M4 - Oral
Strategies for Modeling Systems BiologyScotland A 10:30 am-Noon
|
| Chair(s): A. Jarabek
|
M4.2 What Fraction of Disease Can be Prevented by Removing Specific Exposures?. Cox T; Cox Associates and University of Colorado tcoxdenver@aol.com
Abstract: Epidemiology articles and textbooks often interpret population attributable fractions (PAFs) based on 2 x 2 tables or logistic regression models exposure-response associations as preventable fractions, i.e., in terms of the fraction of illnesses in a population that would be prevented if exposure were prevented. In general, this widely used causal interpretation is not correct, since statistical association is not necessarily and indication of causation. Yet, the concept of a preventable fraction is crucial in many applied risk assessments that support public health decisions and policies and that guide research on the likely human health consequences of removing specific exposures. The purpose of this paper is therefore to introduce a new concept of preventable fractions (PFs) having valid causal interpretations and to show how to calculate useful and correct estimates of PFs from the types of incomplete but useful causal information that is often available in practice.
We show how to use biologically-motivated models and applied probability theory to obtain useful bounds on preventable fractions from available biological and epidemiological data by incorporating causal-logical constraints, such as that the probability that exposure X causes response Y cannot exceed the probability that exposure X precedes response Y. More generally, we apply ideas from systems reliability theory to the uncertain causal relations in biologically-based stochastic models of carcinogenesis to construct bounds on the probabilities that specific exposures will cause specified response that would not have occurred. The practical value of the approach is demonstrated by applying it to estimate PFs for two very different classes of lung carcinogens – PAHs and cadmium compounds – that occur in cigarette smoke.
|
M4.3 What Inhibition of Gap Junction Intercellular Communication Can Tell Us About Carcinogenic Dose-response. Chrostowski P.C.; CPF Associates, Inc. pc@cpfassociates.com
Abstract: Gap Junction Intercellular Communication (GJIC) plays an important role in the progression of carcinogenesis. Cells chemically communicate through the gap junction by the migration of growth regulatory signal transducing substances that have been associated with cell cycling, cell growth, and cell death. Inhibition of GJIC has been suggested as a mechanism for the tumor promotion process on the basis of biochemical examination of tumors and through in vitro experimentation. Inhibition of GJIC is thought to cause decreased regulation of homeostatic growth control allowing a preneoplastic cell to escape the growth control of normal surrouding cells resulting in clonal expansion. Various researchers have found that non-genotoxic chemical carcinogens that are capable of blocking GJIC also test positive in their ability to promote rodent tumors thus substantiating on an in vivo basis a mechanism based on in vitro data. Numerous chemicals have shown an ability to inhibit GJIC including perfluorinated organics, polychlorinated organics, polycyclic aromatics, and phthalate esters. In vitro experimentation shows the inhibitory process to be expressed as a sigmoid dose-response curve with a marked threshold. At high doses, the dose-response curve plateaus in a region often associated with frank cytotoxicity. At lower doses, the GJIC inhibition process is reversible with cessation of exposure. The implications of GJIC for cancer dose-response quantification will be discussed and illustrated with examples for the chlorinated organic compounds, hexachlorocyclohexane and DDT. When considered as GJIC inhibitors, these compounds are substantially less potent carcinogens than when assessed using linear low dose models.
|
M4.4 Value of Information Approaches to Evaluate Models of Low Dose Radiation Effects on Neurons in the Developing Brain. Griffith W.C. *, DeFrank N.M., Gholke J.M., Gribble E.A., Faustman E.M.; University of Washington griffith@u.washington.edu
Abstract: Substantial evidence has demonstrated that low dose radiation exposures during particular gestational periods can result in permanent neuronal perturbations and eventual abnormalities in behavior and mental activity. It is hypothesized that the mechanisms underlying these effects include perturbed migration patterns of post-mitotic neurons, which lead to insufficient mature neuron production and improper positioning of neurons. Existing data sets were used in the construction of a computational model to describe the altered patterns of neuronal migration in the mouse neocortex following irradiation in utero. The subsequent effects on neuron number and differentiation status at the end of neurogenesis were determined and related to the probability of decreased learning ability. Model results indicate that radiation-induced damage among neuronal precursor cells and post-mitotic neurons leads to variable rates of neuron differentiation in a dose and time dependent manner. For example, model output indicates that the migration of post-mitotic neuronal cells is decreased by 75% following a 25cGy exposure, leading to the improper placement of nearly four million mature neurons. This model is of utility for risk assessment as it includes the evaluation of alterations in specific developmental dynamic processes, incorporates in vitro data, and permits comparisons of toxicant effects across various times and doses. A value of information approach is used to identify how future research could decrease important uncertainties in the model for purposes of risk assessment. This research is directly relevant for evaluation of low dose radiation effects. Supported by US Department of Energy Low Dose Program (DE-FG02-03ER63674), NIEHS (2-P01-ES009601, P50 ES012762, P30 ES07033, U10 ES 11387 and T32 ES07032), US EPA (RD-83170901), and the Center for the Study and Improvement of Regulation at Carnegie Mellon University and the University of Washington.
|
M4.5 Computational Modeling of Signaling Pathways Mediating Cell Cycle and Apoptotic Responses to Ionizing Radiation-Induced DNA Damage. Zhao Y*, Conolly R.B.; CIIT Centers for Health Research, Research Triangle Park, NC, USA. National Center for Computational Toxicology, U.S. EPA, Research Triangle Park, NC, USA. mzhao@ciit.org
Abstract: Insight into the shapes of dose response curves for ionizing radiation- (IR) induced cancers can be obtained by (1) computational modeling of molecular-level signaling pathways linking IR-induced DNA damage to adaptive responses such as cell cycle checkpoint controls and apoptosis and, (2) integrating these molecular-level descriptions into clonal growth models for cancer. Here we present computational modeling of the signaling pathways and adaptive responses, and preliminary results for tumor incidence prediction. The overall IR exposure-tumor response model consists of four modules: (1) IR-induced formation of reactive oxygen species (ROS), ROS-mediated formation of DNA double strand breaks (DSB), and repair of the DSB. (2) A signaling pathway that senses the DSB and links to regulatory proteins in the mammalian cell cycle and the apoptotic pathway. (3) The mammalian cell cycle including G1/S and G2/M checkpoint controls, with the description of the G2/M checkpoint due to Tyson and Novak (J. Theor. Biol, 210, 249-263, 2001). (4) A 2-stage clonal growth model for the prediction of tumor incidence. Three progressively more complex versions of modules 2 and 3 are described that illustrate how pathway architecture determines the robustness of the system. Robustness is the ability to maintain a significant level of functionality when some pathway components are missing or inoperative as, for example, when mutations occur in their related genes. The model provides qualitatively accurate descriptions of the IR-mediated activation of cell cycle checkpoints and of the apoptotic pathway, and of time-course perturbations in activities of regulatory proteins including p53 and p21. This exercise in computational modeling organized and synthesized a large amount of experimental information. The model provides insights into the dynamical behavior of the system and can be iteratively refined as new data become available. Eventually, the approach taken here will provide rigorous basis for understanding how biological adaptive processes influence the shapes of dose response curves for IR-induced cancer. This research was supported by the Office of Science (BER), U.S. Department of Energy, Grant No. DE-FG02-03ER63669 to the CIIT Centers for Health Research. This research may not necessarily reflect the views of the U.S. Environmental Protection Agency and no official endorsement should be inferred.
|
