Aquatic Toxicology 70 (2004) 179–188 Tissue-specific Cu bioaccumulation patterns and differences in sensitivity to waterborne Cu in three freshwater fish: rainbow trout (Oncorhynchus mykiss), common carp (Cyprinus carpio), and gibel carp (Carassius auratus gibelio) Gudrun De Boeck∗ , Wouter Meeus, Wim De Coen, Ronny Blust Ecophysiology, Biochemistry and Toxicology Group, Department of Biology, University of Antwerp, Groenenborgerlaan 171, B-2020 Antwerp, Belgium Received 6 June 2003; received in revised form 26 July 2003; accepted 12 July 2004 Abstract Rainbow trout (Oncorhynchus mykiss), common carp (Cyprinus carpio) and gibel carp (Carassius auratus gibelio) were exposed to copper (1–20 ␮M) in softened Antwerp City tap water at pH 7.3 ± 0.1 and with a water hardness of 292.4 ± 8.1 mg/L CaCO3 (Ca 100.8 ± 3.0 mg/L; Mg 11.0 ± 0.2 mg/L). LC50s (96 h) were determined and copper accumulation in gills, liver, and kidney assessed over a 10-day period. Rainbow trout (96 h LC50: 3.3 ␮M/210 ␮g/L) were three times more sensitive to Cu exposure than common carp (96 h LC50: 10.4 ␮M/661 ␮g/L) and almost seven times more sensitive than gibel carp (96 h LC50: 22.0 ␮M/1398 ␮g/L). After 96 h, the incipient lethal level (ILL) was reached for common carp, and by the end of the experiment (>120 h) also for rainbow trout. The ILL was never reached for gibel carp. Survival analysis confirmed the differences in sensitivity shown by the 96 h LC50 values. At 1 ␮M Cu, the relative risk to die was six to seven times greater for rainbow trout as for common or gibel carp, respectively, while it was 9000 and 19,000 times greater at 5 ␮M Cu. Only the environmental Cu concentrations contributed significantly (P < 0.001) to the Time-To-Death (TTD). Tissue Cu concentrations did not relate to TTD. Among species, a clear difference in metal handling was apparent, with high liver residues and liver accumulation rates for the most sensitive species, the rainbow trout, and lower liver but higher kidney residues and kidney accumulation rates for the most resistant species, the gibel carp. Gill concentrations and accumulation rates were lowest in the sensitive rainbow trout. © 2004 Published by Elsevier B.V. Keywords: Cu accumulation; LC50; Rainbow trout; Common carp; Gibel carp; Crucian carp 1. Introduction ∗ Corresponding author. Tel.: +32 3 2653347; fax: +32 3 2653497. E-mail address: gudrun.deboeck@ua.ac.be (G. De Boeck). 0166-445X/$ – see front matter © 2004 Published by Elsevier B.V. doi:10.1016/j.aquatox.2004.07.001 It is well established from ecotoxicological studies that different organisms, even with the same life 180 G. De Boeck et al. / Aquatic Toxicology 70 (2004) 179–188 history and under identical exposure conditions, show species specific sensitivities for different kinds of pollutants. However, the mechanisms underlying these differences are not well understood. When examining the toxicity of substances, assessing the lethal concentration at which 50% mortality occurs over a 96-h time interval (96 h LC50) is a standard procedure, which allows comparing different fish species without requiring a prior understanding of the mechanisms involved (Connell et al., 1999). For freshwater fish, the toxicity of heavy metals such as copper depends strongly on the water chemistry and temperature, the size and condition of the organisms, and the season (Shaw and Brown, 1974; Howarth and Sprague, 1978; Peres and Pihan, 1991; Vittozzi and De Angelis, 1991). Therefore, reported LC50 values for each organism vary considerably, which makes it difficult to compare sensitivities among different species and studies. One of the first effects that can be observed after Cu exposure of freshwater fish is a reduction in sodium uptake. This decreased uptake is caused by an inhibition of the Na+ /K+ -ATPase at the Mg2+ binding site (Li et al., 1996), with possible additional effects of a competitive nature between sodium and copper at the apical side (Grosell and Wood, 2002) where copper probably passes through a Cu-specific channel or leaks through a sodium channel (Handy et al., 2002). At higher ambient Cu levels, such as the concentrations typically used in LC50 experiments, sodium efflux and respiratory problems can occur. The differences in respiratory physiology, including differences in ventilation rates and volumes, can lead to different exposure doses and thus different toxic responses (Alabaster and Lloyd, 1980). These differences in respiratory responses to a pollutant might be important. Some fish species, such as trout, respond to increased (0.08–0.47 ␮mol/L) water copper levels with increased respiration and ventilation rates (Beaumont et al., 1995), while others like carp seem to switch to a lower metabolic rate (De Boeck et al., 1995) at similar levels (0.34–0.84 ␮mol/L). This will cause a difference in the amount of water, and thus toxicant, that passes over the gills (McKim and Erickson, 1991). In this study, we compared Cu toxicity for three freshwater fish with different responses and resistances towards hypoxia. We used three different freshwater fish species: (1) rainbow trout (Oncorhynchus mykiss), a typical salmonid which requires high ambient oxygen levels, and starts to hyperventilate as soon as oxygen uptake is compromised; (2) common carp (Cyprinus carpio), a typical cyprinid which can tolerate low oxygen levels for extended periods of time; and (3) gibel carp (Carassius auratus gibelio), which is closely related to the crucian carp (Carassius carassius) and the goldfish (Carassius auratus), and can tolerate total anoxia for several hours to weeks due to their exceptional metabolism that produces ethanol instead of lactate as an end product of anaerobic metabolism. These cyprinids do not hyperventilate that easily when confronted with low oxygen uptake, and rather reduce their energy consumption by lowering their activity. Acute 96 h LC50 values for copper toxicity, supplied by the database of the US Environmental Protection Agency (http://www.epa.gov/ecotox), vary between 32 and 5100 ␮g/L total copper for Oncorhynchus mykiss (Wilson et al., 1994; Buhl and Hamilton, 1990, respectively), between 4 and 160,000 ␮g/L total copper for Cyprinus carpio (Kaur and Dhawan, 1994; Deshmukh and Marathe, 1980, respectively), and between 36 and 1380 ␮g/L total copper for Carassius auratus (Pickering and Henderson, 1996; Johnson and Finley, 1980, respectively). From these data, it is impossible to derive a reliable estimate on how these species relate towards each other where Cu sensitivity is concerned. This study reports the results of a simultaneous 96 h LC50 experiment for rainbow trout, common carp and gibel carp, exposed to copper (1–20 ␮M). Copper accumulation in gills, liver, and kidney were determined in the fish that died during the exposure, and in surviving fish after 10 days of exposure. We examined if any quantitative relationship between time-to-death and tissue copper accumulation occurred. 2. Methods and materials 2.1. Animal holding Rainbow trout, Oncorhynchus mykiss, were obtained from a local trout farm ‘La fontaine aux truites’, Gérouville, Belgium; gibel carp, Carassius auratus gibelio, were obtained at ‘Van Stalle Fish Farm’, Grimbergen, Belgium. Common carp, Cyprinus carpio, were obtained from the fish hatchery at the Wageningen University, The Netherlands. All fish were kept at the University of Antwerp at 17 ± 1 ◦ C in softened Antwerp City tap water for at least 2 months before exposure. Fish were fed ad libitum once a G. De Boeck et al. / Aquatic Toxicology 70 (2004) 179–188 day with either ‘Pond Sticks’ (Tetrapond, Henckel, Germany) for common and gibel carp, or ‘Trouvit’ (Trouw Nutrition, Fontaine-les-Vervins, France) for rainbow trout. 2.2. Experimental set-up Experiments were conducted at 17 ± 1 ◦ C in a flowthrough system consisting of six mixing aquaria (50 L) connected to 12 aquaria (120 L), containing the fish. The photoperiod was set at 16-h light period, 8-h dark period. Flow of softened Antwerp City tap water was set to 600 L a day for each aquarium (five volumes). Water was dechlorinated by aeration in two 3600 L tanks for 24 h before use. Water levels of Ca, Mg, Na and K were checked at regular intervals, and remained at the following levels: Ca 100.8 ± 3.0 mg/L; Mg 11.0 ± 0.2 mg/L; Na 36.7 ± 1.2 mg/L and K 4.6 ± 0.1 mg/L (n = 30, mean ± S.D.). This resulted in a water hardness of 292.4 ± 8.1 mg/L CaCO3 . Water quality was checked daily: pH was 7.3 ± 0.1 (n = 15, mean ± S.D.), oxygen concentrations remained above 90% saturation and ammonia concentrations below 0.1 mg/L at all times. For the exposure experiments, fish of comparable size were selected. Mean weights ± S.D. were 64.2 ± 28.3 g for rainbow trout, 60.6 ± 27.8 g for common carp, and 65.6 ± 21.8 g for gibel carp (n = 50 for each species). In a first series of exposures, three groups of 10 fish per species were exposed to 0, 1.0 or 5.0 ␮M of Cu in separate aquaria. For each species, a second series with two additional groups were exposed to different Cu concentrations: rainbow trout were exposed to intermediate concentrations of 2.5 and 4.0 ␮M of Cu, while common and gibel carp were exposed to 15.0 and 20.0 ␮M of Cu. Fish were not fed during the experiment. Copper was added as copper nitrate (Cu(NO3 )2 ·2H2 O, Merck, Darmstadt, Germany) to the mixing aquaria both manually (at the start of the experiment), and with a peristaltic pump (Watson Marlow 505 S) during the entire experiment. 2.3. Sampling procedures During the first 12 h of each experimental period, aquaria were checked for mortality every hour, thereafter every 24th hour for a period of 10 days. Fish that were no longer breathing and did not respond when 181 touched were considered dead and were removed from the aquarium. After the 10-day period remaining fish (survivors) were quickly netted and killed by an overdose of buffered MS 222 (1 g/L at pH 7.5, Across Organics, Geel, Belgium). All fish were weighed, measured and immediately dissected on ice: gill lamellae, liver tissue and kidney were rinsed with physiological saline (0.9% NaCl), frozen in liquid nitrogen, and stored at −20 ◦ C for copper determination. Total Cu concentration was determined using atomic absorption spectrophotometry (Varian SpectrAA 800). Tissue samples were dried for 24 h at 60 ◦ C, weighed, dissolved in 70% HNO3 , placed in a micro wave until total digestion had occurred and diluted with Milli-Q grade water (Millipore, Bedford, MA, USA). Since for all tissues (except common carp gills) Cu levels in the tissues of fish that had died during the experiment were in the same range as Cu levels in survivors, we assume that post-mortem Cu accumulation between time of dead and sampling was minimal. The chemical speciation of copper was calculated using MINTEQA2 (Allison et al., 1991), using stability constants and enthalpies for inorganic copper species extracted from the critically selected stability constants of metal complexes database (Martell, 2001). The following copper species and associated thermodynamic stability constants were considered in the calculations: log K CuCl+ = 0.40, CuSO◦4 = 2.36, CuOH+ = 6.01, Cu(OH)◦2 = 11.75, CuCO◦3 = 6.74, CuHCO◦3 = 12.25, Cu(CO3 )2 2− = 10.57. The effect of ionic strength on the activity coefficients of the chemical species was calculated using the Davies equation. 2.4. Statistics All values are given as mean ± S.D., except for the lethal concentrations in Table 1, which are means with 95% confidence intervals. These acute LC50 and LC10 values, or lethal concentrations at which 50 or 10% mortality occurs over a given time interval, were calculated by probit analysis using Minitab 14 (Minitab Inc., State College, PA, USA). Accumulation rates shown in Table 2 were calculated using the assumption that the average values of the control group represented the levels in non-exposed fish at the start of the experiment. Statistical differences between Cu concentrations and accumulation rates were calculated by one-way 182 G. De Boeck et al. / Aquatic Toxicology 70 (2004) 179–188 Table 1 LC50 and LC10 values with 95% confidence intervals at different points in time using Probit analysis Oncorhynchus mykiss Cyprinus carpio Carassius auratus gibelio LC50 LC10 LC50 LC10 LC50 LC10 24 h 48 h 72 h 96 h 168 h 6.1 (5.7–7.1) 5.0 (4.6–5.2) 24.5 (22.9–29.1) 19.6 (17.8–20.6) ND ND 4.8 (4.6–4.9) 4.4 (4.2–4.5) 20.7 (20.2–21.2) 17.9 (16.5–18.7) ND ND 3.7 (3.6–3.9) 3.7 (3.6–3.9) 12.0 (11.0–12.9) 8.4 (6.6–9.6) ND ND 3.3 (3.2–3.5) 2.4 (2.1–2.6) 10.4 (9.3–11.6) 7.6 (6.1–8.8) 22.0 (20.4–24.2) 11.7 (9.5–13.3) 3.0 (2.9–3.1) 2.4 (2.2–2.6) 10.4 (9.3–11.6) 7.6 (6.1–8.8) 8.1 (7.2–9.1) 4.2 (3.3–4.9) Values are given as ␮M of measured total Cu (ND: not determined, no mortalities occurred). Table 2 Copper concentrations and calculated accumulation rates in different tissues of control fish and fish exposed to 1 ␮M Cu for 240 h Gill Liver Kidney Oncorhynchus mykiss Cyprinus carpio Carassius auratus gibelio Oncorhynchus mykiss Cyprinus carpio Carassius auratus gibelio Oncorhynchus mykiss Cyprinus carpio Carassius auratus gibelio Control (␮g/g dry weight) 1 ␮M Cu (␮g/g dry weight) Accumulation rate (␮g/g/day) 2.6 ± 0.9 4.8 ± 0.4aaa 3.9 ± 0.8aa,b 319.4 ± 143.2 50.5 ± 21.7aaa 18.8 ± 10.4aaa 10.5 ± 1.6 14.0 ± 2.0 26.9 ± 8.2aaa,bbb 14.4 ± 4.1*** 31.1 ± 5.6***,aaa 35.8 ± 11.0***,aaa 454.9 ± 225.4 141.4 ± 27.3***,aaa 29.6 ± 12.5*,aaa,bbb 17.1 ± 5.5*** 24.7 ± 3.6*** 55.3 ± 10.7***,aaa,bbb 1.12 ± 0.39 2.53 ± 0.54aaa 3.05 ± 1.05aaa 12.85 ± 21.39 8.72 ± 2.65 1.03 ± 1.19 0.63 ± 0.52 1.03 ± 0.34 2.71 ± 1.02aaa,bbb Values are mean ± S.D., n = 10; (*) significantly different from control; (a) significantly different from Oncorhynchus mykiss; (b) significantly different from Cyprinus carpio; one symbol: P < 0.05; two symbols: P < 0.01; three symbols: P < 0.001. analysis of variance (ANOVA), followed by a Tukey–Kramer multiple comparisons test (Statistica 6, StatSoft, Tulsa, OK, USA). Statistical significance was assigned at P < 0.05. For time-to-death (TTD) analysis we used the Cox proportional risk model to calculate TTD, factors contributing to TTD, and correlations between Cu concentrations in water and tissues samples and TTD (Statistica 6, StatSoft, Tulsa, OK, USA). 3. Results Measured total [Cu] in each exposure was on average 93.8 ± 8.5% (n = 15) of nominal Cu. Modelled free [Cu2+ ] was on average 64.0 ± 11.9% (n = 15) of measured total [Cu] and differences among exposure tanks were related to small differences in pH. LC50 values as measured total concentrations are given in Table 1. For gibel carp, only 96 and 168 h LC50 values could be determined since no mortality occurred before that time. Rainbow trout appeared to be four times more sensitive to Cu as common carp during the first 2 days of exposure, but only three times as sensitive afterwards. Gibel carp showed to be seven times more tolerant as rainbow trout, and more than twice as tolerant as common carp. However, towards the end of the experiment gibel carp showed to be more sensitive to Cu, with 20% mortality after 168 h at 5 ␮M of Cu (compared to no mortality for common carp). The incipient lethal level (ILL), or the concentration below which 50% of the organisms will live indefinitely, was reached after 96 h of exposure for common carp. This was also true for rainbow trout after 168 h of exposure. No more deaths were reported between 120 h of exposure up to the end of the experiment (240 h). ILL from this study was about 3 ␮M (191 ␮g/L) for rainbow trout and 10 ␮M (635 ␮g/L) for common carp. Since for gibel carp relatively little mortality occurred late in the experiment, ILL was not reached. Survival analysis takes the survival times of a group of subjects, and generates a survival curve that shows, compared to a baseline curve, how many of the members remain alive over time without having to wait until the last member of the group has actually died. It also G. De Boeck et al. / Aquatic Toxicology 70 (2004) 179–188 183 1e+5 Oncorhynchus mykiss Cyprinus carpio Carassius auratus gibelio Relative risk 1e+4 1e+3 1e+2 1e+1 1e+0 1e-1 0 5 10 15 20 25 Total copper (µM) Fig. 1. The relative risk to die at different Cu concentrations, calculated by the Cox proportional risk model (see Section 3 for more details on the model) for rainbow trout, Oncorhynchus mykiss, common carp, Cyprinus carpio, and gibel carp, Carassius auratus gibelio. allows determining whether the survival curve is influenced by one or more factors, called predictors. In our study we used the Cu concentration in the water, the weight and length of the fish, and the tissue Cu concentrations in gill, liver and kidney as predictors. The model yields a β-value that allows the calculation of a ‘relative risk to die’ under each exposure condition. These data are shown in Fig. 1. Although the model does not allow calculation of error estimates, the results illustrate some clear-cut differences between the species. We found that this risk was six to seven times greater for rainbow trout compared to common or gibel carp at 1 ␮M Cu, while it was 9000 and 19,000 times greater at 5 ␮M Cu. Only the environmental exposure Cu concentrations played a significant role in the risk to die or time-to-death (TTD) during this short-term experiment (P < 0.001 for all species). Cu tissue concentrations or size of fish did not play a significant role (P > 0.05 for all species). The analysis allows calculating correlations between TTD and the different predictors. For all three species, we found a significant negative correlation between TTD and the Cu concentration in the water (R = −0.81 for rainbow trout (n = 21); R = −0.98 for common (n = 19) and gibel carp (n = 23)). This corresponds well with the pronounced effect on the survival curve. For rainbow trout, a positive correlation was found be- tween gill Cu concentration and water Cu concentration (R = 0.54, n = 21), but no significant correlation existed between gill Cu concentrations and TTD. In common carp, gill Cu concentrations were correlated with water Cu concentrations (R = 0.59, n = 19), as well as with TTD (negatively correlated: R = −0.54, n = 19). A significant correlation existed between Cu levels in liver and kidney tissue (R = 0.64, n = 19). For gibel carp, the only significant correlation was found between Cu concentrations in gill and liver tissue (R = 0.47, n = 23). No other significant correlations were found. In general, relationships between water Cu concentrations and tissue Cu concentrations were poor or absent. This could originate from the different survival times, which were sometimes rather short. Considering survivors only, we see a more uniform pattern with a better correlation between water and tissue Cu concentrations (results not shown). Fig. 2 shows the survival time in relation to the tissue Cu concentrations. It clearly demonstrates the lack of a relationship between tissue Cu concentrations and time to death, as most of the tissue concentrations in the death fish are in the same range as concentrations in the survivors (killed at 240 h). Some remarkable species-specific differences become apparent when comparing the tissue Cu concentrations. Care must be taken in Survival time (hours) G. De Boeck et al. / Aquatic Toxicology 70 (2004) 179–188 Oncorhynchus mykiss 184 240 240 240 192 192 192 144 144 144 96 96 96 48 48 48 0 0 Survival time (hours) 100 200 300 400 0 0 200 400 600 800 1000 240 240 240 192 192 192 144 144 96 96 48 48 0 0 0 Survival time(hours) Carassius auratus gibelio Cyprinus carpio 0 100 200 300 400 0 200 400 600 800 1000 192 192 144 144 144 96 96 96 48 48 48 0 80 100 0 20 40 60 80 100 0 192 400 60 48 240 100 200 300 Cu concentration (µg/g) Gill 40 96 240 0 20 144 240 0 0 0 200 400 600 800 1000 Cu concentration (µg/g) Liver 0 0 20 40 60 80 Cu concentration (µg/g) Kidney 100 Fig. 2. Survival time in relation to gill, liver and kidney copper concentrations (␮g/g dry weight) for rainbow trout, Oncorhynchus mykiss, common carp, Cyprinus carpio, and gibel carp, Carassius auratus gibelio exposed to different Cu concentrations (0–20 ␮M). Data given for dead (black circles), surviving (grey circles) and control (white diamonds) fish. comparing the data since exposure concentrations for the different fish are not necessarily the same. The results clearly show that liver Cu concentrations for rainbow trout are much higher than for the two other species even before exposure started, and this remains the case despite the fact that rainbow trout were exposed to much lower Cu levels. When comparing common and gibel carp, which were exposed to the same range of Cu levels, a clear difference in gill Cu accumulation occurred, with much higher levels in common carp gills (the only tissue that showed a correlation with TTD). In fact, Cu levels in gibel carp gills were in the same range as those in rainbow trout, while the exposure levels showing mortality were at least three to four times higher. Liver Cu levels hardly increased in gibel carp, while in common carp an increase can be observed compared to the non-exposed control fish. The opposite seems to be true for kidney levels, where Cu levels for gibel carp accumulated up to twice the levels in common carp and rainbow trout. To enable a comparison between all three species, Table 2 gives the tissue Cu levels for the control and lowest exposure concentration (1 ␮M), which was sublethal to all three species. All tissues from the three species accumulated Cu; except for trout liver where variation in Cu levels among individuals was so large that the 42% increase was not significant. Cu levels in control rainbow trout gills were lower than in common and gibel carp gills, and the increase of the Cu level in these gills also remained low. Thus the low Cu accumulation levels demonstrated in Fig. 2 are not only the consequence of the lower exposure concentration, but also of the low accumulation rate. At this low Cu exposure we see no evidence for the massive increase in Cu in common carp gills, as was observed for the fish that had died. Accumulation rates even remained below G. De Boeck et al. / Aquatic Toxicology 70 (2004) 179–188 those for gibel carp gills, which showed only limited Cu accumulation at lethal concentrations. Liver Cu levels of rainbow trout were considerably higher than those of common carp, which in turn were higher than the gibel carp levels. The increase in Cu residues in gibel carp liver were minor compared to the increase observed in rainbow trout and common carp liver. Gibel carp, on the other hand, had higher kidney Cu levels and accumulated more Cu in the kidney during the experiment. These effects correspond with the observations made at the lethal exposure concentrations. 4. Discussion LC50 values (96 h) from this study were 210 ␮g/L (3.3 ␮M) for rainbow trout, 660 ␮g/L (10.4 ␮M) for common carp and 1400 ␮g/L (22.0 ␮M) for gibel carp. In a recent comparative study, Sappington et al. (2001) found 96 h LC50 for rainbow trout of 80 ␮g/L. Other trout species in that study had 96 h LC50 values in the same range, while 96 h LC50 values of the cyprinids tested (fathead minnow, bonytail chub and Colorado pikeminnow) ranged between 220 and 470 ␮g/L. However, fish used in that study were considerably smaller (below 1 g) and the water hardness was lower (160–180 mg/L as CaCO3 ). Considering species mean acute values normalised for a hardness of 50 mg/L CaCO3 as given in the ambient water quality criteria for copper (US EPA, 1985), yields 42.5 ␮g/L for rainbow trout, 156.8 ␮g/L for common carp, and 157.1 ␮g/L for goldfish, a relative of the gibel carp. Normalising our own data to a hardness of 50 mg/L CaCO3 using the same conversion formula (LC50 at 50 mg/L = eln(LC50) − 0.9422 × (ln(hardness) − ln(50)) ) gives 96 h LC50 values of 40 ␮g/L for rainbow trout, 125 ␮g/L for common carp and 265 ␮g/L for gibel carp. These values are comparable for rainbow trout and common carp, but gibel carp clearly differ from the values obtained for goldfish. Thus, it appears that freshwater salmonids and cyprinids in general differ with a factor 3 in sensitivity towards Cu toxicity. Our study shows that gibel carp are considerably more tolerant to Cu during the first few days of exposure. Although LC50 values only differ by a factor 3–10, the actual impact of Cu exposure at a given concentration can be over 10,000 times higher for the more sensitive species, as indicated by the risk analysis. 185 Some remarkable differences in Cu accumulation occur between species. Gill Cu concentrations and Cu accumulation rates in rainbow trout are significantly lower than in common or gibel carp. Normally, gill Cu levels reach a steady state after 2–3 days in rainbow trout (Grosell et al., 1996), and Cu levels in tissues other than gill and liver accumulate only when excess Cu from the liver spills over to other tissues (Laurén and McDonald, 1987a,b). The reason for the low accumulation rates in rainbow trout gills could be twofold. Either Cu enters slowly at the apical site, or the binding capacity for Cu within the cells is low and Cu excretion on the basolateral side is fast. Our measurements show that rainbow trout have excessively high Cu levels in the liver compared to the two cyprinids, even before the Cu exposures started. Also accumulation rates are the highest for rainbow trout liver. This suggests that Cu does enter the bloodstream swiftly, and the low accumulation rate in the gills is not caused by a low apical entry. Other studies confirm the high Cu levels in the liver for juvenile (140–150 g) rainbow trout: 46–47 ␮g/g wet weight (Handy, 1992), or 60 ␮g/g wet weight (Dethloff et al., 1999) with dry weight values being about five-fold the wet weight values (own measurements). The discrepancy between Cu accumulation in common carp gills at lethal and sublethal exposure concentrations might be explained by physical gill damage which allows the Cu to enter the gills in large amounts only at the lethal concentrations where the damage occurs. Moreover, gill histology of common carp during Cu exposure in a 10-day experiment, showed more severe gill damage during the first days of exposure compared to the two other species (De Boeck et al., 2001 and unpublished results). It is remarkable that hardly any negative correlation between TTD and a tissue Cu concentration is found. Such a relationship could indicate that an increase of the Cu level in this tissue is causing the toxic effects and thus reduced survival time. The lack of a clear doseresponse relationship in the gills, or any other tissue, could also indicate that surface related processes play an important role in Cu toxicity. As mentioned in the introduction, Cu can induce respiratory distress in fish, and it is striking that the most hypoxia sensitive species is also the most Cu sensitive species, while the gibel carp shows to be resistant towards anoxia as well as metal exposure. The inhibition of oxygen uptake is attributed to excessive mucus production and epithelial 186 G. De Boeck et al. / Aquatic Toxicology 70 (2004) 179–188 swelling (McDonald and Wood, 1993), but as mentioned above, gill histology showed that gill damage and epithelial swelling was more severe in common carp than rainbow trout (De Boeck et al., 2001 and unpublished results). Second candidate is a disturbed ion regulation in gill cells followed by osmotic influx of water causing oedema and cell rupture, as has been shown for common carp (De Boeck et al., 2001) and rainbow trout (Sola et al., 1995) at elevated Cu levels. Toxic effects of Cu and subsequent acclimation to Cu are consistent with the damage–repair model (McDonald and Wood, 1993). This model includes an initial damage phase during which the pathophysiological effects of the metal (mainly disturbance of ionregulation) are expressed and reach maximal values, followed by a repair phase with attendant bioenergetic cost during which the effects diminish and may even disappear (Taylor et al., 2000). The fast occurrence of mortality in this study confirms the importance of this first shock phase, with ionregulatory and respiratory distress, at least for rainbow trout. Also for common carp, this phase might be important considering the massive Cu influx at the gills observed at the lethal exposure concentrations. Why this shock phase with pronounced mortality during the first hours and days is not observed for gibel carp remains to be investigated. In chronic studies, more subtle effects of Cu toxicity may play a role, including endocrine disrupting effects (Handy, 2003). These effects are partially controlled by neuro-endocrine changes in the Cu exposed fish. This might be the basis for the fact that also in chronic studies the relation between tissue Cu concentrations and effect is weak. In a chronic Cu exposure study with rainbow trout, it was concluded that elevated metal burdens in the gill and liver of exposed fish were measures of chronic copper exposure, but not of effect (Taylor et al., 2000). Changes in gill Cu-binding characteristics offered the most promise for assessing the effect of chronic exposure, but there was no clear-cut relationship between water chemistry, gill Cu binding, and toxicity. Cu accumulation under acutely lethal conditions differs from accumulation under chronic exposure conditions, where at least Cu accumulation was found to depend on dose and time, and effects such as tissue damage and reduced growth can to some extent be correlated with internal dose (Farag et al., 1995; Marr et al., 1996). However, a toxicokinetic study of Cu in rainbow trout suggested that long term water- borne Cu exposure resulting in continued Cu accumulation in slowly exchangeable pools should not result in toxicological consequences (Carbonell and Tarazona, 1994). Thus, even under chronic exposure conditions, care should be taken when using Cu residues as an indicator for toxic effects. Gibel carp, the most resistant species in this study, has the lowest metal residues in the liver. Birge et al. (2000) state that variation in metal body burdens among organisms exposed under similar conditions allow an additional means of identifying or confirming the classification of indicator organisms as sensitive, moderately tolerant, or tolerant. The caddisfly Cheumatopsyche, an opportunistic and metal tolerant organism, sustained appreciable lower metal residues and exhibited substantially greater mean densities in their study. In our study, this is only true for liver Cu levels and not for gill and kidney Cu residues. However, Handy (1992) showed that after Cu exposure, liver Cu levels represented about 60% of the total body burden and thus present the major part of body Cu. It is clear that copper handling between the different species is very different. A comparison between eels and salmonids exposed to comparable levels of Cu also indicated a marked difference between the mechanisms of copper acclimation in both species (Grosell et al., 1996). They concluded that enhanced excretion of metals, rather than detoxified protein-bound storage, would be a biological advantage in the long run. Perhaps the increased kidney residues and high Cu accumulation rates in gibel kidney indicate an active role in Cu handling and excretion in gibel carp compared to common carp and rainbow trout. This might be part of the explanation why gibel carp are more resistant to metal exposure. 5. Conclusion Based on LC50 values, we conclude that Oncorhynchus mykiss is three times more sensitive to Cu exposure than Cyprinus carpio, and seven times more sensitive than Carassius auratus gibelio. However, this results in a considerably larger risk for rainbow trout when exposed to Cu: the relative risk to die at 5 ␮M Cu is up to 10,000 times greater for rainbow trout compared to gibel carp. No clear-cut relationships were found between tissue Cu residues or Cu accumulation rates and toxicity in these 10-day exposures. Surface G. De Boeck et al. / Aquatic Toxicology 70 (2004) 179–188 related processes, such as gill damage, could therefore play an important role in Cu toxicity. At lower Cu exposure levels, differences in metal handling become apparent, with high liver residues and liver accumulation rates for the most sensitive species, the rainbow trout, and lower liver but high kidney residues and kidney accumulation rates for the most resistant species, the gibel carp. Surprisingly, gill concentrations and accumulation rates were lowest for the sensitive rainbow trout. Acknowledgements Gudrun De Boeck a post-doctoral fellow of the Fund for Scientific Research-Flanders (FWO-Vlaanderen). This research is supported by grant 1.5.197.02 from the Fund for Scientific Research-Flanders (FWOVlaanderen) and RAFO/1 DEBOG KP00 from the University of Antwerp. References Alabaster, J.S., Lloyd, R., 1980. Water Quality Criteria for Freshwater Fish. Butterworths, London, UK. Allison, J.D., Brown, D.S., Novo-Gradec, K.J., 1991. Minteq A2/Prodef A2, a geochemical assessment model for environmental systems. EPA/600/3-91/021. US EPA Environmental Research Lab, Athens, GA, USA. Beaumont, M.W., Butler, P.J., Taylor, E.W., 1995. Exposure of brown trout, Salmo trutta, to sublethal copper concentrations in soft acidic water and its effect upon sustained swimming performance. Aquat. Toxicol. 33, 45–63. Birge, W.J., Price, D.J., Shaw, J.R., Spromberg, J.A., Wigginton, A.J., Hogstrand, C., 2000. Metal body burden and biological sensors as ecological indicators. Environ. Toxicol. Chem. 19, 1199–1212. Buhl, K.J., Hamilton, S.J., 1990. Comparative toxicity of Inorganic contaminants released by placer mining to early life stages of Salmonids. Ecotoxicol. Environ. Saf. 20, 325–342. Carbonell, G., Tarazona, J.V., 1994. Toxicokinetics of copper in rainbow trout (Oncorhynchus mykiss). Aquat. Toxicol. 19, 213–221. Connell, D., Lam, P., Richardson, B., Wu, R., 1999. Introduction in Ecotoxicology. Blackwell Science, Malden, MA, USA. De Boeck, G., De Smet, H., Blust, R., 1995. The effect of sublethal levels of copper on oxygen consumption and ammonia excretion in the common carp Cyprinus carpio. Aquat. Toxicol. 32, 127–141. De Boeck, G., Vlaeminck, A., Balm, P.H.M., De Wachter, B., Lock, R.A.C., Blust, R., 2001. Morphological and metabolic changes in common carp, Cyprinus carpio, during copper exposure: direct effects of Cu2+ or indirect effects of plasma cortisol elevation? Environ. Toxicol. Chem. 20, 347–381. 187 Deshmukh, S.S., Marathe, V.B., 1980. Size related toxicity of copper and mercury to Lebistes reticulates (Peter) Labeo rohita (Ham.) and Cyprinus carpio (Linn). Indian J. Exp. Biol. 18, 421– 423. Dethloff, G.M., Schlenk, D., Khan, S., Bailey, H.C., 1999. The effects of copper on blood and biochemical parameters of rainbow trout (Oncorhynchus mykiss). Arch. Environ. Contam. Toxicol. 36, 415–423. Farag, A.M., Stansbury, M.A., Hogstrand, C., MacConnell, E., Bergman, H.L., 1995. The physiological impairment of freeranging brown trout exposed to metals in the Clarck Fork River Montana. Can. J. Fish. Aquat. Sci. 52, 2038–2050. Grosell, M., Wood, C.M., 2002. Copper uptake across rainbow trout gills: mechanisms of apical entry. J. Exp. Biol. 205, 1179– 1188. Grosell, M., Boëtius, I., Hansen, H.J.M., Rosenkilde, P., 1996. Influence of pre-exposure to sublethal levels of copper: 64 Cu uptake and distribution among tissues of the European eel (Anguilla anguilla). Comp. Biochem. Physiol. 114C, 229–235. Handy, R.D., 1992. The assessment of episodic metal pollution. I. Uses and limitations of tissue contaminant analysis in rainbow trout (Oncorhynchus mykiss) after short waterborne exposure to cadmium or copper. Arch. Environ. Contam. Toxicol. 22, 74– 81. Handy, R.D., 2003. Chronic effects of copper exposure versus endocrine toxicity: two sides of the same toxicological process? Comp. Biochem. Physiol. 135A, 25–38. Handy, R.D., Eddy, F.B., Baines, H., 2002. Sodium-dependent copper uptake across epithelia: a review of rationale with experimental evidence from gill and intestine. Biochim. Biophys. Acta 1566, 104–115. Howarth, R.S., Sprague, J.B., 1978. Copper lethality to rainbow trout in water of various hardness and pH. Water Res. 12, 455–462. Johnson, W.W., Finley, M.T., 1980. Handbook of acute toxicity of chemicals to fish and aquatic invertebrates. Resour. Publ. 137 Fish Wildl. Serv. USDI, Washington, DC, USA. Kaur, K., Dhawan, A., 1994. Metal toxicity to different life stages of Cyprinus carpio Linn. Indian J. Ecol. 21, 93–98. Laurén, J.D., McDonald, D.G., 1987a. Acclimation to copper by rainbow trout, Salmo gairdneri: Physiology. Can. J. Fish. Aquat. Sci. 44, 99–104. Laurén, J.D., McDonald, D.G., 1987b. Acclimation to copper by rainbow trout Salmo gairdneri: Biochemistry. Can. J. Fish. Aquat. Sci. 44, 105–111. Li, J., Lock, R.A.C., Klaren, P.H.M., Swartz, H.G.P., Schuurmans Stekhoven, F.M.A.H., Wendelaar Bonga, S.E., Flik, G., 1996. Kinetics of Cu2+ inhibition of Na+ /K+ -ATPase. Toxicol. Lett. 87, 31–38. Marr, J.C.A., Lipton, J., Cacela, D., Hansen, J.A., Bergman, H.L., Meyer, J.S., Hogstrand, C., 1996. Relationship between copper exposure duration, tissue copper concentration, and rainbow trout growth. Aquat. Toxicol. 36, 17–30. Martell, A.E., 2001. NIST 46. NIST critically selected stability constants of metal complexes: version 6.0. National Institute of Standards and Technology Virtual Library, Gaithersburg, MD, USA. McDonald, D.G., Wood, C.M., 1993. Branchial mechanisms of acclimation to metals in freshwater fish. In: Rankin, J.C., Jensen, 188 G. De Boeck et al. / Aquatic Toxicology 70 (2004) 179–188 F.B. (Eds.), Fish Ecophysiology. Chapman & Hall, London, UK, pp. 297–321. McKim, J.M., Erickson, R.J., 1991. Environmental impacts on the physiological mechanisms controlling xenobiotic transfer across fish gills. Physiol. Zool. 64, 39–67. Peres, I., Pihan, J.C., 1991. Copper LC50 to Cyprinus carpio. Influence of hardness, seasonal variation, proposition of maximal acceptable toxicant concentration. Environ. Technol. 12, 161– 167. Pickering, Q.H., Henderson, C., 1996. The acute toxicity of some heavy metals to different species of warm water fishes. Int. J. Air Water Pollut. 10, 453–463. Sappington, L.C., Mayer, F.L., Dwyer, F.J., Buckler, D.R., Jones, J.R., Ellersieck, M.R., 2001. Contaminant sensitivity of threatened and endangered fishes compared to standard surrogate species. Environ. Toxicol. Chem. 20, 2869–2876. Shaw, T.L., Brown, V.M., 1974. The toxicity of some forms of copper to rainbow trout. Water Res. 8, 377–382. Sola, F., Isais, J., Masoni, A., 1995. Effects of copper on gill structure and transport function in the rainbow trout Oncorhynchus mykiss. J. Appl. Toxicol. 15, 391–398. Taylor, L.N., McGeer, J.C., Wood, C.M., McDonald, D.G., 2000. Physiological effects of chronic copper exposure to rainbow trout (Oncorhynchus mykiss) in hard and soft water: evaluation of chronic indicators. Environ. Toxicol. Chem. 19, 2298–2308. US Environmental Protection Agency, 1985. Ambient water quality criteria for copper-1984. EPA 440/5-84-031. Office of Water Regulations and Standards Criteria and Standards Division, Washington, DC, USA. Vittozzi, L., De Angelis, G., 1991. A critical review of comparative acute toxicity data on freshwater fish. Aquat. Toxicol. 19, 167–204. Wilson, R.W., Bergman, H.L., Wood, C.M., 1994. Metabolic costs and physiological consequences of acclimation to aluminum in juvenile rainbow trout (Oncorhynchus mykiss). 1. Acclimation 51, 527–535.