Neural and behavioural correlates of repeated social defeat

Abstract

Dominance hierarchies are common across the animal kingdom and have important consequences for reproduction and survival. Animals of lower social status cope with repeated social defeat using proactive and reactive behaviours. However, there remains a paucity of information on how an individual's coping behaviours changes over time or what neural mechanisms are involved. We used a resident-intruder paradigm in the African cichlid fish Astatotilapia burtoni to investigate the neural correlates of these two opposing behaviour groups. Fish initially used both proactive and reactive behaviours, but had a dramatic increase in use of proactive behaviours during the third interaction, and this was followed by cessation of proactive behaviours and exclusive use of reactive coping. By quantifying neural activation in socially-relevant brain regions, we identify a subset of brain nuclei, including those homologous to the mammalian amygdala, showing higher activation in fish displaying proactive but not reactive behaviours. Fish displaying reactive behaviours had greater neural activation in the superior raphe, suggesting a possible conserved function during social defeat across vertebrates. These data provide the first evidence on the involvement of specific brain regions underlying proactive and reactive coping in fishes, indicating that these nuclei have conserved functions during social defeat across taxa.

Figure 1

A resident-intruder paradigm was used to induce repeated social defeat in Astatotilapia burtoni males. (a) A 37.85 L tank was divided in two compartments by an opaque blue barrier, and views of neighbouring tanks were obstructed. One fish was allowed to acclimate to each territory for 2 days. (b) On the morning of the trial, the blue barrier was removed between the two compartments revealing a clear barrier with a small escape hole. The intruder was quickly netted from his home territory and placed into the resident’s territory and allowed to interact until he either escaped through the hole or 1 hour had passed. (c) On subsequent trials (days 2–5), the intruder was quickly netted from his territory and placed into the resident’s compartment, where he quickly faded his stereotypical “dominant” coloration. These trials continued until either the intruder escaped or 30 minutes had passed. At the conclusion of all trials the intruder was placed back into his home territory and the blue barrier was replaced (as shown in a) until the subsequent trial.

Figure 2

Representative coronal sections in different brain regions showing pS6 antibody specificity. Positive pS6 staining produces clear distinctly stained cells throughout many brain regions (top row), but staining is completely eliminated when the antibody is pre-absorbed with a pS6 blocking peptide on adjacent sections (bottom row). See text for abbreviations. Scale bars represent 100 µm in (a,b and d), and 50 µm in (c).

Figure 3

Astatotilapia burtoni males switch between reactive and proactive coping behaviours following repeated social defeat from the same aggressor. (a) A principal component analysis was used to classify behaviours as either proactive or reactive. Component 1 primarily distinguished resident behaviours (squares) while component 2 was related to intruder behaviours (circles). Behaviours with a positive loading value on component 2 were classified as proactive while negatively loaded values were classified as reactive. (b) Use of reactive and proactive coping behaviours changes with the number of defeat interactions. Proactive behaviours (red) peak on day 3 of repeated social defeat, but fish quickly return to using reactive behaviours (blue). Numbers in parentheses represent sample size. Different upper- and lower-case letters indicate differences in reactive and proactive behaviours, respectively, across days. Asterisks indicate difference between proactive and reactive behaviours each day. (c) The majority of animals had their most proactive trial occur on day 3, but ~30% were most proactive on days 1, 2, or 4 of repeated social defeat. (d,e) A PCA of intruder behaviours on day 3 (d) and day 4 (e) of repeated social defeat represents the relative use of each behaviour, where more frequently used behaviours load positively into PC1. (f) The difference between PC1 values of each individual behaviour from day 3 to day 4 indicates the degree of change between the days.

Figure 4

Node by node approach to examine neural activation levels in fish displaying different coping behaviours. (a) pS6-stained cell density (i.e. neural activation) in the Vv-r, Vd, Vp, ATn, TPp, and Dm-3 was highest in fish collected after 3 days of repeated social defeat (red). (b) In the Vs, fish collected after day 2 (yellow) and day 3 had higher neural activation than control (grey) and day 4 (blue) fish. (c) Control fish and those collected on day-2 had higher activation in the Vc compared to fish collected after 3 and 4 bouts of defeat. (d) Neural activation in the SR was higher in fish collected after 4 days of defeat when compared with those collected after only 2 or 3 days of defeat. (e) Representative photomicrographs of pS6 staining in the ATn, Vs, Vc, and SR from fish collected after 2 (left), 3 (middle), and 4 (right) days of repeated social defeat. Tukey’s box plots were used to plot the data: median is represented by a line and mean by an open circle within the box, the box extends to the furthest data points within the 25^th and 75^th percentile, and whiskers extend to the furthest data points not considered outliers. Different letters represent statistical significance at P < 0.05. N = 4 for all groups. See text for abbreviations. Scale bars in E represent 100 μm for the ATn and Vs panels and 50 μm for the Vc and SR panels.

Figure 5

Escaping and non-escaping fish have different patterns of neural activation independent of collection day. (a,b) Escaping fish have higher activation in Dl-v2 and Dc-5 than non-escaping fish. (c) Non-escaping fish have higher activation in the SR compared to escaping fish. See Fig. 4 for box plot descriptions. Different letters represent statistical significance at P < 0.05. See text for abbreviations. Scale bars in a and b represent 100 μm and 50 μm in c.

Figure 6

Fish displaying proactive and reactive coping behaviours have distinct patterns of neural activation. (a) Hierarchical clustering was used to group brain regions, and Pearson correlation coefficients were used to create a heatmap of neural co-activation across all regions examined, along with searching and freezing behaviours. R-values are represented by colour with red being positive correlations and blue representing negative correlations. Significant correlations (P < 0.05) are signified with an*. Boxes designating “proactive”, “reactive”, and “escape” are based on neural activation data in Figs 4 and 5. (b) Discriminant function analysis of pS6-staining in all brain regions clearly separates fish displaying proactive and reactive coping behaviours from neutral and control fish, indicating distinct neural activation patterns associated with these opposing coping behaviours. Each animal is represented by a circle and group means by a star. (c) Schematic summarizing neural activation results in all regions. Nuclei locations and sizes are approximate. Red represents regions associated with proactive, while blue signifies reactive coping behaviours. Purple represents regions that were different in escaping and non-escaping fish, likely relating to spatial learning. See text for abbreviations.