Aquaculture 298 (2009) 101–110

Contents lists available at ScienceDirect

Aquaculture

j ourna l homepage: www.e lsev ie r.com/ locate /aqua-on l ine

An assessment of organ and intestinal histomorphology and cellular stress responsein Atlantic salmon (Salmo salar L.) fed genetically modified Roundup Ready® soy

N.H. Sissener a,⁎, A.M. Bakke b,c, J. Gu b, M.H. Penn b,c, E. Eie a,d, Å. Krogdahl b,c, M. Sanden a, G.I. Hemre a

a National Institute of Seafood and Nutrition Research (NIFES), Bergen, Norwayb Norwegian School of Veterinary Science, Oslo, Norwayc Aquaculture Protein Centre, a CoE, Norwayd Skretting ARC, Stavanger, Norway

Abbreviations: ALAT, alalanine aminotransferase; ASANF, anti-nutritional factor; Ct, threshold cycle; FFSBgenetically modified; Hct, haematocrit; Hb, haemoglobinlactate dehydrogenase; MCH, mean cell haemoglobin; Mconcentration; MCV, mean cell volume; MF, mitotic figuMMC, melanomacrophage centre; PCR, polymerase chainRRS®, Roundup Ready® soy; SBM, soybean meal; TAG, tria⁎ Corresponding author. NIFES, Postboks 2029 No

Tel.: +47 41462385; fax: +47 55905290.

0044-8486/$ – see front matter © 2009 Elsevier B.V. Aldoi:10.1016/j.aquaculture.2009.10.011

a b s t r a c t

a r t i c l e i n f o

Article history:Received 14 August 2009Received in revised form 7 October 2009Accepted 12 October 2009

Keywords:Atlantic salmonGenetic modificationRoundup Ready® soyStressIntestine

This study was conducted to investigate potential differences between genetically modified (GM) RoundupReady® soy and its near-isogenic maternal line as feed ingredients for Atlantic salmon, with focus onintestinal changes commonly caused by soybean meal, histomorphology of other organs and stress response.A 7-month feeding trial was conducted with an inclusion level of 25% GM soy in the diet. Samples forhistology were collected after 4 months, after 6 months, when a cross-over of the diet groups was conducted,and at the end of the trial of the crossed-over groups. Histomorphology of spleen, head kidney and midintestine exhibited no differences between the diet groups, while glycogen deposits in liver were decreasedin the GM fed fish at the final sampling. Common soybean meal-induced changes of the distal intestine inAtlantic salmon were observed in both diet groups at all sampling points, within levels expected at thecurrent inclusion level of soy in the diets. However, mucosal fold height in the distal intestine was lower inthe GM fed group at one of the three sampling points, and mucosal fold fusion was more pronounced in thisgroup overall in the trial. A stress test conducted at the end of the trial gave responses in haematologicalparameters, plasma nutrients and mRNA transcription of heat shock protein (HSP) 27 in both liver and distalintestine, but responses were similar between the two diet groups, indicating similar ability to handle stress.The cross-over design, implemented to look at reversibility of potential GM-effects, proved to be inadequateas the crossing of diet groups in itself caused responses that would obscure possible minor diet effects. Inconclusion, minor differences were observed between the diet groups; however, GM soy did not appear tocause any adverse effects on organ morphology or stress response compared to non-GM soy.

AT, aspartate aminotransferase;M, fullfat soybean meal; GM,; HSP, heat shock protein; LDH,CHC, mean cell haemoglobin

res; MFH, mucosal fold height;reaction; RBC, red blood cell;cylglycerol.rdnes 5817 Bergen, Norway.

l rights reserved.

© 2009 Elsevier B.V. All rights reserved.

1. Introduction

The potential occurrence of unintended effects of genetic modi-fication is one of the issues to be addressed in safety assessment ofgenetically modified (GM) plants used as feed and food (Kuiper andKleter, 2003). Transgene insertion is an imprecise and poorlyunderstood event, and introduction of superfluous DNA, as well asdeletions and rearrangements of host DNA at the insertion site, arecommon occurrences (Somers and Makarevitch, 2004; Latham et al.,2006). This might disrupt transcription of endogenous genes,

resulting in unintended changes in levels of macro- or micronutrients,anti-nutritional factors (ANFs) or production of toxic compounds(Cellini et al., 2004). The regulatory process is designed to look forthese types of changes in GM plants, but this is a targeted approachthat will never cover all known and unknown compounds in the plant.At present, the EU has approved about 30 GM plant products for use infoods and feeds (http://ec.europa.eu/food/dyna/gm_register/index_en.cfm). However, knowledge regarding possible health effectsin animals and man is still sparse (Pryme and Lembcke, 2003;Domingo, 2007). Most studies conducted with GM plants have beenrelatively short term and focused on production parameters such asgrowth, milk production or meat yield, rather than investigating earlybiomarkers for physiology, health and reproduction parameters.Discussions regarding the necessity of evaluating products apparentlysimilar to traditional counterparts (EFSA, 2008), and lack ofstandardized methods for the evaluation of unintended effects(Kuiper and Kleter, 2003), might be reasons for limited researchactivity in this area.

Processed soybeans are the largest source of protein feed and thesecond largest source of vegetable oil in the world. Of the global

Table 1Full fat soybean meal (FFSBM) composition, formulation and proximate composition ofthe diets.

nGM GM

FFSBM compositionProtein (%) 36.7 38.5Lipids (%) 22.6 20.4Starch (%) 1.8 1.6Ash (%) 4.8 5.1Dry matter (%) 92.6 92.9Residue (%)1 26.7 27.3

Formulation g kg −1(2)

Fishmeal 510 510nGM soy 262GM soy 250Fish oil 150 150Soy oil 8Wheat 75 79Vitamin/mineral mix 3 3

Feed compositionDry matter (%) 94.0 93.8Total protein (%) 46.1 45.8Lipids (%) 24.2 24.6Ash (%) 9.8 9.9Starch (%) 6.1 6.2Residue (%)1 7.9 7.3Vitamin B6 (mg/kg) 14.7 14.8Gross energy (kJ/g)3 21.5 21.6

1Residue was calculated as dry matter-(protein+lipid+starch+ash). 2The fishmealused was Norse-LT 94 Nordic fishmeal, made from 65% blue whiting, 30% sprat and 5%cuttoffs (fish industry byproducts) (Fiskernes Fiskeindustri, Denmark). Wholesoybeans (RRS® and non-GM near-isogenic maternal line) were kindly supplied byMonsanto and ground to FFSBM by Skretting ARC. Wheat was bought from DanskLandbrugs Grovvareselskap and the Vitamin/mineral mix from Trouw Nutrition (theNetherlands). 3Gross energy was calculated according to (Tacon, 1987) using theenergy content of 39.5 kJg−1 for lipid, 23.6 for protein and 17.2 for starch. The resultsare presented as the average of two analytical parallels.

102 N.H. Sissener et al. / Aquaculture 298 (2009) 101–110

acreage planted with soybeans, 64% is now GM (James, 2007). Thedominating GM variety is Roundup Ready® soy (RRS®), which ismodified to be tolerant to the herbicide glyphosate (Padgette et al.,1995). Soy is commonly used in feed for Atlantic salmon at low levelsand has, as a plant protein, a well balanced amino acid profilecompared to the requirements of fish (Gatlin et al., 2007). However,the levels of standard qualities used in diets for Atlantic salmon arelimited due to immunological responses in the intestine which seemto be dose and time dependent (Uran, 2008).

Effects of soybean meal (SBM) on the distal intestine of salmoninclude inflammatory responses (enteritis), which seem to be causedby one or several alcohol-soluble ANFs, such as saponins, phytosterols,oligosaccharides and/or other unidentified components (Van denIngh et al., 1991; Baeverfjord and Krogdahl, 1996; Van den Ingh et al.,1996). Further, there are decreases in both in total weight andmucosal fold height, while lamina propria is widened and infiltratedby a mixed population of leukocyte cells (Van den Ingh et al., 1991;Baeverfjord and Krogdahl, 1996; Nordrum et al., 2000). As salmon areso sensitive to soy inclusion in the diet, even compared to other fishspecies (Evans et al., 2005), unintended alterations in GM varietiescould have implications for the suitability of GM soy as a feedingredient for this species.

In addition to the gastrointestinal tract, which is the first site ofexposure to diet ingredients, the liver is a target organ in dietary studies.The liver is a keymetabolic organ and has an important role in responseto toxicants and immune response (Morin et al., 1993; Benninghoff andWilliams, 2008; Tintos et al., 2008). Histomorphological changes havebeen observed in hepatocytes of mice fed GM soy, involving nuclearmodifications that were shown to be reversible when mice wereswitched to a non-GM soy diet (Malatesta et al., 2002, 2005). Nuclearmodifications of hepatocytes have also been observed in sheep fed Bt-maize (Trabalza-Marinucci et al., 2008). Furthermore, histologicalevaluations of other organs have been used to assess effects of GMingredients in diets for mice and rats (Malatesta et al., 2003; Vecchioet al., 2004; Séralini et al., 2007). However, some of these studies havemet criticism, and there are other feeding trials in which no differencesbetween animals fed conventional or GM ingredients have beendetected (Flachowsky et al., 2007).

Inadequate nutrition or harmful substances might only have aneffect when conditions are suboptimal, as the fish will then struggle tomaintain homeostasis. Thus, it might be interesting to compare stressresponse in fish fed non-GM and GM soy. The physiological stressresponse entails increases in stress hormones followed by activationof metabolic pathways, such as mobilization of energy reserves tomaintain or attempt to re-establish homeostasis, and physiologicalresponses such as alterations in blood chemistry and haematology(Barton and Iwama, 1991). On a cellular level, heat shock proteins(HSPs) have been proposed as an indicator of stressed states in fish(Iwama et al., 2004). Heat shock proteins are a highly conserved groupof proteins found in a wide range of organisms from bacteria tohumans (Morimoto et al., 1990, 1992; Welch, 1993; Feder andHofmann, 1999), including fish (reviewed in Iwarma et al., 1998; Basuet al., 2002). With the exception of those of low molecular weight,such as HSP27, these proteins have constitutive functions in theunstressed cell (Morimoto et al., 1990; Hendrick and Hartl, 1993;Welch, 1993; Fink and Goto, 1998). However, various heat shockproteins are up-regulated in response to a wide variety of stressors, asthey have a role in repair and degradation of misfolded or denaturedproteins (Welch, 1993; Freeman et al., 1999; Rabergh et al., 2000).Increased levels of HSP70 have been observed in salmon fed soy as areplacement for fishmeal (Bakke-McKellep et al., 2007; Sagstad et al.,2008).

The aim of the current study was to assess whether GM RRS® soyaffects organ histomorphology and cellular stress response in Atlanticsalmon differently than near-isogenic non-GM soy, focusing oncommonly observed SBM-induced effects in salmon. This work

completes the evaluation of a 7-month feeding trial, where assess-ments of growth, body composition, organ sizes, haematology, plasmachemistry, lysozyme levels and performance through the parr-smolttransformation (Sissener et al., 2009) and liver proteome analysis(Sissener et al., in press), have been conducted to compare RRS® GMsoy and its near-isogenic line as diet ingredients for Atlantic salmon.

2. Materials and methods

2.1. Experimental design and sampling

The 7-month feeding trial was conducted at the Institute of MarineResearch (Matredal, Norway), and was approved by the NationalAnimal Research Authority in Norway. Atlantic salmon with an initialaverage weight of 39.7 g (SD 4.4) were fed two different diets, withfour replicate tanks of 120 fish for each diet group. Roundup Ready®

soy was used in one diet (GM) and its near-isogenic, non-modifiedmaternal line in the other (nGM). Both lines of soybeans weresupplied by the Monsanto Company, St. Louis, MO, USA. The dietswere compositionally similar in major nutrients (Table 1), and in bothdiets full fat soybean meal (FFSBM) provided 21% of the total protein.Further information on feed analysis, growth data, light regime andother details regarding fish husbandry is given elsewhere (Sisseneret al., 2009). The feeding trial was initiated the 11th of August, and forthe work presented in this paper samples were collected 13th ofDecember (sampling 1), 2nd of February (sampling 2) and 28th ofFebruary (sampling 3). The fish were transferred from freshwater toseawater the day after sampling 1, as they were going through theparr-smolt transformation, which is a natural part of the salmon lifecycle. Histological evaluation of the mid and distal intestine, spleen,head kidney and liver before seawater transfer (sampling 1) are

103N.H. Sissener et al. / Aquaculture 298 (2009) 101–110

reported herein, but all other results reported in this paper, includingfurther histological evaluation, were from samples taken during theseawater phase.

A cross-over of the diet groups was implemented after the secondsampling, with the intent to investigate the reversibility of potentialdiet effects. Twenty fish from each tank were anaesthetised(Benzoak® VET, 0.15 ml/L, ACD Pharmaceuticals, Leknes, Norway),marked by fin-clipping of the adipose fin to distinguish them from theother fish, and moved to a tank containing the opposite diet group, sothat cross-over fish and fish kept on the original diet groups were keptin the same tanks. The four fish groups—fish fed the GM diet through-out the trial (GM), fish fed the non-GM diet throughout the trial(nGM), fish switched from the GM to the non-GMdiet (GM-nGM) andfish switched from the non-GM to the GM diet (nGM-GM)—werefed the respective diets for three weeks before the third and finalsampling.

Furthermore, a “stress test” was conducted at the final sampling,but only on fish that had been fed the same diets throughout the trial(the nGM and GM groups). For this purpose, three replicate tanks perdiet were used, due to limitations in the number of fish available. Tostress the fish, the water level was greatly reduced in the tanks, andthe 12–13 fish in each tank were exposed to handling when theywerenetted out and left in a 10-L bucket. Further stress was presumablyinflicted by crowding. The fish were subjected to this treatment for5 min before they were transferred back into the tank and the waterlevel was restored.

All sampled fish were immobilized by a blow to the head,measured (fork length) and weighed. Samples for histologicalevaluation were collected from two fish per tank at all three samplingpoints. At the sampling following the cross-over, only fish that hadbeen exposed to a change in diet—nGM to GM and GM to nGM—weresampled for histology. Tissue pieces approximately 0.5×0.5 cm fromthemid and distal intestine, spleen, head kidney and liver were placedin 4% buffered formaldehyde solution. At the final sampling, sixrandomly selected fish per tank (all four groups) were sampled forblood, and liver, spleen, head kidney and the different regions of theintestine which were weighed for the calculation of organ indices.Prior to weighing, the intestinal tract was cleaned of visceral fat andluminal content, and then divided into proximal (the part of theintestine from which the pyloric caeca originate), mid and distalintestine. This also served as the “before stress”-sampling point.Before, 3 and 22 h following the stress test, blood samples and tissuefrom liver and distal intestine for mRNA analysis were collected fromfive fish per tank. The tissues were quickly dissected out, flash frozenin liquid nitrogen, and stored at −80 °C until analysis.

The fish were not fasted before the samplings as this reducespotential SBM-induced inflammation reactions in the intestine(Baeverfjord and Krogdahl, 1996), as well as change other anatomicaland physiological parameters (Krogdahl and Bakke-McKellep, 2005).Fish were not fed after the stress test, as it was assumed that theywould not eat.

2.2. Histology

Tissue samples for histological evaluation were fixed in 4%buffered formalin for 24 h and subsequently stored in 70% ethanolat 4 °C until further processing and embedding. The tissues weresubsequently dehydrated according to standard histological techni-ques in a graded ethanol series, ending with absolute ethanol.Sections of 5 μm were cut and stained with haematoxylin and eosin(H and E) before blinded examination under a lightmicroscope. Organmorphology was evaluated according to Amin et al. (1992). Distalintestinal morphology was scored according to the criteria previouslydescribed in Atlantic salmon with SBM-induced enteritis (Baeverfjordand Krogdahl, 1996): (1) widening and shortening of the intestinalfolds; (2) loss of the supranuclear vacuolization in the absorptive cells

(enterocytes) of the intestinal epithelium; (3) cellular infiltration of amixed leucocyte population in the central lamina propria within theintestinal folds as well as in the submucosa. A graded examinationscheme was used for these various characteristics: 5=normal;morphological changes from weak to severe were scored from 4 to 1.

2.3. Haematology, plasma enzymes and nutrients

Blood was drawn from the caudal vein (Vena caudalis) by meansof a heparinized medical syringe. Blood samples were divided so thatabout 200 μl were kept as individual samples for haematology, whilethe remainder was centrifuged at 3000×g for 10 min to obtain theplasma fraction, which was pooled for each tank, flash frozen inliquid nitrogen and stored at −80 °C. Hematocrit (Hct) wasimmediately measured using Vitex Pari micro-haematocrit tubeson a Hettich centrifuge (type 201424, GMI Inc, MI, USA) at13,000 rpm for 5 min. Red blood cells (RBC) and haemoglobin (Hb)were measured on a Cell-Dyn 400 (Sequoia-Turner) according to themanufacturer's instructions, using Para 12 control blood (Streck,Omaha, NE, USA) for calibration. The indices mean cell volume(MCV), mean cell haemoglobin (MCH) and mean cell haemoglobinconcentration (MCHC) were calculated according to Sandnes et al.(1988). Plasma samples were analysed for the enzymes lactatedehydrogenase (LDH), alanine aminotransferase (ALAT), and aspar-tate aminotransferase (ASAT) and for the nutrients glucose, totalprotein and triacylglycerol (TAG). These analyses were performed onMaxmat Biomedical Analyser (SM1167, Maxmat S.A., France), usingMaxmat reagents and the appropriate calibrators and controls for thedifferent methods.

2.4. mRNA transcription

Total RNA was purified from frozen liver and distal intestine usingthe EZ1 RNAUniversal Tissue Kit on the BioRobot® EZ1 (Qiagen, Hilden,Germany), including the optional DNase treatment step in the protocol.Homogenisation in QIAzol lysis reagent from the kit was performed onthe bead grinder homogeniser Precellys 24 (Bertin technologies,Montigny-le-Bretonneux, France) for 3×10 s at 6000 rpm for the liversamples and 2×30 s at 6500 rpm for the intestinal samples. RNAquantity and quality were assessed with Nanodrop® ND-1000 UV–Visspectrophotometer (NanoDrop Technologies, Wilmington, USA) andthe Agilent 2100 Bioanalyzer with the RNA 6000 Nano LabChip® kit(both Agilent Technologies, Pao Alto, USA). All analysed samples hadRIN (RNA integrity) values N8, with the majority being N9. Primers andTaqMan probes for the selected reference genes have been used inrecent studies; EF1α (elongation factor 1-alpha) (Moore et al., 2005),beta-actin (Olsvik et al., 2005) andARP (acidic ribosomal phosphoprotein)(Hevrøy et al., 2007). The primers and probe for HSP70 were used bySagstad et al. (2008), while those forHSP27were designed by Pål Olsvik(unpublished). Sequences of the primers (Invitrogen) and probes(Applied Biosystems, Foster City, USA) are given in Table 2.

Constant amounts of 250 ng RNA were reversely transcribed tocDNA on a GeneAmp® PCR 9700machine, using the TaqMan® ReverseTranscriptase kit with oligo(dT) primers (all from Applied Biosys-tems) in 50μL reactions. Each sample was run in duplicate, and a totalof three 96-well plates were used per organ. The plates also containeda five point dilution curve in triplicate, non-template and non-amplification controls, as well as three samples for inter-platecalibration. For real-time PCR, SYBR® Green I Mastermix (RocheApplied Science, Indianapolis, USA), primers, probe and cDNA weremixed in 96-well plates manually for all the liver samples and with aBiomek®3000 Laboratory automation workstation (Beckman Coulter,Fullerton, USA) for the intestinal samples. Thermal cycling wasperformed on a LightCycler® 480 System (Roche) for 40cycles of10 s denaturation at 95 °C and 30 s annealing at 60 °C.

Table 2Primer and probe sequences and the GenBank accession number from which they were designed.

Gene Forward primer Reverse primer Probe Acc.no.

β-actin CCAAAGCCAACAGGGAGAAG AGGGACAACACTGCCTGGAT TGACCCAGATCATGTTT BG933897Ef1α CCCCTCCAGGACGTTTACAAA CACACGGCCCACAGGTACA ATCGGTGGTATTGGAAC AF321836ARP GAAAATCATCCAATTGCTGGATG CTTCCCACGCAAGGACAGA CTATCCCAAATGTTTCATTGTCGGCGC AY255630HSP27 CCAGCTGCCTGAGGATGTG CCTCGGTGCCCAATGATG ACCCCACCTCTGTGACA CV428908HSP70 CCCCTGTCCCTGGGTATTG CACCAGGCTGGTTGTCTGAGT CGCTGGAGGTGTCATG BG933934

Elongation factor 1α, Ef1α; acidic ribosomal phosphoprotein, ARP; heat shock protein, HSP.

104 N.H. Sissener et al. / Aquaculture 298 (2009) 101–110

Cycle threshold (Ct)-values were calculated using the secondmaximum derivative method in the Lightcycler® software. Amplifica-tion efficiency was determined using 2-fold dilution curves with theformula E=10^(−1/slope), with the slope of the linear curve of Ct-values plotted against the log-dilution (Higuchi et al., 1993). Thestability of the reference genes was evaluated using geNormVBA applet(Vandesompele et al., 2002) and NormFinder (Andersen et al., 2004).

2.5. Statistics

Statistical analyses were performed using Statistica™ 8.0 software(Statsoft, Inc. Tulsa, USA). Data fulfilling the assumptions ofhomogeneity of variance and normal distribution (tested by Levene'sand Kolmogorov Smirnov tests respectively) were subjected toparametrical statistics. Nested (hierarchical) ANOVA was used forparameters measured on individual fish (nested within theirrespective tanks), with tank as random effect and diet as fixed effect.Tank-to-tank variation was used as the error term as individual fishfrom the same tank were considered pseudoreplicates, and errordegrees of freedom were computed using the Satterthwaite method.Conventional ANOVA was used for measurements on pooled plasmasamples from each tank, and also for the histology scores. For thelatter, ANOVA was carried out on individual values within samplingtimes, as well as collectively across the three samplings (two-wayANOVA) with sampling time and diet as the two factors. Individualvalues rather than tankmeans were used as Baeverfjord and Krogdahl(1996) determined that histological characteristics of SBM-inducedenteritis did not vary between tanks in feeding salmon. The criticallevel for significance was set at 0.05. However, all p-values below 0.10are given in the tables.

A permutation test was used on themRNA transcription data usingREST©2008 (Relative expression software tool) with 5000 randomi-zations. Efficiency correction and normalization to the referencegenes were done within the REST software, with the average Ct-valuefor each sample as the input variable.

Table 3Mean scores for the morphological changes of distal intestine and liver of Atlantic salmon f

Sampling Diet (DI) MFH (DI) MFF (DI) LW (DI) LPC (D

1 (Dec 13th) nGM 2.6 4.3 3.7 4.1 2.GM 2.7 4.0 4.0 3.8 2.Pooled SE 0.14 0.18 0.13 0.19 0.P value 0.764 0.230 0.118 0.274 1.

2 (Feb 2nd) nGM 3.7 a 3.8 3.9 3.2 2.GM 2.9 b 3.4 3.9 3.4 2.Pooled SE 0.22 0.34 0.11 0.32 0.P value 0.020 0.523 1.000 0.585 0.

3 (Feb 28th) GM to nGM 4.1 4.3 a 3.3 3.6 2.nGM to GM 3.9 3.1 b 3.3 3.3 2.Pooled SE 0.20 0.37 0.15 0.26 0.P value 0.523 0.031 0.770 0.319 0.

For the intestinal scores, lower scores indicate more severe changes (5=normal).Distal intestine, DI; Mucosal fold height, MFH; Mucosal fold fusion, MFF; Lamina width, LW;Enterocyte vacuolization, EV; enterocyte nucleus position, ENP; mitotic figure(frequency),a,bMean values within a column with unlike superscript letters were significantly different

2.6. Calculations

Mean cell volume : MCV = ðHct= rbcÞ⁎10

Mean cell haemoglobin : MCH = ðHb= rbcÞ⁎10

Mean cell haemoglobin concentration : MCHC = ðHb=HctÞ⁎100

3. Results

No morphological changes were observed in mid intestine or headkidney of fish fed any of the diets at any of the sampling times. In thedistal intestine, moderate to pronounced soybean meal-inducedchanges were observed in all fish (Tables 3 and 4). Changes were notrestricted to a single diet group or sampling point, although the degreedid differ somewhat (see below). The changes included shortening ofheights of both simple and complex mucosal folds, fusing of mucosalfolds, widening of lamina propria and submucosa with infiltration ofinflammatory cells, increased number of apoptotic bodies and mitoticfigures, decreaseddegreeof enterocyte vacuolization, shift of enterocytenuclei from basal to a more apical position, as well as large cyst-likestructures within or between enterocytes at the apical part of theintestinal folds (Fig. 1). These cyst-like structures stained positively forcellulose/chitin and were assumed to be microsporidia.

At the sampling before seawater transfer (sampling 1 [Dec 13th] inTable 3), there were no significant differences between the dietgroups in morphological changes in the distal intestine. At thesampling before cross-over (sampling 2 [Feb 2nd] in Table 3), mucosalfold heights in the distal intestine of fish fed the GM diet weresignificantly shorter than in those fed the non-GM diet (P=0.02).After diet cross-over (sampling 3 [Feb 28th] in Table 3), the degree ofmucosal fold fusion (Fig. 2) in fish fed the GM diet was morepronounced than in fish fed the non-GM diet (P=0.031). Consideringall sampling time points together (Table 4), the degree of mucosal fold

ed nGM diet or GM diet at each sampling point.

I) SmC (DI) SmW (DI) EV (DI) ENP (DI) MF (DI) AB (LI) DG

8 4.8 1.1 3.4 4.8 4.4 2.58 4.8 1.0 3.3 4.9 4.3 1.819 0.09 0.04 0.15 0.15 0.19 0.29000 1.000 0.334 0.577 0.554 0.642 0.0906 4.1 1.2 3.1 3.4 3.1 2.04 3.9 1.5 3.1 3.9 3.7 2.324 0.21 0.20 0.12 0.36 0.42 0.35467 0.544 0.287 0.717 0.346 0.310 0.6199 3.3 1.8 3.1 4.6 4.3 3.0 a

8 3.3 1.5 3.1 4.9 4.3 1.8 b

33 0.24 0.31 0.22 0.16 0.09 0.37794 0.857 0.492 1.000 0.278 0.642 0.032

Lamina propria cellularity, LPC; Submucosa cellularity, SmC; Submucosa width, SmW;MF; Apoptotic bodies(frequency), AB; Liver, LI; Deposit of glycogen, DG.(Pb0.05).

Table 4Statistical analysis of data from the histological evaluation of distal intestine, liver andspleen in Atlantic salmon pooled according to sampling points (both diet groupstogether) and according to the GM and nGM diet groups (all sampling points together).

Sampling Diet (DI)MFH

(DI)MF

(DI)AB

(DI)MFF

(LI)DG

(SP)MMCs

1 (Dec 13th) 2.7c 4.8a 4.3a 4.7a

2 (Feb 2nd) 3.3b 3.6b 3.4b 3.8b

3 (Feb 28th) 4.0a 4.8a 4.3a 4.5a

Pooled SE 0.14 0.17 0.19 0.23P value b0.0001 b0.0001 0.001 0.013

nGM 4.1a 2.5a

GM 3.5b 1.9b

Pooled SE 0.18 0.20P value 0.017 0.047

Distal intestine, DI; Mucosal fold height, MFH; Mitotic figure(frequency), MF; Apototicbodies(frequency), AB; Melanomacrophage centers, MMCs ; Mucosal fold fusion, MFF;Liver, LI; Deposit of glycogen, DG; Spleen, SP; Melanomacrophage centre, MMC.a,b,c Mean values within a column with unlike superscript letters were significantlydifferent (Pb0.05).

105N.H. Sissener et al. / Aquaculture 298 (2009) 101–110

fusions in the distal intestine of fish fed the GM diet was significantlymore pronounced than in fish fed the non-GM diet (P=0.017).Sampling time had a significant effect on mucosal fold heights in thedistal intestine, with heights increasing from the first to the thirdsampling. Moreover, more mitotic figures and apoptotic bodies wereobserved in the distal intestine at the second sampling than aftercross-over or before seawater transfer (Pb0.0001 and P=0.001,respectively; Table 4), but there were no differences detectedbetween the diet groups. In addition, the large cyst-like structureswithin or between enterocytes at the apical part of the intestinal folds(Fig. 1), were significantly more numerous at the first and secondsamplings than at the final sampling after cross-over, but there was nosignificant effect of diet. These cyst-like structures are assumed to becaused by a water-borne infection or vertical transmission of anunknown species of microsporidia.

Apart from the degree of glycogen deposition within hepatocytes,no obvious structural changes were observed in liver. A significantlylower degree of glycogen deposition was observed in fish fed the GMdiet compared to those fed the non-GM diet at the final samplingpoint (P=0.032; Table 3 and Fig. 4). Moreover, when data from the

Fig. 1. Cyst-like structures in distal intestine (arrow heads), more prominent at the first angroups. Compared with normal globet cell (a), cyst-like structures filled with mucin-like ma(H&E.×400).

three sampling points were evaluated collectively, the deposits ofglycogen in hepatocytes in the GM diet group were lower than in thenon-GM diet group (P=0.047; Table 4), with no effect of time.

Melanomacrophage (MM) is a characteristic immune cell type ofteleosts. MMs or melanomacrophage centers (MMCs) are prevalent inspleen, but also found in head kidney and liver. MMCs in spleen weresignificantly more numerous and larger at the second sampling thanthose at the first and final samplings (P=0.013), but no differencesbetween diet groups were detected (Table 4, Fig. 3). Apart from theMMCs, no other morphological differences were observed in thespleen.

When the organ indices, blood- and plasma parameters from afterthe cross-over were analysed (all four groups), there were systematicdifferences between fish that had been fin clipped and switched to theopposite diet group and fish that had not (Table 5). These differenceswere unrelated to GM content in the new diet and which diet the fishhad been fed previously, thus the remainder of this paragraphdescribes differences not related to whether the soy in the diet wasGM or not. Despite no differences in growth, the relative sizes(corrected for body weight) of head kidney and proximal- and mid-intestine were significantly lower in the fish that had been crossed-over to the opposite diet group, compared to those that had remainedin the same diet group throughout the trial. Furthermore, RBCnumbers were significantly lower (P=0.01) in the crossed-overfish, while haemoglobin and MCHC (both P=0.09) also showedtendencies to be lower. MCV (P=0.009) and MCH (P=0.005) werehigher in the crossed-over fish. This indicates that the crossed-overfish had fewer but larger red blood cells, with higher total levels ofhaemoglobin in each cell, thereby maintaining haematocrit. However,there was a lower haemoglobin concentration in these larger cells,giving lower haemoglobin in the blood in total. Plasma glucose wasalso higher in the cross-over fish (P=0.03), and there was a similartendency in plasma protein concentration (P=0.10). There were nodifferences detected in the analysed plasma enzymes, but these hadhigh variability.

Likewise, haematological parameters from the stress test exhibitedno differences between the non-GM and GM fed groups, neither inbasal levels nor in the response to stress (Table 6). However, changes(which were similar in both diet groups) were apparent whencomparing the data from before and the two time-points after thestress test. The number of red blood cells (P=0.03), Hct (Pb0.0001)

d second sampling compared to the final sampling, but not different between the dietterial and the nucleus were pushing to the periphery place (b). G, goblet cell; N, nuclei.

Fig. 2. Morphological changes in distal intestine of Atlantic salmon. Mucosal fold height in fish fed with non-GM diet (a) was higher than those fed the GM diet (b) at the secondsampling point. The degree of mucosal fold fusion (MFF) (arrow head) of fish fed the non-GM diet (c) was less severe than those fed the GM diet (d) at the final sampling point.(H&E.×100).

106 N.H. Sissener et al. / Aquaculture 298 (2009) 101–110

and Hb (P=0.003) all decreased following stress, while MCV, MCHand MCHC remained similar to the initial values. This indicates thatchanges in Hct and Hb were due to the reduction in RBC numbers, asthe size of the red blood cells as well as their haemoglobin contentremained the same. For Hct and Hb, the status before stress wassignificantly different from the values both 3 and 22 h after, whilethese two latter values were not significantly different from eachother. RBC before stress was significantly different from 3 h afterstress, while 22 h after the value was intermediate and notsignificantly different from either. Plasma TAG (P=0.0001) andprotein (P=0.002) were reduced after stress. The protein concentra-tion went down by 14% from before to 3 h after stress, and remainedat a similar level 22 h after. The TAG concentration also decreasedfrom before to 3 h after stress, and was reduced even further at 22 h

Fig. 3.Morphological changes in spleen of Atlantic salmon. The number and size of melanomincreased compared to the other two sampling points (a), but there were no differences be

after, a total reduction of 60%. No effects of stress were detected on theanalysed plasma enzymes.

The mRNA transcription of HSP27 and HSP70 in liver and distalintestine did not differ between the two diet groups at any individualsampling point; this was also the case when all three time points wereconsidered together (Table 7). The transcription of HSP70 in both liverand distal intestine remained stable at all three sampling points inboth diet groups, while the transcription of HSP27 exhibited similarchanges after stress in both diet groups. Transcription of HSP27 wasup-regulated by a factor of 1.30 in the liver, whereas in the distalintestine transcription was down-regulated by a factor of 0.74. In bothcases, the samples collected before stress were significantly differentfrom those collected after 3 and 22h, while these two latter samplingsdid not differ.

acrophage centers (MMCs) in spleen at the second sampling point (b) were significantlytween the diet groups. (H&E.×50).

Fig. 4. Deposits of glycogen in hepatocytes were significantly lower in Atlantic salmon fed the GM diet (b) compared to the non-GM diet (a) at the final sampling and overall in thetrial. However, structure of the salmon liver remained normal during whole experiment in both diet groups. (H&E.×100).

107N.H. Sissener et al. / Aquaculture 298 (2009) 101–110

4. Discussion

Feeding salmonwith a high inclusion level (25%) of GM soy causedno apparent differences in growth, mortality or measured healthparameters (haematology, plasma enzymes, lysozyme levels anddifferential count of white blood cells) previously reported from thistrial (Sissener et al., 2009). However, dietary inclusion level of FFSBMdid lead to an inflammatory response in the distal intestine asdescribed previously in Atlantic salmon (Van den Ingh et al., 1991;Baeverfjord and Krogdahl, 1996; Van den Ingh et al., 1996; Nordrumetal., 2000; Bakke-McKellep et al., 2007). The observation that mucosalfold heightswere shorter (significantly so at sampling 2 [Feb 2nd]) anddegree of mucosal fold fusion more pronounced in fish fed GMcompared to non-GM soy may indicate that the GM variety elevatedthe degree of inflammation.

Table 5Cross-over effects.

nGM GM

Weight (g) 187.4 (8.2) 191.6 (7.3)

Organ indicesSSI 0.052 (0.003) 0.051 (0.002)HSI 1.12 (0.07) 1.00 (0.02)H-KSI 0.198 (0.011) 0.186 (0.009)PISI 3.17 (0.18) 3.17 (0.08)MISI 0.270 (0.009) 0.237 (0.006)DISI 0.568 (0.018) 0.550 (0.017)

HaematologyHct (%) 36.2 (0.7) 38.7 (0.6)RBC (1012L−1) 1.14 (0.04) 1.25 (0.03)Hb (g 100 ml−1) 7.41 (0.17) 7.73 (0.15)MCV (1015L−1) 329 (15) 313 (8)MCH (10−6g) 66.4 (1.8) 62.3 (0.7)MCHC (g 100 ml−1) 20.6 (0.5) 20.05 (0.3)

PlasmaTAG (mMol) 2.25 (0.22) 2.45 (0.23)Protein (g L−1) 33.7 (0.8) 35.8 (0.8)Glucose (g L−1) 1.11 (0.06) 1.08 (0.02)ALAT (u L−1) 17.3 (2.0) 17.5 (5.3)ASAT (u L−1) 309 (32) 307 (10)LDH (u L−1) 984 (131) 944 (174)

Mean (SEM). 6 fish from the original diet groups and 6 cross-over fish were sampled fromSpleen somatic index, SSI; hepato-somatic index, HIS; head kidney somatic index H-KSI; prsomatic index, DISI; haematocrit, Hct; red blood cells, RBC; haemoglobin, Hb; mean cell volutriacylglycerol, TAG; lactate dehydrogenase, LDH; alanine aminotransferase, ALAT; aspartat

Differences in the degree of typical SBM-induced changes in thedistal intestine as well as differences in glycogen deposition in theliver between the non-GM and the GM fed salmon could be caused bydifferences in ANF concentrations in the two soy lines. These areknown to vary extensively between different strains of soy (OECD,2001) and can also vary between GM soy and its near-isogenicmaternal line (Sagstad et al., 2008). Atlantic salmon fed dietscontaining 20% SBM grown in six different locations showedsignificant variations in all evaluated SBM-induced changes in thedistal intestine (Uran et al., 2008), confirming that differencesbetween soy strains can affect the severity of enteritis. Nor candifferences in feed intake (not measured in this trial) be entirely ruledout as a contributing factor. Short term reduction of enteritis has beenshown when smolts were transferred to seawater, most likely due todecreased feed intake (Bakke-McKellep et al., 2006).

GM−NnGM nGM−NGM Cross-over effect

187.5 (5.7) 187.8 (5.2) n.s.

0.051 (0.001) 0.057 (0.002) n.s.1.08 (0.03) 1.09 (0.02) n.s.

0.178 (0.005) 0.177 (0.004) P=0.042.67 (0.05) 2.64 (0.07) P=0.001

0.207 (0.009) 0.208 (0.008) P=0.00070.544 (0.033) 0.511 (0.026) n.s.

37.9 (0.9) 37.7 (1.5) n.s.0.99 (0.03) 1.00 (0.04) P=0.017.01 (0.16) 7.12 (0.20) n.s., (P=0.09)386 (13) 385 (21) P=0.00971.3 (1.9) 71.5 (1.3) P=0.00518.6 (0.5) 19.3 (0.8) n.s., (P=0.09)

2.16 (0.17) 2.33 (0.16) n.s.37.0 (1.1) 36.4 (0.9) n.s., (P=0.10)1.15 (0.04) 1.13 (0.03) P=0.0318.2 (2.6) 11.5 (2.3) n.s.365 (37) 299 (29) n.s.

1114 (289) 930 (174) n.s.

each tank (24 fish for each of the four treatment groups).oximal intestine somatic index, PISI; mid intestine somatic index, MISI; distal intestineme, MCV; mean cell haemoglobin, MCH; mean cell haemoglobin concentration, MCHC;e aminotransferase, ASAT.

Table 6Haematology and plasma chemistry from the stress test.

Diet 0 h 3 h 22 h Diet Time

HaematologyHct (%) non-GM 36.2 (0.7) 30.9 (0.9) 32.8 (1.2) n.s. Pb0.0001

GM 38.7 (0.6) 31.6 (1.2) 33.4 (1.2)RBC (1012L−1) nGM 1.14 (0.04) 1.03 (0.02) 1.09 (0.04) n.s. P=0.03

GM 1.25 (0.03) 1.06 (0.03) 1.09 (0.03)Hb (g 100 ml−1) nGM 7.41 (0.17) 6.31 (0.14) 6.78 (0.25) n.s. P=0.003

GM 7.73 (0.15) 6.56 (0.21) 6.90 (0.20)MCV (1015L−1) nGM 329 (15) 299 (4) 304 (9) n.s. n.s.

GM 313 (8) 298 (8) 308 (12)MCH (10−6g) nGM 66.4 (1.8) 61.2 (0.8) 62.4 (0.8) n.s. n.s.

GM 62.3 (0.7) 61.8 (0.9) 63.4 (1.4)MCHC (g 100 ml−1) nGM 20.6 (0.6) 20.5 (0.4) 20.8 (0.6) n.s. n.s.

GM 20.05 (0.3) 20.9 (0.5) 20.9 (0.6)

PlasmaTAG (mMol) nGM 2.25 (0.22) 2.01 (0.25) 0.95 (0.08) n.s. P=0.0001

GM 2.45 (0.23) 1.45 (0.24) 0.92 (0.07)Protein (g L−1) nGM 33.7 (0.8) 30.5 (0.9) 30.7 (0.9) n.s. P=0.002

GM 35.8 (0.8) 29.03 (1.2) 31.8 (0.3)Glucose (g L−1) nGM 1.11 (0.06) 1.19 (0.05) 1.10 (0.03) n.s. n.s.

GM 1.08 (0.02) 1.14 (0.06) 1.00 (0.05)ALAT (uL−1) nGM 17.3 (2.3) 22.7 (5.9) 19.1 (5.2) n.s. n.s.

GM 17.5 (5.3) 20.5 (2.8) 25.7 (2.7)ASAT (uL−1) nGM 309 (32) 437 (77) 315 (49) n.s. n.s.

GM 307 (10) 291 (12) 432 (74)LDH (uL−1) nGM 984 (186) 2010 (972) 1125 (364) n.s.a n.s.a

GM 960 (246) 1433 (129) 1194 (14)

Mean (SEM). Haematocrit, Hct; red blood cells, RBC; haemoglobin, Hb; mean cell volume, MCV; mean cell haemoglobin, MCH; mean cell haemoglobin concentration, MCHC;triacylglycerol, TAG; lactate dehydrogenase, LDH; alanine aminotransferase, ALAT; aspartate aminotransferase, ASAT.

a LDH deviated from a normal distribution, thus was tested with Kruskal–Wallis ANOVA.

108 N.H. Sissener et al. / Aquaculture 298 (2009) 101–110

Hepatocytes of Atlantic salmon are large, regularly shaped cellswith typical central nuclei, and with moderate cytoplasmic glycogencontent. Our evaluation of salmon liver, with no changes apart fromreduced glycogen deposition in GM fed fish in the last samplingperiod, shows minor changes compared to the studies on micementioned in the introduction. Additionally, glycogen deposits in liverof mice fed GM and non-GM soy were not reported to be different(Malatesta et al., 2008b), thus there seems to be no similarities to ourresults. However, the maternal, near-isogenic line, grown undersimilar conditions with the GM variety, which is recommended as thebest possible control (ILSI, 2003; EFSA, 2006), was not used in themice studies, and the authors have shown that the effects might bedue to glyphosate residues rather than the feed ingredients being GM(Malatesta et al., 2008a).

Apart from histology of the distal intestine and glycogen depositsin the liver, no further differenceswere observed between the GM andnon-GM diet groups, consequently, the remainder of this discussiondeals with development over time, cross-over and stress-effects,which are discussed more briefly than the diet effects, as they are notthe main focus of this paper.

Table 7Transcription of heat shock protein mRNA in liver and distal intestine.

Diet 0 h 3 h 22 h Diet Time

HSP70, liver nGM 21.2 (0.5) 21.3 (0.7) 21.5 (1.0) n.s. n.s.GM 21.5 (1.2) 21.7 (0.7) 21.7 (0.7)

HSP70, intestine nGM 21.7 (0.5) 21.8 (0.8) 21.6 (0.9) n.s. n.s.GM 21.7 (1.4) 21.8 (0.7) 21.6 (0.5)

HSP27, liver nGM 33.3 (0.6) 32.9 (0.8) 32.9 (0.8) n.s. P=0.006GM 33.3 (0.4) 32.9 (0.6) 33.1 (1.1)

HSP27, intestine nGM 30.7 (0.8) 31.3 (0.9) 31.1 (0.7) n.s. P=0.002GM 30.8 (1.0) 31.3 (0.6) 31.0 (0.5)

Median Ct-values with the interquartile range in parenthesis.Heat shock protein, HSP.

Development of the fish over time had a greater influence onhistological parameters than diet, as observed for other parametersreported from the same feeding trial (Sissener et al., 2009). Mucosalfold height (MFH) appeared to develop with age and/or body size, andpossibly as a result of transfer from freshwater to seawater.Proliferation (mitotic figures, MF) and apoptosis in distal intestineand the number and size of MMCs in the spleen showed a transitionalincrease in frequency at the second sampling, just after seawatertransfer, in both diet groups. Furthermore, more of the large, cyst-likestructures within or between enterocytes at the apical part of theintestinal fold were observed just before and just after seawatertransfer compared to the final sampling. These differences might beexplained by effects of the parr-smolt transformation and seawatertransfer.

The cross-over design turned out to introduce bias that limits thevalue of the results regarding drawing conclusions on diet effects. Thedifferences between crossed-over fish and fish that were not movedmay have been due to use of anaesthetic, fin-clipping, moving of fishbetween tanks, formation of new fish groups with hierarchicalchanges or a combination of these factors. In a previous study fin-clipping of the adipose did not appear to affect growth or survival ofrainbow trout (Gjerde and Refstie, 1988), while anaesthesia has beenfound to induce stress response in fish (Ortuño et al., 2002; Olsviket al., 2007). Despite its limitations for the original purpose, our cross-over procedure may serve as an additional stress test, with fishshowing similar responses regardless of being crossed-over from non-GM or GM soy diets.

When the haematology results from the cross-over fish (Table 5)were compared to those from the stress test (Table 6), differencesbetween themmay be due to the length of time thefishwere exposed tostress. It seemed that theRBCnumberwas generally decreasedby stress.However, the long term stress putatively experienced by the fish in thecross-over, appeared to lead to either 1) an adaptation with increasedsize andhaemoglobin content of their red blood cells or 2) impairedRBCproduction, since the large average size could indicate fewer immature

109N.H. Sissener et al. / Aquaculture 298 (2009) 101–110

cells, which are generally smaller and contain less haemoglobin thanmature ones (Sandnes et al., 1988). The fish from the stress test had notresponded in this manner and were affected by lowered haemoglobinvalues, whichmight affect oxygen uptake. It seems that stress can affectfish haematology in different ways depending on stressor and possiblyalso on fish species. In sea bass (Dicentrarchus labrax) thermal andphysical stress gave increased haematocrit levels, while chemical stressgave reduced levels, although both resulted in increased plasma cortisol(Roche and Bogé, 1996).

The fact that there was no response in HSP70 transcription afterstress treatment, while HSP27 was regulated in opposite directions inthe liver and distal intestine could be explained by the different rolesof these two HSPs and of the two organs. The main function of HSP27is to protect against protein aggregation by promoting refolding ofdenatured proteins (Garrido et al., 2001), while HSP70 functions inthe folding of novel polypeptide chains, and also mediates the repairor degradation of altered or denatured proteins (Kiang and Tsokos,1998). High molecular weight HSPs are ATP-dependent chaperones,whereas low molecular weight HSPs such as HSP27 are ATP-independent (Parcellier et al., 2003). Varying responses betweendifferent HSPs and different tissues in fish has also been observed byothers (Lele et al., 1997; Stephensen et al., 2002).

5. Conclusion

Effects of the GM Roundup Ready® soy diet compared to non-GMsoy were observed as shorter mucosal fold height in the distalintestine at one sampling point out of three, while mucosal fold fusionwas more pronounced in the GM group overall in the trial, whichcould indicate more pronounced soy-induced changes. The inflam-mation observed in both diet groups was consistent with what isnormally observed in Atlantic salmon fed such high inclusion levels ofSBM. Also, there was lower glycogen deposition in liver of fish fed theGM diet. No other diet-related morphological differences weredetected in any organs, nor were there any differences in themeasured stress parameters between the two diet groups. Thus, GMsoy at an inclusion level of 25% did not appear to cause any adverseeffect of practical importance on organ morphology or stress responsecompared to non-GM soy.

Acknowledgements

We would like to thank The Monsanto Company for kindlysupplying the RRS® as well as the near-isogenic maternal soybeans.We also thank Ivar Helge Matre for skilled technical assistance andLise Dyrhovden for care of the fish during the feeding trial. The projectwas supported by the Norwegian Research Council, grant no. 172151.

References

Amin, A.B., Mortensen, L., Poppe, T.T., 1992. Histology Atlas: Normal Structure ofSalmonids: A Colour Atlas. Akvapatologisk Laboratorium AS, Bodø, Norway.

Andersen, C.L., Jensen, J.L., Orntoft, T.F., 2004. Normalization of real-time quantitativereverse transcription-PCR data: a model-based variance estimation approach toidentify genes suited for normalization, applied to bladder and colon cancer datasets. Cancer Res. 64, 5245–5250.

Baeverfjord, G., Krogdahl, Å., 1996. Development and regression of soybean mealinduced enteritis in Atlantic salmon, Salmo salar L., distal intestine: a comparisonwith the intestines of fasted fish. J. Fish Dis. 19, 375–387.

Bakke-McKellep, A.M., Refstie, S., Stefansson, S.O., Vanthanouvong, V., Hemre, G.-I.,Kroghdahl, Å., 2006. Effects of dietary soybean meal and photoperiod cycle onosmoregulation following seawater exposure in Atlantic salmon smolts. J. Fish Biol.69, 1396–1426.

Bakke-McKellep, A.M., Penn, M.H., Salas, P.M., Refstie, S., Sperstad, S., Landsverk, T.,Ringo, E., Krogdahl, Å., 2007. Effects of dietary soyabean meal, inulin andoxytetracycline on intestinal microbiota and epithelial cell stress, apoptosis andproliferation in the teleost Atlantic salmon (Salmo salar L.). Br. J. Nutr. 97, 699–713.

Barton,B.A., Iwama,G.K., 1991.Physiological changes infish fromstress in aquaculturewithemphasis on the response and effects of corticosteroids. Annu. Rev. Fish Dis. 1, 3–26.

Basu, N., Todgham, A.E., Ackerman, P.A., Bibeau, M.R., Nakano, K., Schulte, P.M., Iwama,G.K., 2002. Heat shock protein genes and their functional significance in fish. Gene295, 173–183.

Benninghoff, A.D., Williams, D.E., 2008. Identification of a transcriptional fingerprint ofestrogen exposure in rainbow trout liver. Toxicol. Sci. 101, 65–80.

Cellini, F., Chesson, A., Colquhoun, I., Constable, A., Davies, H.V., Engel, K.H., Gatehouse,A.M.R., Karenlampi, S., Kok, E.J., Leguay, J.J., 2004. Unintended effects and theirdetection in genetically modified crops. Food Chem. Toxicol. 42, 1089–1125.

Domingo, J.L., 2007. Toxicity studies of genetically modified plants: a review of thepublished literature. Crit. Rev. Food Sci. Nutr. 47, 721–733.

EFSA, 2006. Guidance document of the Scientific Panel on Genetically ModifiedOrganisms for the risk assessment of genetically modified plants and derived foodand feed. in: EFSA (Ed.). The EFSA Journal, pp. 1-100.

EFSA, 2008. Report of the EFSA GMO Panel Working Group on Animal Feeding Trials:safety and nutritional assessment of GM plants and derived food and feed: the roleof animal feeding trials. Food Chem. Toxicol. 46, 1–70.

Evans, J.J., Pasnik, D.J., Peres, H., Lim, C., Klesius, P.H., 2005. No apparent differences inintestinal histology of channel catfish (Ictalurus punctatus) fed heat-treated andnon-heat-treated raw soybean meal. Aqua Nutr. 11, 123–129.

Feder,M.E., Hofmann, G.E., 1999. Heat shock proteins,molecular chaperons, and the stressresponse: evolutionary and ecological physiology. Annu. Rev. Physiol. 61, 243–282.

Fink, A.I., Goto, Y., 1998. Molecular Chaperones in the Life Cycle of Proteins: Structure,Function and Mode of Action. Mercel Dekker, New York.

Flachowsky, G., Aulrich, K., Böhme, H., Halle, I., 2007. Studies on feeds from geneticallymodified plants (GMP)—contributions to nutritional and safety assessment. Anim.Feed Sci. Technol. 133, 2–30.

Freeman,M.E., Borrelli,M.J.,Meredith,M.J., Lepock, J.R., 1999. On the path to the heat shockresponse: destabilization and formation of partially folded protein intermediates, aconsequence of protein thiol modifications. Free Radic. Biol. Med. 26, 737–745.

Garrido, C., Gurbuxani, S., Ravagnan, L., Kroemer, G., 2001. Heat shock proteins:endogenous modulators of apoptotic cell death. Biochem. Biophys. Res. Commum.286, 433–442.

Gatlin, D.M., Barrows, F.T., Brown, P., Dabrowski, K., Gaylord, T.G., Hardy, R.W., Herman, E.,Hu,G., Krogdahl, Å.,Nelson, R., Overturf, K., Rust,M., Sealey,W., Skonberg, D., Souza, E.J.,Stone, D., Wilson, R., Wurtele, E., 2007. Expanding the utilization of sustainable plantproducts in aquafeeds: a review. Aquac. Res. 38, 551–579.

Gjerde, B., Refstie, T., 1988. The effect of fin-clipping on growth rate survival and sexualmaturity of rainbow trout. Aquaculture 73, 383–389.

Hendrick, J.P., Hartl, F.U., 1993. Molecular chaperone functions of heat shock proteins.Annu. Rev. Biochem. 62, 349–384.

Hevrøy, E.M., El-Mowafi, A., Taylor, R.G., Olsvik, P.A., Norberg, B., Espe, M., 2007. Lysineintake affects gene expression of anabolic hormones in Atlantic salmon, Salmosalar. Gen. Comp. Endocrinol. 152, 39–46.

Higuchi, R., Fockler, C., Dollinger, G., Watson, R., 1993. Kinetic PCR analysis: real-timemonitoring of DNA amplification reactions. Nat. Biotech. 11, 1026–1030.

ILSI, 2003. Best practices for the conduct of animal studies to evaluate crops geneticallymodified for output traits, International Food Biotechnology Committee, InternationalLife Sciences Institute, Washington DC. http://www.ilsi.org/NR/rdonlyres/4A2F7C13-B14AA-14BC19-AEDE-696B694B672E693C694/690/BestPracticesGuidelines.pdf.

Iwama, G.K., Afonso, L.O.B., Todgham, A., Ackerman, P., Nakano, K., 2004. Are hspssuitable for indicating stressed states in fish? J. Exp. Biol. 207, 15–19.

Iwarma, G.K., Thomas, P.T., Forsyth, R.B., Vijayan, M.M., 1998. Heat shock proteinexpression in fish. Rev. Fish Biol. Fish 8, 35–36.

James, C., 2007. Executive summary: global status of Commercialized Biotech/GMcrops. ISAAA Brief 37. ISAAA, Ithaca, NY.

Kiang, J.G., Tsokos, G.C., 1998.Heat shockproteins 70 kDa:molecular biology, biochemistryand physiology. Pharmacol. Ther. 80, 183–201.

Krogdahl, Å., Bakke-McKellep, A.M., 2005. Fasting and refeeding cause rapid changes inintestinal tissue mass and digestive enzyme capacities of Atlantic salmon (Salmosalar L.). Comp. Biochem. Physiol. A Mol. Integr. Physiol. 141, 450–460.

Kuiper, H.A., Kleter, G.A., 2003. The scientific basis for risk assessment and regulation ofgenetically modified foods. Trends Food Sci. Technol. 14, 277–293.

Latham, J., Wilson, A.K., Steinbrecher, R.A., 2006. The mutational consequences of planttransformation. J. Biomed. Biotechnol. 25376, 25371–25377.

Lele, Z., Engel, S., Krone, P.H., 1997. hsp47 and hsp70 gene expression is differentiallyregulated in a stress- and tissue specific manner in zebrafish embryos. Dev. Genet.21, 123–133.

Malatesta, M., Caporaloni, C., Gavaudan, S., Rocchi, M.B.L., Serafini, S., Tiberi, C.,Gazzanelli, G., 2002. Ultrastructural morphometrical and immunocytochemicalanalyses of hepatocyte nuclei from mice fed on genetically modified soybean. CellStruct. Funct. 27, 173–180.

Malatesta, M., Biggiogera, M., Manuali, E., Rocchi, M.B., Baldelli, B., GAzzanelli, G., 2003.Fine structural analysis of pancreatic acinar cell nuclei frommice fed on geneticallymodified soybean. Eur. J. Histochem. 47, 358-358.

Malatesta, M., Baldelli, B., Battistelli, S., Manuali, E., Biggiogera, M., 2005. Reversibility ofnuclear modifications in mice fed on genetically modified soybean. Eur. J. Histol. 49,237–242.

Malatesta, M., Perdoni, F., Santin, G., Battistelli, S., Muller, S., Biggiogera, M., 2008a.Hepatoma tissue culture (HTC) cells as a model for investigating the effects of lowconcentrations of herbicide on cell structure and function. Toxicol. in Vitro 22,1853–1860.

Malatesta, M., Boraldi, F., Annovi, G., Baldelli, B., Battistelli, S., Biggiogera, M., Quaglino, D.,2008b. A long-term study on female mice fed on a genetically modified soybean:effects on liver ageing. Histochem. Cell Biol. 130, 967–977.

Moore, L.J., Somamoto, T., Lie, K.K., Dijkstra, J.M., Hordvik, I., 2005. Characterisation ofsalmon and trout CD8[alpha] and CD8[beta]. Mol. Immunol. 42, 1225–1234.

110 N.H. Sissener et al. / Aquaculture 298 (2009) 101–110

Morimoto, R.I., Tissieres, A., Georgopoulus, C., 1990. Stress Proteins in Biology andMedicine. Cold Springs Harbour Press, Cold Springs Harbour.

Morimoto, R.I., Sarge, K.D., Abravaya, K., 1992. Transcriptional regulation of heat shockgenes. J. Biol. Chem. 267, 21987–21990.

Morin, P.P., Hara, T.J., Eales, J.G., 1993. Thyroid hormone deiodination in brain, liver, gill,heart and muscle of Atlantic salmon (Salmo salar) during photoperiodically-induced parr-smolt transformation, I. Outer- and inner-ring thyroxine deiodina-tion. Gen. Comp. Endocrinol. 90, 142–156.

Nordrum, S., Bakke-McKellep, A.M., Krogdahl, Å., Buddington, R.K., 2000. Effects ofsoybean meal and salinity on intestinal transport of nutrients in Atlantic salmon(Salmo salar L.) and rainbow trout (Oncorhynchus mykiss). Comp. Biochem. Physiol.B Biochem. Mol. Biol. 125, 317–335.

OECD, 2001. Consensus Document on Compositional Considerations for New Varietiesof Soybean: Key Food and Feed Nutrients and Anti-Nutrients. Series on the safety ofnovel foods and feeds, No. 2. http://www.olis.oecd.org/olis/2001doc.nsf/LinkTo/NT00005036/$FILE/JT00117705.PDF.

Olsvik, P., Lie, K., Jordal, A.-E., Nilsen, T., Hordvik, I., 2005. Evaluation of potential referencegenes in real-time RT-PCR studies of Atlantic salmon. BMC Mol. Biol. 6, 21.

Olsvik, P., Lie, K., Hevrøy, E., 2007. Do anesthetics and sampling strategies affecttranscription analysis of fish tissues? BMC Mol. Biol. 8, 48.

Ortuño, J., Esteban, M.A., Meseguer, J., 2002. Effects of four anaesthetics on the innateimmune response of gilthead seabream (Sparus aurata L.). Fish Shellfish Immunol.12, 49–59.

Padgette, S.R., Kolacz, K.H., Delannay, X., Re, D.B., LaVallee, B.J., Tinius, C.N., Rhodes, W.K.,Otero, Y.I., Barry, G.F., Eichholtz, D.A., Peschke, V.M., Nida, D.L., Taylor, N.B., Kishore, G.M.,1995. Development, identification, and characterization of a glyphosate-tolerantsoybean line. Crop. Sci. 35, 1451–1461.

Parcellier, A., Gurbuxani, S., Schmitt, E.1, Solary, E., Garrido, C., 2003. Heat shockproteins, cellular chaperones that modulate mitochondrial cell death pathways.Biochem. Biophys. Res. Commun. 304, 505–512.

Pryme, I.F., Lembcke, R., 2003. In vivo studies on possible health consequences ofgenetically modified ingredients consisting of genetically modified plants. Nutr.Health 17, 1–8.

Rabergh, C.M., Airaksinen, S., Soitamo, A., Bjorklund, H.V., Johansson, T., Nikinmaa, M.,Sistonen, L., 2000. Tissue-specific expression of zebrafish (Danio rerio) heat shockfactor 1 mRNAs in response to heat stress. J. Exp. Biol. 203, 1817–1824.

Roche, H., Bogé, G., 1996. Fish blood parameters as a potential tool for identification ofstress caused by environmental factors and chemical intoxication. Mar. Environ.Res. 41, 27–43.

Sagstad,A., Sanden,M.,Krogdahl, Å., Bakke-McKellep,A.M., Frøystad,M.,Hemre,G.-I., 2008.Organsdevelopment, geneexpressionandhealth ofAtlantic salmon(Salmo salarL.) fedgenetically modified soybeans compared to the near-isogenic non-modified parentalline. Aqua Nutr. 14, 556–572.

Sandnes, K., Lie, Ø., Waagbø, R., 1988. Normal ranges of some blood chemistryparameters in adult farmed Atlantic salmon, Salmo salar. J. Fish Biol. 32, 129–136.

Séralini, G.-E., Cellier, D., de Vendomois, J., 2007. New analysis of a rat feeding studywith a genetically modified maize reveals signs of hepatorenal toxicity. Arch.Environ. Contam. Toxicol. 52, 596–602.

Sissener, N.H., Sanden, M., Bakke, A.M., Kroghdahl, Å., Hemre, G.I., 2009. A long term trialwith Atlantic salmon (Salmo salar L.) fed genetically modified soy; focusing generalhealth and performance before, during and after the parr-smolt transformation.Aquaculture 108–117.

Sissener, N.H., Martin, S.A.M., Cash, P., Hevrøy, E.M., Sanden, M., Hemre, G.I., 2009.Proteomic profiling of liver from Atlantic salmon (Salmo salar) fed geneticallymodified soy compared to the near-isogenic non-GM line. Marine Biotechnol.doi:10.1007/s10126-009-9214-1.

Somers, D.A., Makarevitch, I., 2004. Transgene integration in plants: poking or patchingholes in promiscuous genomes? Curr. Opin. Biotechnol. 15, 126–131.

Stephensen, E., Sturve, J., Forlin, L., 2002. Effects of redox cycling compounds onglutathione content and activity of glutathione-related enzymes in rainbow troutliver. Comp. Biochem. Physiol. C Toxicol. Pharmacol. 133, 435–442.

Tacon, A.G.J., 1987. The Nutrition and Feeding of Farmed Fish and Shrimp—A TrainingManual 1. The Essential Nutrients. Food and Agriculture Organization of the UnitedNations, GPC/RLA/075/ITA, Brazil, p. 117.

Tintos, A., Gesto, M., Míguez, J.M., Soengas, J.L., 2008. Beta-Naphthoflavone and benzo(a)pyrene treatment affect liver intermediary metabolism and plasma cortisollevels in rainbow trout Oncorhynchus mykiss. Ecotoxicol. Environ. Saf. 69, 180–186.

Trabalza-Marinucci, M., Brandi, G., Rondini, C., Avellini, L., Giammarini, C., Costarelli, S.,Acuti, G., Orlandi, C., Filippini, G., Chiaradia, E., Malatesta, M., Crotti, S., Antonini, C.,Amagliani, G., Manuali, E., Mastrogiacomo, A.R., Moscati, L., Naceur Haouet, M.,Gaiti, A., Magnani, M., 2008. A three-year longitudinal study on the effects of a dietcontaining genetically modified Bt176 maize on the health status and performanceof sheep. Livest. Sci. 113, 178–190.

Uran, P.A., 2008. Ethiology of soybean-induced enteritis in fish. PhD thesis, WageninigenUniversity.

Uran, P.A., Schrama, J.W., Jaafari, S., Baardsen, G., Rombout, J.H.W.M., Koppe, W.,Verreth, J.A.J., 2008. Variation in commercial sources of soybean meal influencesthe severity of enteritis in Atlantic salmon (Salmo salar L.). Aqua Nutr.

Van den Ingh, T.S.G.A.M., Krogdahl, Å., Olli, J.J., Hendriks, H.G.C.J.M., Koninkx, J.G.J.F.,1991. Effects of soybean-containing diets on the proximal and distal intestine inAtlantic salmon (Salmo salar): a morphological study. Aquaculture 94, 297–305.

Vandesompele, J., De Preter, K., Pattyn, F., Poppe, B., Van Roy, N., De Paepe, A., Speleman, F.,2002. Accurate normalization of real-time quantitative RT-PCR data by geometricaveraging of multiple internal control genes. Genome Biol. 3 0034.0031-0034.0011.

Vecchio, L., Cisterna, B., Malatesta, M., Martin, T.E., Biggiogera, M., 2004. Ultrastructuralanalysis of testes frommice fed genetically modified soybean. Eur. J. Histochem. 48,448–454.

Welch, W.J., 1993. How cells respond to stress. Sci. America 269, 56–64.