Introduction

Growth hormone (GH) is produced by the adenohypophysis, in fish, and functions as a pluripotent endocrine regulator of many physiological processes (Björnsson et al. 2002; Reineke et al. 2005). Most importantly, GH plays a critical role during growth regulation and studies, mainly from salmonids, demonstrate that a functional GH—insulin like growth factor I (IGF-I) axis exists whereby GH stimulates increased levels of IGF-I, in fish tissues (Cao et al. 1989; Duan and Plisetskaya 1993; Sakamoto and Hirano 1993; Shamblott et al. 1995; Perrot and Funkenstein 1999). The actions of GH, on fish cells, occurs by binding to a GH receptor (Pérez-Sánchez et al. 2002) whereas GH secretion, in fish, is under hypothalamic control via a complex array of regulators, including somatostatin (SRIF), GH releasing hormone (GHRH), dopamine (DA), gonadotropin releasing hormone (GnRH) and ghrelin (for review see Björnsson et al. 2002).

To date, GH genes have been cloned and characterized from many fish species, including salmonids (Sekine et al. 1985, 1989; Agellon and Chen, 1986; Nicoll et al. 1987; Agellon et al. 1988; Gonzalez-Villasenor et al. 1988; Rentier-Delrue et al. 1989a; Lorens et al. 1989), sparids (Momota et al. 1988; Funkenstein et al. 1991; Knibb et al. 1991; Tsai et al. 1993; Almuly et al. 2000; Deane and Woo 2006a) and other commercially important fish such as tilapia (Rentier-Delrue et al. 1989b; Ber and Daniel 1992, 1993; Sekkali et al. 1999), grouper (Li et al. 2005), tuna (Sato et al. 1988) and halibut (Einarsdottir et al. 2002). Early and more traditional methods to measure fish GH levels include somatotroph cell morphology, pituitary GH content analysis and radioimmunoassay for plasma GH levels. However, with the wide availability of fish GH genes alongside the use of modern molecular biological approaches, highly sensitive methods utilizing real time reverse transcriptase polymerase chain reaction assays are now applicable for absolute quantification of GH transcripts (Eppler et al. 2005).

Stress is well known to affect GH synthesis, release and circulating levels in mammals (De Feo 1996; Black 2002; Stokes 2003), but a similar consideration of GH as a stress-related hormone, in fish, is not well recognized. In this review, we assess how stressors related to alterations in salinity, changes in temperature, pollutants, handling/confinement, crowding, disease and nutrition modulate GH status in fish. We also provide a section on the emerging role and importance of GH during apoptosis and a final section on present status and future perspectives.

Effects of salinity on GH

Alterations in environmental salinity can result in fish becoming osmotically stressed. Maintaining water and ionic homeostasis is of critical importance as low salinity exposure imposes problems of severe ion depletion, alongside increased water entry, whereas in seawater (or hypersaline environments) excess ion intrusion as well as osmotic water loss must be contended with. The endocrine system of fish is critical in maintaining water and ionic homeostasis with cortisol, GH and IGF-I widely accepted as being important for seawater adaptation, or hypoosmoregulatory function, (McCormick 1995; Sakamoto and McCormick 2006) whereas prolactin has been extensively reported as being critical towards freshwater adaptation, or hyperosmoregulaory function, (Sakamoto and McCormick 2006). Although this review is written around the role and importance of GH during stress, we must not lose sight of these other key hormones, as part of the teleost osmosregulatory strategy, especially when they are faced with osmotic challenges. Early morphological and morphometric studies demonstrated changes in activity of fish somatotrophs during salinity adaptation (Olivereau and Ball 1970; Abraham 1974; Leatherland et al. 1974; Nagahama et al. 1977; Benjamin 1978; Nishioka et al. 1982) and within the past 30 years measurements of plasma GH, pituitary GH and pituitary GH mRNA have been used as indices for assessing the impact of salinity on fish. These data have been summarized and presented in Table 1 and should only be regarded as causal evidence for the importance of GH during seawater exposure and more substantial evidence stems from studies utilizing GH pre-treatment followed by increased salinity exposure. The earliest study to report on the importance of GH as a seawater adapting hormone was described by Smith (1956) whereby pre-treatment of brown trout, with GH, enhanced hypoosmoregulatory ability upon subsequent seawater exposure. Since this initial report, and due to the availability of both mammalian and fish GHs, treatment of fish, with GH, prior to salinity stress, has been widely used as an experimental protocol to test the seawater adapting effects of GH. Implantation of pellets containing ovine GH were found to induce seawater characteristics of Atlantic salmon smolts (Nonnotte and Boeuf 1995) and later studies, also on Atlantic salmon, demonstrated that GH treatment improved seawater adaptability in freshwater parr, pre-smolts (Seddiki et al. 1996) and salinity tolerance upon transfer from 12 ppt seawater to 34 ppt (McCormick 1996). In tilapia, GH treatment was also found to enhance seawater survival (Xu et al. 1997) and maintained hypoosmoregulatory function following hypophysectomy (Sakamoto et al. 1997). Similarly, GH treatment was also found to increase saltwater tolerance in coho salmon (Shrimpton et al. 1995) rainbow trout (Sangiao-Alvarellos et al. 2005) and the euryhaline killifish (Mancera and McCormick 1998a, b, 1999).

Table 1 Effects of salinity on growth hormone (GH) levels in fish (abbreviations used: FW, freshwater; BW, brackish water; SW, seawater; ISO, isoosmotic; ppt, parts per thousand)
Full size table

The key action of GH appears to be primarily associated with branchial osmoregulatory function including the activity and distribution of chloride cells (McCormick 2001; Evans 2002). Consistent with a seawater adapting role, GH treatment of freshwater rainbow trout (Perry 1998; Laurent et al. 1994; Bindon et al. 1994a, b), brown trout (Seidelin and Madsen 1999), sea trout parr (Madsen 1990) and tilapia (Xu et al. 1997) increased chloride cell activity, density and distribution. Associated with chloride cells are a number of ion transport mechanisms, including the sodium pump (Na+–K+-ATPase), a membrane bound enzyme that actively transports K+ into and Na+ out of a cell, against a concentration gradient (Geering 1990; Horisberger et al. 1991). Treatment with GH was found to increase branchial sodium pump activity in Atlantic salmon (Seddiki et al. 1996; Pelis and McCormick 2001; McCormick 1996), rainbow trout (Sangiao-Alvarellos et al. 2005), sea trout parr (Madsen 1990), brown trout (Madsen et al. 1995), Mozambique tilapia (Sakamoto et al. 1997), killifish (Mancera and McCormick 1998a, b, 1999) and air breathing climbing perch (Leena and Oommen 2000). The sodium pump is a heterodimeric enzyme composed of a catalytic alpha subunit and a glycosylated beta subunit, that are encoded for by separate genes, and recently it was shown, from in vitro studies, on isolated sea bream branchial filaments, that GH-mediated increases in sodium pump activity occur via upregulation of both genes (Deane and Woo 2005b). In contrast to the sodium pump, less is known about the role and importance of other ion transporters that are localized in the chloride cell, although GH treatment of Atlantic salmon was found to increase the abundance of a sodium-potassium-chloride ion co-transporter (Na+, K+, 2Cl−) presumably through a corresponding increase in gill chloride cell number (Pelis and McCormick 2001). Due consideration should also be given to the importance of cortisol, as it is plausible that GH may exert its effect synergistically with cortisol. Treatment of sea trout parr (Madsen 1990) killifish (Mancera and McCormick 1999), Atlantic salmon (McCormick 1996) and tilapia (Mancera and McCormick 1998b) with both hormones resulted in enhanced hypoosmoregulatory ability, following seawater transfer, than either hormone alone. In addition to the above studies, chloride cell number was also found to be increased in gills of Atlantic salmon that were treated with both GH and cortisol (Pelis and McCormick 2001) and exposure of tilapia pituitaries to cortisol was found to stimulate GH release implying a potential synergistic action for both cortisol and GH in seawater adaptation (Uchida et al. 2004). The precise mechanism as to how GH and cortisol exert a synergistic effect, on hypoosmoregulatory function in fish, remains to be fully elucidated although this could be related to an increased abundance of gill corticosteroid receptors, following GH treatment, as evidenced from studies on juvenile coho salmon (Shrimpton et al. 1995) and Atlantic salmon (Shrimpton and McCormick 1998). The seawater adaptive effects of GH could also be mediated via upregulation of IGF-I as concomitant increases in plasma GH and IGF-I were found in hatchery released Atlantic salmon smolts (McCormick et al. 2003) and freshwater rainbow trout transferred to 66% seawater (Shepherd et al. 2005). Treatment with exogenous IGF-I has been shown to increase salinity tolerance in killifish, Nile tilapia, Mozambique tilapia, striped bass (Mancera and McCormick 1998a, b, 1999) and Atlantic salmon (McCormick 1996). Also, increased sodium pump alpha subunit transcript and activity, in brown trout, was reported following IGF-I administration (Madsen et al. 1995; Seidelin et al. 1999) whereas in vitro exposure of sea bream branchial filaments, to IGF-I, resulted in transcriptional–translational upregulation of the sodium pump in a manner similar to that observed with GH exposure (Deane and Woo 2005b). Thus far, the evidence as to whether the seawater adapting role of GH is direct and/or mediated via IGF-I expression is not conclusive, however, continued advances in fish cell culture techniques alongside the use of molecular “knockout” protocols such as antisense disruption and RNAi should provide a clearer picture as to the role and importance of GH-mediated IGF-I expression as part of the fish osmoregulatory strategy.

Studies on eel and sparids contradict the widely reported seawater adapting role of GH, in fish, as exogenous GH administration did not modulate osmoregulatory function in eel (Sakamoto et al. 1993), silver sea bream (Deane et al. 1999a; Kelly et al. 1999) and seawater transferred gilthead sea bream (Mancera et al. 2002). In silver sea bream (Deane and Woo 2004, 2006a) and black sea bream (Deane and Woo 2005a), a clear correlation of increased pituitary GH transcript with increased environmental salinity was not apparent. Instead acclimation to conditions of isoosmotic (or near isoosmotic) salinity resulted in significantly highest amounts of pituitary GH transcript in silver sea bream (Deane and Woo 2004, 2006a) and black sea bream (Deane and Woo 2005a,) and increased abundance of pituitary GH in silver sea bream (Deane and Woo 2006a) and pituitary GH cells in gilthead sea bream (Mancera et al. 1995). Environmental salinity is known to influence growth of many fish species and in some species of fish, including sea bream, isoosmotic, conditions have been shown to improve growth (Boeuf and Payan 2001). To date, it has not been established as to how isoosmotic salinity may improve growth, in fish, although an interesting hypothesis has been proposed whereby isoosmotic salinity acclimation could enhance growth through an upwardly shifted pentose phosphate pathway via increased glucose-6-phosphate dehydrogenase activity (Woo and Kelly 1995; Deane and Woo 2005a, c). In addition, isoosmotic salinity can act to reduce stress proteins and stimulate non-specific immune function in sea bream (Deane and Woo 2004; Deane et al. 2002; Narnaware et al. 2000). Given that exogenous GH administration has been reported to increase hepatic glucose-6-phosphate dehydrogenase activity (Deane and Woo 2005c), lower hepatic stress protein abundance (Deane et al. 1999b) and enhanced macrophage phagocytic activity (Narnaware et al. 1997), in sparids, then it is possible that an isoosmotic salinity acclimation may stimulate GH production which in turn enhances growth, reduces stress and stimulates immune function.

Although much information has been gained over the past 25 years we still have much to learn and investigate in relation to the role and importance of GH in fish osmoregulation. Proportionally, only a small fraction of fish species have been studied and a number of unanswered questions still remain. Osmoregulation pertains to both ionic and water homeostasis and with the exception of the sodium pump very little is known regarding the role of GH on many other ionic exchange and water regulation processes. For example, chloride ion transport, via a cystic fibrosis transmembrane conductance regulator (CFTR), is known to be upregulated in fish upon seawater challenge (Marshall and Singer 2002) but we have yet to establish whether GH plays a key role in this process. Similarly, a number of water channel protein families (aquaporins) have been reported in fish with aquaporin 3 being the predominant protein family found in fish gills. As recent studies have shown that aquaporin 3 is reduced upon seawater/hypersaline acclimation in sea bream (Deane and Woo 2006b) and eel (Cutler and Cramb 2002) then we still need to answer whether GH plays a key role in regulating this process. In recent years, genes encoding the fish growth hormone receptor (GHR) have been cloned and sequenced (Calduch-Giner et al. 2001, 2003; Tse et al. 2003; Kajimura et al. 2004; Very et al. 2005; Benedet et al. 2005; Pierce et al. 2007) and from the information available, thus far, two distinct phylogenetic clades, designated as GHR1 and GHR2, have been identified (Saera-Vila et al. 2005; Jiao et al. 2006) Although, it was found that growth activation in Atlantic salmon vertebrae coincided with increased GH receptor gene expression (Wargelius et al. 2005) very little is known regarding how salinity changes can modulate expression of GHRs. To date, a single study on Mozambique tilapia has provided interesting insights into the differential tissue expression patterns of GHR1 and GHR2 in response to salinity as it was found that GHR1 mRNA levels were highest in gills of seawater adapted fish (in comparison to freshwater adapted fish) whereas GH2 mRNA levels were decreased in kidney of seawater adapted fish compared to fish in freshwater (Pierce et al. 2007). Recent evidence, also suggests that GHR expression, in fish tissues, can be modulated by steroid hormones, including cortisol (Jiao et al. 2006). As both cortisol and GH may act synergistically during seawater adaptation then another question that we still need to address is the role and importance of the GH receptor in this regulatory process. The above issues identified are not intended to provide a comprehensive review of what still needs to be understood in terms of the role of GH during salinity stress in fish, but instead aims to show that many factors, particularly at the cellular/molecular level certainly warrant future investigation.

Effects of temperature on GH

As fish are aquatic ectotherms, they have to be able to contend with continuous fluctuations in water temperature and alterations in expression profiles and activities of metabolic enzymes are known to vary in fish when exposed to temperature stress (Crawford and Powers 1989; Lin and Somero 1995; Seddon 1997). Whilst GH has been detected at early stages of embryo development and post hatching (Arakawa et al. 1992; Ayson et al. 1994; Deane et al. 2003) water temperature appears not to modulate differentiation of GH cells (Gabillard et al. 2003a) or GH expression (Gabillard et al. 2005) at these early stages of life. From the limited data available, thus far, it seems more likely that temperature-induced GH alterations occur at post larval development, probably from juveniles onwards, and in Table 2 we have summarized the key studies that have investigated GH changes in response to temperature changes. Seasonal variations in GH have been studied in perch (Swift and Pickford 1965), brown bullhead (Farbridge et al. 1985), coho salmon (Duan et al. 1995), goldfish (Marchant and Peter 1986), gilthead sea bream (Pérez-Sánchez et al. 1994a, b; Mingarro et al. 2002), rainbow trout (Barret and McKeown 1989) and carp (Figueroa et al. 2005) and in all cases peak GH was correlated with spring/summer months when water temperature is likely to be high. Given that day-length will also change on a seasonal cycle and would be longest during spring/summer then the importance of photoperiod, in modulating GH production, in fish, should be given due consideration. Indeed photoperiod has been implicated as key factor regulating fish growth (Boeuf and Le Bail 1999; Boeuf and Falcon 2001). Higher levels of GH have been shown to be correlated with increased photoperiod in Atlantic salmon (McCormick et al. 1995; Björnsson et al. 1997, 2000; Nordgarden et al. 2007) whereas lowered levels of GH in goldfish (Marchant and Peter 1986), coho salmon (Young et al. 1989) and gilthead sea bream (Pérez-Sánchez et al. 1994a, b) were found to be correlated with reduced day-length. Studies on Atlantic salmon also suggest that there could be a photoperiod dependant effect of temperature on GH levels (McCormick et al. 2000, 2002) which is consistent with a concerted temperature/photoperiod regulation on fish endogenous GH. Therefore and based on findings, thus far, temperature and photoperiod should not be considered as mutually exclusive factors controlling fish growth hormone levels. Evidence as to whether the actions of increased levels of GH, are mediated via IGF-I in parallel with elevations observed in warmer months remains equivocal. A parallel increase in GH and IGF-I was reported for amago salmon (Moriyama et al. 1997), whereas in chinook salmon (Pierce et al. 2001), coho salmon (Duan et al. 1995) and gilthead sea bream (Pérez-Sánchez et al. 1994a, b) peak GH levels preceded peak IGF-I levels by approximately 1–2 months. Interestingly, chinook salmon fish lengths were positively correlated with plasma IGF-I in late fall when GH was at its lowest level (Pierce et al. 2001) again indicating disparity between temperature, GH levels and growth and suggesting that the growth effects are really due to IGF-I with GH possibly serving as an early trigger for growth. The findings from the above studies, relating seasonal temperature with GH, lead us to ask, (i), are parallel increases in GH and warm water temperature representative of a stimulated growth function or simply a stress related response due to elevated temperature? and (ii), is the unparallel and delayed IGF-I peak, observed in cooler months, more critical for fish growth than GH? Given that only a few studies have been performed on a limited number of fish species then we have a while to wait before we can answer these questions.

Table 2 Effects of temperature on growth hormone (GH) levels in fish
Full size table

Laboratory temperature acclimation experiments provide for useful information on the effect of temperature, per se, on fish GH levels and to date few studies using this approach have been reported (see Table 2). Tilapia maintained at 26°C had higher plasma GH than those kept at 20°C (Ricordel et al. 1995) and rainbow trout maintained at 16°C also had higher plasma GH levels than groups kept at 4 or 8°C (Gabillard et al. 2003b). However elevated plasma GH levels in rainbow trout may have occurred as a consequence of mechanisms not associated with pituitary GH transcription/translation since pituitary GH transcripts and content remained unchanged with increasing temperature. The positive correlation between increased temperature and increased plasma GH, which has been observed in some fish species, may not manifest into enhanced fish growth but instead may be indicative of stress. For example, Atlantic salmon maintained at 18.9°C had highest amounts of plasma GH but growth increased with increasing temperature from 4.6–14.4°C and decreased between 14.4–18.9°C (Handeland et al. 2000). Whilst most studies, concerning temperature effects on GH, in fish, suggest, warmer water temperature is paralleled by increased GH, very little is known regarding the opposite situation whereby colder temperature enhances GH status. An abrupt transfer of gilthead sea bream from 18 to 9°C demonstrated a rapid decrease of plasma GH (Rotllant et al. 2000a) whereas chronic acclimation of silver sea bream to 12°C resulted in significantly higher amounts of pituitary GH transcript and content, in comparison to sea bream maintained at 25°C (Deane and Woo 2006a). From preliminary evidence obtained thus far, it does appear that colder temperature acclimation, of silver sea bream, is stimulatory towards growth as increased hepatic IGF-I mRNA and plasma levels of thyroid hormones were significantly elevated during 12°C acclimation (Deane and Woo 2005d). From several years of study, on silver and other species of sea bream, we believe that ~32°C may be almost the maximum temperature that these fish species can tolerate (Woo and Fung 1980) and whilst they are found in the warmer waters around Southern China they are also found in much colder waters around Japan. It is possible therefore that colder temperature is stimulatory for growth in this family (Sparidae), as the important enzyme necessary for the generation of reducing power in growth and synthesis, glucose-6-phosphate dehydrogenase is stimulated by GH and cold temperature in sea bream (Deane and Woo 2005c). In fact, various indices for growth enhancement such as elevated protein anabolism (Woo 1990) and higher glucose-6-phosphate dehydrogenase activity (Woo 1990; Deane and Woo 2005c) were apparent in sea bream acclimated to cold temperature.

Recently the importance of temperature in the regulation of GHRs in fish has started to emerge. For rainbow trout it was found that the number of copies of GHR1 and GHR2 were generally higher in embryos reared at 8°C in comparison to those reared at 6°C (Li et al. 2006). During embryonic development it was demonstrated that higher temperature of incubation (12°C vs 4°C) increased the amount of GHR1 transcript up to hatching whereas amounts of GHR2 transcript remained relatively unchanged (Gabillard et al. 2006). In juvenile rainbow trout fed ad libitum, higher temperature of incubation (16°C vs 8°C) resulted in higher levels of GHR1 and GHR2 transcript in liver and GHR1 transcript in muscle, however food restriction counteracted the effect of temperature on GHR expression in both tissues (Gabillard et al. 2006). The interaction of temperature and food restiction was also recently highlighted using late stage embryos/early stage juveniles of rainbow trout incubated at 8.5°C and 6°C (Raine et al. 2007). In this study incubation temperature per se did not alter the expression of GHR1 or GHR2, whereas food restriction caused a significant elevation in GHR1 expression only, at both temperatures.

Effects of pollutants (xenoestrogens and heavy metals) on GH

A large number of endocrine disrupting chemicals are generally grouped as xenoestrogens, that are man-made synthetic chemicals capable of mimicking the action of the natural female sex hormone 17-β estradiol (E2) (Yadetie et al. 1999). To date, most of the studies, whether direct or indirect, concerning xenoestrogen effects on fish GH have used 4-nonylphenol (4-NP) that can bind to E2 receptors and induce vitellogenin synthesis (Christiansen et al. 1998). The occurrence of 4-NP in aquatic ecosystems is via breakdown of nonylphenol polyethoxylates that are commonly found in discharges from plastics, paper and textile industries (Liber et al. 1999) and therefore the environmental impact of this xenoestrogen is of major concern. Injections of 4-NP inhibited the progress of Atlantic salmon smoltification with a concomitant reduction of Na+–K+-ATPase activity, gill chloride cell density and lowered hypoosmoregulatory ability (Madsen et al. 1997). Similarly, Atlantic salmon yolk sac larvae exposed to waterborne 4-NP had reduced Na+–K+-ATPase activity and seawater tolerance during smolt development (Lerner et al. 2007a). The effect of 4-NP treatment was similar to that found with E2 treatment and appeared to be reversible since Na+–K+-ATPase activity was not significantly different to that found with untreated controls, 6–7 weeks after cessation of 4-NP treatment (Madsen et al. 2004). Atlantic salmon smolts that were exposed to low, but environmentally relevant, concentrations of 4-NP did not display weakened hypoosmoregulatory performance (Moore et al. 2003). However, exposure to 4-NP plus the pesticide, atrazine did reduce gill Na+–K+-ATPase activity and impair hypoosmoregulatory ability indicating that combinations of environmental disrupting chemicals in aquatic ecosystems would be more potent. As GH is important in regulating hypoosmoregulatory performance in Atlantic salmon (McCormick 1996) then it may well be the case that 4-NP impairment of processes involved in hypoosmoregulation such as increased Na+–K+-ATPase activity and chloride cell density could have been mediated via disruption of GH regulation but the aforementioned studies on Atlantic salmon do not provide for direct evidence as to the importance of xenoestrogens on fish GH regulation since measurements of plasma or pituitary GH were not reported. To date, only a few studies have provided measurements of GH upon exposure to 4-NP, in fish. Growth hormone transcript from pituitaries of Atlantic salmon remained unchanged (Yadetie and Male 2002) whereas plasma IGF-I levels declined in Atlantic salmon at final stages of parr-smolt transformation in parallel with poor growth (Arsenault et al. 2004). Similarly 4-NP did not alter plasma GH but suppressed plasma IGF-I in Atlantic salmon (McCormick et al. 2005; Lerner et al. 2007b). It may well be the case that the actions of xenoestrogens such as 4-NP are associated with disruption of mechanisms involving the GH receptor as tentative evidence from a recent study on black sea bream has shown that hepatic expression of GH receptors were decreased after treatment with estradiol (Jiao et al. 2006). Given that effects of estradiol and 4-NP are similar in fish and that IGF-I production is regulated upstream via GH binding to its receptor then it is possible that the mechanism of 4-NP effects in fish are associated with disruption in hepatic GH receptor function. The actions of xenoestrogens may also be associated with pituitary GH transcription since the identification of estrogen response elements in the promoter region of the rainbow trout GH gene (Yang et al. 1997) and the modulatory effects of the xenoestrogens o,p’dichlorodiphenyltrichloroethane and tetrachlorodibenzo-p-dioxin on GH mRNA in vitro (Elango et al. 2006) suggest that xenoestrogens could directly affect GH transcription. Alongside 4-NP, recent consideration has also been given to the impact of polychlorinated biphenyls that are manufactured in mixtures called Aroclors. In Atlantic salmon, Aroclor was found to interefere with smolt physiology and behavior but it did not specifically affect plasma GH levels (Lerner et al. 2007b) although oral administration of Aroclor did cause a reduction in plasma GH levels in Arctic charr (Jorgensen et al. 2004).

Heavy metal contamination of aquatic ecosystems represents another major threat to fish growth (Friedmann et al. 1996; Marr et al. 1996; Vosyliene et al. 2003) and development (Friedmann et al. 1996; Chang et al. 1997). In terms of effects on fish endocrine regulation, recent studies have provided evidence that heavy metal exposure can alter cortisol secretion (Hontela et al. 1996; Hontela 1997; Lacroix and Hontela 2004; Gagnon et al. 2006) and interfere with hormone signaling cellular pathways (Marchi et al. 2004, 2005) but much less is known regarding effects of heavy metals on GH in fish. Immunocytochemical and histological studies using Nile tilapia taken from an unpolluted and polluted (containing high levels of zinc, lead and cadmium) lake, showed that those in polluted water had much smaller somatotrophs and weaker GH immunoreactivity than those in unpolluted water (Mousa and Mousa 1999). A similar immunohistochemical study also showed that pituitaries of white sucker, maintained in cages in a lake contaminated with iron-ore mine tailings, had significantly less GH than those maintained in an uncontaminated lake (Goverdina et al. 2005). The environmentally important heavy metal, cadmium, has been shown to cause pituitary necrosis in vertebrates (Hinckle et al. 1987) and reduced abundance of GH transcript in larval rainbow trout following hatching (Jones et al. 2001). In addition, we have recently been able to demonstrate that acute exposure of silver sea bream to cadmium causes a significant reduction in pituitary GH content as determined from immunoblotting (Woo and Man, unpublished data), implying that a possible mode of action of cadmium is related to GH transcription/translation. Although we are presently limited by experimental studies it can be tentatively concluded, from the information available, thus far, that some heavy metals can interfere with pituitary GH production. As to whether heavy metals interfere directly or indirectly with pituitary GH production remains to be elucidated. It could be the case that heavy metal mediated GH suppression is caused simply by pituitary necrosis or it may be possible that several other endocrine pathways, including estrogenic regulation, (Gueval et al. 2000; Jones et al. 2001) are disrupted which in turn could have a negative feedback on GH regulation.

Impact of aquaculture related stress on GH

Fish are widely cultured, many under intensive rearing conditions that would commonly impose stress on fish. Stressors that are most likely to affect fish, under culture conditions, include handling, confinement and crowding which in turn could predispose fish to pathogenic stress. In this section of the review we assess the importance and role of handling/confinement, crowding, disease and nutritional stress on GH levels in fish.

Handling/confinement stress

Rainbow trout that were subjected to an acute (1 h) or chronic (24 h) confinement stress following handling, had decreased amounts of plasma GH (Pickering et al. 1991) whereas reduced plasma GH levels were also reported for Atlantic salmon subjected to 1–24 h confinement (Wilkinson et al. 2006). Similarly Nile tilapia, that were confined for 1 h, had decreased plasma GH but this returned to near control levels within two and a half hours after removal of stress (Auperin et al. 1997). The dynamics of plasma GH, to a handling/confinement stress, was further studied in gilthead sea bream where it was found that plasma GH declined within one hour of stress but returned to near control levels after 4 h (Rotllant et al. 2001). In the above studies, decline in plasma GH levels were always correlated with an increase in plasma cortisol and studies on mammals have shown that glucocorticoids inhibit growth and GH secretion predominantly through inhibition via somatostatin (Giustina and Wehrenberg 1992). It is possible that a similar mechanism may exist in fish, as it has been recently demonstrated that somatostatin significantly decreased GH binding in trout liver (Very and Sheridan 2007).

Crowding stress

Increasing the stocking density of fish, over an optimal level, can result in significant stress through overcrowding. Sexually immature gilthead sea bream that were maintained under crowded (30 kg m−3) conditions for 23 days had lower plasma GH levels than fish maintained in uncrowded (7 kg m−3) conditions (Rotllant et al. 2000b). However plasma GH levels in rainbow trout that were maintained in either crowded (100 g l−1) or uncrowded (25 g l−1) conditions increased in parallel with suppressed growth rates (Pickering et al. 1991). Since the elevations in GH were seen in both uncrowded and crowded conditions Pickering et al. (1991) suggested that changes in water quality, in particular water oxygen content, could be a factor to consider since GH levels increased when oxygen levels were low but returned to basal levels when additional aeration was supplied. In mammals, GH levels have been shown to increase in line with an increase in oxygen requirement and a reduction of oxygen availability (Van Helder et al. 1987) and interestingly, repeated acute stress of Atlantic salmon parr caused an increase in plasma GH (McCormick et al. 1998). As a situation of repeated stress could place a greater demand for oxygen it may be the case that elevated GH, in response to reduced oxygen availability, could be a common feature between mammals and fish. Whilst it is too early to draw any firm conclusions as to the physiological importance of elevated GH during crowding stress it is possible that elevations in GH is a short term protective response against limited oxygen availability during crowding stress. In addition, the role and importance of “crowding factors” that are chemical agents (pheromones) known to adversely affect growth of fish (Francis et al. 1974; Pfuderer et al. 1974; Solomon 1977) should not be overlooked as it is plausible that such factors could also act to regulate fish GH, during crowding.

Pathogenic disease

Aquacultural practices can result in fish being kept under suboptimal conditions compounded by exposure to an array of stressors including hypoxia, adverse water quality, changes in salinity, handling and overcrowding (Snieszko 1974; Hedrick 1998; Deane et al. 2001). Whilst these stressors have been addressed in this review as separate entities they are not mutually exclusive and combinations of these can pre-dispose fish to secondary stressors such as infectious diseases (Walters and Plumb 1980; Robertson et al. 1987). However, disease can also occur, in fish, even without a pre-disposing factor (Hjeltnes and Roberts 1993). Bacterial pathogens such as Vibrio spp., Aeromonas spp., Yersinia spp., Renibacterium spp., and Streptococcus spp. are known causative agents of fish disease and together account for substantial economic loss in fish stocks, worldwide, on a yearly basis. Recent evidence has shown that conditions of disease (vibriosis) can disrupt osmoregulation (Deane and Woo 2005e), cytoprotection (Ackerman and Iwama, 2001; Deane et al. 2004; Deane and Woo 2005e) and immune function (Deane et al. 2001; Li et al. 2003) but much less is known regarding the effect of disease on GH levels in fish. Using an experimental set up to mimic the natural progression of vibriosis, in silver sea bream, it was found that pituitary GH transcript and content rapidly declined from an early onset of disease (Deane and Woo 2006a) an effect that was correlated with reduction in hepatic IGF-I transcript (Deane and Woo 2005f). These findings would imply that a suppression of growth would occur during disease but more importantly the role of GH in immune function and protection against disease may also be adversely affected. Administration of GH was found to enhance the survival of rainbow trout against Vibrio anguillarum and stimulate the non-specific immune response as defined by increased phagocytic activity of macrophages (Sakai et al. 1997). In addition, GH administration to silver sea bream also enhanced macrophage phagocytic activity, as determined by in vitro assays (Narnaware et al. 1997). Whilst it does appear, from the few studies performed thus far, that GH plays a key role as an immunostimulant, overdosage or overproduction may act to induce immunosuppression in some fish. For example, coho salmon that were transgenic for GH had a reduced resistance to V. anguillarum when they were near smolting in comparison to non-transgenic fish (Jhingan et al. 2003).

Nutritional stress

Hormones play a key role during nutrient uptake in fish and GH levels could be altered by food availability, time of feeding and diet composition (MacKenzie et al. 1998). In terms of food availability several studies have provided insight into the effects of fasting (or food deprivation) on GH levels, in fish, and during such periods plasma GH levels increased in tilapia (Weber and Grau 1999), rainbow trout (Sumpter et al. 1991; Takahashi et al. 1991; Farbridge and Leatherland 1992a, b; Holloway et al. 1994; Johnsson et al. 1996), striped bass (Small et al. 2002) and coho salmon (Varnavsky et al. 1995). Although elevated amounts of GH are commonly found in fasted fish it is unlikely that this translates into enhanced growth as hepatic GH receptor numbers have been shown to decrease during food deprivation (Gray et al. 1992; Pérez-Sánchez et al. 1994b) resulting in a reduced tissue sensitivity to GH and a subsequent decline in IGF-I production. The reduced amounts of IGF-I would decrease the negative feedback on GH production leading to increased plasma GH amounts during nutritional stress (Duan and Plisetskaya 1993). In mammals, elevated GH (via nutritional stress), is primarily involved in protein metabolism as well as promoting energy expenditure via increased lipolysis and it has been suggested that GH performs similar functions during fasting in fish (Harmon and Sheridan 1992; Björnsson 1997). Several studies have provided data that does not conform to the pattern described above as plasma GH levels remained relatively unchanged following 21 days fasting in channel catfish (Small 2005) and 8 weeks of fasting in rainbow trout (Pottinger et al. 2003) whereas decreased plasma GH was reported for rainbow trout fed once weekly in comparison to those fed regularly (Farbridge et al. 1992). Under culture conditions, feeding time is also controlled and in salmonids it has been shown that plasma GH levels follow a diurnal pattern encompassing a nocturnal acrophase and a peak post prandial phase (Reddy and Leatherland 1994; Gélineau et al. 1996). Based on the information available, evidence as to whether feeding time has an influence on GH levels in fish is equivocal. Rainbow trout fed in the morning exhibited post prandial GH peaks within 3–6 h following feeding whereas those fed later in the day did not display such a peak in GH levels (Reddy and Leatherland 1994) and rainbow trout fed during the night had higher mean plasma GH amounts when compared to those fed in the morning (Gélineau et al. 1996). However, in rainbow trout, no significant differences in plasma GH levels were observed during feeding in the morning or late afternoon (Gomez et al. 1996) and recently a similar finding was also reported for rabbit fish whereby the daily expression pattern of pituitary GH mRNA remained unchanged regardless of whether food was given in the morning or in the afternoon (Ayson et al. 2007). A final consideration involves the effect of diet composition on fish GH levels. Significantly increased GH amounts were detected in gilthead sea bream fed low protein diets however this was unlikely to be translated into enhanced growth since reduced hepatic GH binding and IGF-1 amounts were also found in gilthead sea bream fed low protein diets (Pérez-Sánchez et al. 1995). In addition a diet high in lipid composition has been shown to increase plasma GH levels for rainbow trout (Holloway and Leatherland 1998; Holloway et al. 1999), Arctic Charr (Cameron et al. 2002) and gilthead sea bream (Martí-Palanca et al. 1996; Company et al. 1999; Pérez-Sánchez 2000). Two explanations currently exist for elevated plasma GH levels when fish are fed diets high in lipid content. Firstly, elevated GH may be a compensatory protein sparing response (Marti-Palanca et al. 1996; Company et al. 1999; Pérez-Sánchez 2000) or secondly it may be indicative of increased lipolytic function of GH (Björnsson 1997).

GH and apoptosis

Apoptosis, or programmed cell death, occurs via a sequence of molecular events that commonly involve activation of cellular aspartate-specific-cysteine protease (caspase) enzyme cascades, DNA fragmentation and chromatin condensation (Oberhammer et al. 1993; Ramachandra and Studzinski 1994; McConkey 1998). The induction of apoptosis can occur in cells that are exposed to stress and recently chemical (Deane and Woo 2005g; DeWitte-Orr and Bols 2005; Liu et al. 2006; Arteeq et al. 2006) and heavy metal (Krumschnabel et al. 2005) exposure were shown to induce apoptosis in fish cells. A growing body of evidence, from studies on mammalian cells, suggests that GH plays a key role in protecting against apoptosis and therefore should be considered as an anti-apoptotic agent. For example, GH treatment reduced the number of apoptotic neuronal cells in rat brains subjected to hypoxic–ischemic stress (Shin et al. 2004), inhibited apoptosis in bovine embryos (Koelle et al. 2002) and endogenous overproduction of GH prevented the occurrence of apoptosis in EL4 lymphoma cells (Robyn and Weigent 2004). Whilst GH can exert its anti-apoptotic effect via IGF-I stimulation (Eisenhauer et al. 1995) it has also been shown that it can act independently of IGF-I through separate signaling pathways (Baixeras et al. 2001). The evidence as to whether GH can also act as an anti-apoptotic agent in fish is equivocal but limited to only two studies thus far. It was reported that in vitro exposure of trout leucocytes to GH did not alter the frequency of apoptotic cells (Yada et al. 2004) whereas exposure of sea bream whole blood preparation to GH protected against chemical induced apoptosis (Deane and Woo 2005g). It is presently unknown as to how GH exerted an antiapoptotic effect in sea bream blood but tentative evidence suggests that this could have occurred via induction of heat shock protein 70 (Deane and Woo 2005g) a cytoprotective protein that has been reported to prevent the onset of apoptosis in mammalian cells (Samali and Orrenius 1998; Beere and Green, 2001). Heat shock proteins are known to be important in the stress response of fish (Iwama et al. 1998; Basu et al. 2002a) and modulated expression by hormones has been reported in fish either in the presence (Basu et al. 2001, 2002b) or absence (Deane et al 1999b, 2000) of stress. As it was found that GH exposure resulted in elevated heat shock protein 70 expression with a concomitant anti-apoptotic action (Deane and Woo 2005g) then it remains a possibility that protective effects of GH could be mediated via heat shock protein 70 regulation.

Present status and future perspectives

We have seen in this review that GH expression can be modulated upon exposure of fish to an array of stressors. In terms of salinity challenge two different situations are now evident (i) an increase in GH during seawater exposure and (ii) an increase in GH during isoosmotic salinity acclimation. However, we are still limited by the number of fish species used for experimental research as the seawater adapting effects of GH have been reported in salmonids, tilapia and killifish whereas isoosmotic stimulated GH has only been reported for a few sparids. Clearly, studies focused on many different fish need to be undertaken in order to extend our knowledge on the role and importance of GH during fish salinity challenge. In addition, studies aimed at understanding how the GH receptor expression is modulated in fish tissues upon salinity challenge as well as the regulatory role of GH on ion transport and water regulation need to be undertaken. In terms of temperature effects and from all studies performed, on a seasonal basis, it has been shown that GH levels peak during warmer months of the year. However, we still need to clearly establish whether such increases are representative of enhanced growth or whether they represent a stress response due to warmer temperature. Environmental pollution of aquatic ecosystems poses a major threat to fish, but thus far a paucity of data exists as to the effects of pollutants such as xenoestrogens and heavy metals on fish GH. Whilst it is apparent that xenoestrogens can disrupt GH mediated processes we need to ascertain whether such effects are mediated via GH receptor binding and/or estrogen responsive elements located on the GH gene promoter. Finally, recent evidence has shown that GH can also function as an anti-apoptotic agent when fish cells are exposed to stress and future studies aimed at investigating the cellular mechanisms related to this cytoprotective role are certainly warranted.