Metals in tissues of seabass and seabream reared in sites with oxic and anoxic substrata and risk assessment for consumers

I. Kalantzia,[1], S.A. Pergantisb, K.D. Blackc, T.M. Shimmieldc, N. Papageorgioua, M. Tsapakisd, I. Karakassisa

a Biology Department, University of Crete, Voutes University Campus, 71003, Heraklion, Crete, Greece

b Chemistry Department, University of Crete, Voutes University Campus, 71003, Heraklion, Crete, Greece

c Scottish Association for Marine Science, Scottish Marine Institute, Oban PA34 1QA, Scotland, United Kingdom

d Hellenic Centre for Marine Research, Institute of Oceanography, P.O. Box 2214, 71003 Heraklion, Crete, Greece

Abstract

Twenty-eight metals and elements were measured in the muscle, liver, gills, bone and intestine of farmed seabass and gilthead seabream from four Mediterranean fish farms. The influence of fish species and the effect of environmental conditions on the metal accumulation in fish tissues was investigated. Most concentrations were lower in muscle and higher in liver and bone than in the other body tissues. Seabass accumulates more elements in its tissues than seabream. Fish reared in coarse, oxic sites accumulate more elements with higher concentrations in muscle, bone and intestine and with lower concentrations in liver and gills than fish reared in silty, anoxic sites. This may be attributed to feed type and sediment properties. According to metal pollution index, hazard quotient, selenium health benefit values, carcinogenic risk of arsenic, maximum safe consumption and the permitted limits, the consumption of both farmed species should be considered as safe for human health.

Keywords: trace elements; metals; farmed marine fish; bioaccumulation; sediment geochemistry; human risk assessment

1. Introduction

Fish consumption is beneficial for the maintenance of good human health as it contains important nutrients, such as omega-3 fatty acids, proteins, carbohydrates, liposoluble vitamins, and important micro- and macroelements such as calcium, magnesium, and selenium (Alasalvar, Taylor, Zubcov, Shahidi, & Alexis, 2002). However, fish can also contain undesirable substances such as organochlorinates, PCBs, PAHs, methyl-mercury and toxic metals.

The demand for fish for human consumption is increasing rapidly and most wild stocks are overexploited (Christensen et al., 2014). Therefore, it is expected that aquaculture will have to further increase production this century to provide food for the rapidly increasing population (Duarte et al., 2009). However, fish farms affect the surrounding environment by increasing the amount of organic material, sulphides and fine particles in sediments thereby reducing the redox potential of sediments (Dean, Shimmield, & Black, 2007).

The main sources of metals and other elements in farmed fish are feed, water and suspended particulate matter (Alhashemi, Sekhavatjou, Kiabi, & Karbassi, 2012; Alam, Tanaka, Allinson, Laurenson, Stagnitti, & Snow, 2002; Ferreira, Caetano, Costa, Pousão-Ferreira, Vale, & Reis-Henriques, 2008). Elements are distributed in tissues depending on the type of exposure, i.e. dietary and/ or aqueous (Alam et al., 2002; Ferreira et al., 2008), and they can be excreted through faeces, urine, and respiratory membranes (Alam et al., 2002; Ikem Egilla, 2008). The degree of elemental accumulation in fish depends on nutritional status, seasonal changes, biological variables (species, swimming behaviour, metabolic activities), uptake and elimination kinetics, chemical speciation/ bioavailability, local environmental conditions (salinity, temperature, pH, alkalinity), and concentrations in water and sediments (Kelly, Ikonomou, Higgs, Oakes, & Dubetz, 2008; Canli Atli, 2003). Metals and other elements can be bound by fine sediment particles, organic matter and sulphides and are made less bioavailable to aquatic animals (Dean et al., 2007; Kalantzi et al., 2013a, 2013b) but re-suspension and redistribution of sediments can re-oxidize sediments and release metals and other elements from the sediments to the water column, making them more bioavailable (Eggleton Thomas, 2004).

Mediterranean seabass and gilthead seabream are the two most widely farmed marine species in the Mediterranean (Alasalvar et al., 2002;). Several studies have measured elemental concentrations in the tissues of these two species (Yildiz, 2008; Minganti, Drava, Pellegrini, & Siccardi, 2010; Alasalvar et al., 2002; Ferreira et al., 2008; Cretì, Trinchella, & Scudiero, 2010). However, none of these studies has correlated elemental accumulation with the environmental conditions prevailing at the farming sites. Furthermore, it has often been argued that the preservation of local environmental quality is important for the industry itself since it ensures the quality of their product (Gowen, Ezzi, Rosenthal, & Maekinen, 1990). In this context, it would be interesting to know whether metal accumulation is actually influenced by the ambient environmental conditions at the farming site.

The main objectives of this study were: (1) to test the hypothesis that sediment sulphide, organic content, redox regime and particle size play an important role in binding a range of elements, making them less available to farmed fish and (2) to assess the human health risks due to the consumption of farmed fish. For these purposes, in the present study, 28 metal and other element concentrations were determined in a large number of samples from different fish farms with different environmental conditions. These data were used to determine the distribution of metals and elements in the different tissues of farmed fish, to compare the content of these elements in tissues of farmed seabass and seabream and to compare elemental accumulation in the tissues of farmed fish reared in sites with oxic and anoxic substrata. To assess the human health risks, multiple approaches were used.

2. Materials and methods

2.1 Study areas and samples collection

Farmed species were collected from four seabass (Dicentrarchus labrax) and gilthead seabream (Sparus auratus) farms in Greece - two located in the Aegean Sea (AEG1 and AEG2) and two in the Ionian Sea (ION1 and ION2). Both species are commercially grown in marine fish cages in these farms but, when the samples were taken, gilthead seabream was not available at ION1. Sampling farms are anonymous in this report because the fish farmers agreed to cooperate in the study on the condition that their identities were not revealed. The farms in AEG1 and ION1 are located in shallow exposed straits ca 200 – 300 m from shore. The farm in AEG2 lies in a semi-exposed area and the fourth fish farm in ION2 is located in a shallow and semi-enclosed bay. Sampling sites are described in detail in Kalantzi et al. (2013a, 2013b). A sub-sample of feed pellets was also collected from feed bags at three different farms (AEG2, ION1, ION2) and stored at –20 oC until analysis.

Immediately after collection, fish were killed in ice and transferred to the laboratory. Total lengths and body weights were recorded for each specimen (Table 1). Fish were dissected using a pre-cleaned stainless steel knife and approximately 2 g of each tissue of interest (muscle, liver, gill, bone, intestine) was sampled (Kalantzi et al., 2013a). Fish samples were stored in labeled, zip-lock bags at –20oC until laboratory analysis. Triplicate sediment core samples (4.5 cm internal diameter) were collected by scuba divers under the cages (0 m) as well as at 5 and 10 m from the edge of the cages downstream in the residual current direction in order to measure the environmental variables (sediment particle size, redox potential, sulphide content, refractory organic matter, labile organic matter, chlorophyll-a, total organic carbon, total organic nitrogen) of the surface layer (0–2 cm) of fish farms. Under the cages, scuba divers deployed sediment traps and retrieved them about 48 h later for sedimentation rate estimation. Detailed description of the methods for the determination of all geochemical variables as well as the exact values of physical and geochemical characteristics of sampling sites have been included in Kalantzi et al. (2013a, 2013b).

2.3 Chemical analysis

Metal and other elements concentrations were determined in each sample separately using a modification of the method described by the USEPA (method 3052) for microwave-assisted acid digestion of siliceous and organically based matrices. All samples were freeze-dried to constant weight, homogenised and stored under a dry atmosphere. After predigestion with concentrated acids, acid digestion followed in a closed, high pressure, microwave system (Multiwave 3000, Anton Paar, Austria). For the measurement of metal concentrations in the sample digests, an Inductively Coupled Plasma – Mass Spectrometer (ICP–MS, Thermo Fischer Scientific, Winsford, United Kingdom; Plasma lab software) was used, according to USEPA (method 6020A). Each sample was analyzed in triplicate. An optimal sample dilution factor of x600 was chosen for samples. Internal standard containing Indium and Bismuth (10 µg L-1 or parts per billion, ppb) was added to each sample and standard. A standard was run for every 10 samples analyzed. Elements concentrations were expressed in wet weight. These protocols are described in detail in Kalantzi et al. (2013a) for fish samples and Kalantzi et al. (2013b) for feed samples.

Dataset quality assurance was performed by analysing one blank and one internationally certified reference materials (CRMs, National Research Council of Canada; Joint Research Centre of European Commission) with every 6 samples digested. Average recovery of all elements of DORM-3 (fish protein) was 95.9 ± 10.1 % (n=36); of LUTS-1 (non-defatted lobster hepatopancreas) was 95.5 ± 10.2 % (n=14); and of BCR-668 (mussel tissue) was 102.8 ± 11.8 % (n=32). The elements concentrations in the digestion blanks were typically very low and were subtracted from the sample values. The limits of detection (LOD) of the procedure were calculated by multiplying the standard deviation of the blanks (n=39) by three and were: 0.05 (Li); 32.29 (Na); 3.49 (Mg); 4.43 (Al); 13.58 (P); 23.26 (K); 35.05 (Ca); 0.02 (V); 0.05 (Cr); 6.99 (Fe); 0.004 (Co); 0.43 (Ni); 0.77 (Cu); 10.75 (Zn); 0.01 (Ge); 0.05 (As); 0.05 (Se); 0.01 (Rb); 0.87 (Sr); 0.002 (Y); 0.01 (Mo); 0.001 (Pd); 0.006 (Cd); 0.003 (Cs); 0.05 (Ba); 0.004 (Hg); 0.04 (Pb) and 0.001 (U) mg/kg dry weight.

2.4 Statistical analysis

For statistics, values of metals below the limit of detection (LOD) were arbitrarily set equal to 0.5*LOD. This was done if more than 50% of samples had a concentration of residue that exceeded the LOD; otherwise these metals were excluded from the analysis (Kalantzi et al., 2013a). Non-metric multidimensional scaling (nMDS) ordinations based on Bray–Curtis similarity of log(x+1) transformed data (Clarke & Warwick, 1994) were used for exploring elemental distribution between the different tissues of farmed fish and fish feed and between the different fish farms in each fish tissue. The significance of differences in overall metal concentrations among samples was tested using the Analysis of Similarities (ANOSIM) of the Bray-Curtis similarity matrices (Clarke & Warwick, 1994). These analyses were performed with PRIMER v6 software (Plymouth Marine Laboratories) (Clarke & Warwick, 1994). Differences in metal concentrations between different feeds, between farmed fish species as well as between farmed fish per site were tested by one-way ANOVA followed by pairwise comparisons based on the Tukey HSD test. Normality and homogeneity of variances were tested using Kolmogorov – Smirnov and Levene tests respectively, and transformation of data was conducted when necessary. In cases where the assumptions for ANOVA were not met even after data transformation, the non-parametric Kruskal-Wallis test was used. Univariate analysis was done using the STATISTICA v.8.0 (StatSoft Inc.) software.

2.5. Human risk assessment analysis

To estimate the human risk from consumption of farmed fish from fish farms in Greece multiple approaches were used: (1) Metal pollution index (MPI, μg/g ww; Khillare, Jyethi, & Sarkar, 2012; Singh, Sharma, Agrawal, Marshall, 2010); (2) Hazard quotient (HQ) and total hazard quotient (THQ) (El-Sadaawy, El-Said, Sallam, 2013; Ru, Feng, He, 2013; Singh et al., 2010; Vieira, Morais, Ramos, Delerue-Matos, Oliveira, 2011; Storelli, 2008); (3) Molar ratio of selenium to mercury (molar ratio Se:Hg) and selenium health benefit values (Se-HBV, mole) (Olmedo et al., 2013; Rozic et al., 2014); (4) Carcinogenic risk of arsenic (As-CR; Vieira et al., 2011); (5) Maximum safe consumption (MSCA, kg ww/d; Metian et al., 2013); (6) Comparison of mean concentrations of metals and elements with permitted limits for edible fish tissues.

Mean concentrations of metals and elements in the edible part of the fish (muscle) were used for the calculation of each factor for each species per site. An average adult body weight of the general population of 70 kg and an average consumption of 5.75 g/day of farmed fish were used for the estimation of the As-CR and of the HQ, THQ and MSCA of metals and elements (FAO, 2005-2012). The As-CR, HQ, THQ and MSCA calculated from reference doses (RfD; μg/kg bw/d) of metals and elements as established by the USEPA (2014). The risk assessment of As was estimated only for the potentially toxic, inorganic form which has been found to be 1 – 10% of the total (Onsanit, Ke, Wang, Wang, & Wang, 2010). The oral carcinogenic slope factor was 1.5 (mg/kg bw/d)-1 for As according to Vieira et al. (2011).

3. Results

3.1. Tissue distribution of metals – Farmed fish feed

Mean metal concentrations in tissues of farmed fish and in feed in each site are summarized in Tables 2 –4. As indicated by nMDS ordination, based on metal content, samples of different tissues form distinct clusters (Appendix 1) and feed samples cluster separately indicating also different metal composition from all tissues. This is consistent with the analysis of similarities (ANOSIM) between muscle, liver, gills, bone, intestine and feed (Appendix 2).

In each site and species, we tested the significance of metal differences between muscle, liver, gills and bone, which are the main tissues that store and accumulate metals and other elements of interest (data not presented). Both farmed fish species accumulate Hg, Rb, As, K and Cs in muscle, Cd, Cu, Mo, Se, Zn, V, Fe, Pd and Ge in liver, U, Na and Pb in gills and Ba, Sr, U, Co, Ni, P, Ca, Mg and Li in bone. In both farmed fish species, U, Ba, Cd, Al, Pb, Cu, Co, Fe, Zn, V, Ni, Mo, Ge, Y, Na, Li and Pd in muscle, Sr, Mg, P and Ca in liver, As, Hg, Rb, Cd, Ba, Cu, K, Ni and Cs in gills and Cd, Se, Mo and Cu in bone showed significantly lower concentrations than in the other tissues. All of the above comparisons were statistically significant (ANOVA and Kruskal-Wallis, p<0.05).

Tissues that store and accumulate elements such as muscle, liver, gills and bone have significant differences in some metals compared to the intestine, which also contains undigested or digested material assumed to be representative of the species’ diet. This was tested by comparing the four body tissues and the intestine (including contents) for each site and species. In the intestines of farmed fish, only Mg, Cr, As, Se, Mo, Pd, Cd, Pb and U showed significantly higher concentrations in at least one site and species than in the other four tissues. However, seabass in ION1 also showed higher Al, V, Fe, Zn, Ge and Y concentrations in the intestine than in the other four tissues. Furthermore, for seabass in ION2 none of the metals showed significantly higher concentrations in the intestine than in the other tissues.