Winning territorial disputes selectively enhances androgen sensitivity in neural pathways related to motivation and social aggression

Abstract

Winning aggressive disputes can enhance future fighting ability and the desire to seek out additional contests. In some instances, these effects are long lasting and vary in response to the physical location of a fight. Thus, in principle, winning aggressive encounters may cause long-term and context-dependent changes to brain areas that control the output of antagonistic behavior or the motivation to fight (or both). We examined this issue in the territorial California mouse (Peromyscus californicus) because males of this species are more likely to win fights after accruing victories in their home territory but not after accruing victories in unfamiliar locations. Using immunocytochemistry and real-time quantitative PCR, we found that winning fights either at home or away increases the expression of androgen receptors (AR) in the medial anterior bed nucleus of the stria terminalis, a key brain area that controls social aggression. We also found that AR expression in brain regions that mediate motivation and reward, nucleus accumbens (NAcc) and ventral tegmental area (VTA), increases only in response to fights in the home territory. These effects of winning were likely exclusive to the neural androgenic system because they have no detectible impact on the expression of progestin receptors. Finally, we demonstrated that the observed changes in androgen sensitivity in the NAcc and VTA are positively associated with the ability to win aggressive contests. Thus, winning fights can change brain phenotype in a manner that likely promotes future victory and possibly primes neural circuits that motivate individuals to fight.

Fig. 1.

(A) Coronal brain sections from which AR-ir and PR-ir were quantified (Left to Right, Top to Bottom: NAcc, LS, BNSTma, AH, VMHvl, mAMY, PMv, VTA, and dPAG). Stereotaxic coordinate of each section (bregma) is provided below each illustration. Boxes depict areas of -ir quantification. (3V) third ventricle, (ac) anterior commissure, (AH) anterior hypothalamus, (Aq) aqueduct, (BNSTma) medial anterior bed nucleus of the stria terminalis, (cp) caudate putamen, (cc) corpus callosum, (D3V) dorsal third ventricle, (fx) fornix, (hpp) hippocampus, (ic) internal capsule, (IPR) rostral interpeduncular nucleus, (LS) lateral septum, (LV) lateral ventricle, (mAMY) medial amygdala, (mfb) medial forebrain bundle, (ml) medial lemniscus, (mm) medial mammillary nucleus, (NAcc) nucleus accumbens, (dPAG) dorsal periaqueductal gray, (PAG) periaqueductal gray, (PMv) ventral premammillary nucleus, (Pn) pontine nuclei, (SNR) reticular part of substantia nigra, (VMHvl) ventrolateral subnucleus of the ventromedial hypothalamus, (VTA) ventral tegmental area. (B) Photomicrographs of AR-ir and PR-ir cells in the NAcc, LS, and BNSTma. Left-most images are animals that received winning experience and a test encounter in the home cage (HH). Right-most images are handled control animals (no fights, CC). (Scale bar, 150 μm.)

Fig. 2.

Amount of (A) AR-ir and (B) PR-ir in limbic and mesolimbic brain areas [HH = three wins and a test encounter in the home cage; UH = three wins in unfamiliar cages and a test encounter in the home cage; UU = three wins and a test encounter in unfamiliar cages; CC = four handling experiences (no-fight controls)]. Certain regions in a single brain were sometimes damaged, causing sample sizes (n) to range from 5–8/group. Data are presented as mean ± SEM *P < 0.05, Tukey HSD post hoc test.

Fig. 3.

Relationships between scores on the winner index and the amount of AR-ir in the (A) NAcc, (B) BNSTma, and (C) VTA. HH, three wins and a test encounter in the home cage; UH, three wins in unfamiliar cages and a test encounter in the home cage; UU, three wins and a test encounter in unfamiliar cages. Best-fit lines depict significant correlations, P < 0.05, Pearson's correlation coefficient.

Fig. 4.

Relative amounts of AR mRNA in the NAcc, LS, and BNSTma. HH/black bars = three wins and a test encounter in the home cage; CC/white bars = four handling experiences (no fight controls). Sample sizes (n) ranged from 5–7/group because one individual was selected at random and used to standardize the other values. Data presented as mean± SEM. *P < 0.05, t test.