Elsevier

Aquaculture

Volume 251, Issues 2–4, 28 February 2006, Pages 280-294
Aquaculture

A biologically based damage assessment model to enhance aquacultural water quality management

https://doi.org/10.1016/j.aquaculture.2005.05.036Get rights and content

Abstract

The lethal concentration for 50% of aquacultural animals (LC50)-based tests determines the external effect concentration (EEC) following certain statistical models, revealing that no biologically based mechanistic information and only statistical interpretations of its model parameters could be made. The purpose of this paper is to determine the survival risk of waterborne metals toward farmed species with respect to lethality based on biologically based mechanistic models. Here we study a biologically based mechanistic damage assessment model (DAM) compared with a pharmacodynamic (PD)-based critical area under the curve (CAUC) model to demonstrate the ability of predicting the internal effect concentration (IEC) and survival rate of farmed species. We tested the proposed models using published acute toxicity and accumulation data for two farmed species, tilapia (Orechromis mossambicus) exposed to arsenic (As) and abalone (Haliotis diversicolor supertexta) exposed to zinc (Zn), to compare observed and predicted LC50 and IEC and, subsequently, to predict the survival rate. Our analyses demonstrate that the DAM- and PD-based survival models performed well and proved its usefulness as a tool in the quantification of risk assessment in aquacultural ecosystems. The study also supports the suggestion that replacing exposure-based EECs by IECs is a first step toward a measure for inherent toxicity and can be used to improve the construction of future environmental quality criteria programs aimed at protecting and restoring the rapidly degrading aquacultural ecosystems.

Introduction

Due to anthropogenic activities or geochemical cycling, heavy metal pollution is a serious problem and is one of the most studied aquacultural ecosystem problems. Aquacultural organisms may bioaccumulate trace metals in their tissues and consequently threaten themselves directly. Standard toxicity tests performed in the laboratory are used extensively to predict the effects of chemicals in aquatic ecosystems. This is the first step in determining the environmental risk of a chemical to aquatic species. We usually use a preselected time span to determine the median lethal concentration (LC50). LC50-based tests determine the external effect concentration (EEC), although the observed effect depends on the intrinsic toxicity or biokinetic behavior of the chemical in the aquatic animal. Therefore, the LC50-based parameters are mostly model-dependent, yet the models usually employed, such as logit and probit, have no biologically based assumptions that allow questions about the relevancy of such models. Hill (1910) has tried to employ a log-logistic model to account for an interaction between the xenobiotic and the receptor-mediated active compound in the organism. However, the conventional analysis of bioassays does not account for biological significance.

Normally, internal effect concentration (IEC) is not measured in toxicity experiments; therefore, we usually treat IEC as a hidden variable in that the tissue concentration is scaled with the bioconcentration factor (BCF) in order to obtain a quantity that is directly proportional to the tissue concentration yet has the dimension of an EEC (Freidig et al., 1999, Lee et al., 2002a). A different approach of measuring the acute toxicity of chemicals is based on IECs, instead of EECs. External LC50 values are then replaced by lethal body concentrations (LBCs) (French-McCay, 2002). The LBC is the concentration of a chemical within an organism at the time of death and can be estimated from experiments in which the increase of mortality with exposure time is observed in conjunction with the concentration of chemicals in the body.

The current procedures in bioassays consist of observing lethality at fixed times, which can lead to the determination of LC50 endpoints, rather than survival curves. There is then a statistical dependence of LC50 data at consecutive times because they concern the same organisms. It is more robust and powerful to use the dose–time–response data than just the LC50 values. Survival models also consider raw experimental data as time to death versus concentration. They have an intrinsically greater statistically power, yet any biological interpretation is limited.

Based on toxicological principles, the mechanisms through which the dose at the target site elicits the ultimate adverse response are described by pharmacodynamic (PD) scheme and referred to as the action of the effect dose at the target site. Recently, Verhaar et al. (1999) and Legierse et al. (1999) have developed a PD-based model, the critical area under the curve (CAUC) model, to describe the time course of LC50 data for chemicals that act through the irreversible interaction between chemicals and receptors. The CAUC model could be applied to depict the acute toxicity and to estimate incipient LC50 values and IECs of waterborne chemicals in organisms. PD-based models have been continuously developed for the understanding of bioassay data (Liao et al., 2002, Liao and Ling, 2004).

One biologically based mechanistic model based on the damage assessment model (DAM) was developed by Lee et al. (2002b) to describe and predict time-dependent toxicity data. DAM depicts the modes of action, including rapid reversible binding to the target site as well as to those that act with irreversible binding. Thus, both of the critical body residue (CBR) and the CAUC models are extreme cases of the DAM (Lee et al., 2002b). DAM assumes that death occurs when the cumulative damage reaches a critical level and was described by a combination of both first-order toxicokinetic and toxicodynamic models. Damage is assumed to accumulate in proportion to the accumulated residue and damage recovery in proportion to the cumulative damage when damage is reversible. The time-dependent LC50 data are determined by both a damage recovery rate and an elimination rate, suggesting that the critical cumulative damage is the determinant of the time–concentration response relationship and not simply the CAUC. DAM is originally based on a mathematical model DEBtox (Bedaux and Kooijman, 1994, Widianarko and van Straalen, 1996), where DEBtox relates survivorship to toxicokinetics by assuming that the probability of dying (i.e., the hazard rate) is related to the concentration of the toxicant in the organism. DEBtox models have also been extensively applied in the fields of ecological risk assessment (Pery et al., 2001, Bonnoment et al., 2002).

The objective of this paper is to determine the survival risk of waterborne metals toward farmed species with respect to lethality based on a biologically based DAM compared with a PD-based CAUC model to demonstrate the ability in predicting IEC and survival rate. Our purposed models are able to describe time-dependent toxicity data, which contain information about the dynamic aspect of the occurrence of effects. The models were adapted to fit the data from toxicity and accumulation experiments simultaneously to reveal mechanistic information on metal toxicity on aquatic animals. We also make an exploratory analysis on the basis of toxicokinetic parameters to predict the IEC and survival of farmed species based on the EEC.

We test the proposed models using published acute toxicity and accumulation data for two farmed species, tilapia (Orechromis mossambicus) exposed to arsenic (As) and abalone (Haliotis diversicolor supertexta) exposed to zinc (Zn), to compare observed and predicted LC50 and IEC and, subsequently, to predict the survival rate. Arsenic and Zn were chosen for practical as well as theoretical reasons, with the availability of reasonable amounts of suitable information as the primary consideration. Generally, as prerequisites for data suitability, we required exposure and whole-body As and Zn levels measured by accepted analytical techniques. In this respect, we considered experimental exposure data to be acceptable only when whole body concentration data were available and when the exposure duration was at least 14 days. Our previous published As–tilapia and Zn–abalone databases meet this principle. On the other hand, As and Zn were chosen in this study because they represent metals of general concern in terms of environmental protection and can span the continuum from nutritionally essential to nonessential.

Section snippets

Damage-based survival model

Bedaux and Kooijman (1994) have developed a biologically based model to investigate the relationships among body residues, cumulative damage, and survival rate in order to describe the time-dependent survival probability that can be expressed as the exponential of cumulative hazard as:S(t)=eH(t),where S(t) is the probability to survival until time t and H(t) is the cumulative hazard (dimensionless). In DAM-based survival modeling, a proportionality constant k3 (dimensionless) is introduced to

Fitting toxicity models to LC50(t) and CL,50(t) data

The optimal fits of the DAM, CAUC, and CBR models to the observed LC50(t) data of As–tilapia as well as LC50(t) and CL,50(t) data of Zn–abalone systems are presented, respectively, in Fig. 2A, C, and D. The estimated model-specific parameters of different chemical-species combinations are listed in Table 2, Table 3.

Fig. 2A and C show that both DAM and CBR models describe the data in a more accurate way, depending on the estimated incipient LC50 values of the DAM (LC50(∞) = 5.96 μg mL 1) and CBR

Characteristics of input and estimated parameters

Because few previous studies have evaluated As toxicity to tilapia and Zn toxicity to abalone, we did not have an a priori estimate of internal lethal body concentrations. When only LC50(t) data are available for the prediction of time-dependent toxicity, the CAUC model is capable of describing the chemical-specific trends of CL,50; however, the CBR model has limitations due to the mathematical formulation and model assumptions. Changing the assumptions could extend further studies.

The

References (32)

  • K.E. Dineley et al.

    Zinc inhibition of cellular energy production: implications for mitochondria and neurodegeneration

    Neurochemistry

    (2003)
  • D.M. Di Toro et al.

    Biotic ligand model of the acute toxicity of metals: 1. Technical basis

    Environ. Toxicol. Chem.

    (2001)
  • H.B.F. Dixon

    The biochemical action of arsenic acids especially as phosphate analogues

    Adv. Inorg. Chem.

    (1997)
  • A.P. Freidig et al.

    Comparing the potency of chemicals with multiple modes of action in aquatic toxicology: acute toxicity due to narcosis versus reactive toxicity of acrylic compounds

    Environ. Sci. Technol.

    (1999)
  • D.P. French-McCay

    Development and application of an oil toxicity and exposure model OilToxEx

    Environ. Toxicol. Chem.

    (2002)
  • L.A. Gaither et al.

    Eukaryotic zinc transporters and their regulation

    BioMetals

    (2001)
  • Cited by (12)

    • Combining groundwater quality analysis and a numerical flow simulation for spatially establishing utilization strategies for groundwater and surface water in the Pingtung Plain

      2016, Journal of Hydrology
      Citation Excerpt :

      Groundwater typically contains a variety of natural and anthropogenic contaminants, so its quality is one of the most crucial factors affecting groundwater utilization for various purposes. For example, the use of the water with high concentrations of manganese (Mn) and iron (Fe) for purposes of agriculture or aquaculture may lead to retardation of growth or even the death of cultivated plants or fish (Tsai et al., 2006; Liao et al., 2008; Ye et al., 2009). The use of groundwater with high level of arsenic (As) is harmful to human health.

    • The long and bumpy journey: Taiwan's aquaculture development and management

      2014, Marine Policy
      Citation Excerpt :

      The multiple uses have given rise to conflicts over resource use. For example, heavy metal pollution is particularly a serious problem for aquaculture organisms [42]. “Green oysters” in marine culture farms located at northwest Taiwan׳s coast was a vivid illustration of heavy metal pollution of copper and zincs [43].

    • Combining bioaccumulation and coping mechanism to enhance long-term site-specific risk assessment for zinc susceptibility of bivalves

      2011, Chemosphere
      Citation Excerpt :

      Zinc (Zn) is an essential micronutrient for almost all aquatic organisms. This transition metal is responsible for many biochemical processes including regulatory, structural, and enzymatic functions (Tsai et al., 2006). Within its optimal range of concentration, Zn plays a critical role in growth, physiology, and development for aquatic organisms (Sappal et al., 2009).

    • Modeling time-dependent toxicity to aquatic organisms from pulsed exposure of PAHs in urban road runoff

      2011, Environmental Pollution
      Citation Excerpt :

      Ashauer combined the damage assessment model and DEBtox concept to form a threshold damage model, which was used to simulate the survival of an aquatic invertebrate after fluctuating and sequential pulsed exposure to pesticides (Ashauer et al., 2007a, 2007b). In addition, the critical area under the curve model and the critical body residue model were also developed to link fluctuating exposure to effects (Tsai et al., 2006; Lee et al., 2002b; Ashauer et al., 2006). From the aspect of runoff discharge, long term integrity of receiving water bodies is dependent on ecological response to the infrequent exposure to a random range of runoff contaminant.

    View all citing articles on Scopus
    View full text