Content uploaded by Monica Sanden
Author content
All content in this area was uploaded by Monica Sanden on Sep 04, 2018
Content may be subject to copyright.
Histological, digestive, metabolic, hormonal and some
immune factor responses in Atlantic salmon, Salmo salar L.,
fed genetically modified soybeans
A M Bakke-McKellep
1,2
, E O Koppang
2
, G Gunnes
2
, M Sanden
3
, G-I Hemre
3
,
T Landsverk
1,2
and Krogdahl
1,2
1 Aquaculture Protein Centre, CoE, Norway
2 Department of Basic Sciences and Aquatic Medicine, Norwegian School of Veterinary Science, Oslo, Norway
3 National Institute of Nutrition and Seafood Research (NIFES), Bergen, Norway
Abstract
The paper reports the second and final part of an
experiment aiming to study physiological and
health-related effects of genetically modified (GM)
soybean meal (SBM) type Roundup Ready
soy-
bean (RRS) in diets for post-smolt Atlantic salmon.
For 3 months salmon were fed diets containing
172 g kg
)1
full-fat SBM from RRS (GM-soy) or
an unmodified, non-isogenic line (nGM-soy), or a
reference diet with fishmeal as the sole protein
source (FM). Slight differences in anti-nutrient
levels were observed between the GM and nGM-
soy. Histological changes were observed only in the
distal intestine of the soy-fed fish. The incidence of
moderate inflammation was higher in the GM-soy
group (9 of 10 sampled fish) compared with the
nGM-soy group (7 of 10). However, no differences
in the concomitant decreases in activities of diges-
tive enzymes located in the brush border (leucine
aminopeptidase and maltase) and apical cytoplasm
(acid phosphatase) of enterocytes or in the number
of major histocompatibility complex class II+
cells, lysozyme activity, or total IgM of the distal
intestine were observed. GM compared with nGM-
soy fed fish had higher head kidney lysozyme
(11 856 vs. 10 456 units g
)1
tissue) and a ten-
dency towards higher acid phosphatase (0.45 vs.
0.39 lmol h
)1
kg
)1
body mass in whole tissue)
activities, respectively. Plasma insulin and thyroxin
levels, and hepatic fructose-1,6-bisphosphatase and
ethoxyresorufin-O-deethylase activities were not
significantly affected. It is not possible, however, to
conclude whether the differences in responses to
GM-soy were due to the genetic modification or
to differences in soy cultivars in the soy-containing
diets. Results from studies using non-modified,
parental line soybeans as the control group are
necessary to evaluate whether genetic modification
of soybeans in diets poses any risk to farmed
Atlantic salmon.
Keywords: enzymes, fish feed, genetically modified,
intestine, major histocompatibility complex, nutri-
tion.
Introduction
In formulated feeds used in aquaculture of Atlantic
salmon, Salmo salar L., inclusion of protein-rich
plant-derived ingredients such as soybean meal
(SBM) and corn gluten at moderate levels is
common. These crops have varieties that have been
genetically modified (GM). The commercial feed
producer is required by EU legislation to label feed
containing GM feed ingredients. At present, Euro-
pean Market Regulations 2003/1829/EC and 2003/
1830/EC govern the use of GM ingredients intended
for food or feed. Threshold labelling at 9 g kg
)1
for
the adventitious presence of approved GM material
and 5 g kg
)1
for GM materials not approved in
Europe is required. The ability of commercial feed
producers to procure non-GM soybeans, maize and
Journal of Fish Diseases 2007, 30, 65–79
Correspondence A M Bakke-McKellep, Department of Basic
Sciences and Aquatic Medicine, Norwegian School of Veterinary
Science, PO Box 8146 Dep., NO-0033 Oslo, Norway
(e-mail: anne.mckellep@veths.no)
65
2007 The Authors.
Journal compilation
2007
Blackwell Publishing Ltd
other crops is diminishing as the global production
and availability of commercial GM plants increases.
The consumer safety of products from fish and other
production animals fed GM feed ingredients has
been a subject of controversy.
Low to moderate levels of SBM of various
qualities may be included in salmonid diets without
causing the fish to show any serious negative effects
on growth or feed utilization (Olli, Krogdahl, van
den Ingh & Bratta
˚s 1994; Olli & Krogdahl 1994;
Storebakken, Shearer & Roem 1998; Refstie,
Storebakken, Baeverfjord & Roem 2001; Krogdahl,
Bakke-McKellep & Baeverfjord 2003). However,
even moderate levels of standard (extracted) and
full-fat SBM have been found to alter intestinal
structure and function. Reductions in digestive and
absorptive capacities (Olli & Krogdahl 1994; Olli
et al. 1994; Nordrum, Bakke-McKellep, Krogdahl
& Buddington 2000; Krogdahl et al. 2003), as well
as an inflammatory response in the distal intestine
(Baeverfjord & Krogdahl 1996; Bakke-McKellep,
Press, Krogdahl & Landsverk 2000; Krogdahl,
Bakke-McKellep, Røed & Baeverfjord 2000; Krog-
dahl et al. 2003) have been reported. The causatory
component(s) of the inflammation is not known.
Few studies so far have been conducted to test the
effects of the use of transgenic soybeans in fish feed
(Hammond, Vicini, Hartnell, Naylor, Knight,
Robinson, Fuchs & Padgette 1996; Brown, Wilson,
Jonker & Nickson 2003) or feed for other animals
(Hammond et al. 1996; Cromwell, Lindemann,
Randolph, Parker, Coffey, Laurent, Armstrong,
Mikel, Stanisiewski & Hartnell 2002; Malatesta,
Caporaloni, Rossi, Battistelli, Rocchi, Tonucci &
Gazzanelli 2002; Brake & Evenson 2004). Results
from a study of GM potatoes fed to rats suggested a
GM or secondary GM effect on organ sizes and
gastric mucosal structure (Ewen & Pusztai 1999),
indicating a need to thoroughly test each GM
plant product on animal models. Herbicide (gly-
phosate)-resistant Roundup Ready
soybean (RRS;
Monsanto Company, St. Louis, MO), modified to
express CP4 enolpyruvylshikimate-3-phosphate
synthase protein, is among the most widely avail-
able commercial GM products today. The Mons-
anto Company has engineered a wide range of crops
and shown that an increase in anti-nutrients may be
expected in RRS (Padgette, Taylor, Nida, Bailey,
MacDonald, Holden & Fuchs 1996). During fish
feed production, extrusion with heat treatment
under high pressure is therefore an important step
in the reduction of heat-labile anti-nutrients found
in soybeans (e.g. protease inhibitors, lectins,
goitrogens and anti-vitamins). However, heat-stable
anti-nutrients such as saponins, tannins, phytates,
allergens, oestrogens and flatulence factors may
cause problems when SBM is added to animal feeds
(Liener 1994).
The present paper reports findings in Atlantic
salmon following a 3-month feeding period with
diets containing commercially available full-fat
GM-soy compared with an unmodified full-fat
soy and a standard fishmeal (FM)-based diet free of
SBM. Results from histological examination of the
gastrointestinal tract, visceral organs, thymus and
brain; immunological investigations including
expression of major histocompatibility complex
class II (MHC class II), lysozyme activity and
immunoglobulin M (IgM) in various tissues;
activities of intestinal digestive enzymes and hepatic
metabolic enzymes; and hormone status are repor-
ted. A previously published paper (Hemre, Sanden,
Bakke-McKellep, Sagstad & Krogdahl 2005)
reported a series of other parameters investigated
during the same experiment. Thus the present
paper concludes the production and health mon-
itoring of post-smolt Atlantic salmon fed GM-soy
from this feeding trial.
Materials and methods
Diets
The formulation and proximate composition of the
experimental diets as well as the chemical composi-
tion of the SBM varieties are shown in Tables 1a
and 1b, respectively (also previously reported in
Hemre et al. 2005). The diets were produced at the
Norwegian Institute of Fisheries and Aquaculture
Research (Titlestad, Norway). A standard fishmeal
(Norsildmel, Bergen, Norway) was used as the
principal dietary source of protein. Total protein in
all three diets was 440 g kg
)1
of diet dry matter,
which is similar to the recommended level for
salmon (Lall & Bishop 1977; NRC 1993). The two
SBMs were included as meal from full-fat beans.
Genetically modified soy (GM-soy; contained
800 g kg
)1
RRS and 200 g kg
)1
unmodified soy-
bean) or non-genetically modified soy (nGM-soy)
replaced fishmeal at a level of 172 g kg
)1
of total dry
matter (130 g kg
)1
of total protein). Suprex corn
(unmodified) was used as the only source of starch in
the diets. Sucrose and soybean oil were included in
the reference (FM) diet, in which fishmeal was the
66
2007 The Authors.
Journal compilation
2007
Blackwell Publishing Ltd
Journal of Fish Diseases 2007, 30, 65–79 A M Bakke-McKellep et al. GM-soybeans in feeds for Atlantic salmon
sole source of protein, to match the levels in the
soybean diets. All diets were supplemented with
vitamins and minerals according to NRC (1993)
and with astaxanthin (carophyll pink; Hoffman,
LaRoche, Germany) according to the recommended
levels for salmon. The finely ground ingredients
were mixed and processed by means of a standard
double screw extrusion technique, as described in
Sørensen, Ljøkjel, Storebakken, Shearer & Skrede
(2002). All diets were dried to dry matter content
above 900 g kg
)1
and stored at room temperature.
Diets were approximately isoenergetic and gross
energy was calculated according to Tacon (1987)
using the following energy densities: lipid
(39.5 kJ g
)1
), protein (23.6 kJ g
)1
) and carbohy-
drate (17.2 kJ g
)1
).
Chemical analyses
Ingredients and feeds were analysed for proximate
composition as described in Hemre et al. (2005),
using standard reference methods. In addition to the
analyses described and reported in Hemre et al.
(2005) for monosaccharide and phytoestrogen lev-
els, and trypsin inhibition activities in the experi-
mental diets and/or SBMs, the analyses of additional
anti-nutritional factors was carried out as follows:
Bowman-Birk inhibitor and Kunitz trypsin
inhibitor (Hajo
´s, Gelencse
´r, Pusztai, Grant, Sakhri
& Bardocz 1995) and soybean lectin were analysed
at the Central Food Research Institute (CFRI) in
Budapest, Hungary by the research groups of Eva
Gelencse
´r and Gyo
¨ngyi Hajo
´s. For lectin detection
a competitive indirect enzyme-linked immunosor-
bent assay (ELISA) for quantification of SBA
(soybean agglutinin, Glycine max lectin) was
performed as previously described (Hajo
´set al.
1995) with some modifications. Plates were coated
overnight at 4 C with 0.5 lgmL
)1
SBA (Sigma,
St Louis, MO, USA) diluted in PBS [in 0.01 m
phosphate-buffered saline (PBS), pH 7.4]. Standard
SBA (Sigma) diluted in GPBS (0.01 mPBS con-
taining 0.1 mgalactose, pH 7.4) or soybean samples
with unknown lectin content were added, followed
by anti-SBA rabbit IgG antibody (CFRI) diluted
1:600 v/v with GPBS. After incubation (30 C,
30 min), anti-rabbit goat IgG horseradish peroxi-
dase conjugate (Jackson Immuno Research, Europe
Ltd., Suffolk, UK) diluted with GPBS (1:3000 v/v)
was added. Following incubation (30 C, 30 min)
a substrate solution of o-phenylene-diamine-H
2
O
2
(Sigma Fast, pH 5.0) was added and the reaction
was stopped by adding 3 mH
2
SO
4
after 15-min
incubation at room temperature. The optical
density was measured at 492 nm against a reference
filter at 630 nm. Between each step the plates
were washed three times with PBST (0.01 m
phosphate-buffered saline containing 0.1% Tween
20, pH 7.4). Results were calculated on the basis of
a half logarithmic regression fitting to the calibra-
tion curve (10
)3
–10
3
lgmL
)1
SBA).
Low molecular carbohydrates (Bach Knudsen
1997), monosaccharide in total and insoluble non-
Table 1a Formulation and proximate composition of the
experimental diets (g kg
)1
dry weight)
Diet code Standard FM nGM-soy GM-soy
Formulation (g kg
)1
)
Fish meal
a
573 488 488
GM full-fat soybean meal
b
– – 172
nGM full-fat soybean meal
c
–172–
Suprex maize
d
164 122 122
Fish oil 198 203 203
Soy oil 34 – –
Sucrose 15 – –
Constant ingredients
e
15 15 15
Carophyll pink 0.8 0.8 0.8
Proximate analysis (g kg
)1
)
Dry matter 933 934 932
Protein 442 441 437
Lipid 254 259 257
Starch 131 100 101
Ash 94 90 92
Residue
f
12 44 45
Gross energy (kJ g
)1
) 22.7 22.3 22.2
Saponins (SD)
Saponin V (ng g
)1
) – 0.0 0.1
Saponin I (lmol g
)1
) – 1.0 0.7
Saponin II (lmol g
)1
) – 0.5 0.3
Saponin aG(lmol g
)1
) – 0.0 0.1
Saponin bG(lmol g
)1
) – 0.2 0.2
Saponin bA(lmol g
)1
) – 0.1 0.1
Total saponin (lmol g
)1
) – 1.7 1.4
Protease inhibition
Total trypsin inhibition
g
(g kg
)1
)
0.0 0.7 0.4
Kunitz trypsin inhibitor
(mg kg
)1
)
– 76.3 32.9
Bowman-Birk inhibitor
(mg kg
)1
)
– 1.5 1.5
Lectin (SBA) (mg kg
)1
) – 0.2 0.1
Fatty acid composition (g kg
)1
of total fatty acids) and total amino acids
were equivalent in all diets. Calcium varied from 18.6 to 19.0 g kg
)1
,
phosphorous from 12.7 to 13.1 g kg
)1
, zinc from 206 to 214 mg kg
)1
,
iron from 154 to 166 mg kg
)1
and selenium from 1.4 to 1.6 mg kg
)1
.
a
Tobismeal (Norsildmel, Bergen, Norway).
b
Roundup Ready
soy registered trademark of Monsanto Technology
LLC.
c
Soyax-aqua LF
, Schouten Industries B.V., Glessen, the Netherlands.
d
Suprex maize (Condrico, the Netherlands).
e
Minerals and vitamins were added according to NRC (1993) recom-
mendations.
f
Residue was calculated as 1000 )(protein +lipid +starch +
ash +moisture).
g
Trypsin inhibitor activity; g trypsin inhibited per kg diet.
67
2007 The Authors.
Journal compilation
2007
Blackwell Publishing Ltd
Journal of Fish Diseases 2007, 30, 65–79 A M Bakke-McKellep et al. GM-soybeans in feeds for Atlantic salmon
starch polysaccharides (Englyst, Quigley, Hudson
& Cummings 1992) and phytic acid (Carlsson,
Bergman, Skoglund, Hasselblad & Sandberg 2001)
were analysed by the research group of Stephan
Sahlstrøm at Matforsk (A
˚s, Norway). Saponin
analyses (according to Hu, Lee, Hendrich &
Murphy 2002) were carried out by the research
group of Patricia A. Murphy at Department of
Food Science and Human Nutrition at Iowa State
University, USA.
Fish, rearing conditions and experimental design
Animals were held and the experiment was
conducted in accordance with the Norwegian
Animal Welfare Act No. 73 of 20 December
1974 and the Regulation on Animal Experimenta-
tion of 15 January 1996. Prior to the start of the
feeding trial, all fish were acclimatized for 6 weeks
and fed commercial fishmeal-based diets (Transfer
Boost, 3 mm pellets; Ewos, Dirdal, Norway). This
diet was analysed and found to be GM-free. On
23 November, 1200 seawater-adapted, vaccinated
(Alpha Ject 5200, Alpharra, Oslo, Norway; against
Aeromonas salmonicida subsp. salmonicida,Vibrio
anguillarum serovar O1 and serovar O2, Vibrio
salmonicida, and Vibrio viscosus; 0.2 mL by intra-
peritoneal injection) Atlantic salmon smolt (mean
weight 104 15 g) were randomly distributed
among 15 square, grey, covered, fibreglass tanks
(1.5 m ·1.5 m ·1.0 m) at the Norwegian Insti-
tute of Fisheries and Aquaculture Research,
Austevoll, Norway. Fish were fed with automated
feeders, and feed intake was monitored. Pit-tagging
(Glass Tag Unique 2, 12 ·12 mm; Jojo Autom-
asjon AS, Sola, Norway) of each fish was used to
control individual weight development. The experi-
ment was conducted over a period of 12 weeks in
which all fish were subjected to a 24 h light regime
(November–March). Each diet was given to five
parallel tanks. Water flow rate was adjusted to
maintain the oxygen content of the outlet water
above 7 mg L
)1
. Temperature was 11 1C and
the salinity of the water was 33 g L
)1
.
Sampling procedure
At the end of the 12-week feeding trial, seven fish
were randomly sampled from each tank. They were
pre-anaesthetized with Aqui-S
TM
(540 g L
)1
isoe-
ugenol; Scan Aqua, A
˚rnes, Norway) and anaesthe-
tized with MS-222 (50 mg L
)1
metacainum; Norsk
Medisinaldepot AS, Bergen, Norway). Their body
weights and fork lengths were measured.
Of the sampled fish, two from each tank were
sampled for histological screening, IgM content and
lysozyme activity of the gastrointestinal tract tissue
– stomach, pyloric caeca, mid intestine and distal
intestine – and the liver, spleen, kidney, head
kidney, thymus and brain, as well as muscle tissue.
For this, the gastrointestinal tract, with the excep-
tion of the pyloric caeca, was opened, contents
gently removed, and the tract divided into the
various regions as described previously (Nordrum
et al. 2000). Approximately 0.5 cm pieces from
each region were placed in Bouin’s fixative (15:5:1
saturated picric acid: 37% formaldehyde in meth-
anol:glacial acetic acid). Approximately 0.2 cm
pieces from each region were frozen in propane
gas chilled to its liquid state in liquid-nitrogen.
Similarly sized pieces from the other organs and
tissues were also fixed in Bouin’s fixative and frozen
in chilled propane. Additional pieces from the mid
and distal intestine were also fixed in PBS-buffered
formalin (pH 7.4; 4%) for 24 h before being
transferred to and stored in 70% ethanol. The
remaining tissue from the four regions of the
gastrointestinal tract and the organs were frozen in
liquid nitrogen. All frozen tissues were stored at
)80 C until analysis.
From the remaining five fish, blood was collected
from the caudal vessels using heparinized, evacuated
glass test tubes (Vacutainers
: Becton Dickinson,
Franklin Lakes, NJ, USA) as well as non-treated
Vacutainers. The blood was stored on ice, and
plasma and serum were separated by centrifugation
at 3000 gfor 10 min before freezing and storage at
)80 C. After blood sampling, these five fish were
dissected, and stomach, intestine (pyloric, mid and
distal regions), liver, spleen, kidney, head-kidney
and brain were immediately frozen in liquid
nitrogen following the removal of stomach and
intestinal content, as well as extraneous fat and
connective tissue surrounding the tissues. These
were stored at )80 C.
Histological techniques and morphological
evaluation
All chemically fixed (Bouin’s or formalin) tissues
were routinely dehydrated in ethanol, equilibrated
in xylene, and embedded in paraffin according to
standard histological techniques. Sections of
approximately 5 lm were cut.
68
2007 The Authors.
Journal compilation
2007
Blackwell Publishing Ltd
Journal of Fish Diseases 2007, 30, 65–79 A M Bakke-McKellep et al. GM-soybeans in feeds for Atlantic salmon
The Bouin-fixed tissue from the gastrointestinal
tract – stomach, pyloric caeca, mid intestine and
distal intestine – and the liver, spleen, kidney, head
kidney, thymus and brain, as well as muscle tissue,
were stained with haematoxylin and eosin (H&E)
before blinded examination under a light micro-
scope. Tissue and organ morphology was evaluated
according to Amin, Mortensen & Poppe (1992).
Distal intestinal morphology was evaluated accord-
ing to the criteria previously described in Atlantic
salmon with SBM-induced enteritis (Baeverfjord &
Krogdahl 1996): (1) widening and shortening of
the intestinal folds; (2) loss of the supranuclear
vacuolization in the absorptive cells (enterocytes) in
the intestinal epithelium; (3) cellular infiltration of
a mixed leucocyte population in the central lamina
propria within the intestinal folds as well as in the
submucosa.
Immunological factors
Major histocompatibility complex class II
The formalin fixed tissue sections from the mid and
distal intestine were stained immunohistochemically
using rabbit antisera recognizing Atlantic salmon
MHC class II bchain as described by Koppang,
Hordvik, Bjerka
˚s, Torvund, Aune, Thevarajan &
Endresen (2003b). Briefly, sections were mounted on
slides coated with 3-aminopropyl triethoxy silane
(Sigma Chemical Company), dried overnight at
50 C, and subsequently stored at room temperature
until staining. Tissue sections were de-waxed,
re-hydrated in graded ethanol baths and autoclaved
in 0.01 mcitrate buffer (pH 6.0) at 121 C for
15 min. To prevent non-specific binding the sections
were incubated in normal caprine serum [1:50 in 5%
bovine serum albumin (BSA) in TBS] for 20 min.
They were then incubated overnight at 4 C with a
rabbit polyclonal antibody against MHC class II b
chain diluted 1:1000 in 1% BSA in TBS. Following
washing in PBS, the sections were incubated for
30 min in biotinylated goat anti-rabbit IgG (Vector
Laboratories, Peterborough, UK) diluted 1:200 in
1% BSA in TBS. Following inhibition of endog-
enous peroxidase activity, the sections were incubated
for 30 min in avidin–biotin complex/horseradish
peroxidase (ABC/HRP-complex, PK-4000; Vector
Laboratories) according to the manufacturer’s
instructions, rinsed in PBS, and bound antibody
detected using 3-amino-9-ethyl-carbazole (AEC;
Sigma) for 17 min at room temperature.
Digital image morphometry was performed to
quantify the number of MHC class II+cells in the
distal intestine. Digital images of the immunohisto-
chemically labelled sections were acquired using a
Leica DM RXA microscope (Leica Microsystems AG,
Wetzlar, Germany) equipped with a Spot RT Slider
camera (Diagnostic Instruments Inc., Sterling
Heights, MI, USA). Using Image-Pro Plus 5.1
software (Media Cybernetics, Silver Spring, MD,
USA), the number of positive cells within the
epithelial cell layer, the lamina propria and the
submucosa above the stratum compactum was
counted separately and related to the surface area
analysed in each tissue section. Only tissue sections
squarely cut across all tissue layers, so that both
compartments were sufficiently represented, were
analysed. Because of these strict criteria, only one to
two individuals from each of the three tanks per
experimental diet were analysed. The mean from each
tank, when applicable, and diet group was calculated.
Lysozyme activity
Lysozyme levels were determined in homogenate
from distal intestine, liver, kidney, head kidney
and spleen using a turbidimetric microplate
technique with the Gram-positive bacteriun Micro-
coccus lysodeikticus (Sigma no. A-7906) suspension
(0.2 mg mL
)1
) calibrated with hen egg white
lysozyme (Ellis 1990). Lysozyme activity was
calculated relative to tissue weight.
Immunoglobulin M
The total amount of IgM was determined in
homogenate from pyloric caeca, mid intestine,
distal intestine, liver, kidney, head kidney and
spleen as described by Lund, Gjedrem, Bentsen,
Eide, Larsen & Røed (1995): ELISA plates were
coated with rabbit antisalmon Ig diluted 1:6000 in
0.05 mol L
)1
carbonate buffer (pH 9.6). The plates
were rinsed and incubated with rinsing buffer
containing 1% BSA for 45 min at 37 C. After a
further rinse, mucosa homogenate (diluted 1:20 in
phosphate buffer containing 0.5 mL L
)1
Tween 20
and 4% horse serum) was added to the wells and
incubated overnight at 4 C. The plates were then
rinsed and incubated with a monoclonal antibody
raised against salmon IgM, diluted in rinsing buffer,
for 1 h at room temperature. After rinsing, peroxi-
dase-labelled antimouse Ig (Amersham, Uppsala,
Sweden) diluted 1:3000 in rinsing buffer was added
69
2007 The Authors.
Journal compilation
2007
Blackwell Publishing Ltd
Journal of Fish Diseases 2007, 30, 65–79 A M Bakke-McKellep et al. GM-soybeans in feeds for Atlantic salmon
and rinsed. The substrate o-phenylenediamine was
added and the plates were incubated for 10 min at
room temperature, followed by addition of
100 lL 1 mol L
)1
H
2
SO
4
. The optical density
was read spectrophotometrically at 492 nm.
Enzyme activities
Acid phosphatase (AcP), leucine aminopeptidase
(LAP) and maltase activities were analysed in tissue
homogenates. Stomach, liver, kidney and head-
kidney tissues were analysed for AcP and intestinal
tissues from the pyloric, mid and distal intestine
were analysed for AcP, LAP and maltase. The
tissues were thawed and homogenized (1:20 w/v)
in ice-cold 2 mmTris/50 mmmannitol, pH 7.1,
containing phenyl-methyl-sulphonyl fluoride (Sig-
ma no. P-7626) as serine protease inhibitor.
Aliquots of homogenates were frozen in liquid
nitrogen and stored at )80 C prior to analysis.
The AcP and LAP activities were measured
colorimetrically with kits (Sigma procedures no.
104 and no. 251, respectively) – using p-nitro-
phenyl phosphate as the substrate for AcP and
l-leucyl-b-naphthylamide as the substrate for LAP.
Disaccharidase (maltase) activity was analysed
according to the methods described by Dahlquist
(1970) using maltose as substrate. Incubations
were performed at 37 C.
Fructose-1,6-bisphosphatase activity in hepatic
tissue was analysed as described by Borrebaek,
Waagbø, Christophersen, Tranulis & Hemre
(1993). Activity was measured in the presence of
2mmNADP +2mmfructose-l,6-bisphosphate
and 0.65 U mL
)1
of both purified glucose-6-
phosphate dehydrogenase and 6-phosphogluconate
dehydrogenase. Activity was assessed spectropho-
tometrically by the formation of NADPH at
30 C.
Enzyme activities are expressed as amount of
substrate hydrolysed as related to time and g tissue,
whole tissue and kg body weight, and mg protein in
the homogenate. Protein was analysed using the
Bio-Rad Protein Assay (Bio-Rad Laboratories,
Munich, Germany).
Xenobiotic metabolizing enzyme
Preparation of microsomes
Preparation of microsomes from liver tissue was
performed on ice. The frozen samples were homo-
genized in a sodium phosphate buffer (0.1 m; pH 7.4;
4mLg
)1
liver) containing KCl (0.15 m), ethylen-
ediaminetetraacetic acid (EDTA; 1 mm), dithiothre-
itol (DTT; 1 mm), phenylmethylsulfphonylfluoride
(PMSF; 0.1 mm) and glycerol (10% v/v). The
homogenate was centrifuged (10 000 g, 15 min at
4C) and the supernatant was filtered through glass
wool before it was centrifuged for 60 min at
100 000 g(4 C). The resulting microsomal frac-
tion was obtained by re-suspending the pellet in a
sodium phosphate buffer (0.1 m, pH 7.4; 1 mL g
)1
liver) containing EDTA (1 mm), DTT (1 mm),
PMSF (0.1 mm) and glycerol (20% v/v).
Enzyme assays
Ethoxyresorufin-O-deethylase (EROD) activity
was assayed fluorimetrically in the microsomal
fraction as described by Stagg & Addison (1995).
Replicates of every sample were run. Briefly, a
1.5 mL incubation mixture containing 0.1 m
sodium phosphate buffer (pH 7.8), 5 lL7-
ethoxyresorufin (0.5 mm, dissolved in DMSO),
and 10 lL NADPH (10 mm) was prepared.
Reactions were initiated by adding 50 lLof
microsomal protein fraction (corresponding to
approximately 0.4 mg protein) and stopped with
1.5 mL of ice-cold methanol. The incubation
time was 15 min. The assay vials were centrifuged
and the supernatant was transferred to high-
performance liquid chromatography (HPLC)
vials. Resorufin was then quantified against
known standards by the use of HPLC (5 lL
injections, mobile phase: 40:60 v/v acetonitri-
le:water flow, 1.0 mL min
)1
). The column used
was a Symmetry
C18 (3.9 ·150 mm, 5 lm;
Waters, Milford, MA, USA) and resorufin was
detected on a Perkin-Elmer LS-4 (Perkin-Elmer,
Buckinghamshire, UK) fluorescence detector.
Excitation was at 535 nm and fluorescence emis-
sion was measured at 585 nm. Protein was
analysed using the Bio-Rad Protein Assay (Bio-
Rad Laboratories).
Hormone levels
Insulin
Endocrine pancreatic response was determined by
analysing plasma insulin by radioimmunoassay as
described by Plisetskaya, Dickhoff, Paquette &
Gorbman (1986).
70
2007 The Authors.
Journal compilation
2007
Blackwell Publishing Ltd
Journal of Fish Diseases 2007, 30, 65–79 A M Bakke-McKellep et al. GM-soybeans in feeds for Atlantic salmon
Total thyroxin (TT4)
Thyroid status was estimated by measuring total
thyroxin levels in serum using the Immulite
2000
Canine Total T4 kit (Diagnostic Products
Corporation Norway AS, Drammen, Norway), a
solid-phase, chemiluminescent competitive immu-
noassay. The results were read with an Immulite
2000 Analyzer (Diagnostic Products Corporation,
Los Angeles, CA, USA).
Statistics
Analyses of variance and Duncan’s multiple range
test were used in the evaluation of most of the
results (tank means) with diet as class variable. For
most variables the analysis was run for each
intestinal section and organ separately due to
significant differences in variances. Transformation
of the results to obtain similar variances was not
performed as the regions and organs seemed to
respond differently both to diet and, hence, to
require separate evaluation. The software package
SAS (Release 8.02 TS Level 02MO; SAS Institute
Inc., Cary, NC, USA) was used in the evaluation.
For results of the histological evaluation, a numer-
ical score was assigned to changes observed and
contingency (frequency) analysis was run using the
software JMP 5.0.1a (SAS Institute Inc.). In all tests
performed, differences were considered statistically
significant when P<0.05. Critical range was used,
where applicable, to indicate the minimum differ-
ence required between the highest and lowest values
for them to be significantly different.
Results
As reported by Hemre et al. (2005), no mortality
was recorded and no significant differences in
growth were detected between the treatment groups
during the 3-month feeding trial. The more
extensive analysis of anti-nutritional factors in the
SBMs (Table 1a) revealed some differences between
the nGM and GM soy: non-starch polysaccharide,
monosaccharide, phytate and phytoestrogen levels
were higher in the GM-soy whereas low molecular
carbohydrates, trypsin inhibitor, lectin and total
saponin levels were higher in the nGM-soy. For the
latter three, differences were not as marked in the
experimental diets (Table 1b), although the same
trends of higher levels in the nGM-soy diet were
observed. Lower values in the experimental diets
compared with in the soy alone were considered a
result of dilution effect and heat caused by extrusion
during diet preparation.
Histological examination of the stomach, pyloric
caeca, mid intestine, liver, spleen, kidney, head
kidney, thymus, brain and muscle tissue revealed no
differences that could be attributed to variations in
feed composition. The only tissue that showed diet-
related variation was the distal intestine (Table 2),
in which the tissue from fish fed the two soy-
containing diets, in comparison with fish fed
standard FM diet, showed marked shortening of
the villous folds, decreases in the presence of
Table 1b Chemical composition of the full-fat soybean meal
varieties: nGM-soy and GM-soy
nGM-soy GM-soy
Proximate composition (g kg
)1
)
Dry matter 910 900
Protein (N ·6.25) 344 344
Lipid 186 190
Starch 4 4
Ash 47 57
Moisture 90 100
Residue
a
329 305
Non-starch polysaccharides (g kg
)1
)
Total 197 256
Insoluble 163 185
Low molecular carbohydrates (g kg
)1
)
Stacchyose 17.1 14.4
Raffinose 4.6 4.4
Sucrose 23.1 15.2
Glucose 0.6 0.9
Fructose 0.1 0.1
Monosaccharides
b
(g kg
)1
) 0.42 0.47
Inositol phosphates (IP; phytates) (mg kg
)1
)
IP4 0.1 0.2
IP5 0.9 0.9
IP6 7.7 11.1
Phytoestrogens (g kg
)1
)
Daidzein 0.25 0.30
Genistein 0.41 0.51
Protease inhibition
Total trypsin inhibition
c
(g kg
)1
) 6.7 5.2
Kunitz trypsin inhibitor (lgg
)1
) 309.9 100.9
Bowman-Birk inhibitors (lgg
)1
) 3.4 2.1
Lectin (SBA) (lgg
)1
) 58.6 7.4
Saponins
Saponin V (ng g
)1
) 0.2 0.1
Saponin I (lmol g
)1
) 2.6 2.0
Saponin II (lmol g
)1
) 1.4 1.1
Saponin aG(lmol g
)1
) 0.3 0.3
Saponin bG(lmol g
)1
) 1.6 1.3
Saponin bA(lmol g
)1
) 1.0 0.8
Total saponin 7.0 5.6
a
Residue was calculated as 1000 )(protein +lipid +starch +ash +
moisture).
b
Total monosaccharides ¼rhamnose +fucose +arabinose +xylose
+mannose +galactose +uronic acid +glucose.
c
Total trypsin inhibitor activity; g trypsin inhibited per kg sample.
71
2007 The Authors.
Journal compilation
2007
Blackwell Publishing Ltd
Journal of Fish Diseases 2007, 30, 65–79 A M Bakke-McKellep et al. GM-soybeans in feeds for Atlantic salmon
supranuclear vacuoles, and thickening of the vil-
lous foldsÕcentral stroma (lamina propria) and
submucosa (Fig. 1). The frequency of the occur-
rence of moderate changes was higher in the fish fed
the GM-soy (9 of 10 fish) compared with those fed
nGM-soy (7 of 10 fish) and standard FM (0 of 10
fish). Contingency analysis revealed a significant
effect of diet on histological score (v
2
¼12.848; 4
degrees of freedom; P>0.0120).
Lysozyme and IgM levels in the pyloric caeca,
mid intestine, distal intestine, liver, kidney, head
kidney and spleen (Table 3) did not significantly
differ between feeding groups. For these parame-
ters, variation between replicate tank means (as well
as individual values) were high. Lysozyme levels
showed a slight but not significant increase in the
head kidney of fish fed GM-soy (P¼0.0608).
Immunohistochemical detection (Fig. 2) indicated
MHC class II+cells in the epithelial cell layer,
possibly a population of intraepithelial leucocytes,
and lamina propria with possible qualitative differ-
ences in staining between dietary groups. Staining
of the brush border of apically located enterocytes
was visible in some soy-fed individuals. Image
morphometric quantification (Table 3) of the
number of MHC class II+cells in these mucosal
compartments, however, revealed no significant
differences between feeding groups. However, vari-
ation was high due to the low number of tissue
sections qualitatively adequate for analysis.
In the pyloric and mid intestinal regions, LAP,
maltase and AcP activities were not significantly
affected by diet (Table 4). Digestive enzyme activ-
ities (LAP and maltase) in the distal intestine,
whether expressed relative to g tissue, whole tissue
and kg body weight or mg protein (Table 4), were
all significantly (P<0.0001) lower in the fish fed
the soybean-containing diets compared with the
standard fishmeal-based diet. However, no differ-
ences in activities between GM and nGM-soy
groups were detected. Although there was a similar
(a)
(b)
(c)
Figure 1 Histological detail of the Bouin-fixed distal intestinal
villous folds of Atlantic salmon fed the standard FM (a), nGM-
soy (b) and GM-soy (c) diets. The tissue was considered normal
in (a); slightly to moderately changed in (b) as vacuoles were still
present in the supranuclear cell compartment (*), normally
present in the enterocytes (e), but the width of the lamina propria
(lp) was somewhat increased; and moderately changed in (c)
whose enterocytes mostly lacked vacuoles (H&E, ·400; photos
by A M Bakke-McKellep).
Table 2 Histological changes in the distal intestine for each
dietary group
Degree of
changes
observed
Standard
FM nGM-soy GM-soy Statistics* (P-value)
Normal 6 0 0 0.0120
Slight 4 3 1
Moderate 0 7 9
Severe 0 0 0
The values for the histological changes depict the number of individuals in
each experimental diet group (n¼10) displaying various degrees of
changes (see text and Fig. 1). *Statistical test performed was contingency
(frequency) analysis. The P-value indicates differences between diet groups
and considered significant when P<0.05.
72
2007 The Authors.
Journal compilation
2007
Blackwell Publishing Ltd
Journal of Fish Diseases 2007, 30, 65–79 A M Bakke-McKellep et al. GM-soybeans in feeds for Atlantic salmon
diet-related trend observed in AcP activities in the
distal intestine, it was not significant, nor were there
diet-related differences in the stomach, liver and
kidney (Table 4). In the head kidney, however, AcP
activity relative to kg body weight (Table 4) of fish
fed GM-soy was significantly (P¼0.0362) higher
than in those fed the standard FM with values in
the nGM-soy group intermediate. No significant
effect of diet was detected for fructose-1,6-bisphos-
phatase activity in the liver (Table 2). Diet means
(n¼5 per tank; three replicate tanks per diet) for
EROD activity, a xenobiotic metabolizing enzyme,
in the liver (Fig. 3) was elevated but not signifi-
cantly so (P¼0.3808) in the fish fed the
(a) (b)
(d)(c)
(f)(e)
Figure 2 Immunohistochemical detection of MHC class II in formalin-fixed sections of the distal intestine of Atlantic salmon fed the
standard FM (a, b), nGM-soy diet (c, d), and the GM-soy (e, f) diets. MHC class II+cells are stained brown and appear to comprise a
population of intraepithelial leucocytes (arrows), as well as some cells in the submucosa (a, c and e; arrowheads) and the brush border of
enterocytes (f; arrowhead) of some soy-fed individuals [H&E, ·200 (a, c, e) and ·400 (b, d, f); photos by A M Bakke-McKellep].
73
2007 The Authors.
Journal compilation
2007
Blackwell Publishing Ltd
Journal of Fish Diseases 2007, 30, 65–79 A M Bakke-McKellep et al. GM-soybeans in feeds for Atlantic salmon
soy-containing diets (34.2 and 35.4 for the nGM
and GM-soy, respectively) compared with the
reference FM diet (26.8 pmol min
)1
mg
)1
pro-
tein). The critical range was 15.66.
Insulin levels did not differ significantly between
the feeding groups. The mean (n¼5 per tank; five
replicate tanks per diet) for the standard FM group
was 10.0; for the nGM-soy group 9.5; and for the
GM-soy group 7.6 ng mL
)1
(critical range 4.2;
P¼0.4027).
Most plasma total thyroxin (TT
4
) levels meas-
ured were below the detection limit of 6 nmol L
)1
.
When this occurred, and assuming normal distri-
bution, the individual value was set at 1/2 the
detection limit. ANOVA then revealed no signifi-
cant effect of diet on TT
4
values. Mean (n¼5 per
tank; five replicate tanks per diet) value for the
standard FM group was 3.8; for the nGM-soy
group 4.7; and for the GM-soy group
5.7 nmol L
)1
(critical range 3.1; P¼0.4331).
Discussion
Because of differences in the levels of anti-
nutritional factors in the experimental diets con-
taining GM-soy and a non-parental, unmodified
counterpart, we can only conclude from the
present study that different soybean varietals may
cause qualitative and quantitative differences in the
inflammatory response observed in the distal
intestine as well in the head kidney of SBM-fed
Atlantic salmon. No apparent correlation was
observed in the incidence of moderate changes in
the distal intestine and trypsin inhibitor, lectin or
saponin levels, thus indicating that these anti-
nutritional factors may not be the direct cause of
Table 3 Mean lysozyme activity, total immunoglobulin M (IgM) and MHC class II-positive cells in various tissues for each dietary
group
Standard FM nGM-soy GM-soy Critical range ANOVA (P-value)*
Lysozyme (units g
)1
tissue)
Pyloric intestine n.d. n.d. n.d. – –
Mid intestine n.d. n.d. n.d. – –
Distal intestine 3223 3921 4668 2129 0.3355
Liver 4344 3658 4548 2918 0.7712
Kidney 15598 15856 20311 7758 0.3313
Head kidney 10425
b
10456
b
11865
a
1404 0.0608
Spleen 5004 4823 3912 1398 0.2035
Immunoglobulin M (optical density)
Pyloric intestine 0.228 0.149 0.179 0.195 0.6560
Mid intestine 0.326 0.211 0.199 0.159 0.1752
Distal intestine 0.319 0.231 0.236 0.184 0.4908
Kidney 0.357 0.418 0.390 0.134 0.6267
Head kidney 0.482 0.363 0.351 0.180 0.2261
Spleen 0.459 0.358 0.301 0.281 0.4529
MHC class II-positive cells (cell numbers in each mucosal compartment per mm
2
)
Distal intestine
Epithelium 124 105 157 118 0.5517
Lamina propria 46 43 49 50 0.9593
Means were calculated from the mean of each replicate tank (n¼5 fish per tank; five replicate tanks per diet for lysozyme and immunoglobulin M;
n¼1–2 fish per tank; three replicate tanks per diet for MHC class II). Different superscript letters indicate significant differences within each row-wise
comparison. Critical range indicates the minimum difference required between the highest and lowest values for them to be significantly different.
n.d. ¼not detected. *Significant when P<0.05.
0
15
30
45
FM nGM-soy GM-soy
EROD activity
(pmol min–1 mg–1 protein)
Figure 3 Mean ethoxyresorufin-O-deethylase (EROD; pmol -
min
)1
mg
)1
protein) activities in the liver of Atlantic salmon fed
the standard FM (white columns), nGM-soy (light grey
columns) and GM-soy (dark grey columns) diets (n¼5 per
tank; three replicate tanks per diet). Differences between diet
groups were not significant (P¼0.0984). Critical range,
indicating the minimum difference required between the highest
and lowest values for them to be significantly different, was 15.7.
74
2007 The Authors.
Journal compilation
2007
Blackwell Publishing Ltd
Journal of Fish Diseases 2007, 30, 65–79 A M Bakke-McKellep et al. GM-soybeans in feeds for Atlantic salmon
the SBM-induced enteritis in Atlantic salmon.
This is the first study in which any indication that
dietary SBM for a salmonid may result in an
increased immune response in the head kidney,
as indicated by lysozyme and acid phosphatase
activity levels, has been observed. Reduced macro-
phage activity has previously been reported in the
head kidney of rainbow trout fed high levels (600–
800 g kg
)1
) of SBM (Burrells, Williams, South-
gate & Crampton 1999). Thus the systemic
immune response to different levels of dietary
SBM merits further investigation. However, due to
the high variability between individuals within
each tank, as well as mean values between parallel
tanks, a higher number of replicates are needed to
lower variation. With the methods employed, this
appears to be particularly relevant for not only
immunological responses, but also for the xeno-
Table 4 Mean enzyme activities in tissue homogenates, as related to mol substrate hydrolysed per unit time and g tissue, whole tissue
and kg body mass (BM), or mg protein, for each dietary group
Enzyme tissue Units Standard FM nGM-soy GM-soy Critical range ANOVA (P-value)*
Acid phosphatase
Stomach lmol h
)1
g
)1
tissue 0.09 0.10 0.11 0.02 0.3257
lmol h
)1
kg
)1
BM 0.64 0.64 0.77 0.23 0.3555
lmol h
)1
mg
)1
protein 1.24 1.23 1.36 0.27 0.5087
Pyloric intestine lmol h
)1
g
)1
tissue 0.23 0.22 0.23 0.03 0.6735
lmol h
)1
kg
)1
BM 5.76 5.60 5.77 0.86 0.8893
lmol h
)1
mg
)1
protein 1.68 1.63 1.60 0.15 0.4353
Mid intestine lmol h
)1
g
)1
tissue 0.19 0.19 0.19 0.03 0.9533
lmol h
)1
kg
)1
BM 0.50 0.53 0.51 0.11 0.7799
lmol h
)1
mg
)1
protein 1.66
a
1.46
b
1.63
ab
0.18 0.0552
Distal intestine lmol h
)1
g
)1
tissue 0.11 0.10 0.10 0.03 0.8054
lmol h
)1
kg
)1
BM 0.51
a
0.33
b
0.34
b
0.17 0.0530
lmol h
)1
mg
)1
protein 1.16 1.26 1.40 0.27 0.1600
Liver lmol h
)1
g
)1
tissue 0.25 0.25 0.24 0.04 0.7166
lmol h
)1
kg
)1
BM 3.35 3.08 2.96 0.53 0.2546
lmol h
)1
mg
)1
protein 1.23 1.12 1.14 0.17 0.3103
Kidney lmol h
)1
g
)1
tissue 0.28 0.28 0.28 0.02 0.8131
lmol h
)1
kg
)1
BM 2.16 2.10 2.13 0.28 0.8756
lmol h
)1
mg
)1
protein 1.70 1.79 1.78 0.15 0.3640
Head kidney lmol h
)1
g
)1
tissue 0.24 0.23 0.25 0.04 0.4435
lmol h
)1
kg
)1
BM 0.36
b
0.39
ab
0.45
a
0.07 0.0362
lmol h
)1
mg
)1
protein 1.22 1.10 1.21 0.15 0.1937
Leucine aminopeptidase
Pyloric intestine mmol h
)1
g
)1
tissue 10.4 10.6 10.4 3.4 0.9879
mmol h
)1
kg
)1
BM 263.0 288.8 265.2 100.7 0.8135
lmol h
)1
mg
)1
protein 192.4 198.8 183.0 63.7 0.8525
Mid intestine mmol h
)1
g
)1
tissue 8.6 8.6 7.0 2.2 0.2067
mmol h
)1
kg
)1
BM 22.6 22.8 19.8 7.8 0.6279
lmol h
)1
mg
)1
protein 183.0 155.6 156.6 43.7 0.3042
Distal intestine mmol h
)1
g
)1
tissue 23.0
a
6.0
b
6.0
b
3.1 <0.0001
mmol h
)1
kg
)1
BM 116.8
a
18.8
b
22.0
b
19.5 <0.0001
lmol h
)1
mg
)1
protein 646.0
a
186.4
b
206.2
b
121.5 <0.0001
Maltase
Pyloric intestine lmol min
)1
g
)1
tissue 0.40 0.35 0.30 0.13 0.2496
lmol min
)1
kg
)1
BM 10.05 9.35 7.66 3.21 0.2558
nmol min
)1
mg
)1
protein 7.36 6.37 5.33 2.73 0.2775
Mid intestine lmol min
)1
g
)1
tissue 0.32 0.27 0.25 0.08 0.2606
lmol min
)1
kg
)1
BM 0.83 0.73 0.69 0.28 0.5424
nmol min
)1
mg
)1
protein 7.04 5.07 5.51 1.98 0.0990
Distal intestine lmol min
)1
g
)1
tissue 0.73
a
0.13
b
0.12
b
0.12 <0.0001
lmol min
)1
kg
)1
BM 3.54
a
0.41
b
0.45
b
0.77 <0.0001
nmol min
)1
mg
)1
protein 19.89
a
4.19
b
4.14
b
3.69 <0.0001
Fructose-1,6-bisphosphatase
Liver pmol min
)1
g
)1
tissue 16.9 15.1 17.3 2.8 0.2101
pmol min
)1
kg
)1
BM 229.2 190.9 214.6 48.1 0.2263
pmol min
)1
mg
)1
protein 21.4 17.6 20.3 4.0 0.1263
Means are calculated from the mean of each replicate tank (n¼5 per tank; five replicate tanks). Different superscript letters indicate significant differences
within each row-wise comparison. Critical range indicates the minimum difference required between the highest and lowest values for them to be
significantly different. *Significant when P<0.05 (in bold).
75
2007 The Authors.
Journal compilation
2007
Blackwell Publishing Ltd
Journal of Fish Diseases 2007, 30, 65–79 A M Bakke-McKellep et al. GM-soybeans in feeds for Atlantic salmon
biotic metabolizing enzyme EROD and hormonal
responses. The critical ranges presented may be
used as a basis for power analysis and the number
of replicates needed to identify valid differences.
Interestingly, the increased frequency of moder-
ate changes in the distal intestine of the fish fed the
SBM-containing diets did not coincide with
significantly increased presence of lysozyme, IgM
or MHC class II+cells, the latter considered to be
expressed in lymphocytes, macrophages, epithelial
cells and dendritic-like cells in non-stimulated
Atlantic salmon (Koppang et al. 2003b). Given
the high individual and tank variations observed, a
larger number of tissue sections qualitatively
adequate for morphometric quantification of
MHC class II+cells may have given different
expression results between the groups of fish, as
indicated by possible small differences observed in
Fig. 2. Also, vaccination by intra-peritoneal injec-
tion has previously been shown to cause granulo-
matous inflammation, characterized by MHC class
II+cells and other inflammatory infiltrate, in
distantly located muscles comprising the fillet
(Koppang, Haugarvoll, Hordvik, Aune & Poppe
2005) and sites in the eye (Koppang, Haugarvoll,
Hordvik, Poppe & Bjerka
˚s 2004) of Atlantic
salmon. We cannot exclude the possibility that
similar reactions may have occurred in the distal
intestinal mucosa of the vaccinated fish used in this
study, although the absence of any inflammatory
changes in the mucosa of the pyloric caeca and mid
intestine, as well as in the distal intestine of the
standard FM fed fish, appears to contradict this
possibility. Previous studies have demonstrated
upregulation of immunological factors in the distal
intestine of salmon fed higher levels of SBM
(Bakke-McKellep et al. 2000; Krogdahl et al.
2000), as well as following immune induction in
other immune-privileged tissues (Koppang, Bjerka
˚s,
Bjerka
˚s, Sveier & Hordvik 2003a). The intestinal
epithelium defines a barrier between the host and
the external environment consisting of physical,
innate and adaptive barriers (Sansonetti 2004). The
gut aims at keeping an immunological homeostasis,
but a number of different factors may influence and
disrupt the preferred status of equilibrium. For
instance, anti-nutritional factors, antigens and/or
toxins in soybeans may be active agents (Sissons
1982). Thus, the MHC class II response in the
distal intestine of salmon fed higher levels of SBM,
including the possibility of enterocytes acting as
antigen-presenting cells as indicated by the brush
border membrane staining in the soy-fed fish,
merits investigation.
The lack of significant effect of the soy-contain-
ing diets on AcP activity in the distal intestine is
likely to be a result of a decrease in AcP-containing
supranuclear lysosome or vesicle numbers and/or
size with a simultaneous increase in the number of
AcP-containing immune cells in the lamina propria
and submucosa, as previous histochemical studies in
SBM-fed salmon suggest (Bakke-McKellep et al.
2000). A loss of supranuclear vacuolization of the
enterocytes in soy-fed salmon has been observed in
numerous earlier studies (van den Ingh & Krogdahl
1990; van den Ingh, Krogdahl, Olli, Hendricks &
Koninkx 1991; Baeverfjord & Krogdahl 1996;
Krogdahl et al. 2003) and has been linked with a
concomitant decrease in AcP-staining in this cellu-
lar compartment (Bakke-McKellep et al. 2000).
This previous study by our group also showed
increased numbers of cells in the lamina propria
and submucosa that stained positively for AcP,
assumed to be macrophages (Press, Dannevig &
Landsverk 1994), as well as immunohistochemically
for neutrophilic granulocytes. Cells of monocytic
lineage (macrophages) as well as neutrophilic
granulocytes and lymphocytes may contribute to
total tissue AcP activity. Neutrophils and lympho-
cytes have been shown to contain AcP activity in
rainbow trout (Afonso, Ellis & Silva 1997)
although this does not appear to have been
specifically studied in Atlantic salmon. Thus, AcP
activity measured biochemically in whole tissue
homogenate appears to be useful as an indicator of
the degree of phagocyte infiltration only when there
are not a large number of AcP-containing cellular
structures such as lysosomes, vacuoles and/or
vesicles otherwise present in other non-immune
cells of the tissue, as is the case in the head kidney
(Table 2; Press et al. 1994) but not the intestine
(Bakke-McKellep et al. 2000). Lysozyme activity
appears to be generally a more specific gauge of
phagocyte activity/presence in whole tissue homo-
genate as indicated by the parallel trend of raised
values in the head kidney, kidney and distal
intestine of the GM-soy group.
The lack of significant effect of the soy-contain-
ing diets on distal intestinal lysozyme activity and
the decrease in IgM content compared with the
standard FM diet apparently contradicts our earlier
studies (Krogdahl et al. 2000). In this previous
experiment, however, the levels of both heat-stable
as well as heat-labile anti-nutrients, toxicants and/or
76
2007 The Authors.
Journal compilation
2007
Blackwell Publishing Ltd
Journal of Fish Diseases 2007, 30, 65–79 A M Bakke-McKellep et al. GM-soybeans in feeds for Atlantic salmon
antigens in the cold-pelleted diet containing 10.8%
alcohol extract of soybeans (soybean molasses), was
probably considerably higher than levels found in
the present extruded diets containing 172 g kg
)1
full-fat SBM.
The decreased activities of brush border digestive
enzymes, LAP and maltase in the distal intestine of
nGM and GM-soy fed salmon, as well as the lack of
response in more proximal intestinal regions, is in
agreement with earlier observations (Krogdahl et al.
2003). We found no effect at the soy levels included
in the experimental diets of the present study on the
liver enzymes fructose-1,6-bisphosphatase, an indi-
cator of metabolic function, and EROD, a xeno-
biotic metabolizing enzyme. In rainbow trout,
however, recent studies have indicated that dietary
SBM may cause changes in histomorphology
(Ostaszewska, Dabrowski, Palacios, Olejniczak &
Wieczorek 2005) and protein profile (Martin,
Vilhelmsson, Me
´dale, Watt, Kaushik & Houlihan
2003) of the liver. Further study with higher dietary
levels is needed to elucidate whether hepatic
changes may be a problem in SBM-fed Atlantic
salmon.
The results presented here support the findings of
the first paper reporting results from this 3-month
feeding trial (Hemre et al. 2005); up to 130 g kg
)1
protein from GM (RRS) soybeans can probably be
safely used as a substitute for traditional SBMs in
practical diets for Atlantic salmon. As we were
unable to acquire the non-GM parent line of the
GM-soy used in this study, it is not possible to
attribute the cause of any differences observed
between the nGM and GM-soy groups to the
genetic modification or differing nutrient and anti-
nutrient compositions of the two soybean varietals
caused by genetic and/or environmental influences.
Thus, pending results from studies carried out using
non-modified, near-isogenic parental line soybeans
as the control group may give an indication of
whether genetic modification of soybeans poses any
risk to farmed Atlantic salmon when added to their
feed.
Acknowledgements
We would like to acknowledge and thank Ellen
Hage, Elin Valen, Kristin Vekterud, Gunn Østby,
and Birgit Røe, our laboratory technicians at the
Norwegian School of Veterinary Science, without
whom we could not have completed this study. Ivar
Hordvik of the University of Bergen (Norway)
contributed in providing the MHC class II bchain
antisera. The authors would also like to acknow-
ledge the following for their help with analyses: Eva
Gelencer and Gyo
¨ngyi Hajos at the Central Food
Research Institute (Budapest, Hungary) for trypsin
inhibitor and lectin, Stephan Sahlstrøm at Matforsk
(A
˚s, Norway) for carbohydrate fraction and phytate,
Patricia Murphy at Iowa State University (USA) for
saponin, Anne Sundby (Norwegian School of
Veterinary Science) for insulin analyses, Marianne
Theodorsen Werner at the National Veterinary
Institute (Oslo, Norway) for phytoestrogen, and
Helene Mathisen, currently at the Norwegian
University of Science and Technology (Trondheim,
Norway), for EROD analyses. Careful fish main-
tenance by Kent Olav Mikkelsen (Norwegian
Institute of Fisheries and Aquaculture Research,
Austevoll, Norway) is highly appreciated. The
research was funded by the Norwegian Research
Council, grants no. 142474/140 and 145949/120.
References
Afonso A., Ellis A.E. & Silva M.T. (1997) The leucocyte
population of the unstimulated peritoneal cavity of rainbow
trout (Oncorhynchus mykiss). Fish and Shellfish Immunology 7,
335–348.
Amin A.B., Mortensen L. & Poppe T.T. (1992) Histology Atlas:
Normal Structure of Salmonids: A Colour Atlas. Akvapatologisk
Laboratorium AS, Bodø, Norway.
Bach Knudsen K.E. (1997) Carbohydrate and lignin contents of
plant materials used in animal feeding. Animal FeedScience
Technology 67, 319–338.
Baeverfjord G. & Krogdahl A. (1996) Development and re-
gression of soybean meal induced enteritis in Atlantic sal-
mon, Salmo salar L., distal intestine: a comparison with the
intestines of fasted fish. Journal of Fish Diseases 19, 375–
387.
Bakke-McKellep A.M., Press C.M., Krogdahl A. & Landsverk T.
(2000) Changes in immune and enzyme histochemical phe-
notypes of cells of the intestinal mucosa of Atlantic salmon,
Salmo salar L., with soybean-meal induced enteritis. Journal of
Fish Diseases 23, 115–127.
Borrebaek B., Waagbø R., Christophersen B., Tranulis M.A. &
Hemre G.-I. (1993) Adaptable hexokinase with low affinity
for glucose in the liver of Atlantic salmon (Salmo salar).
Comparative Biochemistry and Physiology 106B, 833–836.
Brake D.G. & Evenson D.P. (2004) A generational study of
glyphosate-tolerant soybeans on mouse fetal, postnatal, pub-
ertal and adult testicular development. Food and Chemical
Toxicology 42, 29–36.
Brown P.B., Wilson K., Jonker Y. & Nickson T.E. (2003)
Glyphosate tolerant canola meal is equivalent to the parental
line in diets fed to rainbow trout. Journal of Agricultural and
Food Chemistry 51, 4268–4272.
77
2007 The Authors.
Journal compilation
2007
Blackwell Publishing Ltd
Journal of Fish Diseases 2007, 30, 65–79 A M Bakke-McKellep et al. GM-soybeans in feeds for Atlantic salmon
Burrells C., Williams P.D., Southgate P.J. & Crampton V.O.
(1999) Immunological, physiological and pathological
responses of rainbow trout (Oncorhynchus mykiss) to increasing
dietary concentrations of soybean proteins. Veterinary
Immunology and Immunopathology 72, 277–288.
Carlsson N.G., Bergman E.L., Skoglund E., Hasselblad K. &
Sandberg A.S. (2001) Rapid analysis of inositol phosphates.
Journal of Agricultural and Food Chemistry 49, 1695–1701.
Cromwell G.L., Lindemann M.D., Randolph J.H., Parker G.R.,
Coffey R.D., Laurent K.M., Armstrong C.L., Mikel W.B.,
Stanisiewski E.P. & Hartnell G.F. (2002) Soybean meal from
roundup ready or conventional soybeans in diets for growing-
finishing swine. Journal of Animal Science 80, 708–715.
Dahlquist A. (1970) Assay of intestinal disaccharidases. Enzym-
ologia Biologica et Clinica 11, 52–66.
Ellis A.E. (1990) Lysozyme assays. In: Techniques in Fish
Immunology (ed. by J.S. Stolen, T.C. Fletcher, D.P. Anderson,
B.S. Robertsen & W.B. van Muiswinkel), pp. 101–104. SOS
Publications, Fair Haven, NJ, USA.
Englyst H.J., Quigley M.E., Hudson G.J. & Cummings J.H.
(1992) Determination of dietary fiber as non-starch poly-
saccharides by gas-liquid chromatography. Analyst 117,
1707–1714.
Ewen S.W.B. & Pusztai A. (1999) Effect of diets containing
genetically modified potatoes expressing Galanthus nivalis
lectin on rat small intestine. The Lancet 354, 1353–1354.
Hajo
´s G., Gelencse
´r E., Pusztai A., Grant G., Sakhri M. &
Bardocz S. (1995) Biological effects and survival of trypsin
inhibitors and the agglutinin from soybean in the small
intestine of the rat. Journal of Agricultural and Food Chemistry
43, 165–170.
Hammond B.G., Vicini J.L., Hartnell G.F., Naylor M.W.,
Knight C.D., Robinson E.H., Fuchs R.L. & Padgette S.R.
(1996) The feeding value of soybeans fed to rats, chickens,
catfish and dairy cattle is not altered by genetic incorpora-
tion of glyphosate tolerance. Journal of Nutrition 126, 717–
727.
Hemre G.-I., Sanden M., Bakke-McKellep A.M., Sagstad A. &
Krogdahl A. (2005) Growth, feed utilization and health of
Atlantic salmon Salmo salar L. fed genetically modified com-
pared to non-modified commercial hybrid soybeans. Aqua-
culture Nutrition 11, 157–167.
Hu J., Lee S.-O., Hendrich S. & Murphy P.A. (2002) Quanti-
fication of the group B soyasaponins by high-performance
liquid chromatography. Journal of Agricultural and Food
Chemistry 50, 2587–2594.
van den Ingh T.S.G.A.M. & Krogdahl A. (1990) Negative effects
of anti-nutritional factors from soybeans in Salmonidae.
Tijdschrift Diergeneeskd 15, 935–938 (in Dutch).
van den Ingh T.S.G.A.M., Krogdahl A
˚., Olli J., Hendricks
H.G.C.J.M. & Koninkx J.F.J.G. (1991) Effects of soybean-
containing diets on the proximal and distal intestine in
Atlantic salmon (Salmo salar): a morphological study. Aqua-
culture 94, 297–305.
Koppang E.O., Bjerka
˚s E., Bjerka
˚s I., Sveier H. & Hordvik I.
(2003a) Vaccination induces major histocompatibility com-
plex class II expression in the Atlantic salmon eye. Scandina-
vian Journal of Immunology 58, 9–14.
Koppang E.O., Hordvik I., Bjerka
˚s I., Torvund J., Aune L.,
Thevarajan J. & Endresen C. (2003b) Production of rabbit
antisera against recombinant MHC class II bchain and
identification of immunoreactive cells in Atlantic
salmon (Salmo salar). Fish and Shellfish Immunology 14,
115–132.
Koppang E.O., Haugarvoll E., Hordvik I., Poppe T.T. &
Bjerka
˚s I. (2004) Granulomatous uveitis associated with vac-
cination in the Atlantic salmon. Veterinary Pathology 41,
122–130.
Koppang E.O., Haugarvoll E., Hordvik I., Aune L. & Poppe
T.T. (2005) Vaccine-associated granulomatous inflammation
and melanin accumulation in Atlantic salmon, Salmo salar L.,
white muscle. Journal of Fish Diseases 28, 13–22.
Krogdahl A
˚., Bakke-McKellep A.M., Røed K.H. & Baeverfjord
G. (2000) Feeding Atlantic salmon Salmo salar L. soybean
products: effects on disease resistance (furunculosis), and ly-
sozyme and IgM levels in the intestinal mucosa. Aquaculture
Nutrition 6, 77–84.
Krogdahl A
˚., Bakke-McKellep A.M. & Baeverfjord G. (2003)
Effects of graded levels of standard soybean meal on intestinal
structure, mucosal enzyme activities, and pancreatic response
in Atlantic salmon (Salmo salar L.). Aquaculture Nutrition 9,
361–371.
Lall S.P. & Bishop F.J. (1977) Studies on mineral and protein
utilization by Atlantic salmon (Salmo salar L.) grown in sea
water. Fisheries and Marine Service Research and Development
Technical Report 688, 16.
Liener I.E. (1994) Implications of anti-nutritional components
in soybean foods. Critical Reviews in Food Science and Nutri-
tion 34, 31–67.
Lund T., Gjedrem T., Bentsen H.B., Eide D.M., Larsen H.J.S.
& Røed K.H. (1995) Genetic variation in immune parameters
and associations to survival in Atlantic salmon. Journal of Fish
Biology 46, 748–758.
Malatesta M., Caporaloni C., Rossi L., Battistelli S., Rocchi
M.B.L., Tonucci F. & Gazzanelli G. (2002) Ultrastructural
analysis of pancreatic acinar cells from mice fed on ge-
netically modified soybean. Journal of Anatomy 201, 409–
415.
Martin S.A.M., Vilhelmsson O., Me
´dale F., Watt P., Kaushik S.
& Houlihan D.F. (2003) Proteomic sensitivity to dietary
manipulations in rainbow trout. Biochimica et Biophysica Acta
– Proteins and Proteomics 1651, 17–29.
Nordrum S., Bakke-McKellep A.M., Krogdahl A
˚. & Buddington
R.K. (2000) Effects of soybean meal and salinity on intestinal
transport of nutrients in Atlantic salmon (Salmo salar L.) and
rainbow trout (Oncorhynchus mykiss). Comparative Biochem-
istry and Physiology 125B, 317–335.
NRC (1993) Nutrient Requirement of Fish (ed. by R.P. Wilson).
National Research Council, National Academy Press,
Washington, DC., USA.
Olli J.J. & Krogdahl A. (1994) Nutritive-value of 4 soybean
products as protein-sources in diets for rainbow-trout
(Oncorhynchus-mykiss, Walbaum) reared in fresh-water. Acta
Agriculturae Scandinavica (Animal Science) 44, 185–192.
Olli J.J., Krogdahl A
˚., van den Ingh T.S.G.A.M. & Bratta
˚s L.E.
(1994) Nutritive value of four soybean products in diets for
78
2007 The Authors.
Journal compilation
2007
Blackwell Publishing Ltd
Journal of Fish Diseases 2007, 30, 65–79 A M Bakke-McKellep et al. GM-soybeans in feeds for Atlantic salmon
Atlantic salmon (Salmon salar L.). Acta Agriculturae Scandi-
navica (Animal Science) 44, 50–60.
Ostaszewska T., Dabrowski K., Palacios M.E., Olejniczak M.
& Wieczorek M. (2005) Growth and morphological chan-
ges in the digestive tract of rainbow trout (Oncorhynchus
mykiss) and pacu (Piaractus mesopotamicus) due to casein
replacement with soybean proteins. Aquaculture 245, 273–
286.
Padgette S.R., Taylor N.B., Nida D.L., Bailey M.R.,
MacDonald J., Holden L.R. & Fuchs R.L. (1996) The
composition of glyphosate-tolerant soybean seeds is equiva-
lent to that of conventional soybeans. Journal of Nutrition
126, 702–716.
Plisetskaya E., Dickhoff W.D., Paquette T.L. & Gorbman A.
(1986) The assay of salmon insulin by homologous radio-
immunoassay. Fish Physiology and Biochemistry 1, 37–43.
Press C.M., Dannevig B.H. & Landsverk T. (1994) Immune and
enzyme histochemical phenotypes of lymphoid and non-
lymphoid cells within the spleen and head kidney of Atlantic
salmon (Salmo salar L.). Fish and Shellfish Immunology 4,
79–93.
Refstie S., Storebakken T., Baeverfjord G. & Roem A.J. (2001)
Long-term protein and lipid growth of Atlantic salmon (Salmo
salar) fed diets with partial replacement of fish meal by soy
protein products at medium or high lipid level. Aquaculture
193, 91–106.
Sørensen M., Ljøkjel K., Storebakken T., Shearer K.D. & Skrede
A. (2002) Apparent digestibility of protein, amino acids and
energy in rainbow trout (Oncorhynchus mykiss) fed a fish meal
based diet extruded at different temperatures. Aquaculture
211, 215–225.
Sansonetti J.P. (2004) War and peace at mucosal surfaces. Nature
Reviews Immunology 4, 953–964.
Sissons J.W. (1982) Ethanol denaturation of soya bean protein
antigens. Journal of the Science of Food and Agriculture 33, 706.
Stagg R.M. & Addison R.F. (1995) An inter-laboratory com-
parison of measurements of ethoxyresorufin O-de-ethylase
activity in dab (Limanda limanda) liver. Marine Environmental
Research 40, 93–108.
Storebakken T., Shearer K.D. & Roem A.J. (1998) Availability
of protein, phosphorus and other elements in fish meal, soy-
protein concentrate and phytase-treated soy-protein-con-
centrate-based diets to Atlantic salmon, Salmo salar.Aqua-
culture 161, 365–379.
Tacon A.G.J. (1987) The Nutrition and Feeding of Farmed Fish
and Shrimp – A Training Manual: 1. The Essential Nutrients,p.
117. Food and Agriculture Organization of the United
Nations, Brazil, GCP/RLA/075/ITA.
Received: 26 April 2006
Revision received: 18 September 2006
Accepted: 18 September 2006
79
2007 The Authors.
Journal compilation
2007
Blackwell Publishing Ltd
Journal of Fish Diseases 2007, 30, 65–79 A M Bakke-McKellep et al. GM-soybeans in feeds for Atlantic salmon
