- Email: [email protected]
6J mice
Behavioural Brain Research 314 (2016) 215–225
Contents lists available at ScienceDirect
Behavioural Brain Research journal homepage: www.elsevier.com/locate/bbr
Research report
Effects of environmental enrichment on anxiety-like behavior, sociability, sensory gating, and spatial learning in male and female C57BL/6J mice Taylor R. Hendershott, Marie E. Cronin, Stephanie Langella, Patrick S. McGuinness, Alo C. Basu ∗ Department of Psychology, College of the Holy Cross, 1 College Street, Worcester, MA 01610, United States
h i g h l i g h t s • • • • •
Female mice showed greater preference for EPM open arms regardless of housing. EE attenuated sensory gating in male and female mice. EE enhanced spatial learning in male and female mice. EE resulted in greater use of spatially precise strategies in the water maze. Swim speed and escape latency in SE females were slow relative to the other groups.
a r t i c l e
i n f o
Article history: Received 25 June 2016 Received in revised form 29 July 2016 Accepted 2 August 2016 Available online 3 August 2016 Keywords: Environmental enrichment Elevated plus maze Prepulse inhibition Water maze Spatial strategies Sex differences
a b s t r a c t The influence of housing on cognition and emotional regulation in mice presents a problem for the study of genetic and environmental risk factors for neuropsychiatric disorders: standard laboratory housing may result in low levels of cognitive function or altered levels of anxiety that leave little room for assessment of deleterious effects of experimental manipulations. The use of enriched environment (EE) may allow for the measurement of a wider range of performance in cognitive domains. Cognitive and behavioral effects of EE in male mice have not been widely reproduced, perhaps due to variability in the application of enrichment protocols, and the effects of EE in female mice have not been widely studied. We have developed an EE protocol using common laboratory equipment that, without a running wheel for exercise, results in significant cognitive and behavioral effects relative to standard laboratory housing conditions. We compared male and female wild-type C57BL/6J mice reared from weaning age in an EE to those reared in a standard environment (SE), using common measures of anxiety-like behavior, sensory gating, sociability, and spatial learning and memory. Sex was a significant factor in relevant elevated plus maze (EPM) measures, and bordered on significance in a social interaction (SI) assay. Effects of EE on anxiety-like behavior and sociability were indicative of a general increase in exploratory activity. In male and female mice, EE resulted in reduced prepulse inhibition (PPI) of the acoustic startle response, and enhanced spatial learning and use of spatially precise strategies in a Morris water maze task. © 2016 Elsevier B.V. All rights reserved.
1. Introduction Since Donald Hebb’s original observations of increased spatial learning ability in rats reared in an enriched environment (EE) [1,2], laboratory controlled enrichment has been associated with widespread molecular and morphological changes through-
∗ Corresponding author. E-mail address: [email protected] (A.C. Basu). http://dx.doi.org/10.1016/j.bbr.2016.08.004 0166-4328/© 2016 Elsevier B.V. All rights reserved.
out the brain in rodents, mainly in comparison to subjects reared in isolation (for review see Ref. [3,4]). Further, these molecular and morphological changes have been associated with behavioral changes [5]. Although some effects on the nervous system have been replicated by different research groups, behavioral and cognitive effects of EE have not been widely reproducible (Table 1). The mechanisms of enrichment-related changes are of interest for the understanding of functional neuroplasticity, and the influence of housing conditions on rodent behavior is a serious concern for the reproducibility of pre-clinical research findings [6].
216
Table 1 Variability in reports of EE effects on cognition and behavior. Species, strain
Sex
Number of Animals
Age at Enrichment
Age at Testing
Enrichment Duration
Running Wheel
EPM behavior
Sensory gating
[8]
Mouse, C57BL/6J
F
3 weeks
12 weeks
40 days
Yes, No
↓, ↓ % time in open arms
↑, ↓ PPI
[9]
F
3 weeks
25 weeks, 51 weeks
∼22 weeks, 48 weeks
Yes
↓ % open arm entries
[10]
Mouse, 129SvJ/C57BL/6 hybrids Mouse, NMRI
SE = 13 EE = 14 EER = 14 SE = 8 EER = 7
4 weeks
17 weeks
∼13 weeks
Yes
n.c.
[16]
Mouse, ICR
M
4 weeks
6 weeks
Mouse, NMRI
M
4 weeks
8 weeks
2 weeks, 4 weeks 4 weeks
Yes
[17]
SE = 8 EER = 8 SE = 12 EER = 12 SE = 16 EER = 16
Yes
[18]
Mouse, C57BL/6J, M 129S6/SvEv/Tac
3 weeks
10 weeks
7 weeks
Yes
[19]
Mouse, C57BL/6J
8 weeks
14 weeks
40 days
Yes
↓ latency
[21]
Mouse, C57Bl/6
SE B6 = 8 EER B6 = 8 SE 129 = 8 EER 129 = 8 SE = 24 EER = 24 SE = 10 EER = 11 SE-M = 11 SE-F = 12 EE-M = 10 EE-F = 11
3 weeks
20 weeks
17 weeks
Yes
↑ probe
4 weeks
8 weeks
4 weeks
No
Present Study Mouse, C57BL/6J
M
M and F (combined) Female M and F (sex as factor)
Change in EE housed mice relative to SE housed mice: n.c. no change; ↓ decrease; ↑ increase; *interaction of housing and sex, R running wheel.
Social behavior
MWM Spatial learning and memory n.c.
↑, ↓ PPI ↑social behavior ↑aggressive behavior n.c.
↑ latency to enter ↓ PPI open arm ↑ open arm entries
n.c.
↓ latency* ↓ duration ↓ thigmotaxis ↑ spatial precision
T.R. Hendershott et al. / Behavioural Brain Research 314 (2016) 215–225
Ref.
T.R. Hendershott et al. / Behavioural Brain Research 314 (2016) 215–225
Important factors to consider in the interpretation and application of findings from studies using EE include: variability in laboratory-controlled enrichment protocols, developmental stage and duration of enrichment, and species, strain, sex, and age of animal subjects. Of particular concern for interpretation of behavioral outcomes is the use of a running wheel in some enrichment protocols. Exercise alone has been shown to improve object recognition memory and to induce brain derived neurotrophic factor (BDNF) expression in male rats [7], and interaction of enrichment and exercise on sensory gating and spatial memory has been previously documented in mice [8]. There is also a significant gap in this literature regarding sex: few studies of cognition have focused on female rodents or compared them directly to males. We selected a battery of behavioral and cognitive tests that are frequently used to study mouse models of psychiatric disorders and administered it to EE and SE housed male and female mice of the C57BL/6J strain. We used the elevated plus maze (EPM) to assess anxiety-like behavior. We hypothesized that animals reared in EE would display reduced anxiety-like behavior in the EPM, though not all previous research is consistent with this hypothesis (Table 1, [9,10]). We noted the possibility that EE might in some cases result in anxietylike behavior due to increased stress, or changes in activity (for review, see Ref. [11]), and therefore planned to analyze multiple measures for this test that would reveal such effects. We measured PPI of the acoustic startle response, a sensorimotor gating phenomenon observed widely in vertebrates, because it is considered a psychiatric endophenotype. Attenuated PPI has been observed in disorders in which neurodevelopmental factors are implicated, such as schizophrenia [12] and autism [13]. The measurement of this phenomenon in rodent models of neuropsychiatric disorders is widespread, and the field has been cautioned against leaving assumptions of cross-species comparability in neural substrates untested [14]. Results of previous studies of EE and PPI in rodents are mixed; in some cases EE attenuated PPI and in other cases there was enhancement (Table 1, [8,15,16]). We hypothesized that enrichment would result in an increase in social behavior, but the results of past literature have been, again, mixed (Table 1, [17,18]). Previous studies of male mice housed subjects in larger social groups than is standard (up to 8 per cage), and used social behavior assays that allowed free interaction between the subject and the target mouse. We chose to control for complexity of social experience in our housing conditions by keeping the number of cage mates equivalent between EE and SE. We also chose to use a limited social interaction assay that sacrifices information about the dyadic social behaviors in favor of a focus on the behavioral choices of the subject. We used a standard MWM test of spatial learning and memory to evaluate the cognitive function of male and female mice following EE. Bonaccorsi et al. [19] found that EE male and female C57BL/6 mice reached the hidden platform faster than SE mice late in training. Although data from male and female mice were included, sex was not included as a factor in the statistical analysis. Thus, it was not possible to ascertain whether males and females responded differently to behavioral and cognitive assays as a result of the experimental condition [20]. There are previous reports in the literature of enhanced MWM performance in female mice following EE, but without direct comparison to males (Table 1, [8,21]). The effects of EE on emotional regulation and cognition in male and female wild type (WT) mice are, thus, yet to be firmly established. We have summarized published results of comparable studies in Table 1, selecting studies of young (≤6 months) WT mice that used similar behavioral assays and SE (as opposed to socially isolated) controls. The reproducibility of these effects may be improved by the standardization of enrichment protocols, in particular if they can be obtained using common rather than specialized laboratory equipment. Therefore, we conducted a study
217
that uses a simple EE, does not introduce exercise effects by way of a running wheel, and that uses WT mice reared in standard social housing conditions as the control group. We measured emotional and cognitive outcomes commonly of interest in neuropsychiatric disease research in both males and females, including sex as a factor in data analysis to allow the direct identification of sex effects [20]. We hypothesized that EE mice would exhibit less anxiety-like behavior, increased social behavior, increased PPI, and improved spatial learning and memory relative to SE WT mice.
2. Materials and methods 2.1. Animal subjects, general vivarium conditions, and handling procedures Animal subjects were bred in the College of the Holy Cross vivarium from WT C57BL/6J parents within 5 generations of import from Jackson Laboratory, Bar Harbor, ME, USA. Breeding procedure was to co-house a single male breeder with 3 virgin female breeders, checking weekly for pregnancy. Pregnant dams were isolated in the final week of pregnancy and left undisturbed until the end of the first postnatal week. Mice were weaned and randomly assigned to four different housing groups at approximately 4 weeks of age (P2430). Littermates were separated and assigned to different housing conditions to the extent possible to mitigate potential litter effects. Forty-four WT C57BL/6J mice obtained from nine litters were used in this study. Final experimental group sizes were: 11 SE males, 10 EE males, 12 SE females, and 11 EE females. All animals had access to food and water ad libitum and caging was changed weekly. Ambient temperature was maintained at 21 ± 1 ◦ C. The vivarium lighting was regulated on a 12 h:12 h cycle. Behavioral testing was conducted during the light phase, avoiding the phase transition by 2 h. The study was conducted in 3 separate batches, nearly balanced for experimental groups. All husbandry and experimental procedures were approved by the College of the Holy Cross Institutional Animal Care and Use Committee and carried out in accordance with the National Institutes of Health Guide for the Care and Use of Laboratory Animals.
2.2. Enrichment protocol SE control mice were reared in standard laboratory conditions in clear plastic caging. The inner cage dimensions were approximately: 28 cm × 17.2 cm × 12.2 cm (ANCARE, Inc. standard mouse cage). Mice were housed in groups of 2–5 per cage. SE cages were lined with corn cob bedding and provided with a 5.1 cm × 5.1 cm square of cotton nesting material, changed weekly. The EE cages had larger inner dimensions, approximately: 45.4 cm × 23.5 cm × 15.7 cm (ANCARE, Inc. standard rat cage). EE cages were furnished with two squares of nesting material, a plastic in-cage shelter, and a variety of novel objects that were changed weekly during the first 4 weeks post-weaning (Fig. 1). To facilitate consistent application of this protocol, we selected common laboratory items as novel objects: bottle caps and segments of PVC tubing in week 1, severed sample tubes and beakers (polystyrene, 50 ml) in week 2, pipette tips (1000 ml), a polystyrene funnel and a stationary plastic disk of a running wheel in week 3, and half of a wire test tube rack in week 4. All cages were fitted with a wire top rack that held food and a single water bottle. All cages were non-ventilated and kept covered with a filter lid. The larger caging, nesting material, and in-cage shelter were continued for the duration of the experiment.
218
T.R. Hendershott et al. / Behavioural Brain Research 314 (2016) 215–225
Fig. 1. Experimental timeline and enrichment protocol. For 4 weeks following weaning, EE mice were exposed to novel objects changed weekly. Extra nestlet, plastic dome, and larger cages were provided for the duration of the experiment. SE mice were littermates of EE mice, weaned at the same time, reared in parallel and tested together with EE mice.
2.3. Elevated plus maze (EPM) The EPM test was conducted at age 8 weeks ± 4 days. The EPM was constructed of opaque plexiglass. The four arms were 30.5 cm long × 5.7 cm wide and extended from a center area (7 cm diagonal); the maze floor was elevated 52.7 cm above the ground. The walls of the two opposing closed arms were 21 cm high. The light levels for testing were ∼100 lx in the open arms and ∼35 lx in the closed arms. Test subjects were habituated to the testing room for at least half an hour prior to testing. At the start of each EPM trial, the test subject was placed in the center of the maze facing an open arm. The mouse was allowed to explore the maze freely for 5 min. The location of the mouse in the maze was tracked continuously using data capture hardware and EthovisionXT software (Noldus Information Technology, Wageningen, the Netherlands). 2.4. Prepulse inhibition (PPI) of the acoustic startle response (ASR) The ASR was measured using the SR-Lab Startle Response System for mice (San Diego Instruments, Inc., San Diego, California, USA). PPI of the ASR was conducted at 8–9 weeks of age using a previously published protocol [22]. Testing was conducted with lights off in the sound attenuating testing chambers. In brief, each session consisted of 5 min of white noise at a background level of 70 dB followed by 72 test trials. The first six “leader” trials and the last six “trailer” trials delivered only a 120 dB startle stimulus, to allow for stabilization of the ASR and measurement of within-session habituation. The remaining trials were of the following types, in equal numbers, and presented in pseudorandom order: no stimulus, 120 dB test stimulus, 120 dB stimulus with prepulse of 73 dB (+3 dB), 120 dB stimulus with prepulse of 76 dB (+6 dB), 120 dB stimulus with prepulse of 82 dB (+12 dB). All test subjects exhibited the ASR, as determined by comparison of the baseline reading in the no stimulus trials to the 120 dB stimulus test trials by unpaired Stu-
dent’s t-test for each subject. Percent PPI was calculated for each prepulse intensity for each test subject using the following equation: (Average response to 120 dB test stimulus − Average response to 120 dB stimulus preceded by prepulse)/Average response to 120 dB test stimulus. 2.5. Social interaction (SI) A SI assay was conducted at 8–11 weeks of age, using a nonautomated apparatus in an adaptation of a published protocol [23]. The test apparatus, constructed of clear plexiglass, consisted of three 40.5 cm × 20.3 cm × 30.5 cm chambers separated by doors. The lighting in all three chambers was kept at ∼20 lx for testing. Social targets were sex-matched adult mice of the same strain that were not familiar to the subject (not parent or littermate). Target mice were habituated to the holders for 30 min on the day before testing. Test subjects were habituated to the testing room for at least half an hour prior to testing. At the beginning of each trial, the subject was restricted to the center chamber for 5 min for habituation to the apparatus. After 5 min, a novel sex-matched target mouse was placed in a cylindrical holder in one of the side chambers and the doors were opened, giving the subject access to all three chambers. Each subject was given 10 min to freely explore the apparatus. The test subject was tracked using the automated system described in Section 2.3. 2.6. Morris water maze (MWM) A nine-day MWM protocol was performed at age 9–12 weeks to test spatial learning and memory as previously described [22]. Briefly, mice were trained to find a hidden platform using intra- and extra-maze cues for 7 days (4 trials/day, pseudorandomized point of entry to maze, maximum trial duration 90 s, inter-trial interval approximately 10 min). On the 8th day, the hidden platform was
T.R. Hendershott et al. / Behavioural Brain Research 314 (2016) 215–225
removed and subjects were allowed to explore the maze for 60 s (Probe test). On the 9th day, the hidden platform was placed in a new location within the maze (New Platform), and the test subjects were trained to find it for 4 trials. The testing apparatus was a commercially available water maze for mice (San Diego Instruments, Inc., San Diego, California, USA). The lighting in the maze was kept at ∼20 lx for testing. Non-toxic white paint was used to make the water opaque, and water temperature was maintained at 22 ± 1 ◦ C. The test subjects were tracked using the automated system described in section 2.3. Scoring of search strategies as spatially precise or spatially imprecise was conducted post-testing by observers blind to experimental group, according to previously published criteria [24]. Briefly, to be classified as Spatially Precise (SP), a search fit one of the following three descriptions: the subject took a direct route to the platform from entry (direct swim), the subject showed a preference for a corridor from the point of entry towards the platform (directed search), or the subject performed a highly localized search within the target quadrant (focal search). If a search did not meet any of these criteria, it was classified as Spatially Imprecise (SI), a label that encompassed several patterns of behavior such as chaining, scanning, thigmotaxis, perseverance, and random search [24]. 2.7. Statistical analysis Data were managed using Microsoft Excel software. Statistical analyses were conducted using IBM SPSS Statistics 23 software. Data were analyzed by analysis of variance (ANOVA) with housing and sex as between-subjects factors and any repeated measures as within-subjects factors. Restricted ANOVAs or Bonferroni corrected t-tests were conducted post hoc as noted in the results section. 3. Results 3.1. Elevated plus maze (EPM) EPM data were analyzed by 2-way ANOVA using housing and sex as between-subjects factors. All significant main effects of housing and sex are reported in this section. There were no significant interactions between housing and sex. Mice housed in EE entered an open arm of the EPM sooner than mice housed in SE, main effect of housing, F(1,40) = 4.49, p = 0.040 (Fig. 2A). Mice housed in EE also made more entries to the open arms, main effect of housing, F(1,40) = 7.33, p = 0.010 (Fig. 2B). EE mice made more arm entries in total, main effect of housing, F(1,40) = 8.92, p = 0.005 (Fig. 2C), and exhibited greater locomotor activity within the maze, main effect of housing, F(1,40) = 14.08, p < 0.001 (Fig. 2D), than mice housed in SE. Important measures of anxiety-like behavior revealed significant effects of sex, but not housing: Female mice spent a higher percentage of arm time in the open arms, main effect of sex, F(1,40) = 6.44, p = 0.015 (Fig. 2E), and a greater raw amount of time in the open arms, main effect of sex, F(1,40) = 6.03, p = 0.019 (Fig. 2F). There were borderline effects of sex on the percentage of arm entries made to open arms, F(1,40) = 3.45, p = 0.071 (Fig. 2G), and on the total raw time spent in arms, F(1,40) = 3.67, p = 0.063 (Fig. 2H). 3.2. Prepulse inhibition (PPI) of the acoustic startle response (ASR) Startle data were analyzed by 3-way ANOVA using housing and sex as between-subjects factors and trial type as a withinsubjects factor. All significant main effects of housing, sex, and trial type are reported in this section. There were no significant interactions. There were no significant effects of housing or sex on startle reactivity, though there was significant within-session startle habituation evident in the decrease in startle amplitude between the leader and trailer 120 dB stimulus trials, main effect of trial
219
type, F(1,40) = 27.356, p < 0.001 (Fig. 3A). As expected, percent PPI increased with prepulse intensity, main effect of prepulse intensity, F(2,80) = 88.541, p < 0.001 (Fig. 3B). There was a significant effect of housing on the percent PPI such that mice housed in EE exhibited lower percent PPI than mice housed in SE, F(1,40) = 6.655, p = 0.014 (Fig. 3B). According to a 2-way ANOVA, startle reactivity to the 120 dB startle stimulus in the trials that were interspersed with the prepulse test trials was not affected by housing or sex. 3.3. Social interaction Social interaction data were analyzed by 3-way ANOVA using housing and sex as between-subjects factors and chamber (target or empty chamber within the testing apparatus) as a within-subjects factor. There were no significant effects of housing or sex on time spent in the different chambers in the apparatus, though there was a significant main effect of chamber type, F(1,40) = 6.096, p = 0.018, indicating that the mice were sociable (Fig. 4A). The effect of housing was marginal on this measure, F(1,40) = 3.225, p = 0.08, as was the effect of sex, F(1,40) = 3.160, p = 0.083. However, housing did affect the percentage of trial time spent in the center of the apparatus, in that EE mice spent less time in the center, and therefore more time exploring the side chambers: 2-way ANOVA revealed a significant main effect of housing, F(1,40) = 4.399, p = 0.042 (Fig. 4B). The effect of sex was marginal on this measure, F(1,40) = 3.458, p = 0.07. 3.4. Morris water maze All MWM data were analyzed by 3-way ANOVA with housing and sex as between-subjects factors and either day or trial as a repeated-measures factor unless otherwise noted. Post hoc analyses were conducted as noted. No significant effects of housing or sex were observed on average daily latency to reach the hidden platform during the 7 days of acquisition. However, 3-way ANOVA revealed a significant main effect of day, F(6,240) = 41.55, p < 0.001 (Fig. 5A). The measure of average path length showed a significant main effect of day, F(6,240) = 55.39, p < 0.001, and also a significant interaction between day and housing, F(6,240) = 3.516, p = 0.002. Follow-up within-day univariate ANOVA revealed significant effects of housing on Day 1, F(1,42) = 19.98, p < 0.001, and Day 2, F(1,42) = 4.80, p = 0.034, such that EE mice took shorter paths to reach the hidden platform (Fig. 5B). Percentage of time spent in thigmotaxis was affected by day, F(6,240) = 90.46, p < 0.001, decreasing as subjects learned the location of the hidden platform. There was a significant interaction between day and housing in the thigmotaxis measure, F(6,240) = 4.217, p < 0.001. Follow-up withinday univariate ANOVA revealed a significant effect of housing on thigmotaxis on Day 1, F(1,42) = 11.258, p = 0.002 (Fig. 5C). The spatial strategy scores showed significant main effects of both day, F(6,240) = 30.18, p < 0.001, and housing, F(1,40) = 9.25, p = 0.004, reflecting higher use of spatially precise strategies by EE mice throughout hidden platform acquisition (Fig. 5D). The housing-related differences in acquisition were not followed by differences in spatial reference memory according to the results of the probe test. All test groups demonstrated preference for the target quadrant above chance, and 3-way ANOVA with housing and sex as between-subjects factors and quadrant as a within-subjects factor revealed only a significant main effect of quadrant, F(3,120) = 21.302, p < 0.001 (Fig. 5E). However, in the probe test, there was an interaction of housing and sex with respect to swim velocity, F(1,40) = 6.018, p = 0.02, such that a sex difference was observed in SE but not EE mice. Post hoc Bonferroni corrected t-tests revealed that SE females had lower swim velocity than EE females, t(21) = 2.83, p < 0.05, and SE males, t(21) = 2.81, p < 0.05, but there were no significant differences between EE and SE males, or between male and female EE mice (Fig. 5F). There were no sig-
220
T.R. Hendershott et al. / Behavioural Brain Research 314 (2016) 215–225
Fig. 2. Behavior of SE and EE mice in the EPM. (A) EE mice entered an open arm sooner than SE mice. (B) EE mice entered open arms more often than SE mice. (C) EE mice made more total arm entries than SE mice. (D) EE mice travelled a greater distance within the maze than SE mice. (E) Female mice spent a greater percentage of arm time in the open arms than males regardless of housing condition. (F) Females also spent more time overall in open arms than male mice. (G) There was no significant difference in percentage of entries to open arms between groups. (H) There was no significant difference in the total time spent in arms between groups. Asterisks indicate statistical significance of reported main effects: *p < 0.05, **p < 0.01, ***p < 0.001.
nificant effects of sex or housing on any daily-averaged outcome measures on Day 9, when a new platform location was introduced (Figs. 5A–D). The finding that housing-related differences were generally observed in Day 1 of water maze acquisition prompted us to conduct a closer analysis of trial-by-trial performance in that session, in parallel with that planned for new platform acquisition on Day 9. Trial-by-trial analyses of water maze performance on Day 1, the first day of hidden platform training, and Day 9, the reversal learning day, revealed effects of housing on within-session trialby-trial improvement in performance. EE mice were similar to their SE counterparts in Trial 1, but improved in performance more rapidly. On Day 1, the escape latency measure revealed a main
effect of trial, F(3,120) = 6.19, p = 0.001, a main effect of housing F(1,40) = 10.89, p = 0.002, and an interaction of trial and housing, F(3,120) = 3.39, p = 0.02. Post hoc univariate ANOVAs revealed significant effects of housing at Trial 2, F(1,42) = 13.37, p = 0.001, and Trial 3, F(1,42) = 10.07, p = 0.003 (Fig. 6A). The path length measure showed a similar pattern on Day 1: a significant main effect of trial, F(3,120) = 7.90, p < 0.001, a significant main effect of housing, F(1,40) = 18.66, p < 0.001, and a significant interaction of trial and housing, F(3,120) = 4.5, p = 0.005. Post hoc univariate ANOVAs revealed significant effects of housing at Trial 2, F(1,42) = 12.73, p = 0.001, and Trial 3, F(1,42) = 15.43, p < 0.001 (Fig. 6B). The thigmotaxis measure revealed no effect of trial, but a significant main effect of housing, F(1,40) = 10.95, p = 0.002, and a significant inter-
T.R. Hendershott et al. / Behavioural Brain Research 314 (2016) 215–225
221
Fig. 3. Acoustic startle and sensory gating in SE and EE mice. (A) Startle amplitude decreased significantly between the first six “leader” and last six “trailer” 120 dB stimulus trials. (B) There was a significant decrease in the percent PPI in EE mice as compared to SE mice. Asterisks indicate statistical significance of reported main effects: *p < 0.05, ***p < 0.001.
Fig. 4. Sociability in SE and EE mice. (A) Mice preferred the target chamber to the empty chamber in the social interaction assay. (B) EE mice spent less time in the center chamber than SE mice. Asterisks indicate statistical significance of reported main effects: *p < 0.05.
action between trial and housing, F(3,120) = 2.84, p = 0.041. Post hoc univariate ANOVAs revealed significant effects of housing at Trial 2, F(1,42) = 10.51, p = 0.002, and Trial 3, F(1,42) = 8.41, p = 0.006 (Fig. 6C). EE mice (sexes collapsed) also made significantly greater use of spatially precise strategies on Trial 3, 2-tailed Fisher Exact Probability Test p = 0.002 (Fig. 6D). On Day 9, though the hidden platform was moved to a new location to test reversal learning, the conventional performance measures of escape latency and path length were lower and closer between experimental groups than on Day 1. Thigmotaxis was much lower than on Day 1. The overall decrease in these measures was expected given the amount of MWM training that transpired between the two sessions. The use of spatial strategy analysis was
thus particularly helpful in revealing an advantage of EE in withinsession learning. On Day 9, the escape latency measure revealed a main effect of trial, F(3,120) = 9.51, p < 0.001, a 2-way interaction of trial and housing, F(3,120) = 3.75, p = 0.014, and a 3-way interaction between trial, sex, and housing, F(3,120) = 3.11, p = 0.03. The 3-way interaction appears to be driven by the poor improvement in performance of SE females in particular. Post hoc within-trial 2-way ANOVAs using housing and sex as between subjects factors revealed a significant interaction of housing and sex at Trial 2, F(1,40) = 5.136, p = 0.029, and a significant effect of housing at Trial 3, F(1,40) = 5.23, p = 0.028 (Fig. 6E). The path length measure showed a significant main effect of trial, F(3,120) = 8.872, p < 0.001 and a significant interaction of trial and housing, F(3,120) = 4.186,
222
T.R. Hendershott et al. / Behavioural Brain Research 314 (2016) 215–225
Fig. 5. Spatial learning and memory in SE and EE mice. (A) Mean latency to escape to the hidden platform decreased over the 7 days of training. (B) Mean path length decreased over the 7 days of training, and there was an interaction between housing and day. EE mice travelled shorter distances to escape than SE mice on Day 1 and Day 2 of training. (C) Mean time spent exhibiting thigmotaxis decreased over the 7 days of training, and there was an interaction between housing and day. EE mice displayed less thigmotaxis on Day 1 compared to SE mice. (D) Use of spatially precise search strategies increased over the 7 days of training, and there was a main effect of housing (p < 0.01). EE mice made greater use of spatially precise strategies than SE mice. (E) In the 60 s probe test, mice showed a preference for the target quadrant. (F) During the probe test, SE females swam slower than SE males and EE females. Asterisks indicate level of statistical significance of post hoc ANOVA or Bonferroni corrected t-tests: *p < 0.05, **p < 0.01, ***p < 0.001.
p = 0.007. Post hoc univariate ANOVAs revealed a significant effect of housing at Trial 3, F(1,42) = 4.918, p = 0.032 (Fig. 6F). The thigmotaxis measure revealed no significant main effects of trial, housing, or sex, but a significant interaction between trial and housing, F(3,120) = 2.80, p = 0.043 (Fig. 6G). EE mice (sexes collapsed) also made significantly greater use of spatially precise strategies on Trial 4, 2-tailed Fisher Exact Probability Test p = 0.001 (Fig. 6H). 4. Discussion and conclusions The present findings show that the exposure of mice to a simple laboratory-based EE protocol beginning at weaning age affects
emotionally modulated behavior and cognition. We found positive effects of enrichment on anxiety-like behavior and spatial learning, although the effects on sensorimotor gating were seemingly paradoxical. Important measures of anxiety-like behavior were different in males and females, but there were no apparent sex effects on sensory gating, or sociability. Spatial learning showed effects of enrichment, though there was an interaction with sex on measures dependent on swim velocity.
T.R. Hendershott et al. / Behavioural Brain Research 314 (2016) 215–225
223
Fig. 6. Within-session learning in SE and EE mice. (A) On Day 1, EE mice found the hidden platform faster than SE mice in the 2nd and 3rd trials. (B) On Day 1, EE mice swam shorter distances to reach the hidden platform than SE mice in Trial 3. (C) On Day 1, EE mice exhibited more thigmotaxis than SE mice in Trials 2 and 3. (D) On Day 1, more EE mice made use of spatially precise strategies than SE mice in the 3rd trial. (E) On Day 9, EE mice found the new platform faster than SE mice in Trial 3. (F) On Day 9, EE mice swam shorter distances to reach the new platform than SE mice in Trial 3. (G) Levels of thigmotaxis were low for all groups on Day 9. (H) On Day 9, strikingly more EE mice used spatially precise strategies than SE mice on Trial 4 of the new platform task. Asterisks indicate level of statistical significance of post hoc ANOVA or Fisher Exact tests: *p < 0.05, **p < 0.01, ***p < 0.001.
4.1. Anxiety-like behavior Our EPM results indicate an enhancement of activity with EE, more clearly than a specific reduction of anxiety-like behavior. The
housing-related difference in the overall pattern of EPM behavior should be interpreted carefully, as the decrease in latency to enter an open arm and increase in number of open arm entries do not support the conclusion that EE mice exhibited a greater preference
224
T.R. Hendershott et al. / Behavioural Brain Research 314 (2016) 215–225
for open arms over closed arms. Thus, there is a need to systematically explore the effects of EE on emotional regulation in male and female mice using a variety of assays, at least some of which are less dependent on locomotor activity or general exploratory behavior. Based on the observed sex-related behavioral differences in the EPM, we caution against the analysis of combined data from male and female mice without accounting for sex as a factor for this test. Strong sex effects were most evident in the specific EPM measures of open arm preference, which are the classic measures of benzodiazepine-sensitive anxiety-like behavior [25]. In the present study, female mice, regardless of housing, preferred open arms to a greater degree than males. This issue is particularly relevant if the experimental groups are not balanced for sex, but may be so even if they are balanced, given the relatively large variability observed within female mice. Whether this sex difference reflects a difference in anxiety is a matter for debate regarding the construct validity of this assay [26]. While we found no statistically significant change in percent open arm time with EE, we note that the mean percent open arm time of EE females was lower than that of SE females, as has been previously reported: a detailed study of only female mice reported decreased percent of open arm time in the EPM, hypolocomotion, and increased food neophobia associated with EE [8]. Detailed analysis of important considerations in understanding the effects of EE on emotionality in female rodents has been undertaken by previous authors [11].
4.2. Sensory gating The decrease in PPI following EE is contrary to our hypothesis, which was based on the association of PPI deficits with neuropsychiatric disorders and the widespread use of PPI as a phenotype for preclinical testing of antipsychotic drugs. Previous studies found mixed effects of EE on PPI in mice depending on duration of enrichment [16] or presence of a running wheel [8]. A study of male and female rats reported an attenuating effect of EE on PPI in both sexes [15]. This paradoxical result does not seem to be related to a change in the amplitude of the ASR in our study, and does, as such, appear to be indicative of a specific effect on sensory gating. Effects of EE on sensory gating should be systematically examined in future studies, using controlled variations in EE protocol as well as PPI testing protocol, as the discordant effects reported to date raise serious questions about how the PPI phenomenon should fit into the understanding of diasthesis-stress models of psychiatric vulnerability. Mouse models of neuropsychiatric disorders that make use of EE should account for the effects of EE on PPI in the absence of other manipulations in the interpretation of pharmacologic and/or genetic effects.
4.3. Sociability EE did not result in specific effects on sociability in this study, as effects of housing and sex did not reach statistical significance. The pattern of results in the SI assay hint at the possibility that male EE mice spend less time in the center of the apparatus and more time with the target mouse, but this data set is not sufficient to support that conclusion. Future studies should use larger sample sizes for statistical power to detect subtle effects in this domain. Furthermore, the measure of sociability we employed was limited in that the social target mouse was contained in a holding cell. Thus, the measures obtained do not give any information about specific social behaviors or the valence of social interactions, as would a free social interaction assay. A social interaction assay that allowed for more specific behaviors may allow for greater insight into EE effects on affiliative and aggressive behaviors [17]. It is possible that large
differences in the complexity of the home cage social structure or other social exposures may affect these measures as well. 4.4. Spatial learning and memory The results of the seven-day hidden platform spatial learning task show that mice enriched using our EE protocol exhibit enhanced spatial learning evident in the early stages of training. The results of the probe test did not indicate a difference in the strength of the spatial reference memory acquired after the 7 days of training. This lack of a housing effect may have been due to the high amount of training we conducted. However, the enhanced spatial learning seen in the early stages of training could indicate that EE mice were able to develop and utilize reference memories more quickly than SE mice. Although the probe test is one measure of reference memory, the hidden platform task is dependent on spatial reference memory in a more general sense throughout training. A previous report of increased spatial memory in female mice following EE, which used a 5-day acquisition protocol, may have revealed a housing-related difference because the subjects were not similarly over trained; the SE controls in that study performed at chance levels in the probe test [21]. Detailed analysis of Day 1 and Day 9 data from the present study showed that, while EE and SE mice performed similarly on the first trial of Day 1, EE mice exhibited greater trial-to-trial improvement in performance in the first session of training to a new platform location on both days. This difference is indicative of both enhanced exploratory behavior and enhanced cognitive flexibility in EE mice. Unlike EE mice, SE mice did not exhibit trial-to-trial improvement in this study, as in a previous study that failed to find an effect of a potentially deleterious genetic mutation on this measure [22]. Thus, for measures such as this, EE may be necessary to enable the observation of within-session learning. Female SE mice demonstrated longer latencies to escape in trialto-trial learning, which was consistent with decreased velocities observed during the probe trial. Sex should therefore be included as a factor in the analysis of latency in particular, though similar performance on other measures may belie underlying sex differences in neural processes [27–30]. Though latency is the classical and most commonly reported outcome measure for MWM tasks, the measures of distance and spatial strategy were more informative of sex and housing effects on spatial cognition. In addition to the standardization and detailed documentation of housing conditions, reporting of multiple outcome measures may help to reconcile research findings of different studies. 4.5. Conclusions Several studies have shown evidence that EE is beneficial in counteracting the effects of aging, pharmacological challenges, and deleterious mutations (for recent reviews, see [4,31]). However, the effects of EE on emotion and cognition in mice are difficult to extract from the published literature due to lack of standardization in EE and legitimate variation in behavioral testing decisions to suit specific research interests. We believe that this study provides an easily reproducible EE protocol and detailed analysis of several behavioral results of common interest. We observed strong effects of housing on trial-to-trial spatial learning, which involves working memory and requires cognitive flexibility. As we hypothesized, EE improved the cognitive function of male and female mice, increasing the range over which the effects of experimental manipulations might be quantitated. The specific pattern of MWM results we observed is consistent with actions of EE upon the circuitry of the hippocampus and frontal cortex, as suggested by previously described neural correlates [4].
T.R. Hendershott et al. / Behavioural Brain Research 314 (2016) 215–225
Recent emphasis in the research community on the need to include female animal subjects stems from a collective desire for improved scientific understanding of basic neural processes underlying behavior. This desire is coupled with a mandate to prevent the propagation of assumptions of generalizability from preclinical research that may eventually create or exacerbate gender-based health disparities. This shift has highlighted the deficit of information in the published literature on sex differences, even in commonly applied tests. We have provided direct analyses of behavioral sex differences, with caveats: larger samples sizes are required to detect subtle differences, and similarities in behavioral measures do not preclude differences at the level of neural processing [30]. We note that exercise is an important aspect of animal health, and our decision to omit a running wheel from our housing conditions is not a rejection of the need to provide greater access to exercise to laboratory animals. On the contrary, exercise is necessary for normal metabolic health [32]. Indeed, the effects of our EE protocol, though it did not include a running wheel, might nonetheless have been mediated in part by the promotion of physical activity. Future studies will hopefully address the necessity and sufficiency of components of this protocol for the production of behavioral effects, the importance of duration and developmental stage of enrichment, and the neural processes that underlie these effects in male and female mice. Acknowledgments We thank Angelo DeNofrio, Christopher Flynn, Kevin Moriarty, Caitlin Pollard, Samantha Speroni, Catarina Teves, and Nicole Parentela for assistance with animal testing. We thank Gary Chalifoux for assistance with the maintenance of the vivarium and equipment. We are grateful to Michael Drebot for advice on quantitative data processing. We thank Grace Cavanaugh, Joseph Cristiciello, and Michael Keane for careful editing of the manuscript. This work was supported by the College of the Holy Cross. References [1] D.O. Hebb, The Organization of Behavior: A Neuropsychological Theory, Wiley & Sons Inc, New York, 1949. [2] D.O. Hebb, The effects of early experience on problem solving at maturity, Am. Psychol. 2 (1947) 306–307. [3] W.T. Greenough, J.E. Black, C.S. Wallace, Experience and brain development, Child Dev. 58 (1987) 539–559. [4] H. Hirase, Y. Shinohara, Transformation of cortical and hippocampal neural circuit by environmental enrichment, Neuroscience 280 (2014) 282–298. [5] M.R. Rosenzweig, Environmental complexity, cerebral change, and behavior, Am. Psychol. 21 (1966) 321–332. [6] L.A. Toth, The influence of the cage environment on rodent physiology and behavior: implications for reproducibility of pre-clinical rodent research, Exp. Neurol. 270 (2015) 72–77. [7] R.G. Bechara, A.M. Kelly, Exercise improves object recognition memory and induced BDNF expression and cell proliferation in cognitively enriched rats, Behav. Brain Res. 245 (2013) 96–100. [8] S. Pietropaolo, J. Feldon, E. Alleca, F. Cirulli, B.K. Yee, The role of voluntary exercise in enriched rearing: a behavioral analysis, Behav. Neurosci. 4 (2006) 787–803.
225
[9] S. Pietropaolo, J. Feldon, B.K. Yee, Environmental enrichment eliminates the anxiety phenotypes in triple transgenic mouse model of Alzheimer’s disease, Cogn. Affect. Behav. Neusosci. 3 (2014) 996–1008. [10] P. Mesa-Gresa, M. Ramos-Campos, R. Redolat, Behavioral effects of different enriched environments in mice treated with the cholinergic agonist PNU-282987, Behav. Processes 103 (2014) 117–124. [11] C. Girbovan, H. Plamondon, Environmental enrichment in female rodents: considerations in the effects on behavior and biochemical markers, Behav. Brain Res. 253 (2013) 178–190. [12] D.L. Braff, N.R. Swerdlow, M.A. Geyer, Symptom correlates of prepulse inhibition deficits in male schizophrenic patients, Am. J. Psychol. 156 (1999) 596–602. [13] G.M. McAlonan, E. Daly, V. Kumari, H.D. Critchley, T. van Amelsvoort, J. Suckling, A. Simmons, T. Sigmundsson, K. Greenwood, A. Russell, N. Schmitz, F. Happe, P. Howlin, D.G. Murphy, Brain anatomy and sensorimotor gating in Asperger’s syndrome, Brain 125 (2002) 1594–1606. [14] N.R. Swerdlow, M. Weber, Y. Que, G.A. Light, D.L. Braff, Realistic expectation of prepulse inhibition in translational models for schizophrenia research, Psychopharmacology (Berl.) 3 (2008) 331–388. ˜ M. Prunell, D. Rotlland, A. Armario, R.M. Escoriheula, Enduring effects [15] Y. Pena, of environmental enrichment from weaning to adulthood on pituitary-adrenal function, pre-pulse inhibition and learning in male and female rats, Psychoneuroendocrinology 34 (2009) 1390–1404. [16] Y. Chen, Y. Mao, D. Zhou, J. Wang, Y. Ma, Environmental enrichment and chronic restraint stress in ICR mice: effects on prepulse inhibition of startle and Y-maze spatial recognition memory, Behav. Brain Res. 212 (2010) 49–55. [17] P. Mesa-Gresa, A. Perez-Martinez, R. Redola, Environmental enrichment improves novel object recognition and enhances agonistic behavior in male mice, Aggress. Behav. 39 (2013) 269–279. [18] I. Heinla, E. Leidmaa, T. Visnapuu, M. Philips, E. Vasar, Enrichment and individual housing reinforce the differences in aggressiveness and amphetamine response in 129S6/SvSv and C57BL/6 strains, Behav. Brain Res. 267 (2014) 66–73. [19] J. Bonaccorsi, S. Cintoli, R. Mastrogiacomo, S. Baldanzi, C. Braschi, T. Pizzorusso, M.C. Cenni, N. Berardi, System consolidation of spatial memories in mice: effects of enriched environment, Neural Plast. (2013) 1–12. [20] W.E. Crusio, D. Goldowitz, A. Holmes, D. Wolfer, Standards for the publication of mouse mutant studies, Genes Brain Behav. 8 (2009) 1–4. [21] J.E. Gresack, K.M. Kerr, K.M. Frick, Life-long environmental enrichment differentially affects the mnemonic response to estrogen in young, middle-aged, and female mice, Neurobiol. Learn. Mem. 88 (4) (2007) 393–408. [22] A.C. Basu, G.E. Tsai, C.L. Ma, J.T. Ehmsen, A.K. Mustafa, L. Han, S.I. Jiang, M.A. Benneyworth, M.P. Froimowitz, N. Lange, S.H. Snyder, R. Bergeron, J.T. Coyle, Targeted disruption of serine racemase affects glutamatergic neurotransmission and behavior, Mol. Psychiatry 14 (2009) 719–727. [23] J.N. Crawley, Mouse behavioral assays relevant to the symptoms of autism, Brain Pathol. 17 (2007) 448–459. [24] J. Gil-Mohapel, P.S. Brocardo, W. Choquette, R. Gothard, J.M. Simpson, B.R. Christie, Hippocampal neurogenesis levels predict watermaze search strategies in the aging brain, PLoS One 8 (2013) 1–19, e75125. [25] U. Rudolph, F. Crestani, D. Benke, I. Brünig, J.A. Benson, J. Fritschy, J.R. Martin, H. Bluethmann, H. Möhler, Benzodiazepine actions mediated by specific ␥-aminobutyric acidA receptor subtypes, Nature 401 (1999) 796–800. [26] C. Belzung, G. Griebel, Measuring normal and pathological anxiety-like behavior in mice: a review, Behav. Brain Res. 125 (2016) 141–149. [27] J.M. Juraska, C. Henderson, J. Müller, Differential rearing experience, gender, and radial maze performance, Dev. Psychobiol. 17 (1984) 209–215. [28] J.M. Juraska, J.M. Fitch, C. Henderson, N. Rivers, Sex differences in the dendritic branching of dentate granule cells following differential experience, Brain Res. 333 (1985) 73–80. [29] K. Frick, N. Stearns, J. Pan, J. Berger-Sweeney, Effects of environmental enrichment on spatial memory and neurochemistry in middle-aged mice, Learn. Mem. 10 (2003) 187–198. [30] K. Frick, J. Tuscher, Sex steroid hormones matter for learning and memory: estrogenic regulation of hippocampal function in male and female rodents, Learn. Mem. 22 (2015) 472–493. [31] P. Sampedro-Piquero, A. Begega, Environmental enrichment as a positive behavioral intervention across the lifespan, Curr. Neuropharmacol. 14 (2016) 1–11. [32] B. Martin, S. Ji, S. Maudsley, M.P. Mattson, Control laboratory rodents are metabolically morbid: why it matters, Proc. Natl. Acad. Sci. U. S. A. 107 (2010) 6127–6133.
Contents lists available at ScienceDirect
Behavioural Brain Research journal homepage: www.elsevier.com/locate/bbr
Research report
Effects of environmental enrichment on anxiety-like behavior, sociability, sensory gating, and spatial learning in male and female C57BL/6J mice Taylor R. Hendershott, Marie E. Cronin, Stephanie Langella, Patrick S. McGuinness, Alo C. Basu ∗ Department of Psychology, College of the Holy Cross, 1 College Street, Worcester, MA 01610, United States
h i g h l i g h t s • • • • •
Female mice showed greater preference for EPM open arms regardless of housing. EE attenuated sensory gating in male and female mice. EE enhanced spatial learning in male and female mice. EE resulted in greater use of spatially precise strategies in the water maze. Swim speed and escape latency in SE females were slow relative to the other groups.
a r t i c l e
i n f o
Article history: Received 25 June 2016 Received in revised form 29 July 2016 Accepted 2 August 2016 Available online 3 August 2016 Keywords: Environmental enrichment Elevated plus maze Prepulse inhibition Water maze Spatial strategies Sex differences
a b s t r a c t The influence of housing on cognition and emotional regulation in mice presents a problem for the study of genetic and environmental risk factors for neuropsychiatric disorders: standard laboratory housing may result in low levels of cognitive function or altered levels of anxiety that leave little room for assessment of deleterious effects of experimental manipulations. The use of enriched environment (EE) may allow for the measurement of a wider range of performance in cognitive domains. Cognitive and behavioral effects of EE in male mice have not been widely reproduced, perhaps due to variability in the application of enrichment protocols, and the effects of EE in female mice have not been widely studied. We have developed an EE protocol using common laboratory equipment that, without a running wheel for exercise, results in significant cognitive and behavioral effects relative to standard laboratory housing conditions. We compared male and female wild-type C57BL/6J mice reared from weaning age in an EE to those reared in a standard environment (SE), using common measures of anxiety-like behavior, sensory gating, sociability, and spatial learning and memory. Sex was a significant factor in relevant elevated plus maze (EPM) measures, and bordered on significance in a social interaction (SI) assay. Effects of EE on anxiety-like behavior and sociability were indicative of a general increase in exploratory activity. In male and female mice, EE resulted in reduced prepulse inhibition (PPI) of the acoustic startle response, and enhanced spatial learning and use of spatially precise strategies in a Morris water maze task. © 2016 Elsevier B.V. All rights reserved.
1. Introduction Since Donald Hebb’s original observations of increased spatial learning ability in rats reared in an enriched environment (EE) [1,2], laboratory controlled enrichment has been associated with widespread molecular and morphological changes through-
∗ Corresponding author. E-mail address: [email protected] (A.C. Basu). http://dx.doi.org/10.1016/j.bbr.2016.08.004 0166-4328/© 2016 Elsevier B.V. All rights reserved.
out the brain in rodents, mainly in comparison to subjects reared in isolation (for review see Ref. [3,4]). Further, these molecular and morphological changes have been associated with behavioral changes [5]. Although some effects on the nervous system have been replicated by different research groups, behavioral and cognitive effects of EE have not been widely reproducible (Table 1). The mechanisms of enrichment-related changes are of interest for the understanding of functional neuroplasticity, and the influence of housing conditions on rodent behavior is a serious concern for the reproducibility of pre-clinical research findings [6].
216
Table 1 Variability in reports of EE effects on cognition and behavior. Species, strain
Sex
Number of Animals
Age at Enrichment
Age at Testing
Enrichment Duration
Running Wheel
EPM behavior
Sensory gating
[8]
Mouse, C57BL/6J
F
3 weeks
12 weeks
40 days
Yes, No
↓, ↓ % time in open arms
↑, ↓ PPI
[9]
F
3 weeks
25 weeks, 51 weeks
∼22 weeks, 48 weeks
Yes
↓ % open arm entries
[10]
Mouse, 129SvJ/C57BL/6 hybrids Mouse, NMRI
SE = 13 EE = 14 EER = 14 SE = 8 EER = 7
4 weeks
17 weeks
∼13 weeks
Yes
n.c.
[16]
Mouse, ICR
M
4 weeks
6 weeks
Mouse, NMRI
M
4 weeks
8 weeks
2 weeks, 4 weeks 4 weeks
Yes
[17]
SE = 8 EER = 8 SE = 12 EER = 12 SE = 16 EER = 16
Yes
[18]
Mouse, C57BL/6J, M 129S6/SvEv/Tac
3 weeks
10 weeks
7 weeks
Yes
[19]
Mouse, C57BL/6J
8 weeks
14 weeks
40 days
Yes
↓ latency
[21]
Mouse, C57Bl/6
SE B6 = 8 EER B6 = 8 SE 129 = 8 EER 129 = 8 SE = 24 EER = 24 SE = 10 EER = 11 SE-M = 11 SE-F = 12 EE-M = 10 EE-F = 11
3 weeks
20 weeks
17 weeks
Yes
↑ probe
4 weeks
8 weeks
4 weeks
No
Present Study Mouse, C57BL/6J
M
M and F (combined) Female M and F (sex as factor)
Change in EE housed mice relative to SE housed mice: n.c. no change; ↓ decrease; ↑ increase; *interaction of housing and sex, R running wheel.
Social behavior
MWM Spatial learning and memory n.c.
↑, ↓ PPI ↑social behavior ↑aggressive behavior n.c.
↑ latency to enter ↓ PPI open arm ↑ open arm entries
n.c.
↓ latency* ↓ duration ↓ thigmotaxis ↑ spatial precision
T.R. Hendershott et al. / Behavioural Brain Research 314 (2016) 215–225
Ref.
T.R. Hendershott et al. / Behavioural Brain Research 314 (2016) 215–225
Important factors to consider in the interpretation and application of findings from studies using EE include: variability in laboratory-controlled enrichment protocols, developmental stage and duration of enrichment, and species, strain, sex, and age of animal subjects. Of particular concern for interpretation of behavioral outcomes is the use of a running wheel in some enrichment protocols. Exercise alone has been shown to improve object recognition memory and to induce brain derived neurotrophic factor (BDNF) expression in male rats [7], and interaction of enrichment and exercise on sensory gating and spatial memory has been previously documented in mice [8]. There is also a significant gap in this literature regarding sex: few studies of cognition have focused on female rodents or compared them directly to males. We selected a battery of behavioral and cognitive tests that are frequently used to study mouse models of psychiatric disorders and administered it to EE and SE housed male and female mice of the C57BL/6J strain. We used the elevated plus maze (EPM) to assess anxiety-like behavior. We hypothesized that animals reared in EE would display reduced anxiety-like behavior in the EPM, though not all previous research is consistent with this hypothesis (Table 1, [9,10]). We noted the possibility that EE might in some cases result in anxietylike behavior due to increased stress, or changes in activity (for review, see Ref. [11]), and therefore planned to analyze multiple measures for this test that would reveal such effects. We measured PPI of the acoustic startle response, a sensorimotor gating phenomenon observed widely in vertebrates, because it is considered a psychiatric endophenotype. Attenuated PPI has been observed in disorders in which neurodevelopmental factors are implicated, such as schizophrenia [12] and autism [13]. The measurement of this phenomenon in rodent models of neuropsychiatric disorders is widespread, and the field has been cautioned against leaving assumptions of cross-species comparability in neural substrates untested [14]. Results of previous studies of EE and PPI in rodents are mixed; in some cases EE attenuated PPI and in other cases there was enhancement (Table 1, [8,15,16]). We hypothesized that enrichment would result in an increase in social behavior, but the results of past literature have been, again, mixed (Table 1, [17,18]). Previous studies of male mice housed subjects in larger social groups than is standard (up to 8 per cage), and used social behavior assays that allowed free interaction between the subject and the target mouse. We chose to control for complexity of social experience in our housing conditions by keeping the number of cage mates equivalent between EE and SE. We also chose to use a limited social interaction assay that sacrifices information about the dyadic social behaviors in favor of a focus on the behavioral choices of the subject. We used a standard MWM test of spatial learning and memory to evaluate the cognitive function of male and female mice following EE. Bonaccorsi et al. [19] found that EE male and female C57BL/6 mice reached the hidden platform faster than SE mice late in training. Although data from male and female mice were included, sex was not included as a factor in the statistical analysis. Thus, it was not possible to ascertain whether males and females responded differently to behavioral and cognitive assays as a result of the experimental condition [20]. There are previous reports in the literature of enhanced MWM performance in female mice following EE, but without direct comparison to males (Table 1, [8,21]). The effects of EE on emotional regulation and cognition in male and female wild type (WT) mice are, thus, yet to be firmly established. We have summarized published results of comparable studies in Table 1, selecting studies of young (≤6 months) WT mice that used similar behavioral assays and SE (as opposed to socially isolated) controls. The reproducibility of these effects may be improved by the standardization of enrichment protocols, in particular if they can be obtained using common rather than specialized laboratory equipment. Therefore, we conducted a study
217
that uses a simple EE, does not introduce exercise effects by way of a running wheel, and that uses WT mice reared in standard social housing conditions as the control group. We measured emotional and cognitive outcomes commonly of interest in neuropsychiatric disease research in both males and females, including sex as a factor in data analysis to allow the direct identification of sex effects [20]. We hypothesized that EE mice would exhibit less anxiety-like behavior, increased social behavior, increased PPI, and improved spatial learning and memory relative to SE WT mice.
2. Materials and methods 2.1. Animal subjects, general vivarium conditions, and handling procedures Animal subjects were bred in the College of the Holy Cross vivarium from WT C57BL/6J parents within 5 generations of import from Jackson Laboratory, Bar Harbor, ME, USA. Breeding procedure was to co-house a single male breeder with 3 virgin female breeders, checking weekly for pregnancy. Pregnant dams were isolated in the final week of pregnancy and left undisturbed until the end of the first postnatal week. Mice were weaned and randomly assigned to four different housing groups at approximately 4 weeks of age (P2430). Littermates were separated and assigned to different housing conditions to the extent possible to mitigate potential litter effects. Forty-four WT C57BL/6J mice obtained from nine litters were used in this study. Final experimental group sizes were: 11 SE males, 10 EE males, 12 SE females, and 11 EE females. All animals had access to food and water ad libitum and caging was changed weekly. Ambient temperature was maintained at 21 ± 1 ◦ C. The vivarium lighting was regulated on a 12 h:12 h cycle. Behavioral testing was conducted during the light phase, avoiding the phase transition by 2 h. The study was conducted in 3 separate batches, nearly balanced for experimental groups. All husbandry and experimental procedures were approved by the College of the Holy Cross Institutional Animal Care and Use Committee and carried out in accordance with the National Institutes of Health Guide for the Care and Use of Laboratory Animals.
2.2. Enrichment protocol SE control mice were reared in standard laboratory conditions in clear plastic caging. The inner cage dimensions were approximately: 28 cm × 17.2 cm × 12.2 cm (ANCARE, Inc. standard mouse cage). Mice were housed in groups of 2–5 per cage. SE cages were lined with corn cob bedding and provided with a 5.1 cm × 5.1 cm square of cotton nesting material, changed weekly. The EE cages had larger inner dimensions, approximately: 45.4 cm × 23.5 cm × 15.7 cm (ANCARE, Inc. standard rat cage). EE cages were furnished with two squares of nesting material, a plastic in-cage shelter, and a variety of novel objects that were changed weekly during the first 4 weeks post-weaning (Fig. 1). To facilitate consistent application of this protocol, we selected common laboratory items as novel objects: bottle caps and segments of PVC tubing in week 1, severed sample tubes and beakers (polystyrene, 50 ml) in week 2, pipette tips (1000 ml), a polystyrene funnel and a stationary plastic disk of a running wheel in week 3, and half of a wire test tube rack in week 4. All cages were fitted with a wire top rack that held food and a single water bottle. All cages were non-ventilated and kept covered with a filter lid. The larger caging, nesting material, and in-cage shelter were continued for the duration of the experiment.
218
T.R. Hendershott et al. / Behavioural Brain Research 314 (2016) 215–225
Fig. 1. Experimental timeline and enrichment protocol. For 4 weeks following weaning, EE mice were exposed to novel objects changed weekly. Extra nestlet, plastic dome, and larger cages were provided for the duration of the experiment. SE mice were littermates of EE mice, weaned at the same time, reared in parallel and tested together with EE mice.
2.3. Elevated plus maze (EPM) The EPM test was conducted at age 8 weeks ± 4 days. The EPM was constructed of opaque plexiglass. The four arms were 30.5 cm long × 5.7 cm wide and extended from a center area (7 cm diagonal); the maze floor was elevated 52.7 cm above the ground. The walls of the two opposing closed arms were 21 cm high. The light levels for testing were ∼100 lx in the open arms and ∼35 lx in the closed arms. Test subjects were habituated to the testing room for at least half an hour prior to testing. At the start of each EPM trial, the test subject was placed in the center of the maze facing an open arm. The mouse was allowed to explore the maze freely for 5 min. The location of the mouse in the maze was tracked continuously using data capture hardware and EthovisionXT software (Noldus Information Technology, Wageningen, the Netherlands). 2.4. Prepulse inhibition (PPI) of the acoustic startle response (ASR) The ASR was measured using the SR-Lab Startle Response System for mice (San Diego Instruments, Inc., San Diego, California, USA). PPI of the ASR was conducted at 8–9 weeks of age using a previously published protocol [22]. Testing was conducted with lights off in the sound attenuating testing chambers. In brief, each session consisted of 5 min of white noise at a background level of 70 dB followed by 72 test trials. The first six “leader” trials and the last six “trailer” trials delivered only a 120 dB startle stimulus, to allow for stabilization of the ASR and measurement of within-session habituation. The remaining trials were of the following types, in equal numbers, and presented in pseudorandom order: no stimulus, 120 dB test stimulus, 120 dB stimulus with prepulse of 73 dB (+3 dB), 120 dB stimulus with prepulse of 76 dB (+6 dB), 120 dB stimulus with prepulse of 82 dB (+12 dB). All test subjects exhibited the ASR, as determined by comparison of the baseline reading in the no stimulus trials to the 120 dB stimulus test trials by unpaired Stu-
dent’s t-test for each subject. Percent PPI was calculated for each prepulse intensity for each test subject using the following equation: (Average response to 120 dB test stimulus − Average response to 120 dB stimulus preceded by prepulse)/Average response to 120 dB test stimulus. 2.5. Social interaction (SI) A SI assay was conducted at 8–11 weeks of age, using a nonautomated apparatus in an adaptation of a published protocol [23]. The test apparatus, constructed of clear plexiglass, consisted of three 40.5 cm × 20.3 cm × 30.5 cm chambers separated by doors. The lighting in all three chambers was kept at ∼20 lx for testing. Social targets were sex-matched adult mice of the same strain that were not familiar to the subject (not parent or littermate). Target mice were habituated to the holders for 30 min on the day before testing. Test subjects were habituated to the testing room for at least half an hour prior to testing. At the beginning of each trial, the subject was restricted to the center chamber for 5 min for habituation to the apparatus. After 5 min, a novel sex-matched target mouse was placed in a cylindrical holder in one of the side chambers and the doors were opened, giving the subject access to all three chambers. Each subject was given 10 min to freely explore the apparatus. The test subject was tracked using the automated system described in Section 2.3. 2.6. Morris water maze (MWM) A nine-day MWM protocol was performed at age 9–12 weeks to test spatial learning and memory as previously described [22]. Briefly, mice were trained to find a hidden platform using intra- and extra-maze cues for 7 days (4 trials/day, pseudorandomized point of entry to maze, maximum trial duration 90 s, inter-trial interval approximately 10 min). On the 8th day, the hidden platform was
T.R. Hendershott et al. / Behavioural Brain Research 314 (2016) 215–225
removed and subjects were allowed to explore the maze for 60 s (Probe test). On the 9th day, the hidden platform was placed in a new location within the maze (New Platform), and the test subjects were trained to find it for 4 trials. The testing apparatus was a commercially available water maze for mice (San Diego Instruments, Inc., San Diego, California, USA). The lighting in the maze was kept at ∼20 lx for testing. Non-toxic white paint was used to make the water opaque, and water temperature was maintained at 22 ± 1 ◦ C. The test subjects were tracked using the automated system described in section 2.3. Scoring of search strategies as spatially precise or spatially imprecise was conducted post-testing by observers blind to experimental group, according to previously published criteria [24]. Briefly, to be classified as Spatially Precise (SP), a search fit one of the following three descriptions: the subject took a direct route to the platform from entry (direct swim), the subject showed a preference for a corridor from the point of entry towards the platform (directed search), or the subject performed a highly localized search within the target quadrant (focal search). If a search did not meet any of these criteria, it was classified as Spatially Imprecise (SI), a label that encompassed several patterns of behavior such as chaining, scanning, thigmotaxis, perseverance, and random search [24]. 2.7. Statistical analysis Data were managed using Microsoft Excel software. Statistical analyses were conducted using IBM SPSS Statistics 23 software. Data were analyzed by analysis of variance (ANOVA) with housing and sex as between-subjects factors and any repeated measures as within-subjects factors. Restricted ANOVAs or Bonferroni corrected t-tests were conducted post hoc as noted in the results section. 3. Results 3.1. Elevated plus maze (EPM) EPM data were analyzed by 2-way ANOVA using housing and sex as between-subjects factors. All significant main effects of housing and sex are reported in this section. There were no significant interactions between housing and sex. Mice housed in EE entered an open arm of the EPM sooner than mice housed in SE, main effect of housing, F(1,40) = 4.49, p = 0.040 (Fig. 2A). Mice housed in EE also made more entries to the open arms, main effect of housing, F(1,40) = 7.33, p = 0.010 (Fig. 2B). EE mice made more arm entries in total, main effect of housing, F(1,40) = 8.92, p = 0.005 (Fig. 2C), and exhibited greater locomotor activity within the maze, main effect of housing, F(1,40) = 14.08, p < 0.001 (Fig. 2D), than mice housed in SE. Important measures of anxiety-like behavior revealed significant effects of sex, but not housing: Female mice spent a higher percentage of arm time in the open arms, main effect of sex, F(1,40) = 6.44, p = 0.015 (Fig. 2E), and a greater raw amount of time in the open arms, main effect of sex, F(1,40) = 6.03, p = 0.019 (Fig. 2F). There were borderline effects of sex on the percentage of arm entries made to open arms, F(1,40) = 3.45, p = 0.071 (Fig. 2G), and on the total raw time spent in arms, F(1,40) = 3.67, p = 0.063 (Fig. 2H). 3.2. Prepulse inhibition (PPI) of the acoustic startle response (ASR) Startle data were analyzed by 3-way ANOVA using housing and sex as between-subjects factors and trial type as a withinsubjects factor. All significant main effects of housing, sex, and trial type are reported in this section. There were no significant interactions. There were no significant effects of housing or sex on startle reactivity, though there was significant within-session startle habituation evident in the decrease in startle amplitude between the leader and trailer 120 dB stimulus trials, main effect of trial
219
type, F(1,40) = 27.356, p < 0.001 (Fig. 3A). As expected, percent PPI increased with prepulse intensity, main effect of prepulse intensity, F(2,80) = 88.541, p < 0.001 (Fig. 3B). There was a significant effect of housing on the percent PPI such that mice housed in EE exhibited lower percent PPI than mice housed in SE, F(1,40) = 6.655, p = 0.014 (Fig. 3B). According to a 2-way ANOVA, startle reactivity to the 120 dB startle stimulus in the trials that were interspersed with the prepulse test trials was not affected by housing or sex. 3.3. Social interaction Social interaction data were analyzed by 3-way ANOVA using housing and sex as between-subjects factors and chamber (target or empty chamber within the testing apparatus) as a within-subjects factor. There were no significant effects of housing or sex on time spent in the different chambers in the apparatus, though there was a significant main effect of chamber type, F(1,40) = 6.096, p = 0.018, indicating that the mice were sociable (Fig. 4A). The effect of housing was marginal on this measure, F(1,40) = 3.225, p = 0.08, as was the effect of sex, F(1,40) = 3.160, p = 0.083. However, housing did affect the percentage of trial time spent in the center of the apparatus, in that EE mice spent less time in the center, and therefore more time exploring the side chambers: 2-way ANOVA revealed a significant main effect of housing, F(1,40) = 4.399, p = 0.042 (Fig. 4B). The effect of sex was marginal on this measure, F(1,40) = 3.458, p = 0.07. 3.4. Morris water maze All MWM data were analyzed by 3-way ANOVA with housing and sex as between-subjects factors and either day or trial as a repeated-measures factor unless otherwise noted. Post hoc analyses were conducted as noted. No significant effects of housing or sex were observed on average daily latency to reach the hidden platform during the 7 days of acquisition. However, 3-way ANOVA revealed a significant main effect of day, F(6,240) = 41.55, p < 0.001 (Fig. 5A). The measure of average path length showed a significant main effect of day, F(6,240) = 55.39, p < 0.001, and also a significant interaction between day and housing, F(6,240) = 3.516, p = 0.002. Follow-up within-day univariate ANOVA revealed significant effects of housing on Day 1, F(1,42) = 19.98, p < 0.001, and Day 2, F(1,42) = 4.80, p = 0.034, such that EE mice took shorter paths to reach the hidden platform (Fig. 5B). Percentage of time spent in thigmotaxis was affected by day, F(6,240) = 90.46, p < 0.001, decreasing as subjects learned the location of the hidden platform. There was a significant interaction between day and housing in the thigmotaxis measure, F(6,240) = 4.217, p < 0.001. Follow-up withinday univariate ANOVA revealed a significant effect of housing on thigmotaxis on Day 1, F(1,42) = 11.258, p = 0.002 (Fig. 5C). The spatial strategy scores showed significant main effects of both day, F(6,240) = 30.18, p < 0.001, and housing, F(1,40) = 9.25, p = 0.004, reflecting higher use of spatially precise strategies by EE mice throughout hidden platform acquisition (Fig. 5D). The housing-related differences in acquisition were not followed by differences in spatial reference memory according to the results of the probe test. All test groups demonstrated preference for the target quadrant above chance, and 3-way ANOVA with housing and sex as between-subjects factors and quadrant as a within-subjects factor revealed only a significant main effect of quadrant, F(3,120) = 21.302, p < 0.001 (Fig. 5E). However, in the probe test, there was an interaction of housing and sex with respect to swim velocity, F(1,40) = 6.018, p = 0.02, such that a sex difference was observed in SE but not EE mice. Post hoc Bonferroni corrected t-tests revealed that SE females had lower swim velocity than EE females, t(21) = 2.83, p < 0.05, and SE males, t(21) = 2.81, p < 0.05, but there were no significant differences between EE and SE males, or between male and female EE mice (Fig. 5F). There were no sig-
220
T.R. Hendershott et al. / Behavioural Brain Research 314 (2016) 215–225
Fig. 2. Behavior of SE and EE mice in the EPM. (A) EE mice entered an open arm sooner than SE mice. (B) EE mice entered open arms more often than SE mice. (C) EE mice made more total arm entries than SE mice. (D) EE mice travelled a greater distance within the maze than SE mice. (E) Female mice spent a greater percentage of arm time in the open arms than males regardless of housing condition. (F) Females also spent more time overall in open arms than male mice. (G) There was no significant difference in percentage of entries to open arms between groups. (H) There was no significant difference in the total time spent in arms between groups. Asterisks indicate statistical significance of reported main effects: *p < 0.05, **p < 0.01, ***p < 0.001.
nificant effects of sex or housing on any daily-averaged outcome measures on Day 9, when a new platform location was introduced (Figs. 5A–D). The finding that housing-related differences were generally observed in Day 1 of water maze acquisition prompted us to conduct a closer analysis of trial-by-trial performance in that session, in parallel with that planned for new platform acquisition on Day 9. Trial-by-trial analyses of water maze performance on Day 1, the first day of hidden platform training, and Day 9, the reversal learning day, revealed effects of housing on within-session trialby-trial improvement in performance. EE mice were similar to their SE counterparts in Trial 1, but improved in performance more rapidly. On Day 1, the escape latency measure revealed a main
effect of trial, F(3,120) = 6.19, p = 0.001, a main effect of housing F(1,40) = 10.89, p = 0.002, and an interaction of trial and housing, F(3,120) = 3.39, p = 0.02. Post hoc univariate ANOVAs revealed significant effects of housing at Trial 2, F(1,42) = 13.37, p = 0.001, and Trial 3, F(1,42) = 10.07, p = 0.003 (Fig. 6A). The path length measure showed a similar pattern on Day 1: a significant main effect of trial, F(3,120) = 7.90, p < 0.001, a significant main effect of housing, F(1,40) = 18.66, p < 0.001, and a significant interaction of trial and housing, F(3,120) = 4.5, p = 0.005. Post hoc univariate ANOVAs revealed significant effects of housing at Trial 2, F(1,42) = 12.73, p = 0.001, and Trial 3, F(1,42) = 15.43, p < 0.001 (Fig. 6B). The thigmotaxis measure revealed no effect of trial, but a significant main effect of housing, F(1,40) = 10.95, p = 0.002, and a significant inter-
T.R. Hendershott et al. / Behavioural Brain Research 314 (2016) 215–225
221
Fig. 3. Acoustic startle and sensory gating in SE and EE mice. (A) Startle amplitude decreased significantly between the first six “leader” and last six “trailer” 120 dB stimulus trials. (B) There was a significant decrease in the percent PPI in EE mice as compared to SE mice. Asterisks indicate statistical significance of reported main effects: *p < 0.05, ***p < 0.001.
Fig. 4. Sociability in SE and EE mice. (A) Mice preferred the target chamber to the empty chamber in the social interaction assay. (B) EE mice spent less time in the center chamber than SE mice. Asterisks indicate statistical significance of reported main effects: *p < 0.05.
action between trial and housing, F(3,120) = 2.84, p = 0.041. Post hoc univariate ANOVAs revealed significant effects of housing at Trial 2, F(1,42) = 10.51, p = 0.002, and Trial 3, F(1,42) = 8.41, p = 0.006 (Fig. 6C). EE mice (sexes collapsed) also made significantly greater use of spatially precise strategies on Trial 3, 2-tailed Fisher Exact Probability Test p = 0.002 (Fig. 6D). On Day 9, though the hidden platform was moved to a new location to test reversal learning, the conventional performance measures of escape latency and path length were lower and closer between experimental groups than on Day 1. Thigmotaxis was much lower than on Day 1. The overall decrease in these measures was expected given the amount of MWM training that transpired between the two sessions. The use of spatial strategy analysis was
thus particularly helpful in revealing an advantage of EE in withinsession learning. On Day 9, the escape latency measure revealed a main effect of trial, F(3,120) = 9.51, p < 0.001, a 2-way interaction of trial and housing, F(3,120) = 3.75, p = 0.014, and a 3-way interaction between trial, sex, and housing, F(3,120) = 3.11, p = 0.03. The 3-way interaction appears to be driven by the poor improvement in performance of SE females in particular. Post hoc within-trial 2-way ANOVAs using housing and sex as between subjects factors revealed a significant interaction of housing and sex at Trial 2, F(1,40) = 5.136, p = 0.029, and a significant effect of housing at Trial 3, F(1,40) = 5.23, p = 0.028 (Fig. 6E). The path length measure showed a significant main effect of trial, F(3,120) = 8.872, p < 0.001 and a significant interaction of trial and housing, F(3,120) = 4.186,
222
T.R. Hendershott et al. / Behavioural Brain Research 314 (2016) 215–225
Fig. 5. Spatial learning and memory in SE and EE mice. (A) Mean latency to escape to the hidden platform decreased over the 7 days of training. (B) Mean path length decreased over the 7 days of training, and there was an interaction between housing and day. EE mice travelled shorter distances to escape than SE mice on Day 1 and Day 2 of training. (C) Mean time spent exhibiting thigmotaxis decreased over the 7 days of training, and there was an interaction between housing and day. EE mice displayed less thigmotaxis on Day 1 compared to SE mice. (D) Use of spatially precise search strategies increased over the 7 days of training, and there was a main effect of housing (p < 0.01). EE mice made greater use of spatially precise strategies than SE mice. (E) In the 60 s probe test, mice showed a preference for the target quadrant. (F) During the probe test, SE females swam slower than SE males and EE females. Asterisks indicate level of statistical significance of post hoc ANOVA or Bonferroni corrected t-tests: *p < 0.05, **p < 0.01, ***p < 0.001.
p = 0.007. Post hoc univariate ANOVAs revealed a significant effect of housing at Trial 3, F(1,42) = 4.918, p = 0.032 (Fig. 6F). The thigmotaxis measure revealed no significant main effects of trial, housing, or sex, but a significant interaction between trial and housing, F(3,120) = 2.80, p = 0.043 (Fig. 6G). EE mice (sexes collapsed) also made significantly greater use of spatially precise strategies on Trial 4, 2-tailed Fisher Exact Probability Test p = 0.001 (Fig. 6H). 4. Discussion and conclusions The present findings show that the exposure of mice to a simple laboratory-based EE protocol beginning at weaning age affects
emotionally modulated behavior and cognition. We found positive effects of enrichment on anxiety-like behavior and spatial learning, although the effects on sensorimotor gating were seemingly paradoxical. Important measures of anxiety-like behavior were different in males and females, but there were no apparent sex effects on sensory gating, or sociability. Spatial learning showed effects of enrichment, though there was an interaction with sex on measures dependent on swim velocity.
T.R. Hendershott et al. / Behavioural Brain Research 314 (2016) 215–225
223
Fig. 6. Within-session learning in SE and EE mice. (A) On Day 1, EE mice found the hidden platform faster than SE mice in the 2nd and 3rd trials. (B) On Day 1, EE mice swam shorter distances to reach the hidden platform than SE mice in Trial 3. (C) On Day 1, EE mice exhibited more thigmotaxis than SE mice in Trials 2 and 3. (D) On Day 1, more EE mice made use of spatially precise strategies than SE mice in the 3rd trial. (E) On Day 9, EE mice found the new platform faster than SE mice in Trial 3. (F) On Day 9, EE mice swam shorter distances to reach the new platform than SE mice in Trial 3. (G) Levels of thigmotaxis were low for all groups on Day 9. (H) On Day 9, strikingly more EE mice used spatially precise strategies than SE mice on Trial 4 of the new platform task. Asterisks indicate level of statistical significance of post hoc ANOVA or Fisher Exact tests: *p < 0.05, **p < 0.01, ***p < 0.001.
4.1. Anxiety-like behavior Our EPM results indicate an enhancement of activity with EE, more clearly than a specific reduction of anxiety-like behavior. The
housing-related difference in the overall pattern of EPM behavior should be interpreted carefully, as the decrease in latency to enter an open arm and increase in number of open arm entries do not support the conclusion that EE mice exhibited a greater preference
224
T.R. Hendershott et al. / Behavioural Brain Research 314 (2016) 215–225
for open arms over closed arms. Thus, there is a need to systematically explore the effects of EE on emotional regulation in male and female mice using a variety of assays, at least some of which are less dependent on locomotor activity or general exploratory behavior. Based on the observed sex-related behavioral differences in the EPM, we caution against the analysis of combined data from male and female mice without accounting for sex as a factor for this test. Strong sex effects were most evident in the specific EPM measures of open arm preference, which are the classic measures of benzodiazepine-sensitive anxiety-like behavior [25]. In the present study, female mice, regardless of housing, preferred open arms to a greater degree than males. This issue is particularly relevant if the experimental groups are not balanced for sex, but may be so even if they are balanced, given the relatively large variability observed within female mice. Whether this sex difference reflects a difference in anxiety is a matter for debate regarding the construct validity of this assay [26]. While we found no statistically significant change in percent open arm time with EE, we note that the mean percent open arm time of EE females was lower than that of SE females, as has been previously reported: a detailed study of only female mice reported decreased percent of open arm time in the EPM, hypolocomotion, and increased food neophobia associated with EE [8]. Detailed analysis of important considerations in understanding the effects of EE on emotionality in female rodents has been undertaken by previous authors [11].
4.2. Sensory gating The decrease in PPI following EE is contrary to our hypothesis, which was based on the association of PPI deficits with neuropsychiatric disorders and the widespread use of PPI as a phenotype for preclinical testing of antipsychotic drugs. Previous studies found mixed effects of EE on PPI in mice depending on duration of enrichment [16] or presence of a running wheel [8]. A study of male and female rats reported an attenuating effect of EE on PPI in both sexes [15]. This paradoxical result does not seem to be related to a change in the amplitude of the ASR in our study, and does, as such, appear to be indicative of a specific effect on sensory gating. Effects of EE on sensory gating should be systematically examined in future studies, using controlled variations in EE protocol as well as PPI testing protocol, as the discordant effects reported to date raise serious questions about how the PPI phenomenon should fit into the understanding of diasthesis-stress models of psychiatric vulnerability. Mouse models of neuropsychiatric disorders that make use of EE should account for the effects of EE on PPI in the absence of other manipulations in the interpretation of pharmacologic and/or genetic effects.
4.3. Sociability EE did not result in specific effects on sociability in this study, as effects of housing and sex did not reach statistical significance. The pattern of results in the SI assay hint at the possibility that male EE mice spend less time in the center of the apparatus and more time with the target mouse, but this data set is not sufficient to support that conclusion. Future studies should use larger sample sizes for statistical power to detect subtle effects in this domain. Furthermore, the measure of sociability we employed was limited in that the social target mouse was contained in a holding cell. Thus, the measures obtained do not give any information about specific social behaviors or the valence of social interactions, as would a free social interaction assay. A social interaction assay that allowed for more specific behaviors may allow for greater insight into EE effects on affiliative and aggressive behaviors [17]. It is possible that large
differences in the complexity of the home cage social structure or other social exposures may affect these measures as well. 4.4. Spatial learning and memory The results of the seven-day hidden platform spatial learning task show that mice enriched using our EE protocol exhibit enhanced spatial learning evident in the early stages of training. The results of the probe test did not indicate a difference in the strength of the spatial reference memory acquired after the 7 days of training. This lack of a housing effect may have been due to the high amount of training we conducted. However, the enhanced spatial learning seen in the early stages of training could indicate that EE mice were able to develop and utilize reference memories more quickly than SE mice. Although the probe test is one measure of reference memory, the hidden platform task is dependent on spatial reference memory in a more general sense throughout training. A previous report of increased spatial memory in female mice following EE, which used a 5-day acquisition protocol, may have revealed a housing-related difference because the subjects were not similarly over trained; the SE controls in that study performed at chance levels in the probe test [21]. Detailed analysis of Day 1 and Day 9 data from the present study showed that, while EE and SE mice performed similarly on the first trial of Day 1, EE mice exhibited greater trial-to-trial improvement in performance in the first session of training to a new platform location on both days. This difference is indicative of both enhanced exploratory behavior and enhanced cognitive flexibility in EE mice. Unlike EE mice, SE mice did not exhibit trial-to-trial improvement in this study, as in a previous study that failed to find an effect of a potentially deleterious genetic mutation on this measure [22]. Thus, for measures such as this, EE may be necessary to enable the observation of within-session learning. Female SE mice demonstrated longer latencies to escape in trialto-trial learning, which was consistent with decreased velocities observed during the probe trial. Sex should therefore be included as a factor in the analysis of latency in particular, though similar performance on other measures may belie underlying sex differences in neural processes [27–30]. Though latency is the classical and most commonly reported outcome measure for MWM tasks, the measures of distance and spatial strategy were more informative of sex and housing effects on spatial cognition. In addition to the standardization and detailed documentation of housing conditions, reporting of multiple outcome measures may help to reconcile research findings of different studies. 4.5. Conclusions Several studies have shown evidence that EE is beneficial in counteracting the effects of aging, pharmacological challenges, and deleterious mutations (for recent reviews, see [4,31]). However, the effects of EE on emotion and cognition in mice are difficult to extract from the published literature due to lack of standardization in EE and legitimate variation in behavioral testing decisions to suit specific research interests. We believe that this study provides an easily reproducible EE protocol and detailed analysis of several behavioral results of common interest. We observed strong effects of housing on trial-to-trial spatial learning, which involves working memory and requires cognitive flexibility. As we hypothesized, EE improved the cognitive function of male and female mice, increasing the range over which the effects of experimental manipulations might be quantitated. The specific pattern of MWM results we observed is consistent with actions of EE upon the circuitry of the hippocampus and frontal cortex, as suggested by previously described neural correlates [4].
T.R. Hendershott et al. / Behavioural Brain Research 314 (2016) 215–225
Recent emphasis in the research community on the need to include female animal subjects stems from a collective desire for improved scientific understanding of basic neural processes underlying behavior. This desire is coupled with a mandate to prevent the propagation of assumptions of generalizability from preclinical research that may eventually create or exacerbate gender-based health disparities. This shift has highlighted the deficit of information in the published literature on sex differences, even in commonly applied tests. We have provided direct analyses of behavioral sex differences, with caveats: larger samples sizes are required to detect subtle differences, and similarities in behavioral measures do not preclude differences at the level of neural processing [30]. We note that exercise is an important aspect of animal health, and our decision to omit a running wheel from our housing conditions is not a rejection of the need to provide greater access to exercise to laboratory animals. On the contrary, exercise is necessary for normal metabolic health [32]. Indeed, the effects of our EE protocol, though it did not include a running wheel, might nonetheless have been mediated in part by the promotion of physical activity. Future studies will hopefully address the necessity and sufficiency of components of this protocol for the production of behavioral effects, the importance of duration and developmental stage of enrichment, and the neural processes that underlie these effects in male and female mice. Acknowledgments We thank Angelo DeNofrio, Christopher Flynn, Kevin Moriarty, Caitlin Pollard, Samantha Speroni, Catarina Teves, and Nicole Parentela for assistance with animal testing. We thank Gary Chalifoux for assistance with the maintenance of the vivarium and equipment. We are grateful to Michael Drebot for advice on quantitative data processing. We thank Grace Cavanaugh, Joseph Cristiciello, and Michael Keane for careful editing of the manuscript. This work was supported by the College of the Holy Cross. References [1] D.O. Hebb, The Organization of Behavior: A Neuropsychological Theory, Wiley & Sons Inc, New York, 1949. [2] D.O. Hebb, The effects of early experience on problem solving at maturity, Am. Psychol. 2 (1947) 306–307. [3] W.T. Greenough, J.E. Black, C.S. Wallace, Experience and brain development, Child Dev. 58 (1987) 539–559. [4] H. Hirase, Y. Shinohara, Transformation of cortical and hippocampal neural circuit by environmental enrichment, Neuroscience 280 (2014) 282–298. [5] M.R. Rosenzweig, Environmental complexity, cerebral change, and behavior, Am. Psychol. 21 (1966) 321–332. [6] L.A. Toth, The influence of the cage environment on rodent physiology and behavior: implications for reproducibility of pre-clinical rodent research, Exp. Neurol. 270 (2015) 72–77. [7] R.G. Bechara, A.M. Kelly, Exercise improves object recognition memory and induced BDNF expression and cell proliferation in cognitively enriched rats, Behav. Brain Res. 245 (2013) 96–100. [8] S. Pietropaolo, J. Feldon, E. Alleca, F. Cirulli, B.K. Yee, The role of voluntary exercise in enriched rearing: a behavioral analysis, Behav. Neurosci. 4 (2006) 787–803.
225
[9] S. Pietropaolo, J. Feldon, B.K. Yee, Environmental enrichment eliminates the anxiety phenotypes in triple transgenic mouse model of Alzheimer’s disease, Cogn. Affect. Behav. Neusosci. 3 (2014) 996–1008. [10] P. Mesa-Gresa, M. Ramos-Campos, R. Redolat, Behavioral effects of different enriched environments in mice treated with the cholinergic agonist PNU-282987, Behav. Processes 103 (2014) 117–124. [11] C. Girbovan, H. Plamondon, Environmental enrichment in female rodents: considerations in the effects on behavior and biochemical markers, Behav. Brain Res. 253 (2013) 178–190. [12] D.L. Braff, N.R. Swerdlow, M.A. Geyer, Symptom correlates of prepulse inhibition deficits in male schizophrenic patients, Am. J. Psychol. 156 (1999) 596–602. [13] G.M. McAlonan, E. Daly, V. Kumari, H.D. Critchley, T. van Amelsvoort, J. Suckling, A. Simmons, T. Sigmundsson, K. Greenwood, A. Russell, N. Schmitz, F. Happe, P. Howlin, D.G. Murphy, Brain anatomy and sensorimotor gating in Asperger’s syndrome, Brain 125 (2002) 1594–1606. [14] N.R. Swerdlow, M. Weber, Y. Que, G.A. Light, D.L. Braff, Realistic expectation of prepulse inhibition in translational models for schizophrenia research, Psychopharmacology (Berl.) 3 (2008) 331–388. ˜ M. Prunell, D. Rotlland, A. Armario, R.M. Escoriheula, Enduring effects [15] Y. Pena, of environmental enrichment from weaning to adulthood on pituitary-adrenal function, pre-pulse inhibition and learning in male and female rats, Psychoneuroendocrinology 34 (2009) 1390–1404. [16] Y. Chen, Y. Mao, D. Zhou, J. Wang, Y. Ma, Environmental enrichment and chronic restraint stress in ICR mice: effects on prepulse inhibition of startle and Y-maze spatial recognition memory, Behav. Brain Res. 212 (2010) 49–55. [17] P. Mesa-Gresa, A. Perez-Martinez, R. Redola, Environmental enrichment improves novel object recognition and enhances agonistic behavior in male mice, Aggress. Behav. 39 (2013) 269–279. [18] I. Heinla, E. Leidmaa, T. Visnapuu, M. Philips, E. Vasar, Enrichment and individual housing reinforce the differences in aggressiveness and amphetamine response in 129S6/SvSv and C57BL/6 strains, Behav. Brain Res. 267 (2014) 66–73. [19] J. Bonaccorsi, S. Cintoli, R. Mastrogiacomo, S. Baldanzi, C. Braschi, T. Pizzorusso, M.C. Cenni, N. Berardi, System consolidation of spatial memories in mice: effects of enriched environment, Neural Plast. (2013) 1–12. [20] W.E. Crusio, D. Goldowitz, A. Holmes, D. Wolfer, Standards for the publication of mouse mutant studies, Genes Brain Behav. 8 (2009) 1–4. [21] J.E. Gresack, K.M. Kerr, K.M. Frick, Life-long environmental enrichment differentially affects the mnemonic response to estrogen in young, middle-aged, and female mice, Neurobiol. Learn. Mem. 88 (4) (2007) 393–408. [22] A.C. Basu, G.E. Tsai, C.L. Ma, J.T. Ehmsen, A.K. Mustafa, L. Han, S.I. Jiang, M.A. Benneyworth, M.P. Froimowitz, N. Lange, S.H. Snyder, R. Bergeron, J.T. Coyle, Targeted disruption of serine racemase affects glutamatergic neurotransmission and behavior, Mol. Psychiatry 14 (2009) 719–727. [23] J.N. Crawley, Mouse behavioral assays relevant to the symptoms of autism, Brain Pathol. 17 (2007) 448–459. [24] J. Gil-Mohapel, P.S. Brocardo, W. Choquette, R. Gothard, J.M. Simpson, B.R. Christie, Hippocampal neurogenesis levels predict watermaze search strategies in the aging brain, PLoS One 8 (2013) 1–19, e75125. [25] U. Rudolph, F. Crestani, D. Benke, I. Brünig, J.A. Benson, J. Fritschy, J.R. Martin, H. Bluethmann, H. Möhler, Benzodiazepine actions mediated by specific ␥-aminobutyric acidA receptor subtypes, Nature 401 (1999) 796–800. [26] C. Belzung, G. Griebel, Measuring normal and pathological anxiety-like behavior in mice: a review, Behav. Brain Res. 125 (2016) 141–149. [27] J.M. Juraska, C. Henderson, J. Müller, Differential rearing experience, gender, and radial maze performance, Dev. Psychobiol. 17 (1984) 209–215. [28] J.M. Juraska, J.M. Fitch, C. Henderson, N. Rivers, Sex differences in the dendritic branching of dentate granule cells following differential experience, Brain Res. 333 (1985) 73–80. [29] K. Frick, N. Stearns, J. Pan, J. Berger-Sweeney, Effects of environmental enrichment on spatial memory and neurochemistry in middle-aged mice, Learn. Mem. 10 (2003) 187–198. [30] K. Frick, J. Tuscher, Sex steroid hormones matter for learning and memory: estrogenic regulation of hippocampal function in male and female rodents, Learn. Mem. 22 (2015) 472–493. [31] P. Sampedro-Piquero, A. Begega, Environmental enrichment as a positive behavioral intervention across the lifespan, Curr. Neuropharmacol. 14 (2016) 1–11. [32] B. Martin, S. Ji, S. Maudsley, M.P. Mattson, Control laboratory rodents are metabolically morbid: why it matters, Proc. Natl. Acad. Sci. U. S. A. 107 (2010) 6127–6133.