Ethical approval for this study was obtained from the Animal Care and Use Committee of Tokyo Medical and Dental University (approval numbers 0140163C and 0150033A). Litters of postnatal day 5 C57BL/6 male mice (average body weight 2.7 g) along with their mothers were purchased from SLC Japan, Inc., Shizuoka, Japan. The animals were housed under a 12-h light-dark cycle (lights on 08:00 to 20:00) at 21 ± 1℃. Thirty of the male pups were placed under anesthesia with sevoflurane on postnatal day 6 and then returned to their original litters. The same number of pups from each litter was used for the experiments to reduce variability related to the use of different litters. At the age of 3–4 weeks, the male pups were weaned and housed four animals to a cage. The mice had ad libitum access to water and food. Behavioral testing was performed in adulthood (11–13 weeks of age). This study used as few animals as possible. Only male mice were used to avoid potential variability caused by the estrous cycle [8].

On postnatal day 6, male pups (SEVO group; n = 10) were placed in a humid chamber (180 × 180 × 200 mm) with a mat heated to 38 ± 1℃, and 3% sevoflurane was delivered using a calibrated flow meter (Shimano, Tokyo, Japan) as described previously [67]. The gas was administered at 1 L/min in 40% O₂ for 6 h. Pups exposed to only 40% O₂ for 6 h were used as controls (NA group; n = 10). Thirty minutes prior to sevoflurane exposure, the NADPH oxidase inhibitor apocynin was administered intraperitoneally at 50 mg/kg for a total volume of 10 µl (apocynin + SEVO group; n = 10) as previously described [6]. The pups from each experimental group were returned to their litter and allowed to mature.

At the age of 11–13 weeks, mice were subjected to the contextual fear conditioning test. Their movement was monitored using a computerized video tracking system designed by O'Hara & Co., Ltd. (Tokyo, Japan), as described previously [67]. The test consisted of three conditioned-unconditioned stimulus pairings separated by 1-min intervals. Each pairing was as follows: unconditioned stimulus, 0.5-mA foot shock of 1-s duration; and conditioned stimulus, 60-dB white noise of 20-s duration. The unconditioned stimulus was delivered during the last seconds of the conditioned stimulus. The contextual test was performed 24 h after conditioning. The duration of freezing in response to the stimuli was automatically recorded to measure fear memory. The freezing time during total stimulation (5 min) was presented as the freezing percentage, which was compared between each group.

Two hours after the contextual fear conditioning test, the mice (n = 6 in each group) were deeply anesthetized with pentobarbital (50 mg/kg, intraperitoneal injection) and transcardially perfused with 4% paraformaldehyde (PFA) for 5 min. Thereafter, their brain tissue was removed and postfixed in the same buffer overnight at 4℃. Fifty-micrometer-thick floating coronal vibratome sections at 250-µm intervals (starting at bregma 2.20 mm to bregma −2.70 mm) were obtained for immunohistochemistry as previously described [7].

c-Fos staining was performed to investigate neuronal activity after the fear conditioning test. Sections were washed with phosphate-buffered saline (PBS) with 0.3% triton three times (10 min each), and endogenous peroxidase activity was blocked with 1% H₂O₂ (30 min) [91011]. Then, the sections were incubated with anti-c-Fos antiserum (1:1000; Santa Cruz Biotechnology, Santa Cruz, Dallas, TX, USA) diluted with 1% blocking reagent in PBS with Tween 20 (Katayama Chemistry, Osaka, Japan) overnight at 4℃. Next, the sections were incubated with peroxidase-conjugated secondary antibodies (1:1000; Abcam, Inc., Cambridge, UK) for 2 h at room temperature and washed with PBS. The c-Fos signal was visualized using a 3,3′-Diaminobenzidine (DAB)-nickel substrate (0.05% DAB, 0.05% NiSO₄, 0.015% H₂O₂, 0.05% 1 M Tris-HCl, pH 6.7). The sections were then mounted on slides, brushed, and dried overnight. The number of c-Fos-expressing cells in the BLA was counted as described previously [7]. Briefly, four sections in the BLA from each group were carefully matched among subjects using anatomical landmarks from a brain atlas. The images were counted by a single experimenter who was blind to the treatment conditions. The number of c-Fos-expressing cells in the BLA was stated as the number of c-Fos-expressing cells/10⁻¹ mm² [12].

To investigate the neuron type and distribution of cells expressing c-Fos, the sections were co-incubated with anti-c-Fos antiserum and monoclonal mouse anti-Glutamic Acid Decarboxylase (GAD)67 (1:200, MAB5406; Chemicon, Temecula, CA, USA) as the GABAergic neuron marker, or with polyclonal goat anti-Excitatory Amino Acid Carrier 1 (EAAC1) (1:1000, AB1520; Chemicon) as the glutamatergic neuron marker, and then diluted in 1% blocking reagent in PBS with Tween 20 overnight at 4℃. Next, the sections were washed with PBS and co-incubated with Alexa-594-conjugated anti-rabbit IgG (1:1000; Invitrogen, Carlsbad, CA, USA) and Alexa-488 (1:1000; Invitrogen)-conjugated anti-mouse IgG or Alexa-488-conjugated anti-goat IgG (1:1000; Invitrogen) secondary antibody for 2 h at room temperature. Sections were counterstained with 4,6-diamidino-2-phenylindole (DAPI; Thermo Fisher Scientific Corp., Carlsbad, CA, USA) and then mounted on a slide with Fluoromount (Diagnostic BioSystems, Pleasanton, CA, USA). The positive cells in the BLA of each group were counted in a blind manner as previously described [7]. Briefly, positive cells in the BLA were counted and the four sections were fluorescently stained. The fluorescent images were examined with a microscope using the Zeiss LSM Image Browser (Carl Zeiss MicroImaging, Jena, Germany). Settings were adjusted based on the fluorescence intensity of tissue from the control group and were identical for all sections.

Statistical analysis was performed using the SPSS statistical software package (SPSS ver. 24.0; IBM, Corp., Armonk, NY, USA). The Kruskal-Wallis omnibus test was used to compare data, followed by Bonferroni post tests. Values of P < 0.05 were considered statistically significant, and data are presented as mean ± SEM.

c-Fos staining was performed to investigate neuronal activity after the fear conditioning test. c-Fos expression in the BLA was lower in the SEVO group than in the NA group (Fig. 1) [7]. Apocynin treatment preserved neuronal activity by maintaining c-Fos expression in the BLA (Fig. 1D, SEVO vs. apocynin + SEVO, P = 0.046).

Sevoflurane exposure significantly reduced the freezing time in the contextual fear conditioning test, while the administration of apocynin increased long-term memory (Fig. 1E, SEVO vs. Apocynin + SEVO, P = 0.022), as previously described [6].

Recently, we showed that the cells expressing c-Fos in the BLA were glutamatergic neurons (Fig. 2) [7]. There were fewer c-Fos-expressing glutamatergic BLA neurons in the SEVO group than in the NA group, as previously described (Fig. 1) [7]. Apocynin treatment preserved neuronal activity by maintaining the glutamatergic neurons in the BLA (Fig. 3B, SEVO vs. Apocynin + SEVO, P = 0.014). Recently, we found that the NADPH oxidase inhibitor apocynin reduces neurotoxicity by decreasing the superoxide levels and preventing mitochondrial dysfunction [6]. Furthermore, apocynin improved long-term memory impairment in neonatal mice exposed to sevoflurane [6]. Together with the results of our recent studies, this study demonstrated that the administration of apocynin not only prevents the neuroapoptosis that occurs immediately after neonatal sevoflurane exposure, but also preserves the glutamatergic neurons in the BLA long after sevoflurane exposure. The number of GAD67-positive cells remained unchanged in both groups, indicating that the number of GABAergic neurons was not affected by neonatal sevoflurane exposure (Fig. 3A).