Principal Components and Hierarchical Cluster Analyses of Trace Metals and Total Hydrocarbons in Gills, Intestines and Muscles of Clarias gariepinus (Burchell, 1822)

The aim of the study was to comparatively analyze the interrelationships among iron (Fe), manganese (Mn), zinc (Zn), copper (Cu), lead (Pb), cadmium (Cd), chromium (Cr) and total hydrocarbons (THCs) in the gills, intestines and muscles of Clarias gariepinus collected from Osse River, Nigeria, between the periods of April, 2013 to September, 2014. The trace metals in the fish tissues were analyzed using Atomic Absorption Spectrophotometer (AAS, Philips model PU 9100), while total hydrocarbons were analyzed using High Performance Liquid Chromatograph (HPLC,Prominence Dual brand from HGE) equipped with a detector Shimadzu UV-Visible (UV-Vis Prominence SPD 20 A). The concentrations of trace metals and THCs in the tissues were subjected to principal component analysis (PCA), in conjunction with hierarchical cluster analysis (HCA), backed up by correlation analysis (CA). In the most prioritized component among the hierarchies of contaminants, characterized as principal component 1, results of communality extractions and rotated component matrices revealed the order of contaminants was Mn > Cu > Zn > Fe > Cr in the intestines, Cr > Cu > THCs > Mn > Fe in the muscle, while Pb > Cr > Fe > Mn was the order in the gills of the fish. Iron inhibited accumulation of the other trace metals in the gills, where its threshold of essentiality was maximal. Noteworthy is the fact that Mn and Cu were the most active components in the muscle and concurrently of excess concentrations in the tissue, which is the major edible part of fish, and constitutes its main body weight, hence holds its nutritional and economic values. High level of variability which occurred in the toxicant profile across the tissues of C. gariepinus is a function of uptake route, varied organ functions and specificity of tissue permeability of the compared organs. The study demonstrated variability in organ accumulation capacity and toxicant’s competitiveness irrespective of bioavailability. The study provides data useful for future ecotoxicological studies and safety of consumers of the fish.

decrease in white blood cells, red blood cells counts and hemoglobin concentration was observed in Tilapia sparrmanii exposed to chromium concentration of 0.0098 mg/L at pH 7.4-9.0 11,37 . Khangarot et al. 23 detected increases in spleen to body ratio, hemoglobin, white blood cells, packed cell volume (PCV), splenocytes, and antibody production, hence susceptibility to bacterial infection, when Saccobranchus fossilis were exposed to 0.1-3.2 mg/L chromium. Periophthalmus dipes exposed to chromium 5-15 mg/L exhibited decrease in ion-dependent ATPase activity 29 . Vutukuri 38 demonstrated decrease in glycogen content, total lipid content, and protein contents of muscle, gill and liver of Labeo rohita at 39.4 mg Cr L-(96 h-LC50). Carassius auratus exhibited decrease in cell viability and increase in reactive oxygen species when exposed to 250 µM Cr 39 . While Colisa fasciatus showed decrease in liver glycogen content 37 .
Studies have shown that high concentrations of Fe in solution can negatively affect Zn absorption 11 . Manganese-induced alterations in metal absorption and competition for metal transporters and regulatory proteins were also reported by Foster 7 . An inverse relationship between hepatic manganese and iron concentrations has been previously observed 6,19,28 . This is due to the fact that manganese and iron transports are regulated by transferrin receptor and divalent metal transporter.
Fatty fish, such as C. gariepinus are capable of accumulating hydrocarbons in their tissues to concentrations much higher than in the water and sediment of their habitat [40][41][42][43] . This owes to high biomagnification capacity of the fish, owing to its high lipid content. The lipophilic nature of hydrocarbons also plays a role in its accumulation in fish [43][44][45][46] . Furthermore, hydrocarbons readily adsorb on particulates which aid their precipitation to the bottom sediment, thus further exposing demersal fish like catfish and other bottom dwellers to the toxicants [46][47][48] . For example, 14-315 ng/g polycyclic aromatic hydrocarbons (PAHs) was detected in demersal fish [41][42][43][44] , 54-2803 ng/g was detected in mollusk 22 , 52-1600 ng/g in crustacean 49 , and 284-4665 ng/g in algae 48 . All these levels are of ecotoxicological concern. Law and Klungsoyr 22,43 earlier posited that hydrocarbons are notable for high affinity for sediments and lipid tissues of fish, hence may pose carcinogenic risk to humans who consume such fish.
Studies of interactions among toxicants have shown that mere comparison of levels of toxicants with regulatory standards may not provide satisfactory backdrop for safe decisions. Hence, such generalized approaches used in toxicological studies are often concluded with recommendations for further studies to ascertain the actual constituents in environmental media which may be responsible for the contamination 46,47 .
Principal component analysis (PCA) could be employed as a statistical tool to analyze the interactions among a set of toxicants, in order to assign case-specific grades to the toxicants according to their toxicity potentials. Hierarchical cluster analysis (HCA) may further depict the correlation in the interaction of the toxicants and their interrelationships within a given environmental medium. Previously, PCA and HCA have been widely applied in evaluation and classification of toxins 15,[50][51][52][53][54] .
Clarias gariepinus was chosen as the candidate for investigation in this study due to its ecological abundance and economic significance to the populace. Moreover, the fish species has potential for toxicant accumulation due to its pelagic feeding habit and high adipose content which readily biaccumulates lipophilic contaminants 21 . The study was aimed at comparatively analyzing the interrelationships and priorities in the absorption of trace metals (Fe, Mn, Zn, Cu, Pb, Cd and Cr) and total hydrocarbons (THCs) in the gills, intestines and muscles of Clarias gariepinus using multivariate analyses. Multivariate analyses such as PCA and HCA may provide better understanding on the relative activities of the toxicants, beyond the bioaccumulation factors. These findings may add knowledge to the toxicokinetics and toxicodynamics of these toxicants.

Materials and Methods
Description of study area. The study area is a stretch of Osse River (06° 13.432′N, 05° 20.826′E), between Benin River and Ughoton stream which transverses from Nikorowa, through Ekehuan and Gelegele to Iziedema communities. It is located in Osse North-East local Government Area of Edo State; within the tropical rainforest belt of Southern Nigeria (Fig. 1). It is an oligotrophic 55 freshwater river with a thick tropical vegetation cover along its bank. The river is the major source of domestic water for the inhabitants of Gelegele and other communities in the catchment area. The vegetation of the study area comprises mainly of thick trunk trees amidst shrubs and grasses. Most of the plants which dominate the surrounding environment of the study area include; varieties of bamboo trees (Bambusa sp.), water hyacinth (Eichhornia crassipes) and palm trees (Elaeis guineensis).
An oil company named Dubri Oil Company Limited (DOCL) has a flow station close to the Gelegele section of the river; where it has produced about 7.778 million barrels of light crude oil (API 45-50) since its inception in August, 1987 47 . Logging activities were predominant in the forest within the catchment area. The logs were further sliced into marketable sizes at a saw mill at the bank of the river. Such logging activities may be a source of particulates, which may serve as adsorptive substrates to toxicants, thus aggravating fish exposure as they clog up the gills.
Sample collection. Surface water samples were collected at the four (4) locations of distinct anthropogenic activities at the river. The water samples were collected in clean sampling bottles and transported immediately to the laboratory for analysis of the concentrations of the trace metals and THCs. The results were documented on monthly basis. A total of 72 samples of Clarias gariepinus representatives were collected at the four locations along the river course; making a total of 288 fish samples (118.3-218.5 g). The fish were collected by informed artisanal fishermen with the aid of fishing nets, baskets and hooks. They were preserved in a chest chamber at −4 °C, and taken to the laboratory of Animal and Environmental Biology, University of Benin, Nigeria for proper identification using manuals such as Idodo-Umeh 56 and Olaosebikan and Raji 57 . These procedures spanned a period of April, 2013 to September, 2014.
Dissection of fish and excision of organ. The gills, intestines, and muscles of the fish were chosen on the basis of the significant roles the organs play in the animal. The gills were chosen for investigation being the main organ which actively interacts with toxicants in the aqueous phase. It also mediates ion regulation, by the help of the chloride cells. Furthermore, most restaurants do not discard the organ before processing; hence the toxicants accumulated in the gills might pollute the entire food. The intestines are highly adapted for maximal absorption of ingested nutrients alongside toxicants, owing to its large surface area offered by the numerous villi. The muscles of fish constitute its major nutritional and economic values 58 , hence needs to be screened for fitness of consumption.
To excise the organs, the fish was pinned onto a clean dissecting board with underside of the fish facing up. The fish was then dissected ventrally in a triangular pattern. A cross sectional incision traversed between the left and right pectoral fins and tapered at the anus. The intestine was located adjacent to the gonads; identified as the coiled organ which extends from the stomach to the anus. Afterwards, the operculum was removed to expose the gills which were collected in an appropriately labeled sampling bottle.
For convenience, the muscle tissue was obtained from a muscular part of the fish; between the pectoral fin and the dorsal fin. The epithelial layers of the fish were examined using a binocular dissecting microscope (American Optical Corporation, Model 570) and the layers were removed using a sterile surgical blade. The epidermis layer and the basal membrane beneath were carefully peeled off to avoid mechanical damage to the subsequent layers. Afterwards, the dermis was identified by 2 layers which include the stratum laxum and stratum compactum which were carefully removed to expose the subcutis. The subcutis was carefully scrapped off to expose the muscles. The muscles was then excised and kept in a properly labeled sampling bottle for tissue analysis.
Analysis of trace metals. Trace metals in fish tissues. The gills, intestines and muscles of fish were eviscerated using sterile blades. Wet weight (10 g) of each tissue was extracted using 20 mL of HNO 3 : HClO 4 (5:1) at 105 °C for about 24 h 59 . The extract was made up to 25 mL with HNO 3 (70%) and diluted with deionized water. Atomic Absorption Spectrophotometer (AAS, Philips model PU 9100) was used in reading trace metals concentration in mg/kg; wet weights. The meter was calibrated for each metal by dissolving 1 g analar grade metal salt in 1 L of distilled water according to the guidelines of Aderinola et al. 26 .
Trace metals in water samples. Ten (10) mL of water sample was put in a beaker and 2 mL nitric acid was added. The content of the beaker was then heated to evaporation and allowed to cool afterwards. The mixture www.nature.com/scientificreports www.nature.com/scientificreports/ was transferred into volumetric flask. It was then allowed to stand for 24 h, after when it was centrifuged until clear. The sample was screened for suspended solids which were filtered prior to analysis. The trace metals in the mixture were then read using an AAS at 250-350 Volts using the ABS knob.
Analysis of THCs. THCs in fish tissues. Extraction and analysis of THCs in fish tissues were performed according to International Standard Organization (2004). A sample of 2.5 g (wt weight) of fish intestine or reference sample was introduced into a centrifuge tube, we then added 10 mL of acetonitrile/acetone (V/V, 60/40). Whole tissue was homogenized by vortexing 30 sec and 5 min in ultrasonic bath before being centrifuged for 5 min at 4000 rpm. We removed the supernatant and transferred it into a conical tube and the solvent was heated at 35 °C. The extraction was repeated twice with 10 mL of acetonitrile/acetone. The extract was then purified on the cartridges of bonded phase C18 (Waters Sep Pack). Purification was done by adding 2 mL of acetonitrile/acetone are introduced into a conical tube containing the sample. The mixture was vortexed for 15 sec and centrifuged for 30 sec. The upper phase was transferred into a tube and the operation was repeated twice. The supernatants were transferred onto a C18 cartridge previously conditioned with 12 mL of methanol and 12 mL of acetonitrile. The elution was performed with 5 mL of acetonitrile/acetone at atmospheric pressure. The eluent was thereafter concentrated to 50 mg using a rotary evaporator at 35 °C. The purified extract was recovered in 1 mL of hexane. The tube was crimped and stored at −18 °C before analysis.
PAHs were determined using a High Performance Liquid Chromatograph (HPLC) (Prominence Dual brand from HGE) equipped with a detector Shimadzu UV-Visible (UV-Vis Prominence SPD 20 A), a pump (LC 20 AD Prominence Chromatograph), an automatic injector (Autosampler Prominence SIL 20AC) and a column (Column Prominence Owen CTO 20A) type Prevail C18 (15 × 4.6 mm × 5 microns). The cartridge temperature was 40 °C, flow rate was1mL/minute and at wavelength of 254 nm 41 .
THCs in water samples. Analysis of THCs in the water samples was carried out by extracting 1 L of water sample using an AccuBond ODS C18 SPE cartridge (Agilent p/n 188-1356). The cartridge (0.5 g) was conditioned by sequentially rinsing it with four 10-mL aliquots of dichloromethane (DCM), methanol, and two 10-mL aliquots of water (HPLC grade). Afterwards, 1-L of collected water sample was passed through the SPE cartridge at a flow rate of 2.5 mL/min using an automated solid phase extractor. The cartridge was then rinsed thoroughly with 10 mL of HPLC water and then dried with nitrogen for about 10 min. THCs was eluted from the cartridge with 5-mL DCM twice and then added together. The elute was then evaporated with a stream of nitrogen to a volume of 1 mL, and 3.0 mL of acetonitrile was added, which was then concentrated to a final volume of 1.0 mL (USEPA, 1990). THCs concentration was determined using a High Performance Liquid Chromatograph (HPLC).
Quality control and quality assurance measures. Before use, the dissecting instruments and sampling containers were pre-cleaned using 80% ethanol and sterilized 121 °C for 15 min, using a pressure steam sterilizer (Model: SM280E) by Surgifriend Medicals, England. One surgical blade was used per tissue sample, after when it was discarded safely. To avoid hand contamination of the samples, sterile laboratory gloves and nose masks were used throughout the experimental session. All readings were taken in triplicates to minimize error.
Validation of trace metals. Precision and accuracy of the AAS and HPLC methods were validated by repeating each procedure 3 times. Certified reference materials (CRM) by Federal Environmental Protection Agency 60 were employed as a guide. The recovery rates ranged from 93-97%, relative standard deviation (SD) was <6% (determined by Microsoft Excel, 2010), indicating a high data integrity. The reference solutions used to obtain the calibration curves were prepared from a stock solution containing 1000 mg L-1 of each trace metal analyzed. The blanks and reference solutions were also analyzed using the same method that was applied to the samples. The concentrations were expressed in mg/kg. Validation of THCs. The quantification of THCs was validated using reference guidelines of Ake et al. 41 and readings of the HPLC were verified by the certificate (No. TPH-R3-SET, by AccuStandard). To ensure accuracy of equipment we conducted a study of the linearity of the calibration range. The LODs and LOQs were determined and further calculated the coefficient of variation for the tests of repeatability and reproducibility. The calculation of the percentage recovery for testing accuracy was also conducted. The linearity was tested between 0 and 10 μg/kg. The LODs and LOQs were calculated from the blank and analyzed by the HPLC. Limit of detection was calculated with 3 separate trials thus; Compliance statement. Fish used for the experiment were purchased fresh but lifeless from fishermen who provided the information on standard fishing techniques and handling procedures adopted. Effective collaboration with the fishermen was aimed at ascertaining quality control in the study. All experimental procedures were (2020) 10:5180 | https://doi.org/10.1038/s41598-020-62024-9 www.nature.com/scientificreports www.nature.com/scientificreports/ performed in accordance with standard scientific guidelines and regulations. Methods were carried out in accordance with relevant guidelines and regulations.
Statistical analysis. The concentrations of the trace metals and THCs in the tissues of 288 fish samples collected over the period of 18 months were reported as mean values and standard deviation (SD), which were subjected to analysis of variance (ANOVA) to test for the significant differences. They were further subjected to Duncan Multiple Range (DMR) test at 95% confidence interval and probability level of 0.05 using SPSS (version 20). The Kaiser-Meyer-Olkin (KMO) analysis was conducted on the concentrations of toxicants in the tissues to assess the sampling adequacy and suitability for use in the study. The KMO output for the gills at 0.6 was declared as significant variance in data, attributable to the varied accumulation tendencies of the tissues and competition for sites among the toxicants. Furthermore, the sphericity of the data was analyzed using the Bartlett's test which gave a p-value of 0.032 (significant), indicating that a factor/factors may be responsible for the significant variability detected in the data analyzed.
Although power calculation was not conducted for data in this study, the sample size being greater than 100 however guarantees reliable statistical power (Aschard et al. 61 , Jolliffe 23 ). The robustness of the sample size most likely negates the inadequacies associated with natural factors 62 .
Principal component analyses were separately conducted for toxicant concentrations in the different tissues using SPSS (version 20) and verified by Microsoft Excel, 2010 (XLSTAT statistical software). Eigenvalues were obtained for the 288 readings of trace metals and THCs concentrations in the gills, intestines, and muscles of the fish samples through communality extractions. Eigenvalue 1 was adopted as the significant benchmark for the components. Components 1, 2, and 3 were above 1 in the cases of the intestines and muscles (see Fig. 2I,B,C respectively), hence were extracted for further analysis. Components 1 and 2 were extracted for the gills as they were above 1 as indicated by the horizontal line (see Fig. 3I,A). Afterwards, direct oblimin was used in analyzing the significant components obtained from the PCA, following oblique output from KMO and Bartlett's analyses. The components were represented by a coordinate axis which was projected onto two (for 2 components as in the case of the gills) and three (for 3 components as in the cases of the intestines and muscles) dimensional graphs for ease of spatial visualization of the statistical distribution of the data. Exclusion of values was based on the position on the list and the coefficient display format was sorted by size.
For clarification of actual value, PCA was complemented with hierarchical cluster analysis (HCA) which was determined using average linkage hierarchical clustering to evaluate inter-relatedness of toxicant concentrations in the tissues and correlation analysis was employed to further ascertain the interrelationships among the toxicants analyzed. ethical approval. Fish used for this study were fresh but lifeless, however all procedures used conform to standard scientific research guidelines. The source and pre-purchase handling of the specimens were thoroughly investigated for quality assurance purpose.

Results
As presented in Table 1, concentrations of iron and total hydrocarbons in the gills of C. gariepinus exceeded the set standard limit of Federal Environmental Protection Agency (FEPA) 60 . They were also significantly higher than the concentrations in the intestines and muscles (p ≤ 0.05). The concentration of Zn in the intestines exceeded the standard limit of FEPA and was significantly higher than the concentrations in the gills and muscles (p ≤ 0.05). The concentrations of Mn and Cu in the muscle exceeded the regulatory limit and were significantly higher than the concentrations in the gills and intestines at p ≤ 0.05. In the water, only the concentration of manganese was above the established limit. The concentrations of every other toxicant in the tested media were below the FEPA limits 60 .
In the intestines, the principal components 1, 2, and 3 were incorporated into the spatial rotation as they were the components of significant Eigenvalues in the intestine (Fig. 3I,B). In rotated space, Cd exhibited distinct affinity for component 2, while THCs > Fe > Zn also exhibited affinities for component 2 but they rather projected in the opposite direction to Cd (Fig. 3II,B). Conversely, Cu and Mn showed weak affinities for component 3, with some tendencies for the derived component 2. Although Cr and Pb were in proximity to the borderline between components 1 and 3, they however slightly spiked towards components 1 and 3 respectively (Fig. 3II,B).
Projection of Mn towards component 1 was weak, with some tendencies for component 2. The spike direction of Cd was consistent in both the intestine and muscles (towards component 2). Conversely, Pb > THCs > Cr had affinities for the derived component 2, while Fe > Cu exhibited marked affinities for component 3. Importantly, Zn occurred at the borderline between components 1 and 3 (Fig. 3II,C).   www.nature.com/scientificreports www.nature.com/scientificreports/ and THCs in the gills of the fish. Results show that only percentage variance of initial Eigenvalues and extraction sum of squares of components 1(53.470%), followed by 2 (18.812%) were significant, amounting to cumulative percentages of 53.470% and 72.282% respectively ( Table 2). The percentage variance of rotation sums of square were 47.550% and 24.733% for components 1 and 2 respectively, amounting to cumulative percentages of 47.550% and 72.282% respectively.

Analysis of trace metals and tHcs in
After the Principal component analysis (PCA) extraction, we employed direct oblimin with Kaiser Normalization in analyzing the 2 significant components earlier detected. The benchmark significant rotation components matrix (RCM) was set at 0.600. Results showed that the parameters with significant component matrices in component 1 were Pb (0.924) > Cr (0.920) > Fe (0.894) > Mn (0.793) > Zn (0.676) > Cd (0.666), while THCs was the only parameter with a significant RCM in component 2 (Table 3).
Hierarchical cluster analysis (HCA) and correlation analysis. Results of HCA showed Pb, Cd, THCs, Cr, Cu, Mn, and Zn were all clustered together, while Fe was distant from others by 25 units (Fig. 4). This implies that the concentration of iron in the gill of C. gariepinus did not correlate with other toxicants.
A significant negative correlation (−0.7432) occurred between Fe and Zn (Table 4). No significant correlation occurred between iron and other parameters analyzed. High positive correlations occurred between manganese and Cu (0.8045), Pb (0.7726), and Cr (0.8680). A high positive correlation also occurred between Pb and Cr (0.9298). Significant positive correlation also occurred between Zn and Pb (0.6630), Cu and Cr (0.6569). These results correspond with the results of the cluster analysis.    (Table 6).
Hierarchical cluster and correlation analyses. Results of HCA showed Pb, Cd, THCs, and Cr formed a cluster. In close association with this cluster is another cluster formed between Mn and Cu. In this cluster, Pb, Cd, THCs, and Cr formed a sub-cluster and Mn and Cu formed another sub-cluster. Iron and Zn occurred as independent pollutants and in isolation in the intestine dendogram (Fig. 5).
The correlation results (Table 7) of trace metals and THCs in the intestine of C. gariepinus conform to the results of the dendogram. A significant correlation occurred between Fe and Zn (0.63327), and a highly significant correlation occurred between Mn and Cu (0.89189). There were moderate but significant correlations between Cu and Cr (0.65685). Lead also exhibited high correlations with Cr (0.92979), and THCs (0.71123).   www.nature.com/scientificreports www.nature.com/scientificreports/ Cadmium concentration also highly correlated with chromium (0.72126). THCs had significant correlations with contaminants in same dendogram cluster i.e. Pb (0.71123), Cd (0.6542), and Cr (0.6622).
Furthermore, PCA extractions were converged in three iterations using direct oblimin (with Kaiser Normalization) in analyzing the 3 significant components. Results showed that the parameters of significance in  (Table 9).
Hierarchical cluster analysis. In the muscles, the HCA showed that Pb, Cd, THCs, Cu, Cr, and Mn formed a cluster, which was distant from Zn, and further distant from Fe, which was distinctively isolated from others (Fig. 6). The dendogram of contaminants in the muscle tissue of C. gariepinus was buttressed by results of the correlation (Table 10). Significant correlations occurred among contaminants found in the same dendogram cluster i.e. manganese significantly correlated with copper (0.64045), lead (0.67257), cadmium (0.63168), and THCs (0.64704). Copper had significant correlation with Cd (0.6334), and Cr (0.6569). Lead exhibited a significant relationship with cadmium (0.7238) and chromium (0.9298). A highly significant correlation (0.83802) also occurred between cadmium and chromium.   www.nature.com/scientificreports www.nature.com/scientificreports/ Of high importance is the fact that THCs had significant correlations with all other contaminants in the same dendogram cluster. No correlations occurred between Fe and Zn and any other toxicant analyzed in the muscle.

Discussions
High level of variability which occurred in the contaminant profile across the tissues of C. gariepinus is a function of uptake route, varied organ functions and specificity of tissue permeability of the compared organs 63 . However, iron being dissociated from other contaminants in the dendogram of the gills and being associated with same contaminants in dendogram of the intestine showed that organ-specificity dominated contaminant bioavailability among factors that influenced accumulation in the fish [64][65][66]67,68 .
The gill of a fish is an organ which is always in contact with the water medium, hence has the highest level of exposure to the contaminants in the aqueous phase. The concentrations of Fe and THCs in the fish gill were higher than FEPA limits. Iron's dissociation from other metals accompanied low concentrations of Zn and Mn as expected. High concentrations of iron can be linked to its high threshold of essentiality in the hemoglobin, required for respiration in the gills.
The dendogram analysis further showed that the absorption of Fe in the gills varied appreciably compared to that of other contaminants (Pb, Cd, THCs, Cr, Cu, Mn, and Zn) among which high conformity occurred. This was supported by the significant negative correlation which occurred between Fe and most of the contaminants analyzed which indicates that iron inhibited accumulation of the other trace metals in the gills owing to its indispensable threshold of essentiality in the organ. Inverse relationship of Fe concentration with Zn absorption has previously been reported 14,62 . Previous studies have also shown an inverse relationship between hepatic Mn and Fe concentrations 6,19,69 . This is due to the fact that manganese and iron transport is regulated by transferrin receptor and divalent metal transporter. Hence, competition for binding sites may occur. In the gills of the fish, the actual order of contamination was Pb > Cr > Fe > Mn in principal component 1. Occurrence of iron as an active contaminant in principal component 1 of the gills is attributable to its function as an active component of hemoglobin during respiration. Although concentration of Pb in the gill was lower than the FEPA limit, it was significantly higher than the concentrations in other tissues. Furthermore, the concentration was not far below the regulatory limit. This implies that persistent bioaccumulation may be a prognosis of hazardous concentration of Pb. Lead being the leading contaminant in the gill may elicit teratogenic effects in the fish. Lead poisoning may also cause inhibition of the synthesis of hemoglobin, dysfunctions in the kidneys, bone joints, reproductive systems, cardiovascular system, and damage to the nervous system 5,20 . The concentration of THCs in the gills exceeded the FEPA limit, implying that teratogenic, mutagenic and carcinogenic effects may ensue 24,70 . The association of THCs with most of the trace metals in the dendogram indicates that the toxicants might have been released from same source which is attributable to the petroleum activities of the oil company, considering evidences previously provided by Imoobe and Adeyinka 55 , Isibor 31 , and Isibor & Imoobe 47 .
Comparison of accumulation among the tissues showed that the gills accumulated the most toxicants, particularly Fe which had a markedly high BCF of 145.44 (>1) and actual concentration (181.8 ± 42.3 mg/kg) greater than the established limit (100 mg/kg). Iron was also one of the most active toxicants in the tissue. If unregulated in the aquatic habitat, excess iron may elicit tissue injury, characterized by gill tissue necrosis, impairment of the excretory function of the gills and elevation of transaminases activities in the blood plasma of the fish. The gills being an unpalatable part of the fish is discarded before processing in some Nigerian homes. A few restaurants also process the whole fish and discard the unwanted parts, including the gills and internal organs, prior to cooking. Thus, these contaminants may not be transferred into the entire food in such cases. Conversely however, many restaurants process the whole fish without discarding the unwanted and toxicologically implicated organs. Such process may expose the consumers to Fe toxicity which may result in multi-system organ failures, coma, convulsion and ultimately death in extreme cases 51,71 . The outstandingly high concentration of the contaminants detected in the gills is attributable to the organ's physiological and anatomical properties that maximize absorption efficiency. The organ also outstandingly plays physiological roles of iono-regulation, osmoregulation and respiration; hence interacts more with the toxicants in the water than the intestines and muscles.
In the intestine, results of communality extractions and rotated component matrix revealed that the order of contaminants was Mn > Cu > Zn in component 1. The trend is the reverse of the observations of Omar et al. 72 and Silva et al. 62 in Paralichthys brasiliensis. Zinc plays an essential role as an element required for metabolic digestion in the intestine. Fish is a good source of Zn which is a trace element in biological systems. Zinc is one of the essential elements for the metabolism of carbohydrates, proteins, lipids, and nucleic acids. However the concentration of Zn detected in the intestine far exceeded the concentrations in other organs investigated. It also exceeded the set standard of FEPA 60 by a margin of 13.9 mg/kg wt w. Excess Zn in diet may cause immunosuppressive effects, thereby reducing the stimulation responses of lymphocytes and alterations in cholesterol metabolism 27 . Its deficiency in the body can lead to loss of appetite, growth retardation, and immunological problems 40 . Antagonistic relationship between Zn and Fe observed in this study corroborates the observations in previous studies 6,7,19,26,50 .
The results of communality extractions and rotated component matrix showed the order of contaminants in the muscles was Cr > Cu in component 1; and Zn and Pb in component 2. Investigations also showed that Mn (6.5 mg/kg wtw) and Cu (4.6 mg/kg wt w) in the muscles were significantly higher than the concentrations observed in the other organs and the FEPA 60 limits which are 1 mg/kg and 3 mg/kg respectively. The high concentrations of Cu in the muscles are attributable to its high bioconcentration factor. Its high concentrations in the muscle of the fish which is the major edible part 58 may expose the consumers of the fish to risks of liver cirrhosis 72 because essential elements may be toxic when in excess 47 . Similarly, high concentration of Cu (33.7 μg g −1 ) was detected in carapeba (Diaptereus rhombeus) by Silva et al. 62 . Similar concentrations have also been detected in muscles of other freshwater teleosts 33,63,73 and marine species 62 .
Higher accumulation of the toxicants in the muscles than the intestine is at variance with the expected order of gills > intestines > muscles 74 . Higher accumulation was expected in the intestines than the muscle because the intestine is more metabolically active. Moreover, Clarias gariepinus being a bottom feeder is expected to readily accumulate contaminants in its intestine from the repository riverbed. Similar deviant order was earlier observed by Aderinola et al. 26 in Chrysichthys nigrodigitatus, Tilapia zillii and Macrobrachium macrobrachion of Badagry creek in Lagos State, Nigeria, while Ekeanyanwu et al. 75 observed same in Chrysichthys nigrodigitatus and Tilapia nilotica of Okumeshi river in Delta State, Nigeria.
Manganese toxicity effects occur mainly in the respiratory and nervous systems. Symptoms of manganese poisoning are mainly hallucinations, forgetfulness characterized by poor cognitive performance and nerve damage. Manganese can also cause Parkinson's disease, lung embolism, bronchitis, schizophrenia, dullness, weak muscles, headaches and insomnia 20 . Prolonged exposure of the male sex organ to Mn may result in sterility and impotence 73 .
Manganese and copper being the most bioconcentrated toxicants in the muscle of the investigated fish and at the same time Cu being of excess concentrations in the tissue warrant a stringent remediation of the aquatic habitat. This is due to the fact that the muscle is the major edible part of fish and constitutes the major part of its body weight 58 ; hence holds its nutritional and economic values 21 . conclusions High variability in the toxicant profile across the tissues of C. gariepinus is a function of uptake route, varied organ functions and specificity of tissue permeability of the compared organs. The study demonstrated variability in organ accumulation capacity and toxicant's competitiveness irrespective of bioavailability. The study provides data useful for future ecotoxicological studies and safety of consumers of the C. gariepinus in Osse River. A constant check on the river through integrated biomonitoring studies using sentinel species such as the plankton and (2020) 10:5180 | https://doi.org/10.1038/s41598-020-62024-9 www.nature.com/scientificreports www.nature.com/scientificreports/ benthic organisms are promising tools which may contribute effectively to the information on the condition of the aquatic habitat. In the meantime, removal of the gills and intestines of the fish before processing for consumption is recommended.