SUMMARY Seasonally breeding animals exhibit profound physiological and behavioural responses to changes in ambient day length (photoperiod), including changes in reproductive function and territorial aggression. Species where aggression persists when gonads are regressed and circulating levels of gonadal hormones are low, such as Siberian hamsters (Phodopus sungorus) and song sparrows (Melospiza melodia), challenge the well-established framework that gonadal hormones are important mediators of aggression. A solution to this apparent paradox is that a season-specific increase in sensitivity to hormones in brain areas associated with aggression offsets low levels of gonadal hormones during periods of reproductive quiescence. To test this hypothesis, we manipulated photoperiod to induce natural fluctuations in seasonal phenotype across multiple stages of the annual reproductive cycle in female Siberian hamsters that display increased aggression during short-day reproductive quiescence, suggesting that behaviour persists independent of gonadal steroids. Females were housed in long “summer” days or short “winter” days for 10, 24 or 30 weeks to capture gonadal regression, transition back to a reproductively functional state and full gonadal recrudescence, respectively. Long-day animals maintained reproductive functionality and displayed low aggression across all time points. By week 10, short-day reproductively responsive females underwent gonadal regression and displayed increased aggression; non-responsive animals showed no such changes. At week 24, animals were in a transitional period and displayed an intermediate phenotype with respect to reproduction and aggression. By week 30, short-day females were fully recrudesced and returned to long-day-like levels of aggression. Consistent with our hypothesis, gonadally regressed females displayed decreases in 17β-oestradiol (oestradiol) levels, but site-specific increases in the abundance of brain oestrogen receptor-alpha (ERα) in regions associated with aggression, but not reproduction. Increased site-specific ERα may function as a compensatory mechanism to allow increased responsiveness to oestradiol in regulating aggression in lieu of high circulating concentrations of hormones. Collectively, these results broaden our understanding of how breeding phenology maps onto social behaviour and the mechanisms that have evolved to coordinate behaviours that occur in non-breeding contexts.

Many animal species exhibit year‐round aggression, a behaviour that allows individuals to compete for limited resources in their environment (eg, food and mates). Interestingly, this high degree of territoriality persists during the non‐breeding season, despite low levels of circulating gonadal steroids (ie, testosterone [T] and oestradiol [E2]). Our previous work suggests that the pineal hormone melatonin mediates a ‘seasonal switch’ from gonadal to adrenal regulation of aggression in Siberian hamsters (Phodopus sungorus); solitary, seasonally breeding mammals that display increased aggression during the short, ‘winter‐like’ days (SDs) of the non‐breeding season. To test the hypothesis that melatonin elevates non‐breeding aggression by increasing circulating and neural steroid metabolism, we housed female hamsters in long days (LDs) or SDs, administered them timed or mis‐timed melatonin injections (mimic or do not mimic a SD‐like signal, respectively), and measured aggression, circulating hormone profiles and aromatase (ARO) immunoreactivity in brain regions associated with aggressive or reproductive behaviours (paraventricular hypothalamic nucleus [PVN], periaqueductal gray [PAG] and ventral tegmental area [VTA]). Females that were responsive to SD photoperiods (SD‐R) and LD females given timed melatonin injections (Mel‐T) exhibited gonadal regression and reduced circulating E2, but increased aggression and circulating dehydroepiandrosterone (DHEA). Furthermore, aggressive challenges differentially altered circulating hormone profiles across seasonal phenotypes; reproductively inactive females (ie, SD‐R and Mel‐T females) reduced circulating DHEA and T, but increased E2 after an aggressive interaction, whereas reproductively active females (ie, LD females, SD non‐responder females and LD females given mis‐timed melatonin injections) solely increased circulating E2. Although no differences in neural ARO abundance were observed, LD and SD‐R females showed distinct associations between ARO cell density and aggressive behaviour in the PVN, PAG and VTA. Taken together, these results suggest that melatonin increases non‐breeding aggression by elevating circulating steroid metabolism after an aggressive encounter and by regulating behaviourally relevant neural circuits in a region‐specific manner.

No abstract