Acute stress alters intestinal function of rainbow trout, Oncorhynchus mykiss (Walbaum)

Aquaculture 250 (2005) 480 – 495 www.elsevier.com/locate/aqua-online

Acute stress alters intestinal function of rainbow trout, Oncorhynchus mykiss (Walbaum) Rolf Erik Olsena,T, Kristina Sundellb, Terry M. Mayhewc, Reidar Myklebustd, Einar Ringøa,e a Institute of Marine Research, Matre Aquaculture Research Station, N-5984 Matredal, Norway Fish Endocrinology Laboratory, Department of Zoology/Zoophysiology, University of Go¨teborg, Go¨teborg, Sweden c School of Biomedical Sciences, Queens Medical Centre, University of Nottingham, UK d Institute of Anatomy and Cell Biology, University of Bergen, Bergen, Norway e Section of Veterinary Medicine, Department of Food Safety and Infection Biology, Norwegian School of Veterinary Science, Tromsø, Norway b

Received 5 November 2004; received in revised form 28 February 2005; accepted 2 March 2005

Abstract Groups of rainbow trout (Oncorhynchus mykiss Walbaum) in feeding (guts filled with digesta) or food-deprived (3 days of diet deprivation) states were subjected to 15 min of acute stress. Blood samples and intestinal tissue were collected and prepared for physiological, chemical and ultrastructural analysis immediately before stress, and at 4 and 48 h post-stress. Haematocrit, plasma cortisol and lactate levels increased following stress, and the response appeared to be more pronounced in food-deprived than in fed fish. Plasma glucose appeared to increase more in fed fish. Stress did not seem to cause massive tissue damage as measured by cellular leakage of transaminase enzymes into the blood. Furthermore, the plasma oxidative stress marker malondialdehyde did not increase markedly following stress. The content of malondialdehyde did not change following stress neither in midgut nor hindgut sections, and nor did membrane lipid class and fatty acid compositions. Ultrastructural studies showed that stress caused significant widening of the tight junctions between enterocytes in the midgut, with little effect seen in the hindgut. Fed fish appeared to experience more damage than food-deprived fish as judged by the ultrastructural analysis. But these changes were to a large extent transient and cellular organization in the midgut had returned to normal values within 2 days. Intestinal paracellular permeability of the midgut increased 4 h post-stress and was further increased 48 h post-stress of food-deprived fish, whereas no effect was seen in fed trout. In hindgut of food-deprived fish, the permeability appeared to increase sometime after 4 h and was still elevated 48 h after stress. No change in permeability occurred in fed fish. The adherent microbial population level and composition in hindgut was significantly reduced following stress, while the level increased in faeces. This suggests that substantial amounts of mucus are peeled off following stress. It is concluded that, under the present experimental conditions, acute stress causes cellular alteration in rainbow trout gastrointestinal tract. Ultrastructural damages are mainly observed in midgut, but most changes appear to be transient returning to normal levels within 48 h. Changes in adherent intestine microbial population level and compositions together with increased

T Corresponding author. E-mail address: [email protected] (R.E. Olsen). 0044-8486/$ - see front matter D 2005 Elsevier B.V. All rights reserved. doi:10.1016/j.aquaculture.2005.03.014

R.E. Olsen et al. / Aquaculture 250 (2005) 480–495

481

intestinal paracellular permeability following acute stress suggest a possible increased susceptibility to bacterial infections after stress and that active feeding may have a protective role. D 2005 Elsevier B.V. All rights reserved. Keywords: GI tract; Enterocytes; Histology; Ultrastructure; Damage; Bacteria; Stress metabolism; Blood chemistry; Starvation; Fasting; Barrier function; Permeability

1. Introduction The stress response in fish appears to have many similarities to that of terrestrial vertebrates in both primary and secondary stress responses (Wendelaar Bonga, 1997). From mammalian research it is well known that the gastrointestinal (GI) tract is responsive and sensitive to a wide range of stressors. Some of the more common features are degeneration of the mammalian intestinal mucosa and perturbation of its barrier function and uptake mechanisms (Meddings and Swain, 2000; Santos et al., 2000; Hollander, 1999; Saunders et al., 1994; Sengupta and Sharma, 1993). Although most of these effects have been attributed to endocrine mechanisms (So¨derholm and Perdue, 2001; Yates et al., 2001), the involvement of free radicals and oxidative stress has also been suggested (Prabhu et al., 2000; Bagchi et al., 1999, 1998; Wang and Johnson, 1991). The observed damages may lead to enhanced uptake of potentially noxious materials, e.g. increased bacterial translocation (Tamayo et al., 1996), which may facilitate the induction of infectious diseases. Since fish live in water, perturbation of surface barriers may significantly influence the exchange of substances including facilitated uptake of pathogenic bacteria. Although relatively scarce, most available information seems to point towards a GI tract that is sensitive to at least some types of stressors. In European eel, Anguilla anguilla L., social stress damages the mucosa of the stomach by flattening and disappearance, epithelium atrophies and gastric glands degenerate (Peters, 1982). Furthermore, this study demonstrated that cells developed large intracellular cavities, swollen mitochondria and their cellto-cell contacts were compromised. In carp (Cyprinus carpio, Cyprinus carpio haematopterus), catching and transportation stress cause loss of intestinal goblet (mucus-producing) cells and detachment of columnar absorptive epithelial cells (Szakolczai, 1997). Social

stress is also known to reduce fish appetite and growth in the salmonid fish Arctic char, Salvelinus alpinus L. (Olsen and Ringø, 1999), while acute stress leads to mucus loss, damage to junctional complexes and increased cellular detachment in Atlantic salmon, Salmo salar (L.) (Olsen et al., 2003). Furthermore, stress has been shown to alter membrane fatty acid composition in gilthead seabream (Sparus aurata) (Montero et al., 2001, 1999). The basis for the above-mentioned changes in epithelial morphology and composition as well as the extent to which these alterations may influence intestinal function and permeability are still uncertain. The aims of the present studies were to evaluate the effects of acute stress on the GI tract morphology, paracellular permeability, lipid composition, and gut and faeces microbiota of fed and starved rainbow trout. Possible involvement of free radical mediated processes was also studied.

2. Material and methods Rainbow trout (LA strain—Norwegian breeding programme) were hatched and kept in freshwater (FW) under a continuous simulated daylight regime at Matre Aquaculture Research Station, Norway until a weight of 360 g was achieved. A total of 160 individuals was selected at random, anaesthetized in 0.4% benzocaine and transferred into eight 1.5  1.5  0.6 m standard fibreglass tanks supplied with 800 l of aerated freshwater (10 F 1.5 8C). The fish were allowed to adapt to the new conditions for 6 weeks prior to initiation of the experiments. All fish were fed standard commercial diets (Nor-Aqua Innovation Ltd.) with pellet sizes adjusted to fish size and weight. Before transfer to the smaller tanks, fish were fed to satiation using 24-h disc feeders. After transfer, fish were fed to satiation once every morning using a window of 2 h. This procedure ensured that

482

R.E. Olsen et al. / Aquaculture 250 (2005) 480–495

fish were adapted to eating every morning and that they were under the same level of feeding metabolism (their guts filled) 24 h after the last meal. Every day after feeding, the tanks were cleaned manually for around 30 s as part of the general maintenance. This procedure also habituated the fish to brief handling. Three days prior to initiation of the experiment, fish in four tanks were deprived of diets until initiation of the experiment (food-deprived group), while fish in the other four tanks continued to be fed as usual (fed group). The experiment was initiated by subjecting the fish to acute stress by lowering the water level to 5–10 cm and subsequently chasing the fish with a pole for 15 min (Olsen et al., 2003). Full water flow was maintained throughout the procedure. The fish were then allowed to recover with a minimum of handling. At 4 and 48 h post-stress, three fish from each tank were collected, anaesthetized in 0.4% benzocaine and sampled for blood and intestines as described below. An additional experiment was set up at Department of Zoology/Physiology, Go¨teborg University, for analysis of paracellular permeability in fed and food-deprived fish without stress exposure. Rainbow trout (120 individuals) were raised under commercial conditions at Antens laxodling (Alings3s, Sweden), transferred and kept at the fish facility at Department of Zoology/Zoophysiology, Go¨ teborg University, Sweden, under similar conditions as described above, except that no stress was applied to the fish. At day 0, 12 fish were sampled and examined for paracellular permeability of the midgut (posterior to the pyloric caecae to the ileorectal valve) and hindgut (from ileorectal valve to anus) as described below. Half of the remaining fish (48 individuals) were deprived of food and the rest were fed as before. Sampling and analysis of paracellular permeability (n = 12 for each intestinal region and sampling point) was performed at day 3 (equal to the time point 4 h post-stress) and day 5 (equal to time point 48 h post-stress) in both fed and food-deprived groups. The studies were approved by the Animal Welfare Committee, Norway and the Ethical Committee of Animal Research in Go¨teborg, Sweden (license 265/2002). At 0 (at onset of the experiment), 4 and 48 h poststress, 4 fish from each tank were randomly selected and immediately anaesthetized in 0.4% benzocaine. The netting of the fish took 3–5 s. The approach to the tank was as similar as possible to the procedure used

for tank cleaning, in order to avoid additional stress responses. Blood from two fish was withdrawn from the dorsal aorta using heparinized needles. Plasma was immediately prepared by centrifugation at 11,000 rpm for 1 min, frozen in liquid nitrogen, and further stored at 80 8C prior to analysis. Haematocrits (Hct) were obtained using heparinized microcapillary tubes and a Compur M1100 haematocrit centrifuge. The fish were then killed by a sharp blow to the head. A lateral incision was made and the intestines dissected out. Samples from midgut and hindgut of 2 fish from each tank were collected and fixed in McDowell’s fixative (McDowell and Trump, 1976) for electron microscopic examination as described elsewhere (Olsen et al., 1999, 2000). The remaining parts of the intestines were opened longitudinally with the luminal side facing upwards and gently washed 3 times in ice-cold saline solution to remove luminal contents. The mucosa was then lightly scraped off using a glass slide, transferred to cryotubes and immediately frozen in liquid nitrogen and further stored at 80 8C prior to analysis. All treatments were performed on ice and took ca. 2 min to complete after removal of the intestine from the fish. The paracellular permeability of two midgut and two hindgut segments from each experimental tank was assessed by measurement of the apparent permeability (Papp) of the hydrophilic marker molecule 14C mannitol in Ussing chamber systems. The intestinal segments were mounted in Ussing chambers as described by Sundell et al. (2003). The chambers were filled with salmon Ringer solution (Sundell et al., 2003) and the temperature was kept at 10 8C by a cooling mantle. Mixing and oxygenation was achieved by gas-lift using gas-mixture of 0.3% CO2 in air. The exposed tissue surface area was 0.75 cm2 and the half-chamber volume 5 ml. The chambers were equipped with one pair of Pt-electrodes for current passage and one pair of Ag/AgCl-electrodes (Radiometer, Copenhagen) for measurement of transepithelial potential differences (TEP) and further calculation of transepithelial resistance (TER) and short circuit current (SCC) for continuous monitoring of preparation viability throughout the experiment. All preparations were allowed 60 min of equilibration before the experiments was initiated by exchanging the Ringer solution on both sides of the epithelium to fresh Ringer solution, with the addition of 14C-

R.E. Olsen et al. / Aquaculture 250 (2005) 480–495

Mannitol (MW:184; spec. act. 0.02–0.03 MBq  ml 1, Amersham, St. Louis, USA) to the mucosal side. Samples of 50 Al were withdrawn from the serosal compartment every 10 min for 90 min and replaced by the same volume of fresh Ringer. Five milliliter of scintillation cocktail (Opti fluor, Packard) was added to each sample and the radioactivity assessed in a liquid scintillation counter and reported as disintegrations per minute (dpm). Upon analysis, mucosa scrapings were homogenized in three volumes of ice-cold saline using an Ultra Turrax homogeniser. Total lipids were extracted from the homogenate according to the method of Folch et al. (1957). Lipid class composition was assessed using the HPTLC double development method of Olsen and Henderson (1989). Total polar lids were isolated using the same method, except that only the neutral lipids were developed using the neutral solvent system and conditions of Olsen and Henderson (1989). The origin containing total polar lipids was scraped off and subjected to sulphuric acid catalysed transesterification (Christie, 1982), extracted into hexane and stored at 80 8C until analysed. Quantitative analysis of fatty acid methyl esters were carried out by gas liquid chromatography using a HP 5890 gas chromatograph equipped with a J and N Scientific Inc DB-23 fused silica column (30 m * 0.25 mm id) as described by Olsen et al. (2004). Protein content of the homogenate was assessed using the method of Lowry et al. (1951). Plasma and enterocyte homogenate malondialdehyde (MDA) was quantified using the HPLC thiobarbituric acid adduct method of Wong et al. (1987). Cortisol was analysed in unextracted plasma using a radio immunoassay procedure according to Young (1986) and validated by Bisbal and Specker (1991). Cortisol antibodies were obtained from Endocrine Sciences, CA (lot 345102280). Plasma glucose was assessed using the Trinder reagent procedure (Sigma No. 315), glutamic oxaloacetic transaminase (GOT) and glutamic pyruvic transaminase (GPT) by the combined colorimetric assay of Sigma (No. 505) and plasma lactate by Sigma No. 735. Microbial examination of adherent bacteria in hindgut and faeces of five fed fish was determined prior to and 4 h post-stress as described by Ringø (1993). Food-deprived fish were not analysed. Identification of adherent bacteria colonising the hindgut

483

and bacteria isolated from faeces was performed according to the criteria given by Ringø and Olsen (1999). Pantoea spp. and Rhodococcus spp. were identified according to Buller (2004). Numerical results are expressed as group means and standard errors of means (SEM). Data were subjected to Levene’s test for homogeneity of variances and in cases of data sets with unequal variance, a log-transformation was performed resulting in equal variances. Lipid and fatty acid compositions in intestinal segments were compared using general linear models, while differences in other parameters were compared by means of ANOVA, and, if significant differences were observed, Tukeys multiple range test. Null hypotheses were rejected at a probability level of P b 0.05.

3. Results 3.1. Biochemical markers Blood chemistry and intestinal permeability data are given in Fig. 1. In food-deprived fish haematocrit (Hct; Fig. 1A) increased from 24% before stress to a maximum of approximately 41% 4 h post-stress, after which it decreased to 30% at 48 h. The same pattern was seen for fed fish, except for a higher pre-stress Hct of 33.5%. Plasma glucose (Fig. 1B) increased following stress in both groups. The response seemed to be more pronounced in fed fish, reaching 14.7 mg dl 1 by 4 h compared to 9.8 mg dl 1 for fed fish. All values returned to initial levels by 48 h. Fig. 1 C shows that feeding status had no effect on plasma GOT, being around 50 UI l 1 for both fed and food-deprived fish prior to stress. Subjecting rainbow trout to stress appeared to increase the level in both groups, being more pronounced for food-deprived fish. At 48 h, GOT appeared to the approach initial values in food-deprived fish, whereas the plasma content continued to increase in fed fish reaching 85 IU l 1 48 h post-stress. Exposing fasted fish to acute stress led to a tendency of increased GPT levels 4 h post-stress, but the values returned to initial levels after 48 h. In contrast, in fed fish the high initial level of GPT was maintained throughout the experiment. Plasma lactate (Fig. 1E) increased significantly in both fed and food-deprived fish from an initial value

484

R.E. Olsen et al. / Aquaculture 250 (2005) 480–495

18

50

c

A

c

Haematocrit (%)

40

a

b

d d

30

b 20

a

16

Plasma glucose (mg dl-1)

Fed Deprived

a

10

Fed Deprived

B

c

14 12

b

10 8 6

a

a a

a

4 2

0

0 Start

4 hours

48 hours

Start

Time after stress

4 hours

48 hours

Time after stress

120

C

Fed Deprived

bc

6

c

80

ab 60

D

Fed Deprived

a

Plasma GPT (UI l-1)

Plasma GOT (IU l-1)

100

ab

a a

40

a ab

4

ab

bc c 2

20 0

0 Start

4 hours

48 hours

Start

Time after stress

c

20

b 15 10 5

48 hours

700

E

Fed Deprived

Plasma cortisol (ng ml-1)

Plasma lactate (mg dl-1)

30 25

4 hours

Time after stress

a a

a a

0

600

Fed Deprived

F

c

500 400

b 300 200

a a

100

a a 0

Start

4 hours

48 hours

Time after stress

Start

4 hours

48 hours

Time after stress

Fig. 1. Blood chemistry data of food-deprived (deprived) and fed (fed) rainbow trout before, 4 and 48 h post handling stress. (A) blood haematocrit (Hct), (B) plasma glucose, (C) plasma GOT, (D) plasma GPT, (E) plasma lactate, (F) plasma cortisol, (G) plasma malondialdehyde, (H) midgut malondialdehyde, (I) hindgut malondialdehyde, (J) midgut apparent permeability for mannitol, (K) hindgut apparent permeability for mannitol, (L) apparent permeability for mannitol in midgut and hindgut for food-deprived and fed rainbow trout without stress exposure. For plasma samples; n = 12; for intestinal region samples n = 8. Values are presented as meansFSEM.

R.E. Olsen et al. / Aquaculture 250 (2005) 480–495

485

Plasma MDA (nmol ml-1)

14 12

G

Fed Deprived

a a ab ab

b a

10 8 6 4 2

Midgut MDA (nmol mg protein-1)

2,5 16

0

2,0

a

1,5

4 hours

a

1,0

0,5

48 hours

Start

4 hours

48 hours

Time after stress

2,5

6

I

Fed Deprived

2,0

1,5

a a 1,0

a

a

a

a

0,5

Papp3H-mannitol*106(cm*sec-1)

Hindgut MDA (nmol mg protein-1)

a

a

Time after stress

0,0

5

J

Midgut fed Midgut deprived

c

4

b

3 2

a

1

a

a

a

0 Start

4 hours

48 hours

Start

Time after stress

K

4 3 2

a

a

a a

48 hours

6

b Hindgut fed Hindgut deprived

1

4 hours

Time after stress

a

0

Papp3H-mannitol*106(cm*sec-1)

6

Papp3H-mannitol*106(cm*sec-1)

a

a

0,0 Start

5

H

Fed Deprived

5

L

Midgut fed Midgut deprived Hindgut fed Hindgut deprived

4 3 2 1 0

Start

4 hours

48 hours

0

Time after stress

72

120

Days of study Fig. 1 (continued).

of 2.5–16 mg dl 1 in fed fish and 23.2 mg dl 1 in food-deprived fish 4 h post-stress. Regardless of dietary regime, all values returned to initial values

after 48 h. Plasma cortisol (Fig. 1F) increased following stress reaching a maximum of 295 ng ml 1 after 4 h in fed fish. The stress response

– 0.587 0.012 0.234 0.187 0.168 0.252 0.520 0.587 – *** * *** – *** * *** *** *p b 0.05, **p b 0.001, ***p b 0.0001, no: homogeneity of variances not obtained.

– – – – * – – – – Fed Food Fed Food Food

Fed

Food

Fed

Food

Fed Fed Deprived

C 22.6 F 0.8 21.6 F 0.7 21.5 F 1.3 23.5 F 0.8 22.1 F 0.7 22.4 F 0.8 23.1 F1.2 22.3 F 0.9 20.4 F 0.9 22.1 F 0.4 21.3 F 1.2 20.9 F 0.8 – CB 1.3 F 0.1 1.3 F 0.2 1.6 F 0.2 1.5 F 0.3 1.6 F 0.3 1.9 F 0.2 0.9 F 0.1 1.1 F 0.2 1.1 F 0.1 1.0 F 0.1 1.5 F 0.2 1.0 F 0.1 – SP 0.5 F 0.1 0.5 F 0.1 0.4 F 0.0 0.5 F 0.1 0.5 F 0.1 0.5 F 0.1 0.3 F 0.1 0.4 F 0.0 0.4 F 0.1 0.4 F 0.0 0.5 F 0.0 0.3 F 0.1 – PE 22.5 F 0.8 22.9 F 1.0 24.4 F 1.4 20.3 F 0.9 22.8 F 1.1 22.2 F 1.1 26.7 F 1.3 25.0 F 0.7 25.6 F 0.8 25.0 F 0.8 24.5 F 1.6 27.0 F 0.8 – CL 3.4 F 0.1 3.4 F 0.2 3.1 F 0.2 2.4 F 0.2 3.5 F 0.2 3.0 F 0.1 3.1 F 0.1 3.1 F 0.2 2.9 F 0.1 3.0 F 0.1 2.9 F 0.2 3.2 F 0.1 no PI 9.1 F 0.3 9.9 F 0.3 9.4 F 0.4 8.7 F 0.6 9.7 F 0.4 9.1 F 0.6 7.8 F 0.2 7.9 F 0.3 8.6 F 0.3 8.0 F 0.3 8.8 F 0.4 8.3 F 0.3 – PS 7.9 F 0.3 6.9 F 0.3 7.0 F 0.3 6.2 F 0.4 7.5 F 0.3 7.1 F 0.3 7.9 F 0.2 8.3 F 0.3 8.1 F 0.2 8.3 F 0.3 7.7 F 0.6 8.1 F 0.1 no PC 29.8 F 1.3 30.4 F 1.5 29.3 F 2.1 33.4 F 1.0 28.8 F 1.1 30.7 F 1.4 22.2 F 1.2 23.1 F 0.8 24.6 F 0.6 22.4 F 0.8 25.4 F 1.6 23.4 F 0.5 – LPC/ 2.7 F 0.3 3.1 F 0.4 3.1 F 0.3 3.3 F 0.3 3.3 F 0.4 3.0 F 0.3 5.4 F 0.4 5.4 F 0.4 5.6 F 0.5 6.7 F 0.5 4.9 F 0.6 5.3 F 0.3 – SM

4

4

48

48

0 0 0

4

4

48

48

Deprived Hours Mid-Hind-gut r 2 vs. Fed post-stress 0

Membrane lipid class composition of midgut and hindgut mucosa of rainbow trout subjected to stress is given in Table 1. Feeding status or stress had no effect on lipid composition. However, a regional difference was observed between the intestinal segments. In particular, midgut had higher levels of phosphatidylcholine (PC) and phosphatidylinositol (PI) compared to hindgut, while the opposite was true for phosphatidylethanolamine (PE), phosphatidylserine (PS) and lysophosphatidylcholine/sphingomyelin (LPC/SM). Likewise, stress had no influence on the membrane lipid fatty acid composition (total polar lipid) of the intestinal mucosal segments (Table 2). Feeding status did have an effect on the overall fatty acid composition. The content of saturated fatty acids tended to increase following feeding, concomitant with a reduction in n 6 polyunsaturated fatty acids (PUFA) (mostly 20:4 n 6) and thereby total PUFA. It is also interesting to note that there was an interchange

Hindgut

3.3. Lipid composition

Hours after stress

The paracellular permeability of the intestine was influenced by both stress and feeding status. The permeability of the midgut increased 4 h post-stress in food-deprived but not in fed fish (Fig. 1J). This increase was further enhanced 48 h post-stress. In the hindgut, no effect on paracellular permeability was evident 4 h post-stress in either food-deprived or fed fish, whereas increased paracellular permeability was apparent in food-deprived fish 48 h post-stress, at this time point the fish had been deprived of food for 72 h (Fig. 1K). No effect on paracellular permeability was demonstrated by food deprivation alone (Fig. 1L).

Hours after stress

3.2. Intestinal permeability

Midgut

appeared to be more pronounced in food-deprived fish reaching plasma levels of 484 ng ml y1 after 4 h. At 48 h post-stress, plasma cortisol had returned to initial values in both groups. Plasma malondialdehyde (MDA) levels (Fig. 1G) were not significantly influenced by stress. A tendency towards increased levels was observed after 4 h in both experimental groups but these were not significantly different. Likewise, midgut (Fig. 1H) and hindgut (Fig. 1I) malondialdehyde content was not significantly influenced by stress.

Significance ( p)

R.E. Olsen et al. / Aquaculture 250 (2005) 480–495

Table 1 Lipid class composition of middle intestine and distal intestine in rainbow trout in food-deprived (deprived) fish (3 days) before, and 4 and 48 h after subjecting them to acute stress

486

Table 2 Total polar lipid fatty acid composition of middle intestine and distal intestine enterocytes in rainbow trout in starved fish (3 days) and fed fish before, 4 and 48 h after subjecting them to acute stress Hindgut

Significance ( p)

Hours after stress

Hours after stress

Deprived Hours Midr2 vs. Fed post-stress Hind-gut

0

0

4

4

48

48

0

0

4

4

48

48

Deprived

Fed

Deprived

Fed

Deprived

Fed

Deprived

Fed

Deprived

Fed

Deprived

Fed

0.8 F 0.1 19.1 F 0.4 0.9 F 0.1 7.5 F 0.2 4.7 F 0.2 3.0 F 0.1 0.7 F 0.1 0.5 F 0.1 1.9 F 0.1 2.5 F 0.1 6.4 F 0.5 1.6 F 0.1 44.5 F 0.7 28.4 F 0.4 13.1 F 0.5 58.5 F 0.3 53.5 F 0.4 5.0 F 0.2 10.9 F 0.5

0.7 F 0.1 19.2 F 0.9 1.3 F 0.1 6.9 F 0.1 4.3 F 0.2 3.0 F 0.1 0.8 F 0.1 0.5 F 0.1 2.6 F 0.1 2.1 F 0.1 8.9 F 0.7 1.5 F 0.1 41.6 F 1.0 28.1 F1.1 13.7 F 0.4 58.2 F 1.1 53.3 F 1.1 4.9 F 0.1 10.8 F 0.3

0.8 F 0.1 19.4 F 0.7 0.9 F 0.1 7.8 F 0.3 4.5 F 0.2 3.0 F 0.1 0.6 F 0.1 0.7 F 0.1 2.1 F 0.1 2.6 F 0.1 6.0 F 0.5 1.6 F 0.1 44.3 F 0.5 29.1 F 0.7 12.7 F 0.3 58.2 F 0.8 53.3 F 0.8 4.9 F 0.1 11.0 F 0.3

0.8 F 0.1 20.3 F 0.7 1.3 F 0.1 7.2 F 0.2 4.3 F 0.2 3.0 F 0.1 0.7 F 0.1 0.5 F 0.1 2.5 F 0.1 2.1 F 0.1 9.1 F 0.3 1.7 F 0.0 39.7 F 1.1 29.6 F 0.7 13.9 F 0.5 56.5 F 1.1 51.9 F 1.0 4.6 F 0.2 11.5 F 0.5

0.7 F 0.1 16.7 F 0.5 0.7 F 0.1 8.6 F 0.5 5.2 F 0.4 3.1 F 0.1 0.7 F 0.1 0.4 F 0.1 1.8 F 0.1 2.9 F 0.1 5.0 F 0.4 1.8 F 0.2 46.5 F 0.9 27.0 F 0.5 13.2 F 0.4 59.8 F 0.7 54.4 F 0.6 5.4 F 0.1 10.0 F 0.2

0.8 F 0.1 19.6 F 0.6 1.1 F 0.1 7.2 F 0.2 3.9 F 0.1 2.7 F 0.1 0.7 F 0.1 0.5 F 0.1 2.4 F 0.1 2.1 F 0.1 8.9 F 0.4 1.6 F 0.0 42.4 F 1.0 28.8 F 0.7 12.4 F 0.2 58.8 F 0.8 54.0 F 0.7 4.8 F 0.2 11.5 F 0.6

0.9 F 0.1 16.8 F 0.6 0.7 F 0.0 10.8 F 0.4 6.5 F 0.3 3.4 F 0.1 1.0 F 0.0 0.6 F 0.1 1.5 F 0.0 2.8 F 0.1 6.6 F 0.2 2.4 F 0.2 40.7 F 0.7 29.3 F 0.7 14.3 F 0.3 56.5 F 0.9 50.9 F 0.6 5.6 F 0.2 9.2 F 0.4

1.1 F 0.1 17.8 F 0.5 0.8 F 0.1 10.4 F 0.4 6.0 F 0.1 3.6 F 0.1 1.0 F 0.1 0.5 F 0.1 1.6 F 0.1 2.7 F 0.1 6.5 F 0.2 2.3 F 0.2 39.5 F 0.6 30.2 F 0.5 14.6 F 0.3 55.2 F 0.5 49.5 F 0.6 5.7 F 0.2 8.8 F 0.4

0.7 F 0.1 16.0 F 0.5 0.7 F 0.0 10.8 F 0.4 6.8 F 0.2 3.5 F 0.1 0.9 F 0.0 0.7 F 0.1 1.4 F 0.0 3.0 F 0.1 6.7 F 0.1 2.6 F 0.1 40.2 F 0.6 28.4 F 0.6 14.9 F 0.5 56.8 F 0.6 50.9 F 0.5 5.9 F 0.1 8.6 F 0.2

0.9 F 0.1 15.8 F 0.6 0.7 F 0.1 9.9 F 0.2 6.4 F 0.2 3.8 F 0.1 1.0 F 0.1 0.3 F 0.1 1.5 F 0.1 2.8 F 0.1 7.0 F 0.2 2.7 F 0.2 41.2 F 1.0 27.4 F 0.8 15.0 F 0.5 57.6 F 1.1 51.9 F 1.1 5.7 F 0.1 9.1 F 0.3

0.9 F 0.2 15.4 F 0.7 0.7 F 0.1 10.3 F 0.4 7.0 F 0.3 3.3 F 0.1 0.8 F 0.1 0.8 F 0.1 1.5 F 0.1 2.9 F 0.1 6.1 F 0.2 2.4 F 0.2 41.5 F 1.1 27.4 F 0.8 15.4 F 0.8 57.2 F 1.2 51.7 F 1.2 5.5 F 0.2 9.4 F 0.3

1.3 F 0.1 17.9 F 0.2 1.0 F 0.1 10.1 F 0.3 6.1 F 0.2 3.7 F 0.2 0.9 F 0.1 0.4 F 0.1 1.7 F 0.1 2.4 F 0.1 7.1 F 0.4 2.4 F 0.2 38.3 F 0.8 30.2 F 0.3 15.4 F 0.6 54.4 F 0.8 49.0 F 0.8 5.4 F 0.2 9.2 F 0.4

* ** *** *** *** – – * *** *** *** – *** * – * – * –

– – – – – – – * – – – – – – – – – – –

* *** *** *** *** *** *** – *** *** *** *** *** – *** *** *** *** ***

0.149 0.468 0.498 0.716 0.720 0.543 0.379 0.149 0.745 0.590 0.610 0.501 0.454 0.181 0.337 0.218 0.298 0.406 0.465

R.E. Olsen et al. / Aquaculture 250 (2005) 480–495

14:0 16:0 16:1 n 9 18:0 18:1 n 9 18:1 n 7 18:2 n 6 18:4 n 3 20:1 n 9 20:4 n 6 20:5 n 3 22:5 n 3 22:6 n 3 SAT MONO PUFA N 3 N 6 N 3/N 6

Midgut

*p b 0.05, **p b 0.001, ***p b 0.0001.

487

488

R.E. Olsen et al. / Aquaculture 250 (2005) 480–495

between an increase in 16:0 following feeding combined with a reduction in 18:0. The same applies to 20:5 n 3 that increased and 22:6 n 3 that decreased following feeding causing a stable total n 3 PUFA content. Overall, the most pronounced difference was that the hindgut contained more monounsaturated fatty acids and n 6 PUFA than the midgut whereas the reverse was true for n 3 PUFA, total PUFA and n 3 / n 6 ratio. 3.4. Electron microscopy Midgut enterocytes of food-deprived fish had normal appearances displaying numerous microvilli at the apical borders. Epithelial integrity appeared

normal with the presence of tight intercellular junctional complexes (tight junctions and desmosomes, Fig. 2A). Some ruptures of the tight junctions could be observed, but they were mainly observed close to goblet cells that seemed to traverse through to the epithelial cell layer to the luminal surface. In fed fish, several midgut enterocytes appeared to be in an absorptive phase as indicated by numerous small lipid vacuoles and droplets in the cytoplasm with many being surrounded by endoplasmic reticulum. Although a few disrupted junctional complexes could be observed, the most notable finding was the tendency to widening of the intercellular space 1–2 Am below the junctional complexes, whereas the complexes as such appeared intact (Fig. 2B).

Fig. 2. (A) Midgut enterocytes of food-deprived rainbow trout before stress. G: Goblet cell, mw: microvilli, arrows: intact junctional complexes. (B) Midgut enterocytes of fed rainbow trout before stress. G: Goblet cells, L: lipid vesicles and droplets most surrounded by membrane, others are not, arrows: intact junctional complexes, arrowheads; widening of the intercellular space below the junctional complex. (C) Hindgut of fooddeprived rainbow trout before stress. Arrows: pinocytotic vesicles, arrowheads: intracytoplasmic vacuoles. (D) Hindgut of fed rainbow trout before stress. Arrows: widening in intercellular space, arrowheads: intracytoplasmic vacuoles. X = 3750.

R.E. Olsen et al. / Aquaculture 250 (2005) 480–495

Hindgut enterocytes also appeared normal, but differed from those in the midgut by displaying extensive pinocytotic activity at their apical borders and containing large numbers of intracytoplasmic lysosome-like vacuoles. The only major difference between food-deprived (Fig. 2C) and fed fish (Fig. 2D) was a widening of the intercellular space in fed fish, which appeared to be caused by nutrient transport into the lamina propria. Subjecting the fish to stress increased the incidence of disrupted junctional complexes in the midgut. Damage was due mostly to dissociations of desmosomes and nexuses that in some instances led to a total loss of cell-to-cell contacts. In food-deprived fish, the most notable finding was an increased number of widening of the intercellular space 1–2 Am below the junctional complexes while damages to the complexes themselves were relatively rare (Fig. 3A). In fed fish, there was increased prevalence of rupture of junctional complexes as well as widening of the intercellular space below the junctional complexes (Fig. 3B). No major changes were observed in hindgut of fish 4 h after stress. Of the images studied, only one opening of the junctional complex was observed in a fed fish (data not shown). Forty-eight-hour post-stress most junctional complexes of midgut appeared to have resumed their normal form in food-deprived fish. Although occasional widening in the intercellular space 1–2 Am below the junctional complexes was observed, this did not appear to be more prevalent than in prestressed fish (Fig. 4A). In fed fish, numerous lipid droplets could still be observed after 2 days of recovery despite the fact that these fish had not been fed since onset of the stress. Furthermore, occasional damages to the junctional complexes could still be seen although not as frequently as after 4 h. There was also a tendency of fed fish to have more widening of the intercellular space below the junctional complexes as observed in the fish before subjecting them to stress (Fig. 4B). No further effect was observed in hindgut of both food-deprived and fed trout 48 h after stress. However, the tendency to have widening in the intercellular space towards the lamina propria in fed fish as observed before stress was not found, indicating that the nutrients had been removed by this time (data not shown).

489

Fig. 3. (A) Midgut enterocytes of food-deprived rainbow trout 4 h after stress. G: Goblet cell, mw: microvilli, arrow: widening of the intercellular space below the junctional complex, arrowhead: intact junctional complexes. (B) Midgut enterocytes of fed rainbow trout 4 h after stress. L: lipid vesicles and droplets most surrounded by membrane, others are not, arrows: ruptured junctional complexes. X = 3750.

3.5. Microbiology Five individual fish were used to determine the bacterial population level in the hindgut and faeces. Plates from hindgut and faeces of one fish prior to stress and plates from faeces of one fish post-stress were contaminated with fungi. The current study clearly demonstrates that acute stress affected the population level of culturable aerobic bacteria colonising the hindgut and faeces of fed rainbow trout (Table 3). The population level of adherent bacteria in the hindgut decreased from 2  104 (log 4.30) prior to

490

R.E. Olsen et al. / Aquaculture 250 (2005) 480–495

nas, Rhodococcus, Staphylococcus and Microbacterium were isolated from faeces prior to stress, with Alcaligenes (1.5  104) as the dominant bacterial genus. Unidentified pink coloured colonies of Grampositive cocci were also isolated in faeces before stress, but the isolates died prior to positive identification. Four hours post-stress, the genera Acinetobacter, Alcaligenes, Arthrobacter, Pseudomonas, Rhodococcus, Staphylococcus, Microbacterium and Micrococcus were identified in faeces. The dominant genus was Pseudomonas (1.4  105) followed by Staphylococcus (9  104). Moreover, unidentified small transparent colonies of Gram-negative rods Table 3 Log total viable counts (log TVC) g 1 wet mass, total number of isolates and changes in log TVC of bacterial species isolated from the digestive tract of five individual rainbow trout prior to and 4 h post-stress Prior to stress

Fig. 4. (A) Midgut of food-deprived rainbow trout 48 h after stress. mw: microvilli, arrows: intact junctional complexes. (B) Midgut of fed rainbow trout 48 h after stress. mw: microvilli, L: lipid vesicles and droplets most surrounded by membrane, others are not, arrows: widening of the intercellular space below the junctional complex, arrowheads: damage to the junctional complexes. X = 3750.

stress to 5  102 (log 2.70) 4 h post-stress. In contrast, in faeces the aerobic bacterial population level increased from 5  104 (log 4.70) before stress to 4  105 (log 5.60) 4 h after stress. A total of 313 strains of bacteria was isolated from hindgut and faeces of rainbow trout before (114) and after stress (199). Prior to stress, the following adherent bacterial genera; Acinetobacter, Pseudomonas, Rhodococcus and Staphylococcus were isolated from the hindgut, where Staphylococcus (1  104) was the dominant genus. However, in hindgut 4 h poststress only Pseudomonas, Microbacterium and Staphylococcus were isolated. The dominant genus was at this time Pseudomonas (4.1  102). Strains belonging to Acinetobacter, Alcaligenes, Pantoea, Pseudomo-

a

Hindgut No. of isolates Acinetobacter spp. Pseudomonas spp. Microbacterium spp. Rhodococcus spp. Staphylococcus spp. Unidentified*

4.30 75 2.72 3.79 n.d. 3.03 4.04 3.12

Faeces No. of isolates Acinetobacter spp. Arthrobacter spp. Alcaligenes spp. Pantoea spp. Pseudomonas spp. Rhodococcus spp. Staphylococcus spp. Microbacterium spp. Micrococcus spp. Unknown Gram-positive cocci** Unknown Gram-negative rods*** Unidentified*

4.70a 39 3.80 (3) n.d. 4.18 (4) 3.41 (3) 3.71 (4) 3.11 (1) 3.89 (4) 3.41 (2) n.d. 3.89 (4) n.d. 3.11 (1)

(1) (4) (1) (4) (4)

Post-stress 2.70 40 n.d. 2.62 1.10 n.d. 1.10 1.80

(4) (1) (1) (3)

5.60a 159 4.74 (3) 4.18 (2) 3.88 (1) n.d. 5.13 (4) 4.30 (4) 4.96 (4) 4.18 (4) 4.10 (3) n.d. 4.35 (4) 4.40 (4)

a Mean values from four individual fish as plates from one fish were contaminated with fungi. n.d.—not detected. Unidentified*—died prior to positive identification. Unknown Gram-positive cocci**—pink coloured colonies, died prior to positive identification. Unknown Gram-negative rods***—smaller transparent colonies, died prior to positive identification. The number of fish from which bacteria were isolated is given in brackets.

R.E. Olsen et al. / Aquaculture 250 (2005) 480–495

were isolated in faeces after stress, but these isolates died prior to positive identification.

4. Discussion Subjecting fish to stress elicits a series of welldefined stress responses. Immediate release of the catecholamines, adrenaline and noradrenaline, from the chromaffin cells as well as cortisol from the interrenal gland occurs after acute severe stress (Wendelaar Bonga, 1997). One of the main functions of these stress hormones is to release carbohydrate stored in liver as glycogen causing plasma hyperglycaemia (Barton and Iwama, 1991). As acute stress may generate an anaerobic environment, lactate will be generated in the cells by glycolysis and released into the blood. Furthermore, the haematocrit tends to increase in fish subjected to stress due to a combination of erythrocyte swelling and spleen contractions, which will release erythrocytes into the blood stream (Wendelaar Bonga, 1997). All these events were demonstrated in the present experiment 4 h poststress, showing that the chasing procedure evoked significant stress in the fish. Furthermore, all values returned to basal levels after 48 h indicating that in this respect the fish had fully recovered by this time. Feeding status did have a significant effect on the magnitude of the response 4 h after stress. Plasma levels of cortisol and lactate were higher in fooddeprived fish compared to fed fish, while the opposite was true for glucose. Higher glucose levels in fed compared to food-deprived fish has also been demonstrated in fed Atlantic salmon (S. salar L.) for up to 12 h after stress (Olsen et al., 2003). The reason for this discrepancy is probably the relatively larger liver glycogen store in fish under active feeding metabolism, helped by continuous supply of glucose from the intestine as long as diet is available. The higher levels of cortisol and lactate in food-deprived fish 4 h after stress indicates a stronger stress response when food deprived. For cortisol this is in agreement with data on Atlantic salmon where a more pronounced cortisol response was observed when fooddeprived fish were subjected to acute stress (Olsen et al., 2003). The higher lactate content of food-deprived compared to fed fish, at 4 h post-stress, is however not in agreement with the observations on Atlantic

491

salmon. In salmon the lactate content showed the highest values 4 h post-stress in both groups but with a lower level in the food-deprived group. In a comparable study, the peak in plasma lactate levels on food-deprived fish appeared 4 h post-stress for rainbow trout and 1 h post-stress for Atlantic salmon, whereas the corresponding values for fed fish were 1– 2 h post-stress in both species (Olsen and Ringø, unpublished data). Thus, it appears as the sampling at 4 h registered the peak in plasma lactate levels for food-deprived fish while the peak levels of fed fish had already passed and the levels were returning towards basal levels. Although significant differences were observed between fed and food-deprived fish, the effects were of relatively small magnitude compared to the actual response to the stress and can thus be viewed as relatively marginal. In juvenile chinook salmon (Oncorhynchus tswawytscha) (Barton et al., 1988) and Arctic char (S. alpinus) (Jørgensen et al., 2002), there was a tendency towards a lower cortisol response in food-deprived fish compared to fed fish. But the food-deprived fish in these experiments were starved over relatively long periods (20 days and 5 months, respectively) and had passed the point of mobilization of energy reserves in response to fasting. This is a different physiological state than in the present study where both groups of fish could be regarded as being in a bfeedingQ state with the main difference that one group had their guts filled with feed, while the others did not. Subjecting an animal to exhaustive exercise sets off a series of physiological responses that, depending on the severity and duration of stressor can lead to serious consequences including cellular damage, lowered antioxidant capacity and impaired immune status. General cellular damage in mammals is usually monitored by means of analysing leakage of cellular enzymes like the transaminases GOT and GPT into the blood (Chevion et al., 2003; Sanchez et al., 2002). The effects seen in the present study were generally minor and in many cases non-significant which is in agreement with previous studies on channel catfish (Ellsaesser and Clem, 1987) and white sucker (Catostomus commersoni) (McMaster et al., 1994). However, there were some differences between fed and food-deprived fish. Plasma GPT, generally assumed a marker of liver damage, was always higher in fed fish. This probably reflects the

492

R.E. Olsen et al. / Aquaculture 250 (2005) 480–495

higher liver metabolism in this group during feeding metabolism. The transient increase in plasma GOT levels of food-deprived fish was not reflected in the fed fish, where the increase occurred 4 h post-stress and was further elevated after 48 h. It is possible that this observation reflects a late damage or delayed tissue repair in the fed fish. In mammals, intense physical exercise is accompanied with increased oxygen consumption that will increase the production of oxygen reactive species (Alessio, 1993). This will, in turn, increase oxidative stress and the production of lipid peroxidation products and intermediates both in blood and tissues (Bagchi et al., 1998, 1999; Benderitter et al., 1996; Saitoh et al., 1995). As seen for the transaminases, no increase in malondialdehyde could be detected in plasma or intestinal segments. The only tendency appeared to be a minor increase in plasma of both groups 4 h poststress. This agrees with a previous study in Atlantic salmon (Olsen et al., 2003) and may indicate that acute stress in fish does not, to a major degree, involve free radical mechanisms as evaluated through malondialdehyde content. Reductions of the membrane PUFA content in mammals is often regarded as a part of a free radical mediated process (Sumikawa et al., 1993). The lack of effect of stress on the lipid class and fatty acid compositions of intestinal membranes of the present study further strengthens the hypothesis that free radicals are not produced in major amounts following acute stress in salmonids. Although lowering of the membrane n-3 PUFA content has been observed in gilthead seabream (Sparus aurata) juveniles, (Montero et al., 2001, 1999) the stressor used in those studies was of chronic nature (high density), which may involve free radical mechanisms. The mammalian GI tract appears to be very sensitive to stress. Several studies have shown substantial mucosal degeneration and barrier dysfunction following different forms of stress (So¨derholm and Perdue, 2001; Meddings and Swain, 2000; Prabhu et al., 2000; Santos et al., 2000; Saunders et al., 1994; Sengupta and Sharma, 1993). Although less studied, stress also seems to influence the GI tract of fish. In European eel, Anguilla anguilla L. social stress has been associated with cellular degeneration and loss of cell–cell contacts between gastric cells (Peters, 1982) and loss of intestinal goblet cells and detachment of enterocytes in carp (Cyprinus carpio

haematopterus) subjected to catching and transportation stress (Szakolczai, 1997). Recently, Olsen et al. (2003) showed that Atlantic salmon subjected to exhaustive stress had damage to the intercellular junctional complexes of the midgut while little effects were seen in hindgut segments. Furthermore, fed fish appeared to be more susceptible to stress induced damage than food-deprived fish. However, regardless of dietary regime, cellular integrity returned to prestress conditions within 24 h without leaving much permanent damage. The present study is in general agreement with these observations. However, fed fish were observed to have more events of ruptures to junctional complexes and widening of the intercellular space 1–2 Am below these complexes, than fooddeprived fish both before and after stress (0, 4 and 48 h post-stress). This could indicate a species-dependent difference in timing of the response and that salmonids are less prone to acute stress related gut damage compared to mammalian species. One consequence of stress in mammals is the loss of intestinal integrity causing enhanced epithelial permeability that may lead to enhanced uptake of macromolecules, bacterial products and antigens across the epithelium (Saunders et al., 1994; Tamayo et al., 1996; Kiliaan et al., 1998; So¨derholm and Perdue, 2001). These processes are suggested to be mediated via neuronal and humoral factors as well as through mucosal mast cell activation (Santos et al., 2000; So¨derholm and Perdue, 2001; Yates et al., 2001). A focus on the role of corticotropin releasing hormone (CRH), as being the initiation of the hypothalamic–pituitary–interrenal axis, have revealed importance of CRH for the stress induced epithelial damage but probably not mediated via adrenal steroid actions (Santos et al., 1999; Yates et al., 2001). However, there is direct evidence for a role of adrenal glucocorticoids in the stress-induced increase of intestinal permeability as the glucocorticoid receptor antagonist RU-486 abolishes the effect and dexamethasone is able to mimic it (Spitz et al., 1996; Meddings and Swain, 2000). In line with this, acute stress increases the paracellular permeability of rainbow trout midgut 4 h poststress an effect that is further attenuated 48 h after stress. In the hindgut an increase in the paracellular permeability is apparent 48 h after stress. This assessment of paracellular permeability, showed a

R.E. Olsen et al. / Aquaculture 250 (2005) 480–495

more prolonged effect of stress than the ultrastructural examination where most of the visual damage to the junctional complexes was restored within 48 h poststress. In rats, the maximal effect of restraint stress on intestinal barrier functions is suggested to occur 4 h post-stress (cf. Santos et al., 2001, 1999) and most studies are concluded within that time frame. However, the effects on intestinal epithelial conductance were sustained for 24 but not 72 h post-stress and macromolecule transport was elevated for more than 72 h post-stress (Santos et al., 2001, 1999). The even longer pronounced effect of stress on intestinal paracellular permeability of rainbow trout could be an effect of temperature. The in vivo acclimation temperature and the in vitro experimental temperature during which the permeability is measured were 10 8C, probably resulting in slower physiological and biochemical events then in the mammalian system at 37 8C. The apparent discrepancy between the paracellular permeability and the visual damages to junctional complexes as judged by ultrastructural examination, is well in line with previous studies on mammals, where several stress models induce increased paracellular permeability without causing visible changes in gut morphology (Kiliaan et al., 1998; Santos et al., 1999; Meddings and Swain, 2000). The response to stress on paracellular permeability is only apparent in food-deprived and not fed fish. Further, food-deprived fish showed a more pronounced cortisol response to stress than fed fish. In mammals, glucocorticoids are suggested to have a role in the impairment of paracellular permeability as indicated by the capacity of RU486 to abolish the stress induced effect on intestinal permeability (Meddings and Swain, 2000). Thus, the more pronounced increase in plasma cortisol levels after stress in fooddeprived fish may be the reason for the similarly more pronounced effect on paracellular permeability in food-deprived fish. Dexamethasone treatment, as a mimic of stress related glucocorticoid effects, impaired the intestinal permeability only in starved and not in fed rats (Spitz et al., 1996). Thus, for both rainbow trout and rat, the presence of feed in the GI tract seems to be protective during glucocorticoid mediated stress responses. The population level of aerobic bacteria isolated from hindgut and faeces of rainbow trout was affected by acute stress. The population level of adherent

493

microbiota decreased in the hindgut with a concomitant increase in faeces. This is in accordance with observations from Atlantic salmon (S. salar L) and is most likely due to a peel-off effect of mucus (Olsen et al., 2003). The elimination of the adherent bacteria Acinetobacter ssp. and Rhodococcus spp., and the reduction of population level of Pseudomonas spp. and Staphylococcus spp. in the hindgut post-stress with concomitant increase in faeces post-stress might have some relevance in the protection of other autochthonous or allochthonous bacteria to colonise the intestinal mucosa. The increase in population level of Arthrobacter, Microbacterium and Micrococcus in faeces 4 h after stress and their presence below detection level in the hindgut before stress might indicate that these bacteria genera either colonise the midgut or they have weaker attachment to the digestive tract mucosa and are lost during preparation as the intestine rinsed three times in sterile 0.9% saline to remove nonadherent bacteria. The notion that acute stress influences the gut microbiota is supported by the fact that the gastrointestinal microbiota is sensitive to stress-like conditions such as dominance hierarchy as observed by Ringø et al. (1997). In the present study, we have demonstrated that acute stress to rainbow trout has only a marginal effect on indices of tissue damage and oxidative stress parameters, both with regard to blood chemistry and the ultrastructure of intestinal segments. Furthermore, stress does not seem to influence the lipid or fatty acid composition of the intestine as would be expected if subjected to high levels of oxidative stress. Ultrastructural damages to the enterocytes are evident only in midgut following stress, and they appear, to a large extent, transient returning to normal values within 2 days. Moreover, feeding state seems to have some impact on the stress response. Although fish with their guts filled with diet are more susceptible to ultrastructural damage in midgut, they seem less prone to stress than their food-deprived counterparts with regard to plasma cortisol, lactate and paracellular permeability. Furthermore, there seems to be a prolonged stress response in the GI tract of fooddeprived fish that causes substantial impairment of the barrier function lasting for at least 2 days following stress. Stress also causes loss of intestinal mucus,

494

R.E. Olsen et al. / Aquaculture 250 (2005) 480–495

change in intestinal microbiota content and composition. These changes may, in addition to the altered membrane permeability have major impacts on the disease resistance in fish. Acknowledgement The authors thank Henrik Sundh for excellent technical assistance. This work was supported by the Swedish Research Council for Environment, Agricultural Sciences and Spatial Planning (dr.nr: 22.3/ 2003-1241), and the Norwegian Research Council (Grant No. 122851/122). References Alessio, H.M., 1993. Exercise-induced oxidative stress. Med. Sci. Sports Exerc. 25, 218 – 224. Bagchi, D., Carryl, O.R., Tran, M.X., Krohn, R.L., Bagchi, D.J., Garg, A., Bagchi, M., Mitra, S., Stohs, S.J., 1998. Stress, diet and alcohol-induced oxidative gastrointestinal mucosal injury in rats and protection by bismuth subsalicylate. J. Appl. Toxicol. 18, 3 – 13. Bagchi, D., Carryl, O.R., Tran, M.X., Bagchi, M., Garg, A., Milnes, M.M., Williams, C.B., Balmoori, J., Bagchi, D.J., Mitra, S., Stohs, S.J., 1999. Acute and chronic stress-induced oxidative gastrointestinal mucosal injury in rats and protection by bismuth subsalicylate. Mol. Cell. Biochem. 196, 109 – 116. Barton, B.A., Iwama, G.K., 1991. Physiological changes in fish from stress in aquaculture with emphasis on the responses and effects of corticosteroids. Annu. Rev. Fish Dis. 1, 3 – 26. Barton, B.A., Schreck, C.B., Fowler, L.G., 1988. Fasting and diet content affect stress-induced changes in plasma glucose and cortisol in juvenile chinook salmon. Prog. Fish-Cult. 50, 16 – 22. Benderitter, M., Hadj-Saad, F., Lhuissier, M., Maupoil, V., Guilland, J.C., Rochette, L., 1996. Effects of exhaustive exercise and vitamin B6 deficiency on free radical oxidative process in male trained rats. Free Radic. Biol. Med. 21, 541 – 549. Bisbal, G.A., Specker, J.L., 1991. Cortisol stimulates hypo-omsoregulatory ability in Atlantic salmon Salmo salar L. J. Fish Biol. 39, 421 – 432. Buller, N.B., 2004. Bacteria from Fish and Other Aquatic Animals: A Practical Identification Manual. CABI Publishing, Oxfordshire, UK (361pp.). Chevion, S., Moran, D.S., Heled, Y., Shani, Y., Regev, G., Abbou, B., Berenshtein, E., Stadtman, E.R., Epstein, Y., 2003. Plasma antioxidant status and cell injury after severe physical exercise. Prod. Natl. Acad. Sci. U. S. A. 100, 5119 – 5123. Christie, W.W., 1982. The preparation of derivatives of lipids. In: Christie, W.W. (Ed.), Lipid Analysis. Isolation, Separation, Identification and Structural Analysis of Lipids, 2nd edition. Pergamon Press, Oxford, UK, pp. 51 – 62.

Ellsaesser, C.F., Clem, L.W., 1987. Blood serum chemistry measurements of normal and acutely stressed channel catfish. Comp. Biochem. Physiol. 88A, 589 – 594. Folch, J., Lees, M., Sloane-Stanley, G.H., 1957. A simple method for the isolation and purification of total lipids from animal tissues. J. Biol. Chem. 226, 497 – 509. Hollander, D., 1999. Intestinal permeability, leaky gut, and intestinal disorders. Curr. Gastroenterol. Rep. 1, 410 – 416. Jørgensen, E.H., Vijayan, M.M., Aluru, N., Maule, A.G., 2002. Fasting modifies Arocolor 1254 impact on plasma cortisol, glucose and lactate response to a handling disturbance in Arctic charr. Comp. Biochem. Physiol., C 132, 235 – 245. Kiliaan, A.J., Saunders, P.R., Bijlsma, P.B., Berin, M.C., Taminiau, J.A., Groot, J.A., Perdue, M.H., 1998. Stress stimulates transepithelial macromolecular uptake in rat jejunum. Am. J. Physiol.: Gasterointest. Liver Physiol. 275, G1037 – G1044. Lowry, O., Rosebrough, N., Farr, A., Randall, R., 1951. Protein measurement with the Folin-Phenol reagents. J. Biol. Chem. 193, 265 – 275. McDowell, E.M., Trump, B.R., 1976. Histological fixatives suitable for diagnostic light and electron microscopy. Arch. Pathol. Lab. Med. 100, 405 – 414. McMaster, M.E., Munkittrick, K.R., Luxon, K.R., Van der Kraak, G.J., 1994. Impact of low-level sampling stress on interpretation of physiological responses of white sucker exposed to effluent from a bleached kraft pulp mill. Ecotoxicol. Environ. Saf. 27, 251 – 264. Meddings, J.B., Swain, M.G., 2000. Environmental stress-induced gastrointestinal permeability is mediated by endogenous glucocorticoids in the rat. Gastroenterology 119, 1019 – 1028. Montero, D., Izquierdo, M.S., Tort, L., Robaina, L., Vergara, J.M., 1999. High stocking density produces crowding stress altering some physiological and biochemical parameters in gilthead seabream Sparus aurata, juveniles. Fish Physiol. Biochem. 20, 53 – 60. Montero, D., Robaina, L.E., Socorro, J., Vergara, J.M., Tort, L., Izquierdo, M.S., 2001. Alteration of liver and muscle fatty acid composition in gilthead seabream (Sparus aurata) juveniles held at high stocking density and fed an essential fatty acid deficient diet. Fish Physiol. Biochem. 24, 63 – 72. Olsen, R.E., Henderson, R.J., 1989. The rapid analysis of neutral and polar marine lipids using double-development HPTLC and scanning densitometry. J. Exp. Mar. Biol. Ecol. 129, 189 – 197. Olsen, R.E., Ringø, E., 1999. Dominance hierarchies formation in Arctic char (Salvelinus alpinus L.). Nutrient digestibility of subordinate and dominant fish. Aquac. Res. 30, 667 – 671. Olsen, R.E., Myklebust, R., Kaino, T., Ringø, E., 1999. Lipid digestibility and ultrastructural changes in the enterocytes of Arctic char (Salvelinus alpinus L.) fed linseed oil and soybean lecithin. Fish Physiol. Biochem. 21, 35 – 44. Olsen, R.E., Myklebust, R., Ringø, E., Mayhew, T.M., 2000. The influences of dietary linseed oil and saturated fatty acids on caecal enterocytes in Arctic char (Salvelinus alpinus L.): a quantitative ultrastructural study. Fish Physiol. Biochem. 22, 207 – 216. Olsen, R.E., Sundell, K., Hansen, T., Hemre, G.-I., Myklebust, R., Mayhew, T.M., Ringø, E., 2003. Acute stress alters the intestinal

R.E. Olsen et al. / Aquaculture 250 (2005) 480–495 lining of Atlantic salmon, Salmo salar L. An electron microscopical study. Fish Physiol. Biochem. 26, 211 – 221. Olsen, R.E., Dragnes, B.T., Myklebust, R., Ringø, E., 2004. Effect of soybean oil and soybean lecithin on intestinal lipid composition and lipid droplet accumulation of rainbow trout Oncorhynchus mykiss Walbaum. Fish Physiol. Biochem. 29, 181 – 192. Peters, G., 1982. The effect of stress on the stomach of European eel. Anguilla anguilla L. J. Fish Biol. 21, 497 – 512. Prabhu, R., Anup, R., Balasubramanian, K.A., 2000. Surgical stress induces phospholipid degradation in the intestinal brush border membrane. J. Surg. Res. 94, 178 – 184. Ringø, E., 1993. The effect of chromic oxide (Cr2O3) on aerobic bacterial populations associated with the intestinal epithelial mucosa of Arctic charr Salvelinus alpinus (L.). Can. J. Microbiol. 39, 1169 – 1173. Ringø, E., Olsen, R.E., 1999. The effect of diet on aerobic bacterial flora associated with the intestine of Arctic charr Salvelinus alpinus (L.). J. Appl. Microbiol. 86, 22 – 28. Ringø, E., Olsen, R.E., averli, a., Løvik, F., 1997. Effect of dominance hierarchy formation on aerobic microbiota associated with epithelial mucosa of subordinate and dominant individuals of Arctic charr Salvelinus alpinus (L.). Aquac. Res. 28, 901 – 904. Saitoh, D., Kadota, T., Okada, Y., Masuda, Y., Ohno, H., Inoue, M., 1995. Direct evidence for the occurrence of superoxide radicals in the small-intestine of the burned rat. Am. J. Emerg. Med. 13, 37 – 40. Sanchez, O., Arnau, A., Pareja, M., Poch, E., Ramirez, I., Soley, M., 2002. Acute stress-induced tissue injury in mice: differences between emotional and social stress. Cell Stress Chaperones 7, 36 – 46. Santos, J., Saunders, P.R., Hanssen, N.P.M., Yang, P.-C., Yates, D., Groot, J.A., Perdue, M.H., 1999. Corticotropin-releasing hormone mimics stress-induced colonic epithelial pathophysiology in the rat. Am. J. Physiol. 277, G391 – G399. Santos, J., Benjamin, M., Yang, P.C., Prior, T., Perdue, M.H., 2000. Chronic stress impairs rat growth and jejunal epithelial barrier function: role of mast cells. Am. J. Physiol.: Gasterointest. Liver Physiol. 278, G847 – G854. Santos, J., Yang, P.-C., So¨derholm, J.D., Benjamin, M., Perdue, M.H., 2001. Role of mast cells in chronic stress induced colonic epithelial barrier dysfunction in the rat. Gut 48, 630 – 636. Saunders, P.R., Kosecka, U., McKay, D.M., Purdue, M.H., 1994. Acute stressors stimulate ion secretion and increase epithelial

495

permeability in rat intestine. Am. J. Physiol.: Gasterointest. Liver Physiol. 30, G794 – G799. Sengupta, A., Sharma, R.K., 1993. Acute heat stress in growing rats: effect on small intestinal morphometry and in vivo absorption. J. Therm. Biol. 18, 145 – 151. So¨derholm, J.D., Perdue, M.H., 2001. Stress and gastrointestinal tract: II. Stress and intestinal barrier function. Am. J. Physiol. 280, G7 – G13. Spitz, J.C., Ghandi, S., Taveras, M., Aoys, E., Alverdy, J.C., 1996. Characteristics of the intestinal epithelial barrier during dietary manipulation and glucocorticoid stress. Crit. Care Med. 24, 635 – 641. Sumikawa, K., Mu, Z., Inoue, T., Okochi, T., Yoshida, T., Adachi, K., 1993. Changes in erythrocyte membrane phospholipid composition induced by physical training and physical exercise. Eur. J. Appl. Physiol. Occup. Physiol. 67, 132 – 137. Sundell, K., Jutfelt, F., Augustsson, T., Olsen, R.E., Sandblom, E., Hansen, T., Bjørnsson, B.T., 2003. Intestinal transport mechanisms and plasma cortisol levels during normal and out-ofseason parr–smolt transformation of Atlantic salmon Salmo salar. Aquaculture 222, 265 – 285. Szakolczai, J., 1997. Histopathological changes induced by environmental stress in common carp, Japanese coloured carp, European eel, and African catfish. Acta Vet. Hung. 45, 1 – 10. Tamayo, F.G., Valdes, L.I.T., Lopez, N.M., Alfonso, L.B., 1996. Bacterial translocation and waiting in stressed mice. Arch. Med. Res. 27, 115 – 121. Wang, J.Y., Johnson, L.R., 1991. Polyamines and ornithine decaroxylase during repair of duodenal mucosa after stress in rats. Gastroenterology 100, 333 – 343. Wendelaar Bonga, S.E., 1997. The stress response in fish. Physiol. Rev. 77, 591 – 625. Wong, H.Y., Knight, J.A., Hopfer, S.M., Zaharia, O., Leach, C.N Jr., William Sunderman, F. Jr., 1987. Lipoperoxides in plasma as measured by liquid-chromatographic separation of malondialdehyde–thiobarbituric acid adduct. Clin. Chem. 33, 214 – 220. Yates, D.A., Santos, J., Soderholm, J.D., Perdue, M.H., 2001. Adaptation of stress-induced mucosal pathophysiology in rat colon involves opoid pathways. Am. J. Physiol. 281, G124 – G128. Young, G., 1986. Cortisol secretion in vitro by the interregnal of Coho salmon (Oncorhynchus kisutch) during smoltification: relationship with plasma thyroxine and plasma cortisol. Gen. Comp. Endocrinol. 63, 191 – 200.