Stress coping style predicts aggression and social dominance in rainbow trout

Abstract

Social stress is frequently used as a model for studying the neuroendocrine mechanisms underlying stress-induced behavioral inhibition, depression, and fear conditioning. It has previously been shown that social subordination may result in increased glucocorticoid release and changes in brain signaling systems. However, it is still an open question which neuroendocrine and behavioral differences are causes, and which are consequences of social status. Using juvenile rainbow trout of similar size and with no apparent differences in social history, we demonstrate that the ability to win fights for social dominance can be predicted from the duration of a behavioral response to stress, in this case appetite inhibition after transfer to a new environment. Moreover, stress responsiveness in terms of confinement-induced changes in plasma cortisol was negatively correlated to aggressive behavior. Fish that exhibited lower cortisol responses to a standardized confinement test were markedly more aggressive when being placed in a dominant social position later in the study. These findings support the view that distinct behavioral–physiological stress coping styles are present in teleost fish, and these coping characteristics influence both social rank and levels of aggression.

Introduction

In many animal species, social defeat is a powerful stressor that can lead to drastic alterations in physiology and behavior. Behavioral effects of social defeat include appetite inhibition Kramer et al., 1999, Meerlo et al., 1997, Øverli et al., 1998, reduced aggression Blanchard et al., 1995, Höglund et al., 2001, decreased reproductive behavior D'Amato, 1988, Perret, 1992, and increased submissive and defensive behaviors towards conspecifics Blanchard et al., 1993, Siegfried et al., 1984. Social defeat is also frequently used as a paradigm to study the neuroendocrine mechanisms underlying stress-induced behavioral inhibition, depression, and acquisition and expression of fear conditioning (Berton et al., 1999, Fuchs and Flügge, 2002, Gould et al., 1997, Jasnow and Huhman, 2001, Keeney and Hogg, 1999, Koolhaas et al., 1990, Shively et al., 1997; but see also Marrow et al., 1999). It is, however, sometimes difficult to discern whether the physiological and behavioral differences between dominant and subordinate individuals are causes or consequences of social rank. Furthermore, the time course with which these changes occur is often of fundamental interest Summers, 2002, Summers et al., 2003, but with most experimental techniques data acquisition cannot be carried out until contests for dominance are already settled and animals with established dominant or subordinate status can be sacrificed. It would therefore be of interest to investigate whether the outcome of contests for social dominance among individuals with similar social histories can be predicted from observations made before social interaction.

Both in the wild and in captivity, a range of factors such as age, body size, sex, kinship, secondary sexual characteristics, prior dominance experience, or residence in a particular territory have been shown to affect social rank Abbott et al., 1985, Beacham, 1988, Beaugrand and Cotnoir, 1996, Cote, 2000, Renison et al., 2002, Sprague, 1998. The majority of these studies do not address what innate or acquired traits initially helped individuals gain the characteristics that accompany dominance. Some studies, however, suggest that differences in behavior and physiology predispose an animal for a certain social rank, at least in a predictable and stable experimental environment McCarthy, 2001, Morgan et al., 2000, Plusquellec et al., 2001, Pottinger and Carrick, 2001.

In a breeding program with rainbow trout (Oncorhynchus mykiss), two lines of fish with divergent cortisol responsiveness have been created (high-responding, HR; and low-responding, LR) (Pottinger and Carrick, 1999). Fish from the LR line show a strong tendency to become dominant over HR individuals in staged fights for social dominance between pairs (Pottinger and Carrick, 2001). The LR fish also show a reduced locomotor response to a simulated territorial intrusion test and a more rapid resumption of feed intake after transfer to a new environment (Øverli et al., 2002a). Thus, the HR and LR lines of rainbow trout might represent selection for different behavioral–physiological stress coping styles, as defined by Koolhaas et al. (1999). In the study reported here, we investigate whether individual differences in behavioral and physiological components of stress coping style predict social rank even in an unselected population (i.e., no prior identification of LR or HR lineage) of rainbow trout.