Voluntary Exercise Rescues the Spatial Memory Deficit Associated With Early Life Isoflurane Exposure in Male Rats : Anesthesia & Analgesia

Secondary Logo

Journal Logo

Neuroscience and Neuroanesthesiology: Original Laboratory Research Report

Voluntary Exercise Rescues the Spatial Memory Deficit Associated With Early Life Isoflurane Exposure in Male Rats

Chinn, Gregory A. MD, PhD; Sasaki Russell, Jennifer M. PhD; Banh, Esther T. BS; Lee, Saehee C. BA; Sall, Jeffrey W. PhD, MD

Author Information
doi: 10.1213/ANE.0000000000004418
  • Free



  • Question: Is exercise sufficient to rescue the cognitive deficit associated with early life anesthesia exposure in rats?
  • Findings: Exercise can rescue a spatial memory deficit from perinatal isoflurane exposure and is associated with an increase in proliferating cells in the adult hippocampus.
  • Meaning: While previous work has shown the deficit caused by early life anesthesia can be rescued with environmental enrichment, this study specifically demonstrates that postweaning voluntary exercise is sufficient to rescue this phenotype.


ANOVA = analysis of variance; ARRIVE = Animal Research: Reporting of In Vivo Experiments; BDNF = brain-derived neurotrophic factor; BrdU = bromodeoxyuridine; CO2 = carbon dioxide; Con = control; Ct = Cycle threshold; EE = environmental enrichment; Fio2 = fraction of inspired oxygen; GAPDH = glyceraldehyde 3-phosphate dehydrogenase; Iso = isoflurane; MAC = minimum alveolar concentration; mRNA = messenger ribonucleic acid; PFA = paraformaldehyde; P(x) = postnatal day (x); qRT-PCR = quantitative real-time polymerase chain reaction; RNA = ribonucleic acid; RT = reverse transcriptase

Early life exposure to anesthesia in humans continues to be a concern given the mounting evidence of a cognitive deficit in animal models including those of non–human primates.1,2 This deficit is lifelong and highly reproducible across anesthetic agents and several species.3,4 The results from clinical studies on the other hand have been mixed, complicated by the nature of studying children and the inherent confounding variables of subjects who require surgery/anesthesia.

In a rodent model of early life anesthesia exposure, we have previously identified environmental enrichment (EE) as an important modulator of this phenotype.5 Others have subsequently shown that voluntary exercise is a key component of EE,6,7 and it can improve cognitive ability in both human and animal models. In rodents, this beneficial effect of exercise is thought to result from increasing adult hippocampal neurogenesis, the process of a neuroprogenitor cell differentiating into a fully integrated hippocampal neuron8 (although in humans, the dogma of adult neurogenesis has recently been called into question9,10). To understand this process, in our animal model, we hypothesized that voluntary exercise alone is sufficient to rescue this anesthetic-induced phenotype and we investigated the role of exercise on several different memory tasks after early life isoflurane (Iso) exposure in rats.


This manuscript adheres to the relevant Animal Research: Reporting of In Vivo Experiments (ARRIVE) guidelines for reporting animal research.

Iso Exposure

Figure 1.:
Exercise treatment after perinatal isoflurane exposure. A, Outline of experiments undertaken with animal numbers. B, Images of 2 housing paradigms. Exercising animals had access to a running wheel with an attached odometer to measure daily exercise. Sedentary animals were singly housed with a fixed half-wheel in which they could explore but not exercise. C, Treatment animals exercised on average 2.37 km/d for control animals (n = 67) and 2.73 km/d for isoflurane-exposed animals (n = 85). There was no statistical difference between groups (Mann-Whitney P = .428). Error bars represent standard deviation. BDNF indicates brain-derived neurotrophic factor; BrdU, bromodeoxyuridine; Iso, isoflurane; mRNA, messenger ribonucleic acid; ns, not significant; P, postnatal day; qRT-PCR, quantitative real-time polymerase chain reaction.

All animals were housed and treated according to institutional policy with a protocol approved by the UCSF institutional animal care use committee. Postnatal day 2 (P2) male Sprague Dawley rats were purchased (Charles River, South San Francisco, CA) and cross-fostered with dams until P7, the day of Iso exposure. Animals from each litter were chosen at random to be exposed to anesthesia or sham (controls [Con]) and returned to their respective dam postexposure. Iso was administered in a step-down protocol over 6 hours (2% Iso hours, 0–2; 1.4% Iso hours, 2–4; 0.8% Iso hours, 4–6) using a custom-built anesthesia exposure box and monitored with a Datex-Ohmeda (Capnomac Ultima, West Bloomfield, MI) gas analyzer (Supplemental Digital Content, Figure 1, https://links.lww.com/AA/C929). Unlike previous experiments in which tail clamping was used in conjunction with measured volatile anesthetic concentration to determine the minimum alveolar concentration (MAC) of anesthesia in P7 male rats, we instead used the measured Iso concentration to guide the exposure because we did not find a difference between tail clamping and nontail clamping in past volatile anesthetic exposures.5 In addition, the step-down protocol here was developed to decrease the mortality sometimes observed when using a MAC-based protocol. Skin temperature was measured by infrared thermometer every 15 minutes and controlled with a heating pad. Fresh gas was approximately 1 L/min with an fraction of inspired oxygen (Fio2) of 0.4–0.6. Carbon dioxide (CO2) absorbent pellets, Litholyme (Allied Healthcare, St Louis, MO), were placed in the chamber, and CO2 levels were continuously monitored. On completion of Iso exposure, the chamber was flushed with atmospheric air and animals were allowed to emerge from anesthesia. Con were separated from dams for 30 minutes and reunited with their Iso-exposed littermates on emergence from anesthesia after return of righting reflex. Out of the 146 pups exposed to anesthesia, 2 died during exposure. One hundred fourteen Con were utilized for the experiments (Figure 1A).

Exercise Treatment

Animals were weaned at P21 and placed in a solitary cage either with a running wheel and associated odometer or a cage with a fixed half-wheel with which the animal could explore but could not exercise (Figure 1B). Animals were given food and water ad libitum and were subjected to a 12-hour reverse light/dark cycle. Daily exercise distance was recorded.

Barnes Maze

Barnes maze was conducted as previously described.11 Briefly, animals were exposed to the standard rat maze with 20 holes around the perimeter. The testing arena included external spatial cues on each wall, a fan in a fixed position, bright lights in fixed positions, in an otherwise darkened room. The experimenter was hidden from view. Acquisition trials of up to 4 minutes were performed daily for 4 consecutive days, while movements were tracked using a camera (Basler aca1280; Basler, Inc, Exton, PA) and tracking software (Ethovision XT 11.5; Noldus Information Technology, Inc, Leesburg, VA). Probe trials were conducted 1 week after fourth acquisition day and lasted 90 seconds. Cumulative time spent at each hole and time in each quadrant of the maze were recorded. A cohort of young adults was studied after 3 weeks of exercise, beginning training on P43. A different cohort, which had first undergone recognition memory testing, started Barnes maze training on P112. Animals which never found the goal hole during the learning phase were not included in the probe trial analysis (nP43 = 5; nP112 = 1) as well as animals which did not investigate any hole during the probe (nP43 = 2; nP112 = 0).

Recognition Memory Tasks

The 4 specific recognition memory tasks were performed on 2 cohorts of naive animals, starting at P43 (after 3 weeks of exercise). Tasks were performed as previously described in detail.12 Briefly, animals were placed with a set of objects in a testing box and given 4 minutes to explore. Animals were removed for 2 minutes and the box was cleaned and objects were moved or switched depending on the paradigm, then animals were observed for an additional 4 minutes. Movements were tracked with the aforementioned software. Social recognition memory tests were performed as described12 using juvenile males in small cages, with a 4-minute exposure then 2-hour delay, with 4-minute test. Discrimination index was calculated by subtracting the time spent exploring the nongoal object from the goal object, divided by the total time exploring both objects.

Bromodeoxyuridine Pulse

Animals at P43 from each of the groups studied were injected with 50 mg/kg of bromodeoxyuridine (BrdU) intraperitoneally and killed after 4 hours. Animals were briefly anesthetized with Iso then perfused with intracardiac injection of ice-cold phosphate buffered saline followed by 4% paraformaldehyde (PFA). Brains were dissected and fixed overnight in 4% PFA then placed in 30% sucrose until they sunk. Sixty-micron thick sections were obtained on a freezing microtome and collected serially in a 12 well plates. A single well representing 1/12 of the brain was taken for each animal and stained as floating sections using a rat anti-BrdU primary antibody (Invitrogen #MA1_82088, Carlsbad, CA) after exposing tissue to 30 minutes of 2N hydrochloric acid at 38°C, then 0.5N boric acid at room temperature for 30 minutes. Alexa Fluor goat anti-rat 594 (Fisher A11007, Waltham, MA) was used as the secondary. Images were obtained on an upright Nikon epi-fluorescent microscope (Eclipse 80i; Nikon Instruments, Inc, Melville, NY) with a 10× objective using Stereo Investigator software (MBF Biosciences, Williston, VT). Images of each hippocampal section were acquired using consistent exposure times for each round of immunostaining. Images were exported to Image J (National Institutes of Health, Bethesda, MD), and the counting tool was used to count positive cells after the dentate gyrus area was defined. The person acquiring the images and analyzing the images was blinded to the groups.

Quantitative Real-Time Polymerase Chain Reaction

Animals at P21 (weaning) and at P43 (beginning of behavior experiments) were killed and brains were immediately dissected into cortex and hippocampus and homogenized with 500-μL pipette in 250-μL Trizol (Invitrogen). Total ribonucleic acid (RNA) was extracted with chloroform according to manufacturer’s instructions. Concentrations were measured by nanodrop spectrophotometer (ND1000; Themo Fisher Scientific, Wilmington, DE), then samples were treated with DNAse (Invitrogen), and reverse transcriptase reaction was performed (high-capacity cDNA reverse transcriptase [RT] Kit; Life Technologies, Carlsbad, CA). TaqMan brain-derived neurotrophic factor (BDNF) (Rn02531967_s1) and glyceraldehyde 3-phosphate dehydrogenase (GAPDH) probes (Rn01775763_g1) were utilized for quantitative polymerase chain reaction (qPCR) reaction with TaqMan Universal Polymerase Chain Reaction (PCR) master-mix (Applied Biosystems, Foster City, CA) on a Step-One Plus (Applied Biosystems) real-time PCR machine. Samples were run in duplicate with no-RT and no-sample (water) Con. Cycle thresholds (Cts) were manually set to the same values across plates, ensuring that they were in the exponential amplification phase. Biological replicates that did not reach the Ct by cycle 35 were excluded (n = 1). Data were analyzed using the Δ-Δ Ct method13 with GAPDH as the housekeeping gene and median sedentary Con Ct value as the reference point. Data were displayed as the relative fold change on a log axis, with geometric means and 95% confidence intervals calculated.

Statistical Analysis

GraphPad Prism 7 statistical software (GraphPad Software Inc, San Diego, CA) was used for statistical analyses. Before intergroup testing, data were subjected to D’Agostino and Pearson normality testing. If the data were normally distributed, then parametric tests were performed, and if not, nonparametric tests were used. For all analyses, α was set at .05 using 2-tailed tests.

For the comparison of daily exercise, an unpaired Mann-Whitney U test was performed. For the Barnes maze, 3 separate analyses were completed. First, learning was assessed with a 2-way repeated-measures analysis of variance (ANOVA) to analyze the effect of daily training versus group. Second, quadrant analysis used 1-sample t test, asking if the time spent in the goal quadrant was significantly different from chance (0.25). Groups that met this criterion were compared with a 1-way ANOVA. Third, the probe trial was analyzed by comparing the time spent at the goal hole to positions relative to the goal using Dunnett multiple comparison test with adjusted P values.

The analysis of the recognition memory task used the discrimination index to determine if animals were spending more time with the goal object than chance (theoretical mean = 0). This dataset was not normally distributed, so Wilcoxon ranked test was used with a theoretical mean = 0.

BrdU incorporation was analyzed using a 2-way ANOVA to determine if there was an effect of exercise or Iso exposure. Tukey post hoc analysis was used to compare between groups.

The quantitative real-time PCR (qRT-PCR) experiments were subjected to an unpaired Mann-Whitney U test for the P21 cohort, which had only 2 groups. The P43 cohort with 4 groups was tested with a 2-way ANOVA to detect an effect of exercise or Iso exposure.

Given that these experiments had not been done, we did not perform a power analysis a priori, but rather we based our sample size on previous experience with these assays and the literature.11,12,14 We purposefully weighted the groups heavily for the Iso exposure because previously significant mortality would occur at the time of anesthesia. In addition, we weighted the exercise cohort over the sedentary cohort because we were unsure how consistently the animals would exercise. In the end, we did not eliminate animals for low daily exercise. Finally, we were constrained by physical space required within our vivarium for these specially designed cages and could not accommodate larger cohorts.


Iso Causes a Spatial Memory Deficit That Is Rescued With Voluntary Exercise

Figure 2.:
Spatial memory acquisition and recall with the Barnes maze (P43). A, Animals were trained during the acquisition of the Barnes maze task to find the escape box, and escape latency was measured over the 4 separate trials on consecutive days. Two-way repeated-measures ANOVA showed a significant decrease in latency over the 4 days (F [3117 ] = 10.2; P < .001), but there was no significant difference among groups (Con/Sedentary n = 9, Con/Exercise n = 11, Iso/Sedentary n = 10, Iso/Exercise n = 14). B, Quadrant analysis of probe trial, 7 d after the last training session, showed all groups except the Iso/Sedentary animals spent significantly more time at the goal quadrant than chance (2-tailed 1-sample t test, chance = 25%). C, No significant difference in time exploring holes for Iso/Sedentary animals between the “goal” hole and the “nongoal” holes by Dunnett multiple comparisons test (P values adjusted for multiple comparisons). Con/Sedentary and Con/Exercise showed significant discrimination in the holes explored with holes ±4, ±8, ±9, and ±10 and ±3, ±4, ±5, ±6, ±7, ±8, and ±10, respectively. Exercise rescued the spatial memory in the Iso/Exercise group with 10 significantly different holes. Error bars represent standard deviation. *P < .05, **P < .01, ***P < .001. ANOVA indicates analysis of variance; Con, control; Iso, isoflurane; P, postnatal day.

Animals voluntarily ran in the exercise wheels with an average distance of 1.87 and 2.22 km in the Con and Iso-exposed groups, respectively (Figure 1; P = .125). Over the course of the 4-day learning acquisition phase, the animals from all groups showed a decrease in the latency to find the goal in the Barnes maze demonstrating learning is intact in all groups (Figure 2A; 2-way repeated-measures ANOVA, F[3117] = 10.2, P < .001; but no effect by group or interaction). The probe trial (Figure 2B), performed 1 week after completion of the learning phase, was used to assess long-term memory of the target location and was analyzed by the time spent in the target quadrant. Iso-sedentary animals spent no more time than chance in the goal quadrant (1-sample t test, P = .524), while all other groups were able to discriminate the goal (PCon-sedentary = .004; PCon-exercise = .012; PIso-exercise = .033). Subsequent comparison of the 3 groups that did demonstrate memory of the goal revealed no difference in time spent searching in that quadrant (ANOVA, F[2,29] = 0.065; P = .937) and suggesting that voluntary exercise can rescue the deficit that occurs after early exposure to Iso. Additional within-group analysis was then performed comparing time at the goal to all other holes (Figure 2C). The Iso-sedentary group spent no more time at the goal than any other hole, while all other groups did.

No Recognition Memory Deficit After Early Life Anesthesia Exposure

Figure 3.:
Recognition memory preserved after early life Iso exposure. A–C, Recognition memory tasks testing specific domains of recognition memory with increasing difficulty from Novel Object Recognition, Object Place Recognition to Allocentric Object Place Recognition (Con/Sedentary n = 16, Con/Exercise n = 23, Iso/Sedentary n = 23, Iso/Exercise n = 30). Displayed is the discrimination index for individual animals which is the time spent investigating the goal object minus the time investigating the nongoal object, divided by the total time exploring both objects. A 2-way nonparametric test (Wilcoxon signed rank) was performed for each group which found a significant difference from chance (zero) for all groups in the 3 variations of recognition memory tested. D, Social recognition was tested with juvenile male rats as targets. Like the other recognition tasks, all groups had a significant difference from chance (Wilcoxon signed rank). Solid bar represents mean value. Error bars represent standard deviation. *P < .05, **P < .01, ***P < .001. Con indicates control; Iso, isoflurane.

In contrast to previous studies from our laboratory, we found no evidence of a deficit in recognition memory testing in the Iso-sedentary animals from 2 separate cohorts which are shown pooled. There was no cognitive deficit detectable in any of the groups for any of the tests (Novel Object, Object Place, Allocentric Object Place, and Social recognition), which is different from previous results12 because every group had a discrimination index >0 indicating correct discrimination (Figure 3). In addition, there was no significant improvement with exercise in either group when we performed an analysis based on correlation of distance and discrimination index (not shown).

Iso Spatial Memory Deficit Is Not Present at P112

Figure 4.:
Barnes maze acquisition and recall in mature adult (P112 cohort) is preserved. A, Animals that previously underwent recognition memory battery and were naive to the Barnes maze learned the spatial location of the escape box over the course of 4 consecutive days as evidenced by a decrease in latency (2-way repeated-measures ANOVA, F [3138 ] = 12.27; P < .001) (Con/Sedentary n = 11, Con/Exercise n = 12, Iso/Sedentary n = 12, Iso/Exercise n = 13). Like the P43 cohort, there was no difference in learning between groups. B, In the probe trial, quadrant analysis showed discrimination of the goal with significant time spent in the goal quadrant compared to chance for all groups (2-tailed 1-sample t test, chance = 25%). C, The Con/Sedentary and the Con/Exercise groups had significant differences between the time spent exploring the goal hole versus every other space ± from the goal by Dunnett multiple comparison test (P values adjusted for multiple comparisons). Different from the P43 cohort, the Iso/Sedentary P112 group spent significant differences in the goal compared to ±4, ±5, ±6, ±7, ±8, ±9, and ±10. Like the P43 cohort, the P112 Iso/Exercise group performed slightly better that than the Iso/Sedentary group with differences ±3, ±4, ±5, ±6, ±7, ±8, ±9, and ±10. Error bars represent standard deviation. *P < .05, **P < .01, ***P < .001, ****P < .0001. Con indicates control; Iso, isoflurane; P, postnatal day.

A single cohort that was exposed to the exercise/sedentary paradigm and completed the recognition memory battery underwent Barnes maze training at P112. Like the P43 cohorts, these older animals learned the location of the goal hole by training day 4 (2-way repeated-measures ANOVA, F[3138] = 12.27, P < .001; but no effect by group or interaction) (Figure 4A). In contrast to animals tested at P43, all groups in this P112 cohort successfully discriminated the correct goal quadrant from chance, including the Iso-sedentary group (Figure 4B) (1-sample t test, PIso-sedentary < .001; PCon-sedentary = .016; PCon-exercise < .001; PIso-exercise = .001). Subsequent within-group analysis of the time spent at each hole reveals strong ability to discriminate even within the goal quadrant among this cohort of older animals (Figure 4C).

Proliferating Cells Decrease in the Adult Hippocampus After Early Life Anesthesia, but Increase With Voluntary Exercise

Figure 5.:
BrdU labeling at P43 in perinatally isoflurane (Iso)-exposed animals is decreased in sedentary animals but is increased with exercise. A, Representative immunohistochemistry processed images from the hippocampus of P43 animals after a 3 h BrdU pulse. B, Quantification reveals significant differences between Iso/Sedentary animals and every other group by Tukey multiple comparisons suggesting that early life Iso exposure reduces proliferating cells in the hippocampus, but can be reversed with exercise in adulthood (P values adjusted for multiple comparisons) (Con/Sedentary n = 8, Con/Exercise n = 9, Iso/Sedentary n = 8, Iso/Exercise n = 9). Solid bars represent mean, and error bars represent standard deviation. **P < .01, ***P < .001, ****P < .0001. BrdU indicates bromodeoxyuridine; Con, control; Dapi, 4´,6-diamidino-2-phenylindole; Iso, isoflurane; P, postnatal day.

To explore the effect of early life Iso exposure on proliferating cells in the adult brain, we performed a 3-hour BrdU pulse in a separate cohort of animals after 3 weeks of exercise and just before the start of the behavior testing protocol. Two-way ANOVA showed an effect for both Iso exposure and exercise, but no interaction (FIso[1,30] = 10.52, P = .003; FExercise[1,30] = 19.25, P < .001; FInteraction[1,30] = 3.26, P = .08). Comparing groups post hoc, we found that Iso exposure on P7 led to a decrease in BrdU incorporation on P43 in the hippocampus compared to sedentary Con (Figure 5A, B) (Tukey P = .005). Exercise increased the BrdU incorporation of the Iso-exercise group relative to the Iso-sedentary group (Tukey P < .001) and was not different from Con-sedentary animals (Tukey P = .850).

BDNF Messenger RNA Is Not Upregulated After Exercise and Is Not Affected by Early Life Anesthesia in the Hippocampus and Cortex at P21 and P43

Figure 6.:
BDNF mRNA levels in the cortex and hippocampus of animals after perinatal isoflurane exposure. A and C, P21 (time of weaning) relative levels of BDNF mRNA in cortex and hippocampus, respectively, displayed as fold change. Neither brain region was significantly different (cortex [n = 5], P = .222; hippocampus [n = 5], P = .547, unpaired 2-tailed Mann-Whitney test). B and D, P43 (3 wk of ± exercise) relative levels of BDNF mRNA in cortex and hippocampus, respectively. Unlike BrdU incorporation, there was no significant effect of exercise (Con/Sedentary n = 8, Con/Exercise n = 9, Iso/Sedentary n = 10, Iso/Exercise n = 10) (cortex 2-way ANOVA F [1,32 ] = 0.236, P = .631; hippocampus 2-way ANOVA F [1,33 ] = 1.186, P = .284). Geometric mean and 95% confidence intervals displayed. ANOVA indicates analysis of variance; BDNF, brain-derived neurotrophic factor; BrdU, bromodeoxyuridine; Con, control; Iso, isoflurane; mRNA, messenger ribonucleic acid; P, postnatal day.

The increase in neural precursor proliferation with exercise has been shown in other contexts to be linked to upregulation of BDNF,15,16 which also has implications for learning and memory. To better understand the effect of exercise on spatial memory in our model, we quantified messenger RNA (mRNA) levels of BDNF at P21 which was the time of weaning (Figure 6A). There were no significant differences in BDNF message between Iso-exposed and nonexposed animals in either the cortex or hippocampus (Mann-Whitney Pcortex = .222; Phippocampus =.547). Even after 3 weeks of exercise, there was no significant effect of exercise or Iso exposure on BDNF expression (2-way ANOVA; cortex: FExercise[1,32] = 0.236, P = .631; FIso[1,32] = 0.038, P = .847; FInteraction[1,32] = 1.543, P = .223; and hippocampus: FExercise[1,33] = 1.186, P = .284; FIso[1,33] = 1.46, P = .236; FInteraction[1,33] = 1.78, P = .191) (Figure 6B).


We have previously reported that EE can rescue the deficit observed after early life exposure to volatile anesthesia. Here we tested a single component of that enrichment, voluntary exercise, to determine whether it alone can reverse the deficit we reported. The primary finding of this experiment is that in rats, a spatial memory task is sensitive to early life anesthesia exposure and can be rescued by voluntary exercise beginning at the time of weaning. In addition, we show that there is a decrease in proliferating hippocampal cells in adulthood after perinatal anesthesia, which may underlie the memory deficit. Like other reports, we found an increase in proliferating cells in the hippocampus with exercise, but unlike other reports, we did not find differential BDNF mRNA levels.

This result is similar to the spatial memory deficit we reported previously in a study on EE.5 The EE protocol in the original study was composed of 3 components: exercise, social housing, and a complex living space. The experiments here isolate one component of that EE, exercise, which alone was able to rescue the deficit. To quantify the amount of exercise each animal performed, they were housed individually thereby eliminating the social aspect of the original EE experiment. The environment for both exercising and nonexercising groups contains either a wheel or half-wheel and is indeed more complex than a standard cage but comparable between groups. We do not believe that group housing alone is enough to rescue the deficit because we have reported several other experiments where a deficit was seen in animals that were group housed in otherwise nonenriched cages.11,17,18 One difference between our previous EE study5 and this study is that the volatile anesthetic was different, sevoflurane versus Iso, yet the improvement of spatial memory with enrichment is similar between studies.

In contrast to our own previous studies, we did not find a recognition memory deficit after early life Iso exposure. There are a number of possible explanations for the differences in recognition test results in our study compared to previous experiments. One possibility is the difference in Iso exposure. We had previously exposed animals to 4 hours of Iso at 1 MAC, defined by tail clamping, which resulted in a down titration of Iso concentration from 4% to 1.1% over the 4 hours.12,14 We experienced significant animal deaths during that anesthetic exposure. To counteract this, we attempted to give that same amount of Iso but spread out over 6 hours, in a step-down fashion. Although this is roughly the same area under the curve, or MAC hours of exposure in both protocols, it is possible that this longer anesthetic is overall less potent than the previously published 4-hour anesthetic. Alternatively, the differences could also be affected by the social housing of the animals. In our previous studies, the animals were jointly housed, but because of the exercise protocol and to isolate the effect of exercise, all animals were singly housed for these experiments. In other publications, social exposure has been linked to differences in behavioral outcomes.19

These studies also raise the possibility that spatial memory is more sensitive to early life anesthesia than recognition memory because we did not find a difference in recognition memory but did find deficits in the Barnes maze probe trial. The Barnes maze may be a more sensitive test compared to the recognition memory battery, or it may take a more substantial insult to induce a long-term deficit in recognition memory because we have previously reported a deficit using a different exposure paradigm.12,14 Others have noted that age, training, and handling can affect Barnes maze performance,20 which might explain differences in performance of the spatial memory tests between the P43 and P112 cohorts. Despite a lack of difference between groups in the quadrant analysis, Iso-exposed animals seem less able to discriminate between nearby holes in the maze.

How the benefits of exercise are extolled to the brain is the subject of continued study, but the mechanism in rodents includes mediation by BDNF and hippocampal neurogenesis which both solidify and strengthen memories.21 Recently, exogenous BDNF has been injected into the hippocampus after early anesthesia and was shown to improve spatial memory.22 Exercise in humans has been specifically linked to a positive relationship between BDNF and recognition memory.23 In rodents, neurogenesis is doubled after exercise which correlates with improved plasticity and memory.8 While recent observational studies in humans have called into question the role of adult hippocampal neurogenesis,9,10 there is an overwhelming body of evidence that support the positive effects of exercise. BDNF has been a presumed critical factor in improving learning and memory given the number of studies showing exercise-induced upregulation at mRNA and protein levels.24–26 Our study found no significant changes in BDNF mRNA, suggesting that there may be non-BDNF–dependent pathways critical for strengthening long-term learning and memory. We did not quantify BDNF protein; however, it is possible that this could be discordant with the mRNA results although many studies report an increase in mRNA with exercise. This is an area that could be the subject of future investigation. It is significant that while perinatal anesthesia exposure has a lifelong behavioral effect, voluntary exercise is sufficient to rescue this phenotype. This is similar to our previous findings that EE, instituted at the time of weaning, could rescue the deficit5 which argues for EE or exercise alone to be used as a postexposure treatment.

Limitations of this study include the inherent constraints of the P7 rodent model which includes the immaturity of the animals, the lack of control of ventilation, and the difficulty in correlating animal age with human age. However, similar methods have been used in multiple non–human primate studies with behavioral deficits observed which suggests conservation of the mechanism across species and evolution.2,27 Here we used male rats exclusively as was done in the EE enrichment study because we have also described sex-specific differences in susceptibility to anesthesia at P714, so the conclusions drawn are limited to males. While we have no reason to suspect that a similar beneficial effect of exercise would not be present in females, this hypothesis remains untested.

In summary, we show that early life anesthesia exposure in rodents leads to a spatial memory deficit and a decrease in proliferating cells in the hippocampus in the adult. Voluntary exercise can improve spatial memory function and rescue the loss of hippocampal proliferation. This effect cannot be attributed to BDNF mRNA levels, which in this model, appeared unchanged despite daily exercise. Exercise is a simple, inexpensive, low-risk intervention that can offset potential deleterious side effects of early life anesthesia exposure and improve spatial memory.


The authors thank Jason Leong (University of California [UC], San Francisco, San Francisco, CA) for technical help in conducting the behavioral experiments and maintenance of the running wheel cages.


Name: Gregory A. Chinn, MD, PhD.

Contribution: This author helped conceive, design, and perform the experiments; analyze the data; and write the manuscript.

Name: Jennifer M. Sasaki Russell, PhD.

Contribution: This author helped perform the experiments, analyze the data, and edit the manuscript.

Name: Esther T. Banh, BS.

Contribution: This author helped perform and analyze the data from behavioral experiments and edit the manuscript.

Name: Saehee C. Lee, BA.

Contribution: This author helped perform and analyze the data from behavioral experiments and edit the manuscript.

Name: Jeffrey W. Sall, PhD, MD.

Contribution: This author helped conceive and design the experiments and write the manuscript.

This manuscript was handled by: Gregory J. Crosby, MD.


1. Paule MG, Li M, Allen RR, et al. Ketamine anesthesia during the first week of life can cause long-lasting cognitive deficits in rhesus monkeys. Neurotoxicol Teratol. 2011;33:220–230.
2. Raper J, Alvarado MC, Murphy KL, Baxter MG. Multiple anesthetic exposure in infant monkeys alters emotional reactivity to an acute stressor. Anesthesiology. 2015;123:1084–1092.
3. Stratmann G. Review article: neurotoxicity of anesthetic drugs in the developing brain. Anesth Analg. 2011;113:1170–1179.
4. Sinner B, Becke K, Engelhard K. General anaesthetics and the developing brain: an overview. Anaesthesia. 2014;69:1009–1022.
5. Shih J, May LD, Gonzalez HE, et al. Delayed environmental enrichment reverses sevoflurane-induced memory impairment in rats. Anesthesiology. 2012;116:586–602.
6. Kobilo T, Potter MC, van Praag H. Neurogenesis and exercise. Encycl Behav Neurosci. 2010:404–409.
7. Mustroph ML, Chen S, Desai SC, Cay EB, DeYoung EK, Rhodes JS. Aerobic exercise is the critical variable in an enriched environment that increases hippocampal neurogenesis and water maze learning in male C57BL/6J mice. Neuroscience. 2012;219:62–71.
8. Voss MW, Vivar C, Kramer AF, van Praag H. Bridging animal and human models of exercise-induced brain plasticity. Trends Cogn Sci. 2013; 17:525–544.
9. Sorrells SF, Paredes MF, Cebrian-Silla A, et al. Human hippocampal neurogenesis drops sharply in children to undetectable levels in adults. Nature. 2018;555:377–381.
10. Boldrini M, Fulmore CA, Tartt AN, et al. Human hippocampal neurogenesis persists throughout aging. Cell Stem Cell. 2018;22:589–599.e5.
11. Sasaki Russell JM, Chinn GA, Maharjan D, Eichbaum Y, Sall JW. Female rats are more vulnerable to lasting cognitive impairment after isoflurane exposure on postnatal day 4 than 7. Br J Anaesth. 2019;122:490–499.
12. Lee BH, Chan JT, Hazarika O, Vutskits L, Sall JW. Early exposure to volatile anesthetics impairs long-term associative learning and recognition memory. PLoS One. 2014;9:e105340.
13. Livak KJ, Schmittgen TD. Analysis of relative gene expression data using real-time quantitative PCR and the 2(-Delta Delta C(T)) method. Methods. 2001;25:402–408.
14. Lee BH, Chan JT, Kraeva E, Peterson K, Sall JW. Isoflurane exposure in newborn rats induces long-term cognitive dysfunction in males but not females. Neuropharmacology. 2014;83:9–17.
15. Cotman CW, Berchtold NC. Exercise: a behavioral intervention to enhance brain health and plasticity. Trends Neurosci. 2002;25:295–301.
16. Berchtold NC, Chinn G, Chou M, Kesslak JP, Cotman CW. Exercise primes a molecular memory for brain-derived neurotrophic factor protein induction in the rat hippocampus. Neuroscience. 2005;133:853–861.
17. Lee BH, Hazarika OD, Quitoriano GR, et al. Effect of combining anesthetics in neonates on long-term cognitive function. Int J Dev Neurosci. 2014;37:87–93.
18. Ramage TM, Chang FL, Shih J, et al. Distinct long-term neurocognitive outcomes after equipotent sevoflurane or isoflurane anaesthesia in immature rats. Br J Anaesth. 2013;110suppl 1i39–i46.
19. Hendershott TR, Cronin ME, Langella S, McGuinness PS, Basu AC. Effects of environmental enrichment on anxiety-like behavior, sociability, sensory gating, and spatial learning in male and female C57BL/6J mice. Behav Brain Res. 2016;314:215–225.
20. Rosenfeld CS, Ferguson SA. Barnes maze testing strategies with small and large rodent models. J Vis Exp. 2014;84:e51194–e51194.
21. Liu PZ, Nusslock R. Exercise-mediated neurogenesis in the hippocampus via BDNF. Front Neurosci. 2018;12:52.
22. Wu J, Bie B, Naguib M. Epigenetic manipulation of brain-derived neurotrophic factor improves Memory Deficiency Induced by Neonatal Anesthesia in Rats. Anesthesiology. 2016;124:624–640.
23. Whiteman AS, Young DE, Budson AE, Stern CE, Schon K. Entorhinal volume, aerobic fitness, and recognition memory in healthy young adults: a voxel-based morphometry study. Neuroimage. 2016;126:229–238.
24. Vaynman S, Ying Z, Gomez-Pinilla F. Hippocampal BDNF mediates the efficacy of exercise on synaptic plasticity and cognition. Eur J Neurosci. 2004;20:2580–2590.
25. Russo-Neustadt A, Beard RC, Cotman CW. Exercise, antidepressant medications, and enhanced brain derived neurotrophic factor expression. Neuropsychopharmacology. 1999;21:679–682.
26. van Praag H, Kempermann G, Gage FH. Running increases cell proliferation and neurogenesis in the adult mouse dentate gyrus. Nat Neurosci. 1999;2:266–270.
27. Coleman K, Robertson ND, Dissen GA, et al. Isoflurane anesthesia has long-term consequences on motor and behavioral development in infant rhesus macaques. Anesthesiology. 2017;126:74–84.

Supplemental Digital Content

Copyright © 2019 International Anesthesia Research Society