Empathy is moderated by genetic background in mice

Abstract

Empathy, as originally defined, refers to an emotional experience that is shared among individuals. When discomfort or alarm is detected in another, a variety of behavioral responses can follow, including greater levels of nurturing, consolation or increased vigilance towards a threat. Moreover, changes in systemic physiology often accompany the recognition of distressed states in others. Employing a mouse model of cue-conditioned fear, we asked whether exposure to conspecific distress influences how a mouse subsequently responds to environmental cues that predict this distress. We found that mice are responsive to environmental cues that predict social distress, that their heart rate changes when distress vocalizations are emitted from conspecifics, and that genetic background substantially influences the magnitude of these responses. Specifically, during a series of pre-exposure sessions, repeated experiences of object mice that were exposed to a tone-shock (CS-UCS) contingency resulted in heart rate deceleration in subjects from the gregarious C57BL/6J (B6) strain, but not in subjects from the less social BALB/cJ (BALB) strain. Following the pre-exposure sessions, subjects were individually presented with the CS-only for 5 consecutive trials followed by 5 consecutive pairings of the CS with the UCS. Pre-exposure to object distress increased the freezing responses of B6 mice, but not BALB mice, on both the CS-only and the CS-UCS trials. These physiological and behavioral responses of B6 mice to social distress parallel features of human empathy. Our paradigm thus has construct and face validity with contemporary views of empathy, and provides unequivocal evidence for a genetic contribution to the expression of empathic behavior.

Figure 1. The fear conditioning apparatus and conditioning schedules.

(a) Subject mice from the BALB (white mouse) and B6 (dark grey mouse) strains were separated from 2 novel, F1 object mice (brown mice) that received a 2-s electrical shock (UCS _objects) and/or a 30-s tone (CS) under different conditioning schedules (pre-exposure sessions). Both subjects and objects were exposed to the CS, but only objects directly experienced the UCS. Subjects had access to object distress cues (UCS _subjects) that were emitted as a consequence of receiving the UCS. (b) Photograph of the fear conditioning apparatus from an overhead perspective. (c–f) Each conditioning schedule is representative of one 120-s trial.

Figure 2. Freezing responses of test mice during direct presentation of the CS-only or the paired CS-UCS.

(a) When mice were not pre-exposed to objects or conditioned themselves, there was no strain difference in freezing to the CS (N = 12 mice/genotype; age = 6–7 wks). (b) BALB mice expressed longer freezing responses than B6 mice on trials 5-10 when they had not been pre-exposed to objects, but were directly exposed to the CS-UCS contingency (N = 16 mice/genotype; age = 7–8 wks). Asterisks represent significant differences between BALB and B6 mice as assessed by a Bonferroni step-down procedure on a trial-by-trial basis (∝ = 0.005 for each trial). All data were scored in duplicate by 2 independent raters (see the figure legend corresponding to Figure 3 for the inter-rate reliability) and are presented as the mean±s.e.m.

Figure 3. Freezing responses of subject mice to the CS-only and the paired CS-UCS following pre-exposure to object mice under different conditioning schedules.

The top of each panel contains a graphical representation of a single conditioning trial of the type that objects received during the pre-exposure sessions. The data in each panel depict the freezing behavior of subjects after pre-exposure sessions in response to the CS before (trials 1-6) and after (trials 7-10) it was forward paired with the UCS. (a) When previously exposed to objects receiving the paired CS-UCS, B6 subjects expressed more freezing than BALB subjects on 9 (of 10) test trials (N = 26 subjects/genotype; age = 5–7 wks). (b) B6 subjects expressed more freezing than BALB subjects on 5 (of 10) test trials when they were previously exposed to objects receiving the unpaired CS-UCS (N = 18 subjects/genotype; age = 5–7 wks). (c) B6 subjects expressed more freezing than BALB subjects on 3 (of 10) test trials when they were previously exposed to objects receiving the UCS-only (N = 14 subjects/genotype; age = 5–6 wks). (d) Freezing responses of BALB and B6 subjects were similar on all test trials when they were previously exposed to objects receiving the CS-only (N = 16 subjects/genotype; age = 5–7 wks). Note that in all cases the UCS was presented during the final 2 s of trial 6, which was not included in the measurements of freezing. Asterisks represent significant differences between BALB and B6 mice as assessed by a Bonferroni step-down procedure on a trial-by-trial basis (∝ = 0.008 and 0.013 for trials 1-6 and trials 7-10, respectively). All data were scored in duplicate by 2 independent raters and are presented as the mean±s.e.m. Inter-rater reliability, Pearson's correlation, r_p = 0.94, d.f. = 1,984.

Figure 4. Head orientations and freezing responses of subjects during exposure to object distress.

(a) Estimates of subject head orientations were extracted from freeze-frames of video recordings at 2 time-points with respect to UCS presentation to objects (30 s pre-UCS and 1 s post-UCS). The longitudinal axis of the subject's head, running parallel with the sagittal midline, was referenced at 15° increments. A 0° head orientation was defined by the longitudinal axis of the subject's head forming a right angle with the steel dowels that separated the observation and demonstration compartments. Prior to presentation of the UCS to objects, BALB (left panels; N = 16 mice; age = 7–8 wks) and B6 (right panels; N = 8 mice, age = 7–8 wks) subjects were orientated towards the demonstration compartment (mean head angle = dotted black lines, std. dev. = grey shaded area). Following presentation of the UCS to objects, subject head orientations were more consistently directed at objects (depicted by red dotted lines and pink shading), as indicated by a ≈2-fold reduction in variability of head orientations (Brown-Forsythe ANOVA, P<0.001 for both BALB and B6 subjects). (b) The freezing responses of these subjects were minimal and strain-independent during the 30-s period following presentation of the UCS to objects. Data in panel (b) are presented as the mean±s.e.m.

Figure 5. Distress vocalization of an object during UCS presentation.

(a) Representative sonogram of an object distress vocalization recorded in real-time without the demonstration compartment enclosed by Plexiglas®. (b) Sonogram of the same vocalization played back through a speaker without the demonstration compartment enclosed. (c) Sonogram of the same vocalization played back through a speaker with the demonstration compartment enclosed. Color-coding indicates the relative decibel level for the call, with blues representing low-intensity energy and red/yellow representing high-intensity energy.

Figure 6. Freezing responses of subjects after pre-exposure to the CS paired with playback of object distress vocalizations.

(a) Subjects were exposed to objects during conditioning, but the CS-UCS was not presented to objects (N = 10 subjects/genotype; age = 5–7 wks). (b) Subjects were exposed to objects receiving the paired CS-UCS during conditioning (N = 8 subjects/genotype; age = 5–6 wks). (c) Subjects were not exposed to objects, but received the CS forward paired with the playback of object distress vocalizations during conditioning (N = 16 subjects/genotype; age = 5–7 wks). Asterisks represent a significant (P<0.05) difference between subjects from the BALB and B6 strains. All data were scored in duplicate by 2 independent raters and are presented as the mean±s.e.m. Inter-rater reliability, r_p = 0.96, d.f. = 659.

Figure 7. Heart rate changes of mice after a single exposure to the UCS.

During habituation to the conditioning apparatus, HR was measured for 5 min following presentation of a single UCS. Mice from the experiments illustrated in Figures 8 and 9 were treated in the same way during the habituation period, and were therefore pooled for graphical presentation and analysis (N = 24 mice/genotype; age = 7–13 wks). Each data point represents a 10-s bin. HR is presented as a change (Δ) from the baseline HR of each mouse (see Table 2). All data presented as the mean±s.e.m.

Figure 8. Heart rate changes of mice during direct conditioning with the CS-UCS and testing with the CS-only.

HR was measured during (a) the first and second conditioning session in mice that directly received the paired CS-UCS, and during (b) testing, which entailed 5 repeated presentations of the CS-only. HR changes in mice from both strains were similar across the 2 conditioning sessions and therefore pooled for graphical presentation. Each data point represents a 10-s bin. HR is presented as a change (Δ) from the baseline HR of each mouse (see Table 2). (c) Conditioned freezing responses of mice during presentation of the CS-only (N = 9 subjects/genotype; age = 12–13 wks). Asterisks represent significant (P<0.05) HR differences between BALB and B6 mice. Numeric symbols represent significant (P<0.05) differences in freezing between BALB and B6 mice. Behavioral data were scored in duplicate by 2 independent raters and all data are presented as the mean±s.e.m. Inter-rater reliability, r_p = 0.90, d.f. = 74.

Figure 9. Heart rate changes of subjects during pre-exposure to objects that were conditioned with the CS-UCS and during testing with the CS-only.

HR was measured during (a) the first and second conditioning session in subjects that experienced objects receiving the paired CS-UCS, and during (b) testing, which entailed 5 repeated presentations of the CS-only. HR changes in subjects from both strains were similar across the 2 conditioning sessions and therefore pooled for graphical presentation. Each data point represents a 10-s bin. HR is presented as a change (Δ) from the baseline HR of each mouse (see Table 2). (c) Conditioned freezing responses of subjects during presentation of the CS-only (N = 15 subjects/genotype; age = 7–9 wks). Asterisks represent significant (P<0.05) HR differences between BALB and B6 subjects. Numeric symbols represent significant (P<0.05) differences in freezing between BALB and B6 subjects. Behavioral data were scored in duplicate by 2 independent raters and all data are presented as the mean±s.e.m.. Inter-rater reliability, r_p = 0.94, d.f. = 143.

Figure 10. Strain-dependent development of sociability in adolescent mice.

(a) Following 24 h of social isolation, early-adolescent (age = ≈4 wks) B6 test mice expressed more social investigation (SI) towards a novel, opposite-sex F1 mouse compared to age-matched BALB mice. There was no difference in SI among late-adolescent mice (age = 7–8 wks) from the BALB and B6 strains (N = 14 mice/genotype/age). The numeric symbol represents a significant (P = 0.01), planned orthogonal contrast between BALB and B6 mice. (b) Following 8 days of conditioning during early adolescence or late adolescence, B6 mice expressed a social conditioned place preference (SCPP), whereas BALB mice expressed social indifference at both time points (N = 14 mice/genotype/age). SCPP scores were calculated by subtracting the time a mouse spent in an isolation-paired environment from the duration it spent in a socially paired environment (see Materials & Methods). The asterisks represent significant (P<0.05), planned orthogonal contrasts between BALB and B6 mice. All data are presented as the mean±s.e.m.

Figure 11. Freezing responses of socially and isolate-housed subjects pre-exposed to objects receiving the CS-UCS contingency.

Subjects were either isolated or remained in a social housing context outside of their observations of objects receiving the paired CS-UCS during the pre-exposure session. Since freezing responses of socially housed subjects in this experiment (N = 10 subjects/genotype; age = 6–8 wks) did not differ (P = 0.20) from responses of mice that were evaluated in the experiment presented in Figure 3a, the data were pooled for graphical presentation and statistical analysis. Social isolation did not alter the freezing responses of BALB subjects, but it depressed fear acquisition to presentation of the paired CS-UCS in B6 subjects (N = 20 subjects/genotype; age = 7–8 wks). Socially housed B6 subjects expressed longer freezing responses than all other groups on trials 7-10 (P = 0.006). Asterisks represent significant differences between socially housed B6 subjects and all other groups as assessed by a Bonferroni step-down procedure on a trial-by-trial basis (∝ = 0.013 for trials 7-10). All data were scored in duplicate by 2 independent raters and are presented as the mean±s.e.m. Inter-rater reliability (not including pooled data from Fig. 3a), r_p = 0.94, d.f. = 547.