Effects of stress or infection on rat behavior show robust reversals due to environmental disturbance

Animals

Adult male and female Wistar Han rats (7 to 8 weeks at the start of the experiments) were procured from InVivos, Singapore. Rats were housed in groups of two per cage (males and females were housed separately) with ad libitum access to food and water (24–26°C; 60–70% relative humidity; 12h light-dark cycle with lights on at 0700h). For all tests, animals were allocated to groups in a random manner. Experiments were conducted by SA-S and AHN who were blind to group allocations. Analysis was conducted by AV who was also blind to group allocations. All procedures were approved by the Institutional Animal Care and Use Committee of the Nanyang Technology University. All efforts were made to ameliorate any suffering of animals. None of our procedure involved induction of sustained pain requiring pharmacological interventions. Animals were daily observed to confirm lack of sickness related behaviors and weighed weekly. The behavior tests do not involve any use of shock or other painful stimuli. The dose of parasites used in this study does not result in weight loss or sickness behavior in this strain of rats.

At the end of all experiments, animals used in the Toxoplasma infection paradigm were sacrificed by decapitation and their brains were removed and flash frozen. In the case of the stress paradigm, animals were sacrificed by cardiac perfusion using cold phosphate buffered saline (PBS) followed by cold 4% paraformaldehyde.

Toxoplasma gondii infection and quantification of aversion to cat odor

Female rats were either injected with tachyzoites of type 2 Prugniaud strain of Toxoplasma gondii (5x10⁶ tachyzoites in 500 µl phosphate buffered saline, i.p.;) or mock injected with the buffer alone between 2pm and 4pm. Parasites needed for the infection were maintained in vitro in human foreskin fibroblast cultures and were harvested using syringe lysis. Behavioral experiments were conducted seven weeks post-infection; a time-window consistent with chronic phase of the infection. Toxoplasma gondii infection did not cause significant change in body weight of animals (179.1 ± 4.708, n = 8 for uninfected; 183.1 ± 2.706, n = 7 for infected; p = 0.5, independent sample t-test).

Aversion to cat odor was quantified in two different manners. For each run of the experiment, there was one uninfected group and one Toxoplasma-infected group. Fifteen animals were used in total for experiment 1 (8 uninfected, 7 infected) and nineteen animals were used in total for experiment 2 (10 uninfected, 9 infected).

Aversion was first quantified in a rectangular arena with two opposite and identical arms (Figure 1A; 76 x 9 cm each), separated by a central part (9 x 9 cm in size; white Perspex). Animals were habituated to the arena for three consecutive days for 20 minutes each day. On the subsequent day, cat odors were presented in one bisect of the maze (1 ml each; bobcat urine from Maine Outdoor Solutions, USA). Animals were placed in the center of the maze and exploration time in both bisects of the arena was measured for 20 minutes. Trials were video recorded with offline analysis conducted using AnyMaze (Stoelting, USA). In this batch of animals, each received 500 µl of buffered saline intraperitoneally thirty minutes before the behavioral test.

Aversion to cat odor was also quantified in a circular arena (Figure 1A; diameter = 1 m) that was arbitrally divided into four quadrants. Animals were habituated to the arena for three consecutive days for 20 minutes each day. On the subsequent day, cat odor, vanilla essence, water and the bedding from the animal’s home cage were presented in each quadrant of the maze. Animals were placed in the center of the maze and exploration time in all quadrants of the arena was measured for 20 minutes. Trials were video recorded with offline analysis conducted using AnyMaze (Stoelting, USA).

Stress paradigm and quantification of anxiety

Eight week old breeders obtained from InVivos were allowed to acclimatize for at least 5 days before setting up breeding pairs (one male and one female per cage). Breeding cages were changed once a week as per normal, but with gentle handling of female, in case of pregnancy. Once pregnancy was certain (approx. 2 weeks), male was removed. 19 days after breeding pairs were set up (or if visually heavily pregnant), cages were checked daily for litters. Day of birth is assigned P0.

Maternal separation was used as the stress model (P2-P14, daily). 16 animals were used in total; 8 stressed, 8 unstressed. On each of these days, the dam was removed from the cage and placed in a new cage with unsoiled bedding. Pups were then retrieved into another cage with unsoiled bedding, transported to a separate room and put on a heating pad for three hours every morning. At the end of the separation period, pups and then dam were sequentially returned to the original soiled cage. Also, soiled bedding was changed on postnatal day 2, 9 and 14; by returning pups to a clean cage that had been supplemented with a scoop of soiled bedding and nesting material from the original cage. This practice was repeated on postnatal day 18 if the bedding was considered significantly soiled in case of large litter sizes. Pups were weaned on postnatal day 21. Anxiety was quantified when the male pups reached adulthood (7–8 weeks of age). Anxiety was measured using home cage emergence assay (adapted from 12) and elevated plus-maze¹³.

In the home cage emergence assay, a rat placed in its home cage was transported to a well-lit room and habituated for five minutes. The cage was then left open by removal of the lid. The rat was offered a possibility of emerging from the home cage through a wire grid placed within the cage. The latency of emergence was recorded. Emergence was defined when all four limbs of the rat were placed on the grid. Trials were terminated at the emergence or at five minutes, whichever occurred earlier. Trials were video recorded and scored manually.

The elevated plus-maze consisted of a plus-shaped arena with two open (75 × 11cm, 1cm wall, 3–4 lux illumination) and two enclosed arms (75 × 11 cm, 26 cm wall). The arena was elevated to a height of 60 cm above the ground. The animal was placed at the center at the start of the trial. Exploration in open and enclosed arms was quantified for five minutes each.

All experiments for the stress paradigm were done using two groups of mice: stressed and unstressed.

Statistical analysis

The probability of type 1 error was calculated using unpaired two-tailed Student’s t-test. The standardized effect size was calculated using Cohen’s d¹⁴; with values above the magnitude of 0.8 interpreted as being of robust scale. Negative d values correspond to the comparisons where mean of experimental treatment was greater than that of respective controls. Mean inter-group difference was also calculated with 95% confidence intervals. Data is graphically presented as mean and standard error of the mean (SEM), along with individual values for each animal for each endpoint. Number of animals in each experimental group is noted in the figures. All statistical analysis was conducted using Graphpad Prism.

Toxoplasma gondii infection increased aversion to cat odor

In the first set of animals, aversion to cat odor was quantified as percentage time in bisect containing cat odor relative to total trial duration. Rabbit odor was placed in the opposing bisect as a novel non-predator odor. Inter-group differences did not reach pre-determined threshold for statistical significance (Figure 1B; t₁₃ = 1.78, p = 0.098). Despite the lack of sufficiently low type 1 error, the effect on mean was of robust magnitude (Cohen’s d = 0.949; Δ = -11.61% with 95% confidence intervals -25.68 to 2.46%). The maximum of animals from the infected group was below the median of the uninfected animals. The robust effect size and the observation that infected mean was lower than uninfected animals in contrast to the multitude of published studies, led us to plan a further experiment to increase the statistical power.

Figure 1.

Toxoplasma gondii-infected female rats showed increased aversion to bobcat odor in two sequential experiments. Panel A depicts schematics of test arenas used in the experiments later depicted in panels B and C. Ordinate for panel B and panel C depicts time spent by female rats chronically infected with Toxoplasma gondii near bobcat odor. Line graphs depict mean and standard error of the mean for control (black) and infected (gray) female rats. *, p < 0.05; unpaired two-tailed Student’s t-test (n = 8 uninfected and 7 infected animals for panel A; and 10 uninfected and 9 infected animals for panel B).

In this second set of animals, aversion to cat odor was quantified in a circular arena congruent to the initial design of reported infection effects. One quadrant contained soiled bedding from home cage of the animal, serving as the home base for exploratory sorties. Cat odor and a novel vanilla odor were placed in two adjoining quadrants. The ratio of time spent in cat quadrant relative to sum time spent in both cat and novel odor quadrants was calculated (chance = 50%). Toxoplasma infection, in contrast to earlier observations in the similar design, reduced percentage time spent near cat urine (Figure 1C; t₁₇ = 2.70, p = 0.015). The effect of infection on innate aversion was of robust magnitude (Cohen’s d = 1.239; Δ = -16.46% with 95% confidence intervals -29.31 to -3.61%). The maximum of animals from the infected group was again observed to be below the median of the uninfected animals.

Serological examination confirmed that all animals in the infected groups sustained chronic infection with Toxoplasma gondii.

Early life maternal separation stress resulted in anxiolytic behavior in male rats

Animals subjected to early life maternal separation stress were tested in the elevated plus maze and home cage emergence test to determine the effect of maternal separation on anxiety behavior during adulthood.

Stressed animals, in contrast to earlier observations in a similar design, exhibited significantly less anxiety compared to unstressed controls. This was evident as increased percentage entries into anxiogenic open arms of elevated plus-maze (Figure 2A; t₁₄ = 4.21, p = 0.0009). Stress-induced anxiolysis was of robust magnitude (Cohen’s d = -2.104; Δ = 27.77% with 95% confidence intervals 13.62 to 41.92%). The minimum of animals from the stressed group was higher than all but one animal from the unstressed group. Experimental treatment did not cause significant differences in number of entries made into non-anxiogenic enclosed arms of the maze (t₁₄ = 1.98, p = 0.07). To preclude effects of entries in enclosed arms on open arm exploration, we further conducted a univariate analysis of variance for percentage open arm entries while employing number of enclosed entries as a covariate. Furthermore, stressed animals spent more time in open arms during the test duration compared to unstressed counterparts (Figure 2B; t₁₄ = 6.69, p < 0.001). This analysis revealed a significant increase in open arm exploration due to the stress independent of inter-group differences in enclosed arm entries (F_1,13 = 14.898, p = 0.002). This is congruent with significant increase in number of head dips made during the trial by stressed animals (t₁₄ = 3.41, p = 0.0042; Δ = 13.25 with 95% confidence intervals 4.94 to 21.56).

Figure 2. Early life maternal separation stress resulted in increased anxiolytic behavior in male rats.

Ordinate depicts number of entries and into the open arm relative to total entries in open and enclosed arms of the elevated plus maze (A and B, respectively) and latency to emerge from the home cage into a novel environment (C). Line graphs depict mean and standard error of the mean for unstressed (black) and stressed (gray) male rats. *, p < 0.05; **, p < 0.01; unpaired two-tailed Student’s t-test (n = 8 animals in each experimental group).

Stress-induced anxiolysis was also confirmed by home cage emergence test. In this assay, anxiolysis manifests as reduced latency to emerge into a novel environment from home cage. Stress significantly decreased the latency of home cage emergence (Figure 2C; t₁₄ = 3.14, p = 0.0072). Stress-induced anxiolysis was also of robust magnitude in this assay (Cohen’s d = -1.57; Δ = -121.4s with 95% confidence intervals -204.2 to -38.5s). The maximum latency of animals from the stressed group was lower than median latency from the unstressed group.