Neonatal Exposure to Sevoflurane in Mice Causes Deficits in Maternal Behavior Later in Adulthood

ACCUMULATING evidence indicates that exposure to general anesthetics at clinically effective concentrations induces widespread increases in neuronal apoptosis in the developing brain of a variety of animals ranging from rodents to rhesus monkeys.^1–10  Furthermore, long after the anesthetic exposure, learning deficits are manifested later in adulthood^1–5  even though a significant increase in neuronal apoptosis is no longer evident.¹¹  The primary cause of these learning deficits in the adults is not fully understood, which hinders identification of the underlying pathophysiology. The impairment of brain function caused by neonatal exposure to sevoflurane is not specific to the learning deficit. We previously reported that neonatal exposure to sevoflurane induced a disturbance in social behaviors in mice that resembles those observed in subjects with autism.³  This evidence suggests that pervasive deficits in brain functioning may be induced. However, there is not enough evidence to support this hypothesis.

Animals must adapt rapidly to changing environmental conditions. In mammals, pregnant females undergo fundamental behavioral changes: the pattern of care by mothers to their offspring, which is called maternal behavior, is induced in female mothers. Maternal behavior is critical in rodents because pups are born deaf and blind. Neural mechanisms for maternal behavior have been studied most extensively in rodents,^12–14  and it may be the most complex behavioral task for mice in normal laboratory conditions. Although it is believed that maternal behavior is influenced by hormones,^15,16  accumulating evidence suggests a specific hormonal condition is not necessary to induce maternal behaviors. For instance, even nonpregnant nulliparous mice can exhibit maternal behaviors when extensively exposed to pups¹⁷  although they are rarely maternal spontaneously and actively avoid pups.¹⁸  This evidence indicates that sensory stimuli provided by newborns are important in the rapid onset of maternal behaviors in mammals.^19–22 

Recently, we found a high rate of mortality in pups born to female mice that were exposed to sevoflurane in their early stage after birth. In this study, to examine the deficient survival of pups, we investigated whether maternal behavior was impaired in these dams. These mice showed serious impairment of maternal behaviors when they fostered their pups although parturition was normal, which severely impaired the survival of their offspring. We hypothesized two possible mechanisms for impairment of maternal behaviors in these dams: circulating neurohormones, vasopressin and oxytocin, which are related to maternal behaviors,^23–26  could have been altered by neonatal exposure to general anesthetics. Alternatively, activation of central circuits underlying maternal behaviors could have been altered without a change in release of vasopressin and oxytocin into the circulation. Accumulating evidence suggests the existence of central circuit for maternal behaviors in the medial preoptic area (MPOA) in the rostral hypothalamus.^19,27,28  In addition, we hypothesized that neonatal exposure to sevoflurane could have altered maternal behavior by a mechanism reversible by antioxidant: antioxidant reportedly mitigated behavioral deficits caused by neonatal exposure to general anesthetics.^29,30 

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). Inbred C57BL/6 mice were used in this study and maintained as described previously.⁵ 

Sevoflurane anesthesia was carried out as described previously.⁵  In brief, on postnatal day 6 (P6), pups were placed in a humid chamber immediately after removal of mice from the maternal cage. A 3% concentration of sevoflurane was administered in 30% oxygen as the carrier gas. Control mice were exposed to 30% oxygen. Hydrogen gas (1.3%) was supplied as described previously.³⁰  Total gas flow rate was 2 l/min.

In each experiment, siblings from the same litter were randomly allocated into one of the following groups so that each group was balanced on littermate. No obvious differences (e.g., body size and weight) were observed within the litters, and there was no significant difference in mean body weight among the groups (data not shown).

1. Survival rate of delivered pups: control, sevoflurane, and sevoflurane + hydrogen groups (n = 17–19 dams for each group); a minimum biologically important difference was set at a 30% decrease from the baseline level in the control group.

2. Pup exchange test: control and sevoflurane groups (n = 6 dams for each group); a minimum biologically important difference was set at a 30% decrease from the baseline level in the control group.

3. Behavioral studies: control, sevoflurane, and sevoflurane + hydrogen groups (n = 10–11 dams for each group); the primary outcome measure was latencies for pup retrieval; in the pup retrieval test, a minimum biologically important difference was set at a 30% increase from the baseline level in the control group.

4. Hormonal assay: control, hydrogen, sevoflurane, and sevoflurane + hydrogen groups (n = 4–5 dams for each group); a minimum biologically important difference was set as 30% decrease from the baseline level in the control group.

5. Immunohistochemical study: control and sevoflurane groups (n = 5 dams for each group); a minimum biologically important difference was set at a 30% decrease from the baseline level in the control group.

In total, we prepared 160 female pups, which received anesthesia or hydrogen treatment at P6 (55 of control, 56 of sevoflurane, 40 of sevoflurane + hydrogen, and 9 of hydrogen groups). Among them, eight pups with sevoflurane and one pup with sevoflurane + hydrogen died during the treatment. Then, these siblings from the same litter were reunited and cohoused till the experiment (mice were similarly caged and housed in all groups). At 3 weeks of age, mice were weaned and allowed to further mature. At 7–9 weeks of age, female mice were mated with healthy males that had not been exposed to any anesthetic. Among them, 23 female mice did not get pregnant (eight of control, six of sevoflurane, five of sevoflurane + hydrogen, and four of hydrogen groups) and 1 control mouse died due to failure of delivery. These mice were excluded from the final analysis. Thus, for first delivery experiments, we used 46 control dams, 42 sevoflurane-treated dams, 34 sevoflurane + hydrogen–treated dams, and 5 hydrogen-treated dams. These mice were allocated as described above (1–5 in this section).

Among them, some dams were further analyzed for behavioral studies of parous dams: the same sets of mice for behavioral studies in first-time delivery were reused in behavioral studies in second-time (parous) delivery (control: 7 for survival rate and 11 for behavioral studies; sevoflurane-treated: 8 for survival rate and 10 for behavioral studies).

For paternal study experiments, 26 age-matched male mice were either subjected to anesthesia (n = 13) or control (n = 13) treatment at P6 (no mice died during the treatment). Siblings from the same litter were allocated into each group almost equally (i.e., groups were balanced on littermate).

Plasma concentrations of oxytocin and vasopressin in dams at 10 weeks of age were examined by enzyme-linked immunosorbent assay using commercially available kits (Oxytocin enzyme-linked immunosorbent assay kit and arg⁸-Vasopressin enzyme-linked immunosorbent assay kit; Enzo Life Sciences, Farmingdale, NY). Assays were performed according to the manufacturer’s instructions. Blood samples were collected from the inferior vena cava within 6 h after parturition.

Immunohistochemical studies using the anti-c-Fos antibody (rabbit polyclonal; sc-52; Santa Cruz Biotechnology, Santa Cruz, CA) were performed as previously described.³⁰  Samples were obtained within 6 h after parturition. The numbers of immunoreactive cells were counted by an observer blinded to the groups.

On the morning of parturition, maternal behaviors were examined. Maternal behavioral studies using first-time mothers were performed at 10–12 weeks of age. The same sets of female mice were reused in the maternal behavioral studies for second-time (parous) mothers: those mice were mated again at 19–25 weeks of age, and maternal behaviors were examined at 22–28 weeks of age. Paternal behavioral studies using male mice were performed at 11 weeks of age. Survival rate (percentage of the number of pups at the indicated day compared with that at birth) was recorded until P6. In each experiment, observation was made by the same observer who was blinded to the groups. All apparatus used in this study was made by O’Hara & CO., LTD. (Tokyo, Japan).

Pregnant females were individually housed for a few days before parturition and examined for maternal behavior on the morning of parturition. The number of pups with milk in their digestive tract and that of poorly cleaned pups (with placenta, amniotic membrane, or umbilical cords) was recorded on that day. Nest quality was also evaluated at the same time using the score system described previously³¹  with some modifications: grade 3, shaped like a deep hollow surrounded by high banks; grade 2, a hollow with medium-height banks; grade 1, flat with low banks, but still discrete; grade 0, no depression in bedding with no banks. Each new dam was also evaluated for time spent crouching over pups and the percentage of newborns scattered for 20 min with minimal disturbance as described previously.³²  The percentage of scattered pups was expressed as a percentage with respect to time. We calculated the percentage of scatter as follows for each pup: (duration of scatter/total time observed (20 min) × 100). We then calculated the average for each group. These evaluations were carried out before the pup retrieval test.

The pup retrieval test was performed essentially as described previously.¹⁴  Before the test, pups were separated from dams for 30 min. At the beginning, each mouse was put in one corner of a cage and three of her pups were placed in different corners of the same cage. The cages were continuously observed for 10 min with minimal disturbance. Latencies to sniff a pup for the first time and to return each pup to the nest were evaluated.

Parental behavior of virgin male mice toward pups was evaluated for 20 min. At the beginning, each mouse was put in one corner of a cage and three new born pups were placed in different corners of the same cage as described in the pup retrieval test. Latencies to sniff a pup for the first time and the numbers of males which committed attacks toward pups were evaluated. If any of the pups was attacked during the test, all pups were removed immediately and this subject was considered as “attack.”

The pup exchange test was conducted as described previously with some modifications.¹⁴  Pups born to a female dam couple (a dam with sevoflurane exposure at P6 and a control), which were born on the same day, were exchanged within 12 h after delivery. The number of surviving pups was evaluated for 6 days after birth.

The olfactory test was conducted as described previously.³ 

Statistical analysis was performed using GraphPad Prism 5 (GraphPad Software Inc., La Jolla, CA). Comparisons of the means of each group were performed using Student t test, one-way ANOVA followed by Bonferroni post hoc test, and two-way ANOVA followed by Bonferroni post hoc test. Comparisons of the survival rate until P6 were performed using a log-rank (Mantel-Cox) test. We did not exclude any data in this study. P values of less than 0.05 were considered statistically significant. Values are presented as the mean ± SEM in bar graphs.

Female mice were exposed to 3% sevoflurane for 6 h and allowed to mature. These female mice appeared to grow normally and could bear pups. But we found that about half of their pups died within 2 days after birth, whereas pups born to control mice (without exposure to sevoflurane) showed more than 80% survival rate at 6 days after birth (fig. 1). A log-rank analysis confirmed the difference, indicating a significantly lower survival rate in pups from sevoflurane-treated dams compared with those from control dams (P < 0.0001). The number of delivered pups in sevoflurane-treated dams was not significantly differ from control dams (sevoflurane group, n = 110 from 17 dams; control group, n = 124 from 19 dams), indicating that parturition was normal in sevoflurane-treated dams.

The number of pups that did not have milk in their digestive tracts was increased in pups born to sevoflurane-treated dams when compared with pups born to control dams (fig. 2, A and B). However, the ratio of poorly cleaned pups was indistinguishable between them (fig. 2C).

Furthermore, to assess whether the excess mortality of pups born to sevoflurane-treated dams was caused by defects in nurturing, we investigated maternal behaviors in sevoflurane-treated dams. Around the time of delivery, mice usually prepare a high-walled, corner nest, which is significantly different from the flat, centrally located, sleeping pad of the nonpregnant female.²⁰  Because these features are characteristic of maternal females, the evaluation of nest quality is often used as an indicator of maternal behavior.²⁰  We found that sevoflurane-treated dams exhibited incomplete nest building compared with control dams (fig. 2, D–G). Comparison of the nest quality scores confirmed the difference, indicating the significant difference between sevoflurane-treated dams and controls (fig. 2G; t test, t = 3.32, P < 0.01). In addition, we examined the ratio of scattered pups as another indicator of maternal behavior.^32,33  We found that the ratio of scattered pups out of the nest was significantly higher in the sevoflurane-treated dams (fig. 2H; t test, sevoflurane-treated group vs. control group, t = 2.17, P < 0.05).

When all pups are in the nest, the dam normally hovers over them, allowing pups to suckle (crouching). We found that sevoflurane-treated dams exhibited a significantly shorter duration of crouching compared with control dams (fig. 2I; t test, sevoflurane-treated dams vs. control dams, t = 3.84, P < 0.01).

Maternal behavior was further evaluated by the pup retrieval test, which is frequently used to measure maternal behavior.^14,32,34,35  In the test, we monitored the response of dams to three pups placed in different corners of the cage for 10 min. Pups that wander from the nest are usually returned to the nest by a dam using mouth grip. Sevoflurane-treated dams displayed a significantly longer latency to retrieve the pups than control dams (fig. 3, A–C; t test, sevoflurane-treated dams vs. control dams, t = 2.25, P < 0.05 [first retrieval]; t = 2.33, P < 0.05 [second retrieval]; and t = 2.78, P < 0.05 [complete retrieval]). The impairment of retrieval was not caused by failure of the dams to detect the pups, because latency to approach and sniff a pup for the first time was indistinguishable between sevoflurane-treated and control dams (fig. 3D; t test, sevoflurane-treated dams vs. control dams, t = 0.52, P > 0.05). Furthermore, the olfactory test showed that olfactory function was also normal in sevoflurane-treated dams (fig. 3E; t test, sevoflurane-treated dams vs. control dams, t = 0.10, P > 0.05). These results indicated that exploratory and investigative behaviors toward pups were normal in sevoflurane-treated dams, yet they were unable to effectively perform maternal behaviors.

The reduced pup survival was likely caused by the impairment of maternal behaviors in sevoflurane-treated dams. To confirm whether the high mortality rate of pups born to sevoflurane-treated dams was caused by dams or pups, we carried out the pup exchange test. In this test, pups born to sevoflurane-treated dams were successfully fostered when nursed by control dams (fig. 4A). But, more than half of pups born to control dams died when nursed by sevoflurane-treated dams (fig. 4A). A log-rank analysis indicated that the survival rate of pups nurtured by sevoflurane-treated dams was significantly lower than the survival rate of pups nurtured by control dams during the 6 days after birth (P < 0.0001). In addition, the pups born to 3% sevoflurane-treated dams but nursed by control dams had milk in their digestive tracts, whereas pups born to controls but nursed by sevoflurane-treated mice did not (fig. 4B). Thus, we concluded that the sevoflurane-treated dams caused the survival deficit of the pups.

Subsequent to the first delivery, mice usually show a permanently enhanced rate of induction of maternal behaviors²⁰  although the underlying mechanism is largely unknown. Therefore, we investigated whether maternal behaviors were also impaired in second-time parous mice that were exposed to sevoflurane at P6. We found that pup survival rate was indistinguishable between the sevoflurane-treated second-time parous dams and the second-time parous control dams during the 6 days after birth (fig. 5). A log-rank analysis indicated that the survival rate of pups nurtured by sevoflurane-treated second-time parous dams was not significantly different from those of control dams during the 6 days after birth (P > 0.05). Majority of pups born to parous dams exposed to sevoflurane were cleaned and had milk in their digestive tracts, similar to pups born to parous control dams (fig. 6, A and B). Sevoflurane-treated parous dams also exhibited nest quality scores similar to parous control dams (fig. 6C; t test, sevoflurane-treated dams vs. control dams,t = 0.61, P > 0.05). The ratio of scattered pups was not significantly different between sevoflurane-treated parous dams and parous control dams (fig. 6D; t test, sevoflurane-treated group vs. control group, t = 0.93, P > 0.05). Time spent crouching was also indistinguishable between them (fig. 6E; t test, sevoflurane-treated dams vs. control dams, t = 0.36, P > 0.05).

In the pup retrieval test, sevoflurane-treated parous dams displayed latencies to retrieve the pups similar to controls (fig. 7A; t test, sevoflurane-treated dams vs. control dams, t = 0.07, P > 0.05 [first retrieval]; t = 0.06, P > 0.05 [second retrieval]; and t = 0.04, P > 0.05 [complete retrieval]). The latency to approach and sniff a pup for the first time was also indistinguishable between them (fig. 7B; t test, sevoflurane-treated dams vs. control dams, t = 0.83, P > 0.05). These results indicated that maternal behaviors of parous dams were indistinguishable regardless of sevoflurane exposure in their early stage after birth.

Accumulating evidence suggests that the neural circuit for maternal behavior exists in males as well as female mice.²⁰  In some rodents, behavior toward the young is essentially the same for each sex.^36,37  To investigate whether parental behavior is impaired by neonatal exposure to sevoflurane, sevoflurane-treated virgin male mice were exposed to three newborn pups for 20 min. The latency to approach and sniff a pup for the first time was indistinguishable regardless of neonatal exposure to sevoflurane (fig. 8A; t test, sevoflurane-treated male vs. control male, t = 0.37, P > 0.05). But we observed that some of sevoflurane-treated males bit pups within a few minutes after exposure to newborn pups, whereas control males did not commit attacks (fig. 8B). Thus, nurturing behaviors were impaired in male mice similar to first-time mothers.

We recently showed that the antioxidative effects of molecular hydrogen gas suppressed neurotoxicity caused by neonatal exposure to anesthetics in the developing brain.³⁰  Hydrogen gas can be easily supplied as part of the carrier gas mixture during anesthesia. Thus, we sought to investigate whether coadministration of hydrogen gas with sevoflurane could effectively suppress impairments of maternal behaviors caused by neonatal anesthetic exposure.

The survival deficits of pups born to dams exposed to sevoflurane at P6 was prevented by coadministration of 1.3% hydrogen gas with sevoflurane (fig. 9). A log-rank analysis indicated that pup survival rate was indistinguishable between the control dams and the sevoflurane + hydrogen–treated dams (P > 0.05). Furthermore, a log-rank analysis also indicated that the pup survival rate of the sevoflurane + hydrogen–treated dams was significantly higher than that of the sevoflurane-treated dams (P < 0.0001).

The number of pups that did not have milk in their digestive tracts was indistinguishable between the control group and sevoflurane + hydrogen–treated dams (fig. 10A). The ratio of poorly cleaned pups was also indistinguishable between them (fig. 10B). Furthermore, we found that the sevoflurane + hydrogen–treated dams exhibited a normal nest-building score (fig. 10C), a normal ratio of scattered pups in their home cage (fig. 10D), and a normal duration of crouching (fig. 10E) compared with those of the control dams. A one-way ANOVA followed by Bonferroni post hoc test confirmed these findings, indicating no significant differences between sevoflurane + hydrogen–treated dams and controls in these tests (fig. 10, C–E; F and P values are presented below each panel; post hoc test, P > 0.05 for each test). In the pup retrieval test, the sevoflurane + hydrogen–treated dams performed similar to the control dams (fig. 11A). A one-way ANOVA followed by Bonferroni post hoc test confirmed this, indicating no significant difference between sevoflurane + hydrogen–treated dams and controls in each retrieval (fig. 11A; F and P values are presented below each panel; post hoc test, P > 0.05 for each retrieval). We did not detect significant differences in the analysis for latencies to approach and to sniff a pup for the first time among groups (fig. 11B; one-way ANOVA, F = 0.19, P > 0.05). Olfaction abilities were also indistinguishable among groups (fig. 11C; one-way ANOVA, F = 0.008, P > 0.05). Together, it can be concluded that concomitant hydrogen inhalation significantly mitigated impairment of maternal behaviors caused by neonatal exposure to sevoflurane.

Oxytocin and vasopressin are known to be implicated in the induction of maternal behaviors.^23–26  It is possible that neonatal exposure to anesthetics might alter the release of these hormones, leading to the impairment of maternal behavior. Therefore, we sought to examine whether the plasma concentration levels of oxytocin and vasopressin could be disrupted in sevoflurane-treated dams by the use of enzyme-linked immunosorbent assay. Furthermore, to exclude the possibility that the exposure to hydrogen gas per se counteracts perturbations in hormonal levels, we also examined the plasma concentration levels of these hormones in hydrogen-treated mice. We found that there was no significant change in plasma oxytocin concentration on the day of parturition when comparing control, sevoflurane, hydrogen, and sevoflurane + hydrogen groups (fig. 12A). A two-way ANOVA confirmed this, indicating no significant main effect of sevoflurane treatment (F = 0.005, P > 0.05) and of hydrogen treatment (F = 0.012, P > 0.05). The interaction between sevoflurane and hydrogen treatment was not significant as well (F = 0.446, P > 0.05).

Similarly, there was no significant change in plasma vasopressin concentration on the day of parturition when comparing control, sevoflurane, hydrogen, and sevoflurane + hydrogen groups (fig. 12B). A two-way ANOVA confirmed this, indicating no significant main effect of sevoflurane treatment (F = 0.018, P > 0.05) and of hydrogen treatment (F = 1.194, P > 0.05). The interaction between sevoflurane and hydrogen treatment was not significant as well (F = 0.206, P > 0.05). Therefore, we concluded that neither neonatal exposure to sevoflurane nor to hydrogen altered the blood concentration levels of oxytocin and vasopressin when fostering their pups.

Modification of neurons in the MPOA in the rostral hypothalamus is required to express maternal behaviors.^19,27,28  It was reported that when a mouse takes care of pups, c-Fos is induced in the MPOA.^38,39  Thus, we set out to quantify the number of c-Fos–immunopositive cells in the MPOA from maternal dams when fostering their pups. In sevoflurane-treated dams, the number of c-Fos–immunopositive cells was significantly reduced in the MPOA when compared with control dams (fig. 13, A and B; t test, sevoflurane-treated dams vs. control dams, t = 6.92, P < 0.001). This finding suggested that neonatal exposure to sevoflurane disrupted neural mechanism to induce maternal behavior.

In this study, we showed that sevoflurane exposure in the developing brain of female mice caused serious impairment of maternal behaviors when fostering their pups although parturition was normal. Furthermore, survival rate was severely impaired in pups born to sevoflurane-treated dams. Pup exchange test showed that the survival impairment of the pups lay entirely with dams. Taken together, deficits in maternal behavior of sevoflurane-treated dams caused the impaired survival of their offspring. The deficits in maternal behavior in sevoflurane-treated virgin females were not attributed to the secondary effect of other changes such as locomotor or olfactory functions because exploratory and investigative behaviors toward pups were normal in these dams. It should be noted that all dams, irrespective of sevoflurane treatment, promptly approached pups when the pups were placed outside the nest, indicating that neonatal exposure to sevoflurane might not cause an inability to detect sensory cues emanating from pups. We did not find significant difference in pup-cleaning behavior between sevoflurane-treated dams and controls. At delivery, normal dams usually devote an inordinate amount of attention to birth material and ingest the afterbirth (placentophagia).²⁰  Because these behaviors are characteristic of maternal females, some researchers classify them as one of maternal behaviors. However, from the dam’s perspective, afterbirth materials contain attractive substances such as placental opioid-enhancing factor that potentiate the antihyperalgesic properties of endogenous opioids.²⁰  Thus, cleaning pup is not a maternal behavior in the sense of pup-directed caretaking behavior.

We found that blood concentrations of oxytocin and vasopressin were not significantly changed in sevoflurane-treated dams compared with controls when fostering their pups. However, some studies reported that neonatal exposure to general anesthetics increased proinflammatory cytokines and stress hormones shortly after the anesthesia.^40–44  Because some hormones are known to play important roles in neuronal development,^45,46  we cannot exclude the possibility that neonatal exposure to general anesthetics could cause impairment to hormone dynamics shortly after the anesthesia, which might affect neuronal development.

We found that the number of c-Fos–immunoreactive cells was significantly decreased in the MPOA from sevoflurane-treated dams when compared with controls. Accumulating evidence indicates that the MPOA in the rostral hypothalamus plays critical roles in the induction of maternal behavior.^19,27,28  Stimulation from pups converges on the MPOA, and it activates neurons in the MPOA. Then, the activated neurons induce the expression of transcription factors such as c-Fos and its homolog FosB. These transcription factors are known to be essential for the facilitation of maternal behaviors.^13,31,34  Thus, the reduced induction of c-Fos in the MPOA suggested that neonatal exposure to anesthetics impaired neuronal mechanism that plays critical role in maternal behavior.

The hippocampus plays critical roles in multiple brain functions including working memory, which is required to do complex cognitive tasks.^47,48  Hippocampus is known to be vulnerable to neonatal exposure to general anesthetics. For instance, it was reported that neonatal exposure to general anesthetics reportedly decreased the expression of postsynaptic density protein-95 in the hippocampus.⁴⁰  Because postsynaptic density protein-95 is a candidate molecule implicated in synaptic plasticity,^49,50  neonatal exposure to anesthetics might impair synaptic plasticity, which is required to modulate neuronal circuit. Thus, one might speculate that the impairment of the hippocampus was involved in the impairment of maternal behaviors in sevoflurane-treated dams. Furthermore, it was reported that lesions in the hippocampus disrupted maternal behavior in rats although the underlying mechanism was largely unknown.⁵¹  However, some studies reported that limbic structures including the hippocampus are important but not essential for maternal behavior.^52–54  One might also speculate that the defect in maternal behaviors was secondary to learning or memory deficits in sevoflurane-treated mice. However, a defect in maternal behavior is not necessarily indicative of these deficits. Indeed, there are mice that have deficits in maternal behaviors but not in learning and memory. For instance, mice lacking the immediate early gene fosB showed deficits in maternal behaviors but not in learning ability.¹³ 

How does neonatal exposure to general anesthetics cause impairment of maternal behaviors in mice? Although the molecular, cellular, and neurological mechanisms are largely unknown, accumulating evidence indicated that expression of maternal behaviors seems to require change of the neural circuit in the brain.³⁴  Generally, once virgin females are sensitized by extensive exposure to pups, maternal behaviors last for at least several days without further pup exposure. Therefore, the experience of pup exposure may be important to elicit changes in the neural circuit for maternal behaviors, which modify maternal responsiveness in a long-lasting manner. Taking this into consideration, one possible explanation for the impairment of maternal behavior in sevoflurane-treated mice is that the neural circuit indispensable for this type of behavior might be stunted in these mice. But the mothering experience at the first delivery might add an additional redundant route, via memory, to the process of activating the maternal neural circuit. Thus, second-time parous dams could foster pups irrespective of sevoflurane exposure. An alternative explanation is that stimuli emanating from newborn pups might not sufficiently activate the maternal neural circuit in the sevoflurane-treated mice: because of the failure to change the neural circuit effectively in sevoflurane-treated dams, they could not overcome the threshold to express maternal behavior. However, the partial change of the maternal neural circuit induced at the first delivery might contribute to overcome the threshold at the second delivery. In both cases, the adaptation to induce maternal behavior might depend on finely tuned neuronal mechanisms in the circuit for maternal behavior. It should be noted that the capacity to learn still remained in sevoflurane-treated mice because maternal behavior deficit was no longer apparent after the birth of the second litter. This is consistent with a report that environmental enrichment reversed anesthetic-induced memory impairments in the rat developing brain almost completely even when instituted with substantial delay.⁵⁵ 

It was reported that oxidative stress was involved in anesthetic-induced neurotoxicity in the developing brain.²⁹  Hydrogen has recently received attention as an effective antioxidant because of its small size and electrically neutral properties, enabling it to reach target organs easily, to diffuse across cell membranes rapidly, and to penetrate the blood–brain barrier for the protection of neurons.⁵⁶  We previously showed that coadministation of hydrogen gas significantly suppressed the increase in neuroapoptosis and subsequently mitigated the deficits in social behaviors as well as learning deficits caused by neonatal exposure to sevoflurane.³⁰  In the current study, we show that coadministration of hydrogen gas significantly reduced impairment of maternal behaviors caused by neonatal exposure to sevoflurane, suggesting further potential of hydrogen coadministration for therapeutic use.

Similarities in the maternal behaviors of humans and rodents suggest an existence of a general neural circuit for these behaviors. In humans, child abuse and neglect are global problems with serious life-long consequences, including increased likelihood for a wide range of mental disorders. Although there are interpretative limitations to translate animal models to humans, an understanding of the neurobiological basis for the deficits of maternal behavior caused by neonatal exposure to sevoflurane is important in psychiatric medicine and would be helpful for ensuring safer anesthesia in pediatric medicine.

In conclusion, sevoflurane exposure in the mouse developing brain causes serious impairment of maternal behaviors when the females foster their pups although parturition is normal. We previously showed that neonatal exposure to sevoflurane caused impairments of social behaviors that resemble those observed in autism.³  Together with the previous reports, it may be concluded that in an animal model, neonatal exposure to general anesthetics causes pervasive deficits in brain functioning including even an innate behavior that is essential to species survival. Further molecular neuropathological investigations are necessary to fully explain the diverse behavioral alterations caused by neonatal exposure to general anesthetics.

The authors thank Ms. Kiyoko Takamiya and Mrs. Yuko Ogura (Department of Anesthesiology, National Defense Medical College, Tokorozawa, Saitama, Japan) for excellent technical help in this study.

This work was supported by Japan Society for the Promotion of Science (JSPS; Tokyo, Japan); grant numbers 22500304, 23791734, 25861404, and 25293331.

The authors declare no competing interests.