Assessment of fight outcome is needed to activate socially driven transcriptional changes in the zebrafish brain

Abstract

Group living animals must be able to express different behavior profiles depending on their social status. Therefore, the same genotype may translate into different behavioral phenotypes through socially driven differential gene expression. However, how social information is translated into a neurogenomic response and what are the specific cues in a social interaction that signal a change in social status are questions that have remained unanswered. Here, we show for the first time, to our knowledge, that the switch between status-specific neurogenomic states relies on the assessment of fight outcome rather than just on self- or opponent-only assessment of fighting ability. For this purpose, we manipulated the perception of fight outcome in male zebrafish and measured its impact on the brain transcriptome using a zebrafish whole genome gene chip. Males fought either a real opponent, and a winner and a loser were identified, or their own image on a mirror, in which case, despite expressing aggressive behavior, males did not experience either a victory or a defeat. Massive changes in the brain transcriptome were observed in real opponent fighters, with losers displaying both a higher number of differentially expressed genes and of coexpressed gene modules than winners. In contrast, mirror fighters expressed a neurogenomic state similar to that of noninteracting fish. The genes that responded to fight outcome included immediate early genes and genes involved in neuroplasticity and epigenetic modifications. These results indicate that, even in cognitively simple organisms such as zebrafish, neurogenomic responses underlying changes in social status rely on mutual assessment of fighting ability.

Keywords: fighting; gene expression; mutual assessment; social dominance; social genomics.

Fig. 1.

Behavioral paradigm used to promote different social experiences in zebrafish. (A) Experimental setup used to promote the four social experiences: (Left) control group (no social interaction); (Center) mirror elicited fight (animals fought their own image on the mirror), and (Right) real opponent fights (animals fought a real opponent and experienced a victory or a defeat). (B) Behavioral profiles of each social phenotype (i.e., socially isolated, mirror fighters, winners, and losers) as illustrated by the frequency of aggressive and submissive behaviors (average ± SEM; n = 3 for each condition) expressed in the last 10 min of each type of social treatment.

Fig. 2.

Socially driven changes in gene expression in the brain of zebrafish. (A) Venn diagram showing the number of DE genes between each social experience (mirror fighting and winners and losers of a real opponent fight) and the reference group (social isolation). Only in winners (red) and losers (blue) were found DE genes. Genes with known effect in the neurosystem are also given (up-regulated = arrow up; down-regulated = arrow down). (B) Hierarchical clustering of the individuals from each social treatment (columns) and of DE genes (lines). Heatmap represents normalized gene expression levels (red, low expression; green, high expression).

Fig. 3.

Coexpression gene modules for different social experiences. (A) Associations between patterns of expression in the 12 identified modules across social phenotypes (noninteracting, mirror fighters, winners, and losers) and observed behavioral traits (aggression and submission). The colors of the boxes are scaled with the value of correlation coefficients, ranging from red (r = −1) to green (r = 1). On the right side of the heat map are the numbers of genes in each module and a dendrogram showing the inferred relationships among modules. (B) Eigengene values of samples separated by group (isolated, mirror, loser, and winner) for gene modules significantly associated to social phenotypes (light green, dark gray, pale turquoise, dark green, violet, and light yellow).

Fig. S1.

TF motifs enriched in differentially expressed genes for different social experiences. (A) Single motifs found in significant meta-associations. (B) TF motif pairs involving Arnt found in significant meta-associations using flexible cis-METALYSIS analyses (only positive interactions are presented). (C) TF motif pairs involving FOXP2 found in significant meta-associations using flexible cis-METALYSIS analyses (only positive interactions are presented). The social experiences considered are winning (winners) and losing (losers) a real opponent fight. For single motifs, orange/blue indicates TF motifs associated with up-regulated/down-regulated genes; for motif pairs, orange/blue indicates that the strongest association in each social experience was with the up-regulated/down-regulated genes. Gray cells represent associations that were not significant. P values present the meta-associations calculated using cis-METALYSIS. False discovery rate (FDR) presents P values corrected for multiple comparisons following the Benjamini–Hochberg procedure. All TF motifs are from the Jaspar Core Vertebrates database.

Fig. S2.

Comparison between the expression levels of differentially expressed genes with high fold changes in the microarrays with their expression in confirmatory quantitative RT-PCR (qPCR). Black bars, qPCR; checkered bars, microarray technique (mean± SEM).