This article appeared in a journal published by Elsevier. The attached copy is furnished to the author for internal non-commercial research and education use, including for instruction at the authors institution and sharing with colleagues. Other uses, including reproduction and distribution, or selling or licensing copies, or posting to personal, institutional or third party websites are prohibited. In most cases authors are permitted to post their version of the article (e.g. in Word or Tex form) to their personal website or institutional repository. Authors requiring further information regarding Elsevier’s archiving and manuscript policies are encouraged to visit: http://www.elsevier.com/copyright

Author's personal copy Aquaculture 296 (2009) 337–346

Contents lists available at ScienceDirect

Aquaculture j o u r n a l h o m e p a g e : w w w. e l s ev i e r. c o m / l o c a t e / a q u a - o n l i n e

Sustained exercise improves vertebral histomorphometry and modulates hormonal levels in rainbow trout Marie-Hélène Deschamps a,⁎, Laurent Labbé b, Sylvie Baloche c, Martine Fouchereau-Péron d, Sylvie Dufour c, Jean-Yves Sire a a

Université Pierre & Marie Curie-Paris 6, UMR 7138, Évolution et développement du squelette, 7 quai St-Bernard, Case 5, 75 252 Paris Cedex 5, France Pisciculture Expérimentale INRA des Monts d'Arrée, Barrage du Drennec, 29450 Sizun, France Muséum National d'Histoire Naturelle, Biologie des Organismes et Ecosystèmes Aquatiques, UMR CNRS 7208, 75231 Paris Cedex 5, France d Station de Biologie Marine, BP 225, 29182 Concarneau cedex, France b c

a r t i c l e

i n f o

Article history: Received 30 May 2008 Received in revised form 2 July 2009 Accepted 20 July 2009 Keywords: Swimming Vertebrae Mineralization Skeletal abnormalities Thyroid hormones Calcitonin Onchorhynchus mykiss

a b s t r a c t Abnormal compressions and fusions of vertebral bodies are frequently observed in reared rainbow trout and could result from chronic and unbearable muscle pressures acting on the axial skeleton during intensive growth. Sustained swimming at moderate speeds was shown to induce many positive effects on growth and swimming performances in salmonids, but yet little is known about its effects on vertebral remodeling processes and related hormonal regulation. Rainbow trout were subjected to three different swimming speeds (0, 1.0 and 1.5 body length (BL) s− 1), starting one month after they were first fed (65 mm) and ending when they reached 260 mm in size (market-size of 275 g). At the end of the experiment, 20 trout were sampled in each lot (N = 60) and blood samples were taken. Vertebrae abnormalities were assessed by radiological examinations. Vertebrae from the middle axial region (V32–38) were selected to evaluate bone mineralization (BM) and total bone area (Tt-B.Ar.) on radiographed transverse sections (125 ± 10 μm). Assays were performed to evaluate mineral homeostasis (calcemia and phosphatemia), bone cell activities (alkaline phosphatase, ALP, and tartrate-resistant acid phosphatase, TRAP) and bone regulating hormones (calcitonin, CT and thyroid hormones, THs). Sustained exercise reduced the appearance of fused vertebrae, enhanced vertebral BM and decreased vertebral Tt-B.Ar., while it increased circulating CT and TH levels. No variations were observed on mineral homeostasis and bone cell activities. Increasing the swimming speed up to 1.5 BL s− 1 had positive effects on the vertebral skeleton, and therefore, seems to be a suitable approach to prevent aggravation of vertebral abnormalities in juvenile trout. The changes observed in vertebral features are interpreted as a compromise between the necessity to mobilize vertebral mineral ions in response to various physiological demands and to maintain vertebral strength against mechanical constraints. © 2009 Elsevier B.V. All rights reserved.

1. Introduction In fish farming, skeletal abnormalities are a major concern as they result in a decrease of production yield (by affecting survival and morphology), an increase in sorting costs, and impede filet production (Divanach et al., 1997; Boglione et al., 2001; Koumoundouros et al., 1997, 2001; Witten et al., 2006). In reared rainbow trout, a recent study revealed that 22% of normally-shaped individuals showed discreet abnormalities in their vertebrae (Deschamps et al., 2008). This percentage is far above the 2–3% observed in wild salmonids (Gill and Fisk, 1966; Poynton, 1987). Vertebral abnormalities are generally related to inappropriate rearing conditions during early development stages (for reviews see Cahu et al. (2003) and Lall and Lewis-McCrea (2007)). However, vertebra defaults can either appear, be aggravated or contained during ⁎ Corresponding author. Tel./fax: +33 1 44 27 35 72. E-mail address: [email protected] (M.-H. Deschamps). 0044-8486/$ – see front matter © 2009 Elsevier B.V. All rights reserved. doi:10.1016/j.aquaculture.2009.07.016

further growth (Witten et al., 2006). Determining the factors that could play a role in vertebral anomalies is not easy, as bone metabolism is still poorly known in teleosts. In particular, data characterizing specific changes of vertebral bone histology in relation to various rearing conditions are not available (Lall and Lewis-McCrea, 2007). There is thus a need for experimentations aiming to better understand the effects of various farming methods and environmental conditions on bone metabolism, their impact on the vertebra bone properties, and to try to link these events to the occurrence of skeletal abnormalities. Herein, we performed experiments to evaluate the effects of swimming on vertebral bone. Controlling the water flow/velocity in tanks induces changes in the swimming speed of trout. This is commonly used in fish farm to maintain water quality (O2, CO2, pH, ammonia, etc.). However, inappropriate water quality can also lead to vertebral abnormalities as reported in some species (Divanach et al., 1997; Kihara et al., 2002; Sfakianakis et al., 2004). In salmonids, a significant percentage of hyper-mineralized, compressed and/or fused vertebrae is suspected to be related to swimming and, notably, to impaired processes during spinal

Author's personal copy 338

M.-H. Deschamps et al. / Aquaculture 296 (2009) 337–346

development and growth (Kvellestad et al., 2000; Kacem et al., 2004; Wargelius et al., 2005a; Helland et al., 2006; Witten et al., 2006). In fact, swimming can have a positive or negative effect on vertebra growth and tissue composition, and hence on their mechanical properties. An inappropriate swimming speed during intensive growth could substantially increase red muscle development, and therefore results in strong mechanical forces (pressure) leading to vertebral tissue transformation (chondrogenesis), vertebral fractures and/or dysfunctions in the intervertebral junction (Witten et al., 2005, 2006). Moreover, an excessive exercise could weaken the axial skeleton in inducing mineral mobilization from the vertebrae in order to fulfill the demands of muscular activity or to buffer acid–base unbalance (Hermann-Erlee and Flik, 1989; Gil Martens et al., 2006; Helland et al., 2005). In contrast to these negative issues, it is generally admitted that sustained exercise has positive effects on salmonids, as for instance in improving growth rate, food conversion efficiency, blood oxygen-carrying capacity and muscle buffering ability (resulting in smaller decrease in pH) without affecting the composition of the carcass and/or of the viscera (for review, see Jobling et al. (1993) and Davison (1997)). Sustained swimming also stimulates fish metabolism (metabolic and anabolic states) by increasing growth hormone and thyroxine levels in blood serum while reducing circulating stress hormone (Davison, 1997). All these hormones are known to affect bone metabolism (Takagi et al., 1994; Wargelius et al., 2005b; Sbaihi et al., 2007). In addition to this, in fish and other vertebrates, an increase of strain, frequency and duration of pressure on bone through mechanical stimuli is known to improve bone strength (e.g., increase of mineral deposition and bone matrix synthesis) and to impede fracture (Turner, 1998; Kranenbarg et al., 2005; Porter et al., 2006). However, although numerous studies aimed to understand the effects of swimming on fish physiology and body composition, there are only a few reports on the consequences of sustained swimming on bone metabolism and vertebra properties (Kranenbarg et al., 2004, 2005). In the present paper we report the effects of different swimming speeds on (i) the occurrence of vertebral abnormalities, vertebral mineralization and vertebral total bone area, and (ii) the variation of circulating hormones known to be involved in bone metabolism (thyroid hormones and calcitonin) and of enzymes reflecting bone cell activities (alkaline phosphatase and tartrate-resistant acid phosphatase), in the blood of the specimens analyzed.

biomass) for each lot were calculated as the product of daily feeding rates [ln(Wfinal) − ln(Winitial) days)⁎ 100] and feed conversion ratios [FCR = amount of food distributed (Winitial ⁎Wgain)], where Wgain = [(Wfinal ⁎Efinal) − (Winitial ⁎Einitial) +W of deceased and sampled fishes]. The condition factor (K = 100,000W/L3) was also estimated for each group. 2.2. Experimental protocol One and a half month after first feeding (1st Mar. 2005), three groups of 200 trout [1.9 g and 6.5 cm total length (TL) in average] were placed in three sub-squared tanks (0.19 m3: 70 × 70 cm, depth: 40 cm). In such tanks there was no water turbulence and trout did not tend to take refuge in the corners of the tanks. Each tank was equipped with a horizontal inlet located at the surface, in order to induce current, and a central outlet (20 cm in diameter) surrounded by an oblong mesh (23 × 23 cm) at the bottom (Fig. 1). Water velocity was estimated by taking the mean value of the measurements obtained from three different positions in the tanks (at 4.0, 10.5 and 23.0 cm from the edge and at 20 cm from the bottom; Fig. 1B) using a small current meter (‘Ott’ type V C2 ‘10.150’ No. 46447). The tanks were supplied with spring water at a constant temperature of 11.0 ± 0.5 °C. Three months later (4th May 2005), trout were transferred outdoor in circular basins of 2 m3 (200 cm in diameter, depth: 65 cm). Each basin was equipped with a horizontal inlet at the surface, and a central outlet surrounded by a cylinder-shaped mesh (both of 55 cm in diameter) at the bottom. Water velocity measurements were then obtained at 10, 35 and 68 cm from the edge and 27.5 cm from the bottom of the basins. The basins were supplied with water from a river source subjected to seasonal temperature variations, that ranged from 11.3 to 18.5 °C during the experiment. Three water velocities (still-water, low and moderate exercise) were chosen, resulting in swimming speeds of 0.0, 1.0 and 1.5 body length per second (BL s− 1), respectively. In the tank with still-water (0.0 BL s− 1), the input current was established to ensure self-cleaning and water quality, while not inducing behavioral orientation of trout swimming (Fig. 1A). In tanks with slow and medium water velocity,

2. Materials and methods Fertilization, hatchery procedures and swimming experiments were carried out at the experimental fish farm of INRA (PEIMA, Sizun, France). All experiments were done in conformity to French animal welfare laws, guidelines and policies. 2.1. Fish Rainbow trout genitors (Oncorhynchus mykiss, Walbaum) were 5 females and 5 neo-males from INRA Autumnal strain (i.e., broodstock selected to spawn in autumn). Spawning was obtained in early December 2004 and the eggs incubated until ‘eyed’ stage. To ensure that all lots were replicates, the eggs were evenly dispatched into three groups of 1200 eggs, with the same incubation conditions [spring water in stacked racks with oxygen saturation close to 100%, controlled temperature (11.4 to 11.7 °C); constant pH 6.2-6.3]. No disease outbreak occurred between eyed stage and first feeding, Mortality during this period was estimated at 7.7 ± 1.4%. First feeding started on 17 Jan. 2005. Fry and young trout were fed to satiety with commercial diets (Bio Optimal Start, then ECOLIFE 66-67, Biomar). Food distribution was evaluated every week using ECU 2000 software (a modified version of ECUREUIL software; IFREMER, 1992) based on Muller Feuga growth model that takes into account water temperature, fish weight (W), fish effective (E) and specific rationing tables for the species. At the end of the experiments, feeding rates (FR, %

Fig. 1. Schematic drawings of the experimental tanks, showing the location of the measurement of water velocity. A. Still water (0.0 BL s− 1) induced no behavioral orientation of rainbow trout swimming. B. Slow and medium water velocity (1.0 and 1.5 BL s− 1) induced counter-current swimming.

Author's personal copy M.-H. Deschamps et al. / Aquaculture 296 (2009) 337–346

trout were homogenously distributed in the water column and always swam in a counter-current mode (Fig. 1B). Every three weeks, water velocity in each tank (i.e., the mean value of the three measurements) was updated, taking into account the mean total length of 50 sampled trout. As a consequence, the swimming speed increased from 6.4 ± 1.0 and 10.0 ± 1.5 cm s− 1 to 23.2 ± 3.5 and 32.2 ± 4.8 cm s− 1 for 1.0 and 1.5 BL s− 1 experiments, respectively. 2.3. Sampling procedure At the end of the experiment, i.e., when trout reached a mean weight of 275 g (26.0 cm in average; on 29 Aug. 2005, approximately 9 months after spawning), trout were fished from each tank and the externally malformed individuals were removed (i.e., identified with the naked eye, such as curvature of the body). Twenty trout were then randomly sampled (i.e., a total of 60 trout was studied). The fish were anaesthetized in 2-phenoxyethanol (0.9 ml l− 1) and promptly euthanized. Body weight and total length were measured. Then, blood was punctured (1 to 3 ml ind. − 1) at the level of the dorsal aorta using heparinized syringes, and plasma was separated by centrifugation (4400 rpm, 15 min at 4 °C) and stored, with their respective carcass, at −18 °C until processing.

339

2.4.2.3. Total vertebral bone area. The three middle adjacent vertebrae from the selected region (i.e., generally V34–36) were embedded in resin (98% stratyl, 2% Luperox catalysor) and sectioned into 125 ± 10 μm transverse sections using a Leitz 1600 Saw Microtome. The single section through the mid-region of the vertebrae, in which the notochordal canal is the narrowest and the bone tissue area the largest, was retained for the analysis of total bone area (Tt-B.Ar.). Sections were radiographed using a CGR Sigma 2060 generator, adjusted to 8 kV and 6 mA, on a Kodak Industrex film Ready Pack set at 30 cm of the source. The microradiographs (35×) were photographed with an Olympus Camedia digital camera mounted on an Olympus SZX12 binocular microscope. The pictures were enlarged and transformed into binary images with Adobe Photoshop 7.0 software using a threshold adjusted to take into consideration the bone in the entire section thickness. Vertebral mineralized area (Md.Ar) and total vertebral area (Tt.Ar.) were obtained from binary images using Bone Profiler 3.23, a software designed for bone tissue analysis (Girondot and Laurin, 2003), and vertebral total bone area (Tt-B.Ar., %) was calculated as the mineralized tissue area contained in the total area of a vertebra section: Md.Ar. Tt.Ar. × 100. On microradiographs, total bone area (Tt-B.Ar.) refers to the relative amount of whole mineralized bone (sensu Parfitt et al., 1987), i.e. the proportion of all mineralized tissues within a vertebra section (2D measurements).

2.4. Vertebra analyses 2.5. Blood sample assays 2.4.1. Typology of vertebral abnormalities The sampled trout were radiographed in a Faxitron@Cabinet X-ray system model 43 855C, adjusted to 90 kV and 160 ms. X-rays were scanned and digitized using Digital Linear X-ray Scanner EZ 320 set at 60 cm from the source and iX-Pect for EZ 320 X-ray acquisition software. Each radiograph was enlarged and the axial skeleton examined in detail for identification of vertebral abnormalities. Two categories of abnormalities were distinguished: 1 — vertebrae that were either asymmetrical, compressed, hypo- or hyper-mineralized were considered malformed vertebrae (weak abnormalities); 2 — vertebrae that were either juxtaposed, i.e. lacking intervertebral space but still exhibiting a distinguishable body and one pair of neural/hemal arches (incomplete vertebral fusion), or possessed a single vertebral body showing several pairs of neural/hemal arches (complete vertebral fusion), were considered fused vertebrae (strong abnormalities). 2.4.2. Vertebral bone condition parameters 2.4.2.1. Selection and preparation of vertebrae. Soft tissues were removed and seven adjacent vertebrae from the mid region of the axial skeleton (V32 to 38) were selected, as they possess the largest bone volume (Kacem et al., 2004). When vertebral abnormalities were assessed in the selected area on radiographs, the nearest adjacent vertebrae were taken instead (for instance, when V37–38 were shown to be abnormal, V30–V36 were chosen instead of V32–V38). However, only a few individuals were concerned and their numbers were similar in the three groups (i.e., 3, 4 and 3 individuals for 0.0, 1.0 and 1.5 BL s− 1, respectively). In this way, our analyses were performed on normal vertebrae. The vertebrae were dehydrated in a graded series of ethanol (70°, 90°, 100°; 24 h/bath), and lipids were removed in acetone (2 baths of 24 h), then in trichloroethylene (3 baths of 24 h). 2.4.2.2. Vertebral bone mineralization. The first 2 vertebrae and last 2 vertebrae from the selected region (i.e., generally V32–33 and V37– 38) were pooled, dried for 72 h at 37 °C, and weighed (Wdry) to the nearest mg. Then, they were incinerated for 8 h at 800 °C, ashes weighed (Wash) to the nearest mg, and vertebral bone mineralization (BM, %) calculated: (Wash / Wdry) × 100.

In order to work with sufficient serum volumes (2 ml) for all assays, plasma samples from the same experimental group were randomly paired leading to a total of 10 samples per experiment. 2.5.1. Calcium and phosphate determination Calcium and phosphate levels were measured in a spectrophometer following the manufacturer protocol (bioMerieux: Ca-kit 61041 and PHOS-UV 61571, respectively). Each sample was measured in duplicate for calcium and phosphate and expressed in mg l− 1. 2.5.2. Determination of alkaline phosphatase (ALP) and tartrate-resistant acid phosphatase (TRAP) activities ALP and TRAP activities were determined using colorimetric assays on microplates (ALP: Souvannavong et al., 1996; TRAP: Mukherjee et al., 2004). ALP: 100 μl of plasma sample were incubated at 37 °C for 1 h with p-nitrophenyl phosphate (pNPP: 1 mg/ml, Sigma N9389) in an alkaline buffer (Sigma A9226). Reaction was stopped by adding 50 μl of 3 N NaOH. Absorbance was measured at 405 nm against a blank and converted into the amount of produced p-nitrophenyl (pNP) using a standard dilution-curve (pNP: 10 mM, Sigma N7660). TRAP: 100 μl of plasma sample were incubated at 37 °C for 1 h in 50 mM of citrate buffer (pH 5.25), in the presence or absence of 10 μl of sodium tartrate (50 mM). Each sample was measured in duplicate and ALP or TRAP expressed in nmoles of pNP produced ml− 1 h− 1. 2.5.3. Radioimmunoassay (RIA) of calcitonin 2.5.3.1. Plasma extraction. Plasma was purified by absorption on Sep Pak C18 (Waters, Milford, MA) reverse phase sample cartridge (Fouchereau-Peron et al., 1990). Bulk molecules were first removed by washing with 0.05 M phosphate buffer supplemented with EDTA (10 mM) and BSA (0.3%) (pH 7.4). Then, peptides were eluted with acetic acid/acetone/distilled water (20:30:50 v/v). After acetone evaporation, the samples were freeze-dried. Peptides were finally diluted in 2 ml of phosphate buffer. 2.5.3.2. Calcitonin (CT) RIA. Extracts were assessed using RIA specific for CT (Fouchereau-Peron et al., 1990). This extraction method was previously tested in trout and the recovery determined for CT was

Author's personal copy 340

M.-H. Deschamps et al. / Aquaculture 296 (2009) 337–346

75%. Anti-salmon CT (sCT) antibodies were kindly donated by Dr. Julienne (U349, INSERM, Paris). Plasma extracts or standard were incubated with sCT antibodies at 30 °C for 18 h. Then, 125I radiolabeled sCT was added, and the mixture was incubated at 4 °C for 12 h. Free and bound hormones were separated by absorption onto dextrancharcoal. Standard curves were prepared using synthetic sCT (Sigma, ST Louis, Mo). Non-specific absorption was estimated by incubating plasma extracts in the absence of antibody. Radiolabeled hormone was measured using a gamma counter, and plasma extract CT levels were calculated using a logit-log transformed standard dilution-curve. Each extract was measured in duplicate and expressed in ng ml− 1. 2.5.4. Radioimmunoassays of thyroid hormones (T4, T3) Thyroxine (T4) and triiodothyronine (T3) levels in unextracted plasma samples were determined using commercial radioimminoassay kits (RIA-gnost®-T4 and RIA-gnost®-T3, Cisbio international). We are therefore measuring total plasma hormones. These kits are developed for unextracted human plasma samples. The assay can be applied on unextracted samples because reagents of the assay contain 8-anilino-1sulfonic acid, which releases protein-bound hormone. Briefly, unextracted plasma samples, or standard, and 125I radiolabeled T4 or T3 were added in tubes coated with anti- T4 or T3 antibodies. The tubes were incubated at 37 °C for 2 h. The free fraction was removed by decanting. The antibody-bound fraction was then measured using a gamma counter, and T4 and T3 concentrations were calculated from the log-transformed standard dilution-curve. For each plasma sample, T4 and T3 levels were measured in duplicate in both assays, and expressed in ng ml- 1. For each sample, total thyroid hormone (TH) was calculated in summing the value obtained for T4 and T3 contents for each extract. The T3:T4 ratio was also calculated. 2.6. Data analysis Bone mineralization (BM) and total bone area (Tt-B.Ar.) are percentages; therefore, prior to statistical analyses, an arcsine transformation was performed to ensure normality, using the formula: pffiffiffi p′ = arcsin p × 100 (Zar, 1999). However, to allow comparisons, BM and Tt-B.Ar. were expressed in their initial values (%) rather than in p′ values. Normality and homoscedasticity of the data were tested using Agostino–Pearson K2 and Pearson tests, respectively (Zar, 1999). For each parameter (malformations, fusions, total abnormalities, calcium, phosphate, ALP, TRAP, CT, T4, T3, TH and T3:T4 ratio), the data were statistically evaluated using ‘one-way’ analyses of variance (ANOVAs). In order to demonstrate the absence of noise disturbance of deformed vertebrae, two-way ANOVAs were performed on BM and Tt-B.Ar. measurements, taking normal and malformed trout separately. Tukey's honestly significant difference test was used a posteriori to detect differences among the three swimming speeds. All statistics were performed using JMPTM Statistical Software, version 5.1, with a significance level of 5%. Results are expressed in mean values with their standard errors of the mean (±s.e.m.). 3. Results 3.1. Fish growth No difference on growth rate was observed between trout reared at different swimming speeds (i.e., 0.0, 1.0 and 1.5 BL s− 1) (Fig. 2). Therefore, the three experimental groups were sampled at the same time, i.e., 9 months after spawning (180 days of experiments), when trout reached the following size: 0.0 BL s− 1: 257.4 ± 18 mm in total length (255.8 ± 48 g in total weight); 1.0 BL s− 1: 251.7 ± 19 mm (255.0 ± 54 g); and 1.5 BL s- 1: 252.1 ± 14 mm (240.5 ± 43 g). Mortality (3.0, 5.0 and 2.5% for 0, 1.0 and 1.5 BL s- 1, respectively) was negligible in the 3 experimental groups. Daily feeding rates (DFR)

Fig. 2. Growth (mean weight in g ± s.d.) of rainbow trout in the three swimming speed experiments. BL s- 1 = body length/second.

were of 2.74, 2.73 and 2.70% and feed conversion ratios (FCR) of 0.72, 0.73 and 0.76 for 0.0, 1.0 and 1.5 BL s- 1, respectively. Thus, at the end of the experiments, feeding rates were estimated at 1.97, 1.99 and 2.05 for 0.0, 1.0 and 1.5 BL s- 1, respectively. The condition factor (K) was of 1.5, 1.6 and 1.6 for 0.0, 1.0 and 1.5 BL s- 1, respectively. 3.2. Vertebra abnormalities Radiographic analysis of all sampled trout showed that most of the abnormal vertebrae were found in the caudal region (V34 to V60), but some were also observed in the anterior region, i.e. the first postcranial vertebrae (V1-V15) (Fig. 3). The distribution of the vertebral abnormalities was similar in the three experimental groups, although the number of abnormal vertebrae was highly variable among individuals, ranging from 0 to 32. In average, 60% of the trout had at least one abnormal vertebra. The number of affected trout was similar in the three groups (n = 12, 11 and 13 in the 0.0, 1.0 and 1.5 BL s- 1 group, respectively; Fig. 3). The occurrence of abnormal vertebrae in trout was found to be similar in the 3 groups (p N 0.05), even though it was a bit higher in trout swimming at 1.5 BL s- 1 (individual number of abnormal vertebrae = 6.3 ± 1.5, 6.2 ±1.8 and 7.6 ± 1.0 in the 0.0, 1.0 and 1.5 BL s- 1 group, respectively; Fig. 3). Interestingly, trout subjected to sustained exercise showed more malformed vertebrae (that are considered weak abnormalities) (1.0 BL s- 1 =4.6±1.6 and 1.5 BL s- 1 =5.2± 1.3), but less fused vertebrae (that are considered strong abnormalities) (1.0 BL s- 1 = 1.5 ± 0.5 and 1.5 BL s- 1 = 2.4 ± 1.6) than trout reared in still water (0.0 BL s- 1 = 1.7 ± 0.8 malformations and 4.7 ± 1.3 fusions; Fig. 3). Therefore, although vertebral abnormalities appeared to have been induced in all experimental groups, the strongest abnormalities, i.e. fused vertebrae, were more frequently present in trout not subjected to sustained exercise. 3.3. Vertebral bone mineralization (BM) and vertebral total bone area (Tt-B.Ar.) The results of the two-way ANOVAs for each bone condition parameters showed that swimming speeds have significant effects on vertebral bone mineralization (BM: p = 0.0001) and total bone area (Tt-B.Ar.: p=0.0019), and that these trends were not affected when taking normal and affected trout separately (BM and Tt-B.Ar.: pN 0.05; Table 1). The vertebrae of all trout reared in still water had a higher Tt-B.Ar. (33.6± 1.5%) and a lower BM (55.0±0.3%) than trout subjected to sustained exercise (Tt-B.Ar.=30.5±1.3% and 31.2±1.0% and BM=57.2±0.4% and 57.3±0.4%, for 1.0 and 1.5 BL s- 1 experiments, respectively) (BM: Fig. 4). However, no difference in Tt-B.Ar. or in BM was found between trout swimming at 1.0 and 1.5 BL s- 1. Sustained exercise decreased vertebral TtB.Ar and enhanced vertebral BM. Moreover, plotting BM against Tt-B.Ar. for the 60 trout analyzed (simple regression: r2 =0.18) revealed that BM tends to increase when Tt-B.Ar. decreases. This result could indicate that an increase in bone matrix mineralization (BM) compensates for the decrease in bone volume (Tt-B.Ar.). However, swimming either at 1.0 or 1.5 BL s- 1 did not change these parameters. This suggests that vertebral

Author's personal copy M.-H. Deschamps et al. / Aquaculture 296 (2009) 337–346

341

Fig. 3. Occurrence of vertebral abnormalities in reared rainbow trout. A. Location and number of abnormal vertebrae along the vertebral column of the 60 trout analyzed. B. Individual number of vertebrae with different types of abnormalities in the three swimming speed experiments; n = number of trout having at least one abnormal vertebrae on the total number of trout randomly sampled in each experiments. Bars with same colors and with different letters/symbols are significantly different. BL s- 1 = body length/second.

bone plasticity is limited and indicates that a water flow of 1.0 BL s- 1 seems necessary and sufficient to induce these changes in vertebral bone parameters.

3.4. Calcemia and phosphatemia Trout swimming at 1.5 BL s- 1 had a lower calcemia (91.4 ± 5.3 mg l- 1) compared to trout reared in still water (0.0 BL s- 1 = 100.2 ± 3.9 mg l- 1) or submitted to low water velocity (1.0 BL s- 1 = 105.1 ± 2.9 mg l- 1) (p = 0.0101). No difference in phosphatemia was found in the three experiments (p = 0.5157; overall mean= 152.1 ± 3.8 mg l- 1). Swimming activity did not induce an unbalance in calcium and phosphorus homeostasis in trout, though it reduced slightly calcemia in trout submitted to a water velocity of 1.5 BL s- 1.

3.5. ALP and TRAP activities No differences in ALP (p = 0.4537) and TRAP (p = 0.2148) values were found when comparing trout reared at the various swimming speeds. In average, the p-nitrophenyl values (pNP ml- 1 h- 1) were of 2444 ± 204 and 38.9 ± 7.4 nmol for ALP and TRAP, respectively. This result means that the variations observed in the vertebral bone parameters (BM and Tt-B.Ar.) did not result in significant variations of osteoblast and/or osteoclast activities (at least at the time of sampling). 3.6. Calcitonin (CT) Trout reared in still water had a lower amount of CT in the blood plasma (0.0 BL s- 1 = 0.52 ± 0.04 ng ml- 1) than trout submitted to

Table 1 Mean (± s.e.) vertebral bone mineralization (BM, %) and vertebral total bone area (Tt-B.Ar., %) of rainbow trout either affected or not (normal trout) by abnormal vertebrae (as observed by radiography) in the three swimming experiments. BM (%, mean ± s.e.m.)

Tt-B.Ar. (%, mean ± s.e.m.)

Swimming speed

Normal

Affected

All trout

Normal

Affected

All trout

0 BL s- 1

55.2 ± 0.3b (n = 8) 57.2 ± 0.4a (n = 9) 57.2 ± 0.4a (n = 7)

54.9 ± 0.3b (n = 12) 57.7 ± 0.4a (n = 11) 57.4 ± 0.4a (n = 13)

55.0 ± 0.3b (N = 20) 57.2 ± 0.4a (N = 20) 57.3 ± 0.4a (N = 20)

34.0 ± 1.4a (n = 8) 29.3 ± 1.4b (n = 9) 31.5 ± 0.8b (n = 7)

33.4 ± 1.6b (n = 12) 31.6 ± 1.2a (n = 11) 31.1 ± 1.0a (n = 13)

33.6 ± 1.5a (N = 20) 30.5 ± 1.3b (N = 20) 31.2 ± 1.0b (N = 20)

1.0 BL s- 1 1.5 BL s- 1

For each parameter, the mean values with a similar letter are not significantly different (MANOVAs p-value followed by Tukey's tests, p b 0.05). BL s- 1 = body length/second.

Author's personal copy 342

M.-H. Deschamps et al. / Aquaculture 296 (2009) 337–346

4. Discussion 4.1. Growth Rainbow trout reared in still water or subjected to sustained exercise showed no growth rate variation (same length and same weight, 9 months after spawning). Several studies had shown that swimming speeds up to 1.5 BL s− 1 improved growth in salmonids, although such a positive effect was not always revealed according to environmental rearing conditions (for a review, see Jobling et al., 1993; Davison, 1997). Our study indicates that a water flow of up to 1.5 BL s− 1 had no significant effect on growth rate in rainbow trout, in our rearing conditions. 4.2. Vertebra abnormalities

Fig. 4. Mean (± s.e., N total = 60) bone mineralization (BM, %) and total bone area (Tt-B. Ar., %) of rainbow trout vertebrae in the three swimming speed experiments. Bars with a similar letter are not significantly different (ANOVAs p-value followed by Tukey's tests, p b 0.05). BL s- 1 = body length/second.

sustained exercise (1.26 ± 0.04 ng ml- 1 and 1.31 ± 0.04 ng ml- 1, for 1.0 and 1.5 BL s- 1, respectively) (p = 0.0002; Fig. 5). No difference in CT values was found between trout in the 1.0 and 1.5 BL s- 1 groups. Swimming activity led to a rise of CT level in the plasma, but increasing swimming speed up to 1.5 BL s- 1 did not result in higher CT values.

The vertebral abnormalities reported in this study were estimated from a single replicate. Therefore, although significant, the observed differences in the deformity incidence between the different experiments discussed in this section have to be considered with caution and further experiments, conducted with several replicates, should be performed to confirm these results. The percentage (60%) of trout showing vertebral abnormalities on radiographs is high when compared to previous assessments in wild salmonids (Gill and Fisk, 1966; Poynton,1987), albeit not exceptional for reared rainbow trout. In fish farms, high occurrence of vertebral abnormalities was already reported for rainbow trout (Kacem et al., 2004; Madsen et al., 2001; Deschamps et al., 2008) and Atlantic salmon (Kvellestad et al., 2000; Helland et al., 2005, 2006; Witten et al., 2005, 2006; Sullivan et al., 2007a, 2007b). For instance, up to 55% (22% in average) of normally-shaped rainbow trout (i.e., showing no external malformation) were shown to be affected by vertebral anomalies in French farms (Deschamps et al., 2008). In addition, the number of abnormal vertebrae was found to vary greatly among individuals (from only a single abnormal vertebra to various compressed and/or fused

3.7. Thyroid hormones (T4, T3, TH and T3:T4 ratios) The plasma levels of thyroid hormones varied significantly in the three experimental groups (T4: p = 0.0213; T3: p = 0.0148; TH: p = 0.0156) (Fig. 6). In trout reared in still water, the amount of circulating thyroid hormones (T4, T3 and TH) was lower than in trout submitted to a sustained exercise. TH, T4 and T3 levels were not significantly different in the plasma of trout swimming at either 1.0 or 1.5 BL s- 1. The T3:T4 ratio was similar in trout submitted to different swimming speeds (p = 0.8176; Fig. 6). It appears that sustained exercise enhanced the production of thyroid hormones without affecting the conversion rate of thyroxine into tri-iodothyronine. Here again, the rise in thyroid hormones observed in trout swimming at 1.0 BL s− 1 was not higher when increasing water flow to 1.5 BL s− 1.

Fig. 5. Mean (±s.e., N total=30) values of calcitonin (CT, ng ml−1) measured in trout plasma (RIA) in the three swimming speed experiments. Bars with a similar letter are not significantly different (ANOVAs p-value followed by Tukey's tests, pb 0.05). BL s- 1 =body length/second.

Fig. 6. Mean (± s.e., N total = 30) values of thyroid hormones (T4, T3) levels (ng ml− 1) and T3:T4 ratio in trout plasma in the three swimming speed experiments. Bars with a similar letter are not significantly different (ANOVAs p-value followed by Tukey's tests, p b 0.05). BL s- 1 = body length/second.

Author's personal copy M.-H. Deschamps et al. / Aquaculture 296 (2009) 337–346

vertebrae). This reveals a wide range of plastic response of the axial skeleton to environmental factors as already reported in other salmonids (Kvellestad et al., 2000; Helland et al., 2005, 2006; Witten et al., 2005, 2006; Sullivan et al., 2007a, 2007b). The trout were sampled at only one time (market size) giving us no information that could indicate when vertebral abnormalities arisen. In addition, such anomalies could appear at various stages of development and/or during further growth (reviews in Cahu et al. (2003) and Lall and Lewis-McCrea (2007)). However, as the occurrence of vertebral abnormalities was similar in the three experimental groups, we think that they could have been induced during early development stages, when all trout were subjected to similar environmental factors. Indeed, most abnormalities reported in the literature are related to suboptimal rearing conditions during early developmental stages (Cahu et al., 2003; Lall and Lewis-McCrea, 2007). Therefore, a sustained swimming up to 1.5 BL s− 1 did not induce additional occurrence of vertebral abnormalities in rainbow trout. Most anomalies were encountered in the caudal region of the axial skeleton, a finding that confirms previous observations in reared rainbow trout (Kacem et al., 2004; Deschamps et al., 2008, 2009). Such location is thought to be in relation with the subcarangiform type of swimming of trout, in which the muscles located in this region ensure propulsion (Coughlin et al., 2004; Meunier and Ramzu, 2006; Ramzu and Meunier, 1999; Westneat and Wainwright, 2001). Vertebra compression and fusion could arise from intervertebral joint failures imputable to strong mechanical forces exerted by muscles on this region (Witten et al., 2005, 2006). During trout development and subsequent growth, continuous overexertion or short-term overload could lead to the failure of notochord cells maintaining a correct vertebral development in enhancing, for instance, chondrogenesis instead of osteogenesis. The chronic pressure could aggravate developmental abnormalities during growth (Witten et al., 2005, 2006). If we follow this reasoning, we would expect that trout subjected to sustained swimming showed more fused vertebrae than resting trout. This was not the case and we can deduce that a water flow up to 1.5 BL s− 1 does not result in an over-exercise for rainbow trout. These conditions seem rather advantageous as pre-existing abnormalities were not aggravated into more severe defaults (i.e., incomplete or complete vertebral fusions). Moreover, sustained swimming reduced the number of fused vertebrae. The explanation may rely on the same mechanism, i.e. the need of an appropriate mechanical load required to maintain the integrity (matrix composition and cellularity) of the intervertebral junction (Walsh and Lotz, 2004). The aggravation of existing abnormalities in resting trout could also arise from frequent short-term overloads due to escape-responses ahead of aggressions, as salmonids reared in still water generally show increased dominance hierarchies (Jobling et al., 1993; Davison, 1997). The lack of histological evidence impedes the identification of specific mechanisms involved in worsening of vertebral abnormalities. Taken together, our results bring evidence that sustained swimming up to 1.5 BL s− 1 favors appropriate vertebra growth (at least until market-sized of 26.0 cm TL) compared to resting specimens. 4.3. Bone remodeling and mineral balance In rainbow trout, sustained swimming significantly increased vertebral bone mineralization (BM) and decreased vertebral total bone area (Tt-B.Ar.). The vertebral bone condition at the time point of market size is the result of bone remodeling processes (bone deposition and resorption), which act to provide (i) a well-formed and solid endoskeleton able to resist to mechanical constraints, and (ii) a reservoir of minerals that can be mobilized to maintain the homeostasis balance (Persson et al., 1994; Fleming, 1996; Persson, 1997; Vielma and Lall, 1998; Yamada et al., 2001; Lall and Lewis-McCrea, 2007). It is now well established that increasing load and/or exercise enhances bone development and BM (Turner, 1998). Living in water, fish are not

343

subjected to gravitational forces, but their vertebrae play a major biomechanical role in ensuring muscle anchoring and in resisting lateral compressions while transmitting the forces during propulsion (Westneat and Wainwright 2001; Gemballa and Vogel, 2002; Coughlin et al., 2004; Meunier and Ramzu, 2006). In order to adequately fulfill their role, the vertebrae must provide an optimal mechanical competence (i.e., ‘stiffness’) that can be obtained in maximizing bone matrix formation and/or BM (Meunier and Ramzu, 2006; Paxton et al., 2006; Porter et al., 2006; Deschamps et al., 2009). Species subjected to swimming efforts (e.g., pelagic teleosts) show a higher vertebral BM than those adapted to benthic life (Ramzu, 1998; Meunier and Ramzu, 2006). Experimentally exposed vertebrae to high loads are more mineralized (Kranenbarg et al., 2004, 2005). In trout and other subcarangiform swimmers, the highest vertebral features (in length, volume, mineralization, etc.) are observed in the mid region of the vertebral column, where muscle forces and biomechanical constraints are the strongest (Ramzu and Meunier, 1999; Coughlin et al., 2004). Therefore, we interpret the increased BM as an appropriate adaptation of trout vertebrae to the experimental conditions, in which sustained swimming is concomitant to rapid growth. Similarly, the reduced vertebral bone area observed in the same individuals appears as the result of architectural adaptation of growing vertebrae to these conditions. The amount of mineralized collagen matrix, as assessed on microradiographs, depends mainly on calcium and phosphorus availability, as these minerals are involved in the formation of hydroxyapatite crystals (Francillon-Vieillot et al., 1990). However, these minerals are also important in various physiological processes (e.g., osmo-regulation, muscular activity, and reproduction; Fleming, 1996; Kacem et al., 1998, 2000; Kacem and Meunier, 2000, 2003; Witten and Hall, 2003). During rapid growth in active conditions, dietary uptake and absorption of minerals from the surrounding water could not fulfill all body requirements (Lall, 2002; Helland et al., 2005; Kaushik, 2005) and may lead to a reduction of vertebral bone area (for a review, see Sugiura et al., 2004). The total amount of mineralized bone area (Tt-B.Ar) resulting from bone modeling rate, i.e., bone matrix deposition (and mineralization), and resorption processes (mainly through osteoclast activity), a decrease of Tt-B.Ar. indicates that bone resorption is more important than bone deposition. We postulate that the vertebral architecture is modeled in such a way to reach the best compromise to resist biomechanical constraints with a limited amount of Tt-B.Ar. In youngest reared rainbow trout (from 190 to 235 mm TL), the vertebrae undergo major structural changes during growth: extended resorption of the mid region of the trabeculae and bone apposition restricted to their basal and peripheral regions (Deschamps et al., 2009). This important modeling could take part as ‘normal’ growth of vertebral features. Altogether, these findings indicate an optimization of the allocation of Ca and P (i.e., changes in bone repartition, which allow to not increase much Tt-B.Ar.) in order to maintain the mechanical function of the vertebrae while allowing mineral ion mobilization for other physiological demands during growth. Such a modeling supports the inverse relationship observed in our experimental groups between BM and Tt-B.Ar., and could indicate that an increase in BM (resulting from mechanical stimuli) improves the stiffness (rigidity) of the remaining bone, a mechanism compensating the reduction of Tt-B.Ar. in swimming trout. Conversely, in resting trout, sub-optimal mechanical stimuli could alter and/or delay vertebra mineralization during growth, prevent normal remodeling processes and also lead to inappropriate bone condition, to resist strong and sudden mechanical pressure. Moreover, it is worth to note that the increase of swimming effort from 1.0 to 1.5 BL s− 1 did not induce changes in vertebral bone mineralization and in vertebral bone area. This could reveal a limit to vertebral bone plasticity. Therefore, the vertebral bone condition observed in trout swimming actively seems to be a good compromise between the necessity

Author's personal copy 344

M.-H. Deschamps et al. / Aquaculture 296 (2009) 337–346

to mobilize mineral ions for vertebral growth and for many physiological demands, while maintaining a vertebra architecture permitting to resist mechanical constraints throughout growth. The changes observed in vertebral bone condition parameters were achieved although calcemia and phosphatemia levels were not found significantly different in the three experimental groups, at least at the only time point of market size. Similarly, we found that swimming activity did not induce any measurable change in bone cell activities (as revealed by unchanged plasma ALP and TRAP levels. Plasma ALP and TRAP are closely associated with the metabolism of Ca and P, and are often used as markers, where plasma ALP is used for chondrogenic and osteoblastic activities, while plasma TRAP used for osteoclastic activities, respectively. In teleost, TRAP was related to bone (and scale) resorption (Witten and Villwock, 1997; Witten et al., 2007; Rotllant et al., 2005; Nemoto et al., 2007) and inversely related to mineralization (Suzuki et al., 2000; Mukherjee et al., 2004). Nevertheless, little and contradictory information is currently available for what concerns ALP activities in relation to remodeling processes. Indeed, although plasma ALP was found to increase when P availability increased in rainbow trout (Skonberg et al., 1997) and Japanese sea bass (Zhang et al., 2006), this relation was not always confirmed in rainbow trout (Shearer and Hardy, 1987; present study) as well as in black (Shao et al., 2008) and red (Sakamoto and Yone, 1980) sea breams. Moreover, calcitonin was found to suppress plasma ALP (Rotllant et al., 2005). Therefore, to perform comparative analyses, there is a strong need to assess the normal ranges of plasma ALP and to evaluate its variations under various physiological and environmental factors. The levels of calcemia, phosphatemia, ALP and TRAP probably varied earlier in trout ontogeny, in particular at the time point when the modeling processes are activated in trout subjected to active swimming. Mineral homeostasis should be measured at several time points during trout growth with appropriate protocols allowing the dynamics of the process to be monitored. 4.4. Hormonal bone regulation In teleost fish, as in other vertebrates, bone metabolism is regulated through the activity of several hormones. Calcitonin (CT), an endocrine hormone secreted by the ultimobranchial body, regulates calcium metabolism in promoting bone apposition and mineralization, while suppressing bone and scale resorption (Lopez, 1973; Suzuki et al., 2000; Mukherjee et al., 2004). In our study, there is evidence of a relationship between mineral deposition (BM) and CT production. Indeed, sustained exercise simultaneously enhanced CT production and vertebral BM in trout, while both parameters did not change when swimming speed increased from 1.0 to 1.5 BL s− 1. According to our current knowledge and our observations, this suggests that circulating CT must have been one of the major factors of the regulatingmechanisms improving bone mineralization in trout subjected to sustained exercise. Thyroid hormones (TH), combined to growth hormone (GH), are involved in various biological processes such as development and growth (Eales and Brown, 1993; Power et al., 2001). Circulating L-thyroxine (T4) levels are generally used to evaluate the thyroid status (i.e., synthesis activity; Eales and Brown, 1993). Once released in blood, T4 is deiodinated by outer-ring deodinases (ORD) to produce 3,5,3′triiodo-L-thyronine (T3), a bioactive form having a 10-fold higher affinity to TH receptors. Plasma T3 level is believed to regulate energyrequiring processes (Eales and McLatchy, 1989; Edeline et al., 2004). The T3:T4 ratio reflects the rate of T4 ORD activity, a mechanism that allows an increased thyroid hormone activity despite a decreased thyroid gland synthesis-activity as, for instance, during a physical stress (Johnston et al., 1996; Todd and Eales, 2002). In this study, trout subjected to sustained exercise had higher TH levels than trout reared in still water, which demonstrates a better physiological status of the former, associated to an increased energy demand of the tissues.

Moreover, the positive influence of swimming exercise on trout physiology was not associated to change in T3:T4 ratio. As expected (Jobling et al., 1993; Davison, 1997), sustained swimming had positive effects on trout physiology during our experiments. In teleosts, thyroid hormones may have either positive or negative dose-dependent effects in modulating the balance between bone formation and resorption according to the physiological status (Lopez, 1973; Takagi et al., 1994; Sbaihi et al., 2007). Indeed, in juvenile trout, thyroid hormone treatments (T4 or T3) were shown to stimulate skeletal growth by activating two processes, bone formation by osteoblasts and osteoclast resorption (Lopez, 1973; Takagi et al., 1994). In addition, T4 treatment in juvenile trout did not enhance periosteocytic osteolysis (Lopez, 1973), a process known to reduce bone mineral content. According to our knowledge and the present results, the variations of the vertebral bone parameter (either high BM or low Tt-B.Ar. in swimming trout compared to resting ones) seem to be the result of normal and general growth regulation process. Actually, several other hormones as growth hormone (Wargelius et al., 2005b), sex steroids (Carragher and Sumpter, 1991; Persson, 1997) and corticoid hormones (Flik and Perry, 1989; Sbaihi, 2001; Guerreiro et al., 2006) could have played specific roles during our experiments. In the near future, further studies could be conducted in order to elucidate the respective role of these molecules during vertebra growth. 5. Conclusion There is a great need to improve the rearing conditions in order to reduce the number and severity of vertebral abnormalities in reared trout, particularly juxtaposed and fused vertebrae, as they impede filet production and lead to external deformations. Water flow in tanks is an abiotic factor that can be easily controlled by producers. Increasing swimming speed up to 1.5 BL s− 1 enhances calcitonin and thyroid hormone production, which seems to have a positive effect on the vertebral skeleton. This is a suitable method to prevent aggravation of vertebral abnormalities (at least in juvenile trout from 65 to 260 mm). Further histological studies should reveal more precisely tissular and cellular features related to bone remodeling, in relation to mechanical strengths affecting vertebral growth. Acknowledgements We gratefully acknowledge François J. Meunier (MNHN) for his advices, Joël Aubin (INRA) for discussions aiming to define the experiments, and the expertise of Vincent Gayet (PEIMA) in the setting and control of swimming experiments. We thank Mehboob Chilwan (Erasmus, University of Keele, UK) for English corrections. This research was funded by the Office national interprofessionnel des produits de la mer et de l′Aquaculture (OFIMER, Contract No. 050/04/ C), the European Union (IFOP/DPMA, Contract No. 2005/010) and the Comité Interprofessionnel des Produits de l′Aquaculture (CIPA). References Boglione, C., Gagliardi, F., Scardi, M., Cataudella, S., 2001. Skeletal descriptors and quality assessment in larvae and post-larval of wild-caught and hatchery-reared gilthead sea bream (Sparus aurata L. 1758). Aquaculture 192, 1–22. Cahu, C., Zambonino Infante, J., Takeuchi, T., 2003. Nutritional components affecting skeletal development in fish larvae. Aquaculture 227, 245–258. Carragher, J.F., Sumpter, J.P., 1991. The mobilization of calcium from calcified tissues of rainbow trout (Oncorhynchus mykiss) induced to synthesize vitellogenin. Comp. Biochem. Physiol. 99, 169–172. Coughlin, D.J., Spiecker, A., Schiavi, J.M., 2004. Red muscle recruitment during steady swimming correlates with rostral-caudal patterns of power production in trout. Comp. Biochem. Physiol. Part A 137, 151–160. Davison, W., 1997. The effects of exercise training on teleost fish, a review of recent literature. Comp. Biochem. Physiol. Part A 117, 67–75. Deschamps, M.-H., Kacem, A., Ventura, R., Courty, G., Haffray, P., Meunier, F.J., Sire, J.-Y., 2008. Assessment of “discreet” vertebral abnormalities, bone mineralization and bone compactness in farmed rainbow trout. Aquaculture 279, 11–17.

Author's personal copy M.-H. Deschamps et al. / Aquaculture 296 (2009) 337–346 Deschamps, M.-H., Girondot, M., Labbé, L., Sire, J.-Y., 2009. Changes in vertebral structure during growth of reared rainbow trout Oncorhynchos mykiss (Walbaom): a new approach using modeling of vertebral bone profiles. J. Fish Diseases 32, 233–246. Divanach, P., Papandroulakis, N., Anastasiadis, P., Koumoundouros, G., Kentouri, M., 1997. Effect of water currents on the development of skeletal deformities in sea bass (Dicentrarchus labrax L.) with functional swimbladder during postlarval and nursery phase. Aquaculture 156, 145–155. Eales, J.G., Brown, S.B., 1993. Measurements and regulation of thyroidal status in teleost fish. Rev. Fish Biol. Fish. 3, 199–347. Eales, J.G., McLatchy, D.L., 1989. The relationship between T3 production and energy balance in salmonids and other teleosts. Fish Physiol. Biochem. 7, 289–293. Edeline, E., Dufour, S., Briand, C., Fatin, D., Elie, P., 2004. Thyroid status is related to migratory behavior in Anguilla anguilla glass eels. Mar. Ecol. Prog. Ser. 282, 261–270. Fleming, I.A., 1996. Reproductive strategies of Atlantic salmon: ecology and evolution. Rev. Fish Biol. Fish. 6, 379–416. Flik, G., Perry, S.F., 1989. Cortisol stimulates whole body calcium uptake and the branchial calcium pump in freshwater rainbow trout. J. Endocrinol. 120, 75–82. Fouchereau-Peron, M., Arlot-Bonnemains, Y., Taboulet, J., Milhaud, G., Moukhtar, M.S., 1990. Distribution of calcitonin gene-related peptide and calcitonin-like immunoreactivity in trout. Regul. Peptides 27, 171–179. Francillon-Vieillot, H., de Buffrénil, V., Castanet, J., Géraudie, J., Meunier, F.J., Sire, J.-Y., Zylberberg, L., de Ricqlès, A., 1990. Microstructure and mineralization of vertebrate skeletal tissues. In: Carter, J.G. (Ed.), Skeletal Biomineralization: Patterns Processes and Evolutionary Trends. Van Nostrand Reinhold, New York, pp. 471–530. Gemballa, S., Vogel, F., 2002. Spatial arrangement of white muscle fibers and myoseptal tendons in fishes. Comp. Biochem. Physiol. Part A 133, 1013–1037. Gil Martens, L., Witten, P.E., Fivelstad, S., Huysseune, A., Saevareid, B., Vikesa, V., Obach, A., 2006. Impact of high water carbon dioxide levels on Atlantic salmon smolts (Salmo salar L.): effects on fish performance, vertebrae composition and structure. Aquaculture 261, 80–88. Gill, C.D., Fisk, D.M., 1966. Vertebral abnormalities in sockeye, pink, and chum salmon. Trans. Amer. Fish. Soc. 95, 177–182. Girondot, M., Laurin, M., 2003. Bone Profiler: a tool to quantify, model and statistically compare bone section compactness profile. J. Vertebr. Paleontol. 23, 458–461. Guerreiro, P.M., Rotland, J., Fuentes, J., Power, D.M., Canario, A.V.M., 2006. Cortisol and parathyroid hormone-related peptide are reciprocally modulated by negative feedback. Gen. Comp. Endocrinol. 148, 227–235. Helland, S., Refstie, S., Espmark, A., Hjelde, K., Baeverfjord, G., 2005. Mineral balance and bone formation in fast-growing Atlantic salmon parr (Salmo salar) in response to dissolved metabolic carbon dioxide and restricted dietary phosphorus supply. Aquaculture 250, 364–376. Helland, S., Denstadli, V., Witten, P.E., Hjelde, K., Storebakken, T., Skrede, A., Åsgård, T., Baeverfjord, G., 2006. Hyper dense vertebrae and mineral content in Atlantic salmon (Salmo salar L.) fed diets with graded levels of phytic acid. Aquaculture 261, 603–614. Hermann-Erlee, M.P.M., Flik, G., 1989. Bone: comparative studies on endocrine involvement in bone metabolism. In: Pang, P.K.T., Schreibman, M.P. (Eds.), Vertebrate Endocrinology; Fundamentals and Biomedical Implications. Academic Press, New York, pp. 211–243. Jobling, M., Baardvik, B.M., Christiansen, J.S., Jørgensen, E.H., 1993. The effects of prolonged exercise training on growth performance and production in fish. Aquacult. Int. 1, 95–111. Johnston, C.E., Gordillo, C., Eales, J.G., 1996. Transition from hatchery to a laboratory environment induces inner-ring monodeiodination pathways for thyroid hormones in the liver of rainbow trout, Oncorhynchus mykiss. Can. J. Zool. 74, 2178–2183. Kacem, A., Meunier, F.J., 2000. Mise en évidence de l′ostéolyse périostéocytaire vertébrale chez le saumon atlantique Salmo salar (Salmonidae, Teleostei) au cours de sa migration anadrome. Cybium 24, 105–112. Kacem, A., Meunier, F.J., 2003. Halastatic demineralization in the vertebrae of Atlantic salmon, during their spawning migration. J. Fish. Biol. 63, 1122–1130. Kacem, A., Meunier, F.J., Baglinière, J.L., 1998. A quantitative study of morphological and histological changes in the skeleton of Salmo salar during its anadromous migration. J.Fish Biol. 53, 1096–1109. Kacem, A., Gustafsson, S., Meunier, F.J., 2000. Demineralization of the vertebral skeleton in Atlantic salmon Salmo salar L. during spawning migration. Comp. Biochem. Physiol. Part A 125, 479–484. Kacem, A., Meunier, F.J., Aubin, J., Haffray, P., 2004. Caractérisation histo-morphologique des malformations du squelette vertébral chez la truite arc-en-ciel (Oncorhynchus mykiss) après différents traitements de triploïdisation. Cybium 28, 15–23. Kaushik, S.-J., 2005. Besoins et apport en phosphore chez les poissons. INRA. Prod. Anim. 18, 203–208. Kihara, M., Ogata, S., Noriaki, K., Kubota, I., Yamaguchi, R., 2002. Lordosis induction in juvenile red sea bream, Pagrus major, by high swimming activity. Aquaculture 212, 149–158. Koumoundouros, G., Gagliardi, F., Divanach, P., Boglione, C., Cataudella, S., Kentouri, M., 1997. Normal and abnormal osteological development of caudal fin in Sparus aurata L. fry. Aquaculture 149, 215–226. Koumoundouros, G., Divanach, P., Kentouri, M., 2001. The effect of rearing conditions on development of saddleback syndrome and caudal fin deformities in Dentex dentex (L.). Aquaculture 200, 285–304. Kranenbarg, S., Waarsing, J.H., Muller, M., Weinans, H., van Leeuwen, J.L., 2004. Lordotic vertebrae in sea bass (Dicentrarchus labrax L.) are adapted to increased loads. J. Biomech. 38, 1239–1246. Kranenbarg, S., van Cleynenbreugel, T., Schipper, H., van Leeuwen, J.L., 2005. Adaptive bone formation in acellular vertebrae of sea bass (Dicentrarchus labrax L.). J. Exp. Biol. 208, 3493–3502.

345

Kvellestad, A., Høie, S., Thorud, K., Tørud, B., Lyngøy, A., 2000. Platyspondyly and shortness of vertebral column in farmed Atlantic salmon Salmo salar in Norway — description and interpretation of pathologic changes. Dis. Aquat. Org. 39, 97–108. Lall, S.P., 2002. The minerals. In: Halver, J.E., Hardy, R.W. (Eds.), Fish nutrition, Third edition. Elsevier, San Diego, USA, pp. 259–308. Lall, S.P., Lewis-McCrea, L.M., 2007. Role of nutrients in skeletal metabolism and pathology in fish — an overview. Aquaculture 267, 3–19. Lopez, E., 1973. Étude morphologique et physiologique de l′os cellulaire des poissons téléostéens. Mem. MNHN 80, 1–90. Madsen, L., Arnbjerg, J., Dalsgaard, I., 2001. Radiological examination of the spinal column in farmed rainbow trout Oncorhynchus mykiss (Walbaum): experiments with Flavobacterium psychrophilum and oxytetracycline. Aquacult. Res. 32, 235–241. Meunier, F.-J., Ramzu, M.-Y., 2006. La régionalisation morphofonctionnelle de l′axe vertébral chez les Téléostéens en relation avec le mode de nage. C. R. Palevol 5, 499–507. Mukherjee, D., Sen, U., Bhattacharyya, S.P., 2004. The effects of calcitonin on plasma calcium levels and bone metabolism in the fresh water teleost Channa punctatus. Comp. Biochem. Physiol. 138, 417–426. Nemoto, Y., Higuchi, K., Baba, O., Kudo, A., Takano, Y., 2007. Multinucleate osteoclasts in medaka as evidence of active bone remodeling. Bone 40, 399–408. Parfitt, A.M., Drezner, M.K., Glorieux, F.H., Kanis, J.A., Malluche, H., Meunier, P.J., Ott, S.M., Recker, T.R., 1987. Bone histomorphometry: standardization of nomenclature, symbols, and units. J. Bone Min. Res. 2, 595–610. Paxton, H., Bonser, R.H.C., Winwood, K., 2006. Regional variation in the microhardness and mineralization of vertebrae from brown and rainbow trout. J. Fish. Biol. 68, 481–487. Persson, P., 1997. Calcium Regulation During Sexual Maturation of Female Salmonids: Estradiol-17B and Calcified Tissues. Department of Zoophysiology, Göteborg. Persson, P., Sundell, K., Björnsson, B.T., 1994. Estradiol-17b-induced calcium uptake and resorption in juvenile rainbow trout Oncorhynchus mykiss. Fish Physiol. and Biochem. 13, 379–386. Porter, M.E., Beltrán, J.L., Koob, T.J., Summers, A.P., 2006. Material properties and biochemical composition of mineralized vertebral cartilage in seven elasmobranch species (Chondrichthyes). J. Exp. Biol. 209, 2920–2928. Power, D.M., Llewellyn, L., Faustino, M., Nowell, M.A., Björnsson, B.T., Einarsdottir, I.E., Canario, A.V.M., Sweeney, G.E., 2001. Thyroid hormones in growth and development of fish. Comp. Biochem. Physiol. Part C 130, 447–459. Poynton, S.L., 1987. Vertebral column abnormalities in brown trout, Salmo trutta L. J. Fish. Dis. 10, 53–57. Ramzu, M., 1998. Vertebral histology in four teleostean species having different modes of axial swimming. Ital. J. Zool. 65, 449–454. Ramzu, M., Meunier, F.J., 1999. Descripteurs morphologiques de la zonation de la colonne vertébrale chez la truite arc-en-ciel Oncorhynchus mykiss (Walbaum, 1792) (Teleostei, Salmoniforme). Ann. Sci. Nat. 3, 87–97. Rotllant, J., Redruello, B., Guerreiro, P.M., Fernandes, H., Canario, A.V.M., Power, D.M., 2005. Calcium mobilization from fish scales is mediated by parathyroid hormone related protein via the parathyroid hormone type 1 receptor. Regul. Peptides 132, 33–40. Sakamoto, S., Yone, Y., 1980. A principal source of deposited lipid in phosphorus deficient red sea bream. Bull. Japan. Soc. Sci. Fish 46, 1227–1230. Sbaihi, M., 2001. Interaction des stéroïdes sexuels et du cortisol dans le contrôle de la reproduction et du métabolisme calcique chez un téléostéen migrateur, l′anguille (Anguilla anguilla L.). Département de physiologie de la reproduction. Université Pierre et Marie Curie, Paris. pp. 214. Sbaihi, M., Kacem, A., Aroua, S., Baloche, S., Rousseau, K., Lopez, E., Meunier, F., Dufour, S., 2007. Thyroid hormone-induced demineralisation of the vertebral skeleton of the eel, Anguilla anguilla. Gen. Comp. Endocrinol. 151, 98–107. Sfakianakis, D.G., Koumoundouros, G., Divanach, P., Kentouri, M., 2004. Osteological development of the vertebral column and of the fins in Pagellus erythrinus (L. 1758). Temperature effect on the developmental plasticity and morpho-anatomical abnormalities. Aquaculture 232, 407–424. Shao, Q., Ma, J., Xu, Z., Hu, W., Xu, J., Xie, S., 2008. Dietary phosphorus requirement of juvenile black seabream, Sparus macrocephalus. Aquaculture 277, 92–100. Shearer, K.D., Hardy, M., 1987. Phosphorus deficiency in rainbow trout fed a diet containing deboned scrap. Prog. Fish-Cult. 49, 192–197. Skonberg, D.I., Yogev, L., Hardy, R.W., Dong, F.M., 1997. Metabolic response to dietary phosphorus intake in rainbow trout (Oncorhynchus mykiss). Aquaculture 157, 11–24. Souvannavong, V., Lemaire, C., De Nay, D., Brown, S., Adam, A., 1996. Expression of alkaline phosphatase by a B-cell hybridoma and its modulation during cell growth and apoptosis. Immunol. Lett. 47, 163–170. Sugiura, S.H., Hardy, R.W., Roberts, R.J., 2004. The pathology of phosphorus deficiency in fish — a review. J. Fish. Dis. 27, 255–265. Sullivan, M., Guy, D.R., Roberts, R.J., Manchester, N.J., 2007a. The aetiology of spinal deformity in Atlantic salmon, Salmo salar L.: influence of genetic factors on the frequency and severity in freshwater stages. J. Fish. Dis. 30, 753–758. Sullivan, M., Hammond, G., Roberts, R.J., Manchester, N.J., 2007b. Spinal deformation in commercially cultured Atlantic salmon, Salmo salar L.: a clinical and radiological study. J. Fish. Dis. 30, 745–752. Suzuki, T., Srivastava, A.S., Kurokawa, T., 2000. Experimental induction of jaw, gill and pectoral fin malformations in Japanese flounder, Paralichthys olicaceus, larvea. Aquaculture 185, 175–187. Takagi, Y., Tanabe, J.H., Yamada, J.Pleasecheckcapturedauthor.H.,1994. Stimulation of skeletal growth by thyroid hormone administrations in the rainbow trout, Oncorhynchus mykiss. J. Exp. Zool. 268, 229–238. Todd, K.J., Eales, J.G., 2002. The effect of handling and blood removal on plasma levels and hepatic deiodination of thyroid hormones in the rainbow trout Oncorhynchus mykiss. Gen. Comp. Endocrinol. 85, 376–384.

Author's personal copy 346

M.-H. Deschamps et al. / Aquaculture 296 (2009) 337–346

Turner, C.H., 1998. Three rules for bone adaptation to mechanical stimuli. Bone 23, 399–407. Vielma, J., Lall, S.P., 1998. Phosphorus utilization by Atlantic salmon (Salmo salar) reared in freshwater is not influenced by higher dietary calcium intake. Aquaculture 160,117–128. Walsh, A.L., Lotz, J.C., 2004. Biological response of intervertebral disc to dynamic loading. J. Biomech. 2004, 329–337. Wargelius, A., Fjelldal, G., Hansen, T., 2005a. Heat shock during early somitogenesis caudal vertebral column defects in Atlantic salmon (Salmo salar). Dev. Genes. Evol. 215, 350–357. Wargelius, A., Fjelldal, G., Benedet, S., Hansen, T., Björnsson, B.T., Nordgarden, U., 2005b. A peak in GH-receptor is associated with growth activation in Atlantic salmon vertebrae, while upregulation of IGF-I receptor expression is related to increased bone density. Gen. Comp. Endocrinol. 142, 163–168. Westneat, M.W., Wainwright, S.A., 2001. Mechanical design for swimming: muscle, tendon, and bone. In: Block, B., Stevens, E. (Eds.), Fish Physiology. Academic Press, pp. 271–311. 7. Witten, P.E., Hall, B.K., 2003. Seasonal changes in the lower jaw skeleton in male Atlantic salmon (Salmo salar L.): remodelling and regression of the kype after spawning. J. Anat. 203, 435–450. Witten, P.E., Villwock, W., 1997. Growth requires bone resorption at particular skeletal elements in a teleost fish with acellular bone (Oreochromis niloticus, Teleostei: Cichlidae). J. Appl. Ichthyol. 13, 149–158.

Witten, P.E., Villwock, W., Peters, N., Hall, B.K., 2000. Bone resorption and bone remodelling in juvenile carp, Cyprinus carpio L. J. Appl. Ichthyol. 16, 254–261. Witten, P.E., Gil-Martens, L., Hall, B.K., Huysseune, A., Obach, A., 2005. Compressed vertebrae in Atlantic salmon Salmo salar: evidence for metaplastic chondrogenesis as skeletogenic response late in ontogeny. Dis. Aquat. Org. 64, 237–246. Witten, P.E., Obach, A., Huysseune, A., Baeverfjord, G., 2006. Vertebrae fusion in Atlantic salmon (Salmo salar): development, aggravation and pathways of containment. Aquaculture 258, 164–172. Yamada, Y., Okamura, A., Tanaka, S., Utoh, T., Horie, N., Mikawa, N., Oka, H.P., 2001. The roles of bone and muscle as phosphorus reservoirs during the sexual maturation of female Japanese eels, Anguilla japonica Temminck and Schlegel (Anguilliformes). Fish Physiol. Biochem. 24, 327–334. Zar, 1999. Biostatistical Analysis4th edition. Simon & Schuster/A Viacom Company, Upper Saddle River. 663pp. Zhang, C., Mai, K., Ai, Q., Zhang, W., Duan, Q., Tan, B., Ma, H., Xu, W., Liufu, Z., Wang, X., 2006. Dietary phosphorus requirement of juvenile Japanese seabass, Lateolabrax japonicus. Aquaculture 255, 201–209.