Toxicity testing in the 21 century: defining new risk assessment approaches based on perturbation of intracellular toxicity pathways

Abstract

The approaches to quantitatively assessing the health risks of chemical exposure have not changed appreciably in the past 50 to 80 years, the focus remaining on high-dose studies that measure adverse outcomes in homogeneous animal populations. This expensive, low-throughput approach relies on conservative extrapolations to relate animal studies to much lower-dose human exposures and is of questionable relevance to predicting risks to humans at their typical low exposures. It makes little use of a mechanistic understanding of the mode of action by which chemicals perturb biological processes in human cells and tissues. An alternative vision, proposed by the U.S. National Research Council (NRC) report Toxicity Testing in the 21(st) Century: A Vision and a Strategy, called for moving away from traditional high-dose animal studies to an approach based on perturbation of cellular responses using well-designed in vitro assays. Central to this vision are (a) "toxicity pathways" (the innate cellular pathways that may be perturbed by chemicals) and (b) the determination of chemical concentration ranges where those perturbations are likely to be excessive, thereby leading to adverse health effects if present for a prolonged duration in an intact organism. In this paper we briefly review the original NRC report and responses to that report over the past 3 years, and discuss how the change in testing might be achieved in the U.S. and in the European Union (EU). EU initiatives in developing alternatives to animal testing of cosmetic ingredients have run very much in parallel with the NRC report. Moving from current practice to the NRC vision would require using prototype toxicity pathways to develop case studies showing the new vision in action. In this vein, we also discuss how the proposed strategy for toxicity testing might be applied to the toxicity pathways associated with DNA damage and repair.

Figure 1. Schematic of components required to implement the new vision of toxicity testing in the 21^st century (needs permission from NRC TT21C report).

The core components of the TT21C vision relate to testing for biological activity of compounds in toxicity pathway assays and to dose-response and extrapolation modeling using computational systems biology and pharmacokinetic tools. In the period of transition from current practice to the new vision some targeted testing in animals might be required. Targeted testing might also allow evaluation of target pathways and identification of metabolites. Assessing likely metabolism remains a challenge for the full implementation of the TT21C vision.

Figure 2. Comparison of current (A) and proposed (B) toxicity testing paradigms.

The current approach (Panel A) to setting regulatory standards involves interpretation of the most sensitive end point observed in animal studies. Low-dose extrapolation requires obtaining a point of departure from the results of the animal studies and the use of either linear or threshold extrapolation plus application of uncertainty factors to the point of departure. Other extrapolations between species or across exposure routes are sometime conducted with pharmacokinetic modeling of the tissue doses that are associated with adverse effects. A similar sequence of steps can be envisioned for setting standards based on pathway assays (Panel B). Likely hazards are determined by the sensitivity of the various toxicity pathway assays; extrapolations require assessing adverse consequences of in vitro exposures and use of computational systems biology pathway (CSBP) and pharmacokinetic modeling to set a standard for human exposure related to mg/kg/day ingested or ppm (parts per million) in the inhaled air.

Figure 3. A schematic figure illustrating progressive activation of a prototype toxicity pathway, with attendant discrete phenotypic transitions.

Text on right of figure shows proposed approaches for characterizing discrete transitions through multiple cellular phenotypes, i.e., (1) basal function, (2) minimally perturbed cellular states, (3) upregulation of adaptive, homeostatic gene batteries, and finally (4) overtly adverse states with excessive pathway perturbations. The structure of the circuitry with various embedded, nonlinear feedback loops is schematized in the middle panels. In the context of in vitro assay design, varying free concentrations of a test compound in the test media lead to increasing activation of pathway component (signaling protein) A, thereby driving pathway perturbations of signaling components B and C, and further signaling events downstream of C. Each of these steps is expected to be associated with specific alterations in gene expression, phenotypic read-out and pathway activation, identified by CSBP modeling. As the concentration increases, various portions of the network would be sequentially activated until full activation was achieved (indicated by progressively darker shading of pathway components). Full pathway activation triggers a robust response throughout the pathway circuitry, leading to adverse outcomes measurable in the cellular assay.

Figure 4. Stress response pathways and network motifs.

(A) Typical structure of a stress response pathway (adapted from Simmons et al. [26]). The so-called eight canonical stress response pathways, conserved broadly across eukaryotes, have a common structure (common motifs) for sensing damage and mounting a transcriptional response to counteract the stress. (B) Common network motifs in intracellular response pathways. Three elements, (genes/proteins) X, Y, and Z, in a pathway can regulate each other to form: (i) a negative feedback loop; (ii) a positive feedback loop; (iii) a coherent feed-forward loop, where X activates Y, and both X and Y activate Z; and (iv) an incoherent feed-forward loop, where X activates both Y and Z, but Y suppresses Z. Two transcription factors X and Y can regulate each other through, for instance: (v) a double-negative feedback loop; or (vi) a double-negative feedback loop with positive autoregulation. Sharp arrows denote activation; flat arrows denote suppression.

Figure 5. p53-mediated perturbation of DNA damage response pathways that affect mutagenesis.

A negative feedback motif composed of p53, Mdm2, and others (not shown) can produce undamped oscillations in response to radiation-induced double strand breaks. A partial AND (pAND) gate is used to indicate that to produce inheritable mutations, both DNA damage and cell proliferation (DNA replication) are required, and that the two processes likely contribute to mutagenesis in a multiplicative manner. Activation of the apoptosis pathway works to mitigate mutagenesis by killing cells with severe or unrepairable DNA damage. Pointed arrows indicate activation and blunted arrows indicate inhibition.

Figure 6. Using prototype pathways and compounds: the fast track for implementing the 2007 NRC TT21C vision.

The next step in moving forward will be to see the entire 2007 TT21C vision placed in practice with a group of pathways and specific prototype compounds preferentially affecting the target pathways. The use of compounds with robust animal toxicity profiles will allow a mapping of in vivo responses associated with apical end points and intermediate end points against new in vitro toxicity pathway assay results. The comparison will permit a better sense of the relationship between in vitro and in vivo responses in key pathways and the relevant estimates of risk. Prototype compounds for the DNA-damage pathway could include DNA-damaging compounds – polyaromatic hydrocarbons, formaldehyde, and alkylating agents, etc. – and compounds causing DNA damage indirectly through oxidative stress pathways, such as flavonoids. The output of the two approaches in relation to risk estimates and estimates of safe region of exposure could also be compared to assess correspondence or lack of correspondence in the approaches.