Neonatal Exposure to Sevoflurane Induces Abnormal Social Behaviors and Deficits in Fear Conditioning in Mice

The experiments were approved by the Committee for Animal Research at National Defense Medical College (Tokorozawa, Saitama, Japan). Pregnant C57BL/6 mice were purchased from SLC (SLC Japan Inc., Shizuoka, Japan). The animals were illuminated with a 12-h light–dark cycle (light from 07:00 to 19:00), and room temperature was maintained at 21°± 1°C. At the age of 3 weeks, the mice were weaned and housed in groups of 4 animals in a room. Mice had ad libitum  access to water and food.

Previous studies reported that there is litter variability in the rate of apoptosis that occurs spontaneously in neonate mice.11Therefore, a balanced number of control and experimental animals were drawn from the same litters, so that each experimental condition had its own group of littermate controls. Only the male offspring were used in this study. A total of 51 litters, 101 control and 103 treated pups, were used in this study.

Postnatal day 6 (P6) male mice were placed in an acrylic box and exposed to 3% sevoflurane or no anesthetics for 6 h. The total gas flow was 2 l/min, using air as a carrier. During anesthetic exposure, the mice were kept warm on a plate heated to 38°C. Control and experimental animals were under the same treatment and environment except that the control animals were exposed only to air.

Arterial blood analysis was performed essentially as described previously.5,12Briefly, the pups underwent a quick arterial blood sampling from the left cardiac ventricle, and the samples were transferred into heparinized glass capillary tubes. A single sample (55 μl) was analyzed immediately after blood collection by blood gas analyzer (ABL800; Radiometer, Copenhagen, Denmark). Samples were obtained immediately after removal from the maternal cage (0 h) or at the end of anesthesia (6 h). At the time of blood sampling, the experiments were terminated by decapitation.

Cerebral blood flow (CBF) was measured by a laser-Doppler blood perfusion imager (Peri Scan PIM II; Perimed, Stockholm, Sweden). Mice were taken out of the chamber before and every hour during anesthetic treatment and were placed face down on the floor while being continuously exposed to sevoflurane via  a tube with its opening positioned at the nose of the animals. Their head skins were peeled for scanning CBF, and data were captured using appropriate software (LDPIwin version 2.6; Lisca, Linköping, Sweden). The perfusion response is presented in arbitrary perfusion units. Because the arbitrary perfusion units values are not absolute blood flow, the magnitude of the difference in perfusion was calculated as the ratio between the area of maximum peak perfusion and areas of baseline perfusion. Arbitrary perfusion unit values were compared between anesthetized animals and those with mock anesthesia at baseline and at 1-h intervals for 6 h.

Animals from both treatment and control groups were perfused transcardially with 0.1 m phosphate buffer containing 4% paraformaldehyde immediately after 6 h of sevoflurane anesthesia, and then the brains were exposed to immersion fixation for 24 h at 4°C. The brains were histologically analyzed using paraffin-embedded sections (5 μm thick). For immunohistochemistry, anti–active caspase-3 antiserum (D175; Cell Signaling Technology, Beverly, MA) was used at dilutions of 1:400 in antibody diluent (Dako, Glostrup, Denmark). Before to use, sections were dewaxed in xylene and hydrated using a graded series of ethanol. Antigenic 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,13followed by blocking with a nonspecific staining blocking reagent (Dako) for 1 h to reduce background staining. The sections were then incubated overnight in a humidified chamber at 4°C. Subsequently, peroxidase-conjugated secondary antibody (DAKO En Vision + system; Dako) and 3,3-diaminobenzine-tetrachloride (DAB; Vector Laboratories, Burlingame, CA) were used according to the manufacturer’s instructions. Finally, the sections were counterstained with Nissl. Activated caspase-3–positive cells were counted by the investigator who was blinded to the treatment conditions.

Terminal deoxynucleotidyl transferase–mediated deoxyuridine 5-triphosphate–biotin nick end labeling staining was performed using an in situ  apoptosis detection kit (ApopTag fluorescein; CHEMICON, Temecula, CA) according to the manufacturer’s protocol. Sections were counterstained with 4′,6-diamidino-2-phenylindole (DAPI). Fluorescein was histochemically examined with a fluorescent microscope (TE-2000E; Nikon, Tokyo, Japan) equipped with interlined charge-coupled device camera (DS-U1; Nikon).

Mice forebrain was quickly removed and were homogenized in four volumes of 20 mm Tris-HCl, pH 7.4, 2 mm EDTA, 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 rest of the 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 in each sample was measured using a protein assay kit (BCA; Pierce, Rockford, IL).

The homogenate proteins were subjected to sodium dodecyl sulfate polyacrylamide gel electrophoresis. The proteins were transferred onto polyvinylidene fluoride membranes (Immobilon-P; Millipore, Bedford, MA). The blots were immunoreacted with anti–cleaved poly(adenosine diphosphate–ribose) polymerase (PARP; 1:1,000, rabbit polyclonal, Asp214; Cell Signaling) or anti–β-actin (1:5,000, mouse monoclonal, AC-15; Sigma, St. Louis, MO) antibodies, and the protein bands were visualized by chemiluminescence detection system (SuperSignal West Pico; Pierce).

As described previously for CBF and histopathologic studies, some sets of mice for behavioral studies were exposed to 3% or 0% sevoflurane for 6 h at P6. They were allowed to mature, and at the appropriate ages, sevoflurane and control mice underwent behavioral tests, namely, open-field, elevated plus-maze, Y-maze, fear conditioning, social recognition, social interaction, olfactory, and novelty tests. The movement of each mouse was monitored and analyzed using a computer-operated video tracking system (SMART, Barcelona, Spain). In the tasks using 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 apparatus 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 8-week-old mice, by a previously described method.14Activity was measured as the total distance traveled (meters) in 10 min.

The elevated plus-maze test was performed as previously described.14The elevated plus maze consisted of two open arms (25 × 5 cm) and two enclosed arms being elevated to a height of 50 cm above the floor. Normally, mice prefer a closed environment to an open area. 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 aged 8 weeks.

This study was performed as previously described.14This study allowed us to assess spatial working memory. The symmetrical Y maze made of acrylic consists 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 the total number of arms entered were recorded. The percentage of alternation is 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 aged 11 weeks. The motion of the animals was manually recorded.

This is a simple and sensitive test of hippocampal-dependent and hippocampal-independent learning as previously described.14Briefly, the conditioning trial for contextual and cued fear conditioning consisted of a 5-min exploration period followed by three conditioned stimulus–unconditioned stimulus pairings separated by 1 min each: unconditioned stimulus, 1 mA foot shock intensity, 1 s duration; conditioned stimulus, 80 db white noise, 20 s duration; 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 at 24 h after conditioning. A cued test (for the same set of mice) was performed by presentation of a cue (80 db white noise, 3 min duration) in alternative context with distinct visual and tactile cues. The rate of freezing response (absence of movement in any parts of the body during 1 s) was scored automatically and used to measure fear memory. The test was performed on mice of two different age groups: 8 weeks or between 14 and 17 weeks.

Social recognition test was conducted as described previously.15We transferred 18-week-old mice from group to individual housing for 7 days before testing to permit establishment of a home cage territory. Testing began when a stimulus female mouse was introduced into the home cage of each male mouse for 1-min confrontation. At the end of the 1-min trial, the stimulus animal was removed and returned to an individual cage. This sequence was repeated for four trials with 10-min intertrial intervals, and each stimulus was introduced to the same male resident in all four trials. In a fifth trial, another stimulus mouse was introduced to a resident male mouse.

Caged social interaction for social versus  inanimate targets was performed in an open field using two cylinder cages allowing olfactory and minimal tactile interaction as described previously.16The cylinder cages were 10 cm in height, with a bottom diameter of 9 cm and bars spaced 7 mm apart.

Fifteen-week-old mice were habituated to the flavor of a novel food (blueberry cheese) for 3 days before testing. On the fourth day, after 24 h of food deprivation, a piece of blueberry cheese was buried under 2 cm of bedding in a clean cage. The mice were placed in the cage, and the time required to find the food was measured manually.

Activity was measured as the total duration of interaction with an inanimate novel object (red tube) in 10 min.

The same set of mice underwent social recognition (at 19 weeks of age), social interaction (at 14 weeks of age), olfactory (at 15 weeks of age), and novelty (at 15 weeks of age) tests. In other analyses, each test was conducted with a new set of animals.

Statistical analysis was performed using Statview software (SAS, Cary, NC). Comparisons of the means of two groups were performed using the Student t  test. In the Y-maze task, comparisons of group performance relative to random levels were performed using the one-sample t  test. Data of the social recognition task were analyzed by repeated-measures two-way analysis of variance. Values are presented as mean ± SEM.

To examine the effect of neonatal exposure to sevoflurane, we exposed P6 mice to 3% sevoflurane for 6 h. Hypoxia is a known cause of neuronal cell death.17To assess the adequacy of ventilation and oxygenation, we examined the blood gas data in mice during the anesthesia. Control samples were obtained from pups exposed to air during the same period. We found that pH, arterial carbon dioxide tension, arterial oxygen tension, and arterial oxygen saturation did not differ significantly from sham control (table 1). These results, together with the fact that pups looked pink throughout the 6 h of gas exposure, led us to conclude that it was unlikely that apoptosis in this protocol was caused by hypoxia/ hypoventilation.

Further, to assess the adequacy of cerebral perfusion, we measured CBF during anesthesia using a laser-Doppler blood perfusion imager. Control mice were exposed to air for corresponding period of the sevoflurane treated mice. Anesthesia treatment with sevoflurane did not affect CBF compared with control mice at any point during the 6 h of anesthesia (figs. 1A and B).

Sevoflurane anesthesia significantly increased cleaved caspase-3 apoptosis in the mice immediately after exposure (table 2and figs. 2–4). Figures 2B and Dshowed that the increased apoptosis was most robust in the caudate/putamen, retrosplenial cortex, dorsal hippocampal commissure, and neocortex in the brains of pups with sevoflurane exposure. In other sections, thalamus, subiculum, inferior colliculus, and pontine nuclei were also shown to be damaged severely (figs. 3B and C). Figure 4Bindicated that severe neuronal damage occurred in the extrahippocampal circuit, which is believed to be important for mediating learning and memory functions,18including dorsal hippocampal commissure as well as thalamus and retrosplenial cortex. Apoptosis were also observed in amygdala (fig. 4D), hippocampus (fig. 4F), frontal cortex (fig. 4H), and mammillary complex (fig. 4J). These data indicated that the apoptotic response to sevoflurane was robust and followed a pattern that was characteristic of the pattern reported for other anesthetic drugs or ethanol.18 

It was reported that caspase-3 cleavage may occur under a condition that is a nonapoptotic event,19raising questions about the reliability of cleaved caspase-3 immunostaining for detecting cell death. Therefore, to verify that the cleaved caspase-3 immunoreactivity represent authentic apoptosis, we also performed terminal deoxynucleotidyl transferase–mediated deoxyuridine 5-triphosphate–biotin nick end labeling as another independent measure of apoptotic cell death. We found the same pattern of staining as observed by cleaved caspase-3 staining (figs. 4L, N, P, and R).

Further, to verify that the previously described immunohistochemically detectable reactivity represented authentic apoptosis and to quantify the apoptosis response, we examined cortex extracts from control and sevoflurane-treated pups by Western blot analysis using antibody specific for cleaved PARP. PARP is one of the main cleavage targets of caspase-3 in vivo , and the cleavage is readily detected in many apoptosis model.20Western immunoblotting with anticleaved PARP antibody detected immunoreactivity in sevoflurane-exposed pup extracts, whereas the band was under detection level in control pub brain extracts (fig. 4S).

To examine responses to a novel environment, mice with neonatal exposure to sevoflurane were assayed in an open-field test. These mice did not differ from control animals in their exploratory behavior (fig. 5A; t  test, t = 1.24, P > 0.05). To study whether anxiety-related behavior of mice with neonatal exposure to sevoflurane was affected, mice underwent an elevated plus-maze test. Anxiety-related behavior was assessed by the percentage of time spent in the open arms of the test equipment. Anesthetized mice did not differ significantly in the percentage of time spent in the open arms (fig. 5B; t  test, t = 0.76, P > 0.05). These results indicate that the emotional state of mice with neonatal exposure to sevoflurane did not differ grossly from controls under the conditions of this study.

Further, to examine whether exposure of the developing brain to sevoflurane was associated with changes in spatial working memory, mice underwent a Y-maze spontaneous alternation task. Working memory refers to a cognitive function that provides concurrent temporary storage and manipulation of the information necessary for complex cognitive tasks. This test examines whether mice remember the position of the arm selected in the preceding choice. Mice with and without sevoflurane exposure performed this task with 64.3 ± 7.3% and 60.0 ± 8.3% correct choices, respectively, which were well above the expected results of random choices (random choice = 50%; one-sample t  test, P < 0.05), there being no difference compared with control (fig. 5C; t  test, t = 1.20, P > 0.05). This result suggests that spatial working memory was not affected by exposure of the developing brain to sevoflurane.

To assess the influence of neonatal exposure to sevoflurane on long-term memory, mice underwent contextual/cued fear conditioning. In this paradigm, mice learn to associate previous neutral auditory cues and the apparatus (context) with electric foot shock in a single training session, such that robust long-term memory was established for an experimental context (hippocampus dependent) and an auditory cue (hippocampus independent).21,22Long-term memory was assessed based on the freezing reaction of the mice in response to the context or the conditioned cue. The freezing response to the same context in mice with neonatal exposure to sevoflurane was reduced significantly compared with controls after a 24-h retention delay at 8 weeks of age (fig. 6A; t  test, t = 3.10, P < 0.01). The response of mice with sevoflurane to the cued fear conditioning was also reduced significantly after a 48-h retention delay compared with control mice (fig. 6B; t  test, t = 3.16, P < 0.01).

It was reported that exposure of infant mice to ethanol induced neuroapoptosis and subsequent memory impairments that were very severe at P30 and less severe at P75.18This result provided evidence favoring the interpretation that recovery of some type of learning functions might occur in later adulthood in ethanol-treated mice. Therefore, further to examine whether the neonatal exposure to sevoflurane causes permanent neurocognitive deficits in mice and how it evolves over time, we undertook an assessment of hippocampal function at later time point using another set of mice.

We found that neonatal exposure to sevoflurane caused deficits in the fear conditioning test at 14–17 weeks of age similar to the results at 8 weeks. The freezing response of mice with sevoflurane was reduced significantly in contextual tests compared with that of controls after a 24-h retention delay (fig. 6C; t  test, t = 3.48, P < 0.01). The freezing response of sevoflurane-exposed mice to cued fear was also reduced significantly compared with that of controls after a 48-h retention delay at 14–17 weeks of age (fig. 6D; t  test, t = 2.11, P < 0.05). These results strongly suggested that exposure of P6 mice to sevoflurane caused hippocampal-dependent and -independent neurocognitive deficits that persisted for relatively long time periods (at least from 8 to 14–17 weeks of age) of the mice’s lifespan.

Mice are a social species and exhibit social interaction behavior.23Therefore, we investigated whether mice with neonatal exposure to sevoflurane display abnormal social behaviors. First, we investigated social memory, which depends predominantly on olfactory cues. This ability is needed for social familiarity and can be identified as a consistent decrease in olfactory investigation during repeated encounters with a female in the social recognition test. Control mice showed a significant decline in the time spent in investigating a female with subsequent presentation of the same female in trials 3 and 4, as compared with trial 1 (fig. 7A). This decrease was not due to a general decline in olfactory investigation, because presentation of a novel female during trial 5 resulted in a similar amount of investigation as trial 1 with the original female. In contrast, mice with neonatal exposure to sevoflurane showed high levels of sustained investigation at each encounter with the same female and the same level of investigation when presented with a new female at trial 5, significantly different from the response of controls (analysis of variance, F = 14.51, P < 0.001 [between control and sevoflurane administration]; F = 28.34, P < 0.0001 [between trials]; F = 16.64, P < 0.0001 [interaction between trials and sevoflurane administration]; fig. 7A). These data suggest that mice with neonatal exposure to sevoflurane do not develop social memory.

In a test for social versus  inanimate preference, control mice spent significantly more time interacting with the social target than with the inanimate target (fig. 7B; t  test, t = 7.30, P < 0.0001). In contrast, mice with neonatal exposure to sevoflurane spent a similar amount of time interacting with both targets (fig. 7B; t  test, t = 1.77, P > 0.05). Furthermore, mice with neonatal exposure to sevoflurane exhibited decreased interaction with a social target compared with controls (fig. 7B; t  test, t = 2.38, P < 0.05), indicating a social interaction deficit. We did not attribute the abnormalities in social recognition and interaction to impairment in general interest in novelty or olfactory sensation, because we did not detect significant differences between groups in tests for novel inanimate object interaction (fig. 7C; t  test, t = 0.21, P > 0.05) or for olfaction (fig. 7D; t  test, t = 0.12, P > 0.05). Therefore, it can be concluded that mice with neonatal exposure to sevoflurane demonstrated deficits in social behavior.

This study showed that single administration of sevoflurane to neonatal mice caused a significant increase in neuroapoptosis in the brain compared with littermate controls exposed only to air. Consistent with other recent evidence that apoptotic neurodegeneration can be induced by exposure during the brain growth spurt to drugs that block NMDA receptors and/or hyperactivate GABA^Areceptors,5,6,18neonatal exposure to sevoflurane was shown to induce widespread apoptosis in several major brain regions, leading to impaired learning later in adulthood. This finding for sevoflurane is consistent with recent finding by Johnson et al. ,24who found that a neonatal exposure to isoflurane triggers a significant neuroapoptosis response in the mouse brain. Our results of fear conditioning strongly suggested that exposure of P6 mice to sevoflurane caused learning deficits, although we could not rule out the possibility of the effects of sevoflurane on sensitization despite true conditioning to the auditory cue. Furthermore, this study also showed that exposure of neonatal mice to inhaled sevoflurane caused deficits in social behavior. To our knowledge, this is the first study to show that single administration of sevoflurane, which is a commonly used anesthetic in pediatric surgery throughout the world, causes a robust neuroapoptosis response in infant mouse brain and behavioral deficits in both cognitive and social spheres. The minimum alveolar anesthetic concentration that prevents purposeful movement in response to supramaximal noxious stimulation in 50% of animals (minimum alveolar concentration) of sevoflurane in human neonates is 3.3 ± 0.2%.8Therefore, the concentration of sevoflurane (3%) used in this mice study would be comparable to clinically used ranges.

Most anesthetics act via  NMDA and/or GABA^Areceptors. It was demonstrated that neonatal coadministration of an NMDA antagonist and a GABA^Aagonist was much more detrimental than either of these used alone.5,6Although the precise mechanism for this is yet to be understood, this evidence suggested that more severe neurodegeneration was induced when both NMDA and GABA^Areceptors were simultaneously altered in the developing brain. Therefore, an anesthetic that has NMDA antagonist and GABA^Aagonist properties would be of concern when administered to the developing brain.

Autism spectrum disorders (ASDs) are a group of common neuropsychiatric disorders characterized primarily by impairments in social, communicative, and behavioral functioning,25of unknown mechanism. Epidemiologic studies have shown that the prevalence of ASDs is 3–6 per 1,000 children. If neonatal exposure to anesthetics induces deficits in social behavior, a causal link between ASDs and neonatal exposure to anesthetics could be suggested, because deficits in social behaviors are a core feature of ASDs. Our findings revealed that neonatal exposure to sevoflurane induced deficits in social memory and social interaction in mice. These tests were thought to be core paradigms to test autistic behavior in mice and have been used to measure autistic behavior in other ASD models as well.16,17,26,27This study is the first to indicate the potential risk of general anesthetics to induce disturbances in social behaviors that resembles those seen in ASDs.

However, there is a caveat that the relevance of these mouse findings to the human situation is unknown and requires clarification. It is too early to say whether anesthetics have the same effect in humans. There may be species differences in the detrimental effects of anesthetic agents on the developing brain. Physicians could reduce any potential risks by limiting the duration of anesthetic administration in neonates.

In any case, the current results suggest the potential hazards of neonatal sevoflurane exposure in causing social behavioral alterations. We would like to insist on the need for further research to determine whether a correlation exists between anesthesia exposure during development and ASDs in human populations. What would be the potential association between the sevoflurane-induced apoptosis and sevoflurane-induced changes in social behaviors? Several authors have proposed that glutamate and GABAergic system disturbance in cortical network in ASDs may be characterized by an imbalance between excitation and inhibition in neuronal networks.28,29Such a change may lead to hyperexcitability or unstable neuronal networks, which may alter oscillatory rhythms in brain.30,31The excitatory/inhibitory balance in cortical networks may be controlled by the relative numbers and activities of glutamatergic and GABAergic neurons.29Indeed, reduced GABAergic inhibition by mutations of genes encoding subunits of the GABA^Areceptors is associated with ASDs.28,29,32Neural loss by a drug that may violate NMDA and GABA^Areceptors in the critical period, seen in the current study, might interfere with the developmental mechanisms patterning the balance between excitation and inhibition system and cause ASD-like behaviors.

Our results showed that neuronal degeneration was particularly severe in several of the specific brain regions that comprise the extrahippocampal circuit, which is believed to be important for mediating learning and memory functions. This pattern was similar to the pattern for neonatal exposure to ethanol, which has NMDA antagonist and GABA^Aagonist properties.18These pattern of brain damage and subsequent learning deficits are described in ethanol-treated mice.18In this regard, it is noteworthy that there is evidence that prenatal exposure to ethanol may be a factor in social difficulties in humans.33Further research, including electrophysiologic study, will be needed to address the molecular mechanisms that explain the relevance between neonatal exposure to sevoflurane and deficits in social behavior. It would also be necessary to study whether other drugs that cause neuroapoptosis in the critical period would induce deficits in social behavior.

The authors thank Kyoko Takeuchi, Ph.D. (Assistant Professor), Kenji Miura, Ph.D. (Lecturer), Kazuyo Kuroki (Technician, all from the Department of Developmental Anatomy and Regenerative Biology, National Defense Medical College, Tokorozawa, Saitama, Japan), Kunihito Takahashi, Ph.D., (Senior Researcher, Hakuju Institute for Health Science Co. Ltd., Tokyo, Japan), and Tatsuyo Harasawa (Technician, Central Research Laboratory, National Defense Medical College) for the participation in this study; and Kouichi Fukuda, Ph.D. (Associate Professor, Center for Laboratory Animal Science, National Defense Medical College), for the assistance in animal administration.