. 5B and D). Isoflurane induced greater cleaved PARP expression than did sevoflurane (fig. 5B and D). Similarly, sevoflurane or isoflurane induced fewer TUNEL⁺cells compared with equivalent exposure to desflurane with the same treatment duration (6 h) in some regions (fig. 9). Furthermore, sevoflurane induced fewer TUNEL⁺cells than did isoflurane (fig. 9) with the same treatment duration in some regions. These results indicated that the order of neurotoxic potencies in the developing brain was: desflurane > isoflurane > sevoflurane.

To examine behavioral activity in a novel environment, mice treated with 8% desflurane, 3% sevoflurane, or 2% isoflurane for 6 h as neonates were examined in an open-field test as adults. Compared with control animals, mice treated with anesthetics at the neonatal stage did not show abnormal activity as evaluated by the total distance traveled (fig. 10A, one-way ANOVA, F = 2.16, P = 0.10). To study whether anxiety-related behavior was altered in association with neonatal exposure to these anesthetics, mice were subjected to an elevated plus-maze test. Anxiety-related behavior was assessed based on the percentage of time spent in the open arms of the test equipment. Mice treated with desflurane, sevoflurane, and isoflurane did not differ significantly from control animals (fig. 10B, one-way ANOVA, F = 0.34, P = 0.80).

Working memory is the ability to hold information temporally to do complex cognitive tasks; it involves both the hippocampus and prefrontal cortex.19–20To examine whether exposure of the developing brain to desflurane, sevoflurane, and isoflurane was associated with changes in spatial working memory, the mice were tested in a Y-maze task. This task examines whether mice remember the position of the arm selected in the preceding choice. By nature, rodents seek a new arm, different from that selected in the preceding choice, but if working memory is impaired, the number of correct choices would be reduced in the Y-maze task. Control, desflurane, sevoflurane, and isoflurane mice performed this task with 66.0 ± 1.9%, 52.2 ± 1.7%, 62.5 ± 2.3%, and 60.3 ± 1.9% correct choices, respectively (fig. 11). Values of mice in the control, sevoflurane, and isoflurane groups were well above the expected results of random choices (random choice = 50%; one-sample Student t  test, P < 0.05 in the above cases). However, the value of desflurane was not significantly different from the expected level of random choice in the above case (random choice = 50%; one-sample Student t  test, P = 0.21). Mice with 8% desflurane treatment had significantly impaired performance compared with that of control mice (fig. 11, one-way ANOVA, drugs, F = 8.97, P < 0.0001; Neuman-Keuls multiple comparison post hoc  test, P < 0.001: control vs.  8% desflurane). Previously, we reported that neonatal exposure to sevoflurane did not affect performance in the Y-maze task,1consistent with results here. These results indicate that 8% desflurane has greater effects on working memory in adulthood, but equivalent exposures to sevoflurane or isoflurane do not. A Neuman-Keuls multiple comparison post hoc  test showed values for desflurane were significantly lower than those for sevoflurane or isoflurane (fig. 11).

To assess effects of neonatal exposure to desflurane, sevoflurane and isoflurane on long-term memory, mice were examined in a contextual/cued fear conditioning test. The freezing responses of mice exposed to desflurane, sevoflurane, and isoflurane were significantly reduced in the contextual test compared with those of controls after a 24-h retention delay (fig. 12A, one-way ANOVA, F = 7.31, P = 0.003; Neuman-Keuls multiple comparison post hoc  test, P < 0.001, < 0.01, and less than 0.001 for control vs.  8% desflurane, 3% sevoflurane, and 2% isoflurane, respectively). There were no significant differences among drugs (Neuman-Keuls multiple comparison post hoc  test, all P  values more than 0.05). Freezing responses of the mice after treatment with desflurane, sevoflurane, and isoflurane to cued fear were also significantly reduced compared with controls after a 48-h retention delay (fig. 12B, one-way ANOVA, F = 15.63, P < 0.0001; Neuman-Keuls multiple comparison post hoc  test, all P  values less than 0.001 for control vs.  8% desflurane, 3% sevoflurane, and 2% isoflurane). There were no significant differences among drugs (Neuman-Keuls multiple comparison post hoc  test, all P  values more than 0.05). These results suggest that mice treated with desflurane, sevoflurane, or isoflurane as neonates had impaired long-term memory as adults.

In this study, we observed that the MAC values at 1-h anesthesia of P6 mice were 11.5% for desflurane, 3.8% for sevoflurane, and 2.7% for isoflurane, respectively. Thus, the concentrations of desflurane (8%), sevoflurane (3%) or isoflurane (2%) used in this study were estimated within the range of 0.70 to 0.79 MAC multiples (desflurane: 0.70 MAC multiple, sevoflurane: 0.79 MAC multiple, isoflurane: 0.74 MAC multiple) in P6 mice. In addition, our data suggest that MAC values in 6-day-old mice for these anesthetics decrease comparably as a function of anesthesia duration. Therefore, we concluded that concentrations of desflurane, sevoflurane, or isoflurane used in this study were almost equipotent.

Our results indicated that in mice, neonatal exposure to desflurane induces increased apoptotic neurodegeneration both in time- and dose-dependent manners. Exposure to desflurane for 6 h induced far more than twice the amounts of apoptotic cells compared with 3-h exposure. As shown in figure 1, MAC values for 6-day old mice are unstable and decreased progressively, which is consistent with a previous report.14Thus, compared with data from 6-h anesthesia, data obtained from 3-h exposure groups represent relatively lower MAC multiples. Statistical analysis also revealed that the cell number of caspase-3⁺cells after 6-h anesthesia in the developing brain was significantly different among the anesthetics, although there would be a multiplicity problem (type I error) because a large number of inferences were conducted. To further verify the differences of neurotoxicity among anesthetics, we comparably examined cortical extracts from control and anesthetic-treated pups by Western blot analysis, using antibody specific for cleaved PARP. Cleaved PARP is a main cleavage target of caspase-3 in vivo ; this cleavage is readily detected in many apoptosis models.21Western blot analysis showed that levels of cleaved PARP immunoreactivity was significantly different among the anesthetics and the order of the immunoreactivity level was desflurane > isoflurane > sevoflurane.

In this study, we performed working memory tests as well as long-term memory tests. Several previous studies reported that neonatal exposure to volatile anesthetics led to deficits in several types of memory in adulthood.1,6,22However, limited studies comparatively investigated the effects of these drugs on cognitive abilities. Our results indicated that, although desflurane, sevoflurane, or isoflurane all induce deficits in long-term memory, only desflurane exposure induces working memory deficits. On the other hand, these mice exhibited normal performance in the open-field and the elevated plus-maze tests, suggesting that memory deficit does not result from secondary effects of impaired general activity. These findings suggest that desflurane has greater neurotoxicity and causes more severe cognitive impairment in the developing brain than do sevoflurane or isoflurane, although the current data do not allow us to draw any conclusions about the cause-effect relationship between neurotoxicity and behavioral deficits.

We previously indicated the potential risk of sevoflurane to induce disturbances in social behaviors that resemble those seen in autism,1although we did not investigate such deficits here. It would be interesting to investigate whether these behaviors could be induced more severely in desflurane-treated mice than those treated with sevoflurane or isoflurane.

Working memory refers to a cognitive function that provides concurrent temporary storage and manipulation of the information necessary to perform complex cognitive tasks.23Working memory is thought to be involved in higher executive functioning such as planning and sequential behavior; deficits in working memory are directly related to deficits in behavioral flexibility. In the current study, neonatal exposure to desflurane induced the most severe apoptosis in the brain and deficits in spatial working memory among three halogenated ethers. We cannot rule out the possibility that working memory is selectively vulnerable to desflurane. This possibility, however, is unlikely because neonatal exposure to a combination of drugs (midazolam, N²O, and isoflurane) in infant rats causes working memory deficits in the 8-arm radial maze task,2suggesting that this working memory deficit is not specific to desflurane. Our results indicate that sevoflurane and isoflurane as well as desflurane cause deficits in long-term memory, consistent with previous reports.1,15 

Recently, it was reported that isoflurane anesthesia induces persistent, progressive memory impairment, a loss of neural stem cells, and reduced neurogenesis in young but not adult rodents.24Neurogenesis in the brain proceeds throughout adulthood, but the molecular mechanisms underlying the regulation of stem cell proliferation and differentiation have not been fully elucidated. Many reports indicate that impaired adult neurogenesis is associated with deficits in some hippocampal-dependent memory, including working memory.25,–,27In this regard, it is notable that isoflurane caused reduced neurogenesis in juvenile animals.24 

Consistent with our study, Stratmann et al.  showed that a 4-h exposure to 1 MAC of isoflurane led to impaired hippocampal functions.22,28On the other hand, Loepke et al.  showed that a 6-h exposure to 1.5% isoflurane in P7 mice did not lead to any impairment in neurocognitive function.29We speculate that the following factors might contribute to the differences of these results from our study. First, our isoflurane exposure concentration (2%) was higher than that used by Loepke et al.  Although the difference in concentrations might appear small, there could be a critical point between them. In this regard, the degree of neurodegeneration in the study of Loepke et al.  seems considerably less than in our study. Second, the treatment ages differed: P6 versus  P7. Vulnerability of regions responsible for learning functions might differ considerably between these ages.

Liang et al.  30recently reported that isoflurane caused greater neurodegeneration in the developing brain of neonatal mice than did equivalent exposure to sevoflurane. Consistent with their finding, our results indicated that isoflurane caused greater neurodegeneration than an equivalent exposure to sevoflurane in the developing brain of neonatal mice. Furthermore, our study extends the findings of Liang et al.  30by demonstrating that at almost equivalent potencies, desflurane triggered a significantly greater increase in apoptotic neurodegeneration in neonatal mouse brain compared with isoflurane and sevoflurane. What mechanism might underlie the differential effects of these halogenated ethers? One possibility is that the different potentials of these drugs to cause apoptosis are correlated with the different potentials to inhibit NMDA receptor activity in the developing brain, although we did not examine this finding in the current study. Future studies are necessary to determine the relationship between the effects of anesthetics on neurodegeneration and the inhibition of NMDA receptors. Another possibility is that differential effects of halogenated ethers might be related to their different abilities to induce calcium release from the endoplasmic reticulum through activation of inositol-1,4,5-trisphosphate receptors. Isoflurane reportedly induces more severe damage on cultured cells than does sevoflurane, which might result from isoflurane's greater ability to induce calcium release from the endoplasmic reticulum.31However, our results do not seem to support the possibility that different apoptosis induction levels are caused by different potencies of the drugs with respect to calcium release from the endoplasmic reticulum, as suggested by Liang et al.  30If this were the case, desflurane would be less potent than isoflurane in inducing neurodegeneration, because isoflurane is more potent than sevoflurane or desflurane in inducing calcium release from endoplasmic reticula in cultured cells.31 

In conclusion, the current study suggests that desflurane has relatively greater neurotoxicity than sevoflurane or isoflurane, as shown by greater neuroapoptosis and additional impaired cognitive function. The order of neurotoxic potencies is desflurane > isoflurane > sevoflurane. A recently published work by Istaphanous et al.  reported that all three anesthetics have similar neurotoxic profiles, which differed from our findings.32Although we do not know the reason for the discrepancy, different mice strains (pups from CD1 and C57BL/6 breeding pairs vs.  inbred C57BL/6) or subtle difference of age (P6 vs.  P7, P8) might contribute to the difference from our results. Additional studies are needed to determine underlying mechanisms for the neurotoxicity of halogenated ethers and the relationship between neurodegeneration and cognitive functions. Nevertheless, our findings provide the basis for differential use of halogenated ethers in obstetric and pediatric medicine.

The authors thank Makoto Ozaki, M.D., Ph.D. (Professor, Department of Anesthesiology, Tokyo Women's Medical University, Tokyo, Japan), for critical discussion of this study; Kiyoko Takamiya and Yuko Ogura (Technicians, Department of Anesthesiology, National Defense Medical College, Tokorozawa, Saitama, Japan) and Tatsuyo Harasawa (Technician, Central Research Laboratory, National Defense Medical College) for excellent technical help in this study; and Kouichi Fukuda, Ph.D. (Associate Professor, Center for Laboratory Animal Science, National Defense Medical College), for assistance in animal administration.