Human twin studies have compared neurocognitive outcomes in twins in which one, both, or neither have undergone procedures involving exposure to general anesthetics. By finding that both exposed and unexposed members of a twin pair have equally poor neurocognitive outcomes, these studies conclude that pre-existing conditions rather than general anesthetics are the cause. We tested in healthy laboratory male rats whether sevoflurane-unexposed rats that were reared together in the same litter or cage with neonatally sevoflurane-exposed rats can develop behavioral deficiencies similar to those developed by their exposed cagemates.

All experimental procedures were approved by the University of Florida Institutional Animal Care and Use Committee (Gainesville, Florida). The study was conducted in accordance with the Animal Research: Reporting of In Vivo Experiments guidelines. The initial testing of this new phenomenon was done in male rats because sevoflurane may induce more robust neurobehavioral phenotype in male rats.  Postnatal day 5 male Sprague–Dawley rats from different litters were mixed together and distributed in three types of newly formed litters, each having a foster dam. The three litter types were as follows: (1) only pups that were not exposed to sevoflurane and remained with their foster dams (control litter); (2) only pups that underwent exposure to 2.1% sevoflurane for 6 h on postnatal day 5, as previously described (sevoflurane litter) and (3) equal numbers of sevoflurane-exposed and sevoflurane-unexposed pups (mixed litter). The study did not involve specific investigation of foster dams’ behavior, but routine daily checkups of all study animals did not suggest any obvious differences in foster dams’ care, and all pups remained alive. After weaning on postnatal day 21, rats from control litters were housed two per cage (control group), as were rats from sevoflurane litters (sevoflurane group). Rats from mixed litters were also housed two per cage after weaning, so that each cage contained one sevoflurane-unexposed and one sevoflurane-exposed rat. Based on sevoflurane exposure status, rats from mixed litters or cages were assigned to two separate study subgroups depending on their exposure status: (1) the mixed–control group and (2) the mixed–sevoflurane group. Behavioral evaluations were done between postnatal days 60 and 90, as previously described. 

As anticipated based on the results of previous studies the sevoflurane group, compared to the control group, spent less time and made fewer entries in open arms of the elevated plus maze (fig. 1, A and B), which is typically interpreted as evidence of an anxiety-like behavior. The sevoflurane group had also lower sensitivity to inhibition of acoustic startle by preceding prepulse (prepulse inhibition of startle; fig. 1C), indicative of impaired sensorimotor gating function. Finally, in the Morris water maze test, the sevoflurane group, compared to the control group, did not differ in escape latencies across the 5-day training period (spatial learning; fig. 2A) but during the probe test spent less and more time in the target and entrance quadrants, respectively (fig. 2, B and C), which is considered as evidence of impairment in spatial memory. To our surprise, the mixed–sevoflurane and the mixed–control groups did not differ from each other in all these parameters (figs. 1 and 2). The mixed–sevoflurane group, compared to the sevoflurane group, showed less anxiety-like behavior, had similarly impaired sensorimotor gating, and did not exhibit evidence of impaired spatial memory (figs. 1 and 2).

Fig. 1.
The behavioral effects of sevoflurane and cohabitation during the elevated plus maze and prepulse inhibition of the acoustic startle tests. (A, B) The percentage of time spent in the open arms and number of entries to the open arms during the elevated plus maze test. The data are means ± SD from 16 rats per group (n = 15, the control group). One rat in the control group was removed from the study because of a fall. (C) The percentage of prepulse inhibition of the startle at prepulse intensities of 3 dB, 6 dB, and 12 dB. The data are means ± SD from 16 rats per group. The color coding of the experimental groups is shown at the bottom of the plots. This color coding is applicable to the entire figure. The P values of the multiple pairwise comparisons are shown in the respective plots above the horizontal lines. The beginning and end of each horizontal line correspond to the compared experimental groups. Two-way ANOVA with treatment (control and sevoflurane) and housing (“paired” indicates rats from the same treatment group housed together, and “unpaired” indicates sevoflurane-exposed and sevoflurane-unexposed rats housed together) as the independent variables were used to assess differences in time spent in the open arms and number of entries to the open arms during the elevated plus maze test. For prepulse inhibition of startle, linear mixed models for repeated measures were used, with prepulse inhibition intensity as repeated measures. The models also included treatment and housing as main effects, as well as interaction terms. Pairwise comparisons were conducted to examine the hypothesis of housing differences for treatment effects. Multiple pairwise comparisons were done with the Holm–Sidak method.

The behavioral effects of sevoflurane and cohabitation during the elevated plus maze and prepulse inhibition of the acoustic startle tests. (A, B) The percentage of time spent in the open arms and number of entries to the open arms during the elevated plus maze test. The data are means ± SD from 16 rats per group (n = 15, the control group). One rat in the control group was removed from the study because of a fall. (C) The percentage of prepulse inhibition of the startle at prepulse intensities of 3 dB, 6 dB, and 12 dB. The data are means ± SD from 16 rats per group. The color coding of the experimental groups is shown at the bottom of the plots. This color coding is applicable to the entire figure. The P values of the multiple pairwise comparisons are shown in the respective plots above the horizontal lines. The beginning and end of each horizontal line correspond to the compared experimental groups. Two-way ANOVA with treatment (control and sevoflurane) and housing (“paired” indicates rats from the same treatment group housed together, and “unpaired” indicates sevoflurane-exposed and sevoflurane-unexposed rats housed together) as the independent variables were used to assess differences in time spent in the open arms and number of entries to the open arms during the elevated plus maze test. For prepulse inhibition of startle, linear mixed models for repeated measures were used, with prepulse inhibition intensity as repeated measures. The models also included treatment and housing as main effects, as well as interaction terms. Pairwise comparisons were conducted to examine the hypothesis of housing differences for treatment effects. Multiple pairwise comparisons were done with the Holm–Sidak method.

Fig. 2.
The behavioral effects of sevoflurane and cohabitation during the Morris water maze test. (A) Plots showing the values of escape latencies during the 5-day training period of the Morris water maze test. (B, C) Histograms showing the time that rats from different treatment groups spent in each quadrant (B) and the time that rats from a specific treatment group spent in each quadrant (C). The color coding of the experimental groups in (A) and (B) is shown at the bottom of the plots. (C) Colors of quadrants and quadrant locations are shown in the inset. The same quadrant locations are applicable to (B). The P values of the multiple pairwise comparisons are shown in the respective plots above the horizontal lines. For clarity purposes, the P values between the times spent in the entrance and target quadrants only are shown in (C). The beginning and end of each horizontal line correspond to the compared experimental groups. The data are means ± SD from 16 rats per group. For the Morris water maze escape latency and time in each quadrant, linear mixed models for repeated measures were used, with day and quadrant modeled as repeated measures, respectively. The models also included treatment and housing as main effects, as well as interaction terms. Pairwise comparisons were conducted to examine the hypothesis of housing differences for treatment effects. Multiple pairwise comparisons were done with the Holm–Sidak method.

The behavioral effects of sevoflurane and cohabitation during the Morris water maze test. (A) Plots showing the values of escape latencies during the 5-day training period of the Morris water maze test. (B, C) Histograms showing the time that rats from different treatment groups spent in each quadrant (B) and the time that rats from a specific treatment group spent in each quadrant (C). The color coding of the experimental groups in (A) and (B) is shown at the bottom of the plots. (C) Colors of quadrants and quadrant locations are shown in the inset. The same quadrant locations are applicable to (B). The P values of the multiple pairwise comparisons are shown in the respective plots above the horizontal lines. For clarity purposes, the P values between the times spent in the entrance and target quadrants only are shown in (C). The beginning and end of each horizontal line correspond to the compared experimental groups. The data are means ± SD from 16 rats per group. For the Morris water maze escape latency and time in each quadrant, linear mixed models for repeated measures were used, with day and quadrant modeled as repeated measures, respectively. The models also included treatment and housing as main effects, as well as interaction terms. Pairwise comparisons were conducted to examine the hypothesis of housing differences for treatment effects. Multiple pairwise comparisons were done with the Holm–Sidak method.

Our findings suggest that through cohabitating, rats can affect each other’s brain development, ameliorating some sevoflurane-induced deficits in exposed rats and inducing some deficits in unexposed rats at behavioral levels. These findings suggest that in twin studies of neurodevelopmental abnormalities in general and anesthesia-induced neurodevelopmental abnormalities in particular, intersibling interactions should be considered a potentially important determinant of outcomes. These findings also suggest a new explanation of why neurodevelopmental disorders are more likely to occur among siblings.

The mechanism(s) whereby cohabitating sevoflurane-exposed and sevoflurane-unexposed rats affect each other’s phenotypes remain to be determined. Our thinking was that such a transmission would be more plausible between human cohabitants because of persistent social (conscious) interaction. In animals, on the other hand, transmission through such persistent conscious interaction seemed less realistic. One potential mechanism for mediation of such behavioral effects between cohabitating rodents is coprophagy. Our preliminary studies do not support this mechanism. In addition to unraveling the mechanisms of cohabitation effects, among other important questions to be addressed in future studies are investigations of this newly described phenomenon in female rats and whether rearing of unexposed and exposed pups in the same litter until weaning is sufficient for transmission of phenotypic alterations. We realize that these findings are controversial. Although we have repeated a number of control measurements, these findings will require re-examination by different laboratories.