CRF and stress in fish

General and Comparative Endocrinology 146 (2006) 36–44 www.elsevier.com/locate/ygcen Review CRF and stress in fish Gert Flik *, Peter H.M. Klaren, Erwin H. Van den Burg, Juriaan R. Metz, Mark O. Huising Department of Animal Physiology, Institute for Neuroscience, Toernooiveld 1, 6525 ED Nijmegen, The Netherlands Received 21 July 2005; revised 8 November 2005; accepted 14 November 2005 Available online 5 January 2006 Abstract The endocrine stress response is pivotal in vertebrate physiology. The stress hormone cortisol—the end product of the endocrine stress axis—(re-)directs energy flows for optimal performance under conditions where homeostasis may be or become at risk. Key players in the continuous adaptation process are corticotropin-releasing factor (CRF) from the hypothalamic nucleus preopticus (NPO), pituitary adrenocorticotropic hormone (ACTH) and cortisol produced by the interrenal cells in the headkidney (adrenal equivalent of fish). CRF is a member of a large family of related peptides that signals through CRF-receptor subtypes specific for central and peripheral actions of the peptide. CRF is ‘‘chaperoned’’ by a unique and phylogenetically very well-conserved binding protein (CRFBP); the functions of the CRFBP can only be speculated on so far, but its mRNA and protein abundance are important indicators of the central CRFsystem activity, and indeed its mRNA levels are altered by restraint stress. Moreover, the unique structure and size of the CRFBP provide good tools in phylogenetic studies, that date the CRF-system to at least one billion years old. Pro-opiomelanocortin is produced and processed to ACTH and endorphin in the hypothalamic NPO and pituitary pars distalis ACTH-cells, to MSH and acetylated endorphins in the pituitary pars intermedia MSH-cells. ACTH is the prime corticotrope in acute stress conditions. In carp, MSH, considered a mild corticotrope in chronic stress responses in other fish, lacks corticotropic effects (in line with the absence of the melanocortin-5 receptor in headkidney); yet, an unknown corticotropic signal susbtance in the pars intermedia of carp awaits elucidation. Interesting observations were made on the CRF control of pituitary cells. CRF stimulates ACTH-cells, but only when these cells experience a mild dopaminergic block. Endorphin, produced in the NPO and transported via axons to the pituitary gland in vivo, reverses the stimulatory CRF action on MSH-cells to a differential inhibition of N-acetyl b-endorphin release in vitro (MSH release is not affected). We speculate that the consistently observed elevation of plasma MSH during chronic stress may exert central actions related to feeding and leptin regulated processes. A BOLD-fMRI study revealed the functional anatomy of the stress response at work in a paradigm, where carp were exposed to a sudden water temperature drop. In carp (and other fish), the endocrine stress axis is already operational in very early life stages, viz., around hatching and comprises hypothalamic, pituitary, and interrenal signaling to adjust the physiology of the hatchling to its dynamically changing environment. Understanding of stress during early life stages is critical as the consequent rises in cortisol may have long lasting effects on survival and fish quality. Ó 2005 Elsevier Inc. All rights reserved. Keywords: CRF; CRFBP; CRFR1; CRFR2; POMC; ACTH; MSH; BOLD-fMRI; Nucleus preopticus; Pituitary gland; Headkidney; Common carp; Cyprinus carpio; Larvae; Aquaculture 1. Introduction For a dynamic interaction with the environment, crucial for survival, vertebrates continuously adjust their physiology (‘‘adaptational responses’’) to an ever-changing environment. To do so, vertebrates may depend on a seemingly endless repertoire of physiological, endocrinological, and * Corresponding author. Fax: +31 24 3653229. E-mail address: [email protected] (G. Flik). 0016-6480/$ - see front matter Ó 2005 Elsevier Inc. All rights reserved. doi:10.1016/j.ygcen.2005.11.005 immunological responses that allow them, indeed, to cope with physical, chemical, and biological disturbances. Fish, in intimate contact with their aqueous environment via the elaborate and delicate epithelium of the gills—where unwanted chemicals and antigens may easily penetrate— were the first vertebrates that developed stress responses, which include interactions between the endocrine stress axis (a key player in the adaptive response with gills as an important target for cortisol, the end product of the stress axis) and the immune system (crucial to eliminate G. Flik et al. / General and Comparative Endocrinology 146 (2006) 36–44 intruding antigens). Bidirectional communication between the endocrine and immune system then is a prerequisite (Engelsma et al., 2002). Indeed, evidence is accruing from molecular biological studies that in fish, representing the earliest vertebrates, the majority of chemical signaling between these systems is already operational. In this topical review, we address the endocrine stress response in carp (Cyprinus carpio). We will focus on the functional anatomy of the stress axis in adult fish, and peculiarities of the CRFand POMC-systems in the stress axis. In adult carp, we were able to visualise the endocrine stress axis at work, in vivo, by blood oxygen level dependent functional magnetic resonance imaging (BOLD-fMRI). The early development of the stress axis activity was investigated in carp around hatching. This mini-review reflects our contribution to the symposium ‘‘Novel functions of the corticotropin-releasing factor system’’ held at the 15th International Congress of Comparative Endocrinology in Boston, May 2005; it is certainly not an exhaustive review, rather a document to stimulate further reading and discussion. 2. Some functional anatomy of the endocrine stress axis in adult fish The focus in this section is on the hypothalamic nucleus preopticus (NPO), pituitary ACTH- and MSH-cells and the headkidney, where cortisol is produced. The hypothalamic nuclei involved in stress responses receive multiple input from sensory systems that continuously monitor and evaluate the internal and external conditions of the fish. Eventually, much of this information converges to the NPO, where it is integrated to direct the pituitary ACTH- and MSH-cells in their control over the interrenal cells and central systems that need adjustment during stress. 3. The CRF-system CRF-producing cells in the NPO are believed to play a crucial role in the process of adaptation to stressors. Importantly, other peptides like arginine vasotocin (AVT), thyrotropin-releasing hormone (TRH) and the POMC-derived ACTH and endorphins are only some other players in this field. The peptides may be produced in the same CRF-producing cells or in seperate cells and intercellular communication with these signals may occur via synaptic (or synaps-like) contacts or be specified via localised receptor expression on the neuron somata. With an antibody to carp ACTH10–23, which recognises the processed ACTH-molecule (not its precursor POMC or the MSH-sequence), we demonstrated (Metz et al., 2004) the presence of ACTH in NPO-cells in the vicinity of CRFcells. This ACTH could function in paracrine mode to modulate CRF-cell activity within the NPO, since we see no ACTH in axon bundles projecting to the pituitary gland. Paracrine release of ACTH in the NPO could be demonstrated by immunohistochemistry at the ultrstruc- 37 tural level on tannic acid treated tissue (to capture exocytosis), but such studies have not yet been carried out. We do not yet know the afferent input to these ACTH-producing cells. A more speculative yet attractive hypothesis is that these ACTH-producing cells are primarily an endorphin source. Processing of POMC to ACTH yields endorphin and, indeed, we do see endorphin-positive axons approaching and entering the pituitary gland. We do not rule out that (part of) the endorphin is N-terminally acetylated during transport towards the pituitary gland (Van den Burg et al., 2001). Acetylation of endorphin results in loss of opioid activity as the acetylated isoforms lose significant affinity for opioid receptors (Barg et al., 1993; Rene et al., 1998) and this may be required to fine-regulate endorphin levels as these peptides are rather powerful biochemicals. A remarkable observation is that endorphin, when given as a tonus in vitro to carp pituitary MSH-cells, alters the response to CRF: CRF inhibits under such conditions the release of acetylated endorphin but not of MSH. Apparently, a differential regulation of processing and release of these peptides, that reside in a 1:1 ratio in the precursor POMC, is part of the complex pituitary response to hypothalamic stress signals (Van den Burg et al., 2005a). Acetylated endorphin has been assigned a potentiating role in corticotropic actions of MSH in the Mozambique tilapia (Balm et al., 1995). Every trial to confirm this concept in carp failed so far and we could not demonstrate mRNA of melanocortin-5 receptor, a predicted vector for MSHsignaling in headkidney (Metz et al., 2005). Little, if any, information is available on specific (activating) receptors for acetylated endorphins. Maybe acetylation of endorphins aims at production of opioid receptor antagonists (Rene et al., 1998). In that case, the search for opioid receptors on fish (tilapia) cortisol-producing cells is indicated. The consequence of such a discovery would be that the potentiating effect of acetylated endorphin on MSH-corticotropic actions relies on antagonist action on opioid receptors. Such a complicated scenario is realistic as it was demonstrated in the mammalian adrenal cortex that the zona fasciculata cells express opioid l- and j-receptors that stimulate corticosteroid production, via ACTH-independent pathways (Kapas et al., 1995). The search for specific, activating receptors for acetylated-endorphin awaits further research and experimentation; that such receptors may exist follows from the consideration that highly specific antisera to acetylated endorphins can be made. This suggests that the acetylated molecule is highly specifically recognised by the immune system and thus is part of a known ligand-receptor combination. Axons of the NPO CRF-cells project directly to the pituitary gland. This is a different situation from that in higher vertebrates (as of amphibia) where the hypothalamic-releasing hormones are freed into the portal circulation of an eminentia mediana. The important difference between fish and higher vertebrates then is that the hypothalamicreleasing hormones depend on receptor expression profiles of targets in higher vertebrates, whereas localised release 38 G. Flik et al. / General and Comparative Endocrinology 146 (2006) 36–44 (via synapses or into the circulation via neurosecretion in the pars nervosa) is typical for fish. Specificity is thus realised in the CRF-sending neurons in fish and in the receiving cells in higher vertebrates, respectively. The CRF signal is eventually determined by at least three factors (Fig. 1): the CRF structure, the presence of CRF binding protein (CRFBP), and the receptor for CRF (at least two known, R1 is the ‘‘central’’ receptor). Carp are tetraploid fish and make two CRF molecules that differ in a proline (neutral) and alanine (amphiphatic) on position four of the mature 41 amino acid peptide (Huising et al., 2004). As this is not a conservative replacement, one wonders whether the peptides have evolved into serving different functions. For sure, the alanine would give the molecule less rigidity that could easily translate into a different tertiary structure. Little is known on this topic but it has been shown that messenger RNA’s for the two very similar copies of POMC in this fish (Arends et al., 1998a) are differentially expressed among different strains of common carp when challenged with varying water temperatures (Arends et al., 1998b). This indicates that even very small differences in peptide molecules potentially yield advantages in adaptation physiology and amplify the (stress) signaling repertoire of the fish. The same could hold for the two CRFs in carp. The high sequence identity between fish and human mature CRF (93%) substantiates a strong evolutionary pressure on conservation of this signal substance over the roughly 450 million years since the rise of fishes in the Ordovicium. The small, yet significant, changes within duplicate genes as occur in the tetraploid carp confirm this notion from observations on a single species; the genome duplication in carp is however more recent and estimated to have occurred 16 million years ago (Larhammer and Risinger, 1994). Another exciting hypothesis to test is whether different CRFs are produced in peripheral organs of the fish (headkidney, immune system, caudal neurosecretory system, gills) with a concomittant target specificity. CRF is projected towards ACTH-cells and towards the pars intermedia MSH-cells. Interestingly, CRF may be a releasing hormone for TSH in fish (as in birds via CRFR2 rather than CRFR1; De Groef et al., 2003) and this warrants studies on thyroid involvement in stress regulation. Intriguingly, CRF-signaling to the pituitary gland could exert combined and concerted control over metabolism and metamorphosis via cortisol and thyroid hormones during early life stages and metabolism alone in adult stages. This notion further warrants studies into interactions of transcription factors for thyroid hormone (T3) and cortisol (and others such as retinoic acid receptors known to heterodimerise for their actions), and what wonderful models fish such as carp provide to study the ontogeny of these aspects of the CRF-system! We are presently screening the carp pituitary gland for CRFR1 and CRFR2 mRNAs to see if the chicken situation was/ maybe an ‘‘invention’’ of fish. The abundant CRF-immunoreactivity in carp (and other fish) pars intermedia (Huising et al., 2004) may indicate that NPO axons terminate on pars nervosa blood vessels and release CRF there for peripheral purposes. This scenario could explain the extraordinarily high levels of CRF seen in fish plasma (Pepels and Balm, 2004a). Whether plasma CRF in fish is ‘‘chaperoned’’ by a binding protein is doubted by these authors, but this seems counterintuitive considering the situation in pregnant women: the high levels of CRF at term are buffered by CRFBP to protect the body from the potent CRF signal. Interestingly, the urophysis of fish is another potential source of CRF that could determine the high plasma levels in fish (Lu et al., 2004); whether the urophysis also contains CRFBP awaits further experimentation. Carp, pufferfish, and in all Fig. 1. The vertebrate CRF-system. Five members of the vertebrate CRF family have been identified, of which sauvagine has only been identified in a single frog species (Phylomedusa sauvagei). Note that urotensin-I is the fish ortholog of urocortin 1, found in tetrapods. All CRF family members signal via CRF-receptors that belong to the superfamily of seven-helix G-protein-coupled receptors. Note that CRFR3 has to date only been identified in a single catfish species (Ameiurus nebulosus). The bioavailability of some CRF family members is modulated by CRF-BP. Recently, splice variants of CRFR1 and CRFR2 genes encoding truncate and soluble CRFRs have been described in human and rodents, which represents a novel way to modulate CRFsignaling. G. Flik et al. / General and Comparative Endocrinology 146 (2006) 36–44 likelihood all fish, produce CRFBP (Huising et al., 2004), which may modulate CRF bioactivity or serve as a reservoir. The impetus for this research was given by data obtained with an antiserum to human CRFBP (donated by Dr. W. Vale, Salk Institute, La Jolla, CA) that allowed us to demonstrate a 35.8 kDa CRFBP in carp hypothalamus and pituitary extracts by Western blot (Huising et al., 2004). As will be discussed below, CRFBP is wellconserved during phylogeny and this explains the cross-reactivity at the basis of these observations. In carp, CRFBP is produced by a separate group of small cells at the periphery of the NPO. The axons of these cells join those of the CRF-producing neurons on their way towards the pituitary gland. This leaves the CRF- and CRFBP-producing cells as two targets for regulation in fish (in higher vertebrates, CRF and CRFBP are often produced by the same cell). Restraint stress elevates mRNA levels of both CRF and CRFBP in carp hypothalamus and results in release of both proteins from nerve terminals in the pituitary gland as shown by a comparison following immunohistochemistry of control and experimental tissue and thus we have to consider both proteins in our appreciation of the CRF-signaling system. When parallels would exist with the IGF-system, we should at least extend our views on CRFBP functions to direct, CRF-independent effects of the binding protein on CRF targets (Bauchat et al., 2001). The rather well-studied IGF-system may fulfil a role model in this search. The IGF- and CRF-systems with their binding proteins share a long evolutionary history: both have been traced back in fishes (Bauchat et al., 2001; Huising et al., 2004). Very recently, CRFBP was demonstrated in the honeybee (Huising and Flik, 2005) and this would date the CRF-system to at least one billion years old. A third player in the CRF-system is of course the receptor moiety on the target cell for CRF. To date, two categories of CRF-receptors (with splice variants; Eckart et al., 2002) are known, CRFR1 and CRFR2. Both are widely expressed in the central nervous system and peripherally. CRFR1 is the important receptor in the stress axis. CRFR2 is involved in a whole host of stress-related physiological and behavioral responses, from alterations in vascular tone and blood pressure to effects on feeding and anxiety-related behaviors; obviously CRFR2-mediated regulation of learning and anxiety control involves brain nuclei and areas outside the endocrine stress axis (Bale et al., 2002a,b). The involvement of the CRFR1 in the stress response of carp was demonstrated by a predicted down-regulation of the messenger RNA for this protein following a 24 h restraint (Huising et al., 2004). Fish (carp) possess an orthologue of the mammalian CRFR2 and its particular involvement in fish physiology and disease/stress awaits further experimentation. With the demonstration of the ‘‘complete CRF system’’ in fish, we hope to convince students in biology that we should look, where possible, for original functions of the components of this system in the early vertebrates (fish) and should realise ourselves that an anthropocentric approach in this particular field of 39 research (extrapolation of data from human studies) may not always be justified, considering the high degree of specialisation of the human being and carp alike, and may leave surprises hidden. The versatility of the CRF-system may be at the basis of the success of fishes that, with an estimated 35,000 extant species, are the most successful vertebrates in this world, and that have explored and occupied essentially every thinkable niche. We only start to understand bits of the highly complex brain/hypothalamus regulation of pituitary output, which involves many more signals than those of the CRF-family (CRF, Urocortins, Urotensins), thyrotropin-releasing hormone (TRH), arginine-vasotocin (AVT), and proopiomelanocortin(POMC-)derived peptides from the NPO which were addressed here; not to speak of similarly complex pituitary cocktails of pleiotropic signals meant to control the peripheral targets of the body for a properly concerted stress response. This complexity must be at the basis of the success of fishes and evolution of later vertebrates. 4. The POMC-system 4.1. ACTH (Fig. 2) The expression of POMC in the NPO was addressed above ‘‘in the margin’’. Our knowledge of ACTH actions in hypothalamic cells is rather limited, not in the least because of the scarcity of sensitive assays for ACTH in fishes (good antibodies are rare due to the extreme conservation of the ACTH molecule in most vertebrates). We succeeded in developing an ACTH-immunoassay for carp ACTH (Metz et al., 2004) that we used mainly to better understand the carp pituitary ACTH-cell. Carp ACTHcells appear to be under negative dopaminergic control in vivo: ACTH-release is unleashed in ectopic pars distalis tissue. This notion was key to further study these cells in vitro. First attempts to measure CRF-regulated output from the pituitary pars distalis (which can easily be separated from the rest of the highly organised pituitary gland in fishes) directly in a perifusion setup enigmatically failed. The release rate of ACTH increased over time and CRF over a wide range of concentrations was without stimulatory effect. We then reasoned that the inhibitory dopamine tonus that these cells normally experience in the intact fish, with its consequences for second messenger make-up of the cells, could be a requirement for CRF action in vitro. Indeed, ACTH-cells do show CRF-dependency for ACTH-release only under mild dopamine block. It seems that the powerful ACTH signal is normally inhibited by dopamine for a balanced stress-axis output. Thus, proper ACTH-cell functioning in vitro is apparently only guaranteed when the cells experience a certain dopamine tonus. Further, it was the peculiar and generally recognised power of the ACTH signal, e.g., reflected by picomolar surges during stress, that underlies the notion that an inhibitory tonus by dopamine is an integral part of stress-axis activity to allow for balanced and strictly controlled stress axis out- 40 G. Flik et al. / General and Comparative Endocrinology 146 (2006) 36–44 Fig. 2. Serial, transversal sections of the pituitary gland of common carp immunostained with antisera raised against the C-terminus of carp b-endorphin (A) and ACTH (B). Obviously, b-endorphin-positive fibers project in the direction of the PI. These fibers most likely originate from the NPO, which is positive for both b-endorphin and ACTH (NPO not visible in these sections). ACTH, however, is not immunoreactive in the pituitary-entering nerve fiber. Thus, although derived from the same precursor, this suggests differential trafficking of the peptides b-endorphin and ACTH. As shown, both POMCderived peptides are also produced in the corticotrope cells of the PD. In the PI, ACTH is immediately further cleaved to a-MSH and CLIP, and thus leaves no ACTH immunoreactivity there. Note that the b-endorphin immunoreactivity shown represents the non-acetylated full-length b-endorphin, which we established immunohistochemically with the use of a N-acetyl b-endorphin-specific antiserum (not shown in this figure). For clarity, nomenclature is shown only in (A). Abbreviations: V, ventricle; NF, nerve fiber; PD, pars distalis; PI, pars intermedia. put and stress adaptation. More studies are needed. ‘‘Do both signals use cAMP?’’, ‘‘are other second messengers involved?’’, ‘‘is there cross-talk or bidirectional communication between their second messengers?’’, these are just some questions that have to be answered to understand the mechanisms underlying this observation. Interestingly, hypothalamic dopamine production decreases in mice experiencing chronic stress (Konstandi et al., 2000). If the same holds for fish, a lower dopamine tonus due to stress could desensitise the ACTH-cell for CRF, impeding acute stress signals. In chronically stressed (not in controls) Mozambique tilapia (Lamers et al., 1997), the pituitary MSH-cells express a stimulatory, high affinity dopamine D1 receptor, compatible with this notion of decreased dopamine levels and an increased MSH production during chronic stress. Apparently, such situations of decreased dopamine activity do occur in fish. 4.2. MSH and N-acetylated endorphin (Fig. 2) As mentioned above, the pituitary gland secretes many post-translationally modified POMC-derived peptides. MSHs and acetylated endorphins are two categories of products of the MSH-cell often considered in stress regulation. In a series of studies on tilapia in the nineties of the last century, MSH was proposed as a mild corticotrope (in concert with endorphin; Balm et al., 1995). The complex picture around the rather pleiotropic MSH-signal was narrowed down to its di-acetylated form as the specific corticotrope next to the non- and mono-acetylated isoforms (Lamers et al., 1992) and thyrotropin-releasing hormone (not CRF) the prime hypothalamic regulator of its differential release (Lamers et al., 1994). The picture that emerged was that of potent pars distalis ACTH in acute stress appearing in short surges to evoke sharp transient rises in cortisol and mild pars intermedia MSH chronically elevated to reset baseline cortisol to somewhat higher levels. A wonderful picture, but maybe a tilapia picture only. We could not confirm a similar concept in carp. Whatever combination of synthetic MSH- and endorphin-isoforms tested, no corticotropic action for MSH could be demonstrated in carp (Metz et al., 2005). Accordingly, no other than the MC2R, the ‘‘ACTH-receptor’’, could be demonstrated in carp headkidney tissue and this receptor was downregulated following stress in line with an anticipated negative feedback control. We anticipated that an effect of MSH would be mediated by an MC5R as this receptor is also expressed in mammalian adrenal cortex (Maia et al., 2002; Metz et al., 2005; MC1R is the melanophore MSH-receptor, MC3R and MC4R are predominantly expressed in brain tissue). The intriguing observation remains that an aqueous extract of pars intermedia (conservatively dissected and thus mainly containing products of MSH- and somatolactin-cells plus CRF and CRFBP in the pars nervosa tissue therein) contains a corticotropic principle at least equipotent to ACTH (Metz et al., 2005). We are investigating the possibilities now. Plasma MSH-levels do rise in chronically stressed carp (Metz et al., 2005) as in tilapia (Lamers et al., 1994), but maybe with another purpose? We consider the possibility of pituitary MSH feedback on central mechanisms to regulate food intake during stress (Cerda-Reverter et al., 2003). We have cloned carp leptin and assuming a rather pregnant role for leptin in control over neuronal activity involved in feeding behaviour, we now can evaluate anorexigenic (MSH, CART, Stat3, Socs3, CRF) and orexigenic (NPY, AgRF) signals to these neurons. Another function of MSH that requires more attention is its peripheral lipolytic activity in mammals (Brennan et al., 2003; Forbes et al., 2001) as well as in fish (Yada et al., 2002). Stress conditions may require an additional energy source beyond glycogen/glucose. The consistent rises in plasma MSH seen during chronic stress in fish would be in line with this hypothesis. G. Flik et al. / General and Comparative Endocrinology 146 (2006) 36–44 4.3. The headkidney 41 5. The endocrine stress axis in action viewed by BOLDfMRI The interrenal tissue of fish represents a peculiar anatomical situation with the cells producing cortisol and those that produce the catecholamines adrenalin and noradrenalin being intermingled and arranged along the cardinal veins in the headkidney of the fish; the endocrine cells lie embedded in haematopoietic tissue. Such organisation suggests that the endproducts of the stress axes (catecholamines and cortisol) have direct paracrine access to the cells of the immune system. Vice versa, chemical signals of the immune system could exert direct paracrine actions on the steroid and chromaffin cells. An example of such a concerted signal response was demonstrated by immersion vaccination of carp: cortisol surges are combined with upregulation of acute phase genes of the crucial innate immune components (interleukin-1b, iNOS, a2-macroglobulin, serum amyloid proteinA and tumor necrosis factor a), all meant to fight the imminent threat of pathogen invasion (Huising et al., 2003). Just with this concept in mind, this headkidney tissue deserves far more attention than it receives right now. We wonder about all the possible feedbacks that the combined headkidney signals may exert at every level of the stress axis. No doubt exists that what we call the endocrine stress axis, i.e., the NPO-pituitary gland-interrenal cell-axis, is involved in the regulation of stress responses and this notion is based on extensive analyses of the physiology, biochemistry, and molecular biology that show more or less predicted responses in stress paradigms. Yet, to the best of our knowledge, there is no report on the visualisation of stress axis activity in vivo, and most certainly not in fish. In a wonderful collaboration with Dr. A. Van der Linden and colleagues, we developed a paradigm to study the stress axis activity in a high resolution MRI setup (Fig. 3; Van den Burg et al., 2005b). For this purpose, of course, we had to find out whether the (mild) anaesthesia required to immobilise the fish for imaging did not interfere with the initiation of the endocrine stress response. We thus set out to test several anaesthetics and succeeded to immobilise carp by irrigation of the gills with water containing low maintenance concentrations of MS222, while keeping the stress response intact. The stress response was monitored i.a. by changes in plasma cortisol levels. The actual stressor chosen was a sudden A B C D poa th pd cortisol 0 1 2 3 4 5 20 30 40 Time after onset of temperature drop (min) Fig. 3. Blood oxygen level dependent functional magnetic resonance imaging (BOLD-fMRI) in carp brain preoptic area (poa) and anterior midbrain tegmentum (mt), the tuberal hypothalamus (th) and in the pituitary gland following a 10 °C temperature drop in the water irrigating the gills. (A) BOLDfMRI shows decreased (blue, mt) and increased (red, poa) BOLD-contrast, showing differential effects of the temperature drop on cellular activity; the area where the NPO CRH-cells reside (poa) is activated. (B) BOLD-fMRI of the tuberal hypothalamus; the area where a multitude of neural pathways between NPO and pituitary runs is activated. (C) BOLD-fMRI of the pituitary gland; opposite responses in the pars distalis (pd; red/pink) and pars intermedia (pi; blue). Oxygen consumption increases in the pd (consistent with activation of ACTH-cells) and decreases in the pi following an acute temperature drop. (D) Cartoon depicting the time-sequential BOLD-fMRI events preceding the release of cortisol following an acute temperature drop (abbreviations as above). The observations are in line with the predicted sequence of events during activation of the endocrine HPI-axis. 42 G. Flik et al. / General and Comparative Endocrinology 146 (2006) 36–44 temperature drop (25 to 15 °C within 5 min), realised by switching between two water tanks from where the gills were irrigated. The fish was equipped with an antenna on top of its head and images were collected in the horizontal plane. This required the construction of a dedicated brain map for carp, which was made under supervision of our colleague Dr. Meek, renowned comparative brain anatomist (Meek and Nieuwenhuys, 1998). The ‘‘horizontal approach’’ was a requirement to avoid anticipated disturbances of the water irrigating the gills below the brain (the plane of gill adherence to the palatum is sufficiently distant from pituitary gland and brain basis). Just one other aspect addressed was the potential effect of noise that goes with MRI-analysis on the responses of the fish. To this end, a sound CD was recorded of the actual setup in Antwerp and played with a gettho blaster at realistic decibels in the Nijmegen laboratory, where a fake MRI setup (but with the temperature switch) was constructed to sample carp for cortisol under conditions with and without noise. Fish certainly can perceive noises (have inner ears as higher vertebrates), and responded with a slight elevation of cortisol levels to handling and noise. However, the lightly anaesthetised carp responded with a very significant cortisol reponse within 5 min to the temperature drop. Two major analyses were then carried out. The first concerned the distribution of blood by the use of injected iron particles as contrast during the temperature drop. A wealth of data was obtained, but the most remarkable observation was a massive drainage of blood from the brain to peripheral sinuses. We have interpreted these data as follows. Considering that the carp is a poikilotherm, the fish may benefit from a temporary insulation of the brain with the still warm blood. It should be noted that the brain is very close to the branchial circulation that can be considered as a heat exchanger system: any change in water temperature will change blood temperature quickly and this change is then reaching the brain first before the cooled blood reaches the body of the fish. In a poikilotherm, the biochemistry and thus the functioning of the brain is strictly dependent on the temperature of the ambient medium. The blood redistribution response to drastically changing ambient water temperature occurs as of 2 min following the onset of the temperature drop, which is 1.5 min after escape and stress responses are initiated (see below). In contrast to the blood volume reduction in the brain, the pars distalis of the pituitary gland received more blood, which not only accomodates increased activity of the ACTH-producing cells, but also facilitates the transport of ACTH to the interrenal tissue to stimulate production and release of cortisol. To find out which areas of the brain were involved in the initiation of this endocrine response, a blood oxygen level dependent functional magnetic resonance imaging (BOLD-fMRI) study was done. The principle of this technique is that haemoglobulin with and without oxygen bound behaves magnetically differently and thus the oxy- genation degree of the haemoglobulin in fact determines the constrast properties of the blood. Activated cells consume more oxygen, so that the local concentration of deoxygenated haemoglobulin increases and the signal intensity in the BOLD-fMRI images changes. It appeared that some areas were inactivated during or following the temperature drop, whereas others were clearly activated (which may sound counterintuitive in a poikilotherm that is cooled down). Remarkable observations include activation of the NPO area (where the CRF-neurons reside), the tuberal hypothalamus (halfway NPO and pituitary gland) and the pituitary pars distalis (the pars intermedia is inactivated in this acute stress paradigm). More importantly, the activation followed a time-sequential pattern in line with activation of the endocrine stress axis: first the NPO (within 30 s), then the tuberal hypothalamus, followed by the pituitary pars distalis and then the interrenal cortisol release (as of 4–5 min). We concluded from these studies (Van den Burg et al., 2005b) that we succeeded to visualise the endocrine stress response in action! Presently, we are working on other of the plethora of data then obtained that concern areas of the brain, where sensory input from the buccal cavity is processed and relayed to the cerebellum for sensory processing and sensorimotor integration. The rationale behind this thought is that the fish would normally try to escape or avoid such cooler water. 6. Stress in very young fish, the onset of the endocrine stress axis Understanding of the neuroendocrine regulation of the stress response of an organism is key to understand its general physiology and well being, as the signals involved in the stress response are powerful and influential. Cortisol (or corticosterone in some vertebrates) the key glucocorticosteroid stress hormone that is released following a cascade of signals from the brain down to the adrenal (or interrenal) cells, is an all-determining factor in steering the energy flows in the organism for optimal performance to cope with stressors; functions such as growth, reproduction, and immune regulation are at risk when stressors persist and evoke too strong and long-lasting elevations of cortisol. Importantly, stress events (through actions of cortisol) at a given point in time may have effects on later development and life. For that very reason it is crucial to know when and how very young fish give their first stress responses (Flik and Wendelaar Bonga, 2001) and how we can avoid such responses or how they become exaggerated in laboratory and aquaculture settings. In our work on carp larvae (Flik et al., 2002) we have observed that embryonic stages, still protected by the egg membrane, are able to produce cortisol when the embryo is disturbed by manipulation of the egg with forceps. We cannot exclude in this case that some direct pressure action of the squeezing procedure startled the cells that produce cortisol (although we very carefully squeezed to elicit three tail beats only), yet we favor a regulated (via CRF and ACTH) basis as also G. Flik et al. / General and Comparative Endocrinology 146 (2006) 36–44 ACTH is produced very early during development of carp. Transient rises in ACTH in the prehatching stage indicate that the synthetic machinery is operational and up-regulated at such early stages. For cortisol, this holds as well as cortisol is never stored but always produced according to needs. Admittedly, the yolk also harbors maternal sources of ACTH and cortisol, but then in the developing egg only a decline in levels could result from consumption, an increase can only be explained by embryonic synthesis. Interestingly, Pepels and colleagues (2004) working on Mozambique tilapia larvae quantified CRF immunoreactivity in the heads of larvae as young as five days post hatching, and a more sensitive immunohistochemistry extrapolated these results to even younger stages (2 days post hatch). An attractive hypothesis to test is that the fish is equipped with an endocrine stress axis some time before but certainly upon hatching. The consequences of such observations are obvious. We have to handle fertilised eggs and fish larvae with care. Experimentally we have shown that carp exposed to copper ions in the water respond differentially to levels of 0.3 and 0.8 mM. Both levels induce elevations in cortisol after hatching (the egg membrane is impermeable to copper ions), but the milder rise at the lower level of copper is enhancing growth and not affecting mortality, while the higher level results in a twenty percent higher cortisol content of the larvae, yet this results in retarded growth and significant increases in mortality. Apparently, the lower concentration of copper evoked a cortisol response within the bandwidth of the carp, the higher concentration of copper presents an allostatic overload (McEwen and Wingfield, 2003). Understanding and establishing this bandwidth of the stress response will be a powerful tool in predicting the success of aquaculture practices. References Arends, R.J., Vermeer, H., Martens, G.J.M., Leunissen, J.A.M., Wendelaar Bonga, S.E., Flik, G., 1998a. Cloning and expression of two proopiomelanocortin mRNAs in the common carp (Cyprinus carpio L.). Mol. Cell. Endocrinol. 143, 23–31. Arends, R.J., Van der Gaag, R., Martens, G.J.M., Wendelaar Bonga, S.E., Flik, G., 1998b. Differential expression of two proopiomelanocortin mRNAs during tempreature stress in common carp (Cyprinus carpio L.). J. Endocrinol. 159, 85–91. Bale, T.L., Lee, K.-F., Vale, W.W., 2002a. The role of corticotropinreleasing factor receptors in stress and anxiety. Integr. Comparat. Biol. 42, 552–555. Bale, T.L., Picetti, R., Contarino, A., Koob, G.F., Vale, W.W., Lee, K.F., 2002b. Mice deficient for both corticotropin-releasing factor receptor 1 (CRFR1) and CRFR2 have an impaired stress response and display sexually dichotomous anxiety-like behavior. J. Neurosci. 22, 193–199. Balm, P.H.M., Hovens, M.L.M., Wendelaar Bonga, S.E., 1995. Endorphin and MSH in concert form the corticotropic principle released by tilapia (Oreochromis mossambicus; Teleostei) melanotropes. Peptides 16, 463–469. Barg, J., Belcheva, M., McHale, R., Levy, R., Vogel, Z., Coscia, C.J., 1993. Beta-endorphin is a potent inhibitor of thymidine incorporation into DNA via mu-, and kappa-opioid receptors in fetal rat brain cell aggregates in culture. J. Neurochem. 60, 765–767. 43 Bauchat, J.R., Busby Jr., W.H., Garmong, A., Swanson, P., Moore, J., Lin, M., Duan, C., 2001. Biochemical and functional analysis of a conserved insulin-like growth factor binding protein isolated from rainbow trout hepatoma cells. J. Endocrinol. 170, 619–628. Brennan, M.B., Costa, J.L., Forbes, S., Reed, P., Bui, S., Hochgeschwender, U., 2003. Alpha-melanocyte-stimulating hormone is a peripheral, integrative regulator of glucose and fat metabolism. Ann. NY Acad. Sci. 994, 282–287. Cerda-Reverter, J.M., Schioth, H.B., Peter, R.E., 2003. The central melanocortin system regulates food intake in goldfish. Regul. Pept. 115, 101–113. De Groef, B., Goris, N., Arckens, L., Kuhn, E.R., Darras, V.M., 2003. Corticotropin-releasing hormone (CRH)-induced thyrotropin release is directly mediated through CRH receptor type 2 on Thyrotropes. Endocrinology 144, 5537–5544. Eckart, K., Jahn, O., Radulovic, J., Stiedl, O., Brauns, O., Tezval, H., Zeyda, T., Spiess, J., 2002. Pharmacology and biology of corticotropin-releasing factor (CRF) receptors. Receptors Channels 8, 163–177. Engelsma, M.Y., Huising, M.O., van Muiswinkel, W.B., Flik, G., Kwang, J., Savelkoul, H.F.J., Verburg-van Kemenade, B.M.L., 2002. Neuroendocrine-immune interactions in fish: a role for interleukin-1. Vet. Immunol. Immunopathol. 87, 467–479. Flik, G., Wendelaar Bonga, S.E., 2001. Stress in very young and adult fish. Life and Environment/Vie et Milieu 51, 229–236. Flik, G., Stouthart, X.J.H., Spanings, F.A.T., Lock, R.A.C., Fenwick, J.C., Wendelaar Bonga, S.E., 2002. Stress response to waterborne Cu during early life stages of carp, Cyprinus carpio. Aquat. Toxicol. 56, 167–176. Forbes, S., Bui, S., Robinson, B.R., Hochgeschwender, U., Brennan, M.B., 2001. Integrated control of appetite and fat metabolism by the leptin-proopiomelanocortin pathway. Proc. Natl. Acad. Sci. 98, 4233–4237. Huising, M.O., Guichelaar, T., Hoek, C., Verburg-van Kemenade, B.M.L., Flik, G., Savelkoul, H.F.J., Rombout, J.H.W.M., 2003. Increased efficacy of immersion vaccination in fish with hyperosmotic pre-treatment. Vaccine 21, 4178–4193. Huising, M.O., Metz, J.R., van Schooten, C., Taverne-Thiele, A.J., Hermsen, T., Verburg-van Kemenade, B.M.L., Flik, G., 2004. Structural characterisation of a cyprinid (Cyprinus carpio L.) CRH, CRH-BP and CRH-R1, and the role of these proteins in the acute stress response. J. Mol. Endocrinol. 32, 627–648. Huising, M.O., Flik, G., 2005. The remarkable conservation of corticotropin-releasing hormone-binding protein (CRH-BP) in the honeybee (Apis mellifera) dates the CRH system to a common ancestor of insects and vertebrates. Endocrinology 146, 2165–2170. Kapas, S., Purbrick, A., Hinson, J.P., 1995. Action of opioid peptides on the rat adrenal cortex: stimulation of steroid secretion through a specific mu opioid receptor. J. Endocrinol. 144, 503–510. Konstandi, M., Johnson, E., Lang, M.A., Malamas, M., Marselos, M., 2000. Noradrenaline, dopamine, serotonin: different effects of psychological stress on brain biogenic amines in mice and rats. Pharmacol Res. 41, 341–346. Lamers, A.E., Flik, G., Atsma, W., Wendelaar Bonga, S.E., 1992. A role for di-acetyl alpha-melanocyte-stimulating hormone in the control of cortisol release in the teleost Oreochromis mossambicus. J. Endocrinol. 135, 285–292. Lamers, A.E., Flik, G., Wendeleaar Bonga, S.E., 1994. A specific role for TRH in release of diacetyl alpha-MSH in tilapia stressed by acid water. Am.J. Physiol. 267, 1302–1308. Lamers, A.E., Ter Brugge, P.J., Flik, G., Wendelaar Bonga, S.E., 1997. Acid stress induces a D1-like dopamine receptor in pituitary MSH cells of Oreochromis mossambicus. Am. J. Physiol. 273, R387–R392. Larhammer, D., Risinger, C., 1994. Molecular genetic aspects of tetraploidy in the common carp Cyprinus carpio. Mol. Phylogenet. Evol. 3, 59–68. Lu, W., Dow, L., Gumusgoz, S., Brierley, M.J., Warne, J.M., McCrohan, C.R., Balment, R.J., Riccardi, D., 2004. Coexpression of corticotropin-releasing hormone and Urotensin I precursor genes in the caudal 44 G. Flik et al. / General and Comparative Endocrinology 146 (2006) 36–44 neurosecretory system of the Euryhaline Flounder (Platichthys flesus): a possible shared role in peripheral regulation. Endocrinology 145, 5786–5797. Maia, M., Vivier, J., Hue, D., Durand, P., Feige, J.J., Defaye, G., 2002. Expression of the melanocortin receptors MC2-R (ACTH-receptor) and MC5-R during embryonic development of ovine adrenals. Endocr. Res. 28, 631–635. McEwen, B.S., Wingfield, J.C., 2003. The concept of allostasis in biology and biomedicine. Horm. Behav. 43, 2–15. Meek, J., Nieuwenhuys, R., 1998. Holosteans and teleosts. In: Nieuwenhuys, R., Ten Donkelaar, H.J., Nicholson, C. (Eds.), The Central Nervous System of Vertebrates, vol. 2. Springer-Verlag, Berlin, pp. 759–937. Metz, J.R., Huising, M.O., Meek, J., Taverne-Thiele, A.J., Wendelaar Bonga, S.E., Flik, G., 2004. Localisation, expression, and control of adrenocorticotropic hormones in the nucleus preopticus and pituitary gland of common carp (Cyprinus carpio L.). J. Endocrinol. 182, 23–31. Metz, J.R., Geven, E.J.W., Van den Burg, E.H., Flik, G., 2005. ACTH, alpha-MSH and control of cortisol release: cloning, sequencing and functional expression of the melanocortin-2 and melanocortin-5 receptor in Cyprinus carpio. Am. J. Physiol. 289, R814–R826. Pepels, P.P.L.M., Balm, P.H.M., 2004a. Ontogeny of t corticotropinreleasing factor and of hypothalamic-pituitary-interrenal axis responsiveness to stress in tilapia (Oreochromis mossambicus; Teleostei). Gen. Comp. Endocrinol. 139, 251–265. View publication stats Pepels, P.P.L.M., Van Helmvoort, H.A.C., Meek, J., Wendelaar Bonga, S.E., Balm, P.H.M., 2004b. Corticotropin-releasing hormone (CRH) content in the teleost stress response: rapid appearance of the peptide in plasma of tilapia (Oreochromis mossambicus; Teleostei). J. Endocrinol. 180, 425–438. Rene, F., Muller, A., Jover, E., Kieffer, B., Koch, B., Loeffler, J.-P., 1998. Melanocortin receptors and delta-opioid receptor mediate opposite signaling actions of POMC-derived peptides in CATH.a cells. Eur. J. Neurosci. 10, 1885–1890. Van den Burg, E.H., Metz, J.R., Arends, J.R., Devreese, B., Vandenberghe, I., Van Beeumen, J., Wendelaar Bonga, S.E., Flik, G., 2001. Identification of beta-endorphins in the pituitary gland and blood plasma of the common carp (Cyprinus carpio L). J. Endocrinol. 169, 271–280. Van den Burg, E.H., Metz, J.R., Spanings, F.A.T., Wendelaar Bonga, S.E., Flik, G., 2005a. Plasma alpha-MSH and Acetylated betaendorphin levels following stress vary according to CRH sensitivity of the pituitary melanotropes in common carp, Cyprinus carpio. Gen. Comp. Endocrinol. 140, 210–221. Van den Burg, E.H., Peeters, R.R., Verhoye, M., Meek, J., Flik, G., Van der Linden, A., 2005b. Brain responses to ambient temperature fluctuations in fish: reduction of blood volume and initiation of a whole-body stress response. J. Neurophysiol. 93, 2849–2855. Yada, T., Moriyama, S., Suzuki, Y., Azuma, T., Takahashi, A., Hirose, S., Naito, N., 2002. Relationships between obesity and metabolic hormones in the ‘‘cobalt’’ variant of rainbow trout. Gen. Comp. Endocrinol. 128, 36–43.