Data are presented as group means + SEM of the ALT and the time spent in each behavioral category (indicated as percentage of the total 10 min test). Treatment effects on these behavioral parameters were statistically tested by repeated measures analyses of variance (ANOVA), using SPSS for Windows (version 18: SPSS Inc, Chicago, IL, USA). The ANOVA design consisted of one within-subjects variable with four levels (compounds or doses).
To account for possible violations of the sphericity assumption for factors with more than two levels, Huynh-Feldt adjusted p-values and the epsilon correction factor are reported, together with the unadjusted degrees of freedom and F-values. For all analyses the partial eta-squared effect sizes are reported with
η
2 < 0.06 reflecting a small effect;η
2 ≥ 0.06 a medium effect; andη
2 ≥ 0.14 a large effect. If overall significance was obtained, post-hoc pairwise comparisons were carried out in order to reveal specific differences in treatment conditions. If these were found, the influence of individual variation in the baseline level of offensive aggression was investigated by entering this baseline level as a covariate into the design.If a significant effect of the covariate on the overall treatment effect was found, Pearson’s correlations were computed to find out whether effects were greater with lower or higher baseline level of aggression. To this end, effect measures were computed as the difference between the behavior observed during the treatment with vehicle and the one observed under different doses of OXT or OXTR antagonist. Furthermore, using a median split approach, the OXT and OXTR antagonist dose-effect curves on aggression were plotted for the low-medium aggressive (i.e., group average of the percentage of time spent in offensive behavior was 35 ± 4% in exp.1 and 18 ± 3% in exp. 3) and for the medium-high aggressive (i.e., group average of the percentage of time spent in offensive behavior was 65 ± 3% in exp.1 and 48 ± 4% in exp. 3) groups of subjects.
To verify the behavioral specificity of OXT treatment, ethograms of OXT-treated animals were compared to those of untreated animals, matched for the level of aggression. These comparisons were carried out by means of t-tests for independent samples.
All comparisons with a p-value ≤ 0.05 were considered to be statistically significant.
P-values between 0.06 and 0.1 were noted as a trend towards significance.
2
RESULTS
Experiment 1: Oxytocin dose-response curve
All three doses of the OXT reduced offensive behavior of the experimental male resident towards the unfamiliar intruder during the social encounter (F3,27 = 9.487, p = 0.001,
η
2 = 0.513,ε
= 0.840). This anti-aggressive effect occurred in a dose-dependent manner (post-hoc pairwise comparisons: vehicle vs. OXT 0.25 μg p = 0.079, vehicle vs. OXT 1 μg p = 0.013, vehicle vs. OXT 4 μg p = 0.003), with both the intermediate and the highest dose being significantly more effective than the lowest dose (OXT 0.25 μg vs. OXT 1 μg p = 0.028, OXT 0.25 μg vs. OXT 4 μg p = 0.006). Concomitantly, social explorative behavior was significantly enhanced (F3,27 = 7.676, p = 0.001,η
2 = 0.460,ε
= 0.967) by OXT 0.25 μg and OXT 1 μg compared to vehicle and OXT 4 μg (vehicle vs. OXT 0.25 μg p= 0.003, vehicle vs. OXT 1 μg p = 0.031, OXT 0.25 μg vs. OXT 4 μg p = 0.002, OXT 1 μg vs. OXT 4 μg p = 0.003) (Figure 1).
Figure 1. Behavioral changes induced by acute pharmacological manipulation of the central oxytocinergic system. Male resident wild-type Groningen rats were exposed to an unfamiliar male intruder Wistar rat after acute icv administration of vehicle (saline 5 μl; empty bars) or oxytocin at the doses of 0.25 μg (light gray bars) – 1.0 μg (gray bars) – 4.0 μg (dark gray bars)/5 μl. Insert graph depicts the treatment effects on the attack latency time. Data are presented as mean + SEM (N = 10). *p < 0.05 indicates a significant difference in comparison with vehicle.
Dose-dependent changes were also found in the modulation of self-grooming (F3,27 = 12.743, p = 0.001,
η
2 = 0.586,ε
= 0.540). All doses resulted in an increased duration of this behavior compared to vehicle (vehicle vs. OXT 0.25 μg p = 0.053, vehicle vs. OXT 1 μg p = 0.021, vehicle vs. OXT 4 μg p < 0.001). Self-grooming after OXT 1 μg (p = 0.026) and OXT 4 μg (p = 0.001) was significantly more intense than after OXT 0.25 μg. Finally,immobility was found to be enhanced by OXT 4 μg compared to all the other treatments (F3,27 = 10.187, p = 0.001,
η
2 = 0.531,ε
= 0.699) (vehicle vs. OXT 4 μg p = 0.011, OXT 0.25 μg vs. OXT 4 μg p = 0.006, OXT 1 μg vs. OXT 4 μg p = 0.004). The overall treatment effect on ALT only showed a trend towards significance (F3,27 = 3.880, p = 0.071,η
2 = 0.301,ε
= 0.394). OXT 4 μg tended to prolong this latency compared to vehicle (p = 0.070) (Figure 1, insert). No treatment effects were found on the category of non-social behaviors.Due to the wide within-group individual variation in intermale aggression, we tested the influence of baseline level of aggression on the significant overall treatment effects.
Analysis of covariance revealed that the anti-aggressive effect induced by OXT 0.25 μg and OXT 1 μg indeed depended upon the baseline level of aggression, as the interaction of treatment with this baseline level was significant (F3,24 = 8.379, p = 0.001,
η
2 = 0.512,ε
= 1.00) and adjustment reduced the previously found effects (F3,24 = 2.756, p = 0.064,η
2 = 0.256,ε
= 1.00) (OXT 0.25μg p = 0.023, OXT 1 μg p = 0.054, OXT 4 μg p = 0.173).Larger treatment effects were observed in the more aggressive animals as revealed by positive correlations between the baseline level of aggression and the effect measures (OXT 0.25 μg r = 0.827, p = 0.003; OXT 1 μg r = 0.835, p = 0.003) (Figure 4A). Baseline aggression level did not significantly interact with the treatment effect of OXT on social explorative behavior, self-grooming, immobility or ALT.
Experiment 2: Co-administration study
In line with the outcome of Exp.1, we replicated the anti-aggressive (F3,48 = 4.400, p = 0.008,
η
2 = 0.216,ε
= 0.914) effect of OXT 1 μg which appeared significantly effective compared to vehicle (p = 0.002), OXTR antagonist (p = 0.010), and OXTR antagonist + OXT (p = 0.037). This dose also markedly intensified social explorative behavior (F3,48 = 5.560, p = 0.002,η
2 = 0.258,ε
= 1.00) compared to vehicle (p = 0.002), OXTR antagonist (p = 0.005), and OXTR antagonist + OXT (p = 0.025). When injected alone the OXTR antagonist treatment failed to show behavioral changes compared to vehicle. However, when applied prior to OXT treatment it completely blocked the synthetic OXT-induced effects. No overall treatment effect was found on self-grooming, or on locomotor activity, or on non-social behaviors, or ALT (Figure 2).Experiment 3: Oxytocin receptor antagonist dose-response curve
Statistical analysis did not reveal any overall significant treatment effects in any of the five behavioral categories. However, a trend effect was found for the ALT (F3,33 = 2.337, p = 0.099,
η
2 = 0.175,ε
= 0.899), where OXTR antagonist 15 μg delayed the first attack compared to OXTR antagonist 3.75 μg (p = 0.036), and OXTR antagonist 7.5 μg (p = 0.023), but not compared to vehicle. Yet, when the influence of the baseline level of aggression on the overall treatment effect was tested (as a covariate in the ANOVA), a significant effect in the category of offensive behavior was revealed (F3,30 = 3.340, p = 0.032,η
2 = 0.250,ε
= 1.00) (vehicle vs. OXTR antagonist 7.5 μg p = 0.060) (Figure 3). The negative correlation between the baseline level of aggression and the effect measures, although not significant,2
Figure 2. Behavioral changes induced by acute pharmacological manipulation of the central oxytocinergic system. Male resident wild-type Groningen rats were exposed to an unfamiliar male intruder Wistar rat after acute icv administration of vehicle (Veh; empty bars), oxytocin (OXT) (Veh 2.5 μl + OXT 1.0 μg/2.5 μl; gray bars), peptidergic oxytocin receptor (OXTR) antagonist (Veh 2.5 μl + OXTR antagonist 0.75 μg/2.5 μl; dark gray bars) or co-administration of both (OXTR antagonist 0.75 μg/2.5 μl + OXT 1.0 μg/2.5 μl; striped bars). Insert graph depicts the treatment effects on the attack latency time. Data are presented as mean + SEM (N = 17). *p < 0.05 indicates a significant difference in comparison with vehicle.
Figure 3. Behavioral changes induced by acute pharmacological manipulation of the central oxytocinergic system. Male resident wild-type Groningen rats were exposed to an unfamiliar male intruder Wistar rat after acute icv administration of vehicle (saline 5 μl; empty bars) or a non-peptidergic oxytocin receptor (OXTR) antagonist L368.899 at the doses of 3.75 μg (light gray bars) - 7.5 μg (gray bars) – 15 μg (dark gray bars)/5 μl. Insert graph depicts the treatment effects on the attack latency time.
Data are presented as mean + SEM (N = 12).
Figure 4. Potency of oxytocin (A) and the oxytocin receptor antagonistnon-peptidergic (B) to change offensive behavior in low-medium (open squares) and medium-high (filled squares) aggressive groups of subjects. Gray area indicates SEM range of the vehicle-treated animals. Data are presented as mean ± SEM.
suggests a greater increase of offensive aggression in animals with a lower level of baseline aggression (OXTR antagonist 7.5 μg r = -0.440, p = 0.153) (Figure 4B).