Intergenerational Perioperative Neurocognitive Disorder in Young Adult Male Rats with Traumatic Brain Injury

All experimental procedures were approved by the University of Florida Institutional Animal Care and Use Committee (Gainesville, Florida). The study was conducted and data are reported in accordance with the Animal Research: Reporting of In Vivo Experiments guidelines.²¹  Sprague-Dawley rats (assigned generation F0) were purchased from Charles River (USA). The F0 female rats were used as breeders only to generate offspring (assigned generation F1). Rats were housed under controlled illumination (12-h light and 12-h 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, F1 litters were culled to 12 pups. At 21 days, pups were weaned and housed two per cage for the rest of the study. Experimental data in this study are from 192 male and 72 female rats. One rat in the injury group died immediately after traumatic brain injury induction, and two other rats in the surgery group were removed from the study because of suture failure during the recovery period. All three rats were excluded from all analyses.

Figure 1 shows an overview of the study design. F0 male rats were randomized into four treatment groups (n = 30 per group). Investigators were blinded to group assignments. Rats in the injury group were subjected to all interventions: (1) surgery under 3% sevoflurane anesthesia for 40 min on postnatal day 60 to conduct a craniectomy and implant an injury hub; (2) a midline fluid percussion-inflicted moderate traumatic brain injury (approximately 2.0 atm average pressure wave; approximately 9 min time from the initial impact until the rat spontaneously rights itself from a supine position)²⁰  on the same day; and (3) exposure to 2.1% sevoflurane for 3 h on postnatal days 62, 64, and 66, to model anesthesia or sedation needed for treatment of conditions associated with traumatic brain injury or unrelated ones.¹³  Rats in the sevoflurane group had only sevoflurane exposure on postnatal days 60, 62, 64, and 66. Rats in the surgery group had a craniectomy and injury hub implantation but not traumatic brain injury on postnatal day 60. They also had sevoflurane exposure on postnatal days 62, 64, and 66. Rats in the control group were placed in a new cage and housed one per cage for an equivalent amount of time on postnatal days 60, 62, 64, and 66.

A subset of F0 male rats from all groups (n = 12 per group) was euthanized 1 h after recovery from sevoflurane anesthesia on postnatal day 66 or at an equivalent timepoint in the control group to study acute effects. The remaining F0 male rats (n = 18 per group) were mated on postnatal day 90 with control female rats to produce offspring. A cohort of 18 breeding pairs of F0 male and female rats were used to produce F1 offspring for a given experimental group (18 F1 rats per group per sex). A given F1 offspring experimental group included one or two rats from a given F1 litter. Within 24 h of delivery, F1 litters were culled to 12 pups. The F0 females were housed individually throughout the entire gestation and postpartum rearing periods. At the age of 21 days, pups were weaned and housed in sex-matched pairs for the rest of the study. F1 rats were not exposed to any treatment and were subjected to animal facility rearing only.

The F0 sires and F1 male and female offspring were sequentially evaluated in the elevated plus maze,²²  for prepulse inhibition of the acoustic startle response,^22,23  and in the Morris water maze^24,25  (fig. 1). To measure corticosterone responses to stress, blood samples were collected immediately before physical restraint for 30 min and 10, 60, and 120 min after restraint in F0 sires, and 30 min after confinement for the prepulse inhibition of acoustic startle test in F1 offspring. Ten days after completing the in vivo studies, F0 and F1 rats were anesthetized and euthanized through decapitation to collect trunk blood and brain tissue samples for enzyme-linked immunosorbent assays,^15,16  reverse transcription–polymerase chain reaction,^15,16  and targeted next-generation bisulfite sequencing studies.²⁶  All methods have been described in the referenced studies^{15,16,22–26}  (see Supplemental Digital Content 1, https://links.lww.com/ALN/D18, for detailed descriptions). In response to peer review, additional experiments were conducted on hippocampal Nr3c1 methylation in F1 females and serum levels of proinflammatory cytokine interleukin-6 in both generations. In addition, the Morris water maze data were reanalyzed to estimate the time spent in each quadrant by a single group.

The primary outcomes in this study were the neuroendocrine and behavioral changes in F0 sires and in their F1 offspring. All other outcome measurements were secondary outcomes. Sample size calculations were done, assuming a range of anticipated differences in mean outcomes and SD based on background data and past experience with similar measurements in Sprague-Dawley rats.^3,8  These analyses indicated that sample sizes of at least n = 16 rats per group for behavioral studies and n = 5 rats per group for measurements in tissue samples were required to detect differences between treatment groups, with effect sizes of d ≥ 0.8, assuming an α level of 0.05. Values are reported as mean ± SD. Boxplots were used to identify outliers. No outliers were detected that were not in the plausible range of values for the outcomes; therefore, all data were maintained in analyses. One-way ANOVA was used to assess F0 data for acute serum corticosterone, interleukin-1β and interleukin-6 levels, acute ionized calcium binding adaptor 1 (microglia/macrophage-specific protein marker) expression, long-term resting corticosterone, total corticosterone concentrations before and after the restraint, interleukin-1β and interleukin-6 levels, changes in gene expression, time spent in and number of entries to the open arms and total distance traveled during the elevated plus maze test, and number of crossings over the former platform during the Morris water maze probe test. Two-way repeated-measures ANOVA with experimental groups and time as the independent variables was run to analyze changes in serum corticosterone levels before and at three timepoints after the restraint. Two-way repeated-measures ANOVA was used to analyze the F0 prepulse inhibition data, with treatment and prepulse intensity as independent variables. Two-way ANOVA with experimental groups and days of training as the independent variables was used to analyze changes in escape latencies to the escape platform during the Morris water maze test in F0 rats. Two-way ANOVA was used to analyze the time spent in each quadrant during the Morris water maze probe test in F0 rats, with the treatment and quadrant as independent variables. Two-way ANOVA with treatment and sex as the independent variables was used to assess F1 data for changes in serum corticosterone levels at rest and after the prepulse inhibition test, changes in serum interleukin-1β and interleukin-6 levels, changes in gene expression, time spent in and number of entries to the open arms and total distance traveled during the elevated plus maze test, and Morris water maze platform location crossing times. For F1 prepulse inhibition, Morris water maze escape latency and time in each quadrant, linear mixed models for repeated measures were used, with prepulse inhibition intensity, days of training, and quadrant modeled as repeated measures, respectively. These analyses account for within-subject corrections across repeated measurements. The models also included treatment and sex as main effects, as well as interaction terms. An independent t test was used to analyze methylation level at each CpG site and overall CpG sites of Nr3c1 gene. Analyses were conducted in SigmaPlot 14.0 software and SPSS v27 (IBM Corp., USA). Multiple pairwise comparisons were done with the Holm–Sidak method. P < 0.05 was considered significant.

Analyses of blood and brain tissue samples collected 1 h after the last exposure to sevoflurane or at equivalent timepoints in the control group found a main effect of treatment (or intervention), i.e., control, sevoflurane, surgery, and traumatic brain injury, on serum levels of corticosterone and the proinflammatory cytokines interleukin-1β and interleukin-6, as well as hippocampal levels of ionized calcium binding adaptor 1, a microglia- or macrophage-specific calcium-binding protein (fig. 2, A to D; Supplemental Table 1, https://links.lww.com/ALN/D19). Rats in the injury group had higher serum levels of corticosterone than the other groups, except the sevoflurane group. Inflammatory markers were greatest in injury rats, both systemically and within the brain.

Assessment of resting corticosterone levels more than 80 days after the interventions (Supplemental Table 1, https://links.lww.com/ALN/D19; fig. 3A) revealed that the injury group had higher corticosterone levels than all the other groups. There was also an interaction between treatment and the time course of corticosterone levels after restraint stress (fig. 3, B and C; Supplemental Table 1, https://links.lww.com/ALN/D19), in that the injury group had higher levels 10 min after restraint than all other groups.

In agreement with the changes in corticosterone levels, there was a main effect of treatment on messenger RNA (mRNA) levels of hypothalamic corticotropin-releasing hormone (Crh), glucocorticoid receptor (Nr3c1), and mineralocorticoid receptor (Nr3c2) (fig. 3, D, F, and H; Supplemental Table 1, https://links.lww.com/ALN/D19). When compared to controls, only rats in the injury group had lower Nr3c1 and Nr3c2 levels, whereas Crh mRNA levels were increased in the injury, surgery, and sevoflurane groups. There were also changes in mRNA levels of hippocampal Nr3c1 and Nr3c2 but not Crh (fig. 3, E, G, and I; Supplemental Table 1, https://links.lww.com/ALN/D19). The injury group had lower Nr3c1 and Nr3c2 mRNA levels than the control and sevoflurane groups, whereas the surgery group had lower levels of these receptors than the control group.

There were significant changes in serum levels of proinflammatory cytokines (fig. 3, J and K; Supplemental Table 1, https://links.lww.com/ALN/D19), which, compared to controls, were increased in the injury and surgery groups (interleukin-1β) and in the sevoflurane, injury, and surgery groups (interleukin-6). There were also main effects of treatment on Bdnf mRNA in the hypothalamus and hippocampus (fig. 3, L and M; Supplemental Table 1, https://links.lww.com/ALN/D19). The Bdnf levels were lower in the injury group than the control and sevoflurane groups (hypothalamus) and the control group (hippocampus), whereas rats in the surgery and sevoflurane groups had lower hypothalamic Bdnf mRNA levels than the control group.

Analyses of F0 rats’ behavior during the elevated plus maze test revealed main effects of treatment on both time spent in the open arms and number of open-arm entries (fig. 4, A and B; Supplemental Table 1, https://links.lww.com/ALN/D19), with no effects on total distance traveled (fig. 4C; Supplemental Table 1, https://links.lww.com/ALN/D19). The control group spent more time in the open arms versus all the other groups. Only the injury group made fewer open-arm entries than the control group.

There were also effects of treatment and prepulse intensity in the prepulse inhibition of the acoustic startle response test (fig. 4D; Supplemental Table 1, https://links.lww.com/ALN/D19). Multiple pairwise comparisons indicated that rats in the injury group exhibited impaired prepulse inhibition of startle at prepulse intensities of 3 dB, 6 dB, and 12 dB versus the control group.

Spatial learning (escape latencies) during the five days of Morris water maze training revealed no differences among treatment groups (fig. 4E; Supplemental Table 1, https://links.lww.com/ALN/D19). During recall testing, however (i.e., in the absence of an escape platform), the injury, surgery, and sevoflurane groups made fewer crossings over the escape platform location than the control group (fig. 4F; Supplemental Table 1, https://links.lww.com/ALN/D19). Furthermore, the injury group exhibited greater impairment in spatial memory by spending more time in the entrance quadrant and less time in the target quadrant (fig. 4G; Supplemental Table 1, https://links.lww.com/ALN/D19). Analysis of the time that F0 rats from each treatment group spent in each quadrant showed that rats from the control, sevoflurane, and surgery groups spent more time in the target quadrant than in any other quadrant (fig. 4H; Supplemental Table 1, https://links.lww.com/ALN/D19). Consistent with an interaction of the detrimental effects of surgery, traumatic brain injury, and sevoflurane, however, there was no difference between the times that F0 injury rats spent in the entry and target quadrants (fig. 4H; Supplemental Table 1, https://links.lww.com/ALN/D19).

Resting corticosterone in F1 offspring was unaffected by paternal treatment, though there was a main effect of sex (greater corticosterone levels in females; (fig. 5A; Supplemental Table 2, https://links.lww.com/ALN/D20). In contrast, stress-induced corticosterone release was affected by paternal treatment in a sex-dependent manner (fig. 5B; Supplemental Table 2, https://links.lww.com/ALN/D20). Specifically, 30 min after completion of the prepulse inhibition test, F1 males of sires from the injury and surgery groups had higher corticosterone levels than both F1 males offspring of control and sevoflurane sires or the respective groups of F1 females. Stress-induced corticosterone levels in F1 females were not affected by paternal treatments.

There was no main effect of paternal treatment, sex, or sex by paternal treatment interaction on Crh mRNA in hypothalamic and hippocampal tissue from nonstressed animals (fig. 5, C and D; Supplemental Table 2, https://links.lww.com/ALN/D20). In contrast, and consistent with the differential response to stress, assessment of hypothalamic Nr3c1 mRNA revealed a main effect of sex and paternal treatment by sex interaction, but no effect of paternal treatment (fig. 5E; Supplemental Table 2, https://links.lww.com/ALN/D20), whereas assessment of hippocampal Nr3c1 mRNA revealed a main effect of paternal treatment, sex, and paternal treatment by sex interaction (fig. 5F; Supplemental Table 2, https://links.lww.com/ALN/D20). Compared to F1 males of control sires, only F1 males of injury sires had reduced Nr3c1 mRNA transcripts in the hypothalamus and hippocampus. The Nr3c1 mRNA in F1 males of the injury and surgery sires (hypothalamus) and the injury, surgery, and sevoflurane sires (hippocampus) were lower than in respective groups of F1 females. Effects of paternal treatment on Nr3c2 mRNA levels in the hypothalamus were not sufficient to achieve significance (fig. 5G; Supplemental Table 2, https://links.lww.com/ALN/D20); however, there was a main effect of paternal treatment, but not sex or treatment by sex interaction on Nr3c2 mRNA levels in the hippocampus (fig. 5H; Supplemental Table 2, https://links.lww.com/ALN/D20). Only F1 males of injury sires had lower hippocampal Nr3c2 mRNA levels than F1 males of control sires.

Consistent with the markers of sustained systemic inflammation in F0 sires, there was a significant between-subjects effect of paternal treatment, sex, and treatment by sex interaction on serum levels of interleukin-1β and interleukin-6 in F1 offspring (fig. 5, I and J; Supplemental Table 2, https://links.lww.com/ALN/D20). F1 males of injury sires had higher serum interleukin-1β levels than F1 males of sires from all the other groups or F1 females in the respective groups. Also, F1 males of surgery sires had higher interleukin-1β levels than male rats of control sires (fig. 5I; Supplemental Table 2, https://links.lww.com/ALN/D20). Serum levels of interleukin-6 were increased in F1 male offspring of sires from sevoflurane, injury, and surgery groups (fig. 5J; Supplemental Table 2, https://links.lww.com/ALN/D20). Levels of interleukin-1β and interleukin-6 in F1 females were not affected by paternal treatment.

There were main effects of paternal treatment, sex, and paternal treatment by sex interaction on the Bdnf mRNA transcripts in both hypothalamus and hippocampus of F1 offspring (fig. 5, K and L; Supplemental Table 2, https://links.lww.com/ALN/D20). F1 males of injury and surgery sires (hypothalamus) and the injury, surgery, and sevoflurane sires (hippocampus) had lower Bdnf mRNA levels than F1 males of control sires. F1 males of injury, surgery, and sevoflurane sires (hypothalamus) and injury and surgery sires (hippocampus) had lower Bdnf mRNA levels than the respective groups of F1 females. The latter were not affected by paternal treatments.

In the elevated plus maze, there were main effects of paternal treatment and sex, but no paternal treatment by sex interaction on percent time in the open arms (fig. 6A; Supplemental Table 2, https://links.lww.com/ALN/D20). F1 males of injury and surgery sires spent less time in the open arms than F1 males of control sires or respective groups of F1 females. The number of open-arm entries and the total distance traveled were not affected (fig. 6, B and C; Supplemental Table 2, https://links.lww.com/ALN/D20).

There were overall differences for paternal treatment, sex, and prepulse intensity for prepulse inhibition of startle in F1 rats (fig. 6D; Supplemental Table 2, https://links.lww.com/ALN/D20). F1 males from injury sires were significantly impaired compared to male offspring of control sires at prepulse intensities of 3 dB, 6 dB, and 12 dB and to the male offspring of surgery sires at a prepulse intensity of 3 dB. There was no significant treatment effect in F1 females.

Paternal treatment did not affect escape latencies during the 5-day training period (fig. 6E; Supplemental Table 2, https://links.lww.com/ALN/D20) or number of crossings over the former escape platform (fig. 6F; Supplemental Table 2, https://links.lww.com/ALN/D20) during the Morris water maze test. There was a within-subjects effect of quadrant, but not sex or paternal treatment, on the time spent in the four quadrants of the maze (fig. 6G; Supplemental Table 2, https://links.lww.com/ALN/D20). Male offspring of control sires spent less time in the entrance quadrant and longer in the target quadrant than F1 males of sires from all the other groups. The analysis of the time that F1 rats from a specific treatment group spent in each quadrant showed that in contrast to F0 sires, only F1 male offspring of control sires spent more time in the target quadrant than in any other quadrant (fig. 6H; Supplemental Table 2, https://links.lww.com/ALN/D20). F1 males from sevoflurane and surgery sires spent similar time in the entry and target quadrants, while F1 male offspring of injury sires spent similar time in all four quadrants (fig. 6H; Supplemental Table 2, https://links.lww.com/ALN/D20). Although F1 female offspring of control, surgery, and injury sires spent more time in the target quadrant than in any other quadrant, the time that F1 female offspring of sevoflurane sires spent in the target quadrant was not sufficient to become statistically different from the time that they spent in the entry quadrant (fig. 6H; Supplemental Table 2, https://links.lww.com/ALN/D20).

To gain insight into potential mechanism(s) by which effects of paternal treatment are transmitted to F1 offspring, we used targeted next-generation bisulfite sequencing to evaluate DNA methylation levels in the Nr3c1 gene in spermatozoa of F0 injury and control sires and the hippocampus of their F1 offspring, as well as in the hippocampus of F0 injury and control sires. We found differentially methylated CpG sites in the chr18: 31,271,681–31,393,375 region of the Nr3c1 gene in F0 spermatozoa (nine hyper- and one hypomethylated), F1 male, but not F1 female hippocampus (five hyper- and one hypomethylated), and F0 hippocampus (five hyper- and one hypomethylated; fig. 7, B and C). Among hypermethylated CpG sites, 3 F1 male hippocampal and 3 F0 spermatozoa had the same genomic coordinates (CpG sites -1745, -486, and -385; fig. 7, B and C). Interestingly, two of these common hypermethylated CpG sites (CpG sites -486 and -385) were also hypermethylated in the hippocampus of injury sires. There was a significant effect of treatment on overall methylation levels of Nr3c1 promoter regions in the spermatozoa and hippocampus of F0 sires and the hippocampus of F1 males, but not F1 females (fig. 7D), in that methylation of 29 CpG sites in the Nr3c1 promoter was increased in injury sires and their male offspring compared to their respective control groups.

A strong cognitive reserve among young adults is a likely reason that the possibility of perioperative neurocognitive disorder has received less attention in patients from this age group. The findings of this study demonstrate that in young adult male rats, the effects of surgery under sevoflurane anesthesia to induce traumatic brain injury, and to an even greater degree the effects of surgery and traumatic brain injury combined, interact with the effects of subsequent repeated sevoflurane exposure to induce abnormalities in hypothalamic-pituitary-adrenal axis functioning, inflammatory markers, and some, but not all, behavioral tests. The findings of this study also demonstrate that male F1 offspring can develop the same types of abnormalities; i.e., an intergenerational perioperative neurocognitive disorder. Importantly, unexposed F1 offspring were in some cases affected even when their exposed sires did not exhibit overt deficiencies, at least in regard to spatial memory in sevoflurane F0 males and their F1 male offspring. Overlapping hypermethylated CpG sites in the Nr3c1 gene in the spermatozoa of F0 injury rats and in the hippocampus of their male but not female offspring (particularly in the proximal promoter) reduced Nr3c1 expression in the F1 male but not female hippocampus, and exacerbated glucocorticoid receptor-dependent hypothalamic-pituitary-adrenal axis responses to stress in F1 males but not females support the involvement of epigenetic mechanisms in the intergenerational transmission of adverse effects of paternal surgery, traumatic brain injury, and sevoflurane exposure. The findings that Nr3c1 was similarly hypermethylated and exhibited similarly reduced expression in the brains of F0 injury sires and that both generations exhibited similar neuroendocrine, inflammatory, and neurobehavioral abnormalities point to overlapping initiating mechanisms of the F0 somatic and germ cell effects of surgery, traumatic brain injury, and sevoflurane.

Studies in animal models have shown that parental preconception treatments with corticosterone or synthetic glucocorticoids induce epigenetic changes in spermatozoa and phenotypic alterations in offspring similar to those induced by parental preconception stress.^27–30  In addition, treatment with glucocorticoid receptor antagonists during paternal preconception exposure to stress ameliorates heritable effects of stressful experiences,²⁹  suggesting that exacerbated glucocorticoid responses are involved in initiation of heritable effects of stressful experiences. Dysregulated stress response systems and elevated corticosterone in particular are likely contributing factors in initiating heritable effects of the combination of surgery, traumatic brain injury, and sevoflurane (this study) as well as sevoflurane alone.^15,16  In support of a role for corticosterone in initiating intergenerational effects of sevoflurane, we recently demonstrated that F0 male and female rats exposed to sevoflurane had high corticosterone levels at the time of exposure, but only F0 males exhibited persistent neurobehavioral deficiencies.¹⁵  Notably, F0 sires and dams had similar epigenomic alterations in the potassium chloride cotransporter 2 gene Slc12a5 in sperm and ovarian tissue, respectively, and both passed abnormalities to F1 male offspring.¹⁵  The hypothalamic-pituitary-adrenal axis effects of sevoflurane in young adult rats can be ameliorated by pretreatments with the sodium potassium chloride cotransporter 1 inhibitor bumetanide (the authors’ unpublished observations). Studies to investigate the effects of bumetanide and glucocorticoid receptor antagonists on intergenerational perioperative neurocognitive disorder in young adult rats with traumatic brain injury are in preparation.

An important difference between the intergenerational effects of sevoflurane alone^15,16  and the combination of surgery, traumatic brain injury, and sevoflurane was that only F1 male offspring of surgery, and even more so injury, sires exhibited abnormal corticosterone responses to stress. Notably, sires from the surgery and injury groups had greater increases in levels of inflammatory markers, suggesting that the interaction between acute stress- and inflammation-like effects of sevoflurane, surgery, and traumatic brain injury may lead to greater germ cell, long-term somatic, and neuroendocrine and neurobehavioral abnormalities in F0 sires. By extension, they may also lead to greater neuroendocrine and neurobehavioral abnormalities in F1 offspring. The interaction of the effects of paternal surgery, traumatic brain injury, and sevoflurane to induce dysregulation of hypothalamic-pituitary-adrenal axis functioning in F0 sires and their F1 male offspring was especially evident in the reduced expression of both hypothalamic and hippocampal corticoid receptor genes. Corticoid receptors are involved in feedback regulation of corticosterone levels within the hypothalamic-pituitary-adrenal axis.³¹  Dysregulated stress response systems and inflammation play key roles in the pathogenesis of neurodegenerative or neurocognitive disorders and aging.^32–35  Therefore, these findings further support the possibility that young adults with traumatic brain injury are at a greater risk of developing accelerated neurocognitive decline after surgery or anesthesia, and that such a risk can be inherited by their male offspring.

Because of persistent stress-like effects of sevoflurane alone^15,16  and even stronger persistent stress- and inflammation-like effects of surgery, traumatic brain injury, and sevoflurane, we speculate that surgery or sedation do not need to be concurrent with traumatic brain injury to result in more severe intergenerational perioperative neurocognitive disorder. In other words, patients who experience traumatic brain injury may develop perioperative neurocognitive disorder because of later distant exposure to surgery or sedation. It will be important to test this possibility by investigating intergenerational perioperative neurocognitive disorder in patients with traumatic brain injury subjected to surgery or prolonged sedation across different time intervals.

Many additional questions still need to be addressed in future studies that more closely model clinical settings, e.g., the roles of number, duration, and depth of paternal anesthesia and the delay between surgery, traumatic brain injury, and anesthesia exposure and mating. The 3-h sevoflurane regimen on three alternating days employed in this study may be long for the most common clinical cases, but it is still applicable to many patients.^36–38  More recent examples of long exposure to general anesthetics include patients with COVID-19 in the intensive care unit who may require general anesthesia levels of sedation for more than 20 days.³⁹  It will be translationally important to elucidate effects of traumatic brain injury of different severity, as well as other models of traumatic brain injury including those that do not involve surgery. It will further be important in future work to investigate a broader range of genes (beyond Nr3c1) that could be involved in intergenerational transmission of the effects of surgery, traumatic brain injury, and sevoflurane.

In this exploratory study, we have chosen to start testing the interaction hypothesis of intergenerational effects of traumatic brain injury or surgery or sevoflurane, in general, and of the role of epigenomic transmission in such effects, in particular, in F0 males, because the sires’ contribution to offspring development in rodents is largely restricted to passing spermatozoa containing genomic and epigenomic information. Testing of this hypothesis in F0 females, which will be done in future studies, will require embryo or ovary transfer to a nonexposed female and fostering of neonates by a nonexposed surrogate mother. This will allow investigation of the role of epigenetic modifications in the maternal germline in heritable effects of traumatic brain injury or surgery or sevoflurane, while excluding potential effects of altered maternal physiology and behavior on the offsprings’ development. Investigations of both maternal and paternal intergenerational perioperative neurocognitive disorders are not only equally important but, additionally, may help to elucidate mechanisms of heritable effects of parental experiences, as intergenerational effects of stress and general anesthetics may be parent-dependent. For example, studies in humans show that distinct, even opposite, changes in methylation of the NR3C1 gene in offspring occur depending on whether they are initiated by paternal or maternal stressful experiences that lead to posttraumatic stress disorder.⁴⁰  On the other hand, we have recently reported that F0 female rats exposed to sevoflurane in young adulthood are less affected neurobehaviorally than F0 males, but F1 male offspring of both sevoflurane-exposed F0 males and females exhibit neurobehavioral deficiencies.¹⁵  Therefore, it is a possibility that the dams with traumatic brain injury exposed to surgery or general anesthesia might have hypermethylated Nr3c1 genes in ovary and act as asymptomatic “silent carriers” of perioperative neurocognitive disorder to offspring.

In the current study, abnormalities in all measured parameters were found almost exclusively in F1 males. These findings closely resemble sex-specific intergenerational effects induced by young adult or neonatal exposure to sevoflurane alone,^15,16  suggesting similar initiating mechanisms of intergenerational effects of parental exposure to sevoflurane alone and surgery or traumatic brain injury or sevoflurane. Future studies will be needed to elucidate such mediating mechanisms.

The confounding effects of social, cultural, educational, and behavioral factors are among the reasons that the heritability of adverse effects of surgery, traumatic brain injury, or general anesthetics has not been evaluated in humans. Nevertheless, the persistent and heritable adverse effects of stress in specific circumstances, for instance, when war or famine affected large groups of people in relatively compact living areas within a defined period of time, have been extensively studied.^40–42  The findings of such human studies along with the findings of stress- and inflammation-like intergenerational effects of surgery or traumatic brain injury or sevoflurane in young adult male rats in this study, which had the advantage of strictly controlled experimental conditions, draw attention to the need for understanding and management of perioperative neurocognitive disorders in millions of young adult patients with traumatic brain injury. Considering that persistent inflammation and excessive stress are associated with accelerated cognitive decline in later life,^32–35  our findings suggest that surgery/traumatic brain injury/sevoflurane-induced perioperative neurocognitive disorder in the F0 sires and their F1 male offspring might worsen as life progresses. These findings also highlight the need to investigate the dynamics of intergenerational perioperative neurocognitive disorder in human patients with traumatic brain injury across the life span of several generations.

Supported in part by the National Institutes of Health (Bethesda, Maryland; R56HD102898 and R01HD107722 to Dr. Martynyuk), the I. Heermann Anesthesia Foundation (Newberry, Florida; L.-S. Ju), and the Jerome H. Modell, M.D., F.A.H.A., Endowed Professorship (University of Florida, Gainesville, Florida; Dr. Gravenstein).

Dr. Morey owns equity in Xhale, Inc. (Gainesville, Florida) In addition, the University of Florida owns equity in Xhale, Inc., a faculty start-up company producing alar pulse oximeters for clinical use in humans. Dr. Gravenstein serves as a medical advisor for Teleflex Medical (Wayne, Pennsylvania). The other authors declare no competing interests.

Supplemental Text 1. Experimental Methods, https://links.lww.com/ALN/D18

Supplemental Table 1. Results of statistical analyses of F0 rat data, https://links.lww.com/ALN/D19

Supplemental Table 2. Results of statistical analyses of F1 rat data, https://links.lww.com/ALN/D20