Mercury (Hg) is among the Lop 10 environmental chemicals of major public health concern (WHO). The Minamata Convention on Mercury (United Nations Environment Program, 2017), commits signing countries to control anthropogenic mercury emissions and reduce human exposure. Human biomonitoring (HBM) programs, are the most straight-forward approaches to get information on the actual exposure levels in the population and assess over time. We report here the results of a HBM study in a nationwide cross-section of Spanish adults (18-65y) as baseline values obtained before the Minamata Convention entered into force. Subsequent follow-ups will show if the Convention has been successful. The study includes 1880 blood samples, 1704 urine samples and 577 hair samples from all Spanish regions collected and analysed under a strictly quality controlled and quality assured protocol. The EU-DEMOCOPHES project demonstrated that fish and seafood are the major sources of mercury exposure and that the Spanish as well as the Portuguese populations have higher levels than other European countries. The data from the present study confirms this pattern at national level and that inhabitants in coastal regions have higher values than from inland regions. The geometric mean (GM) for blood is 635 mu g Hg/l, in urine is 1.11 mu g Hg/l and for hair is 1.91 mu g Hg/g. In an international comparison these values are not exceptional. Spanish concentrations fall into the group of Easter Mediterranean populations. Although information on gender, age, occupational sector, geographical area, sampling period and frequency of fish consumption is reported in the tables, the purpose of this paper has not been to analyse the determinants of exposure in detail but to provide baseline data for future assessments and for regional authorities. (C) 2019 Elsevier B.V. All rights reserved.

The freshwater fish gill forms a barrier against an external hypotonic environment. By culturing rainbow trout gill cells on permeable supports, as intact epithelia, this study investigates barrier property mechanisms. Under symmetrical conditions the apical and basolateral epithelial surfaces contact cell culture media. Replacing apical media with water, to generate asymmetrical conditions (i.e. the situation encountered by the freshwater gill), rapidly increases transepithelial resistance (TER). Proteomic analysis revealed that this is associated with enhanced expression of pre-apolipoprotein AI (pre-apoAI). To test the physiological relevance, gill cells were treated with a dose of 50 μg ml−1 human apolipoprotein (apoAI). This was found to elevate TER in those epithelia which displayed a lower TER prior to apoAI treatment. These results demonstrate the action of apoAI and provide evidence that the rainbow trout gill may be a site of apoAI synthesis. TER does not differentiate between the trans-cellular (via the cell membrane) and para-cellular (via intercellular tight junctions) pathways. However, despite the apoAI-induced changes in TER, para-cellular permeability (measured by polyethylene glycol efflux) remained unaltered suggesting apoAI specifically reduces trans-cellular permeability. This investigation combines proteomics with functional measurements to show how a proteome change may be associated with freshwater gill function.

Gill epithelia from freshwater rainbow trout can be grown in primary culture in Leibovitz's L-15 medium, by seeding freshly isolated gill cells on two successive days, from two different fish, directly onto permeable filter supports (DSI technique). This preparation allows the measurement of transepithelial resistance (TER) and exposure of the apical surface to freshwater, as in vivo. New culture methods were developed and evaluated, using TER as an indicator of epithelial integrity, in an effort to improve the utility of the preparation for proteomic and toxicological research. TER was not related to cell density or protein content in DSI epithelia. To eliminate bovine proteins, the 5% foetal bovine serum (FBS) normally required for epithelial development was replaced with trout plasma. While previously frozen trout plasma proved toxic, freshly collected heparinized plasma, provided by chronically cannulated adult trout, was not. The use of 5% fresh trout plasma supported a TER development curve identical to that with 5% FBS, a useful advance for proteomic research because foreign (bovine) proteins are eliminated. However, 10% plasma reduced TER development, and 100% plasma abolished it. The inhibitory effect on TER of high plasma levels was seen only early in epithelial development, and was exerted from the apical side, likely an effect on tight junction formation. Mature plasma-supplemented preparations mounted a TER rise in response to apical freshwater exposure comparable to that of FBS-supplemented epithelia. Yolk-sac fry extract was inhibitory to TER development, even in the presence of 5% FBS. Transfer of mature epithelia from 18 °C to 4 °C maintained stable TER and extended the useable lifespan by at least ten days, thereby facilitating storage of preparations for toxicity testing. A new method of growing epithelia, involving only a single seeding of cells from a single fish, directly onto filter inserts (SDSI technique), provided mature epithelia with much lower TER, a smaller TER response to apical freshwater, and lower cell density and protein content than DSI epithelia. These SDSI epithelia offer the advantage of multiple preparations grown directly from unique individuals for in vitro toxicity testing.

The induction of 7-ethoxyresorufin O-deethylase (EROD) has been measured in cultured epithelia from rainbow trout gills. Epithelia incubated with water on the apical side and culture media at the basolateral side were exposed to 2,3,7,8-tetrachlorodibenzo-p-dioxin (TCDD), β-naphthoflavone (βNF), benzo[k]fluoranthene (B(k)F), and 3,3′,4,4′,5-pentachlorobiphenyl (PCB#126) from the water. EROD activity was measured as the formation of resorufin from 7-ethoxyresorufin over time in intact epithelia. The EC50 values obtained after 24 h of exposure (mean±S.D.) were for TCDD (n=9) 4.1±3.2×10−11 M, for βNF (n=6) 1.6±3.8×10−9 M, for B(k)F (n=4) 5.4±3.0×10−9 M and for PCB#126 (n=4) 6.15±10.1×10−9 M. When assaying for EROD activity, it was found that the resorufin concentrations differed between the apical and the basolateral compartments, indicating an asymmetrical distribution of the enzymatically formed resorufin molecules. Generally, the resorufin concentration was highest in the basolateral compartment, but there were differences between epithelia obtained from different fish individuals. Of a total of 13 preparations 10 had the highest resorufin concentration in the basolateral compartment, while in three preparations, the resorufin was uniformly distributed or slightly higher in the apical compartment. The reasons for this asymmetrical distribution of substrate metabolites are not known, and the addition of multidrug resistance inhibitors (verapamil and cyclosporin A) did not alter the asymmetrical pattern. The transepithelial electrical resistance (TER) was also measured to diagnose the tightness of the epithelia. The change from culture media to experimental water (containing TCDD, βNF, or DMSO as control) in the apical compartment resulted in a large increase in TER, followed by a decline, measured after 24 h. The cytochrome P450 1A (CYP1A) inducers had no effect on the TER and were judged, therefore, not to affect the tightness of the epithelia.

Cellular approaches for diagnostic effects assessment in ecotoxicology: introductory remarks to an EU-funded project.

Most teleost fish are ammoniotelic, and relatively few are ureotelic, in which the majority of nitrogenous waste is excreted as urea. This study aimed to determine whether the gill ultrastructure of ureotelic fish might have specific, unique characteristics compared with ammoniotelic fish. The gill morphology was studied in three closely related species of the family Batrachoididae: Opsanus beta, the gulf toadfish; Opsanus tau, the oyster toadfish; and Porichthys notatus, the plainfin midshipman, because prior studies have demonstrated that the two former species are ureotelic and excrete urea in unique, short daily pulses, whereas the latter is ammoniotelic. Ultrastructural studies demonstrated significant trafficking of dense-cored vesicles (50-200 nm) between the Golgi apparatus and the apical membrane of epithelial cells surrounding gill filaments and lamellae in these two Opsanus spp. The material constituting the core of these vesicles was intensely stained by lead salt and was unloaded externally when vesicles contacted the apical membrane. Another characteristic of these urea-secreting fish was the presence of numerous large, black-stained lysosomes, which contained cored vesicles, suggesting a second destination for the dense-cored vesicles. As a working hypothesis, the present data suggest that the urea-transporter protein, recently found in toadfish gills, is inserted in the vesicle. Subsequently, it could serve to either sequester cytosolic urea that ultimately is secreted into the water after contact of these vesicles with the pavement cell apical membrane, or it could allow facilitated diffusion of urea across the plasma membrane following insertion into the membrane. As further comparative evidence, the ammoniotelic P. notatus exhibited neither the vesicular trafficking nor the population of lysosomes both found in Opsanus spp.

The influence of a CO2/HCO−3-buffered medium on intracellular pH regulation of gill pavement cells from freshwater rainbow trout was examined in monolayers grown in primary culture on glass coverslips; intracellular pH (pHi) was monitored by continuous spectrofluorometric recording from cells loaded with 2′,7′-bis(2-carboxyethyl)-5(6)-carboxy-fluoroscein. When cells in HEPES-buffered medium at normal pH=7.70 were transferred to normal CO2/HCO−3-buffered medium {PCO2=3.71 mmHg, [HCO−3]= 6.1 mmol l−1, extracellular pH (pHe)=7.70}, they exhibited a brief acidosis but subsequently regulated the same pHi (∼7.41) as in HEPES. Buffer capacity (β) increased by the expected amount (5.5–8.0 slykes) based on intracellular [HCO−3], and was unaffected by most drugs and treatments. However, after transfer to high PCO2=11.15 mmHg, [HCO−3]= 18.2 mmol l−1 at the same pHe=7.70, the final regulated pHi was elevated (∼7.53). The rate of correction of alkalosis caused by washout of this high PCO2, high-HCO−3 medium was unaffected by removal of extracellular Cl−. Removal of extracellular Na+ lowered resting pHi and greatly inhibited the rate of pHi recovery from acidosis. Bafilomycin A1 (3 μmol l−1) had no effect on these responses. However amiloride (0.2 mmol l−1) inhibited recovery from acidosis caused by washout of an ammonia prepulse, but did not affect resting pHi, the latter differing from the response in HEPES where amiloride also lowered resting pHi. Similarly 4-acetamido-4′- isothiocyanatostilbene-2,2′-disulfonic acid, sodium salt (0.1 mmol l−1) did not affect resting pHi but slowed the rate of recovery from acidosis, though to a lesser extent than amiloride. Removal of extracellular Cl− also slowed the rate of recovery but greatly increased β by an unknown mechanism; when this was taken into account, H+ extrusion rate was unaffected. These results are consistent with the presence of Na+-(HCO−3)N co-transport and/or Na+-dependent HCO−3/Cl− exchange, in addition to Na+/H+ exchange, as mechanisms contributing to “housekeeping” pHi regulation in gill cells in CO2/HCO−3 media, whereas only Na+/H+ exchange is seen in HEPES. Both Na+-independent Cl−/HCO−3 exchange and V-type H+-ATPase mechanisms appear to be absent from these cells cultured in isotonic media.

A new double-seeded insert (DSI) technique is described for culture of branchial epithelial preparations from freshwater rainbow trout on filter supports. DSI epithelia contain both pavement cells and mitochondria-rich (MR) cells (15.7+/-2.5 % of total cell numbers). MR cells occur singly or in clusters, are voluminous, open apically to the 'external environment' and exhibit ultrastructural characteristics similar to those found in the 'chloride cells' of freshwater fish gills. After 6-9 days in culture with Leibovitz's L-15 medium on both surfaces (symmetrical conditions), transepithelial resistance (TER) stabilized at values as high as 34 k capomega cm(2), indicative of electrically 'tight' epithelia. The density of MR cells, the surface area of their clusters and transepithelial potential (TEP; up to +8 mV basolateral positive, mean +1.9+/-0.2 mV) were all positively correlated with TER. In contrast, preparations cultured using an earlier single-seeded insert (SSI) technique contained only pavement cells and exhibited a negligible TEP under symmetrical conditions. Na(+)/K(+)-ATPase activities of DSI preparations were comparable with those in gill filaments, but did not differ from those of SSI epithelia. Replacement of the apical medium with fresh water to mimic the in vivo situation (asymmetrical conditions) induced a negative TEP (-6 to -15 mV) and increased permeability to the paracellular marker PEG-4000. Under symmetrical conditions, unidirectional Na(+) and Cl(-) fluxes were in balance, and there was no active transport by the Ussing flux ratio criterion. Under asymmetrical conditions, there were large effluxes, small influxes and evidence for active Cl(-) uptake and Na(+) extrusion. Unidirectional Ca(2+) fluxes were only 0.5-1.0 % of Na(+) and Cl(-) fluxes; active net Ca(2+) uptake occurred under symmetrical conditions and active net extrusion under asymmetrical conditions. Thus, DSI epithelia exhibit some of the features of the intact gill, but improvements in culture conditions are needed before the MR cells will function as true freshwater 'chloride cells'.

The biotransformation of xenobiotics and steroids was investigated in cultured respiratory epithelial cells from rainbow trout (Oncorhynchus mykiss) gills. As a first approach, ethoxyresorufin-O-deethylase (EROD), chosen as a marker of CYP1A activity, was measured in monolayers of adherent cells. The induction of this enzyme was studied in cells exposed to beta-naphthoflavone (BNF) or 2,3,7,8-tetrachlorodibenzo-p-dioxin (TCDD) in concentrations ranging from 10(-6) to 10(-12) M. After 24 h, TCDD showed a maximal induction at a concentration of 10(-9) M while BNF showed a maximal induction at a concentration of 10(-7) M. Concurrently, a variety of substrates involved in cytochrome P450-dependent metabolism as well as phase II reactions, namely ethoxycoumarin, aniline and testosterone were incubated with cultured gill cells for 2 or 8 h and with freshly isolated hepatocytes for comparison. Our results revealed a significant cytochrome P450-dependent activity in gill cells with ethoxycoumarin and aniline, but no hydroxylation was observed with testosterone as substrate. No trace of sulfate conjugate was detected. With 2.5 µM aniline as substrate, 2-hydroxyaniline accounted for 32.1% of the radioactivity after 2 h incubation whereas acetanilide amounted to 6.4%. Significant differences were found between gill cells and isolated hepatocytes in the capacity of these systems to conduct oxidative and conjugating metabolic pathways. Qualitatively, the main difference was observed for testosterone which is hydroxylated in position 6beta and 16beta and conjugated to glucuronic acid in liver cells, whereas reductive biotransformation giving rise to dihydrotestosterone and androstanediol and traces of androstenedione were observed in gill cells. Quantitatively, the biotransformation activity in gill epithelial cells, expressed as pmol/h per mg protein, was between 1.5 and 14% of the activity level observed in isolated hepatocytes, depending on the substrate.

Induction of the cytochrome P4501A (CYP1A)-mediated enzyme activity 7-ethoxyresorufin O-deethylase (EROD) was measured in cultured respiratory epithelial cells from rainbow trout (Oncorhynchus mykiss) gills. Monolayers of adherent cells were exposed to the inducers β-naphthoflavone (β-NF), 2,3,7,8-tetrachlorodibenzo-p-dioxin (TCDD), 3,3′,4,4′,5-pentachlorobiphenyl (PCB#126) and benzo[k]fluoranthene (B[k]F). EROD activity was either measured directly in the adherent cells after exposure to inducer, or the exposed adherent cells were dislodged by trypsination and EROD activity was measured in suspended cells. A time course study of the induction caused by β-NF showed that the EROD activity remained at a similar level from 24 to 72 h after addition of the compound. A second time course study was performed using all four substances. For β-NF, B[k]F and TCDD the induction was similar throughout the time period and for PCB#126 maximal induction was reached at 24 h, followed by a decrease. Concentration–response relationships for EROD induction were established for the two different methods by exposing the cells to 0.036, 0.36 and 3.6×10−6 M β-NF. When measured in suspended cells, EROD activity increased at the two lowest concentrations. At the highest concentration no further increase was detected. The adherent cells showed a different concentration–response to the inducer, with an increase at the low concentration, a maximum at 0.36×10−6 M, and a decrease at the highest concentration. The maximal EROD activity was about two times higher in adherent cells than in suspended cells. All further measurements were therefore performed on adherent intact cells. The potencies and efficacies of the different compounds to induce EROD were compared in a set of experiments using concentrations ranging from 10−11 to 10−6 M. In all cases, EROD activity increased with increasing concentrations up to a maximal level, and thereafter the activity decreased. TCDD caused maximal EROD activity at a concentration of 10−9 M, and all other inducers caused maximal EROD activity at a concentration of 10−7 M. The maximal EROD activities obtained with B[k]F, β-NF and PCB#126 were about 80, 80 and 40%, respectively, of the maximal activity obtained with TCDD. The relative potencies compared with TCDD were 0.008 for B[k]F, 0.0035 for PCB#126 and 0.002 for β-NF, based on EC50 values determined from the concentration–response curves.

In the search for suitable cell models that could be used in a test battery for toxicity screening purposes, we have focused on gill epithelial cells from rainbow trout. Gill epithelial cells are attractive models, since the gill is the primary target and uptake site for many toxicants in the water (Evans, D.H., 1987. Environ. Health Perspect. 71, 54–58). The aim of this study was to compare the sensitivity of gill epithelial cells in different test systems: suspensions, primary cultures and epithelia on filters in order to find out which system would be most suitable in a test battery for toxicity screening. The toxicity of the first 30 reference chemicals from the MEIC (multicenter evaluation of in vitro cytotoxicity) project (Bondesson, I., Ekwall, B., Hellberg, S. et al., 1989. Cell Biol. Toxicol. 5, 331–347) to gill epithelial cells in primary culture was tested using the fluorescent viability probe calcein-AM. Ten of the chemicals were also tested on cells suspended after 6–8 days in culture and five of the chemicals were tested on cells cultured on permeable filters. Decrease in transepithelial resistance was used as the endpoint for cells cultured on filters. The results were compared with toxicity data previously obtained by using freshly isolated gill epithelial cells in suspension. There was no significant difference in sensitivity of gill epithelial cells in suspension, primary cultures or resuspended cells. This observation strongly supports the use of cells in suspension in a test battery for toxicity screening purposes. Cultured epithelia on filters proved to be less suitable for screening purposes due to the considerable variation in resistance between epithelia even from the same fish.

1. Primary cultures of sea bass (Dicentrarchus labrax) gill cells grown on permeable membranes form a highly differentiated tight epithelium composed of respiratory-like cells. This preparation was also found to provide a functional model for investigating the hormonal regulation of Cl- secretion. 2. In control conditions, i.e. in the absence of hormones or other stimuli, the cultured epithelium showed a short-circuit current (Isc) of 8.8 +/- 0.4 microA cm-2, a transepithelial potential (Vt) of 28.6 +/- 0.6 mV (serosal side positive), and a transepithelial resistance (Rt) of 5026 +/- 127 Omega cm2. Addition of 50 nM PGE2 caused a stimulation of Isc, Vt and transepithelial conductance, Gt. The increase in Isc was probably due to the elevation in Cl- secretion, since it could be correlated with the stimulation of serosal to mucosal 36Cl- flux. Application of the neurohypophyseal peptide arginine vasotocin (AVT; 50 nM) or the beta-adrenergic agonist isoproterenol (isoprenaline; 0. 5 microM) evoked a stimulation in Cl- secretion, as was shown by the increases in Isc and Gt. The excitatory effect of isoproterenol followed by the inhibitory action of propranolol, a beta-adrenergic antagonist, suggested the presence of beta-adrenergic receptors. Noradrenaline (0.1 microM) elicited a reduction in Isc, Vt and Gt, which was counterbalanced by the addition of phentolamine, an alpha-adrenergic antagonist. This suggested an activation of alpha-adrenergic receptors. 3. This study provides evidence for hormonal control of the Cl- secretion in sea bass gill respiratory cells in culture, involving AVT, prostaglandin (PGE2), and beta- and alpha-adrenergic receptors.

The permeability of toadfish gills and skin to urea and water has been measured in order to investigate the mechanisms behind the pulsatile excretion of urea previously described in this species. A perfused gill preparation was used in the all studies and isolated pieces of skin mounted in an Ussing chamber in the skin studies. Simultaneously, urea and water permeability was measured in vivo in free swimming fish. In vivo the nonpulsing urea permeability was exceptionally low compared to other teleosts, while the tritiated water permeability was similar to that of other teleosts. The urea permeability increased 30-fold during a pulse while water permeability stayed unaffected. Compared to in vivo, tritiated water permeability was approximately 50% lower in the gills and the skin when measured directly in the isolated preparations. The urea permeability was almost identical between the three preparations. Four out of 20 perfused gill preparation showed a spontaneous urea pulse during perfusion. Several treatments were tested to elicit the pulse artificially but without success. Hormones and drugs tested were: arginine-vasotocin (AVT), 10(-10) M; adrenaline, 10(-7) M; isoprenaline, 10(-5) M; acetylcholine, 10(-7) and 10(-6) M; serotonin, 10(-7) and 10(-6) M; adenosine, 10(-6) M; cortisol, 10(-7) M; and combinations of AVT, adrenaline, and cortisol. Adrenaline and isoprenaline increased tritiated water permeability without affecting urea permeability. Gradually increasing the ammonia levels in the perfusate from 0.1 mM to 1.6 mM caused a slight increase in water permeability but a marked and progressive increase in urea permeability. No indications of an ammonia trapping mechanism in the gills were found. There was no effect of AVT (10(-10) mol l(-1)) in the urea permeability of the skin preparation while cortisol (10(-7) M) led to a modest increase in urea permeability. Based on a comparison between the in vivo and in vitro preparations used here, we conclude that the urea pulse in a urea-pulsing toadfish occurs through the gills and not the skin. We still do not know which internal mechanism or signal triggers the urea pulse in the toadfish. J. Exp. Zool. 283:1-12, 1999. (C) 1999 Wiley-Liss, Inc.

Branchial epithelia of freshwater rainbow trout were cultured on permeable supports, polyethylene terephthalate membranes (“filter inserts”), starting from dispersed gill epithelial cells in primary culture. Leibowitz L-15 media plus foetal bovine serum and glutamine, with an ionic composition similar to trout extracellular fluid, was used. After 6 days of growth on the filter insert with L-15 present on both apical and basolateral surfaces, the cultured preparations exhibited stable transepithelial resistances (generally 1000–5000 Ω cm2) typical of an electrically tight epithelium. Under these symmetrical conditions, transepithelial potential was zero, and unidirectional fluxes of Na+ and Cl− across the epithelium and permeability to the paracellular marker polyethylene glycol-4000 (PEG) were equal in both directions. Na+ and Cl− fluxes were similar to one another and linearly related to conductance (inversely related to resistance) in a manner indicative of fully conductive passive transport. Upon exposure to apical fresh water, transepithelial resistance increased greatly and a basolateral-negative transepithelial potential developed. At the same time, however, PEG permeability and unidirectional effluxes of Na+ and Cl− increased. Thus, total conductance fell, and ionic fluxes and paracellular permeability per unit conductance all increased greatly, consistent with a scenario whereby transcellular conductance decreases but paracellular permeability increases upon dilution of the apical medium. In apical fresh water, there was a net loss of ions from the basolateral to apical surfaces as effluxes greatly exceeded influxes. However, application of the Ussing flux ratio criterion, in two separate series involving different methods for measuring unidirectional fluxes, revealed active influx of Cl− against the electrochemical gradient but passive movement of Na+. The finding is surprising because the cultured epithelium appears to consist entirely of pavement-type cells.