Time-related changes of the intestinal morphology of Atlantic salmon, Salmo salar L., at two different soybean meal inclusion levels P A Urµn 1,2 , J W Schrama 1 , J H W M Rombout 2 , J J Taverne-Thiele 2 , A Obach 3 , W Koppe 3 and J A J Verreth 1 1 Aquaculture and Fisheries Group, Wageningen Institute of Animal Sciences, Wageningen University, Wageningen, The Netherlands 2 Cell Biology and Immunology Group, Wageningen Institute of Animal Sciences, Wageningen University, Wageningen, The Netherlands 3 Skretting, Aquaculture Research Centre, Stavanger, Norway Abstract Soybean meal (SBM) induces enteritis in the distal intestine of Atlantic salmon. The present study as- sesses the effects of SBM concentrations on the kinetics of the enteritis process. Fish of 300 g, kept at 12 °C, were fed diets with different SBM inclusions: 0%, 10% and 20% SBM for 57 days. Samples of the distal intestine of five fish per treatment were taken for histological and electron microscopical analysis. A semi-quantitative scoring system was used to assess the degree of the mor- phological changes induced by SBM feeding in the distal intestinal epithelium. The first signs of enteritis appeared earlier in the salmon fed the 20SBM diet than in those fed the 10SBM diet. Thereafter, the condition increased steadily, dis- playing no signs of recovery. Furthermore, at the lower concentration, the process marking the onset of enteritis began more gradually than at the higher concentration and it displayed a tendency to stabi- lize after 13–20 days of continuous feeding. Elec- tron microscopy indicated that the endocytosis process was hampered at day 3 of 20SBM and at 7 days of 10SBM. Furthermore, a strong reduction of microvilli was already evident after 7 days of 20SBM feeding, thus indicating a decreased uptake capacity of the distal enterocytes. In addition, transformation and migration of eosinophilic granulocytes was observed, which, in combination with the lysozyme C immunoreactivity supports their protective role during the inflammatory pro- cess in the distal gut of Atlantic salmon. It can be concluded that the severity of enteritis and its kinetics are concentration-dependent, showing no signs of recovery during feeding with diets con- taining SBM. Keywords: endocytosis, enteritis, eosinophilic gran- ulocytes, kinetics, soybean meal, supranuclear vac- uoles. Introduction Soybean is widely known to contain adverse anti- nutritional compounds that may induce intestinal disorders in salmonids, being especially harmful to Atlantic salmon, Salmo salar L. Rainbow trout, Oncorhynchus mykiss (Walbaum), seems to be less affected by the soybean noxious factors that induce enteritis, as documented in previous studies (Nor- drum, Bakke-Mckellep, Krogdahl & Buddington 2000; Refstie, Korsøen, Strorebakken, Baeverfjord, Lein & Roem 2000; Buttle, Burrells, Good, Williams, Southgate & Burrells 2001). When Atlantic salmon is fed on soybean meal (SBM)- based diets, the morphology of the distal intestine is disturbed, described by Baeverfjord & Krogdahl (1996) as Ônon-infectious sub-acute enteritisÕ. The Journal of Fish Diseases 2009, 32, 733–744 doi:10.1111/j.1365-2761.2009.01049.x Correspondence J W Schrama, Aquaculture and Fisheries group, Wageningen University PO Box 338, 6700 AH Wageningen, The Netherlands (e-mail: [email protected]) 733 Ó 2009 The Authors. Journal compilation Ó 2009 Blackwell Publishing Ltd
Transcript
Time-related changes of the intestinal morphology of
Atlantic salmon, Salmo salar L., at two different soybean
meal inclusion levels
P A Ur�n1,2, J W Schrama1, J H W M Rombout2, J J Taverne-Thiele2, A Obach3, W Koppe3
and J A J Verreth1
1 Aquaculture and Fisheries Group, Wageningen Institute of Animal Sciences, Wageningen University, Wageningen,
The Netherlands
2 Cell Biology and Immunology Group, Wageningen Institute of Animal Sciences, Wageningen University,
Wageningen, The Netherlands
3 Skretting, Aquaculture Research Centre, Stavanger, Norway
Abstract
Soybean meal (SBM) induces enteritis in the distalintestine of Atlantic salmon. The present study as-sesses the effects of SBM concentrations on thekinetics of the enteritis process. Fish of 300 g, keptat 12 �C, were fed diets with different SBMinclusions: 0%, 10% and 20% SBM for 57 days.Samples of the distal intestine of five fish pertreatment were taken for histological and electronmicroscopical analysis. A semi-quantitative scoringsystem was used to assess the degree of the mor-phological changes induced by SBM feeding in thedistal intestinal epithelium. The first signs ofenteritis appeared earlier in the salmon fed the20SBM diet than in those fed the 10SBM diet.Thereafter, the condition increased steadily, dis-playing no signs of recovery. Furthermore, at thelower concentration, the process marking the onsetof enteritis began more gradually than at the higherconcentration and it displayed a tendency to stabi-lize after 13–20 days of continuous feeding. Elec-tron microscopy indicated that the endocytosisprocess was hampered at day 3 of 20SBM and at7 days of 10SBM. Furthermore, a strong reductionof microvilli was already evident after 7 days of20SBM feeding, thus indicating a decreased uptake
capacity of the distal enterocytes. In addition,transformation and migration of eosinophilicgranulocytes was observed, which, in combinationwith the lysozyme C immunoreactivity supportstheir protective role during the inflammatory pro-cess in the distal gut of Atlantic salmon. It can beconcluded that the severity of enteritis and itskinetics are concentration-dependent, showing nosigns of recovery during feeding with diets con-taining SBM.
Soybean is widely known to contain adverse anti-nutritional compounds that may induce intestinaldisorders in salmonids, being especially harmful toAtlantic salmon, Salmo salar L. Rainbow trout,Oncorhynchus mykiss (Walbaum), seems to be lessaffected by the soybean noxious factors that induceenteritis, as documented in previous studies (Nor-drum, Bakke-Mckellep, Krogdahl & Buddington2000; Refstie, Korsøen, Strorebakken, Baeverfjord,Lein & Roem 2000; Buttle, Burrells, Good,Williams, Southgate & Burrells 2001). WhenAtlantic salmon is fed on soybean meal (SBM)-based diets, the morphology of the distal intestine isdisturbed, described by Baeverfjord & Krogdahl(1996) as �non-infectious sub-acute enteritis�. The
Journal of Fish Diseases 2009, 32, 733–744 doi:10.1111/j.1365-2761.2009.01049.x
Correspondence J W Schrama, Aquaculture and Fisheries
changes in the distal intestinal mucosa are describedas a deep shortening of the mucosal folds (MF), adecreasing number of supranuclear vacuoles (SNV)in absorptive cells, a widening of the central stromawith a correspondingly high amount of connectivetissue and an increased infiltration of inflammatorycells in the lamina propria (LP) (van den Ingh,Krogdahl, Olli, Hendriks & Koninkx 1991; Bae-verfjord & Krogdahl 1996; van den Ingh, Olli &Krogdahl 1996), an increased number of gobletcells (GC), and a shortening of the microvilli (Mv)(van den Ingh et al. 1991). These pathologicalchanges seem to be particularly present in the distalintestinal segment rather than in the proximalsegment, as already reported in several studies onsalmonids (van den Ingh et al. 1991; Burrells,Williams, Southgate & Crampton 1999; Nordrumet al. 2000; Buttle et al. 2001).
Baeverfjord & Krogdahl (1996) established thatafter 3 weeks of SBM-feeding, the development ofenteritis became critical for those Atlantic salmonfed a relatively high SBM-based diet (c. 33%) whichinduced all the above-mentioned characteristic signsof enteritis within the first week of experimentalfeeding. A more recent study (Krogdahl, Bakke-McKellep & Baeverfjord 2003), using differentsolvent-extracted SBM inclusion levels, indicatedthat the degree of enteritis in the distal intestine,measured at 60 days of SBM feeding, was increasedwith increasing SBM. Furthermore, even at lowerSBM levels (15–25%), growth performance, feedconversion, apparent digestibility/utilization of ma-cronutrients and energy were affected. Almost allstudies on enteritis focus on the more advancedstages of the inflammation process (>20 days afterSBM feeding), with the exception of the kineticstudy of Baeverfjord & Krogdahl (1996) whichused a 33% dietary inclusion solvent extractedSBM. They found the initial signs of morphologicalchanges at day 2 after feeding the SBM diet, and atday 7 all typical characteristics of enteritis werepresent but the severity increased up until the lastsampling point at 20 days. Thus, morphologicalinformation on the onset of enteritis is limited aswell as on the influence of the SBM level on thekinetics of enteritis. Information on the ultrastruc-tural changes of the distal epithelium is lacking,both regarding the kinetics and the influence ofSBM concentrations.
In this study, the kinetics of morphologicalchanges is assessed in fish fed diets containing 10%and 20% of SBM as part of the protein fraction.
The main objective of this study was to describe theprogressive morphological changes in the distalintestine of salmon fed diets containing differentinclusion levels of a SBM variety positively selectedto give a strong reaction. Morphological parameterscharacteristic for inflamed distal intestinal mucosawere assessed at seven different time points over theduration of the experimental period. A previouslyintroduced semi-quantitative scoring system (Uran,Schrama, Rombout, Obach, Jensen, Koppe &Verreth 2008) was used to elucidate the impacton the intestinal morphology. In addition, as well aspaying attention to ultrastructural changes duringthe enteritis process, a preliminary qualitative studyon the role and contribution of the eosinophilicgranulocytes (EG) to the inflammatory process isalso presented.
Materials and methods
Fish and rearing conditions
The experiment was carried out at the SkrettingFish Trials Station, Lerang (Jørpeland, Norway).For this experiment, 300 Atlantic salmon (Aqu-aGen strain) were sampled for gut histology andelectron microscopy (EM). The Atlantic salmonoriginated from a stock of fish already present at theresearch station. The experiment consisted of a2-week acclimatization period and an 8-weekexperimental period. At the start of the acclimati-zation period the fish weighed approximately 300 g.
For the experiment six indoor tanks with adiameter of 1 m each were used. The water volumein the tanks was 400 L with a flow through of12–15 L min)1. The stocking density was 50 fishper tank. Sea water was pumped from a depth of90 m in the fjord, with a salinity of 34& and anoxygen concentration above 9 ppm. Prior to theacclimatization period fish were kept at 8 �C.During the acclimatization and the experimentalperiod, water temperature was kept at 12 �C. Thephotoperiod was 18 L:6D.
Diets and feeding
Feed was produced at the Skretting Feed Technol-ogy Plant (Stavanger, Norway). Three diets wereformulated: a control diet (0SBM) and two exper-imental diets (10SBM and 20SBM) (Table 1). Themajor ingredients in the 0SBM diets were: fishmeal(protein content above 70%), fish oil and wheat.
734
� 2009 The Authors.
Journal compilation
� 2009
Blackwell Publishing Ltd
Journal of Fish Diseases 2009, 32, 733–744 P A Uran et al. Soybean meal and kinetics of enteritis in salmon
This control diet did not contain any SBM. For theexperimental diets, fishmeal and wheat were ex-changed, in the case of the 10SBM diet for 10%SBM, and in the case of the 20SBM diet for 20%SBM (solvent-extracted Hipro SBM) (Table 1). Alldiets were produced to be iso-energetic and iso-nitrogenic on a crude protein and a crude lipidbasis. Diets were supplemented with a standardvitamin and mineral premix. Feed was produced asextruded 4 mm sinking pellets.
Prior to the experiment the fish were fed acommercial salmon diet (Skretting, Stavanger,Norway), which did not contain any soybeanproducts. During the acclimatization period alltanks were fed with the control diet (0SBM). At thestart of the experimental period (day 1), for eachdiet group, two tanks were changed to either a10SBM or a 20SBM experimental diet. Two tanksremained at the control diet. Fish were fed 20% inexcess. Feed was divided into two meals per day andprovided by automatic feeders.
Chemical analysis of diets
The nutrient composition of the experimental dietswas determined using standard techniques forproximate analyses. Crude protein content wasdetermined by the Kjeldahl nitrogen measurementin accordance with the Nordic Committee on Food
Analysis, Method No. 6, 4th edn, 2003. Crude fatcontent was measured by low field nuclear magneticresonance. Moisture content in the samples wasmeasured by drying to constant weight at 102–105 �C for 16–18 h. Ash content was measured bycombustion at 540 �C for 16–18 h, after which theremaining residues were weighed, both in accor-dance with the Nordic Committee on FoodAnalysis Method No. 23, 3rd edn, 1991. Thepreceding analyses were carried out at the SkrettingAquaculture Research Centre (Stavanger, Norway).For chemical composition of diets see Table 1.
Sampling for intestinal morphology
During the experimental period, fish gut wassampled for histological measurements at days 3,5, 7, 13, 20, 36 and 57. At each sampling point,five fish per treatment group were sampled (five fishfrom the control group and five from each of thetwo experimental diets). Removing some of the fishfrom a group will probably induce stress in theremaining fish, which might influence their feedintake. Such stress brought about through excesshandling would influence the actual amount ofSBM eaten by the remaining fish and mightconsequently affect the degree of enteritis at thenext sampling point. As a result of the limitedamount of tanks available for this experiment,together with the short-time interval betweensampling moments, it was decided to sample tankswithin dietary treatments alternatively for the sevendifferent sampling points (i.e. one tank at each dietwas sampled at days 3, 7, 20 and 57 and the othertank on the remaining days). Despite the advantageof reducing the level of stress exposure to the fishand thereby possibly reducing feed intake, thissampling procedure may have resulted in a con-founding between sampling point and tank. How-ever, additional statistical analysis suggested thatthere were no systematic differences between tankswithin dietary treatment for the measured param-eters. There was a slight trend for differencesbetween affected tanks for SNV with the 10SBMtreatment, which was reflected by a very slight butsystematic fluctuating pattern between consecutivesampling points (Fig. 1). These findings indicatethat the presence of a confounding effect betweensampling moment and tank was absent and/ornegligible. However, to be sure that such aconfounding effect did not bias the statisticalanalysis the measured parameters were statistically
Table 1 Formulation and chemical composition of the exper-
imental diets (SBM, soybean meal)
Dietsa
0SBM 10SBM 20SBM
Ingredients (g kg)1)
Fishmealb 564.3 519.8 475.3
Soybean mealc 0 100 200
Wheat 210.6 140 70
Fish oild 222.6 237.3 252.1
Vitamin premix 1.3 1.3 1.3
Mineral premix 1.3 1.3 1.3
Pigment premix
Yttrium oxide 0.1 0.1 0.1
Carophyll pink 0.6 0.6 0.6
Total 1000.8 1000.4 1000.7
Chemical composition by analysis (g kg)1)
Crude protein 429.8 440.2 450.6
Crude fat 277.1 289.2 301.4
Moisture 51.0 62.0 47.0
Ash 72.0 70.0 73.0
aAmount of fishmeal replaced by soybean meal in percent.bLT North Atlantic, Egersund, Norway.cSolvent-extracted soybean meal, protein content above 70%.dNorthern hemisphere.
735
� 2009 The Authors.
Journal compilation
� 2009
Blackwell Publishing Ltd
Journal of Fish Diseases 2009, 32, 733–744 P A Uran et al. Soybean meal and kinetics of enteritis in salmon
analysed, separately, for effect of diet for eachsampling point.
Directly after the morning meal, the fish wereanaesthetized using 0.05 g L)1 metacaine (ArgentChemical Laboratories), and thereafter killed by asharp blow to the head. The distal intestine wasdissected from the point where the intestinaldiameter increases, the mucosa becomes darkerand annular rings are clearly noticeable.
Light microscopy
For light microscopy (LM) analysis a 2-cm sectionof distal intestine of each fish was taken and gentlyrinsed with cold (4 �C) saline. Samples were fixedin 4% phosphate-buffered formalin, pH 7.2, andstored at room temperature. After dehydration, inaccordance with standard procedures, samples wereembedded in paraffin. Transverse sections of 5 lm
thickness were stained using a mixture of haemat-oxylin & eosin and Alcian blue, pH 2.5. Alcian bluestaining enhances the contrast between GC andthe SNV. Slides were blindly evaluated afterrandomization.
Semi-quantitative scoring system
The LM sections were evaluated according to thesemi-quantitative method developed at Wagenin-gen University (Uran et al. 2008), which assessesthe degree of SBM-induced enteritis in the Atlanticsalmon distal intestine using the following param-eters: 1, the morphology of the MF; 2, the presenceand size of SNV; 3, the abundance of GC; 4, theinfiltration of EG into the LP and sub-epithelialmucosa (SM); 5, the degree of widening of the LP;and 6, the degree of thickening of the SM. Sectionswere photographed with an Olympus DP 50 digital
Figure 1 Enteritis parameters scored using the Wageningen semi-quantitative scoring system (Uran et al. 2008), where 1 indicates
normal intestinal morphology and 5 indicates the highest degree of soybean meal (SBM)-induced enteritis in the distal intestine of
Atlantic salmon. The changes in the morphology of the MF, the presence of the supranuclear vacuoles (SNV), the abundance of goblet
cells (GC), the degree of infiltration of eosinophilic granulocytes (EG), the widening of the lamina propria (LP), and the thickening of
the sub-epithelial mucosa (SM), are shown. Three different diets, a control diet containing fishmeal as protein source and 0% soybean
meal (0SBM), and diets containing 10% and 20% soybean meal (10SBM and 20SBM, respectively), both as replacements of fishmeal in
the diet, were fed for a period of 57 days.
736
� 2009 The Authors.
Journal compilation
� 2009
Blackwell Publishing Ltd
Journal of Fish Diseases 2009, 32, 733–744 P A Uran et al. Soybean meal and kinetics of enteritis in salmon
camera connected to a Nikon Microphot-FXA lightmicroscope (Badhoevedorp). Images were processedand analysed using the AnalySiS Extended Pro 3.1
software (Soft Imaging System GmbH). Each of theSBM-induced enteritis parameters was scored on ascale from 1 to 5 (Table 2). An increasing scoredvalue represents a more severe enteritis condition(for illustrations of the different scores, see the listof special publication attachments at http://www.a-fi.wur.nl/UK/Publications/).
Electron microscopy
For EM small pieces of intestine were sampled andfixed with 2.5% glutaraldehyde in 0.1 m phosphatebuffer, pH 7.2, at 4 �C for 4–6 h; then rinsed twicewith 0.1 m phosphate buffer, pH 7.2, plus 0.2 m
sucrose and kept in the same buffer at 4 �C fortransport. Samples were post-fixed on 1% OsO4 for
1–2 h and thereafter embedded in Epon 812(Electron Microscopy Sciences). Semi-thin sections(1 lm) were mounted on glass slides and stainedwith toluidine blue. Ultra-thin sections were cutusing a diamond knife installed on a ReichertUltracut (Leica) and mounted on copper grids.Sections were counter-stained using uranyl acetatefollowed by lead citrate and examined in a PhilipsEM 208 electron microscope (Philips). As a resultof the limitations of this technique, two animals pertreatment and one for the control were embeddedin Epon and further processed. To exclude localdifferences, two different parts of the same samplewere analysed. Images were made using a SISMegaview III digital camera (Soft Imaging SystemGmbH) connected to the Philips EM 208. Theimages were processed and analysed using theAnalySiS Extended Pro 3.1 software.
Immunohistochemistry (lysozyme C)
Transverse sections of distal intestine from controland SBM-fed fish were deparaffinized and hydrated,following standard procedures. Sections were trea-ted with Tris–EDTA buffer, pH 9, and kept at95 �C for 20 min. Sections were then washed inphosphate buffer solution (PBS) with 0.1% TritonX-100 (PBS-t) and kept at room temperature.Afterwards, sections were incubated with 5%normal goat serum for 30 min and then incubatedwith anti-human lysozyme C (1:50; Dako) for 1 h.After washing twice with PBS-t, sections wereincubated with horseradish peroxidase-labelledgoat-anti-rabbit-HRP (GAR-HRP; Dako) for 1 h.Slides were washed twice with PBS and kept in0.05 m, pH 5, sodium acetate buffer for 10 min.Slides were incubated with 3-amino-9-ethyl-carba-zole substrate (Sigma-Aldrich) until a reaction wasobserved and then they were counter-stained withhaematoxylin, rinsed in running tap water, andembedded in Kaiser�s glycerine gelatine (Merck).Sections were analysed and photographed with anOlympus DP 50 digital camera connected to aNikon Microphot-FXA light microscope. Theimages were processed and analysed using theAnalySiS Extended Pro 3.1 software.
Statistical analysis
All statistical analyses were made using the SASSystem (SAS 1999). All measured parameters wereanalysed separately per sampling point for the effect
Table 2 Semi-quantitative scoring system for the different
parameters used to assess the degree of enteritis developed by
Atlantic salmon fed a soybean meal-based diet. From Uran et al.(2008)
Score Description
Mucosal folds
(MF)
1 Basal length
2 Some shrinkage and bloating
3 Diffused shrinkage and onset
of tissue disruption
4 Diffused tissue disruption
5 Total tissue disruption
Supranuclear
vacuoles
(SNV)
1 Basal SNV size
2 Some size reduction
3 Diffused size reduction
4 Onset of extinction
5 No SNV
Goblet cells
(GC)
1 Scattered cells
2 Increased number and sparsely
distributed
3 Diffused number widely spread
4 Densely grouped cells
5 Highly abundant and tightly-packed
cells
Eosinophilic
granulocytes
(EG)
1 Few in SM basal small quantity
2 Increased number in SM and some
migration into LP
3 Increased migration into LP
4 Diffused number in LP and SM
5 Dense EG in LP and SM
Lamina propria
(LP)
1 Normal size LP
2 Increased size of LP
3 Medium size LP
4 Large LP
5 Largest LP
Sub-epithelial
mucosa (SM)
1 Normal SM
2 Increased size SM
3 Medium size SM
4 Large SM
5 Largest SM
737
� 2009 The Authors.
Journal compilation
� 2009
Blackwell Publishing Ltd
Journal of Fish Diseases 2009, 32, 733–744 P A Uran et al. Soybean meal and kinetics of enteritis in salmon
of diet. Preliminary analyses showed that the datawere non-normally distributed; therefore, the effectof diet was tested using the non-parametric Krus-kal–Wallis test. Pair-wise comparison was per-formed using the Mann–Whitney U-test. Resultsare considered statistically significant when P-valuesare <0.05.
Results
Morphological changes of the distal intestine
The control diet (0SBM) did not induce any sign ofenteritis. The MF appeared as long finger-likestructures, composed of enterocytes which wereperfectly aligned along the mucosal epithelium. Thenucleus of the enterocytes was located between themid to basal parts of the cell. The enterocytespresented normal round and SNV. The GC werepresent and scattered among the enterocytes. TheLP was a delicate and single thin layer of cellsunderlying the epithelium. The SM found betweenthe basal part of the folds and the stratumcompactum was composed of a basal EG popula-tion forming the stratum granulosum. All controlsamples for the entire duration of the experimentalperiod followed the same pattern as describedabove. This is reflected in Fig. 1 by the fact that allthe semi-quantitative scores remained at the basallevel of 1 for all the measured parameters, at allsampling points.
Fish fed diets containing SBM showed no visiblemorphological changes up to day 3. The first visiblechanges occurred on day 5 for the 20SBM diet. Atthis point, the SNV started to shrink and theirregular alignment was disturbed. An increasing cellinfiltration into the SM was also noticed. At day 5,for the 10SBM diet, there were no noticeablechanges. At day 7 the morphological changesdescribed above were detectable in both SBM dietsbut to a lesser degree for the 10SBM diet (Fig. 1).The MF started to shrink, with subsequent tissuedisruption visible as damaged and disturbed folds.The LP and SM had increased in size at both SBMinclusions. The SNV showed a significant reductionand the number of GC increased steadily, partic-ularly in the 20SBM diet. At day 13 major changeswere seen between the two SBM diets. The fish fedthe 20SBM diet presented a heavily changedintestinal mucosa, while the fish fed 10SBMshowed only minor changes in MF, LP and SM(Fig. 1). The SNV remained normal in size and
regularly aligned whilst the number of GC and EGincreased. From day 13 onwards, all the scoredparameters, except for GC, tended to stabilize infish fed 10SBM. However, for the 20SBM diet,most parameters measured became more severeduring the experimental period (Fig. 1). The MFshowed a mild increase during the first 20 days offeeding but later diffuse tissue disruption occurreduntil they appeared as swollen mucosal tissue. Therewas a rapid reduction in size of SNV towards day20 when the highest scores were reached wherevacuoles were completely absent. The GC becamedensely grouped cells from day 13 onwards. Thenumber of EG increased from day 13 onwardsduring which there was a very large migratoryincrease and a dense concentration observedtowards the LP and in SM, respectively. At day57, the signs observed in fish fed 10SBM hadstabilized but the morphological features of fish fed20SBM reached maximum scores for most of theparameters (Fig. 1). It was observed that the onsetof the enteritis condition began at between 3 and5 days and at day 7 (Table 3) all the parametersevaluated were significantly different from those infish fed the control diet (P < 0.05).
Ultrastructural changes of the distal Intestine
The influence of SBM on the ultrastructure of thedistal intestinal epithelia was first observed on day 3when evidence of disturbed endocytotic activity wasnoticed, particularly among fish fed the 20SBMdiet. The pinocytotic invaginations at the basal partof the Mv were less frequent than in controls. After7 days of 20SBM feeding, these membrane-boundvesicles were not visible and there was no evidenceof any newly engulfed material at the level of the
Table 3 Significance of the signs of enteritis developed by
Journal of Fish Diseases 2009, 32, 733–744 P A Uran et al. Soybean meal and kinetics of enteritis in salmon
terminal web, whereas in controls it was filled withsmall pinocytotic vesicles pinching off from theapical membrane and moving towards the endo-somes to finally fuse into the SNV (Fig. 2). By theend of the experiment on day 57, virtually no SNVwere observed. With the 10SBM diet, these changesall occurred more slowly than at 20SBM. Disap-pearance of the pinocytotic vesicles was first noticedfrom day 7 at 10SBM and after 57 days ofcontinuous feeding SNV were more numerousthan in fish fed 20SBM.
The Mv of fish fed 20SBM had dramaticallydecreased in size, to about one-third of theirnormal length as early as day 7 compared with thecontrols and remained so until the end of theexperiment with no signs of recovery (Fig. 3). Thebrush border membrane of fish fed 10SBM alsoshowed a decrease but to a much lesser extent. Atday 7, less invaginations were observed but Mvwere of normal size compared with the controlsand the SNV were reduced but still present untilthe end of the trial.
There was a clear difference in the appearance ofthe granular content of the EG observed under theEM. During the first 7 days of 20SBM feeding, thedensity of granules began to change and this processcontinued until day 20 with some appearing aselectron lucent in comparison with the electron
dense granules observed in the controls (Fig. 4).Based on visual estimations the quantity of EGseemed to peak after 20 days, with a majority ofintermediate electron dense granules. In fish fed10SBM the electron dense or intermediate densegranules were observed less frequently.
Detection of lysozyme C in EG
Using immunohistochemistry, lysozyme C immu-noreactivity was observed within the granules of theEG present in the LP and SM of the distal gut(Fig. 5).
Discussion
The main objective of this study was to describe theprogressive morphological changes in the intestineof Atlantic salmon fed different SBM inclusionlevels in their diets. A 20% SBM inclusion level ledto a severe inflammation of the distal intestinalepithelium, whereas a 10% SBM inclusion levelinduced a slow and mild reaction. The increasedinflammation with increasing dietary SBM levels isin agreement with the study of Krogdahl et al.(2003) who measured morphological changes after60 days of SBM feeding. The present study furtherdemonstrated that the kinetics of the different
(a)
B
SNV
Mv
v
(b)
Figure 2 Ultrastructure of the distal intestine epithelium of Atlantic salmon fed the control diet. (a) A normally vacuolated enterocyte
with large numbers of pinocytotic vesicles fusing into supranuclear vacuoles (SNV). The microvilli (Mv) are normal in size (bar = 2lm).
(b) Detail of cytoplasmic invaginations (arrow) pinching off the apical membrane to fuse into bigger vesicles (v) (bar = 0.5 lm).
739
� 2009 The Authors.
Journal compilation
� 2009
Blackwell Publishing Ltd
Journal of Fish Diseases 2009, 32, 733–744 P A Uran et al. Soybean meal and kinetics of enteritis in salmon
(e) (f)
(b)
(c)
Mv
(a)
(d)
Mv
Figure 3 Ultrastructure of the distal
intestine epithelium of Atlantic salmon fed
soybean meal (SBM)-based diets. (a) No
changes can be observed after 3 days of
feeding a diet containing 10% SBM
(10SBM) as the protein source. Arrows show
pinocytotic vesicles. (b) Changes after 3 days
of feeding a diet containing 20% SBM as the
protein source (20SBM). The pinocytotic
vesicles have started to diminish (arrow).
(c) After day 7 the first signs of a disrupted
uptake process are observed with the 10SBM
diet, indicated by the decreased formation of
membrane-bound vesicles, however,
microvilli (Mv) are still of normal size.
(d) After 7 days with the 20SBM diet,
vesicles have disappeared and newly engulfed
material was not observed at the apical
cytoplasm. The Mv are reduced in size to
nearly one-third of their normal length.
(e) After day 57 with 10SBM feeding the
changes at the distal intestine epithelium
were not as severe as with 20SBM but there
is still some reduction in size of the brush
border membrane and the absence of newly
formed vesicles at the terminal web. (f) After
57 days with 20% SBM feeding, a severe
reduction in Mv size and a total absence
of new vesicles and SNV are seen
(bar = 2 lm).
740
� 2009 The Authors.
Journal compilation
� 2009
Blackwell Publishing Ltd
Journal of Fish Diseases 2009, 32, 733–744 P A Uran et al. Soybean meal and kinetics of enteritis in salmon
clinical signs differ depending on the inclusionlevel. The 20SBM inclusion level boosted theinflammatory process, leading to a faster andsteadily increasing reaction with more severe change
for most of the parameters assessed. At 10SBM, thesame parameters appeared to stabilize after 13–20 days of continuous SBM feeding. Thus, theresults from this study showed that SBM-inducedenteritis is a concentration-dependent process withno clear signs of recovery.
All morphological signs of enteritis becamevisible only after 7 days of SBM feeding, anobservation that concurs with the LM observationsof Baeverfjord & Krogdahl (1996). However, in thecurrent study, EM revealed the first signs of distalepithelial disturbances after 3 days of 20SBM-feeding. These findings suggest that the endocytosisprocess might be disturbed resulting in the absenceof small vesicles that normally finally fuse into theSNV. Rombout, Lamers, Helfrich, Dekker &Taverne-Thiele (1985) described this process asthe normal transport mechanism in the distal gut ofcommon carp. King, Pusztai & Clarke (1982)observed changes of the intestinal Mv, disruption ofthe terminal web and swelling of the apicalcytoplasm, possibly explained by disturbances inmembrane-associated transport processes, in rats fedkidney-bean lectins. It remains to be investigated as
(a)
(b)
(d)
(c)
(e)
Figure 4 Eosinophilic granulocytes (EG) of
the distal intestine epithelium of Atlantic
salmon. (a) Intestinal EG of fish fed a
control diet (no inclusion of soybean meal);
note the predominant electron dense
granules. (b) The predominant type of EG
with intermediate and electron-lucent
granules, present in the lamina propria and
sub-epithelial mucosa of the distal intestine
epithelium after 7 days of feeding a diet co-
ntaining 10% SBM (10SBM). (c) Day 7 after
feeding a diet containing 20% SBM (20SB-
M), note there are more intermediate and
electron-lucent granules. (d) Appearance of
EG after 57 days of feeding a 10SBM diet,
with electron-lucent granules now the most
abundant type. (e) EG after 57 days of feed-
ing a 20SBM diet, showing predominantly
electron-lucent granules (bar = 2 lm).
Figure 5 Immunohistochemistry using an anti-lysosyme C
serum on the distal intestinal segment of Atlantic salmon fed
with a soybean meal (SBM)-based diet. Immunoreactive lyso-
zyme C cells (arrows) indicate the presence of lysozyme in
eosinophilic granule cells in the lamina propria of the distal
intestine (bar = 20 lm).
741
� 2009 The Authors.
Journal compilation
� 2009
Blackwell Publishing Ltd
Journal of Fish Diseases 2009, 32, 733–744 P A Uran et al. Soybean meal and kinetics of enteritis in salmon
to whether the reduction in microvillous surfacearea at the brush-border membrane and thedecreased pinocytotic uptake in this study arecorrelated with the disappearance of SNV. Krog-dahl & Bakke-McKellep (2005) noted that changesin enzyme activities may also be the result ofchanges in tissue mass, cell numbers, morphologyand enzyme synthesis in absorptive cells. Thus, itmay be speculated that the atrophy of the Mvobserved in this study is the consequence of anunbalanced enzymatic profile caused by noxiousfactors present in soybean fractions. Knudsen,Uran, Arnous, Koppe & Frøkiær (2007) postulatedsoyasaponins or a combination with factors likeantigenic soybean proteins or even with opportu-nistic intestinal microflora to be the causal factorsthat trigger the inflammatory reaction. Some com-ponents present in the diet were suggested tohamper vacuolization probably by disturbing theendocytosis process. The disturbance of such anessential basic mechanism as endocytosis must havea great impact on the functioning of enterocytes,rapidly resulting in the disappearance of SNV, andprobably also a decrease of the height of the Mv. Itcould be that disturbances of the enterocytes resultin the release of danger signals, such as heat-shockproteins (HSP), which may be responsible for therecruitment of leucocytes, and finally increasinginflammation of the connective tissue. The role ofHSP70 in SBM-induced enteritis has been shownalready in Atlantic salmon by Bakke-McKellep,Penn, Salas, Refstie, Sperstad, Landsverk, Ringø &Krogdahl (2007).
An interesting feature observed in this study wasthe ultrastructural change in the EG. Although theirreal function is not completely understood, it issuggested that they are important contributors tothe inflammation process and therefore to innateimmunity (Kodama, Tijiwa, Moritomo & Nakani-shi 2002). In salmonids these EG were namedbecause of their acidophilic nature, so that thegranules stained red with eosin (Ezeasor & Stokoe1980). These cells are widely distributed in con-nective tissue, especially in the gastro-intestinaltissue and gills (Sveinbjørnsson, Olsen & Paulsen1996). Those located in intestinal tract can form aconsiderable basal cell population arranged as acontinuous layer, the so-called stratum granulosum(see Paulsen, Sveinbjørnsson & Robertsen 2001).From the results of the present study, it is suggestedthat these cells undergo an activation process,changing the granular content from electron dense
to electron lucent, in this case as a response to theSBM-induced inflammation. According to Reite(1996), the EG are part of a protective barrieragainst certain substances, in our case the SBM. Asa result of their similar morphology and enzymaticprofile they are suggested to be mammalian mastcell analogues (Ellis 1985), sharing the presence ofbasic proteins and acidic glycosaminoglycans in thecytoplasmic granules, but differing in the absence ofhistamine, leukotrienes C4 and B4 and prostaglan-din D2 (Irani & Schwartz 1989). Like mammalianmast cells, fish EG appear to be abundantly presentat sites of persistent inflammation (see Reite &Evensen 2006). Their granules seem to containantimicrobial substances which are involved in fishnon-specific defence mechanisms (Sveinbjørnssonet al. 1996). This study supports the presence oflysozyme in the EG granules and suggests at least anantimicrobial function for the EG. In combinationwith a strong increase in GC and probably mucusproduction, the secretion of antimicrobial sub-stances by EG may be a very powerful innateimmune reaction during the inflammation process.In any case, this function is supported by theinfiltration of the EG towards the LP of the MF.However, from EM it is still not clear whether thechange in granular content (from dense to lucent) isrelated to the secretion of the granular content. Apossible explanation could be a change in thephysicochemical properties or simply a difference intheir emptying rates of the granules, forming amixed population of heterogeneous maturationstages where certain types are predominant. In thiscase, the more advanced the condition the moreelectron-lucent granules are present. Previous stud-ies have suggested the presence of other antimicro-bial peptides (AMP)-like piscidin in striped basshybrid skin, gill and gut mast cells (Silphaduang &Noga 2001) and pleurocidin in winter flounder gillEG (Murray, Leggiadro & Douglas 2007). Never-theless, the mechanism through which these sub-stances are released, whether this is an activation ora degranulation process, and the way in which itmay be related to changes in the ultrastructure ofthe granules, remains to be clarified.
In conclusion, SBM-induced enteritis is a con-centration-dependent process that increases overtime with no clear signs of recovery. The morpho-logical changes observed in SNV and Mv areindications of a disturbed endocytosis process that,in the long term, might affect resistance againstcommensals and pathogens present in the hind gut.
742
� 2009 The Authors.
Journal compilation
� 2009
Blackwell Publishing Ltd
Journal of Fish Diseases 2009, 32, 733–744 P A Uran et al. Soybean meal and kinetics of enteritis in salmon
This study, related to the kinetics of the inflamma-tory process, lays down criteria for the diagnosis ofthe early development of SBM-induced enteritis. Assoon as the substances inducing these changes areknown, optimal sources of proteins can be selectedfor the optimization of plant-based feed formula-tions for Atlantic salmon.
Acknowledgements
This research was supported by �Instituto Colom-biano para el Desarrollo de la Ciencia y laTecnologıa� Colciencias and Skretting ARC, Stav-anger, Norway. We would like to acknowledge thestaff at both the Skretting Fish Trials Station andSkretting ARC for their technical assistance, fortheir help during the feeding trial, the collection ofthe samples and for the laboratory analyses. Theresults presented are part of the PhD thesis of P.A.Uran.
References
Baeverfjord G. & Krogdahl A. (1996) Development and
regression of soybean meal induced enteritis in Atlantic
salmon, Salmo salar L., distal intestine: a comparison with
the intestines of fasted fish. Journal of Fish Diseases 19, 375–
387.
Bakke-McKellep A.M., Penn M.H., Salas P.M., Refstie S.,
Sperstad S., Landsverk T., Ringø E. & Krogdahl A. (2007)
Effects of dietary soyabean meal, inulin and oxytetracycline on
intestinal microbiota and epithelial cell stress, apoptosis and
proliferation in the teleost Atlantic salmon (Salmo salar L.).
British Journal of Nutrition 97, 699–713.
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. VeterinaryImmunology and Immunopathology 72, 277–288.
Buttle L.G., Burrells A.C., Good J.E., Williams P.D., Southgate
P.J. & Burrells C. (2001) The binding of soybean agglutinin
(SBA) to the intestinal epithelium of Atlantic salmon Salmosalar and rainbow trout, Oncorhynchus mykiss, fed high levels
of soybean meal. Veterinary Immunology and Immunopathology80, 237–244.
Ellis A.E. (1985) Eosinophilic granular cells (EGC) and
histamine responses to Aeromonas salmonicida toxins in
rainbow trout. Developmental and Comparative Immunology 9,
251–260.
Ezeasor D.N. & Stokoe W.M. (1980) A cytochemical, light and
electron microscopic study of the eosinophilic granule cells in
the gut of the rainbow trout, Salmo gairdneri Richardson.
Journal of Fish Biology 17, 619–634.
van den Ingh T.S.G.A.M., Krogdahl A., 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
in Atlantic salmon (Salmo salar): a morphological study.
Aquaculture 94, 297–305.
van den Ingh T.S.G.A.M., Olli J.J. & Krogdahl A. (1996)
Alcohol-soluble components in soybeans cause morphological
changes in the distal intestine of Atlantic salmon, Salmo salarL. Journal of Fish Diseases 19, 47–53.
