Dose metrics

Monte Carlo simulation, an iterative technique which derives a range of risk estimates, was incorporated into a trichloroethylene risk assessment using the PBPK model developed by Fisher and Allen (1993). The results of this study (Cronin et al. 1995), which used the kinetics of TCA production and trichloroethylene elimination as the dose metrics relevant to carcinogenic risk, indicated that concentrations of 0.09-1.0 pg/L (men) and 0.29-5.3 pg/L (women) in drinking water correspond to a cancer risk in humans of 1 in 1 million. For inhalation exposure, a similar risk was obtained from intermittent exposure to 0.07-13.3 ppb (men) and 0.16-6.3 ppb (women), or continuous exposure to 0.01-2.6 ppb (men) and 0.03-6.3 ppb (women) (Cronin et al. 1995). 

Oberdorster, G Oberdorster, E., and Oberdorster, J. (2007) Concepts of nanoparticle dose metric and response metric. Environmental Health Perspectives, 115 (6), A290. 

Traditional, default approaches for toxicological risk assessment are not based on specific understanding of modes of action and tissue dose metrics (e.g., tissue concentrations, body burdens, AUCs). In recent years, PBPK/PBTK modeling has found frequent application in risk assessments where PBPK models serve as important adjuncts to studies on modes of action of xenobiotics. 

Use animal PK modeling to convert the dose-response relationship seen in toxicity studies (applied dose) to a dose-response relationship based on internal dose, using a dose metric derived from human biomonitoring data. This approach fosters the development of a biomarker-response relationship and biomarker-based toxicity values. 

The doses from exposure are characterized by distributions. For each possible dose level, these distributions quantify the probability that an individual in a specified population or subpopulation will receive that dose level as a result of exposure to atrazine and simazine through drinking water ingestion, dietary consumption, herbicide handling, or a combination of these potential exposure routes. For chronic toxic endpoints, the traditional (default) dose metric summarizing a lifetime of exposure is the lifetime average daily dose (LADD). Distributions of LADDs have been determined, and the corresponding distributions of the MOEs are presented herein. 

When appropriately validated and understood, biomarkers present unique advantages as tools for exposure assessment (Gundert-Remy et al, 2003). Biomarkers provide indices of absorbed dose that account for all routes and integrate over a variety of sources of exposure (IPCS, 1993, 2001a). Certain biomarkers can be used to represent past exposure (e.g. lead in bone), recent exposure (e.g. arsenic in urine), and even future target tissue doses (e.g. pesticides in adipose tissue). Once absorbed dose is determined using biomarkers, the line has been crossed between external exposure and the dose metrics that reflect the pharmacokinetics and toxicokinetics of an agent (see section 5.3.3). 

Exposure or dose metrics are relevant to the type of health effects being investigated... 

When it comes to mixtures, an important development is the use of the internal dose as a dose metric, particularly in human assessments. The internal dose is either measured directly or modeled using PBPK models, for example, as a blood or a target tissue concentration. Application of an internal dose metric makes it possible to account for 1) interindividual variability in toxicokinetics, 2) temporal variations in exposure patterns, and 3) interactions between substances during absorption, metabolism, and transport. In ecological risk assessment, internal doses are sometimes measured but rarely modeled with PBPK models. The awareness is growing that the internal dose is a useful metric but the use in formal risk assessment procedures is still limited, for separate compounds as well as for mixtures. 

Once there is a measure of the concentration of the pesticide in the exposure medium (air, water, food, etc.) in contact with the body or the actual concentration that comes into contact with the body, a daily dose metric can be calculated (e.g. maximum, average, geometric mean, etc.). This typically involves developing a mathematical equation that expresses dose as a function of pesticide concentration and other important parameters referred to as human exposure factors (USEPA, 1999a). In the context of this discussion, the term human exposure factor refers specifically to (a) human characteristics, such as body weight, surface area, life expectancy, inhalation rates for air and consumption rates for food, drinking water and soil (b) human behaviors, such as activity patterns, occupational and residential mobility and consumer product use, which are used by exposure assessors to calculate potential dose. 

Development of an appropriate absorbed dose metric (including time scale) to cumulative toxic-ologically equivalent doses across the chemicals of interest. 

New dose metrics should be used that incorporate age or time dependence on the dose level rather than a lifetime average daily dose or its analog for a shorter time period... 

As shown previously, PBPK models allow the conversion of potential dose or exposure concentration to tissue dose, which can then be used for risk characterization purposes. The choice of an internal dose metric is based principally on an understanding of the mode of action of the chemical species of concern. The internal dose metric (sometimes called the biologically effective dose) is often used in place of the applied dose in quantitative dose-response assessments, in order to reduce the uncertainty inherent in using the applied dose to derive risk values. 

Risk characterization of chemicals Experimental testing is replaced by filling in data gaps needed to derive dose metrics or assessment factors that are combined with exposure information in risk assessment. 

When talking about the NM dose, one has to make a distinction between three dose metrics (1) the administered dose (particle mass, number, or surface area administered per volume media at the onset of an experiment), (2) the delivered dose (particle mass, number, or surface area to reach the cell monolayer via diffusion and sedimentation over the duration of an experiment), and (3) the cellular dose (particle mass, number, or surface area internalized by the cells during the experiment). The determination of the cellular and delivered dose of NMs is essential for accurate interpretation of results derived from in vitro particle-cell interaction studies (e.g., particle uptake, cytotoxicity, biokinetic studies) [36], Using two different cell culture configurations, upright and inverted, Cho et al. [37] have recendy shown that the uptake of NPs is gready sensitive to the position in which cells are cultured and strongly... 

Secondly, the above issues then need to be placed in a concentration-effect relationship. The main issue then is the determination of the appropriate dose metric. What is the amount (or concentration) of the chemical under study that is responsible for the effect In other words how do we determine the appropriate exposure at the site of toxic action related to the primary chemico-biological interaction that forms the basis of the compound s toxicity The commonly used practice is to relate the effects to the nominal concentration, i.e., the amount of compound added to the in vitro system divided by its volume. If data from this exposure-effect relationship are to be the basis of an estimation of risk for an organism, this approach may be a source of errors in those cases where the local exposure of the cells in vitro differs from the exposure of targets in the in vivo situation [9], These differences can result from differences in protein binding in plasma vs. culture medium or other processes that may influence the local exposure at the target, e.g., binding to culture plastic [10, 11], More appropriate dose metrics, depending on the in vitro system as well as on the chemical s mechanism of action, may be the freely available concentration, either as the peak concentration or as the area under the curve (AUC) for the free concentration, or the intracellular concentration [12]. 

Moreover, when an appropriate concentration-effect relationship in vitro has been determined, and when these data need to be applied to the determination of risk in vivo, the next quantification step is the translation of the in vitro dose metric to the dose-effect relationship in the in vivo situation. This process, referred to as the quantitative in vitro-in vivo extrapolation (QIVTVE) [13], also requires a number of considerations [14],... 

Groothuis FA, Heringa MB, Nicol B, Hermens JLM, Blaauboer BJ, Kramer NI (2014) Dose metric considerations in in vitro assays to improve quantitative in vitro-in vivo dose extrapolations. Toxicology. doi 10.1016/j. tox.2013.08.012... 

Broeders JJW, Blaauboer BJ, Hermens JLM (2013) In vitro biokinetics of chlorpromazine and the influence of different dose metrics on effect concentrations for cytotoxicity in Balb/c 3T3, Caco-2 and HepaRG cell cultures. Toxicol In Vitro 27(3) 1057-1064... 

Sixth, and finally, the adequacy of model structure as well as parameter values should be evaluated based on comparison of mode predictions with experimental data that had not been used for calibration purpose. This process essentially evaluates whether the PBPK model is capable of providing reliable predictions of the various dose metrics of potential use in a cancer risk assessement. The model should not only reprodnce consistently the shape of the pharmacokinetic time-course curve (i.e., including bnmps and valleys) and not jnst provide satisfactory fit only to a portion of the cnrve. Evaluation or validation of PBPK models should be regarded... 

Outputs Predictions of pharmacokinetic behavior or tissue dose associated with given set of parameters or structure. Sensitivity coefficients Population distributions (probability) of dose metric for an exposure situation Percentage of variance in the output (e.g., dose metric) attributed to each parameter s variance... 

By facilitating the simulation of the dose metrics for use in cancer dose-response analysis, the PBPK models address the uncertainty associated with interspecies, route-to-route, and high-dose to low-dose extrapolations (Andersen et al. 1993 Andersen and Krishnan 1994 Clewell et al. 2002a Clewell and Andersen 1987 Melnick and Kohn 2000). Since the first demonstration of the application of PBPK models in cancer risk assessment by Andersen and co-workers in 1987, there have been substantial efforts to evaluate the appropriate dose metrics and cancer risk associated with a number of other volatile organic chemicals using the PBPK modeling approach (Table 21.3). These risk assessments have been based on the PBPK model simulations of a variety of dose metrics that reflect the current state... 

TABLE 21.3. Examples of PBPK Models Developed for Assessing Dose Metrics and Cancer Risks... 

Selection of the appropriate dose metric for the risk assessment, based on the results of the preceding step and mode of action of the chemical. 

Characterization of the quantitative relationship between the dose metric and the cancer incidence observed in bioassay(s) to estimate the dose metric-based slope factor or threshold level. 

Back-calculation of the potential dose or exposure concentration associated with the dose metric for a predetermined (i.e., acceptable) risk level (e.g., 1 X 10 ) or threshold level, using the human PBPK model, based on the assumption of equivalent tissue responses for equivalent dose metrics regardless of the species (Andersen et al. 1987 Krishnan and Andersen 1991b). 

Inhalation Exposure (ppm) Lifetime Average Dose Metrics (mg metabolized per liter of Uver) Tumor Probability... 

---