Development of a PBPK Model for Silver Accumulation in Chub Infected with Acanthocephalan Parasites.

Simultaneous presence of metals and parasites in fish might lead to potential risks to human health. Parasites might influence metal accumulation and disturb detoxification in fish, thereby affecting biomarkers of fish responses as well as metal biomagnification in humans. It is, therefore, of importance to take into account parasite infection when investigating metal accumulation in fish. However, mechanisms of metal accumulation and distribution in fish-parasite systems are not integrated into current approaches. The present study proposes a new physiologically based pharmacokinetic model for mechanistic simulation of metal partitioning between intestinal parasites and their hosts. As a particular case, Ag accumulation in the system of chub Squalius cephalus and the acanthocephalan Pomphorhynchus tereticollis was investigated. As a novelty, fish cardiac output and organ-specific blood flow distribution were incorporated in our model. This approach distinguishes the current model from the ones developed previously. It also facilitates model extrapolation and application to varying conditions. In general, the model explained Ag accumulation in the system well, especially in chub gill, storage (including skin, muscle, and carcass), and liver. The highest concentration of Ag was found in the liver. The accumulation of Ag in the storage, liver, and gill compartments followed a similar pattern, i.e., increasing during the exposure and decreasing during the depuration. The model also generated this observed trend. However, the model had a weaker performance for simulating Ag accumulation in the intestine and the kidney. Silver accumulation in these organs was less evident with considerable variations.

Relationships between absorption efficiency of elements in mammals and chemical properties.

Oral absorption efficiency is an important factor to consider in human risk assessment and varies widely between elements. Linking absorption efficiency to chemical properties facilitates the understanding of underlying processes and enables extrapolation across elements. In our study, oral absorption efficiency in humans was predicted for a number of elements based on their ionization energy and electronegativity. Data on oral absorption efficiency in humans were retrieved via a literature survey. A model was developed based on the assumption that ionic species readily react with biotic ligands. Accordingly, ionization energy was presumed to represent the reactivity and absorption of atoms in the gastrointestinal tract. The coefficients of the model were parameterized by fitting the quantitative relationship between absorption efficiency and ionization energy to data collected from well-standardized studies. Generally, absorption efficiency was strongly related to ionization energy, explaining 94% of the variability in absorption efficiency between elements reported by the International Commission on Radiological Protection (ICRP). In addition, the absorption efficiencies predicted based on ionization energy were within a factor of two of those given by the ICRP (ME = -0.05; RMSE = 0.31). However, the model is not applicable to alkaline metals and molybdenum because of the uniquely high solubility of their compounds or the flexible electron configuration of these elements. Approximately 56% of the variability in absorption efficiency between elements could be explained by electronegativity. These strong relationships between absorption efficiency and ionization energy and, to a lesser extent, electronegativity indicate potential for extrapolation across elements using atomic properties.

Modeling toxicity of mixtures of perfluorooctanoic acid and triazoles (triadimefon and paclobutrazol) to the benthic cladoceran Chydorus sphaericus.

Toxicity of perfluorooctanoic acid (PFOA) and triazoles, that is, paclobutrazol (PBZ) and triadimefon (TDF), singly and in binary mixtures was assessed on the benthic cladoceran Chydorus sphaericus. Following single exposure, toxicity of PBZ was not significantly different from that of TDF as shown by the overlapping 95% confidence intervals of their median effective concentrations: EC50(PBZ) = 0.07-0.09 mM; EC50(TDF) = 0.066-0.078 mM. PFOA was far less toxic than these triazoles: EC50(PFOA) = 0.20-0.24 mM. Assuming no interactions between PFOA and the triazoles, the toxicity of PFOA in mixtures with the triazoles estimated by the concentration addition (CA) and independent action (IA) models was significantly lower than its toxicity in single exposure. This contrasted with the results for PBZ and TDF. Based on the assumption that PFOA may interact with the triazoles, different predictions were obtained by the two models. Significant alleviation of PBZ and TDF on the toxicity of PFOA was found by the extended CA model only. According to the extended IA model, PFOA had nonsignificant effects on the toxicity of PBZ while significantly increasing the toxicity of TDF. This difference is possibly related to different modes of actions for various functional groups in PBZ and TDF.

The Asymmetric Response Concept explains ecological consequences of multiple stressor exposure and release.

Our capacity to predict trajectories of ecosystem degradation and recovery is limited, especially when impairments are caused by multiple stressors. Recovery may be fast or slow and either complete or partial, sometimes result in novel ecosystem states or even fail completely. Here, we introduce the Asymmetric Response Concept (ARC) that provides a basis for exploring and predicting the pace and magnitude of ecological responses to, and release from, multiple stressors. The ARC holds that three key mechanisms govern population, community and ecosystem trajectories. Stress tolerance is the main mechanism determining responses to increasing stressor intensity, whereas dispersal and biotic interactions predominantly govern responses to the release from stressors. The shifting importance of these mechanisms creates asymmetries between the ecological trajectories that follow increasing and decreasing stressor intensities. This recognition helps to understand multiple stressor impacts and to predict which measures will restore communities that are resistant to restoration.

What contributes to the metal-specific partitioning in the chub-acanthocephalan system?

Physiologically based pharmacokinetic (PBPK) models have been applied to simulate the absorption, distribution, metabolism, and elimination of various toxicants in fish. This approach allows for considering metal accumulation in intestinal parasites. Unlike "semi" physiologically-based models developed for metals, metal accumulation in fish was characterised based on metal-specific parameters (the fraction in blood plasma and the tissue-blood partition coefficient) and physiological characteristics of the fish (the blood flow and the tissue weight) in our PBPK model. In the model, intestinal parasites were considered a sink of metals from the host intestine. The model was calibrated with data for the system of the chub Squalius cephalus and the acanthocephalan Pomphorhynchus tereticolliis. Metal concentrations in this fish-parasite system were monitored in Ag and Co treatments in duplicate during a 48-day exposure phase (Ag and Co were added to tap water at concentrations of 1 and 2 µg/L, respectively) and a 51-day depuration phase. Their concentrations in the gills increased during the exposure phase and decreased in the depuration phase. A similar pattern was observed for Ag concentrations in other chub organs, while a relatively stable pattern for Co indicates regulations in the accumulation of essential metals by chubs. The metals were taken up by the acanthocephalans at similar rate constants. These results indicate that metal availability to parasites, which is determined by the internal distribution and fate, is critical to metal accumulation in the acanthocephalans. The high concentration of Ag in the liver as well as the high rate of Ag excretion from the liver to the intestine might contribute to higher concentrations of metals in the bile complexes in the intestine, which are available to the parasites, but not to the reabsorption by the host intestine. The opposite pattern might explain the lower availability of Co to the acanthocephalans.

Delineation of the exposure-response causality chain of chronic copper toxicity to the zebra mussel, Dreissena polymorpha, with a TK-TD model based on concepts of biotic ligand model and subcellular metal partitioning model.

A toxicokinetic-toxicodynamic model was constructed to delineate the exposure-response causality. The model could be used: to predict metal accumulation considering the influence of water chemistry and biotic ligand characteristics; to simulate the dynamics of subcellular partitioning considering metabolism, detoxification, and elimination; and to predict chronic toxicity as represented by biomarker responses from the concentration of metals in the fraction of potentially toxic metal. The model was calibrated with data generated from an experiment in which the Zebra mussel Dreissena polymorpha was exposed to Cu at nominal concentrations of 25 and 50 µg/L and with varied Na+ concentrations in water up to 4.0 mmol/L for 24 days. Data used in the calibration included physicochemical conditions of the exposure environment, Cu concentrations in subcellular fractions, and oxidative stress-induced responses, i.e. glutathione-S-transferase activity and lipid peroxidation. The model explained the dynamics of subcellular Cu partitioning and the effect mechanism reasonably well. With a low affinity constant for Na + binding to Cu2+ uptake sites, Na + had limited influence on Cu2+ uptake at low Na+ concentrations in water. Copper was taken up into the metabolically available pool (MAP) at a largely higher rate than into the cellular debris. Similar Cu concentrations were found in these two fractions at low exposure levels, which could be attributed to sequestration pathways (metabolism, detoxification, and elimination) in the MAP. However, such sequestration was inefficient as shown by similar Cu concentrations in detoxified fractions with increasing exposure level accompanied by the increasing Cu concentration in the MAP.

Modelling copper toxicokinetics in the zebra mussel, Dreissena polymorpha, under chronic exposures at various pH and sodium concentrations.

The stenohaline zebra mussel, Dreissena polymorpha, is uniquely sensitive to the ionic composition of its aquatic environment. Waterborne copper (Cu) uptake and accumulation in zebra mussels were examined at various conditions in an environmentally relevant range in freshwater, i.e. Cu exposure levels (nominal concentrations of 25 and 50 µg/L), pH (5.8-8.3), and sodium (Na+) concentrations (up to 4.0 mM). Copper accumulation was simulated by a kinetic model covering two compartments, the gills and the remaining tissues. The Cu uptake rate constant decreased with decreasing pH from 8.3 down to 6.5, indicating interactions between H+ and Cu at uptake sites. The kinetic simulation showed dose-dependent effects of Na+ on Cu uptake. At 25 µg/L Cu, addition of Na+ at 0.5 mM significantly inhibited the Cu uptake rate, while no significant differences were found in the uptake rate upon further addition of Na+ up to a concentration of 4.0 mM. At 50 µg/L Cu, the Cu uptake rate was not influenced by Na+ addition. Calibration results exhibited dose-dependent elimination rates with more profound elimination with increasing exposure levels. With kinetic parameters calibrated at environmentally relevant conditions, in terms of pH and Na+ concentrations, the model performed well in predicting Cu accumulation based on independent data sets. Estimates of the Cu concentration in mussels were within a factor of 2 of the measurements. This demonstrates potential application of kinetic models that are calibrated in environmentally relevant freshwater conditions.

Development of a toxicokinetic-toxicodynamic model simulating chronic copper toxicity to the Zebra mussel based on subcellular fractionation.

A toxicokinetic-toxicodynamic model based on subcellular metal partitioning is presented for simulating chronic toxicity of copper (Cu) from the estimated concentration in the fraction of potentially toxic metal (PTM). As such, the model allows for considering the significance of different pathways of metal sequestration in predicting metal toxicity. In the metabolically available pool (MAP), excess metals above the metabolic requirements and the detoxification and elimination capacity form the PTM fraction. The reversibly and irreversibly detoxified fractions were distinguished in the biologically detoxified compartment, while responses of organisms were related to Cu accumulation in the PTM fraction. The model was calibrated using the data on Cu concentrations in subcellular fractions and physiological responses measured by the glutathione S-transferase activity and the lipid peroxidation level during 24-day exposure of the Zebra mussel to Cu at concentrations of 25 and 50 µg/L and varying Na+ concentrations up to 4.0 mmol/L. The model was capable of explaining dynamics in the subcellular Cu partitioning, e.g. the trade-off between elimination and detoxification as well as the dependence of net accumulation, elimination, detoxification, and metabolism on the exposure level. Increases in the net accumulation rate in the MAP contributed to increased concentrations of Cu in this fraction. Moreover, these results are indicative of ineffective detoxification at high exposure levels and spill-over effects of detoxification.

Modelling chronic toxicokinetics and toxicodynamics of copper in mussels considering ionoregulatory homeostasis and oxidative stress.

Chronic toxicity of copper (Cu) at sublethal levels is associated with ionoregulatory disturbance and oxidative stress. These factors were considered in a toxicokinetic-toxicodynamic model in the present study. The ionoregulatory disturbance was evaluated by the activity of the Na+/K+-ATPase enzyme (NKA), while oxidative stress was presented by lipid peroxidation (LPO) and glutathione-S-transferase (GST) activity. NKA activity was related to the binding of Cu2+ and Na + to NKA. LPO and GST activity were linked with the simulated concentration of unbound Cu. The model was calibrated using previously reported data and empirical data generated when zebra mussels were exposed to Cu. The model clearly demonstrated that Cu might inhibit NKA activity by reducing the number of functional pump sites and the limited Cu-bound NKA turnover rate. An ordinary differential equation was used to describe the relationship between the simulated concentration of unbound Cu and LPO/GST activity. Although this method could not explain the fluctuations in these biomarkers during the experiment, the measurements were within the confidence interval of estimations. Model simulation consistently shows non-significant differences in LPO and GST activity at two exposure levels, similar to the empirical observation.

How does the metallothionein induction in bivalves meet the criteria for biomarkers of metal exposure?

Metallothionein (MT) concentrations in the whole soft tissue or in a particular tissue of bivalves have widely been used in ecotoxicological studies and biomonitoring programmes. This approach is based on the reported results on the enhancement of MT induction in bivalves in response to metal exposure. The validity of using MT induction as a biomarker is briefly assessed in the present study. The sensitivity of MT induction in these organisms is questionable due to the high basal MT level as well as the high natural variability related to the effects of a number of biotic and abiotic factors, which are not well described yet. Moreover, the relationship between exposure to metals, the toxic effects of that exposure, and the appearance of MT in soft tissue, is not well characterized. A variety of factors may influence the appearance and distribution of MT: 1) the uneven distribution of metals in particular portions of the soft tissue and in particular subcellular compartments; 2) pre-exposure to metals, perhaps at non-toxic levels; 3) metal-metal competition and metal-protein interactions; and 4) tissue-specific induction, functions, and isoforms of MT. Therefore, attention is required when using MT induction in bivalves for assessment of metal exposure or consequent toxic effects. The MT concentration can be a reliable indicator only when it is considered in relation with metal uptake kinetics and subcellular partitioning while specifying the isoform of MT synthesised and considering various confounding factors. The kinetic turnover of MT may provide useful information on metal exposure and biological effects since it covers both the synthesis and breakdown of MT as well as the chemical species of metals accumulated and MT.

Development and Validation of a Biodynamic Model for Mechanistically Predicting Metal Accumulation in Fish-Parasite Systems.

Because of different reported effects of parasitism on the accumulation of metals in fish, it is important to consider parasites while interpreting bioaccumulation data from biomonitoring programmes. Accordingly, the first step is to take parasitism into consideration when simulating metal bioaccumulation in the fish host under laboratory conditions. In the present study, the accumulation of metals in fish-parasite systems was simulated by a one-compartment toxicokinetic model and compared to uninfected conspecifics. As such, metal accumulation in fish was assumed to result from a balance of different uptake and loss processes depending on the infection status. The uptake by parasites was considered an efflux from the fish host, similar to elimination. Physiological rate constants for the uninfected fish were parameterised based on the covalent index and the species weight while the parameterisation for the infected fish was carried out based on the reported effects of parasites on the uptake kinetics of the fish host. The model was then validated for the system of the chub Squalius cephalus and the acanthocephalan Pomphorhynchus tereticollis following 36-day exposure to waterborne Pb. The dissolved concentration of Pb in the exposure tank water fluctuated during the exposure, ranging from 40 to 120 µg/L. Generally, the present study shows that the one-compartment model can be an effective method for simulating the accumulation of metals in fish, taking into account effects of parasitism. In particular, the predicted concentrations of Cu, Fe, Zn, and Pb in the uninfected chub as well as in the infected chub and the acanthocephalans were within one order of magnitude of the measurements. The variation in the absorption efficiency and the elimination rate constant of the uninfected chub resulted in variations of about one order of magnitude in the predicted concentrations of Pb. Inclusion of further assumptions for simulating metal accumulation in the infected chub led to variations of around two orders of magnitude in the predictions. Therefore, further research is required to reduce uncertainty while characterising and parameterising the model for infected fish.

Modelling metal accumulation using humic acid as a surrogate for plant roots.

Metal accumulation in roots was modelled with WHAM VII using humic acid (HA) as a surrogate for root surface. Metal accumulation was simulated as a function of computed metal binding to HA, with a correction term (E(HA)) to account for the differences in binding site density between HA and root surface. The approach was able to model metal accumulation in roots to within one order of magnitude for 95% of the data points. Total concentrations of Mn in roots of Vigna unguiculata, total concentrations of Ni, Zn, Cu and Cd in roots of Pisum sativum, as well as internalized concentrations of Cd, Ni, Pb and Zn in roots of Lolium perenne, were significantly correlated to the computed metal binding to HA. The method was less successful at modelling metal accumulation at low concentrations and in soil experiments. Measured concentrations of Cu internalized in L. perenne roots were not related to Cu binding to HA modelled and deviated from the predictions by over one order of magnitude. The results indicate that metal uptake by roots may under certain conditions be influenced by conditional physiological processes that cannot simulated by geochemical equilibrium. Processes occurring in chronic exposure of plants grown in soil to metals at low concentrations complicate the relationship between computed metal binding to HA and measured metal accumulation in roots.

Uncertainties associated with lacking data for predictions of solid-solution partitioning of metals in soil.

Soil properties, i.e., pH and contents of soil organic matter (SOM), dissolved organic carbon (DOC), clay, oxides, and reactive metals, are required inputs to both mechanistic and empirical modeling in assessing metal solid-solution partitioning. Several of these properties are rarely measured in site-specific risk assessment. We compared the uncertainties induced by lacking data on these soil properties in estimating metal soil solution concentrations. The predictions by the Orchestra framework were more sensitive to lacking soil property data than the predictions by the transfer functions. The deviations between soil solution concentrations of Cd, Ni, Zn, Ba, and Co estimated with measured SOM and those estimated with generic SOM by the Orchestra framework were about 10 times larger than the deviations in the predictions by the transfer functions. High uncertainties were induced by lacking data in assessing solid-solution partitioning of oxy-anions like As, Mo, Sb, Se, and V. Deviations associated with lacking data in predicting soil solution concentrations of these metals by the Orchestra framework reached three-to-six orders of magnitude. The solid-solution partitioning of metal cations was strongly influenced by pH and contents of organic matter, oxides, and reactive metals. Deviations of more than two orders of magnitude were frequently observed between the estimates of soil solution concentrations with the generic values of these properties and the estimates based on the measured data. Reliable information on these properties is preferred to be included in the assessment by either the Orchestra framework or transfer functions.

Delineating ion-ion interactions by electrostatic modeling for predicting rhizotoxicity of metal mixtures to lettuce Lactuca sativa.

Effects of ion-ion interactions on metal toxicity to lettuce Lactuca sativa were studied based on the electrical potential at the plasma membrane surface (ψ0 ). Surface interactions at the proximate outside of the membrane influenced ion activities at the plasma membrane surface ({M(n+)}0). At a given free Cu(2+) activity in the bulk medium ({Cu(2+)}b), additions of Na(+), K(+), Ca(2+), and Mg(2+) resulted in substantial decreases in {Cu(2+)}0. Additions of Zn(2+) led to declines in {Cu(2+)}0, but Cu(2+) and Ag(+) at the exposure levels tested had negligible effects on the plasma membrane surface activity of each other. Metal toxicity was expressed by the {M(n+)}0 -based strength coefficient, indicating a decrease of toxicity in the order: Ag(+) > Cu(2+) > Zn(2+). Adsorbed Na(+), K(+), Ca(2+), and Mg(2+) had significant and dose-dependent effects on Cu(2+) toxicity in terms of osmolarity. Internal interactions between Cu(2+) and Zn(2+) and between Cu(2+) and Ag(+) were modeled by expanding the strength coefficients in concentration addition and response multiplication models. These extended models consistently indicated that Zn(2+) significantly alleviated Cu(2+) toxicity. According to the extended concentration addition model, Ag(+) significantly enhanced Cu(2+) toxicity whereas Cu(2+) reduced Ag(+) toxicity. By contrast, the response multiplication model predicted insignificant effects of adsorbed Cu(2+) and Ag(+) on the toxicity of each other. These interactions were interpreted using ψ0, demonstrating its influence on metal toxicity.

Modelling metal-metal interactions and metal toxicity to lettuce Lactuca sativa following mixture exposure (Cu²âº-Zn²âº and Cu²âº-Ag⁺).

Metal toxicity to lettuce Lactuca sativa was determined following mixture exposure based on the concepts of concentration addition (CA) and response addition (RA). On the basis of conventional models assuming no interaction between mixture components, Ag(+) was the most toxic, followed by Cu(2+) and Zn(2+). Furthermore, ion-ion interactions were included in quantitatively estimating toxicity of interactive mixtures of Cu(2+)-Zn(2+) and Cu(2+)-Ag(+) by linearly expanding the CA and RA models. About 80-92% of the variability in the root growth could be explained by this approach. Estimates by the extended models indicate significant alleviative effects of Zn(2+) on Cu(2+) toxicity whereas Cu(2+) did not significantly affect Zn(2+) toxicity. According to the extended CA model, Cu(2+) significantly reduced Ag(+) toxicity while Ag(+) enhanced Cu(2+) toxicity. Similar effects were not found by the extended RA model. These interactions might result from their individual uptake mechanisms and toxic actions as published in literature.

Predicting effects of cations on copper toxicity to lettuce (Lactuca sativa) by the biotic ligand model.

A biotic ligand model (BLM) was developed to estimate Cu toxicity to lettuce (Lactuca sativa) in terms of root elongation after 4 d of exposure. Effects of Na(+), K(+), Ca(2+), and Mg(2+) on Cu toxicity were examined. The addition of these cations resulted in a 50-fold difference in the copper median effective activity (EC50 cu2+). However, these variations could not be interpreted entirely as a function of the concentrations of these cations alone. In particular, only the relationship between EC50 cu2+ and the activity of protons was found to be significant in the whole range of pH examined from 5.0 to 7.0. The addition of K(+), Na(+), Ca(2+), and Mg(2+) at concentrations up to 20 mM resulted in a 16-fold difference in EC50 cu2+ values. This difference was significant, as indicated by non-overlapping standard deviations of the negative logarithm of EC50 cu2+ pEC50 cu2+) obtained with (7.37 ± 0.22) and without (6.76 ± 0.22) additions of K(+), Na(+), Ca(2+), and Mg(2+). The variations were not statistically significantly related to concentrations of these cations; therefore, only protons can be integrated in the BLM predicting Cu toxicity to lettuce L. sativa with the important parameters: log K(HBL) =6.27, log K(CuBL) =7.40, and [formula in text]. The lack of significant relationships between EC50 cu2+ and concentrations of the cations was not in line with the main assumption of the BLM about the competition between cations for binding sites.

Modeling metal bioaccumulation in the invasive mussels Dreissena polymorpha and Dreissena rostriformis bugensis in the rivers Rhine and Meuse.

The metal-specific covalent index and the species-specific size-based filtration rate were integrated into a biokinetic model estimating metal bioaccumulation in mussels from the dissolved phase and phytoplankton. The model was validated for zebra (Dreissena polymorpha) and quagga (Dreissena rostriformis bugensis) mussels in the rivers Rhine and Meuse, the Netherlands. The model performed well in predicting tissue concentrations in different-sized zebra mussels from various sampling sites for (55) Mn, (56) Fe, (59) Co, (60) Ni, (82) Se, (111) Cd, (118) Sn, and (208) Pb (r(2) =0.71-0.99). Performance for (52) Cr, (63) Cu, (66) Zn, (68) Zn, and (112) Cd was moderate (r(2) <0.20). In quagga mussels, approximately 73 to 94% of the variability in concentrations of (82) Se, (111) Cd, (112) Cd, and (208) Pb was explained by the model (r(2) =0.73-0.94), followed by (52) Cr, (55) Mn, (56) Fe, (60) Ni, and (63) Cu (r(2) =0.48-0.61). Additionally, in both zebra and quagga mussels, average modeled concentrations were within approximately one order of magnitude of the measured values. In particular, in zebra mussels, estimations of (60) Ni and (82) Se concentrations were equal to 51 and 76% of the measurements, respectively. Higher deviations were observed for (52) Cr, (59) Co, (55) Mn, (56) Fe, (111) Cd, (63) Cu, and (112) Cd (underestimation), and (66) Zn, (68) Zn, (208) Pb, and (118) Sn (overestimation). For quagga mussels, modeled concentrations of (66) Zn and (68) Zn differed approximately 14% from the measured levels. Differences between predictions and measurements were higher for other metals.