A Way Forward Commentary
Published: December 9, 2007
The concept of toxicity pathways as an organizing principle in hazard assessment has been debated for many years. A recent recommendation from the National Research Council (NRC) to consider the potential of the toxicity pathway concept in the future of toxicity testing comes at a time when there is an intense practical need for predictive models for effects of a wide array of chemicals on endpoints such as skin sensitization and reproduction. A significant conceptual attribute of toxicity pathways (Figures 1 & 2) is that it is possible to illustrate and understand the importance of chemistry and chemical interactions, metabolism, molecular initiating events, and the biological processes involved in producing an adverse outcome(s) meaningful to risk assessment. The biological processes within a toxicity pathway can be organized to represent the cascade of events that occur across biological levels of organization such that measures at the molecular, cell, tissue, organism, and population levels can be distinguished and effectively related to one another. This allows a prediction of likely adverse outcomes based on an understanding of discrete molecular initiating events.
Although the toxicity pathway concept is intuitively appealing, some have questioned how applicable it may be to understanding the risks of chemicals. For example, at an international workshop in Setubal, Portugal (2002) on the validation and regulatory acceptance of QSAR, one participant commented that there are thousands of pathways leading to reproductive impairment, and that it was not viable to propose to identify and model all of those involved in any adverse outcome. This type of reaction is, perhaps, not surprising since the historical approach to chemical risk assessment has focused on collection of large amounts of data for a few chemicals, with an emphasis on aspects of their unique rather than similar properties in terms of biological responses. However, we feel that it is possible to simplify even complex biological responses such as skin sensitization or reproduction such that a relative handful of predictive models can effectively focus and prioritize whole animal testing.
Recent regulatory activities associated with the REACH (Registration, Evaluation, Authorization, and restriction of CHemicals) legislation in Europe highlight the immediate need for predictive approaches to chemical risk assessment. A toxicity pathways-based approach was recently used to simplify the identification of skin/lung-sensitizing chemicals, thereby providing a basis for eliminating one of the more costly tests associated with REACH (Schultz, et al., 2006). In this case, skin sensitization has been associated with a common molecular initiating event of chemical reaction with proteins to form protein adducts. Although this molecular initiating event can be a challenge to model fully, it is probable that reliable QSAR models will be available by 2009 even before in vitro methods are thoroughly validated.
Chemical impacts on reproduction are far more complex than skin sensitization. The hypothalamic-pituitary-gonadal (HPG) axis, which controls reproduction in vertebrates, is comprised of multiple cell types in several different tissues that interact through a variety of signaling molecules and feedback mechanisms. Nonetheless, considering reproductive toxicity from a pathways perspective is useful in simplifying the system and developing appropriate predictive models.
In the case of reproduction, the translation of molecular initiating events into adverse outcomes can be aided by the use of systems biology or toxicology models (Figure 2). Systems models enable consideration of the complex HPG axis in a holistic fashion, whereby interactions between genes, proteins, and biochemical metabolites result in emergent properties that ultimately control key aspects of reproductive development and reproduction. Villeneuve, et al. (2007) described a conceptual systems model for the HPG axis in small fish as a basis for understanding and interpreting genomic data from studies with HPG-active chemicals. Aspects of this understanding include the identification of key nodes in the HPG axis in terms of molecular initiating events and control processes that modulate adverse outcomes. For example, although there are dozens, perhaps hundreds, of genes and proteins associated with the HPG axis that theoretically could be affected by chemicals, in reality only a handful of these are likely to be biologically significant targets for toxicants in terms of reproductive impacts.
This is a function both of the ability of a toxicant to interact with a biomolecule, and the importance of the biomolecule in affecting function of the HPG axis. For example, in the axis there are reactions which are not rate-limiting and, hence, would be less susceptible to perturbation than those that are rate-limiting. In addition, HPG function can be maintained or restored through any of a number of adaptive/compensatory responses. Use of systems biology to identify and understand these types of processes can help identify those initiating events most likely to produce adverse outcomes, thereby focusing developing of appropriate in vitro assays and predictive QSAR models.
A systems-based approach to understanding chemicals that affect HPG function also enables identification of biological “choke points” via which reproduction can/will be affected, irrespective of the exact mechanism of action of a chemical (i.e., the molecular initiating event). An example of this comes from studies with small fish (fathead minnows) exposed to HPG-active toxicants with markedly different mechanisms of action (Miller, et al., 2007). In those studies, chemicals with four different mechanisms of action were shown to cause reproductive toxicity (decrease egg production) through a common biological response, inhibition of the production of egg yolk protein. In this instance, the use of a systems-based approach to understanding effects of the different chemicals enabled identification of a common pathway via which toxicity was manifested, thereby essentially simplifying the system in terms of predictive modeling.
In summary, we feel that the convergence of toxicity pathways with systems biology provides the basis for identifying primary vulnerabilities for disturbance from chemicals, even in relatively complex systems. Systems models can be used to narrow the number of molecular initiating events that should be represented by in vitro assay systems and QSAR models. With appropriate in vitro assays, QSAR models can be developed with no whole animal testing. In addition, through a systems-based approach, it is possible to identify linkages between molecular initiating events and subsequent biological outcomes, giving rise to strategic testing approaches which can optimize animal use, even when animals are required for testing.
©2007 Gilman Veith & Gerald Ankley
Schultz, T.W., Carlson, R.E., Cronin, M.T.D., Hermens, J.L.M., Johnson, R., O’Brien, P.J., Roberts, D.W., Siraki, A., Wallace, K.B. & Veith, G.D. (2006). A conceptual framework for predicting the toxicity of reactive chemicals: modeling soft electrophilicity. SAR QSAR Environ. Res. 17, 1-16.
Villeneuve, D.L., Larkin, P., Knoebl, I., Miracle, A.L., Kahl, M.D., Jensen, K.M., Makynen, E.A., Durhan, E.J., Carter, B.J., Denslow, N.D. & Ankley, G.T. (2007). A graphical systems model to facilitate hypothesis-driven ecotoxicogenomics research on the brain-pituitary-gonadal axis. Environ. Sci. Technol. 40, 321-330.