In mammals, the synthesis of serotonin (5-HT) in the brain is dependent upon availability of its precursor, the amino acid tryptophan (TRP) (Leathwood, Reference Leathwood1987; Fernstrom & Wurtman, Reference Fernstrom and Wurtman1997). This occurs because the rate-limiting enzyme in the biosynthetic pathway of 5-HT, tryptophan hydroxylase, is unsaturated with TRP under normal physiological conditions. Further, TRP enters the brain from the circulation through a protein transporting not only TRP but also other large neutral amino acids (LNAA) (Fernstrom & Wurtman, Reference Fernstrom and Wurtman1997). Dietary manipulations affecting plasma TRP/LNAA ratios have been used as a tool to investigate involvement of the 5-HT system in behaviour, mood and cognition in humans (Markus et al. Reference Markus, Panhuysen, Jonkman and Bachman1999, Reference Markus, Panhuysen, Tuiten and Koppeschaar2000). Such manipulations have also been shown to affect the endocrine and behavioural response to stress in comparative vertebrate models (Winberg et al. Reference Winberg, Øverli and Lepage2001; Lepage et al. Reference Lepage, Tottmar and Winberg2002, Reference Lepage, Vilchez, Pottinger and Winberg2003; Höglund et al. Reference Höglund, Bakke, Øverli, Winberg and Nilsson2005).
A characteristic behavioural response to stress in all vertebrates appears to be a reduction in food intake (Carr, Reference Carr2002). Stress-induced anorexia could in part be an effect of central 5-HTergic activation, since a general consensus is that increased 5-HT neurotransmission inhibits food intake (e.g. Leibowitz & Alexander, Reference Leibowitz and Alexander1998). However, information on the appetite-regulating role of 5-HT in vertebrates other than mammals is meagre. In teleosts, chronic stress, which induces increased brain turnover of 5-HT, is associated with suppressed appetite (Øverli et al. Reference Øverli, Winberg, Damsgård and Jobling1998; Ortega et al. Reference Ortega, Renner and Bernier2005). Further, a central anorectic action of 5-HT was demonstrated by intracerebroventricular injection of 5-HT in goldfish (De Pedro et al. Reference De Pedro, Pinillos, Valenciano, Alonso-Bedate and Delgado1998) and also by intraperitoneal injections of the serotonin-releasing agent fenfluramin in rainbow trout (Ruibal et al. Reference Ruibal, Soengas and Aldegunde2002). TRP treatment, on the other hand, has been shown to act in both a stimulatory and an inhibitory way, and to have no effect on appetite in domestic hens (Rosebrough, Reference Rosebrough1996), pigs (Seve, Reference Seve1999) and rats (Haleem et al. Reference Haleem, Jabeen and Parveen1998).
Studies in rainbow trout show that enhancement of 5-HT signalling by TRP-enriched feed has context-dependent effects on the endocrine stress response. In undisturbed individuals, a 7 d treatment with TRP-enriched feed induced a slight increase in cortisol release, while individuals exposed to an acute stressor showed decreased cortisol release (Lepage et al. Reference Lepage, Tottmar and Winberg2002, Reference Lepage, Vilchez, Pottinger and Winberg2003). Long-term enhancement of central 5-HT signalling by selective serotonin re-uptake inhibitors also inhibits the endocrine stress response in both teleost fish and mammals (Jensen et al. Reference Jensen, Jessop, Harbuz, Mork, Sanchez and Mikkelsen1999; Lepage et al. Reference Lepage, Larson, Mayer and Winberg2005). However, long-term effects of enhanced 5-HT signalling on behavioural responses such as feeding have, to our knowledge, not been investigated in teleosts. The aim of the present study was to investigate if stress-induced changes in feeding behaviour were affected by 7 d of pre-treatment with TRP-enriched feed in brown trout.
Material and methods
Animals
The experimental fish were 1-year-old brown trout (Salmo trutta) (body mass, 19·6 g (sd 2·6), n 16). Before the start of the experiment, fish were kept indoors in a holding tank (200 fish/m2), continuously supplied with aerated water at 8–10°C, for >2 months. The fish were hand-fed with commercial trout pellets (Ewos ST40) at 1–2 % of the body mass per day. The light/dark regime was 12/12 h.
Experimental fish feed
Wet feed was prepared following Winberg et al. (Reference Winberg, Øverli and Lepage2001). In brief, 1 kg of herring (Clupea harengus) fillets and 1 kg of shrimp (Pandalus borealis) were minced together and divided into two 1 kg portions. Hot water (0·5 l) with gelatine (50 g) was added to each portion. One feed batch was supplemented with 3 g of TRP, resulting in a TRP/LNAA ratio of 0·22. The unsupplemented feed batch with a [TRP]/[LNAA] ratio of 0·06 served as control feed. Thin sheets (3 mm thick) of the feed mixes were frozen and kept at − 20°C. Before use, frozen feed was divided in 3 × 3 × 3 mm cubes with a scalpel. Amino acid analysis of the two experimental feeds was carried out according to approved EU methods (OJ, 1998, 2000) on a Biochrom 30 Amino Acid Analyser (Biochrom Ltd, Cambridge, UK) for the LNAA, except TRP which was analysed in a Dionex Summit HPLC system, with the addition of a Shimadzu RF-535 fluorescence detector.
Experimental protocol
The experiment was divided into three phases; acclimation; TRP pre-treatment; and a stress phase. In all these phases, 16 fish were kept individually isolated. Feeding behaviour was quantified daily by the feeding score described by Øverli et al. (Reference Øverli, Sørensen and Nilsson2006). In brief, from the first day after transfer to social isolation, fish were hand-fed, and feeding behaviour was rated according to the following scale:
Score 0: the fish did not show any interest in the presented feed, and was immobile in a corner of the aquarium.
Score 1: the fish showed some interest in the presented feed, but did not move or moved little and slowly to catch feed, and consumed 0–5 % of the offered feed.
Score 2: the fish showed interest in the presented feed and left the location where they were placed before feeding to catch feed, but always returned to this place between consuming feed pieces. Fish given this score consumed 5–50 % of the offered feed.
Score 3: The fish showed a high degree of interest in the presented feed, searched for feed actively and did not return to the spot where they were located before feeding. Fish given this score consumed 50–100 % of the offered feed.
On this scale, 100 % of the offered experimental wet feed equals a feeding ratio at 10 % of the body mass per day (corresponding to approximately 1 % dry feed/kg; Winberg et al. Reference Winberg, Øverli and Lepage2001). The acclimation and the TRP pre-treatment phases were conducted in four 100 cm × 50 cm × 50 cm glass aquaria, continuously supplied with aerated water (1 l/min, 10–12°C). These aquaria were divided into four chambers by grey plastic walls, where the fish were kept individually isolated. The stress phase was conducted in plastic 25 cm × 20 cm × 20 cm aquaria, continuously supplied with aerated water (1 l/min, 10–12°C).
The acclimatisation phase lasted from day 1, after transfer from the holding tank to the observations aquaria, to day 28. At the end of this period, the mean feed response index reached 1·8 (Fig. 1), and all fish were reactive to the presented feed.
Mean feeding score during the acclimatization (A), l-typtophan (TRP) pre-treatment (B) and stress phase (C) in juvenile brown trout. One group (n 7) was fed wet feed supplemented with 3 g/kg TRP 7 d before transfer and unsupplemented control feed during the acclimatization and stress phase, while the control-treated group (n 8) was fed unsupplemented feed throughout the experiment. An arrow indicates transferring the fish to a new environment. Values are means and sem. (■) TRP; (□) control.
In the TRP pre-treatment phase, eight fish were given TRP-supplemented experimental feed for 7 d (day 29–35), whereas the remaining eight fish were given control feed throughout the experiment. One individual in the TRP treated group was excluded from the data analyses, since it did not consume any feed during the last 2 d of the TRP pre-treatment phase.
On the morning before feeding at the start of the stress phase (day 36), fish were netted and lifted up from the glass aquaria and inserted individually into much smaller plastic aquaria. After transfer, both the TRP-pre-treated group and the control group received control feed, and the feeding behaviour was quantified on the day of transfer and for a further 3 d (day 37–39).
Statistical analyses
All data are presented as mean and standard error of the mean (sem) unless stated otherwise. When analysing the effects of TRP pre-treatment on stress-induced change in feeding behaviour, the total feeding score observed during 4 d following transfer was analysed with the Mann–Whitney U test. Statistical analyses were performed using Statistica 5.1 (StatSoft Inc.).
Results and Discussion
The stress of being transferred and put into smaller aquaria resulted in a decrease in feeding score in both TRP pre-treated and control-treated groups. However, a significantly higher total feeding score in the TRP-pre-treated group compared with the control-treated group was observed during the 4 d following transfer (0·8 (sd 0·16) v. 0·4 (sd 0·13); P < 0·05, Z = 1·97, Mann–Whitney U test), suggesting that TRP attenuates stress-induced anorexia.
Previous studies of the effect of TRP treatment on feeding behaviour and/or growth rate in vertebrates show contradictory results (Rosebrough, Reference Rosebrough1996; Haleem et al. Reference Haleem, Jabeen and Parveen1998; Seve, Reference Seve1999; Hussein et al. Reference Hussein, Cantor, Pescatore, Gates, Burnham, Ford and Paton2001; Winberg et al. Reference Winberg, Øverli and Lepage2001; Hseu et al. Reference Huseu, Lu, Su, Wang, Tsai and Hwang2003; Tormo et al. Reference Tormo, de Tejada, Morales, Paredes, Sanchez, Barriga and Hernandez2004; van Hierden et al. Reference van Hierden, Koolhaas and Korte2004; Papoutsoglou et al. Reference Papoutsoglou, Karakatsouli and Chiras2005).
The present study demonstrates an improved feeding after a stressful event in TRP-pre-treated fish compared with control-treated fish. This suggests context-dependent effects of TRP on feeding behaviour. Furthermore, this observation implies that differences in uncontrolled stress during experimentation could be the reason for some of the variation between previous studies on the effect of TRP on feeding behaviour.
The inclusion rate of TRP used in the present study has been shown to enhance 5-HT signalling in juvenile grouper (Epinephelus coioides) and in rainbow trout (Winberg et al. Reference Winberg, Øverli and Lepage2001; Lepage et al. Reference Lepage, Tottmar and Winberg2002, Reference Lepage, Vilchez, Pottinger and Winberg2003; Huseu et al. Reference Huseu, Lu, Su, Wang, Tsai and Hwang2003). This observation is in accordance with the general consensus that increased availability of TRP to the brain enhances central 5-HTergic synthesis in vertebrates (Fernstrom & Wurtman, Reference Fernstrom and Wurtman1997; Aldegunde et al. Reference Aldegunde, Soengas and Rozas2000; Winberg et al. Reference Winberg, Øverli and Lepage2001; Höglund et al. Reference Höglund, Bakke, Øverli, Winberg and Nilsson2005).
5-HT is generally considered to have an anorectic effect (e.g. Leibowitz & Alexander, Reference Leibowitz and Alexander1998) and to act in a stimulatory manner on the endocrine stress response (Chauloff, Reference Chauloff1993). A positive relationship between chronic stress, increased brain turnover of 5-HT and suppressed appetite has also been demonstrated in teleosts (Øverli et al. Reference Øverli, Winberg, Damsgård and Jobling1998; Ortega et al. Reference Ortega, Renner and Bernier2005). Further, an anorectic action of 5-HT has been demonstrated by central administration in goldfish (De Pedro et al. Reference De Pedro, Pinillos, Valenciano, Alonso-Bedate and Delgado1998) and oral administration in European sea bass (Rubio et al. Reference Rubio, Sanchez-Vazquez and Madrid2006). However, the present study and previous studies in rainbow trout show that TRP treatment does not affect feeding behaviour in undisturbed fish. Moreover, studies in rainbow trout demonstrate a suppressive effect on the endocrine stress response by TRP-enriched feed (Lepage et al. Reference Lepage, Tottmar and Winberg2002, Reference Lepage, Vilchez, Pottinger and Winberg2003). Furthermore, this and other effects, such as decreased aggression (Winberg et al. Reference Winberg, Øverli and Lepage2001), which is associated with changes in the 5-HTergic system, were first evident after 7 d of TRP treatment. This suggests that the behavioural and endocrine effects of TRP-enriched feed are mediated by an indirect mechanism related to increased 5-HTergic activity and release. The effects of TRP and 7 d of treatment with the serotonin reuptake inhibitor citalopram on aggression and the endocrine stress response are remarkably similar (Lepage et al. Reference Lepage, Larson, Mayer and Winberg2005). Since both serotonin reuptake inhibitors and TRP are expected to increase 5-HTergic synaptic transmission acutely, Lepage et al. (Reference Lepage, Larson, Mayer and Winberg2005) suggested common long-term mechanisms, such as changes in 5-HTergic receptor expression or sensitivity, for these two treatments. Moreover, brain 5-HT is under the control of and controls the release of hypothalamic neuropeptides, playing important roles in mediating stress-induced inhibition of feeding (Bernier, Reference Bernier2006). Among these are the melanocortins and corticotrophin-releasing hormone-like peptides (Carr, Reference Carr2002; Bernier, Reference Bernier2006). Further studies are needed to investigate if TRP enrichment affects these neuropeptides, and changes in 5-HTergic receptor expression and sensitivity.
In conclusion, the present study demonstrates that pre-treatment with dietary TRP attenuates stress-induced anorexia in brown trout. Further, the present result also shows that the behavioural effect of TRP is context dependent, which could explain some of the variation in results between previous studies on the effect of TRP on feeding behaviour.
