Animals often fight with each other over access to limited resources (mates, breeding sites, food, shelters etc.). As fighting is potentially costly to contestants (e.g.
in time and energy, Neat et al. 1998; higher predation risk, Brick 1999; physical injuries, Austad 1983), it is beneficial for animals to evaluate the potential costs and benefits associated with a contest and adjust their fighting strategy accordingly (Smith
& Price 1973; Smith & Parker 1976). Many studies have shown that animals can assess the quality and quantity of a resource: individuals usually fight longer and more intensely for more valuable resources. For example, male jumping spiders (Euophrys parvula) were more willing to escalate fights when a model female was presented (Wells 1988). Individuals’ assessments of resources can also be influenced by their internal states. For instance, hungrier hermit crabs were more aggressive and more likely to fight for food than less hungry ones (Laidre & Elwood 2008). That potential contest costs can influence animals’ contest decisions has also been demonstrated in numerous studies. Male cichlid fish (Nannacara anomala), for example, used lower-intensity contests and took longer to escalate into mouth wrestling when a model predator was presented, because intensifying a contest would compromise an individual’s ability to monitor predators (Brick 1999). These examples show that animals’ contest behavior is sensitive to the costs and benefits associated with the contest.
Past contest experience has also been found to influence individuals’ behavior in subsequent contests. Individuals with a recent winning experience often become more aggressive, are more likely to initiate future contests and to win again (winner effect),
while individuals with a recent losing experience tend to display the opposite behavior and lose again (loser effect). (See Hsu et al. 2006 for a review.) Winner and loser effects are usually hypothesized to result from prior winning and losing experiences affecting an individual’s assessment of its own fighting ability and therefore its estimation of contest costs (Whitehouse 1997). Winner or loser effects have been reported for animals of all different taxa, such as burying beetle (Nicrophorus humator, Otronen 1990), crab spider (Misumenoides formosipes, Hoefler 2002),
crayfish (Orconectes rusticus, Bergman et al. 2003), Siamese fighting fish (Betta splendens, Baenninger 1970), copperhead snake (Agkistrodon contortrix, Schuett
1997) and blue-footed booby (Sula nebouxii, Drummond & Canales 1998). Despite the prevalence of winner/loser effects, the physiological mechanisms underlying them are still unclear.
1.2 Contest Behavior and the Neuroendocrine System
Contest behavior is modulated by the neuroendocrine system; steroid hormones (e.g. androgens, estrogens and glucocorticoids) and neurotransmitters (e.g. serotonin, dopamine and norepinephrine) are frequently explored because of their associations with aggressiveness and dominance status (see Nelson & Chiavegatto 2001 for a review). There have been many studies of the relationship between contest behavior and steroid hormones and/or neurotransmitters and many of their results are contradictory. Their major findings are summarized below.
1.2.1 Steroid Hormones and Contest Behavior
Androgens
Androgens are steroid hormones which mediate the development and maintenance of male sex organs and secondary sex characteristics in vertebrates by binding to androgen receptors. In vertebrates, androgens are mainly produced by the gonads, but the adrenal cortex also secretes low levels of them. Contest interaction may directly affect gonadotropin-releasing hormone neurons (GnRH neurons) to activate the hypothalamic-pituitary-gonad (HPG) axis and stimulate the secretion of androgens (Francis et al. 1993). While testosterone is the primary androgen in reptiles, birds and mammals, 11-ketotestosterone is the primary androgen in fish (Borg 1994;
Vermeulen et al. 1994).
Androgen levels and receptor expression levels are frequently shown to affect an individual’s contest behavior and social status. For instance, individuals with higher (natural or exogenous) testosterone levels tend to behave more aggressively and/or achieve higher social status in mice (Mus musculus, Zielinski & Vandenbergh 1993) and lambs (Ovis aries, Ruiz-de-la-torre & Manteca 1999); male mice exhibiting a spontaneous mutation that fails to generate the long form of the androgen receptor are less aggressive than normal individuals (Olsen 1983; Maxson 2000). Contest interaction and social status, interestingly, also affect an individual’s androgen levels.
Defeated rhesus monkeys (Macaca mulatta), rodents and birds usually have depressed plasma testosterone levels (Harding 1983; Leshner 1983; Huhman et al. 1991), and repeated winning experiences significantly increased testosterone levels in California mice (Peromyscus californicus, Oyegbile & Marler 2005, 2006).
Estrogens
Estrogens are a group of steroid compounds which promote the development of female secondary sexual characteristics and are involved in regulating the menstrual
cycle. In vertebrates, estrogens are mainly produced in the ovaries by the developing follicles and the corpus luteum. The adrenal cortex and the mammary glands, however, also secrete low levels of them (Goodman 2008).
Research indicates that levels of both estrogen and estrogen receptor gene expression affect an individual’s contest behavior and aggressiveness. Injection with estradiol promotes male-like aggressive behavior and increases the probability of winning contests in female mice (Simon & Gandelman 1978). Of three kinds of estrogen receptors (α-, β- and γ-isoform, Giguère et al. 1988; Hong et al. 1999), estrogen α and β receptors were most well-studied and associated with contest behavior and aggressiveness. After a targeted knockout of the estrogen-α receptor gene (ERαKO), male mice exhibit decreased levels of aggression in resident-intruder paradigm tests, while, after a knockout of the estrogen-β receptor (ERβKO), mice display normal or increased levels of aggression. ERαKO female mice, however, exhibit increased aggression towards other intruder females mice relative to wild-type females (Ogawa et al. 1997; Ogawa et al. 1998; Ogawa et al. 1999; Ogawa et al.
2000). These results revealed that α- and β-isoforms of estrogen receptor might mediate aggression in opposite directions, and that the effects of the α-isoform on aggression might differ between male and female mice. Contest interaction and social status also affect an individual’s estrogen-receptor expression; for example, subordinate or defeated male African cichlids (Astatotilapia burtoni) have lower ERβa and ERβb levels in the anterior brain than dominants or winning males (Burmeister et al. 2007).
Glucocorticoids
Fighting can induce short-term but strong stress responses in animals (Miczek et
al. 2002). In vertervrates, the stress response is mainly mediated by the neuroendocrine system. In mammals, stress can induce changes in the hypothalamic-pituitary-adrenal axis (HPA axis) which activates the adrenal cortex to release glucocorticoids; in fish, stress acts through the hypothalamic-pituitary-interrenal axis (HPI axis) activating the interrenal gland to release glucocorticoids. Glucocorticoids increase an individual’s ability to face a stressful environment (McEwen 2000) by supressing the function of insulin, decreasing the synthesis of glycogen and elevating the level of blood glucose (Genuth 1993). The primary glucocorticoid is corticosterone in amphibians, reptiles, birds and rodents and cortisol in fish and primates (Wendelaar Bonga et al. 1995).
Glucocorticoid levels and receptor-expression levels affect an individual’s contest behavior and aggressiveness. For example, long-term cortisol treatment in rainbow trout (Oncorhynchus mykiss) inhibits aggressive behavior, while short-term cortisol treatment does not (Øverli et al. 2002). In lizards (Anolis carolinensis), blocking cortisol receptors with mifepristone reduces aggressive attacks/displays during the early stages of aggressive interaction (Summers et al. 2005). An individual’s glucocorticoid levels are influenced by contest interaction: losing a fight often induces an increase in glucocorticoid levels (Sakakura et al. 1998; Schuett &
Grober 2000; Øverli et al. 2004) and a loser’s post-fight glucocorticoid level correlates with the amount of aggression it was subjected to in the fight (Winberg &
Lepage 1998; Elofsson et al. 2000; Sloman et al. 2001; Earley & Hsu 2008).
1.2.2 Neurotransmitters and Contest Behavior
Several neurotransmitters (e.g. serotonin, NO, dopamine) are closely associated with animal aggression. Among these, serotonin (5-hydroxytryptamine; 5-HT) has
been associated not only with aggressive behavior (Edwards & Kravitz 1997; Weiger 1997) but also with dominance status (Winberg & Nilsson 1993). Serotonin is synthesized from L-tryptophan via tryptophan hydroxylase (Aldegunde et al. 2000).
Neurons producing and storing serotonin are located primarily in the raphe nuclei of the brainstem. In mammals the projections of the raphe nuclei extend to various brain regions, including the hippocampus, cerebral cortex, amygdala, hypothalamus and pituitary gland. Once released into synapses, serotonin binds to various receptors (Hoyer et al. 2002). Serotonin is mainly metabolized to 5-hydroxyindoleacetic acid (5-HIAA) by monoamine oxidase (MAO). The activity of the serotonergic system or the turnover of serotonin is often reported as the ratio of 5-HIAA: 5-HT (Winberg et al. 1992; Koutoku et al. 2003; Dias & Crews 2006).
Serotonin and related receptor-expression levels influence an individual’s contest behavior and aggressiveness. Acute serotonin treatment, for example, decreases aggressive behavior (latency to attack and duration of attacks) in Siamese fighting fish (Betta splendens, Clotfelter et al. 2007). And, blocking the serotonin reuptake receptors (i.e. facilitating chronic serotonin elevation) of dominant individuals causes these individuals to become less aggressive (fewer attacks and displays) and lose to previous losers in the green anole lizard (Anolis carolinensis, Larson & Summers 2001). Activating 5-HT1A receptors reduces aggressive behavior: acute treatment with a 5-HT1A receptor agonist decreases aggressive behavior in Siamese fighting fish (Clotfelter et al. 2007). Not surprisingly, contest interaction in turn affects an individual’s serotonin level. For instance, subordinate or defeated male lizards (Anolis carolinensis) have significantly increased 5-HIAA/5-HT ratios in the hippocampal
cortex and the nucleus accumbens after a fight, ratios which gradually decrease over a week (Summers et al. 1998).
Because of the close association between the neuroendocrine system and an
individual’s contest decisions, the influence of a recent winning/losing experience on an individual’s contest decisions is probably mediated through this circuit. My research project therefore aimed to explore this possibility.
1.3 Winner/Loser Effects and the Neuroendocrine System
If winner and loser effects are mediated by the neuroendocrine system, winning and losing experiences may temporarily alter an individual’s hormone or receptor-expression levels which may then affect its subsequent contest behavior.
Several studies have investigated the relationship between the winner effect and testosterone levels, while there is no literature regarding the physiological mechanism underlying the loser effect.
In a study of California mice (Trainor et al. 2004), castrated male individuals were given a forced winning experience and then injected with testosterone.
(Members of the control group were injected with saline.) The result was that individuals injected with testosterone displayed a significant winner effect, while the control individuals did not. This study indicated that a winner effect might accompany a temporary increase in testosterone. It is also possible, however, that the behavior changes simply because of the exogenous testosterone rather than because of a winner effect. In another study of male California mice, the authors compared effects of three, two, one or zero winning experiences on the probability of winning a subsequent encounter; testosterone and corticosterone were also measured after the final encounter (Oyegbile & Marler 2005). The results showed that individuals with more winning experiences had a significantly higher probability of winning a subsequent encounter (winner effect) and had significantly higher post-encounter testosterone.
However, because the hormone levels immediately after winning experiences were
not measured, these results could not demonstrate a direct linkage between winning experiences and the circulating testosterone levels.
One further study by Fuxjager et al. (2010) also investigated the relationship between the winner effect and androgen-receptor gene expression. The author provided male California mice with three winning experiences in either their home cage or unfamiliar cages, and then compared the probability of winning a subsequent encounter in their home cage or unfamiliar cages; androgen- and progestin- receptor expression levels in various brain regions were also measured after the final encounter.
The results showed that individuals with winning experiences and a subsequent encounter in thei home cage had significantly higher androgen-receptor gene/protein expression in the nucleus accumbens and ventral tegmental area brain regions. This study indicated that winning experiences in combination with a contest in home cage increased androgen receptor gene/protein expression in specific brain regions (NAcc
& VTA regions). However, because individuals which received three winning experiences in their home cage displayed higher aggression in subsequent contests and the expression levels were measured after the contests (but not after winning experiences), it is possible that the androgen receptor gene/protein expressions increased because of the elevated aggressive interaction in the home-cage contests, rather than because of the winning experiences. The results of the study, unfortunately, could therefore not demonstrate a direct link between winning experiences and the elevated androgen receptor gene/protein expression.
Overall, whether winner/loser effects are mediated by hormone levels or receptor levels remains unclear although both hormones levels and receptor expression levels are highly correlated with contest behavior.
1.4 Previous Studies of Kryptolebias marmoratus
Kryptolebias marmoratus is a species of mangrove killifish. It is an ideal organism for examining the behavioral and the neuroendocrine mechanisms underlying winner and loser effects, because it displays both winner and loser effects and because these effects last at least 48 hours (Hsu & Wolf 1999, 2001). Previous studies also demonstrated that the contest behavior in this fish is highly correlated with hormone levels (Earley & Hsu 2008). Pre-fight cortisol related negatively and pre-fight testosterone related positively to contest initiation and winning probability, particularly in the smaller opponent. In pairs where the larger fish won the contest, winners with higher pre-fight testosterone and lower pre-fight cortisol attacked the losers more frequently. Losers that escalated with winners had significantly higher post-fight levels of all three hormones than losers that retreated without escalation.
Also, winners that attacked losers at higher frequency had higher levels of post-fight cortisol (Earley & Hsu 2008). In this fish, however, levels of testosterone, 11-ketotestosterone, estradiol and cortisol did not change after (randomly assigned) wins or losses (Lee 2009; Lu 2010), which indicated that winner and loser effects are not mediated by the circulating levels of these hormones. Because receptors could also modulate these hormones’ effect on behavior, I further explored the relationship between winner/loser effect and receptors in this fish by investigating the changes in receptors after the fish received randomly assigned winning or losing experiences.
1.5 Objectives
The primary objective of this study was to use K. marmoratus to explore the roles of hormone and serotonin receptors in mediating winner and loser effects. To do this, I examined the temporal changes (0 hr, 3 hr, 48 hr after experience training) in glucocorticoid receptor (GR), androgen receptor (AR), estrogen receptor α (ERα),
estrogen receptor β (ERβ) and 5-HT1A receptor (5-HT1AR) gene expression levels after both wins and losses. Furthermore, the fish were given either one or three winning, losing or no recent contest experiences to examine whether a different number of experiences would have a different effect on these receptors.
If winner and loser effects are mediated by the changes in receptor expressions, I would expect to discover that: (1) For those individuals with winning experience(s), AR and ERα gene expression levels increase in 0 hr and 3 hr after experience training, then gradually decrease in 48 hr, and/or ERβ, 5-HT1AR and GR decrease in 0 hr and 3 hr after experience training, then gradually increase in 48 hr. (2) For those individuals with losing experience(s), AR and ERα gene expression levels decrease in 0 hr and 3 hr after experience training, then gradually increase in 48 hr, and/or ERβ, 5-HT1AR and GR gene expression levels increase in 0 hr and 3 hr after experience training, then decrease in 48 hr.
Materials and Methods
2.1 Study Organism
The mangrove killifish K. marmoratus is one of two vertebrate species known to reproduce by self-fertilization (Costa et al. 2010). It inhabits mangrove areas ranging from south-eastern Brazil, Venezuela, much of the Caribbean, the Bahamas and Yucatan to southern Florida (Kallman & Harrington 1964; Harrington & Kallman 1968). Most populations of the fish exist in nature as isogenic, homozygous strains, although outcrossing heterozygous populations have been discovered in Twin Cays, Belize (Taylor et al. 2001; Mackiewicz et al. 2006a; Mackiewicz et al. 2006b). This fish is aggressive in both the field and the laboratory (Taylor 1990). Because it displays both winner and loser effects (Hsu & Wolf 1999, 2001; Hsu et al. 2009) and the contest behavior in this fish is highly correlated with hormone levels (Earley &
Hsu 2008), it is an ideal organism for exploring these effects’ underlying physiological mechanisms. This study used individuals of three strains of K.
marmoratus from different geographical areas (DAN2K: Dangria, Belize, collected in
2000; RHL: San Salvador, Bahamas, collected in 2001; SLC8E: St. Lucie County, Florida, U.S.A., collected in 1995) which were originally collected from the field by Dr. D. Scott Taylor, (Florida, U.S.A.). Fish were kept individually in a 13 × 13 × 10 cm translucent plastic container (maintenance container). Every container was filled with 800 ml of approximately 25 ppt synthetic sea water (Instant Ocean® powder) and labeled with a unique code for individual identification. Two holes were drilled close to the upper edges of each container for feeding and aeration. Because the fish is capable of respiring through its skin (Grizzle & Thiyagarajah 1987), no extra aeration was needed. Fish were maintained at 25 ± 2 °C on a 12:12 h light:dark cycle and fed
newly hatched brine shrimp (Artemia) nauplii at 3:00 pm every day. The fish can be less than 2 cm in standard length (from the tip of the snout to the end of the caudal peduncle) when they first reach maturity. For ease of handling, only fish larger than 26 mm in standard length (> 6 months old) were used.
2.2 Experimental Design and Procedures
2.2.1. Experimental Design
This study aimed to examine changes in the expression levels of steroid hormone and serotonin receptor genes after one or three recent wins/losses and whether these changes decay with the passage of time. Before they were exposed to any of the treatments, individual fishes’ levels of testosterone (T) and cortisol (CORT) were measured as indicators of their baseline physical condition. A total of six experience types (three winning experiences, ET3W; three losing experiences, ET3L; a control (for ET3W and ET3L) with three no-contest experiences, ET3N; one winning experience, ET1W; one losing experience, ET1L; a control (for ETW and ETL) with one no-contest experience, ET1N) and three time-decay treatments (0 hour delay, 0h; 3 hours delay, 3h and 48 hours delay, 48h) were implemented for this study, for a total of 18 treatment combinations (Table 1).
Results from previous studies (Rosa et al. 2002; Mohanan et al. 2005) indicated that a sample size of 4-6 for each treatment should be enough to detect the differences.
A total of 298 fish were available for this study and were randomly assigned to the 18 treatments (n = 16 or 17 for each of the treatments).
2.2.2. Experimental Procedures (Fig. 1)
Baseline Hormone Collection
On the first day of the experiment (Day 1), I measured each individual’s baseline hormone levels, following the procedures for hormone sample collection and analysis from Earley & Hsu (2008). To minimize the impact of the daily cycle which is likely to exist for steroid hormones (Emata et al. 1991), I collected all baseline hormone samples at the same time of day.
At 9:30 am, each focal individual was placed in a glass beaker with 400 ml clean 25 ppt synthetic sea water and left in the beaker for 1 hour. After 1 hour, the focal individuals were removed from the beakers and returned to their maintenance containers. Hormones were then extracted from the water samples using C18 solid phase extraction columns (Lichrolut RP-18, 500 mg, 3.0 ml; Merck, NJ, USA) fitted to a 24-port manifold. Columns were first primed with 2 consecutive washes with 2 ml HPLC grade methanol (MeOH) followed by 2 consecutive washes with 2 ml ultrapure water. Tubing then was fastened securely to the top of each column and placed into the water sample collected from the fish. A vacuum was engaged and the water sample passed through the tubing into the column. Columns then were frozen (-20 ºC) until further processing. Freeze storage of water samples and columns has been determined not to impact steroid concentrations (Ellis et al. 2004).
Experience Training
On Days 2 to 4, the focal individuals pre-assigned to have three experiences (ET3W, ET3L and ET3N individuals) received one winning, losing or no experience per day for three consecutive days. The standard methods for providing winning and losing experiences to the fish are described below. The focal individuals pre-assigned to receive one experience (ET1W, ET1L, and ET1N individuals) received
their single experience on Day 2.
Fish Dissection and Tissue Collection
Immediately after the completion of the experience training, individuals assigned to the 0 hour delay treatment were decapitated and their heads and gonads preserved in RNAlater® (Applied Biosystems/Ambion Inc., TX, USA) and stored in a freezer at -80 ºC for subsequent examination of the gene expressions of steroid hormone receptors and serotonin receptors (section 2.6). Three hours and 48 hours after the completion of the experience training, individuals assigned to the 3 hour and 48 hour decay treatments were decapitated and their heads and gonads were preserved as above.
2.3 Providing a Losing/Winning Experience
To ensure that focal individuals lost or won, I fought them against much larger/smaller (difference > 2 mm) standard winners/losers from the same clone. I staged contests among several large fish and the one that defeated all others was designated a standard winner, to provide losing experiences to the focal individuals of its strain. Small fish were fought among themselves and the one that lost to all others
To ensure that focal individuals lost or won, I fought them against much larger/smaller (difference > 2 mm) standard winners/losers from the same clone. I staged contests among several large fish and the one that defeated all others was designated a standard winner, to provide losing experiences to the focal individuals of its strain. Small fish were fought among themselves and the one that lost to all others