Intergenerational Effects of Sevoflurane in Young Adult Rats

The U.S. Food and Drug Administration recommends avoiding anesthesia in children under 3 yr old whenever possible, as it may negatively affect their brain development.¹  This recommendation is based largely on an earlier body of work in animals, particularly rodents, showing that neonates are especially susceptible to lasting adverse effects of anesthesia during the first 2 postnatal weeks.²  In rodents, the first 2 weeks of life are characterized by unique physiologic properties, such as excitatory signaling through γ-aminobutyric acid type A (GABA_A) receptors,^3,4  a major substrate for the inhibitory (sedative) effects of γ-aminobutyric acid–ergic (GABAergic) anesthetics in the more mature brain.^5–7  GABA_A receptor signaling is excitatory during early life due to elevated levels of intraneuronal Cl⁻, maintained by relatively low and high levels of the potassium ion-chloride ion cotransporter KCC2 and sodium ion-potassium ion-chloride ion cotransporter NKCC1, respectively.^3,4  During the second postnatal week in rats, GABA_A receptor signaling gradually becomes inhibitory, primarily due to age-dependent increases in KCC2.^3,4  The magnitude of GABA_A receptor excitatory signaling and the proper timing of its transition from excitatory to inhibitory are key for normal brain development, and impairments in KCC2 and shifts in GABA_A receptor signaling toward excitation have been linked to neuropsychiatric disorders in both humans and animal models.^8,9  It appears that exposure to GABAergic anesthetics early in life may induce developmental abnormalities at least in part by enhancing the magnitude of GABA_A receptor excitatory signaling at the time of anesthesia, and by disrupting normal developmental increase in Kcc2 expression.^10–12  In support of this contention, exposure of neonatal rats to GABAergic anesthetics, such as sevoflurane, causes electroencephalography-detectable epileptic seizures and multifold increases in corticosterone secretion at the time of anesthesia.¹²  After maturing to adulthood, these anesthetic-exposed rats have downregulated Kcc2 levels, exacerbated hypothalamic-pituitary-adrenal axis responses to stress, and behavioral abnormalities.^10,11  Inhibition of NKCC1 at the time of anesthesia ameliorates acute seizure activities, as well as many of the lasting developmental effects of GABAergic anesthetics,^11,12  suggesting that anesthetic-exacerbated GABA_A receptor–mediated excitation in neonatal brain is an important initiating step in anesthetic-induced neurobehavioral abnormalities in the exposed animals.

Surprisingly, neonatal parental sevoflurane also impairs Kcc2 expression in the brains of future male offspring by initiating DNA methylation at the 5´ position of cytosine residues adjacent to guanines in the Kcc2 gene in parental gamete cells.¹⁰  These findings suggest that neonatal sevoflurane can induce two distinct types of adverse effects in the exposed animals—excitatory GABA_A receptor signaling–dependent effects in the exposed neonatal brain and epigenetic reprogramming of their germ cells, which can pass adverse effects of neonatal parental sevoflurane to future unexposed offspring.¹⁰  Epidemiologic studies and research in animal models suggest that parental germ cells can be susceptible to reprogramming by environmental factors across the lifespan.^13–15  Therefore, here we tested the hypothesis that sevoflurane can affect future offspring even when administered to adult rats. We further hypothesized that the primarily inhibitory GABA_A receptor signaling in the brain of young adult rats^3,16  renders them resistant to the adverse neuroendocrine and behavioral effects of sevoflurane.

All experimental procedures were approved by the University of Florida Institutional Animal Care and Use Committee (Gainesville, Florida). Sprague-Dawley rats were bred at the University of Florida animal care facility. The rats were housed under controlled illumination (12-h light–dark, lights on at 7:00 am) and temperature (23 to 24°C) with free access to food and water. Within 24 h of delivery, litters were culled to 12 pups. At the age of 21 days, pups were weaned and housed in sex-matched pairs for the remainder of the study.

Male and female rats (generation F0) were randomized for treatment and breeding pairs using a randomization plan generator (http://www.randomization.com/), and the investigators were blind to group assignments. The F0 rats underwent anesthesia on postnatal days 56, 58, and 60. The rats were held in a temperature-controlled chamber to maintain body temperature at +37°C with a continuous supply of 30% oxygen in air (1.5 l/min) during anesthesia, which was induced by 6% sevoflurane for 3 min followed by 2.1% sevoflurane for 177 min for anesthesia maintenance (the Sevoflurane group). Gas monitoring was performed using a calibrated Datex side stream analyzer (Datex-Ohmeda, Finland), which sampled from the animal chamber interior. Rats in the F0 control group (Control group) were not subjected to anesthesia. A subset of the rats in the F0 Control and Sevoflurane groups was sacrificed 1 h after the last exposure to sevoflurane (or equivalent timepoint in the Control group) to collect blood and brain tissue samples to measure serum levels of corticosterone and to evaluate expression of KCC2 in the paraventricular nucleus of the hypothalamus (see fig. 1 for schematic of experimental design). The remaining Control and Sevoflurane F0 male and female rats were mated on postnatal day 85 to produce offspring (generation F1). Only rats from different litters were mated. F1 rats were categorized as the offspring of (1) control males and control females; (2) exposed males and control females; (3) control males and exposed females; and (4) exposed males and exposed females. The females were housed individually throughout the entire gestation and postpartum rearing periods.

Sixty-four F0 rats (32 control and 32 sevoflurane-exposed) were used as breeders to produce 122 F1 rats, which were not exposed to sevoflurane anesthesia and were subjected to animal facility rearing only. The F0 rats were sequentially evaluated in the elevated plus maze starting on postnatal day ~125 and prepulse inhibition of the acoustic startle response on postnatal day ~135 (fig. 1). A subset of these animals was sacrificed on postnatal day ~160 to collect trunk blood, brain, and testis and ovarian tissues for further analyses. The rest of these animals were physically restrained for 30 min on postnatal day ~160 to measure corticosterone responses, followed by collection of tissue samples for further analyses. The F1 rats were evaluated in the elevated plus maze starting on postnatal day ~60 and prepulse inhibition of startle on postnatal day ~70 (fig. 1). One half of these animals were sacrificed on postnatal day ~95 to collect trunk blood and brain tissue samples for further analyses, and the rest of these animals were tested for the corticosterone responses to physical restraint for 30 min on postnatal day ~95 before collecting the brain tissue samples.

The elevated plus maze studies were performed using an elevated plus maze apparatus and BIO-EPM 3C video tracking software (EB Instruments, USA) during the light phase of the dark–light cycle as previously described by our laboratory.^10,12,17  If a fall occurred, the animal was removed from the study (one F0 male rat was removed from the study for this reason).

The prepulse inhibition of startle tests were performed using a SR-Lab startle apparatus (San Diego Instruments, USA) as previously described by our laboratory.^10,12,17  The percentage of prepulse inhibition of startle for each prepulse level was calculated using the following formula: % prepulse inhibition of startle = 100 x [(pulse alone) – (prepulse + pulse)]/pulse alone. Data were collected as the highest voltage during the response window amplitude.

Blood samples (~300 µl) were collected using the “tail clip” method at rest and 10, 60, and 120 min after the restraint. Physical restraint was administered using rodent holders (Kent Scientific Corporation, USA). Serum corticosterone was measured using commercial ELISA kits (Cayman Chemical Company, USA) according to the manufacturer’s instructions.

After decapitation, the trunk blood samples were collected and centrifuged at 4°C, 1,000 g (where g is the relative centrifugal force) for 15 min, and then kept at –80°C for hormone assays. The brain was removed from the skull onto ice pads. The hypothalamus was isolated by making an anterior cut at the level of the optic chiasm, a posterior coronal section anterior to the mammillary bodies, two sagittal cuts parallel to the lateral ventricles, and a dorsal horizontal cut at the level of the anterior commissure, as described previously.¹⁸  The hippocampus was isolated from the respective blocks. All brain tissue samples were placed in vials filled with RNAlater solution (Invitrogen, USA) and then stored at –80°C. Testis tissue was removed and washed with 0.9% saline before weighing. Sperm were isolated from the caudal epididymis of adult males into phosphate-buffered saline with 1% bovine serum albumin solution at 37°C using a swim-up assay. After settling for 30 min, sperm-containing supernatant was centrifuged for 5 min at 4,000 rpm. Sperm pellets were stored at –80°C. After separation from the adipose tissues, ovaries were stored at –80°C.

Hormone analyses in serum samples isolated from F0 and F1 rats were performed using commercially available kits according to the manufacturer’s instructions. Serum concentrations of follicle stimulating hormone (CSB-E06869r) and luteinizing hormone (CSB-E12654r) were quantified using ELISA kits (Cusabio Technology LLC, USA). Serum testosterone and estradiol concentrations were measured using ELISA kits (582701, Cayman Chemical Company, USA, and ES180S-100, Calbiotech, USA, respectively).

Levels of mRNA for Nkcc1 and Kcc2 in the hypothalamus and hippocampus, for corticotropin-releasing hormone, estrogen receptor α, estrogen receptor β, aromatase, gonadotropin-releasing hormone in the hypothalamus, and for glucocorticoid receptor in the hippocampus were analyzed via reverse transcription–polymerase chain reaction in a StepOnePlus Real-Time polymerase chain reaction System (Applied Biosystems, USA) as previously described by our laboratory.¹⁰  RNA was extracted from the samples using an RNeasy Plus Kit (Qiagen, USA), reverse-transcribed with a high-capacity complementary DNA reverse transcription kit (Bio-Rad Laboratories, USA), and then analyzed via reverse transcription–polymerase chain reaction. Taqman probes specific for the genes described at the beginning of this section were obtained from Applied Biosystems: corticotropin-releasing hormone (Rn01462137_m1), glucocorticoid receptor (Rn00561369_m1), estrogen receptor α (Rn01430446_m1), estrogen receptor β (Rn00562610_m1), aromatase (Rn00567222_m1), gonadotropin-releasing hormone (Rn00562754_m1), Nkcc1 (Rn00582505_m1), and Kcc2 (Rn00592624_m1). Data were normalized to glyceraldehyde-3-phosphate dehydrogenase mRNA (Rn01775763_g1). Gene expression was calculated using the “double delta Ct” method for relative quantification, and data are presented as relative fold change from that of control animals.

Genomic DNA was extracted from the sperm pellet and ovaries of adult F0 rats and from hippocampal and hypothalamic tissues of F1 rats using the DNeasy Blood and Tissue kit (Qiagen, Germany). The sodium bisulfite conversion was performed with EZ DNA Methylation kits (Zymo Research, USA) following the manufacturer’s instructions. The primers (Kcc2: forward: GATTGTAAGTGTTTTTATTATTGAGTTGTATATT; reverse: AATAAACTTTTCCCCTTTTATACCC) were designed for the bisulfite-converted DNA sequences, using previously published sequences.¹⁰  Polymerase chain reaction amplification was performed with HotStar Taq (Qiagen, Germany). Amplicons were cloned into pCR4-TOPO vector with the TOPO TA cloning kit for sequencing (Life Technologies, USA). Miniprep was performed on each positive clone using ZR Plasmid Miniprep kit (Zymo Research, USA). Sanger sequencing was done by Genewiz (USA) using M13R primers. The DNA methylation status of all 5´ position of cytosine residues adjacent to guanines sites was analyzed using Benchling Molecular Biology 2.0 Software (Benchling, USA). The bisulfite sequencing analysis of the Kcc2 DNA was performed only in F1 male progeny of control males and control females and exposed males and exposed females matings.

Rats were anesthetized with sevoflurane and transcardially perfused with saline, followed by 4% paraformaldehyde in 0.1 M phosphate buffer. The brains were collected and fixed in the 4% paraformaldehyde overnight and then dehydrated in 30% sucrose solution in phosphate-buffered saline at 4°C for 2 days. The brains were cut into 18-µm-thick coronal sections using a cryostat. After blocking with 10% normal goat serum for 1 h at room temperature, the slices were incubated with the primary antibody (rabbit anti-KCC2, 1:500; MilliporeSigma, USA) in 10% normal goat serum at 4°C overnight. After washing with PBS for 3 × 5 min, the slices were exposed to the secondary antibody, Alexa Fluor 549 goat anti-rabbit (Invitrogen, USA). Slides were then washed with phosphate-buffered saline for 3 × 5 min and incubated for 10 min at room temperature with 4´,6-diamidine-2´-phenylindole dihydrochloride (1:1,000; Sigma, USA) in phosphate-buffered saline for 10 min. The sections were mounted and covered with a coverslip after washing with phosphate-buffered saline for 3 × 5 min. A confocal microscope (IX2-DSU spinning disk confocal fluorescent microscope system, Olympus, Japan) was used to capture the fluorescent images. The immunofluorescence intensity of the KCC2-immunoreactive cells was measured using ImageJ software (National Institutes of Health, USA). Three images from the paraventricular nucleus of the hypothalamus of each animal were taken, and the average intensity values were calculated to use for statistical analysis.

Statistical analyses were conducted on raw data using SigmaPlot 14.0 software (Systat Software, Inc., USA), which automatically checks if a data set meets test criteria (Shapiro–Wilk for normality test and Brown–Forsythe for Equal Variance Test). Values are reported as mean ± SD. The primary outcome in this study was the neurobehavioral changes in offspring of rats exposed to sevoflurane in young adulthood. All other outcome measures were the secondary outcomes. One F0 male rat was removed from the elevated plus maze study because of a fall from the maze. Another F0 male rat in the control group was removed from the behavioral studies because of injury while mating with a female rat. Boxplots were used to identify outliers. No outliers were detected that were in not plausible range of values for the outcome (indicating measurement or recording error). Therefore, all values were maintained in analyses. An independent t test was used to analyze F0 data for serum corticosterone levels and Kcc2 expression 1 h after the last exposure to sevoflurane, elevated plus maze, testis weight, hormone levels for testosterone, estradiol, luteinizing hormone, and follicle-stimulating hormone and gene expression for corticotropin-releasing hormone, glucocorticoid receptor, estrogen receptor α, estrogen receptor β, aromatase, gonadotropin-releasing hormone, Nkcc1, and Kcc2. All comparisons were run as two-tailed tests. To analyze F1 data for elevated plus maze, testis weight, testosterone level, and gene expression for corticotropin-releasing hormone, glucocorticoid receptor, Nkcc1, and Kcc2, one-way ANOVA was used. A two-way repeated-measures ANOVA was used to analyze the prepulse inhibition of startle data, with the treatment and prepulse intensity as independent variables. A two-way repeated-measures ANOVA was used to analyze changes in serum corticosterone levels at rest and at three time points after the restraint, with experimental groups and time as the independent variables. To assess differences in total corticosterone concentrations, area under the curve with respect to baseline (levels of corticosterone at rest), was calculated and compared across experimental groups using one-way ANOVA. Two-way repeated measures ANOVA with treatment as between-subject factor and the 5´ position of cytosine residues adjacent to guanines site as within-subject factor was used to analyze data on the frequency methylation of 5´ position of cytosine residues adjacent to guanines sites in the Kcc2 gene promoter in F0 and F1 tissue samples. An independent t test was used to analyze DNA methylation level at all six 5´ position of cytosine residues adjacent to guanines sites. All multiple pairwise comparisons were done with the Holm–Sidak method. P ≤ 0.05 was considered statistically significant. Statistical details are presented in text and in figure legends. The sample sizes in this study were based on previous experience with the same experimental techniques.^10,11,17  This study was not designed to detect an anesthetic × sex interaction.

In the elevated plus maze test, there was a statistically significant between-subjects effect of young adult parental exposure to sevoflurane on time spent in open arms (F[3,58] = 8.514, P < 0.001; fig. 2A) in F1 males. Specifically, F1 male progeny of sevoflurane exposed fathers and unexposed mothers (P =0.040), unexposed fathers and exposed mothers (P < 0.001), and exposed fathers and exposed mothers (P < 0.001) spent less time in the open arms compared to F1 male offspring of control fathers and control mothers. Also, there were between-subjects effects of parental exposure to sevoflurane on number of crossings (F[3,58] = 4.657, P =0.006; fig. 2B) and distance traveled (F[3,58] = 6.288, P < 0.001; fig. 2C) during the elevated plus maze test. Only F1 male offspring of both exposed parents made fewer crossings (P =0.003) and traveled shorter distances (P =0.001). All groups of F1 females were similar in respect to time spent in open arms (fig. 2D), number of crossings (fig. 2E), and distance traveled (fig. 2F) during the elevated plus maze test.

There was a statistically significant effect of young adult parental sevoflurane exposure on prepulse inhibition of startle responses in F1 male rats (F[3,174] = 7.590, P < 0.001; fig. 2G). Multiple pairwise comparisons indicated that when compared to offspring of unexposed parents, only F1 males of both exposed parents (P =0.009) or those of control fathers and exposed mothers (P =0.027) exhibited reduced prepulse inhibition of startle at prepulse of 3 decibel. In contrast to F1 males, there was no treatment effect on prepulse inhibition of startle in F1 female rats (F[3,168] = 0.584, P =0.627; fig. 2H). The startle amplitudes were similar among all experimental groups of F1 male (F[3,58] = 1.991, P =0.125) and F1 female (F[3,56] = 0.514, P =0.674) rats.

Measurements of serum levels of corticosterone in the postnatal day ~95 F1 male and female rats before and after physical restraint revealed no differences among experimental groups of the same sex (F[3,84] = 0.335, P =0.800, males; fig. 2, I and J; and F[3,84] = 0.142, P =0.934, females; fig. 2, K and L).

There were similar levels of Nkcc1 mRNA in the hypothalamus of all four treatment groups of the postnatal day ~95 F1 male rats (F[3,20] = 0.928, P =0.446; fig. 3A), but there was a between-subjects effect of parental sevoflurane exposure on hypothalamic Kcc2 mRNA levels (F[3,20] = 5.636, P =0.006; fig. 3B). Specifically, F1 male progeny of sevoflurane-exposed fathers and mothers had lower levels of Kcc2 mRNA when compared to F1 male offspring of control fathers and control mothers (P =0.012). In the hypothalamus of F1 females, there were not between-subjects effects of parental sevoflurane exposure on Nkcc1 mRNA levels (F[3,19] = 0.886, P =0.466; fig. 3C) or Kcc2 mRNA levels (F[3,20] = 0.738, P =0.542; fig. 3D).

In the hippocampus of F1 males, there were no effects of parental sevoflurane exposure on Nkcc1 mRNA levels (F[3,20] = 1.357, P =0.284; fig. 3E), but there was a significant effect on Kcc2 mRNA levels (F[3,20] = 41.375, P < 0.001; fig. 3F). Only male offspring of both exposed parents (P < 0.001) and offspring of exposed fathers and control mothers (P < 0.001) had reduced hippocampal Kcc2 mRNA levels compared to offspring of two control parents. In the hippocampus of F1 females, there were not significant between-subjects effects of parental neonatal sevoflurane exposure on Nkcc1 mRNA level (F[3,20] = 0.925, P =0.447; fig. 3G) or Kcc2 mRNA level (F[3,20] = 0.524, P =0.671; fig. 3H).

Consistent with the observations of normal levels of corticosterone (fig. 2, I and J), levels of corticotropin-releasing hormone mRNA in the hypothalamus (F[3,20] = 0.519, P =0.674; fig. 3I) and glucocorticoid receptor in the hippocampus (F[3,19] = 3.293, P =0.043; fig. 3J) of postnatal day ~95 F1 males were not different among all four treatment groups. In addition, there were no significant between-subjects effects of parental sevoflurane exposure on either hypothalamic corticotropin-releasing hormone mRNA (F[3,20] = 1.225, P =0.327; fig. 3K) or hippocampal glucocorticoid receptor mRNA levels (F[3,19] = 0.860, P =0.479; fig. 3L) in F1 female rats.

Increased DNA methylation in the Kcc2 promoter in gamete cells of adult rats exposed to sevoflurane as neonates and in the hypothalamus and hippocampus of their male offspring has recently been described.¹⁰  We therefore sought to determine if similar effects would be observed in response to young adult exposure to sevoflurane. There was a significant effect of treatment (F[1,36] = 30.8, P < 0.001) and within-subjects effect of the 5´ position of cytosine residues adjacent to guanines site (F[5,36] = 20.066, P < 0.001) on the frequency of 5´ position of cytosine residues adjacent to guanines methylation in the Kcc2 promoter region in sperm of the postnatal day ~160 F0 rats (fig. 4A). There was also a statistically significant interaction between the 5´ position of cytosine residues adjacent to guanines site and treatment (F[5,36] = 10.036, P < 0.001). Pairwise multiple comparison analysis showed that young adult exposure to sevoflurane led to increased methylation in F0 male rats at 5´ position of cytosine residues adjacent to guanines site 1 (P < 0.001 vs. control) and 5´ position of cytosine residues adjacent to guanines site 2 (P =0.006 vs. control). Similarly, there was a significant effect of treatment (F[1,36] = 34.714, P < 0.001) and within-subjects effect of 5´ position of cytosine residues adjacent to guanines site (F[5,36] = 42.686, P < 0.001) on the frequency of 5´ position of cytosine residues adjacent to guanines site methylation in the Kcc2 gene promoter in ovarian tissue of F0 rats (fig. 4B). There was also a statistically significant interaction between 5´ position of cytosine residues adjacent to guanines site and treatment (F[5,36] = 12.343, P < 0.001). Pairwise multiple comparison analysis showed that young adult exposure to sevoflurane led to increased methylation in F0 female rats at the 5´ position of cytosine residues adjacent to guanines site 1 (P < 0.001 vs. control) and 5´ position of cytosine residues adjacent to guanines site 2 (P < 0.001 vs. control).

There was a between-subjects effect of parental treatment on 5´ position of cytosine residues adjacent to guanines site methylation in the promoter of Kcc2 gene in the hypothalamus of the postnatal day ~95 male offspring of both exposed parents (F[1,36] = 48.286, P < 0.001), and a within-subject effect of 5´ position of cytosine residues adjacent to guanines site (F[5,36] = 42.629, P < 0.001; fig. 4C). There was also a statistically significant interaction between 5´ position of cytosine residues adjacent to guanines site and treatment (F[5,36] = 15.886, P < 0.001). Pairwise multiple comparison analyses showed that young adult parental exposure to sevoflurane led to increased DNA methylation in the Kcc2 gene promoter in the hypothalamus of F1 male progeny of both exposed parents at 5´ position of cytosine residues adjacent to guanines site 1 (P < 0.001 vs. F1 males from the control males and control females group), 5´ position of cytosine residues adjacent to guanines site 2 (P =0.002 vs. F1 males from the control males and control females group), and 5´ position of cytosine residues adjacent to guanines site 3 (P =0.013 vs. F1 males from the control males and control females group).

There was also a between-subjects effect of parental treatment on 5´ position of cytosine residues adjacent to guanines site methylation in the promoter of Kcc2 gene in the hippocampus of male offspring of both exposed parents (F[1,36] = 21.740, P < 0.001, and within-subject effect of 5´ position of cytosine residues adjacent to guanines site (F[5,36] = 20.852, P < 0.001; fig. 4D). There was also a statistically significant interaction between 5´ position of cytosine residues adjacent to guanines site and treatment (F[5,36] = 5.268, P < 0.001). Pairwise multiple comparison analyses showed that young adult parental exposure to sevoflurane led to increased DNA methylation in the Kcc2 gene promoter in the hippocampus of F1 male progeny of both exposed parents at 5´ position of cytosine residues adjacent to guanines site 1 (P < 0.001 vs. control males and control females), 5´ position of cytosine residues adjacent to guanines site 2 (P =0.018 vs. control males and control females), and 5´ position of cytosine residues adjacent to guanines site 3 (P =0.018 vs. control males and control females).

Unexpectedly, we found that more than 2 months after exposure to sevoflurane anesthesia, F0 male rats exhibited behavioral deficiencies. They spent less time in the open arms of the elevated plus maze (t₍₂₈₎ = 2.18; P =0.038; fig. 5A), but did not differ from their control counterparts in number of crossings (t₍₂₈₎ = 1.456; P =0.157; fig. 5B) or distance traveled (t₍₂₈₎ = –1.351; P =0.188; fig. 5C). In F0 females, there was no significant effect of sevoflurane on any of these elevated plus maze parameters (fig. 5, D–F).

As in the elevated plus maze task, there was a significant effect of sevoflurane exposure on prepulse inhibition of startle in adult F0 male rats (F[1,123] = 6.765; P =0.010; fig. 5G), but not in F0 female rats (F[1,60] = 0.049; P =0.827; fig. 5H). Multiple pairwise comparisons indicated that exposure to sevoflurane led to impaired prepulse inhibition of startle responses in F0 male rats at prepulse intensity of 3 decibel (P =0.042 vs. control). Startle stimuli by themselves caused similar responses in the control and sevoflurane groups of F0 male (t₍₄₁₎ = –0.969; P =0.338) and F0 female (t₍₃₀₎ = 1.465; P =0.153) rats.

Male F0 rats had higher total corticosterone responses to physical restraint on postnatal day ~160 when compared to their control counterparts (t₍₁₄₎ = –6.209; P < 0.001; fig. 5, I and J). This increase was due to higher levels of corticosterone at 10 min (P < 0.001 vs. control) and 60 min (P =0.036 vs. control) after restraint, as serum levels of corticosterone before the restraint (P =0.736 vs. control) and 120 min (P =0.787 vs. control) postrestraint were not different in control and sevoflurane-exposed rats (fig. 5I). There was no difference in serum corticosterone levels between control and sevoflurane-exposed F0 female rats (fig. 5, K and L).

During collection of sperm for the DNA methylation studies, we incidentally found that sevoflurane-exposed F0 males had reduced testis weight (t₍₂₇₎ = 4.494; P < 0.001; fig. 5N) more than 3 months after sevoflurane exposure. Counterintuitively, their serum levels of testosterone were increased (t₍₁₄₎ = –2.839; P =0.013; fig. 5M), although their serum levels of estradiol were normal (t₍₁₄₎ = 0.703; P =0.494; fig. 5O). F0 female rats in the sevoflurane and control groups were not different with respect to serum levels of testosterone (P =0.743; fig. 5P) or estradiol (P =0.600; fig. 5Q).

Our current and recently published findings¹⁰  indicate that rats exposed to sevoflurane in young adulthood and in the early postnatal period develop comparable systemic abnormalities. Given that sevoflurane acts as a stressor in neonatal rats and initiates developmental abnormalities, at least in part, by potentiating excitatory GABA_A receptor signaling,^10–12,17  here we tested whether sevoflurane can induce a stress-like^19–22  reduction in KCC2 expression in young adult rats. To measure acute effects of sevoflurane, rats were exposed to 2.1% sevoflurane anesthesia for 3 h on postnatal day 56, postnatal day 58, and postnatal day 60. Brain tissue and trunk blood samples were collected 1 h after sevoflurane anesthesia on postnatal day 60. Consistent with the stress-like effects of sevoflurane reported in previous work,^10–12,17  the exposed male and female rats had increased serum levels of corticosterone compared to controls (t₍₆₎ = –3.313; P = 0.016, males, fig. 6A; and t₍₈₎ = –3.949; P = 0.004, females; fig. 6B). Despite similar increases in corticosterone levels in sevoflurane-exposed males and females, immunofluorescence evaluations of KCC2 expression in the paraventricular nucleus of the hypothalamus region found reductions in KCC2 level in male (t₍₆₎ = 3.343; P = 0.016), but not female (t₍₆₎ = 0.773; P = 0.469), F0 rats (fig. 6, C–F).

Acute KCC2 and long-term systemic effects of young adult sevoflurane in the exposed males only suggest an involvement of sex steroids. To test whether sevoflurane-induced increases in serum testosterone levels involve changes in functioning of the entire hypothalamic-pituitary-gonadal axis, we measured the expression of gonadotropin-releasing hormone in the hypothalamus and serum levels of follicle-stimulating hormone and luteinizing hormone. Consistent with the finding that sevoflurane increased systemic levels of testosterone in postnatal day ~160 F0 male rats, these same rats also had increased hypothalamic levels of gonadotropin-releasing hormone mRNA (t₍₁₀₎ = –2.519; P =0.030; fig. 7A) and serum levels of luteinizing hormone (t₍₁₄₎ = –4.932; P < 0.001; fig. 7B), while serum levels of follicle-stimulating hormone (t₍₁₀₎ = 1.026; P =0.329; fig. 7C) were not different from those in F0 control male rats. There were no treatment effects on the hypothalamic levels of gonadotropin-releasing hormone mRNA (fig. 7D) or serum levels of luteinizing hormone (fig. 7E) and follicle-stimulating hormone (fig. 7F) in F0 female rats.

The elevated testosterone may modulate the hypothalamic-pituitary-adrenal axis responses to stress through estrogen receptors after testosterone aromatization to estradiol in the brain. One such mechanism includes modulation by estradiol of the glucocorticoid receptor–mediated negative feedback effect of corticosterone on the hypothalamic-pituitary-adrenal axis activity. Estradiol can both inhibit and enhance the negative feedback effects of glucocorticoids by activating estrogen receptor α and estrogen receptor β, respectively.^23,24  In further support of involvement of testosterone and estradiol in exacerbated hypothalamic-pituitary-adrenal axis responses to stress, we found that the hypothalamic levels of aromatase mRNA were increased in F0 sevoflurane-exposed males (t₍₁₀₎ = –4.333; P =0.002; fig. 7G). In addition, the F0 male rats exposed to sevoflurane had increased and decreased hypothalamic levels of estrogen receptor α mRNA (t₍₁₀₎ = –5.144; P < 0.001; fig. 7H) and estrogen receptor β mRNA (t₍₁₀₎ = 3.156; P =0.010; fig. 7I), respectively. Again, consistent with the normal hypothalamic-pituitary-adrenal axis responses to stress in the exposed F0 females, the increase in the hypothalamic aromatase mRNA in sevoflurane-exposed F0 female rats did not achieve statistical significance (t₍₁₀₎ = –1.878; P =0.090; fig. 7J). Moreover, in the F0 females, the hypothalamic levels of estrogen receptor α mRNA (fig. 7K) were not different between Sevoflurane and Control groups, while those of estrogen receptor β mRNA were slightly, though significantly increased in the Sevoflurane group (t₍₁₀₎ = –2.521; P =0.030; fig. 7L) when compared to the Control group.

In agreement with the exacerbated corticosterone responses to stress in sevoflurane-exposed F0 male rats at postnatal day ~160, these rats had increased hypothalamic corticotropin-releasing hormone mRNA levels 2 h after the restraint (t₍₁₄₎ = –3.181; P =0.007; fig. 7M), as well as reduced levels of glucocorticoid receptor mRNA in the hippocampus (t₍₁₀₎ = 2.493; P =0.032; fig. 7N). Neither hypothalamic corticotropin-releasing hormone mRNA levels (fig. 7O) nor hippocampal glucocorticoid receptor mRNA levels (fig. 7P) were different in sevoflurane-exposed and control F0 female rats.

Finally, considering the controlling roles of GABA_A receptor signaling in hypothalamic gonadotropin-releasing hormone and corticotropin-releasing hormone neuronal activity, it is plausible that sevoflurane initiates enhancements of gonadotropin-releasing hormone and corticotropin-releasing hormone neuronal activity and, hence, the entire reproductive and stress axes, by downregulating Kcc2 expression and rendering GABA_A receptor signaling less inhibitory or even excitatory. To assess whether exposure of young adult rats to sevoflurane led to persistent alterations in expression of Cl^– transporters, brain hypothalamic and hippocampal tissue samples were collected more than 3 months after exposure (postnatal day ~160). The F0 male rats from the Sevoflurane group had normal Nkcc1 mRNA levels (t₍₁₀₎ = 1.065; P =0.312; fig 8A), but decreased Kcc2 mRNA levels (t₍₁₀₎ = 2.273; P =0.046; fig. 8B) in the hypothalamus. In contrast, the F0 female rats from the Sevoflurane group had unaltered levels of both Nkcc1 mRNA (t₍₁₀₎ = 0.155; P =0.880; fig. 8C) and Kcc2 mRNA (t₍₁₀₎ = 1.346; P =0.208; fig. 8D) in the hypothalamus. In the hippocampus of F0 male rats from the Sevoflurane group, Kcc2 mRNA levels, but not Nkcc1 mRNA levels, were significantly reduced (t₍₁₀₎ = 2.387, P =0.038, fig. 8, E and F). The hippocampal mRNA levels for Nkcc1 and Kcc2 were similar in control and sevoflurane-exposed F0 female rats (fig. 8, G and H).

The results of these experiments show that sevoflurane administered to young adult rats of either sex can induce abnormalities in their male offspring, as well as in the exposed male rats themselves. The reduction in Kcc2 expression in the hypothalamus and hippocampus of the F1 male progeny of the exposed parents, as well as hypermethylation of the Kcc2 promoter in F0 male and female gamete cells and F1 male hypothalamus and hippocampus, support the involvement of epigenetic mechanisms in transmitting adverse effects of sevoflurane exposure in young adult rats to the next generation. The findings that both the physiology and behavior of sevoflurane-exposed F0 females were normal, while their ovarian Kcc2 promoter was hypermethylated and male progeny of exposed F0 females exhibited neurobehavioral abnormalities, suggest that sevoflurane affects somatic and germ cells via different mechanisms.

Human and animal studies provide evidence that exposure to alcohol, endocrine disruptors, and stress can affect future generations.^{10,12–15,25–27}  Surprisingly, the heritable effects of general anesthetics are poorly studied, despite the fact that anesthetic agents share molecular mechanisms of action with alcohol and may act as both environmental stressors and endocrine disruptors.^{10–12,17,27–29}  Anecdotal observations point to the possibility of heritable effects of anesthesia and surgery in humans.³⁰  Some evidence of heritable effects of general anesthetics comes from clinical surveys indicating that anesthesia care providers may have altered female:male offspring ratios.^31–35  Importantly, sex ratio effects have been linked to alcohol, stress, and endocrine disruptors as well.^36–38  Our recently published¹⁰  and current findings support the possibility that sevoflurane can induce intergenerational effects. Obviously, sevoflurane-induced epigenetic modulation of Kcc2 in F0 gamete cells and F1 male brain may contribute to the intergenerational effects of young adult sevoflurane exposure, but it is unlikely that sevoflurane-induced changes in Kcc2 are the only mediating mechanism. Indeed, this study found that male progeny of exposed dams and control sires had normal Kcc2 mRNA levels in the hypothalamus and hippocampus but exhibited deficiencies in the elevated plus maze and prepulse inhibition of startle behavioral tests. It is possible that Kcc2 DNA methylation in gamete cells affects other epigenetic mechanisms that program offspring brain development, and/or that multiple genes are independently epigenetically modified by sevoflurane in F0 gamete cells and F1 brain. Furthermore, the epigenetic effects of GABAergic anesthetics may not be limited to DNA methylation, as we and others recently reported experimental evidence for involvement of not only DNA methylation,^10,39  but also histone acetylation,^40,41  in adverse effects of neonatal sevoflurane exposure in rats.

Despite different anesthesia regimens in our recent study of intergenerational effects of neonatal anesthesia¹⁰  and the young adult anesthesia in the current study, similarities in the intergenerational effects of sevoflurane outweighed differences between the two models. These findings are also consistent with those reported by Rodgers et al. that 42-day exposure of male rats to stress either throughout puberty or in adulthood leads to similar blunted hypothalamic-pituitary-adrenal axis responses in their progeny.¹⁴  Moreover, repeated exposure of 11-week-old male mice to the anesthetic agent enflurane negatively affected their offspring’s behavior.⁴²  The susceptibility of germ cell maturation to sevoflurane over an extended period of the lifespan, from the early postnatal period¹⁰  through young adulthood (this study), suggests that the gamete cells at different stages of maturation may be susceptible to epigenetic modifications initiated by sevoflurane. It will be important in future studies to analyze epigenetic effects of parental exposure to anesthetics in F0 oocytes, which represent only a minor part of total ovarian tissue that was tested in this study.

Irrespective of whether fathers, mothers, or both parents were exposed to sevoflurane as young adults, only male F1 progeny exhibited epigenetic, gene expression, and systemic abnormalities. Given that Kcc2 was hypermethylated in both male and female gamete cells in F0 rats, it is plausible that parental sevoflurane initiates mechanisms leading to disruption of embryonic DNA methylation reprogramming in F1 males, but not in F1 females, resulting in affected F1 males but normal F1 females. Our findings that the exposed but physiologically unaffected dams, similar to exposed and affected sires, pass deleterious effects of sevoflurane to unexposed male offspring raise the possibility that male offspring may be affected even when anesthesia level or duration is not sufficient to induce significant abnormalities in the exposed parents.

The long-term adverse effects of general anesthesia in early childhood and at advanced ages are a widely recognized health-related concern and the subject of extensive clinical and laboratory research.⁴³  Investigations of such effects in young adults in their prime reproductive age, however, are relatively scarce. Several studies have assessed effects of isoflurane in young adult rats, primarily using young adult rats as comparisons to other age groups. Isoflurane administered to postnatal day 60 rats affected progenitor proliferation and improved spatial memory in one study and had no effect in other one.^44,45  Crosby et al. also observed improvement in spatial memory in rats anesthetized with 1.2% isoflurane–70% nitrous oxide at 6 months of age.⁴⁶  Aside from the fact that two studies found long-term effects of isoflurane in young adult rats, different isoflurane concentrations and exposure regimens make it difficult to compare the effects across these studies. Such comparison is even more problematic with our current findings in F0 males, given that we tested not only a distinct anesthesia regimen, but also a different anesthetic agent. Clearly, further research is needed to elucidate the full range of long-term effects of young adult sevoflurane anesthesia and mechanisms that mediate such effects.

The GABA_A receptor/testosterone/aromatase/estradiol/KCC2 pathway may be a key mediator of sevoflurane’s long-term neuroendocrine effects in F0 male rats. Because many gonadotropin-releasing hormone neurons are excited by GABA_A receptor signaling even under basal conditions,^47,48  sevoflurane may initially stimulate gonadotropin-releasing hormone neuronal activity and, hence, the entire reproductive axis. Male-specific factors may be required for sevoflurane’s actions to acutely reduce KCC2 expression, as similar increases in corticosterone levels in sevoflurane-exposed female rats were not accompanied by a reduction in KCC2 expression. The GABA_A receptor/testosterone/aromatase/estradiol/KCC2 pathway in F0 male rats may function as a system with a positive feedback effect leading to a persistently upregulated hypothalamic-pituitary-gonadal axis. Sevoflurane-induced increases in systemic testosterone and in brain aromatase expression may lead to reductions in Kcc2 expression, and in turn to diminished inhibitory or increased stimulatory control of gonadotropin-releasing hormone neurons by KCC2/GABA_A receptor signaling. The plausibility of this scenario is supported by literature demonstrating that estradiol increases expression of aromatase and decreases expression of KCC2 in the brain.^49,50  Interestingly, consistent with our findings, Galanopoulou and Moshé found that estradiol reduced KCC2 expression in males only.⁵⁰ 

The GABA_A receptor/testosterone/aromatase/estradiol/KCC2 pathway is also likely to be involved in dysregulated (exacerbated) hypothalamic-pituitary-adrenal axis responses to stress and stress-dependent behavioral abnormalities. One such mechanism includes modulation by estradiol of the glucocorticoid receptor–mediated negative feedback effect of corticosterone on hypothalamic-pituitary-adrenal axis activity. Estradiol can both inhibit and enhance the negative feedback effects of glucocorticoids by activating estrogen receptor α and estrogen receptor β, respectively.^23,24  In support of this mechanism are our findings that hypothalamic levels of aromatase and estrogen receptor α mRNA were elevated, whereas hypothalamic levels of estrogen receptor β mRNA and hippocampal levels of glucocorticoid receptor mRNA were downregulated. Higher estradiol concentrations are thought to contribute to higher corticosterone at rest or after stress exposure in adult female rats when compared to their male counterparts.^51–53  Hence, our findings of exacerbated corticosterone responses to stress in F0 males, but not F0 females, suggest that exposure to sevoflurane in young adulthood induces a persistent transformation of the male stress response to a more “female-like” form. Upregulated expression of hypothalamic aromatase and resulting elevated levels of estradiol may also explain why in sevoflurane-exposed males, elevated levels of testosterone were accompanied by exacerbated hypothalamic-pituitary-adrenal axis responses to stress, because it was previously shown that gonadectomy of male rats elevated, while androgen replacement blunted the corticosterone response to stress.⁵⁴  The persistent downregulation of Kcc2 expression and resulting GABA_A receptor depolarizing and stimulatory signaling in the hypothalamic paraventricular nucleus are likely to further contribute to dysregulated stress responses and behavioral abnormalities by impairing fundamental mechanisms of hypothalamic-pituitary-adrenal axis functioning; e.g., the neuroactive steroid and GABA_A receptor negative feedback–based mechanism of desensitization to stress.⁵⁵  Finally, it is plausible that the negative feedback effect of elevated testosterone contributed to the reduced testis weight in exposed F0 males.

In conclusion, our results demonstrate that parental exposure to sevoflurane in young adulthood epigenetically reprograms germ cells, leading to neurobehavioral abnormalities in adult male progeny. These findings also provide evidence that sevoflurane administered even in young adulthood induces neurobehavioral deficits, profound alterations in the hypothalamic-pituitary-gonadal axis, and dysregulated hypothalamic-pituitary-adrenal axis responses to stress in exposed male rats. Exposed young adult female rats exhibited no long-term physiologic or behavioral abnormalities, but together with the exposed males, passed the adverse effects of sevoflurane exposure to their unexposed male offspring. These differential effects of sevoflurane exposure in males and females could suggest that distinct mechanisms mediate the somatic and germ cell effects of young adult sevoflurane exposure. It is important to note, however, that the current study was not powered to detect sex differences, and hence it will be important in future work to conduct direct comparisons between males and females.

The authors would like to thank Terrie Vasilopoulos, Ph.D., Department of Anesthesiology, University of Florida College of Medicine, Gainesville, Florida, for her advice and help with the statistical analyses.

Supported by the National Institutes of Health, Bethesda, Maryland (grant Nos. R01NS091542 and R01NS091542-S, to Dr. Martynyuk), the Escher Autism Fund (San Jose, California; to Dr. Martynyuk), the I. Heermann Anesthesia Foundation (Gainesville, Florida; to Dr. Ju), and the Jerome H. Modell, M.D., F.A.H.A., Endowed Professorship, Gainesville, Florida (to Dr. Gravenstein).

The authors declare no competing interests.