Identification of a Critical Dose Level for Risk Assessment: Developments in Benchmark Dose Analysis of Continuous Endpoints

Approach toward In Vitro-Based Human Toxicity Effect Factors for the Life Cycle Impact Assessment of Inhaled Low-Solubility Particles

Today's scarcity of animal toxicological data for nanomaterials could be lifted by substituting in vivo data with in vitro data to calculate nanomaterials' effect factors (EF) for Life Cycle Assessment (LCA). Here, we present a step-by-step procedure to calculate in vitro-to-in vivo extrapolation factors to estimate human Benchmark Doses and subsequently in vitro-based EFs for several inhaled nonsoluble nanomaterials. Based on mouse data, the in vitro-based EF of TiO₂ is between 2.76 · 10^-4 and 1.10 · 10^-3 cases/(m²/g·kg intake), depending on the aerodynamic size of the particle, which is in good agreement with in vivo-based EFs (1.51 · 10^-4-5.6 · 10^-2 cases/(m²/g·kg intake)). The EF for amorphous silica is in a similar range as for TiO₂, but the result is less robust due to only few in vivo data available. The results based on rat data are very different, confirming the importance of selecting animal species representative of human responses. The discrepancy between in vivo and in vitro animal data in terms of availability and quality limits the coverage of further nanomaterials. Systematic testing on human and animal cells is needed to reduce the variability in toxicological response determined by the differences in experimental conditions, thus helping improve the predictivity of in vitro-to-in vivo extrapolation factors.

Combined in vitro-in vivo dosimetry enables the extrapolation of in vitro doses to human exposure levels: A proof of concept based on a meta-analysis of in vitro and in vivo titanium dioxide toxicity data

Evaluating the potential risks of nanomaterials on human health is fundamental to assure their safety. To do so, Human Health Risk Assessment (HHRA) relies mostly on animal studies to provide information about nanomaterials toxicity. The scarcity of such data, due to the shift of the nanotoxicology field away from a phenomenological, animal-based approach and towards a mechanistic understanding based on in vitro studies, represents a challenge for HHRA. Implementing in vitro data in the HHRA methodology requires an extrapolation strategy; combining in vitro dosimetry and lung dosimetry can be an option to estimate the toxic effects on lung cells caused by inhaled nanomaterials. Since the two dosimetry models have rarely been used together, we developed a combined dosimetry model (CoDo) that estimates the air concentrations corresponding to the in vitro doses, extrapolating in this way in vitro doses to human doses. Applying the model to a data set of in vitro and in vivo toxicity data about titanium dioxide, we demonstrated CoDo's multiple applications. First, we confirmed that most in vitro doses are much higher than realistic human exposures, considering the Swiss Occupational Exposure Limit as benchmark. The comparison of the Benchmark Doses (BMD) extrapolated from in vitro and in vivo data, using the surface area dose metric, showed that despite both types of data had a quite wide range, animal data were overall more precise. The high variability of the results may be due both to the dis-homogeneity of the original data (different cell lines, particle properties, etc.) and to the high level of uncertainty in the extrapolation procedure caused by both model assumptions and experimental conditions. Moreover, while the surface area BMDs from studies on rodents and rodent cells were comparable, human co-cultures showed less susceptibility and had higher BMDs regardless of the titanium dioxide type. Last, a Support Vector Machine classification model built on the in vitro data set was able to predict the BMD-derived human exposure level range for viability effects based on the particle properties and experimental conditions with an accuracy of 85%, while for cytokine release in vitro and neutrophil influx in vivo the model had a lower performance.

Using benchmark dose modeling for the quantitative risk assessment: Carbon nanotubes, asbestos, glyphosate

Benchmark dose method is one of the most famous quantitative approaches available for toxicological risks prediction. However, it is not fully clear how occupational health professionals can use it for specific workplace scenarios requiring carcinogen risk assessment. The paper explores the hypothesis that benchmark dose method allows to effectively approximate dose-response data on carcinogenic response, providing reasonable estimations of risks in the situations when a choice between more complex models is not warranted for practical purposes. Three case studies were analyzed for the agents with different levels of scientific confidence in human carcinogenicity: carbon nanotubes, amosite asbestos, and glyphosate. For each agent, a critical study was determined, and a dose-response slope factor was quantified, based on the weighted average lower bound benchmark dose. The linear slope factors of 0.111 lifetime excess cases of lung carcinoma per mg/m³ of MWCNT-7 (in rats exposure equivalent), 0.009 cases of mesothelioma per f/cc-years of cumulative exposure to amosite asbestos, and 0.000094 cases of malignant lymphoma per mg/kg/day of glyphosate (in mice equivalent) were determined. The correlations between the proposed linear predictive models and observed data points were R = 0.96 (R² = 0.92) for carbon nanotubes, R = 0.97 (R² = 0.95) for amosite asbestos, and R = 0.89 (R² = 0.79) for glyphosate. In all three cases, the linear extrapolation yielded comparable level of risk estimations with the "best fit" nonlinear model; for nanoparticles and amosite asbestos, linear estimations were more conservative. By performing a simulation study, it was demonstrated that a weighted average benchmark dose expressed the highest correlation with multistage and quantal-linear models.

A Review of Recent Advances in Benchmark Dose Methodology

In this review, recent methodological developments for the benchmark dose (BMD) methodology are summarized. Specifically, we introduce the advances for the main steps in BMD derivation: selecting the procedure for defining a BMD from a predefined benchmark response (BMR), setting a BMR, selecting a dose-response model, and estimating the corresponding BMD lower limit (BMDL). Although the last decade has shown major progress in the development of BMD methodology, there is still room for improvement. Remaining challenges are the implementation of new statistical methods in user-friendly software and the lack of consensus about how to derive the BMDL.

Dose-Related Severity Sequence, and Risk-Based Integration, of Chemically Induced Health Effects

Risk assessment of chemical hazards is typically based on single critical health effects. This work aims to expand the current approach by characterizing the dose-related sequence of the development of multiple (lower- to higher-order) toxicological health effects caused by a chemical. To this end a "reference point profile" is defined as the relation between benchmark doses for considered health effects, and a standardized severity score determined for these effects. For a given dose of a chemical or mixture the probability for exceeding the reference point profile, thereby provoking lower- to higher-order effects, can be assessed. The overall impact at the same dose can also be derived by integrating contributions across all health effects following severity-weighting. In its generalized form the new impact metric relates to the probability of response for the most severe health effects. Reference points (points of departure) corresponding to defined levels of response can also be estimated. The proposed concept, which is evaluated for dioxin-like chemicals, provides an alternative for characterizing the low-dose region below the reference point for a severe effect like cancer. The shape and variability of the reference point profile add new dimensions to risk assessment, which for example extends the characterization of chemical potency, and the concept of acceptable effect sizes for individual health effects. Based on the present data the method shows high stability at low doses/responses, and is also robust to differences in severity categorization of effects. In conclusion, the novel method proposed enables risk-based integration of multiple dose-related health effects. It provides a first step towards a more comprehensive characterization of chemical toxicity, and suggests a potential for improved low-dose risk assessment.

Chimeric mice with humanized liver as a model for testing organophosphate and carbamate pesticide exposure

BACKGROUND

Diagnosis of acute intoxication with organophosphate (OP) or carbamate (CM) pesticides in humans is achieved by measuring plasma butyrylcholinesterase (BuChE) activity. However, BuChE activity is not an ideal biomarker in experimental animal models. The aim of this study was to establish an experimental mouse model for evaluating exposure to OP and CM pesticides by monitoring BuChE activity using chimeric mice in which the liver was reconstituted with human hepatocytes.

RESULTS

A single oral administration of acephate (300 mg/kg), chlorpyrifos (10 mg/kg), fenobucarb (300 mg/kg) or molinate (250 mg/kg) in chimeric mice led to inhibition of >95%, > 95%, 28% and 60% of plasma BuChE activity after 7, 0.5, 0.5 and 7 h, respectively. Dose-dependent decreases in plasma BuChE activity were also observed for acephate and chlorpyrifos. A 5-day repeated-dose study with 10 or 30 mg/kg acephate found a constitutive reduction in plasma BuChE activity to 80% and 70% of pre-dose levels, respectively.

CONCLUSION

Changes in plasma BuChE activity in chimeric mice with humanized liver clearly reflected the exposure levels of OP and CM pesticides. These results suggest that the humanized-liver mouse model may be suitable for estimating levels of exposure to these pesticides in humans. © 2017 The Authors. Pest Management Science published by John Wiley & Sons Ltd on behalf of Society of Chemical Industry.

Comparison of Points of Departure for Health Risk Assessment Based on High-Throughput Screening Data

BACKGROUND

The National Research Council's vision for toxicity testing in the 21st century anticipates that points of departure (PODs) for establishing human exposure guidelines in future risk assessments will increasingly be based on in vitro high-throughput screening (HTS) data.

OBJECTIVES

The aim of this study was to compare different PODs for HTS data. Specifically, benchmark doses (BMDs) were compared to the signal-to-noise crossover dose (SNCD), which has been suggested as the lowest dose applicable as a POD.

METHODS

Hill models were fit to > 10,000 in vitro concentration-response curves, obtained for > 1,400 chemicals tested as part of the U.S. Tox21 Phase I effort. BMDs and lower confidence limits on the BMDs (BMDLs) corresponding to extra effects (i.e., changes in response relative to the maximum response) of 5%, 10%, 20%, 30%, and 40% were estimated for > 8,000 curves, along with BMDs and BMDLs corresponding to additional effects (i.e., absolute changes in response) of 5%, 10%, 15%, 20%, and 25%. The SNCD, defined as the dose where the ratio between the additional effect and the difference between the upper and lower bounds of the two-sided 90% confidence interval on absolute effect was 1, 0.67, and 0.5, respectively, was also calculated and compared with the BMDLs.

RESULTS

The BMDL₄₀, BMDL₂₅, and BMDL₁₈, defined in terms of extra effect, corresponded to the SNCD_1.0, SNCD_0.67, and SNCD_0.5, respectively, at the median. Similarly, the BMDL₂₅, BMDL₁₇, and BMDL₁₃, defined in terms of additional effect, corresponded to the SNCD_1.0, SNCD_0.67, and SNCD_0.5, respectively, at the median.

CONCLUSIONS

The SNCD may serve as a reference level that guides the determination of standardized BMDs for risk assessment based on HTS concentration-response data. The SNCD may also have application as a POD for low-dose extrapolation.

Animal-Free Chemical Safety Assessment

The exponential growth of the Internet of Things and the global popularity and remarkable decline in cost of the mobile phone is driving the digital transformation of medical practice. The rapidly maturing digital, non-medical world of mobile (wireless) devices, cloud computing and social networking is coalescing with the emerging digital medical world of omics data, biosensors and advanced imaging which offers the increasingly realistic prospect of personalized medicine. Described as a potential “seismic” shift from the current “healthcare” model to a “wellness” paradigm that is predictive, preventative, personalized and participatory, this change is based on the development of increasingly sophisticated biosensors which can track and measure key biochemical variables in people. Additional key drivers in this shift are metabolomic and proteomic signatures, which are increasingly being reported as pre-symptomatic, diagnostic and prognostic of toxicity and disease. These advancements also have profound implications for toxicological evaluation and safety assessment of pharmaceuticals and environmental chemicals. An approach based primarily on human in vivo and high-throughput in vitro human cell-line data is a distinct possibility. This would transform current chemical safety assessment practice which operates in a human “data poor” to a human “data rich” environment. This could also lead to a seismic shift from the current animal-based to an animal-free chemical safety assessment paradigm.

The adverse outcome pathway for rodent liver tumor promotion by sustained activation of the aryl hydrocarbon receptor

An Adverse Outcome Pathway (AOP) represents the existing knowledge of a biological pathway leading from initial molecular interactions of a toxicant and progressing through a series of key events (KEs), culminating with an apical adverse outcome (AO) that has to be of regulatory relevance. An AOP based on the mode of action (MOA) of rodent liver tumor promotion by dioxin-like compounds (DLCs) has been developed and the weight of evidence (WoE) of key event relationships (KERs) evaluated using evolved Bradford Hill considerations. Dioxins and DLCs are potent aryl hydrocarbon receptor (AHR) ligands that cause a range of species-specific adverse outcomes. The occurrence of KEs is necessary for inducing downstream biological responses and KEs may occur at the molecular, cellular, tissue and organ levels. The common convention is that an AOP begins with the toxicant interaction with a biological response element; for this AOP, this initial event is binding of a DLC ligand to the AHR. Data from mechanistic studies, lifetime bioassays and approximately thirty initiation-promotion studies have established dioxin and DLCs as rat liver tumor promoters. Such studies clearly show that sustained AHR activation, weeks or months in duration, is necessary to induce rodent liver tumor promotion--hence, sustained AHR activation is deemed the molecular initiating event (MIE). After this MIE, subsequent KEs are 1) changes in cellular growth homeostasis likely associated with expression changes in a number of genes and observed as development of hepatic foci and decreases in apoptosis within foci; 2) extensive liver toxicity observed as the constellation of effects called toxic hepatopathy; 3) cellular proliferation and hyperplasia in several hepatic cell types. This progression of KEs culminates in the AO, the development of hepatocellular adenomas and carcinomas and cholangiolar carcinomas. A rich data set provides both qualitative and quantitative knowledge of the progression of this AOP through KEs and the KERs. Thus, the WoE for this AOP is judged to be strong. Species-specific effects of dioxins and DLCs are well known--humans are less responsive than rodents and rodent species differ in sensitivity between strains. Consequently, application of this AOP to evaluate potential human health risks must take these differences into account.

A model for aryl hydrocarbon receptor-activated gene expression shows potency and efficacy changes and predicts squelching due to competition for transcription co-activators

A stochastic model of nuclear receptor-mediated transcription was developed based on activation of the aryl hydrocarbon receptor (AHR) by 2,3,7,8-tetrachlorodibenzodioxin (TCDD) and subsequent binding the activated AHR to xenobiotic response elements (XREs) on DNA. The model was based on effects observed in cells lines commonly used as in vitro experimental systems. Following ligand binding, the AHR moves into the cell nucleus and forms a heterodimer with the aryl hydrocarbon nuclear translocator (ARNT). In the model, a requirement for binding to DNA is that a generic coregulatory protein is subsequently bound to the AHR-ARNT dimer. Varying the amount of coregulator available within the nucleus altered both the potency and efficacy of TCDD for inducing for transcription of CYP1A1 mRNA, a commonly used marker for activation of the AHR. Lowering the amount of available cofactor slightly increased the EC50 for the transcriptional response without changing the efficacy or maximal response. Further reduction in the amount of cofactor reduced the efficacy and produced non-monotonic dose-response curves (NMDRCs) at higher ligand concentrations. The shapes of these NMDRCs were reminiscent of the phenomenon of squelching. Resource limitations for transcriptional machinery are becoming apparent in eukaryotic cells. Within single cells, nuclear receptor-mediated gene expression appears to be a stochastic process; however, intercellular communication and other aspects of tissue coordination may represent a compensatory process to maintain an organism’s ability to respond on a phenotypic level to various stimuli within an inconstant environment.

The point of transition on the dose-effect curve as a reference point in the evaluation of in vitro toxicity data

Dose-effect evaluation is an increasingly important step of health risk assessment. The foreseen increase of in vitro methods argues for the development and evaluation of a clearly defined reference points for dose-effect modelling of in vitro data. In the present study, the traditional use of a concentration corresponding to 10% or 50% of the maximal effect (EC₁₀ or EC₅₀) is compared with a strategy, under which, a reference point (Benchmark dose, BMD(T) ) is calculated that represents the dose where the slope of the dose-effect curve changes the most (per unit log-dose) in the low dose region. To illustrate the importance of the reference point, dose-effect data on CYP1A1 enzyme activity for 30 polychlorinated biphenyl (PCB) congeners were evaluated in order to calculate relative potencies, in relation to 2,3,7,8-TCDD, with confidence intervals (CIs). The present study shows that the interpretation of the results as potency and rank orders potentially depends on the choice and definition of the reference point (BMD(T) , EC₁₀ or EC₅₀). This is important as potency ranking may be used as a method for screening and prioritization, in research, in policy development or in pharmaceutical development. The use of the BMD(T) implies a focus on the change of structure in the parameter's dose-response rather than a particular percentage change in the response in such a parameter. In conclusion, the BMD(T) may be used as an alternative base for evaluation of dose-effect relationships in vitro. It offers an objective geometrical definition of a reference point in the low-dose region of the dose-effect curve.

Variation in benchmark dose (BMD) and the 95% lower confidence limit of benchmark dose (BMDL) among general Japanese populations with no anthropogenic exposure to cadmium

BACKGROUND

The use of benchmark dose (BMD) and the 95% lower confidence limit of benchmark dose (BMDL) have been gaining popularity not only in experimental studies but also in epidemiological studies including those on toxicology of cadmium (Cd), a ubiquitous hazardous element in the environment. However, the reproducibility of BMD and BMDL values has seldom been examined.

OBJECTIVES

This study was initiated to determine whether consistent BMD and BMDL values are obtained for similar non-exposed populations, i.e., the populations with no anthropogenic exposure to Cd in a single nation of Japan.

METHODS

Cd (an exposure marker), α(1)-microglobulin (α(1)-MG), β(2)-microglobulin (β(2)-MG) and N-acetyl-β-D-glucosaminidase (NAG) (three effect markers of tubular dysfunction) levels in the urine of adult Japanese women from five previous publications of this study group were examined. Overall, data were available for 17,375 cases (in 16 prefectures) regarding Cd, α(1)-MG and β(2)-MG, and 6,409 cases (in ten prefectures) regarding NAG. The data were used to calculate BMD and BMDL values taking advantage of the hybrid approach (Budtz-Jǿrgensen et al. in Biometrics 57:698-706, 2001). It was possible to calculate BMD and BMDL values for α(1)-MG and β(2)-MG for all of the 16 prefectures with 17,375 cases, whereas the values for NAG were successfully calculated for nine prefectures with 5,843 cases.

RESULTS

The application gave BMD values of 1.92, 2.46 and 2.32 μg Cd/g cr for α(1)-MG, β(2)-MG and NAG, respectively, and BMDL values of 1.83, 2.32 and 2.09 μg Cd/g cr. Large inter-prefectural variations were observed in the BMD and BMDL; there was about fourfold difference both in BMD and in BMDL calculated for α(1)-MG and β(2)-MG in 16 prefectures, and the variation was greater (i.e., by about sevenfold) in BMD and BMDL for NAG in nine prefectures. A survey of relevant literature revealed variation in BMD and BMDL values of similar folds as observed in the present analyses in five studies of Japanese populations. Multiple regression analyses taking BMD or BMDL as a dependent variable and age, CR concentration and Cd concentration as independent variables showed both BMD and BMDL were significantly influenced by Cd concentration in cases of α(1)-MG and β(2)-MG, whereas BMD and BMDL for NAG was by CR.

CONCLUSIONS

Even when the analysis was conducted in a single nation, both BMD and BMDL for the Cd effect markers varied by ca. fourfold when examining α(1)-MG or β(2)-MG and the values varied by ca. sevenfold for NAG among Cd-non-exposed populations. The most influential factors in the study population may include urine density and Cd levels in the urine.

A signal-to-noise crossover dose as the point of departure for health risk assessment

BACKGROUND

The U.S. National Toxicology Program (NTP) cancer bioassay database provides an opportunity to compare both existing and new approaches to determining points of departure (PoDs) for establishing reference doses (RfDs).

OBJECTIVES

The aims of this study were a) to investigate the risk associated with the traditional PoD used in human health risk assessment [the no observed adverse effect level (NOAEL)]; b) to present a new approach based on the signal-to-noise crossover dose (SNCD); and c) to compare the SNCD and SNCD-based RfD with PoDs and RfDs based on the NOAEL and benchmark dose (BMD) approaches.

METHODS

The complete NTP database was used as the basis for these analyses, which were performed using the Hill model. We determined NOAELs and estimated corresponding extra risks. Lower 95% confidence bounds on the BMD (BMDLs) corresponding to extra risks of 1%, 5%, and 10% (BMDL01, BMDL05, and BMDL10, respectively) were also estimated. We introduce the SNCD as a new PoD, defined as the dose where the additional risk is equal to the "background noise" (the difference between the upper and lower bounds of the two-sided 90% confidence interval on absolute risk) or a specified fraction thereof.

RESULTS

The median risk at the NOAEL was approximately 10%, and the default uncertainty factor (UF = 100) was considered most applicable to the BMDL10. Therefore, we chose a target risk of 1/1,000 (0.1/100) to derive an SNCD-based RfD by linear extrapolation. At the median, this approach provided the same RfD as the BMDL10 divided by the default UF.

CONCLUSIONS

Under a standard BMD approach, the BMDL10 is considered to be the most appropriate PoD. The SNCD approach, which is based on the lowest dose at which the signal can be reliably detected, warrants further development as a PoD for human health risk assessment.

Toward a unified approach to dose-response modeling in ecotoxicology

This study reviews dose-response models that are used in ecotoxicology. The focus lies on clarification of differences and similarities between models, and as a side effect, their different guises in ecotoxicology are unravelled. A look at frequently used dose-response models reveals major discrepancies, among other things in naming conventions. Therefore, there is a need for a unified view on dose-response modeling in order to improve the understanding of it and to facilitate communication and comparison of findings across studies, thus realizing its full potential. This study attempts to establish a general framework that encompasses most dose-response models that are of interest to ecotoxicologists in practice. The framework includes commonly used models such as the log-logistic and Weibull models, but also features entire suites of models as found in various guidance documents. An outline on how the proposed framework can be implemented in statistical software systems is also provided.