Improving Risk Assessment: Toxicological Research Needs

A workshop convened to define research needs in toxicology identified several deficiencies in data and methods currently applied in risk assessment. The workshop panel noted that improving the link between chemical exposure and toxicological response requires a better understanding of the biological basis for inter-and intra-human variability and susceptibility. This understanding will not be complete unless all life stages are taken into consideration. Because animal studies serve as a foundation for toxicological assessment, proper accounting for cross-species extrapolation is essential. To achieve this, adjustments for dose-rate effects must be improved, which will aid in extrapolating toxicological responses to low doses and from short-term exposures. Success depends on greater use of validated biologically based dose-response models that include pharmacokinetic and pharmacodynamic data. Research in these areas will help define uncertainty factors and reduce reliance on underlying default assumptions. Throughout the workshop the panel recognized that biomedical science and toxicology in particular is on the verge of a revolution because of advances in genomics and proteomics. Data from these high-output technologies are anticipated to greatly improve risk assessment by enabling scientists to better define and model the elements of the relationship between exposure to biological hazards and health risks in populations with differing susceptibilities.     

Current status of the problem of quantitative estimation of cytogenetic effects of low doses of radiation

The results of the author's own experimental studies and the literature on the cytogenetic effects in human lymphocyte culture induced by low-level radiation are presented. The quantitative regularities of the occurrence of structural chromosome damages in the genome of human somatic cells under the effect of low doses differ from those induced by high doses of ionizing radiation. The adequacy of extrapolating the risk of harmful after-effects from high-to low-level radiation is discussed.     

Statistical models for low dose exposure

Extrapolation of health risks from high to low doses has received a considerable amount of attention in carcinogenic risk assessment over decades. Fitting statistical dose-response models to experimental data collected at high doses and use of the fitted model for estimating effects at low doses lead to quite different risk predictions. Dissatisfaction with this procedure was formulated both by toxicologists who saw a deficit of biological knowledge in the models as well as by risk modelers who saw the need of mechanistically-based stochastic modeling. This contribution summarizes the present status of low dose modeling and the determination of the shape of dose-response curves. We will address the controversial issues of the appropriateness of threshold models, the estimation of no observed adverse effect levels (NOAEL), and their relevance for low dose modeling. We will distinguish between quantal dose-response models for tumor incidence and models of the more informative age/time dependent tumor incidence. The multistage model and the two-stage model of clonal expansion are considered as dose-response models accounting for biological mechanisms. Problems of the identifiability of mechanisms are addressed, the relation between administered dose and effective target dose is illustrated by examples, and the recently proposed Benchmark Dose concept for risk assessment is presented with its consequences for mechanistic modeling and statistical estimation.     

Issues in Exposure and Dose Assessment for Epidemiology and Risk Assessment

Epidemiologic studies have been effective in identifying human environmental and occupational hazards. However, most epidemiologic data has been difficult to use in quantitative risk assessments because of the vague specification of exposure and dose. Toxicologic animal studies have used applied doses (quantities administered, or exposures with fixed duration) and well characterized end points to determine effects. However, direct use of animal data in human risk assessment has been limited by uncertainties in the extrapolation. The applied dose paradigm of toxicology is not suited for cross species extrapolation, nor for use in epidemiology as a dose metric because of the complexity of human exposures. Physiologically based pharmacokinetic (PBPK) modeling can estimate the time course of tissue concentrations in humans, given an exposure-time profile, and it has been used for extrapolating findings from animals to humans. It is proposed that human PBPK modeling can be used in appropriately designed epidemiologic studies to estimate tissue concentrations. Secondly, tissue time courses can be used to form dose metrics based on the type and time course of adverse effects. These dose metrics will strengthen the determination of epidemiologic dose-response relationships by reducing misclassification. Findings from this approach can be readily integrated into quantitative risk assessment.     

Transgenic rice plants expressing human p450 genes involved in xenobiotic metabolism for phytoremediation

Phytoremediation is the use of plants to remove xenobiotic compounds from the environment. Plants have the inherent ability to detoxify xenobiotic pollutants, but they are generally poor at degrading them. The introduction of genes involved in xenobiotic degradation is aimed at enhancing plants' potential further. Rice (Oryza sativa) is a good candidate for this purpose and has been transformed with genes encoding cytochrome P450 monooxygenases CYP1A1, CYP2B6, and CYP2C19. The transgenic plants were more tolerant to various herbicides than nontransgenic Nipponbare rice plants, owing to enhanced metabolism by the introduced P450 enzymes. Transgenic plants were able to remove atrazine and metolachlor from soil. Field testing and risk assessment are very important for developing transgenic plants for phytoremediation. Transgenic rice plants should become useful as herbicide-tolerant crops and for phytoremediation of xenobiotic pollutants in future.     

Radiation breakage of DNA: a model based on random-walk chromatin structure

Monte Carlo computer software, called DNAbreak, has recently been developed to analyze observed non-random clustering of DNA double strand breaks in chromatin after exposure to densely ionizing radiation. The software models coarse-grained configurations of chromatin and radiation tracks, small-scale details being suppressed in order to obtain statistical results for larger scales, up to the size of a whole chromosome. We here give an analytic counterpart of the numerical model, useful for benchmarks, for elucidating the numerical results, for analyzing the assumptions of a more general but less mechanistic “randomly-located-clusters” formalism, and, potentially, for speeding up the calculations. The equations characterize multi-track DNA fragment-size distributions in terms of one-track action; an important step in extrapolating high-dose laboratory results to the much lower doses of main interest in environmental or occupational risk estimation. The approach can utilize the experimental information on DNA fragment-size distributions to draw inferences about large-scale chromatin geometry during cell-cycle interphase. Received: 2 March 2000 / Revised version: 2 February 2001 / Published online: 21 August 2001     

Risk assessment of neurotoxic pesticides

In order to establish safe exposure levels for toxic chemicals, risk assessment guidelines have been developed. A compilation is given by the author on the elements of risk assessment of hazardous neurotoxic pesticides, using data obtained from human epidemiological studies, from animal experiments, from the international literature and from the author's own experiments as well. Well-controlled laboratory studies of neurotoxicity have the potential to provide adequate exposure and effect data for accurate hazard identification. Animal models of neurotoxicity as highly sensitive behavioral and neurophysiological methods as a function of doses, provide data for human low dose extrapolation by using mathematical models. This procedure might be the basis for reducing risk ("risk management"), therefore some examples are given, how to handle properly neurotoxic pesticides with different- high or low-risk.     

A note on the use of a stochastic mammillary compartmental model as an environmental safety model

Environmental safety testing typically requires procedures for extrapolating from the relatively high experimental to the very low use doses of potentially harmful substances. In the present paper, a stochastic mammillary compartmental model for environmental safety testing is proposed and extrapolation procedures based on its dose-response relationship are developed. The proposed model is a direct generalization of one of the basic safety models, the one-hit model, in that a harmful reaction is assumed to occur if at any time any of the peripheral compartments attains a specified threshold of particles. Consideration of a closed model yields an upper bound on the probability of attaining a certain threshold level, thus providing a conservative procedure for extrapolating to a low dose, while a lower bound obtained from a related open model provides a useful monitoring device as to the sharpness of the upper, bound. The extrapolation procedure is illustrated with simulated data and approximations for initial values are developed.     

Physiologically-Based Toxicokinetic Modeling of Zearalenone and Its Metabolites: Application to the Jersey Girl Study

Zearalenone (ZEA), a fungal mycotoxin, and its metabolite zeranol (ZAL) are known estrogen agonists in mammals, and are found as contaminants in food. Zeranol, which is more potent than ZEA and comparable in potency to estradiol, is also added as a growth additive in beef in the US and Canada. This article presents the development and application of a Physiologically-Based Toxicokinetic (PBTK) model for ZEA and ZAL and their primary metabolites, zearalenol, zearalanone, and their conjugated glucuronides, for rats and for human subjects. The PBTK modeling study explicitly simulates critical metabolic pathways in the gastrointestinal and hepatic systems. Metabolic events such as dehydrogenation and glucuronidation of the chemicals, which have direct effects on the accumulation and elimination of the toxic compounds, have been quantified. The PBTK model considers urinary and fecal excretion and biliary recirculation and compares the predicted biomarkers of blood, urinary and fecal concentrations with published in vivo measurements in rats and human subjects. Additionally, the toxicokinetic model has been coupled with a novel probabilistic dietary exposure model and applied to the Jersey Girl Study (JGS), which involved measurement of mycoestrogens as urinary biomarkers, in a cohort of young girls in New Jersey, USA. A probabilistic exposure characterization for the study population has been conducted and the predicted urinary concentrations have been compared to measurements considering inter-individual physiological and dietary variability. The in vivo measurements from the JGS fall within the high and low predicted distributions of biomarker values corresponding to dietary exposure estimates calculated by the probabilistic modeling system. The work described here is the first of its kind to present a comprehensive framework developing estimates of potential exposures to mycotoxins and linking them with biologically relevant doses and biomarker measurements, including a systematic characterization of uncertainties in exposure and dose estimation for a vulnerable population.     

Human exposure to endocrine disrupters: carcinogenic risk assessment

Human exposure to endocrine disrupters (EDs) is widespread and is considered to pose a growing threat to human health. Recent advances in molecular and genetic research and better understanding of mechanisms of blastic cell transformation have led to efforts to improve cancer risk assessment for populations exposed to this family of xenobiotics. In risk assessment, low dose extrapolation of cancer incidence data from both experimental animals and epidemiology studies has been largely based on models assuming linear correlation at low doses, despite existence of evidence showing otherwise. Another weakness of ED risk assessment is poor exposure data in ecological studies. Those are frequently rough estimates derived from contaminated items of local food basket surveys. Polyhalogenated hydrocarbons are treated as examples. There is growing sense of urgency to develop a biologically based dose response model of cancer risk, integrating emerging data from molecular biology and epidemiology to provide more realistic data for risk assessors, public, public health managers and environmental issues administrators.     

Estimating radiation-induced cancer risks at very low doses: rationale for using a linear no-threshold approach

The possible cancer risks caused by ionizing radiation doses of ~1 mSv or less are too small to be estimated directly from epidemiological data. The linear no-threshold (LNT) approach to estimating such risks involves using epidemiological data at higher (but still low) doses to establish an “anchor point”, and then extrapolating the excess cancer risk linearly down from this point to the low dose of interest. The study in this issue by Professor Tubiana and colleagues, summarizing a French Academy of Sciences report, argues that such LNT extrapolations systematically give substantial overestimates of the excess cancer risk at very low doses. We suggest that, to the contrary, even if there are significant deviations from linearity in the relevant dose range, potentially caused by the effects of inter-cellular interactions or immune surveillance, we know almost nothing quantitatively about these effects. Consequently, we do not know the magnitude, nor even the direction of any such deviations from linearity—the risks could indeed be lower than those predicted by a linear extrapolation, but they could well be higher.     

Animal models in carotenoids research and lung cancer prevention

Numerous epidemiological studies have consistently demonstrated that individuals who eat more fruits and vegetables (which are rich in carotenoids) and who have higher serum β-carotene levels have a lower risk of cancer, especially lung cancer. However, two human intervention trials conducted in Finland and in the United States have reported contrasting results with high doses of β-carotene supplementation increasing the risk of lung cancer among smokers. The failure of these trials to demonstrate actual efficacy has resulted in the initiation of animal studies to reproduce the findings of these two studies and to elucidate the mechanisms responsible for the harmful or protective effects of carotenoids in lung carcinogenesis. Although these studies have been limited by a lack of animal models that appropriately represent human lung cancer induced by cigarette smoke, ferrets and A/J mice are currently the most widely used models for these types of studies. There are several proposed mechanisms for the protective effects of carotenoids on cigarette smoke-induced lung carcinogenesis, and these include antioxidant/prooxidant effects, modulation of retinoic acid signaling pathway and metabolism, induction of cytochrome P450, and molecular signaling involved in cell proliferation and/or apoptosis. The technical challenges associated with animal models include strain-specific and diet-specific effects, differences in the absorption and distribution of carotenoids, and differences in the interactions of carotenoids with other antioxidants. Despite the problems associated with extrapolating from animal models to humans, the understanding and development of various animal models may provide useful information regarding the protective effects of carotenoids against lung carcinogenesis.     

Human and animal hepatocytes in vitro with extrapolation in vivo

Human and animal hepatocytes are now being used as an in vitro technique to aid drug discovery by predicting the in vivo metabolic pathways of drugs or new chemical entities (NCEs), identifying drug-metabolizing enzymes and predicting their in vivo induction. Because of the difficulty of establishing whether the cytotoxic susceptibility of human hepatocytes to xenobiotics/drugs in vitro could be used to predict in vivo human hepatotoxicity, a comparison of the susceptibility of the hepatocytes of human and animal models to six chemical classes of drugs/xenobiotics in vitro have been related to their in vivo hepatotoxicity and the corresponding activity of their metabolizing enzymes. This study showed that the cytotoxic effectiveness of 16 halobenzenes towards rat hepatocytes in vitro using higher doses and short incubation times correlated well with rat hepatotoxic effectiveness in vivo with lower doses/longer times. The hepatic/hepatocyte xenobiotic metabolizing enzyme activities of various animal species and human have been reviewed for use by veterinarians and research scientists. Where possible, recommendations have been made regarding which animal hepatocyte model is most applicable for modeling the susceptibility to xenobiotic induced hepatotoxicity of those humans with slow versus rapid metabolizing enzyme polymorphisms. These recommendations are based on the best human fit for animal drug/xenobiotic metabolizing enzymes in terms of activity, kinetics and substrate/inhibitor specificity. The use of human hepatocytes from slow versus rapid metabolizing individuals for drug metabolism/cytotoxicity studies; and the research use of freshly isolated rat hepatocytes and "Accelerated Cytotoxicity Mechanism Screening" (ACMS) techniques for identifying drug/xenobiotic reactive metabolites are also described. Using these techniques the molecular hepatocytotoxic mechanisms found in vitro for seven classes of xenobiotics/drugs were found to be similar to the rat hepatotoxic mechanisms reported in vivo.     

External radiation doses from <Superscript>137</Superscript>Cs to frog phantoms in a wetland area: in situ measurements and dose model calculations

For assessment of external radiation doses to frogs in a wetland area contaminated with ¹³⁷Cs, frog phantoms were constructed from polymethyl methacrylate (PMMA). The frog phantoms contained thermoluminescence (TL) chips and were used in situ at two study sites to measure doses. To test if higher doses are received by the sensitive skin of frogs, extra-thin TL chips were applied close to the surface of the frog phantoms. In addition, the measured doses were compared with those calculated on the basis of soil sample data from the wetland multiplied with dose-conversion coefficients from the US Department of Energy’s RESRAD-BIOTA code and from the ERICA assessment tool. Measured doses were generally lower than those calculated to ellipsoids used to model frogs. Higher doses were measured at the frog phantoms’ surfaces in comparison to inner parts at one of the two sites indicating that the frogs’ thin skin could receive a higher radiation dose than expected. In the efforts to assure protection of non-human biota, in situ measurements with phantoms provide valuable dose information and input to dose models in site-specific risk assessments of areas contaminated with radionuclides.     

Extrapolating toxic effects on individuals to the population level: the role of dynamic energy budgets

The interest of environmental management is in the long-term health of populations and ecosystems. However, toxicity is usually assessed in short-term experiments with individuals. Modelling based on dynamic energy budget (DEB) theory aids the extraction of mechanistic information from the data, which in turn supports educated extrapolation to the population level. To illustrate the use of DEB models in this extrapolation, we analyse a dataset for life cycle toxicity of copper in the earthworm Dendrobaena octaedra. We compare four approaches for the analysis of the toxicity data: no model, a simple DEB model without reserves and maturation (the Kooijman-Metz formulation), a more complex one with static reserves and simplified maturation (as used in the DEBtox software) and a full-scale DEB model (DEB3) with explicit calculation of reserves and maturation. For the population prediction, we compare two simple demographic approaches (discrete time matrix model and continuous time Euler-Lotka equation). In our case, the difference between DEB approaches and population models turned out to be small. However, differences between DEB models increased when extrapolating to more field-relevant conditions. The DEB3 model allows for a completely consistent assessment of toxic effects and therefore greater confidence in extrapolating, but poses greater demands on the available data.     

Advances in development of transgenic plants for remediation of xenobiotic pollutants

Phytoremediation-the use of plants for cleaning up of xenobiotic compounds-has received much attention in the last few years and development of transgenic plants tailored for remediation will further enhance their potential. Although plants have the inherent ability to detoxify some xenobiotic pollutants, they generally lack the catabolic pathway for complete degradation/mineralization of these compounds compared to microorganisms. Hence, transfer of genes involved in xenobiotic degradation from microbes/other eukaryotes to plants will further enhance their potential for remediation of these dangerous groups of compounds. Transgenic plants with enhanced potential for detoxification of xenobiotics such as trichloro ethylene, pentachlorophenol, trinitro toluene, glycerol trinitrate, atrazine, ethylene dibromide, metolachlor and hexahydro-1,3,5-trinitro-1,3,5-triazine are a few successful examples of utilization of transgenic technology. As more genes involved in xenobiotic metabolism in microorganisms/eukaryotes are discovered, it will lead to development of novel transgenic plants with improved potential for degradation of recalcitrant contaminants. Selection of suitable candidate plants, field testing and risk assessment are important considerations to be taken into account while developing transgenic plants for phytoremediation of this group of pollutants. Taking advantage of the advances in biotechnology and 'omic' technologies, development of novel transgenic plants for efficient phytoremediation of xenobiotic pollutants, field testing and commercialization will soon become a reality.     

Biomarkers as Predictors in Health and Ecological Risk Assessment

Biomarkers are measurable biological parameters that change in response to xenobiotic exposure and other environmental or physiological stressors, and can be indices of toxicant exposure or effects. If the biomarkers are sufficiently specific and well characterized, they can have great utility in the risk assessment process by providing an indication of the degree of exposure of humans or animals in natural populations to a specific xenobiotic or class of xenobiotics. Most biomarkers are effective as indices of exposure, but adequate information is rarely available on the appropriate dose-response curves to have well-described biomarkers of effect that can be widely applicable to additional populations. Specific examples of acetylcholinest-erase inhibition following exposure to organophosphorus insecticides are cited from experiments in both mammals (rats) and fish. These experiments have indicated that the degree of inhibition can be readily influenced by endogenous (e.g., age) and exogenous (e.g., chemical exposures) factors, and that the degree of inhibition is not readily correlated with toxicological effects. Caution is urged, therefore, in an attempt to utilize biomarkers in the risk assessment process until more complete documentation is available on the specificity, sensitivity, and time course of changes, and on the impact of multiple exposures or the time of exposures.     

The Effect of Ongoing Exposure Dynamics in Dose Response Relationships

Characterizing infectivity as a function of pathogen dose is integral to microbial risk assessment. Dose-response experiments usually administer doses to subjects at one time. Phenomenological models of the resulting data, such as the exponential and the Beta-Poisson models, ignore dose timing and assume independent risks from each pathogen. Real world exposure to pathogens, however, is a sequence of discrete events where concurrent or prior pathogen arrival affects the capacity of immune effectors to engage and kill newly arriving pathogens. We model immune effector and pathogen interactions during the period before infection becomes established in order to capture the dynamics generating dose timing effects. Model analysis reveals an inverse relationship between the time over which exposures accumulate and the risk of infection. Data from one time dose experiments will thus overestimate per pathogen infection risks of real world exposures. For instance, fitting our model to one time dosing data reveals a risk of 0.66 from 313 Cryptosporidium parvum pathogens. When the temporal exposure window is increased 100-fold using the same parameters fitted by our model to the one time dose data, the risk of infection is reduced to 0.09. Confirmation of this risk prediction requires data from experiments administering doses with different timings. Our model demonstrates that dose timing could markedly alter the risks generated by airborne versus fomite transmitted pathogens.