Sex and low-level sampling stress modify the impacts of sewage effluent on the rainbow trout (Oncorhynchus mykiss) immune system

(1)

Aquatic Toxicology 73 (2005) 79–90

Sex and low-level sampling stress modify the impacts of sewage effluent on the rainbow trout (Oncorhynchus mykiss)

immune system

Birgit Hoeger

^a

, Bettina Hitzfeld

^b

, Bernd K¨ollner

^c

, Daniel R. Dietrich

^a

, Michael R. van den Heuvel

^d^,∗

aEnvironmental Toxicology, University of Konstanz, Germany

bSwiss Agency for the Environment, Forests andLandscape, Berne, Switzerland

cFriedrich-Loefﬂer-Institute, Greifswald-Insel Riems, Germany

dForest Research, Private Bag 3020, Sala Street, Rotorua, New Zealand Received 10 January 2005; received in revised form 21 March 2005; accepted 22 March 2005

Abstract

The objective of the present study was to investigate the influence of chronic exposure to municipal sewage treatment efflu- ent at environmentally relevant concentrations on immune parameters in rainbow trout (Oncorhynchus mykiss), including the assessment of potential differences in reactivity between sexually mature male and female fish. Trout were exposed to 1.5 and 15% (v/v) secondary treated municipal sewage effluent for 32 weeks. Fish were injected intra-peritoneally either with inactivated Aeromonas salmonicida to simulate an infection or with PBS as control for this immune challenge 6 weeks prior to sampling.

Exposure to effluent resulted in a decrease in A. salmonicida-specific serum antibody level and blood lymphocyte numbers in mature females, but not in male fish. Injection of A. salmonicida resulted in enhanced serum lysozyme activity in mature male trout, which were not exposed to effluent. This stimulating effect of A. salmonicida could not be found in effluent-exposed trout, again potentially revealing a suppressive effect of the effluent. An influence of sampling fish on two consecutive days was observed in many immune parameters, most likely reflecting handling stress. Leucocyte and lymphocyte numbers in peripheral blood were consistently lower in male and female fish on the second sampling day. Phagocytosis in head kidney macrophages from male trout was also influenced by sampling day, whereby a stimulation of this reaction occurred on the second day of sampling. Liver mixed function oxygenase activity was found to be enhanced in mature male trout exposed to 15% effluent. In conclusion, the study showed, that exposure to sewage treatment plant effluent, in surface water relevant concentrations, can lead to potentially adverse effects on selected immune reactions in rainbow trout. However, this study also demonstrated that both handling stress and the sex of mature fish have distinct influences on the immune response detected in male and female fish and are likely to influence measured immune parameters to the extent that subtle effluent induced changes may be difficult to detect.

Keywords: Rainbow trout; Oncorhynchus mykiss; Fish immune system; Immunotoxicology; Aeromonas salmonicida; Handling stress; Sewage treatment plant; Effluent

∗Corresponding author. Tel.: +64 7 343 5899; fax: +64 7 343 5695.

E-mail address: mike.vandenheuvel@forestresearch.co.nz (M.R. van den Heuvel).

doi:10.1016/j.aquatox.2005.03.004

Konstanzer Online-Publikations-System (KOPS) URL: http://www.ub.uni-konstanz.de/kops/volltexte/2008/4987/

URN: http://nbn-resolving.de/urn:nbn:de:bsz:352-opus-49879

(2)

1. Introduction

In the last two decades sewage treatment effluents have been found to contain a broad array of anthropogenic substances, including pharmaceutical residues (Daughton, 2001; Kolpin et al., 2002), chemicals contained in household products, such as various derivatives of alkylphenol (Bennie, 1999), as well as natural and synthetic hormones (Desbrow et al., 1998). In spite of usually being present in low concentrations, some of these environmental pollutants have the potential to affect the health of aquatic organisms, due to their specificity for eliciting effects on certain biochemical functions. Pharmaceuticals in particular are designed to specifically influence subtle biochemical mechanisms and might thus affect similar functions in non-target aquatic species in the low concentrations found in aquatic environments.

Hormone-like activity of several low-level pollutants in our surface waters has already been investigated (Ashfield et al., 1998), while effects on immune mechanisms have not received a similar level of attention.

However, as some of the commonly found pollutants are known to alter immune reactions or, in the case of some pharmaceuticals, are even designed to influence the immune system, it should be assumed that not only reproductive-endocrine mechanisms can be affected in aquatic organisms, but that immune reactions might also be a target for some of the aquatic low-level pollutants. Effects on various immune parameters in fish have already been shown for several substances, well known to be present in sewage effluent, including PCBs (Khan, 2003), PAHs (Weeks and Warinner, 1986; Karrow et al., 2001; Reynaud et al., 2003), pesticides (Dunier and Siwicki, 1993; Gal- loway and Handy, 2003) and antibiotics (Lund´en and Bylund, 2000). In the only study examining exposure to wastewater in the field, caged carp (Cyprinus car- pio) were exposed to river water receiving sewage ef- fluent (Price et al., 1997). Exposed fish displayed a significant reduction in the proliferative ability of T- and B-lymphocytes, and a decrease in serum lysozyme activity in comparison to fish from a reference site. In a laboratory experiment, exposure of goldfish (Carra- sius auratus) to treated sewage resulted in a decrease in erythrocyte, granulocyte and lymphocyte numbers in blood and lower blood cell phagocytic activity (Kakuta, 1997).

A key element to assessing the competency of the immune system is to examine responses with and without the presence of an immune activator. One such approach is to expose the organism to inactivated pathogens. The infection model applying inactivated A.

salmonicida in trout is known to result in an activation of different leukocyte populations e.g. monocytes and B-cells leading to a specific antibody response against bacterial antigens (K¨ollner and Kotterba, 2002). Fur- thermore, MHC II expression was found to be increased in monocytes, B-lymphocytes and a part of the thrombocytes, indications of a general stimulation of the en- tire immune system (Stachowski et al., 2004).

Stress is known to be an important factor in fish health and is also a potential confounding factor in ex- periments investigating the immune functions of fishes.

Effects of stress on immune parameters in fish have been investigated in various studies and e.g. a suppressed antibody response against sheep red blood cells in Atlantic salmon (Salmo salar) (Einarsdottir et al., 2000) and reduced resistance against Trypanoplasma borrelii infection in carp (C. carpio) (Saeij et al., 2003) has been detected after rearing stress due to repeated water level reduction and daily handling, respectively.

Acute stress, however, has also been found to enhance specific immune reactions, as has been shown for lysozyme activity in rainbow trout (O. mykiss) after a 10 min handling stress (Demers and Bayne, 1997).

Depending on the specific situation, handling stress may thus enhance or oppose possible adverse effects of water pollutants on immune reactions in fish.

The aim of the present study was to determine if exposure to a sewage effluent from a modern wastewater treatment plant affects immune functions in rainbow trout (O. mykiss). This species was chosen as a rep- resentative of teleosts because of the large availability of validated test methods and immunological knowl- edge. In order to provide relevance and eliminate confounding environmental factors such as energy intake, long-term exposure of sexually maturing rainbow trout in a mesocosm facility was chosen as the method of exposure. General immune parameters, like differential blood cell counts, reactions of the unspecific immune system, including serum lysozyme activity and head kidney macrophage activity (phagocytosis and oxidative burst), as well as specific antibody produc- tion against the trout pathogen Aeromonas salmonicida salmonicida were determined. The high relevance of

(3)

immune parameters as endpoints chosen in this chronic exposure study with sewage effluent is based on the potential of imminent immune suppression to lead to higher disease susceptibility, thus likely endangering the health of fish populations.

2. Material and methods 2.1. Experimental design

Two-year-old rainbow trout (Oncorhynchus mykiss) were exposed for 32 weeks to a nominal concentration of 1.5% (v/v) or 15% (v/v) municipal sewage treatment plant (STP) effluent in 12,000 L tanks containing 50 individuals. Final treated effluent was obtained from a municipal sewage treatment plant located in Rotorua, New Zealand every 2 days. This STP employs a pre- treatment step with stop screens and a grit trap, a pri- mary treatment step with sedimentation and secondary activated sludge treatment (Bardenpho Reactor).

Activated carbon de-chlorinated tap water (aquifer source) was used as the diluent and as the reference treatment. Two replicate tanks were used for each treatment. Trout exposure tanks were provided with a constant water flow of 10 L/min, which resulted in a 95% replacement time of approximately 45 h. Dis- solved oxygen, pH and conductivity (Radiometer Pa- cific, Auckland, New Zealand) in the fish exposure tanks and in undiluted effluent were measured daily.

Additional aeration was provided in the effluent hold- ing tanks and the trout exposure ponds maintaining dissolved oxygen above 90% saturation for the duration of the experiment as measured on a daily basis. The average pH-values (±S.D.) in the exposure tanks were 7.21±0.25 and 7.22±0.25 in the 15% effluent tanks, 7.15±0.28 and 7.20±0.28 in the 1.5%

effluent tanks and 7.13±0.29 and 7.13±0.28 in the control tanks. Conductivity in each tank, diluent conductivity, and effluent conductivity was used to calcu- late the actual effluent concentration in the fish tanks on a daily basis. The actual mean effluent concentration was 13.48±3.25 and 12.91±3.51 in the 15% effluent tanks and 1.45±0.59 and 1.59±0.64 in the 1.5% effluent tanks. Water temperature ranged between 12.7 and 16.7^◦C, 12.7 and 16.9^◦C and 12.9 and 17.5^◦C in control water, 1.5% effluent and 15% effluent, respectively.

All trout were tagged with individually numbered T-bar

type tags (HallPrint Pty Ltd., Holden Hill, SA, Aus- tralia). Weight and length for each tagged trout were measured at the start of the experiment. Trout were fed daily with commercial feed pellets (Reliance Stock Food, Dunedin, New Zealand) at a ration of 0.7% of the wet body weight and this ration was adjusted monthly to compensate for growth. The exposure was started on September 22, 2001, with immature fish and continued for 32 weeks. After an exposure period of 26 weeks trout were anesthetized with ethyl-3-aminobenzoate methanosulfonate (MS222; Fluka, Switzerland) and 1 mL of blood was taken by syringe from the caudal vein. Fish were then either injected with formalde- hyde inactivated A. salmonicida salmonicida strain MT 423 antigen (1×10⁸ cells in 25␮L per 100 g body weight) (kindly provided by G¨unter Kotterba, Friedrich-Loeffler-Institute, Insel Riems, Germany), or with a corresponding volume of phosphate balanced salt solution (PBS) as a control for the injection. Anti- gen preparation of A. salmonicida followed the descrip- tion byK¨ollner and Kotterba (2002). Fish were exposed for a further 6 weeks and were sampled between May 6 and 14, 2002, just prior to the natural spawning period for this strain of trout. Female trout were sampled over two consecutive days on May 6 and 7. Male trout were sampled over two consecutive days on May 13 and 14, one week later than the females in order to allow them to recover from the short-term stress due to removal of females from the tanks. At the time of sampling, nor- mal aggressive feeding behaviour had resumed in the male trout. This particular strain of trout typically has a proportion of individuals that do not mature at 3 years of age. Immature trout were not examined for immune responses at the termination of the experiment. How- ever, since blood was removed from these fish at the 26 weeks time-point, lysozyme was examined in con- junction with the maturing males and females sampled at this time-point.

2.2. Sampling

Fish were sacrificed with a blow on the head. Blood was removed by caudal puncture and a blood smear was prepared immediately. Blood for serum samples was collected in untreated vacutainers. These samples were kept on ice and subsequently centrifuged at 1000×g (10 min, 4^◦C) to obtain serum, which was then kept at −80^◦C pending analysis. Head kidney

(4)

for macrophage preparation was collected and kept on ice in Leibovitz’s L-15-medium (Invitrogen, Auck- land, New Zealand), containing 53 mU heparin sodium salt/mL (Sigma, St. Louis, USA) and 100 U/mL of Penicillin/Streptomycin (P/S) (Invitrogen, Auckland, New Zealand). A 1 g sample of liver was frozen in liquid nitrogen for subsequent determination of 7- ethoxyresorufin-O-deethylase (EROD) activity.

2.3. Immunological parameters

Blood smears were stained according to Pappen- heim (May-Gruenwald/Giemsa staining). Pictures of the blood smears were taken with an SV Micro Sound Vision color camera (Sound Vision Inc., Boston, USA) on a microscope (Zeiss Axiolab) using Axio Vi- sion Version 2.0.5. (Carl Zeiss Vision GmbH, Hall- bergmoos, Germany). A total of 1500 cells per slide were counted. The different blood cell populations were expressed in percent of total cells counted.

The preparation of head kidney macrophages was performed according to Secombes (1990) using a Percoll gradient (Sigma, St. Louis, USA). The resulting cell suspension was adjusted to a density of 1×10⁶cells/mL in L-15 medium, containing P/S, and 100␮L/well of this suspension was seeded into 96-well black fluorometer plates (BMG Labtechnolo- gies, Offenburg, Germany). After incubation at 18^◦C for 90 min to allow attachment of macrophages, media was removed by inverting the plate on a paper towel.

For the phagocytosis assay, a volume of 100␮L of a 250␮g/mL fluorescein-labeled E. coli suspen- sion (K-12 strain, Molecular Probes, Eugene, USA) was added to each well including eight blank wells, which did not contain macrophages. After incubation at 18^◦C for 2 h, media was removed, followed by the addition of 100␮L trypan blue solution (0.025%). The trypan blue solution was removed after 1 min of incubation and fluorescence was measured in a microplate fluorometer (Polarstar Galaxy, BMG Labtechnologies, Offenburg, Germany; 485 nm excitation, 520 nm emission).

For the oxidative burst assay, either 200 ng/mL phorbol-myristate-acetate (PMA) (Sigma, St. Louis, USA) in HBSS or HBSS alone was pipetted onto the cells. Measurement of the oxidative burst reaction was started 5 min later by the addition of 2-

7-dichlorodihydrofluorescein diacetate (H₂DCFDA) (Molecular Probes, Eugene, USA) at a concentration of 10␮g/mL. The time course of H₂O₂-production was measured in a microplate fluorometer (485 nm excitation, 520 nm emission). The slope was obtained from data measured from 4 to 10 min. Results were calculated as pmol 2-7-dichlorofluorescein (DCF) (Acros, Schwerte, Germany) produced per minute per well.

The method for measuring serum lysozyme fol- lows the description of Ellis (1990) (turbimetric assay). Briefly, 950␮L M. lysodeikticus solution (Sigma, Steinheim, Germany) (0.5 mg/mL in 0.05 M sodium phosphate buffer, pH 6.2) were pipetted in a cuvette and measured in a spectrophotometer at 530 nm, followed by the addition of 50␮L of serum or plasma.

Absorption was then measured after 30 s, 1, 2, 3, 4 and 4.5 min and lysozyme activity was expressed as decrease of optical density (OD) per min.

The ELISA for the detection of A. salmonicida- specific antibodies in trout serum followed the method described by K¨ollner and Kotterba (2002). Briefly, Maxisorp ELISA plates (Nunc, Roskilde, Denmark) were coated with 20␮g/mL A. salmonicida salmoni- cida strain MT 423 antigen (kindly provided by G¨unter Kotterba, Friedrich-Loeffler-Institute, Insel Riems, Germany) overnight at 4^◦C. Plates were blocked with 1% ovalbumin (Sigma, Steinheim, Germany) in PBS for 1 h at 20^◦C, followed by washing three times with PBS containing 0.05% Tween 20 (T-PBS) (Carl Roth GmbH, Karlsruhe, Germany). After incubation with diluted serum samples (1:4000 in T-PBS) for 1 h at 20^◦C and another wash cycle, plates were in- cubated with monoclonal antibody 4C10 (mouse-anti trout immunoglobulin M) for 1 h at 20^◦C. After washing, goat-anti-mouse IgG/IgM-peroxidase conjugate (Pierce, Rockford, USA) was added for 1 h at 20^◦C. Af- ter washing three times with T-PBS, detection was car- ried out with 3,3,5,5-tetramethylbenzidine (TMB) according to manufacturers description (Sigma, St. Louis, USA). The colour reaction was stopped by the addition of 1 M H₂SO₄and absorption was measured at 450 nm in an SLT plate reader 340 ATTC (SLT Labinstruments, Groedig, Austria). As no standards were available for IgM determination, results are given as OD. To enable comparison without a standard curve all samples were measured in parallel in a single ELISA run, which was repeated twice.

(5)

2.4. Liver 7-ethoxyresoruﬁn-O-deethylase (EROD) activity

Hepatic mixed function oxygenase (MFO) activity was assessed in liver post-mitochondrial super- natant (PMS) as EROD activity using a fluorescence plate-reader method. Liver extracts were homogenised in 0.1 M phosphate buffer containing 1 mM EDTA, 1 mM dithiothreitol, and 20% glycerol at pH 7.4. Ho- mogenates were centrifuged at 9000×g to obtain PMS. EROD reaction mixture contained 0.1 M HEPES buffer pH 7.8 (Sigma, St. Louis, MO, USA), 5.0 mM Mg²⁺, 0.5 mM NADPH (Applichem, Darmstadt, Ger- many), 1.5␮M 7-ethoxyresorufin (Sigma, St. Louis, MO, USA), and about 0.5 mg/mL of PMS protein. Re- sorufin was determined using 544 nm excitation and 590 nm emission filters. Protein content was estimated from fluorescamine (Sigma, St. Louis, USA) fluorescence (390 nm excitation, 460 nm emission filters) against bovine serum albumin (Sigma, St. Louis, USA).

2.5. Statistics

Immune parameters in mature male and female trout were analysed independently using ANOVA with data appropriately transformed to meet the assump- tions of homogeneity of variance and normality. The variability of replicate tanks was tested as a variable in all ANOVAs and did not add significant vari- ability for any of the endpoints. Sampling day, A.

salmonicida-injection, effluent concentration and any possible interactions were examined as variables in all subsequent analyses. Where no significant variability due to the tested variables was found, that variable was eliminated from the model. Dunnett’s test for post hoc comparisons of treatment groups to the reference group was performed on the final ANOVA model. Where statistical interactions were observed to occur, the data was split based on the interacting variable (sampling day or injection) and an independent ANOVA was performed for each possible category of that particular variable. EROD data were compared using a non-parametric Kruskal–Wallis one-way analysis of variance with Bonferroni adjustment for multiple comparisons. All statistical testing was completed using the SYSTAT^®software package (Wilkinson, 1990).

The level of confidence used in determining statistical differences for all analyses wasα= 0.05.

3. Results

External examination of trout at 26 and 32 weeks indicated no obvious signs of external opportunistic infections such as fin erosion, fungus, or other skin le- sions, though fins did occasionally show damage consistent with confinement due to aggression or abrasion on the tanks. Slightly increased mortality was experi- enced in the 15% effluent during the summer months, whereby 5% of the initially 100 trout of this group died.

3.1. Innate immune measurements

Data on the different immune parameters investigated were analysed for effects of effluent exposure and A. salmonicida-injection as well as consequences of sampling on two consecutive days possibly leading to handling stress in remaining fish after the first sampling day. Differences in reactivity of immune functions between male and female trout were also assessed.

Leucocyte and lymphocyte numbers in peripheral blood were not influenced by A. salmonicida-injection in either males or females. Leucocyte numbers as well as lymphocyte numbers were, however, consistently and statistically significantly lower in male and female fish on the second sampling day compared to day 1 (Fig. 1). There were consistent trends for reductions in lymphocytes and total leucocytes with increasing effluent concentration, however, a statistically significant effect of effluent exposure on peripheral blood cell numbers could only be found for lymphocytes in females on the first but not on the second sampling day. Also, no significant effects were found on the proportion of circulating thrombocytes, monocytes and granulocytes in either sex due to effluent expo- sure, A. salmonicida-injection, or sampling day (data not shown).

For the oxidative burst assay, males and females showed very different patterns of response (Fig. 2a and b). Females showed an entirely different pattern of response on day 1 as opposed to day 2 as indicated by a significant interaction between effluent exposure and sampling day. When data from sampling days were examined separately, effluent- exposed females showed significantly increased oxidative burst on day 1. There was a non-significant

(6)

Fig. 1. Peripheral blood leucocyte numbers in female (a) and male trout (b) and lymphocyte numbers in female (c) and male fish (d) after 32 weeks of exposure to effluent. Female and male fish were each sampled on two consecutive days (days 1 and 2) and data was accordingly analysed for effects of effluent exposure and sampling day. Shown are mean values with standard error of the means (S.E.M). Data was tested using a three-way ANOVA with Dunnett’s post-test. n≥6;^*p < 0.05.

trend for oxidative burst to decrease with increasing effluent dose in females on day 2. Oxidative burst in males was not altered by effluent exposure, but reactivity on day 2 was significantly lower than on day 1.

As with oxidative burst, kidney macrophage phagocytosis showed very different patterns of response to sampling day in males and females (Fig. 2c and d).

Phagocytosis in females was not affected by sampling day, whereas in males, phagocytosis greatly increased on day 2 of sampling. There were no effects of sewage effluent on phagocytosis in either males or females.

Serum lysozyme activity was measured after 26 weeks of exposure and at the end of the experiment after 32 weeks. Immature fish (including male and female fish) displayed a significant decrease in serum lysozyme activity after exposure to 1.5% or 15% effluent for 26 weeks (Fig. 3a). In mature female trout at 26 weeks, exposure to 1.5 and 15% effluent also resulted in decreased lysozyme activity, however, only

statistically significant in the group exposed to 1.5%

effluent. In mature male fish, exposure to effluent had no effect on serum lysozyme activity after 26 weeks of exposure. Upon termination of the experiment (32 weeks) only data for mature male trout was analysed, as immature fish were not sampled at this time-point, and the majority of serum samples from female fish caused precipitation in the M. lysodeik- ticus solution used for the lysozyme assay. Expo- sure to 1.5% or 15% STP effluent for 32 weeks did not have an effect on serum lysozyme activity. How- ever, in control fish exposed to de-chlorinated tap water, i.p.-injection with the inactivated trout pathogen A. salmonicida resulted in a significant increase in serum lysozyme activity compared to PBS-injected fish (Fig. 3b). In the fish exposed to effluent, no difference in lysozyme activity between pathogen-injected and sham-injected trout could be found. Serum lysozyme activity was not observed to be sensitive to handling stress.

(7)

Fig. 2. Head kidney macrophage oxidative burst in female (a) and male trout (b) and macrophage phagocytosis in female (c) and male fish (d) after 32 weeks of exposure to effluent. Female and male fish were each sampled on two consecutive days (days 1 and 2) and data was accordingly analysed for effects of effluent exposure and sampling day. Shown are mean values with S.E.M. Data was tested using a three-way ANOVA with Dunnett’s post-test, n≥8 for (a) and (c); n≥5 for (b) and (d);^*p < 0.05;^***p < 0.001.

3.2. Aeromonas salmonicida-speciﬁc antibody ELISA

Trout sham-injected with PBS and kept together in the same exposure tanks with A. salmonicida-injected fish did not show detectable serum levels of anti-A.

salmonicida antibodies, excluding a possible cross- contamination. As no detectable antibody production was evident, samples from sham-injected fish were ex- cluded from further analysis. Male and female trout again showed a different pattern of response. In mature female trout, exposure to 1.5% or 15% STP effluent resulted in lower serum levels of specific antibodies, pro- duced against inactivated A. salmonicida. The decrease in serum antibody levels was, however, only statistically significant in the group exposed to 1.5% effluent (Fig. 4). Mature male trout did not display significantly altered levels of specific antibodies after exposure to effluent compared to tap water exposed fish and unlike antibody levels in females this measure tended to in-

crease with effluent concentration. Serum level of anti- A. salmonicida IgM were not observed to be influenced by stress due to consecutive sampling days.

3.3. Liver EROD activity

Exposure to 15% effluent over 32 weeks resulted in a significant increase in liver EROD activity in mature male fish compared to control fish (Fig. 5). In mature female fish, no effect of STP effluent on liver EROD activity could be shown.

4. Discussion

Chronic exposure of rainbow trout to treated sewage effluent resulted in slightly increased mortality, presumably due to increased ammonia levels occasionally detected in the effluent in the course of regular moni- toring at the sewage treatment plant (Tim Charleson,

(8)

Fig. 3. Lysozyme activity in serum samples, taken after exposure to effluent for: (a) 26 weeks (prior to injection with A. salmonicida, n≥21) and (b) after 32 weeks (n≥7), where data was analysed for effects of effluent exposure and A. salmonicida-injection. Shown are mean values with S.E.M. Data was tested using one-way ANOVA with Dunnett’s post-test;^*p < 0.05;^**p < 0.01.

Fig. 4. A. salmonicida-specific antibody level in trout serum sam- ples after exposure to effluent for 32 weeks. Data was tested using one-way ANOVA with Dunnett’s post-test. Shown are mean values with S.E.M. Male trout: n = 12 for the groups exposed to 0 and 15%

effluent; n = 7 for the 1.5% effluent group. Female fish: n = 12 for the groups exposed to 0 and 1.5% effluent; n = 8 for the 15% effluent group;^*p < 0.05.

Fig. 5. Liver EROD activity in trout after exposure to STP effluent for 32 weeks. Shown are median values with S.E.M. Data was tested with Kruskall–Wallis non-parametric one-way ANOVA.

n≥17;^***p < 0.001.

Rotorua District Council, personal communication).

Exposure of rainbow trout to STP effluents induced a suppression of peripheral blood lymphocyte numbers, serum lysozyme activity and specific antibody production, while the production of reactive oxygen species was enhanced only in effluent-exposed female fish.

Significantly reduced peripheral blood lymphocyte numbers were also observed in an earlier experiment with juvenile rainbow trout exposed to high concentrations of effluent from the same STP for 28 days (Hoeger et al., 2004b). Similarly,Kakuta (1997) re- ported lower lymphocyte numbers in goldfish (C. au- ratus) after exposure to treated effluent in a laboratory experiment. A possible cause for the decrease of lymphocyte numbers could be polyaromatic hydrocarbons (PAH), as has been described before byKhan (2003), who found lower circulating lymphocyte numbers in winter flounder (Pleuronectes americanus) sampled at a PAH-polluted boat wharf. The presence of PAHs in the effluent used for the study at hand has been suggested by analysis of bile samples of trout exposed to higher concentrations of effluent from the same STP for 28 days (Hoeger et al., 2004b). As PAHs are known to be strong activators of cytochrome P450 1A1, the significant induction of liver EROD activity in effluent- exposed trout, whether as juveniles for 28 days (Hoeger et al., 2004b) or as adult male fish for 32 weeks (this study), further supports the assumption that these fish were exposed to PAH via the sewage treatment effluent used. At first sight it seems contradictory to explain the reduced lymphocyte numbers observed in female

(9)

trout by exposure to PAHs while an increase in EROD activity was not demonstrated in mature females in the present study. However, in this case the lack of EROD induction does not exclude exposure to PAHs, as EROD activity has previously been found to be suppressed in reproductively mature female trout, presumably due to an estrogen dependent mechanism (van den Heuvel and Ellis, 2002).

The decreased ability of mature female trout to produce A. salmonicida-specific antibodies was most likely a direct consequence of the lower circulating lymphocyte numbers in peripheral blood, after exposure to STP effluent. Immunohistology in spleen samples from the same exposure experiment revealed an activation of B-cells in effluent-exposed female trout, implied by higher surface immunoglobulin M expression on this cell type (Hoeger et al., 2004a). At first sight, this observation seems to be in contradic- tion to the observed lower specific serum antibody levels. However, the activation of B-lymphocytes in the spleen might be the result of effluent-induced unspecific stimulation resulting in a polyclonal stimulation of B-cells and, hence, appear unrelated to the produc- tion of A. salmonicida-specific antibodies. A possi- ble effluent-induced stimulatory effect on head kidney macrophages was observed, as suggested by the increased oxidative burst in female trout. Whether the observed stimulation of B-lymphocytes in spleen is a direct consequence of antigen presentation by these macrophages following effluent exposure cannot be determined with the data obtained in the present study.

A. salmonicida-injected trout in the effluent- exposed groups did not display enhanced lysozyme activity compared to the PBS-injected fish, further suggesting an immune-suppressive effect of this effluent, in this case specifically on the capability of the fish to express this bacteriolytic enzyme. In- deed, A. salmonicida-injected trout in the control groups showed the expected enhanced lysozyme activity (Moyner et al., 1993), thus demonstrating that the formaldehyde inactivated A. salmonicida was capable of stimulating an immune response. The suppressive effect of STP effluent on serum lysozyme activity was also observed at an earlier time-point in this study, i.e.

at the 26-week time-point in non-pathogen-stimulated immature and female fish. Suppression of lysozyme activity has also been shown previously in carp, which were held in an effluent receiving river (Price et al.,

1997). PAHs have moreover been considered to be the active component of creosote leading to reduced plasma lysozyme level in rainbow trout (Karrow et al., 1999). Whether the suppression of lysozyme activity was due to the suggested concomitant PAH exposure with the effluents cannot be determined, as chemical analysis of PAHs in the effluent used were not con- ducted. However, the assumed effect of PAHs in the effluent on lysozyme activity is consistent with the other seemingly PAH-related immune modulations described above and thus suggests that PAHs could be the immune-suppressive component in the STP effluents tested in this study.

The seeming lack of effect of A. salmonicida on im- mune parameters other than lysozyme activity might be due to the time-point chosen for sampling (6 weeks after injection). It is possible that fast reacting innate immune functions had actually been stimulated by pathogen-injection, but have already returned to base- line values at the time of sampling. Effects expected by A. salmonicida may thus have simply been missed due to a time schedule inappropriate to examine fast reacting immune parameters.

Effects of different kinds of stress on immune parameters have been studied extensively, especially in correlation with the role of cortisol as an important stress mediator (Narnaware et al., 1994; Espelid et al., 1996; Rotllant et al., 1997). However, in studies on pollution effects, possible consequences of handling stress, due to sampling procedures are often ignored. In the present study, results were also analysed for effects of handling stress due to the two consecutive sampling days. In agreement with the literature, decreased leucocyte and specifically lymphocyte numbers in both male and female fish on the second sampling day were found.

It could be speculated that the mechanism resulting in decreased lymphocyte numbers is based on handling stress on the first sampling day leading to cortisol re- lease and subsequently apoptosis of B-lymphocytes in peripheral blood. This idea is derived from previous studies in carp (C. carpio) revealing induction of apop- tosis in peripheral blood lymphocytes caused by cortisol (Verburg-van Kemenade et al., 1999; Saeij et al., 2003).

Barton (2002) has stated that stress is not neces- sarily detrimental to fish, as long as homeostasis can be regained by adaptive mechanisms. In this context, only severe or long-lasting stressors may be regarded

(10)

as detrimental to fish health. However, little is known about the combined effects of different kinds of stressors, like e.g. handling stress in combination with poor water quality. Depending on the immune parameter investigated and the stressors that act on the fish, effects may be additive or opposed. In the case of additive effects, adverse effects of aquatic pollution on fish health become exceedingly likely. In the field of environmental toxicology, special consideration should be given to the distinction between the actual pollution effects and the influence of stress, which can hardly be com- pletely avoided in any experimental procedure. Some alleged toxicological effects of a certain environmental contaminant might actually stem from, or at least be influenced by, unconsidered stressors.

The present study also underlines the dependence of different immune parameters in salmonids on sexual maturation and subsequently the sex of the investigated fish, as demonstrated by differences in sensi- tivity of serum lysozyme activity to effluent exposure in immature, male and female fish, as well as sensi- tivity of specific antibody production against effluent exposure in female, but not male trout. Differences in the response to effluent between male and female fish have also been found for macrophage activity, whereby only female fish but not male trout showed increased production of reactive oxygen species. A similar observation has been described byFournier et al. (1998), who found that exposure of mummichogs (F. heterocli- tus) to pulp mill effluent resulted in a decreased oxida- tive burst response in female fish only, but not in male fish. In general, female trout seemed to be more sensitive to effluent exposure than male fish in the study at hand. It is well established that the immune system of fishes is modulated by hormones such as cortisol, growth hormone, reproductive hormones and others (Harris and Bird, 2000). Sex also modulates the stress response as reproductively active female fish have been observed to demonstrate enhanced cortisol responses while males show depressed cortisol responses to handling stress (Pottinger and Carrick, 2000). The sex differences observed in this study were generally consistent with the increased responsiveness of female trout to stress.

The precipitation of the M. lysodeikticus solution caused by addition of serum samples from mature female fish, gained at the end of the experiment (after 32 weeks), is likely also related to sexual develop-

ment. Substances like lectin, C-reactive protein (CRP) and lysozyme are present in fish eggs as a defense against infection (Yousif et al., 1994a, 1994b; Nuno- mura, 1991) and are likely to be present in higher amounts in blood upon maturation of female fish. The capacity to agglutinate bacteria has e.g. been shown for egg lectin in coho salmon (O. kisutch) (Yousif et al., 1994b).

While, in recent years, the applicability of endocrine parameters for the assessment of aquatic pollution has been widely investigated and discussed, immunotoxicology in fish is still undeveloped due to the lack in the understanding of basic immunology in fishes and the immense complexity of the immune system. The experimental design, needed to enable the assessment of immune parameters, is complicated by various charac- teristics of the immune system. As immune reactions are temporally very dynamic it might be necessary to concentrate more on time courses of the investigated reactions, rather than using endpoint determination, which is a snapshot in time that misses most of the events occurring (Koellner et al., 2005). An example of this can be seen in this study: the duration of post- challenge exposure was optimised in order to measure A. salmonicida-specific antibody production, which is slow in trout. However, this overlooked rapid responses, such as oxidative burst, that are well known to respond to A. salmonicida within very short periods of time.

A better understanding of the fish immune system is essential for the interpretation of results, suggesting effects of environmental contamination on immune parameters in fish. Identification of mechanisms and patterns of action of pollutants in the aquatic environment would facilitate estimation of adverse effects on fish health. While the study at hand provides clear evidence for adverse effects of sewage effluents, and most likely the PAHs therein, on the fish immune system and how that response can be modulated by stress and sexual development, further studies are necessary to provide a better assessment of possible toxicological effects of STP effluents on overall immune competence in fish.

Acknowledgements

The study was supported by a travel grant from the Universit¨atsgesellschaft e.V., University of Konstanz

(11)

and by a scholarship to B. Hoeger from the German Federal Environmental Foundation, Osnabr¨uck. Re- search in New Zealand was funded by the Founda- tion for Research Science and Technology and by a Royal Society of New Zealand Travel Grant. The au- thors would like to thank Dr. Willi Nagl, Dr. Rosanne Ellis, Megan Finley, Murray Smith, G¨unter Kotterba and Tim Charleson.

References

Ashfield, L.A., Pottinger, T.G., Sumpter, J.P., 1998. Exposure of female juvenile rainbow trout to alkylphenolic compounds results in modifications to growth and ovosomatic index. Environ. Tox- icol. Chem. 17, 679–686.

Barton, B.A., 2002. Stress in fishes: a diversity of responses with particular reference to changes in circulating corticosteroids. Integ.

Comp. Biol. 42, 517–525.

Bennie, D.T., 1999. Review of the environmental occurance of alkylphenols and alkylphenol ethoxylates. Water Qual. Res. J.

Canada 34, 79–122.

Daughton, C.G., 2001. Emerging pollutants, and communicating the science of environmental chemistry and mass spectrometry: pharmaceuticals in the environment. J. Am. Soc. Mass Spectrom. 12, 1067–1076.

Demers, N.E., Bayne, C.J., 1997. The immediate effects of stress on hormones and plasma lysozyme in rainbow trout. Dev. Comp.

Immunol. 21, 363–373.

Desbrow, C., Routledge, E.J., Brighty, G.C., Sumpter, J.P., Waldock, M., 1998. Identification of estrogenic chemicals in STW effluent. 1. Chemical fractionation and in vitro biological screening.

Environ. Sci. Technol. 32, 1549–1558.

Dunier, M., Siwicki, A.K., 1993. Effects of pesticides and other organic pollutants in the aquatic environment on immunity of fish:

a review. Fish Shellfish Immunol. 3, 423–438.

Einarsdottir, I.E., Nilssen, K.J., Iversen, M., 2000. Effects of rearing stress on Atlantic salmon (Salmo salar L.) antibody response to a non-pathogenic antigen. Aquacult. Res. 31, 923–930.

Ellis, A.E., 1990. Lysozyme assays. In: Van Muiswinkel, W.B. (Ed.), Techniques in Fish Immunology. SOS Publications, Fair Haven, pp. 101–103.

Espelid, S., Lokken, G.B., Steiro, K., Bogwald, J., 1996. Effects of cortisol and stress on the immune system in Atlantic salmon (Salmo salar L.). Fish Shellfish Immunol. 6, 95–110.

Fournier, M., Lacroix, A., Voccia, I., Brousseau, P., 1998. Phago- cytic and metabolic activities of macrophages from mummichog naturally exposed to pulp mill effluents in the Miramichi River.

Ecotoxicol. Environ. Saf. 40, 177–183.

Galloway, T., Handy, R., 2003. Immunotoxicity of organophospho- rous pesticides. Ecotoxicology 12, 345–363.

Harris, J., Bird, D.J., 2000. Modulation of the fish immune system by hormones. Vet. Immunol. Immunopathol. 77, 163–176.

Hoeger, B., Koellner, B., Kotterba, G., Van Den Heuvel, M.R., Hitzfeld, B., Dietrich, D.R., 2004a. Influence of chronic expo-

sure to treated sewage effluent on the distribution of white blood cell populations in rainbow trout (Oncorhynchus mykiss) spleen.

Toxicol. Sci. 82, 97–105.

Hoeger, B., Van den Heuvel, M.R., Hitzfeld, B., Dietrich, D.R., 2004b. Effects of treated sewage effluent on immune function in rainbow trout (Oncorhynchus mykiss). Aquat. Toxicol. 70, 345–355.

Kakuta, I., 1997. Effect of sewage on blood parameters and the resis- tance against bacterial infection of goldfish, Carassius auratus.

Environ. Toxicol. Water Qual. 12, 43–51.

Karrow, N.A., Boermans, H.J., Dixon, D.G., Hontella, A., Solomon, K.R., Whyte, J.J., Bols, N.C., 1999. Characterizing the immuno- toxicity of creosote to rainbow trout (Oncorhynchus mykiss): a microcosm study. Aquat. Toxicol. 45, 223–239.

Karrow, N.A., Bols, N.C., Whyte, J.J., Solomon, K.R., Dixon, D.G., Boermans, H.J., 2001. Effects of creosote exposure on rainbow trout pronephros phagocyte activity and the percentage of lym- phoid B cells. J. Toxicol. Environ. Health 63, 363–381.

Khan, R.A., 2003. Health of flatfish from localities in Placentia Bay, Newfoundland, contaminated with petroleum and PCBs. Arch.

Environ. Contam. Toxicol. 44, 485–492.

Koellner, B., Wasserrab, B., Fischer, U., Kotterba, G., van den Heuvel, M.R., 2005. How can toxic effects of pollution of the aquatic environment on the immuno-competence of fishes be detected? A discussion on the relevance of biomarkers philosophy.

In: Dietrich, D.R., Webb, S.F., Petry, T. (Eds.), Hot spot pollutants: pharmaceuticals in the environment. Academic Press.

K¨ollner, B., Kotterba, G., 2002. Temperature dependent activation of leucocyte populations of rainbow trout, Oncorhynchus mykiss, after intraperitoneal immunisation with Aeromonas salmonicida.

Fish Shellfish Immunol. 12, 35–48.

Kolpin, D.W., Furlong, E.T., Meyer, M.T., Thurman, E.M., Za- ugg, S.D., Barber, L.B., Buxton, H.T., 2002. Pharmaceuticals, hormones, and other organic wastewater contaminants in U.S.

streams, 1999–2000: a national reconnaissance. Environ. Sci.

Technol. 36, 1202–1211.

Lund´en, T., Bylund, G., 2000. The influence of in vitro and in vivo exposure to antibiotics on mitogen-induced proliferation of lym- phoid cells in rainbow trout (Oncorhynchus mykiss). Fish Shell- fish Immunol. 10, 395–404.

Moyner, K., Roed, K.H., Sevatdal, S., Heum, M., 1993. Changes in non-specific immune parameters in Atlantic salmon, Salmo salar L. induced by Aeromonas salmonicida infection. Fish Shellfish Immunol. 3, 253–265.

Narnaware, Y.K., Baker, B.I., Tomlinson, M.G., 1994. The effect of various stresses, corticosteroids and adrenergic agents on phago- cytosis in the rainbow trout Oncorhynchus mykiss. Fish Physiol.

Biochem. 13, 31–40.

Nunomura, W., 1991. C-reactive protein in eel: purification and ag- glutinating activity. Biochim. Biophys. Acta 1076, 191–196.

Pottinger, T.G., Carrick, T.R., 2000. Contrasting seasonal modulation of the stress response in male and female rainbow trout. J. Fish Biol. 56, 667–675.

Price, M.A., Jurd, R.D., Mason, C.F., 1997. A field investigation into the effect of sewage effluent and general water quality on selected immunological indicators in carp (Cyprinus carpio L.).

Fish Shellfish Immunol. 7, 193–207.

(12)

Reynaud, S., Duchiron, C., Deschaux, P., 2003. 3- Methylcholanthrene inhibits lymphocyte proliferation and increases intracellular calcium levels in common carp (Cyprinus carpio L.). Aquat. Toxicol. 63, 319–331.

Rotllant, J., Pavlidis, M., Kentouri, M., Abad, M.E., Tort, L., 1997.

Non-specific immune responses in the red porgy Pagrus pagrus after crowding stress. Aquaculture 156, 279–290.

Saeij, J.P., Verburg-van Kemenade, L.B., van Muiswinkel, W.B., Wiegertjes, G.F., 2003. Daily handling stress reduces resistance of carp to Trypanoplasma borreli: in vitro modulatory effects of cortisol on leukocyte function and apoptosis. Dev. Comp. Im- munol. 27, 233–245.

Secombes, C.J., 1990. Isolation of salmonid macrophages and analysis of their killing activity. In: Van Muiswinkel, W.B. (Ed.), Techniques in Fish Immunology. SOS Publications, Fair Haven, pp. 101–103.

Stachowski, A., Hordvik, I., Koppang, E., K¨ollner, B., 2004. Charac- terization of anti-salmon-MHC class II monoclonal antibodies.

Immunobiology 209, 481.

van den Heuvel, M.R., Ellis, R.J., 2002. Timing of exposure to a pulp and paper effluent influences the manifestation of reproductive effects in rainbow trout. Environ. Toxicol. Chem. 21, 2338–2347.

Verburg-van Kemenade, B.M.L., Nowak, B., Engelsma, M.Y., Weyts, F.A.A., 1999. Differential effects of cortisol on apoptosis and proliferation of carp B-lymphocytes from head kidney, spleen and blood. Fish Shellfish Immunol. 9, 405–415.

Weeks, B.A., Warinner, J.E., 1986. Functional evaluation of macrophages in fish from a polluted estuary. Vet. Immunol. Im- munopathol. 12, 313–320.

Wilkinson, L., 1990. SYSTAT: The System for Statistics. SYSTAT Inc., Evanston, IL, USA.

Yousif, A.N., Albright, L.J., Evelyn, T.P.T., 1994a. In vitro evidence for the antibacterial role of lysozyme in salmonid eggs. Dis.

Aquat. Org. 19, 15–19.

Yousif, A.N., Albright, L.J., Evelyn, T.P.T., 1994b. Purification and characterization of a galactose-specific lectin from the eggs of coho salmon Oncorhynchus kisutch and its interaction with bac- terial fish pathogens. Dis. Aquat. Org. 20, 127–136.