Development of a systems toxicology approach and its application to quantify the biological impact of tobacco smoke in vitro and in vivo
The potential health risks of long-term exposure to biologically active substances such as therapeutic drugs or environmental toxins are frequently determined a posteriori through clinical epidemiology studies. However, disease may take decades to manifest, and changes in therapeutic regime, lifestyle or exposures may impact the risk of disease onset. Moreover, correlative disease risk assessment in epidemiology is not intended to elucidate the mechanisms linking perturbations in molecular signalling to disease and thus provides fewer options for intervention. We propose to quantify the biological network perturbations caused by active substances and thereby identify mechanisms and biomarkers that are modulated in response to exposure and are related to disease onset.
Two distinct metrics were developed: NPA (network perturbation amplitude), which assesses the amplitude of signalling in a network, and BIF (biological impact factor), which aggregates the amplitude scores of multiple networks. These methods rely on the development of causal models of biological networks that include measurable downstream quantities affected by pathway perturbations. Here we present a five-step biological impact evaluation and sample applications to quantify the biological perturbations induced by whole cigarette smoke (CS).
Firstly, we studied transcriptomics in organotypic 3-dimensional cultures of human bronchial epithelium and modelled the biological network perturbations induced by CS exposure. When compared to human bronchial epithelium in vivo, a single exposure to CS in this in vitro system induced biological perturbations similar to those observed in the airway epithelium of human smokers in vivo.
Secondly, we augmented the OECD TG412 28-day inhalation study with transcriptomics and phosphoproteomics to evaluate biological perturbations by CS in vivo and to derive a quantitative BIF of various exposures. The observed histopathological end points correlated with the degree of perturbation of their associated biological networks in vivo, thus providing a powerful approach to investigate disease mechanisms in vivo.