Gull island is located in the eastern suburbs of Guangzhou, China. It is surrounded by the main channel of the Pearl River and the Lotus Mountain Waterway. As one of the main breeding areas in Guangzhou, gull island is mainly used for the cultivation of four famous domestic fishes (Grass carp, Mylopharyngodon piceus, Hypophthalmichthys molitrix and Aristichthys nobilis). About 30,000 tons of fish products are provided for Guangzhou and other regions from the gull island every year. A fish pond selected in the gull island is representative to explore the variations of heavy metals in the fish pond ecosystem during a complete breeding cycle. A total of 670 kg of grass carp fry (weight 150 - 250 g) was put into the pond covering an area of 2160 m² in December 2017 and six months were set as a complete breeding time according to the local breeding experience.

A total of 108 surface water, feeds and sediment samples were collected in triplicate, as well as 46 fish samples were obtained from the pond in gull island, China (Figure 1) during December 2017 and January, March, April, May and June of 2018. Surface water samples (50 cm below the air-water interface) were collected in polyethylene bottles. Feed samples were obtained from the fish farm and stored in polyethylene bags. Surface sediment samples (to a depth of 10 cm) were collected using a Petersen’s grab and sealed in polyethylene bags. At the same time, fish samples were captured using fishing nets and sealed in polyethylene bags in an ice cube box. All samples were transported to the laboratory immediately after collection. The water, feed and fish samples were kept at 4˚C until processing, whereas the sediment samples were frozen at −20˚C before further analysis.

Water was acidified with HNO₃ to PH < 2 and filtered using 0.45 μm nylon filters to estimate the presence of Cr, Ni ,Cu, Zn, As, Cd and Pb [11]. The feed samples were grinded with a mortar and pestle to screen through a 100-mesh nylon sieve. Each feed sample of 0.5 g was digested with 10 mL HNO₃ to determine the target metals [12]. In the case of sediment, it was air-dried, ground in an agate mortar and passed through a 100-mesh nylon sieve. Sediment samples (0.1 g) were digested with the mixture of 3 mL HNO₃, 1 mL HCL and 1 mL HF to measure the metal concentrations [13]. Fish samples were cleaned with deionized water and measured the biometric data (fish length and fish weight) (TableS1). The fish samples were dissected using a scalpel to separate the tissues (including the muscle, skin, bladder, gill, heart, kidney and liver). The separated tissues were washed, freeze-dried at a freeze dryer (Christ ALPHA 1 - 2 LD plus, Germany) for 3 - 4 d and then uniformly ground into powder until the metal concentration was measured. Ultra pure nitric acid (10 mL) was added for every dry tissue sample of 0.5 g to analysis the metals [12]. A 350X inductively coupled plasma mass spectrometer (Perkin Elmer; Waltham, MA, USA) was applied to analysis the studied metals in digested solutions.

All glassware and polyethylene bottles used in this study were pre-soaked with 10% HNO₃ for 24 h, rinsed with ultrapure water and then air dried before use. Three samples including one procedural bank, one matrix spike sample and one bank spike sample were analyzed along with every batch digestion samples (10 samples). The recoveries of Cr, Ni, Cu, Zn, As, Cd and Pb were 94% ± 20%, 92% ± 3%, 86% ± 2%, 91% ± 15%, 92% ± 2%, 87% ± 1% and 76% ± 1% in matrix spike water samples. The standard reference material (GBW10050; Chinese Academy of Geological Sciences) was used to guarantee the determination accuracy for feed and fish samples. The recoveries were 80% ± 20% for Cr, 86% ± 11% for Ni, 90% ± 15% for Cu, 97% ± 8% for Zn, 86% ± 4% for As, 81% ± 10% for Cd and 86% ± 12% for Pb. Recovery tests for the sediment standard reference (GBW07312; Chinese Academy of Geological Sciences) scored 80% ± 10% for Cr, 96% ± 6% for Ni, 102% ± 15% for Cu, 98% ± 6% for Zn, 99% ± 12% for As, 118% ± 14% for Cd and 99% ± 8% for Pb.

The concentrations of heavy metals were presented on a dry weight (dw) as mg·kg⁻¹. For comparison, the wet weight (ww) converted to dw with a conversion rate of 20% (assuming water content of 80%) [14]. Statistical analysis was performed using SPSS 22.0 software and EXCEL 2007. The sampling map was drawn by ArcMap 10.2. The Geoaccumulation index (I_geo), metal pollution index (MPI), bioaccumulation factor (BCF), biota-sediment accumulation factor (BSAF) and the metal concentrations in tissues were performed using SigmaPlot 10.0.

The geoaccumulation index (I_geo) was applied to assess the concentrations of heavy metals in sediments [15] and expressed as Equation (1):

I geo = log 2 C n 1.5 × B n (1)

where C_n is the measured content of metal n in sediments, B_n is the background value of metal n in sediments. The constant 1.5 performs the potential variation about the baseline date caused by lithogenic effects [15]. The geochemical background values in Guanddong Province were 50.5 mg·kg⁻¹ for Cr, 14.4 mg·kg⁻¹ for Ni, 17 mg·kg⁻¹ for Cu, 47.3 mg·kg⁻¹ for Zn, 8.9 mg·kg⁻¹ for As, 0.056 mg·kg⁻¹ for Cd and 36 mg·kg⁻¹ for Pb [16]. The lists of 7 grades of I_geo are shown in TableS2.

The Potential ecological risk index was established by [17] to assess the heavy metals ecological risks in sediment (E_i) and comprehensive heavy metals ecological risks in sediment (RI), which could be calculated with the following Equations (2)-(3):

E i = T i × ( C i × S i ) (2)

RI = ∑ i = 1 7 E i (3)

where C_i and S_i are the measured and background concentrations of metal i, as well as T_i is the toxicity factor of metal i (Cr = 2, Ni = Cu = Pb = 5, Zn = 1, As = 10 and Cd = 30) [18]. The E_i and RI classification are presented in TableS3.

The metal pollution index (MPI) was used to compare the metal concentration in different tissues [19] and Equation (4) was:

MPI = ( C 1 × C 2 × ⋯ × C i ) 1 i (4)

where C_i are the average concentrations of metal i in tissues (dw, mg·kg⁻¹).

Bioaccumulation factor (BCF) is to assess by the ratio between metal concentration in the fish organs and those in the water, while the biota-sediment accumulation (BSAF) is to assess by the ratio between metal concentration in the fish organs and those in the sediment [10]. The calculated Equations (5)-(6) are:

BSAF = C fish / C sediment (5)

BCF = C fish / C water (6)

where C_fish, C_water and C_sediment are the metal concentration in the fish tissue, water and sediment. The grades of BCF values are: less probability of accumulation (BCF < 1000); bioaccumulative (1000 < BCF <5000); highly bioaccumulative (5000 < BCF). As for BASF, if the value > 1, it suggests metal in fish tissue can accumulate from the sediment.

Health risk of metal in fishes was assessed by estimation of daily intake (EDI), target hazard quotient (THQ), hazard index (HI) and carcinogenic risk (CR), which were calculated with the following Equations (7)-(10) [20] [21] [22] [23].

E D I = C N × D C B W (7)

T H Q = C N × D C × E F × E D B W × A T × R f D (8)

H I = ∑ i = 1 7 T H Q (9)

C R = C N × D C × E F × E D × C S F × 10 − 3 B W × A T (10)

where CN is the metal concentration in tissues (ww, mg·kg⁻¹); DC is the daily ingestion rate (51.8 g·person⁻¹·day⁻¹) [24]; BW is the body weight (60 kg); EF is exposure frequency (365 days year⁻¹); ED is the exposure duration (70 years); AT is the average exposure time for non carcinogens (365 days × 70 year) [19]; RfD is the oral reference dose (Cr, Ni, Cu, Zn, As, Cd and Pb are 5, 20, 40, 300, 0.3, 1 and 1.5 µg·kg⁻¹·day⁻¹) [25]; CSF is oral slope factor of carcinogens for As (1.5 mg·kg⁻¹·day⁻¹) and Cd (6.3 mg·kg⁻¹·day⁻¹) [20].

The heavy metal concentrations in the pond water during a complete breeding cycle are given in TableS4. The mean metals concentrations were in the order of Zn > Cu > As > Ni > Pb > Cr > Cd. The concentration of Zn (48.7 ± 21.9 μg·L⁻¹) and Cu (5.72 ± 4.71 μg·L⁻¹) in the water were relatively high, which was consistent with the result from Wen-Rui Tang River [26]. Additionally, the Zn and Cu concentrations in this pond water were higher than the Pearl River [27], which might due to the input of enriched Zn and Cu feeds (TableS5). The concentrations of heavy metals in the pond water were increased in June 2018 than in December 2017, which could be explained that the fish feeds were the major sources of metals to aquaculture [5]. The higher metal concentrations were observed in March 2018, which might be related to the increased feed remains in the water. In comparison to the fisheries standard of China [28], concentrations of metals in the pond water did not exceed the permissible values. The result indicated that the pond water quality was suitable for farming. According to the classification criteria of contamination degree (CD) [29] which was used to assess the heavy metal pollution in water, the calculated results showed that heavy metals in the water showed low pollution (CD < 6).

The metals concentrations analyzed in the fish pond sediments are presented in TableS6. There were no significant differences (P > 0.05) among the heavy metals during the complete breeding cycle. The average metals concentrations in the pond sediments decreased in the order of Zn > Cr > Pb > As > Cu > Ni > Cd. The concentrations of Zn, Cr, As, Cu, Ni and Cd were approximately 2.7, 1.34, 3.4, 2, 2 and 6.25 times higher than that of the background values [16], indicating these metal pollutions came from anthropogenic activities. While the concentration of Pb was comparable to the background value, implying the Pb was mainly from the earth crust.

The concentrations of heavy metals in tissue organs during a complete breeding cycle are depicted in TableS7 and Figure 2. Average values of heavy metals in the muscle are shown in Table1. Zinc had the highest concentration in tissue organs, followed by Cu, Cr, Ni, Pb, As and Cd.

^aChina National Standards (GB 2762-2017) [45]. ^bEuropean Commission [41]. ^cFood and Agriculture Organization, World Health Organization [31]. ^dMeans date are not available.

Zinc is the essential metal to promote metabolism and its shortage can result in some adverse influences, i.e. retarded growth, dysfunction of the immune system and appetite loss [30]. The Zn concentration was the highest compared to target metals analyzed in different tissue. During the breeding cycle, the highest Zn concentration (259 ± 20.5 mg·kg⁻¹, dw) was observed in the skin (May 2018) and the lowest Zn value (18.6 ± 3.79 mg·kg⁻¹, dw) was found in the muscle (January 2018). The accumulation sequence of mean Zn concentrations in all the organs with an order is skin > kidney > gill > heart > liver > bladder > muscle. The skin is the main organ which accumulated higher Zn concentration compared to other target organs. Skin is the organ directly contact with water, Zn had the highest concentration in the water compared to other target metals could cause the higher Zn accumulation in the skin. Interestedly, relatively higher Zn concentrations in all the organs were observed in May 2018, and the concentration of Zn in June 2018 was lower than the background value in organs in December 2017. Since Aril 2018, the intake of feeds increased with the temperature increasing. Subsequently, the rainfall started to increase since May 2018, this might cause the decrease in Zn concentration in organs in June 2018. The mean Zn concentration in the muscle was higher than the acceptable limit (20 mg·kg⁻¹, dw) proposed by World Health Organization [31]. Moreover, the average Zn concentration in this study was higher than those reported studies [19] [32] - [38], but lower than this reported study [10].

Copper is the essential metal to form the hemoglobin and some requisite enzymes, but excess intake of Cu can alter the liver and kidney function [39]. The lowest Cu concentration was observed in the skin (0.79 ± 0.02 mg·kg⁻¹, dw) in June 2018, and the highest concentration of Cu was found in the liver (23.1 ± 1.93 mg·kg⁻¹, dw) in April 2018. The mean concentrations of Cu in organs were in the sequence of: liver > heart > kidney > gill > bladder > muscle > skin. The highest Cu value was found in the liver which was in agreement with the study [40]. The acceptable limit of 8 mg·kg⁻¹ dw for Cu was developed by European Commission [41]. It was worth noting that the mean Cu concentration in the muscle was far higher than those recorded values [19] [32] - [38], but lower than the accorded value [10].

Chromium is regarded as being involved in the metabolism of carbohydrates and lipids [42]. If the intake of Cr is not enough, the risks of cardiovascular diseases and diabetes will be increased [43]. The highest Cr level was detected in the heart (8.01 ± 1.28 mg·kg⁻¹, dw) in January 2018, and the lowest level was detected in the kidney (0.01 ± 0.004 mg·kg⁻¹, dw) in December 2017. The average Cr concentration in organs was arranged in the following order: heart > kidney > gill > bladder > skin > muscle > liver. The concentrations of Cr in organs in March 2018 were higher than in June 2018, probably because the absorption rate of young fish was relatively high but the body burden of Cr was declined with age due to rapid elimination [44]. The average content of Cr in the muscle did not exceed the permissible value established by China National Standards (GB 2762-2017) [45]. Furthermore, the mean Cr content in the muscle was lower than the values reported from the Persian Gulf (11.2 mg·kg⁻¹, dw) [32], Pulau Ketam, Malaysia (0.62 mg·kg⁻¹, dw) [35] and Chengdu, China (6.21 mg·kg⁻¹, dw) [10], but higher than those reported values from the Trabzon, Turkey (0.03 mg·kg⁻¹, dw) [33], River Ganga, India (0.07 mg·kg⁻¹, dw) [34], Mediterranean (0.004 mg·kg⁻¹, dw) [19], Northern Aegean Sea (0.01 mg·kg⁻¹, dw) [36] and Northeast China (0.01 mg·kg⁻¹, dw) [38]. However, the mean Cr content in the muscle was comparable to the value reported from the Chao lake, China (0.18 mg·kg⁻¹, dw) [37].

Nickel can cause adverse health effects to pulmonary, e.g., fibrosis, tumours and lung inflammation [46]. In the present study, the highest amount of Ni was found in the bladder (4.83 ± 1.21 mg·kg⁻¹, dw) in March 2018 and the lowest amount of Ni was observed in the liver (0.01 ± 0.004 mg·kg⁻¹, dw) in June 2018. The order of mean Ni concentration in organs from large to small was as follows: bladder > gill > skin > muscle > kidney > heart > liver. The observed mean value of Ni (0.24 mg·kg⁻¹, dw) in fish muscles was higher than those observed from Pulau Ketam, Malaysia (0.12 mg·kg⁻¹, dw) [35], Northern Aegean Sea (0.14 mg·kg⁻¹, dw) [36] and Northeast China (0.01 mg·kg⁻¹, dw) [38], but lower than those values reported for Ni from the Persian Gulf (14.2 mg·kg⁻¹, dw) [32] and Chengdu, China (1.18 mg·kg⁻¹, dw) [10]. However, the mean Ni concentration was comparable to the value from the Trabzon, Turkey (0.25 mg·kg⁻¹, dw) [33].

Lead is a non-essential element and can cause adverse health effects (neurotoxicity and nephrotoxicity) [47]. The highest Pb content was detected in the gill (1.40 ± 0.14 mg·kg⁻¹, dw) in December 2017, while the lowest content was detected in the liver (0.003 ± 0.001 mg·kg⁻¹, dw) in June 2018. There were significant differences in Pb concentrations among different fish organs (P < 0.05). No Pb concentrations were detected in the muscle. The following decreasing order of mean Pb content in organs was found: gill > kidney > heart ≈ bladder ≈ skin > liver. The average Pb concentration in the muscle did not exceed the maximum levels set by [45], as well as lower than those reported studies [10] [32] - [38].

Arsenic is an ubiquitous in the environment and may be potentially a toxic metal. The highest concentration of As was detected in the kidney (0.13 ± 0.01 mg·kg⁻¹, dw) in June 2018, whereas the lowest concentration was detected in the bladder (0.02 ± 0.01 mg·kg⁻¹, dw) in April 2018. The As concentrations were not detected in all the target organs from December 2017 to January 2018. In addition, the distribution of As in each organ is relatively balanced compared to other target metals. The mean As concentrations in organs could be sequenced as follows: gill ≈ kidney > muscle ≈ skin ≈ heart ≈ liver > bladder. The mean As concentration in the muscle was far lower than the regulated value by [41]. Compared with other previous values, the mean concentration of As was lower than [10] [19] [32] [35] [37], but higher than [38].

Cadmium is deemed as an element with chronic toxicity. The concentrations of Cd were the lowest in organs compared to other target metals. The highest Cd concentration was recorded in the kidney (0.13 ± 0.003 mg·kg⁻¹, dw) in April 2018, and the lowest value was recorded in the heart (0.001 ± 0.0002 mg·kg⁻¹, dw) in June 2018. In December 2017, As concentrations were not detected in any organs. In the case of fish organs, there were not detected in the muscle and skin (except the April 2018) from December 2017 to June 2018. The average Cd concentrations were observed in the order of kidney > liver > gill > heart ≈ bladder > skin. The average Cd concentration was lower than the maximum level set by [45], as well as those reported studies [10] [19] [32] - [38].

TableS8 shows the calculated values of I_geo, E_i and RI for garget metals in the pond sediment in PRD. The average I_geo of heavy metals were ranked in the order of 3 > Cd > 2 > As > 1 > Zn > Ni > Cu > 0 > Cr > Pb. Based on the criteria of sediment I_geo, Cd was moderate to heavily contaminated; As was moderately contaminated; Zn, Ni and Cu were uncontaminated to moderately contaminated; Cr and Pb were uncontaminated. The mean values of E_i for As, Ni, Cu, Pb, Cr and Zn did not exceed 40, implying low risk caused by these metals in the pond sediment. Moreover, Cd contributed high risk in the sediment with the value between 160 and 320. According to the RI values, moderate risk (150 < RI < 300) were presented by these heavy metals in this study area.

The MPI can be visually used to indicate the degree of metal pollution in fish tissues. The calculated MPI results in different tissues during a complete breeding cycle are shown in Figure 3. The highest MPI value was observed in the gill (1.00), while the lowest value found in the liver (0.004). It was worth noting that the highest and lowest MPI values in all the tissues occurred in March 2018 and December 2017, except that the muscle showed the highest MPI value in May 2018 and the skin found the lowest MPI value in January 2018. The average MPI value in tissues was as follows: gill (0.39) > kidney (0.34) > heart (0.15) > bladder (0.13) > skin (0.11) > liver (0.10) > muscle (0.03). The gill and kidney had higher metal pollution compared with other target tissues, which was consistent with the previous study [25]. One way for the metal to enter the fish is through the breath of the gill, and the mucus in the gill can easily absorb metal ions or metal-containing suspensions, etc., and the kidney is the detoxification center of active metabolism of the organism, so the MPI values of the gill and kidney will be higher than other tissues.

The BCF is used to assess the ability of the aquatic organism to accumulate metals from the water. If BCF > 1000, it implies that the organism has a potential to accumulate the metals, and vice versa. The calculated BCF values are presented in Figure 4(a). The highest BCF value was found for Cd (in the kidney), while the lowest value was observed for As (in the bladder). The BCF values of metals in the grasp carp were generally in the order of Zn > Cu > Cd > Cr > Ni > Pb > As. The BCF values of Cu and Zn (essential metals) were higher than Cr, Ni, As, Cd and Pb (non-essential elements), implying that a more important transfer of essential elements than non-essential elements which was inconsistent with previous research [48].

The BSAF is depicted to evaluate the ability of the aquatic organism to accumulate metals from the sediment and the results are shown in Figure 4(b). If BASF > 1, it reflects that the organisms can probably accumulate metals. Among all the metals, only Zn had the value of BASF > 1 in the skin, this result indicated that the skin could accumulate Zn from the sediment. Probably because grass carp mainly live in the lower layer of the water environment enriched with a higher Zn concentration.

The values of EDI for target metals are shown in TableS9. The EDI of all the metals during the breeding cycle was far less than the PTDI, indicating that the exposure risk of heavy metals through the fish consumption within the safe range.

The calculated values of THQ, HI and CR are given in TableS10. The highest value of THQ was found in May 2018, and the THQ of all the target metals was less than 1 during the breeding cycle, which demonstrated that the non-carcinogenic health risks associated with intake of single metal through the consumption of grasp carps. Considering that human exposure to multiple heavy metals may produce synergistic or interactive effects, the individual THQ of the seven heavy metals are added to obtain the HI. The results showed that the maximum HI value (<1) during the breeding cycle was obtained in May 2018, which indicated that there was non-carcinogenic risk associated with intake of multiple metals through the consumption of grasp carps. In general, when the CR value is less than 10⁻⁶, it is considered that the carcinogenic risk can be negligible; when the CR value is higher than 10⁻⁴, the carcinogenic risk is unacceptable; when the value of CR is between 10⁻⁶ and 10⁻⁴, the carcinogenic risk is considered to be acceptable [49]. In TableS10, the Cd CR values were less than 10⁻⁶, implying that the carcinogenic risk caused by Cd could be negligible. The As CR values were less than 10⁻⁶ from December 2017 to January 2018, and the As CR values were between 10⁻⁶ to 10⁻⁴ from March and June of 2018, this results showed that carcinogenic risk could be negligible in December 2017 and January 2018, whereas the carcinogenic risk was considered to be acceptable from the rest sampling month. In summary, the consumption of grass carp during the complete breeding cycle caused non-cancer risk.