(217c) Evaluation of the Toxicity of Nanomaterials Based On Knowledge Extraction From High Throughput Screening of Biological Toxicity Data

Authors: 
Cohen, Y. - Presenter, University of California, Los Angeles
George, S. - Presenter, Center for the Environmental Implications of Nanotechnology (CEIN)
Bradley, K. - Presenter, Center for the Environmental Implications of Nanotechnology (CEIN)
Damoiseaux, R. - Presenter, Center for the Environmental Implications of Nanotechnology (CEIN)
France, B. - Presenter, Center for the Environmental Implications of Nanotechnology (CEIN)


Nano-sized materials are increasingly utilized in many industrial products and processes, with over 1000 commercial products utilizing engineered nanomaterials (eNMs), primarily due to their unique nano-scale properties. Despite the many beneficial uses of nanomaterials, there is a growing public concern regarding the potential release of eNMs to the environment and the potential adverse impacts of exposures to eNMs. Although there are mounting studies on the toxicity of eNMs, understanding of the general principles governing eNMs toxicity potential and the long-term environmental health and safety associated with eNM containing products is in its infancy. In this regard, high throughput toxicity screening is critical for characterization of the potential hazard of eNMs, which is in turn indispensable information for subsequent risk assessment and development of environmental and health regulatory policies. In the present work an approach to knowledge extraction from high throughput screening (HTS) of nanoparticle toxicity is presented, based on RAW and BEAS-2B mammalian cell lines, focusing on identification of toxicity outcomes (i.e., “hit identification”), development of predictive nano-quantitative-structure-activity relations (nano-SARs), identification of pathway linkages and correlation of cell signaling pathways and cytotoxicity. Based on newly developed feature selection approach, a classification based nano-SAR was developed for cytotoxicity of metal and metal oxide nanoparticles. The approach was demonstrated for a library of metal and metal oxide nanoparticles to develop a classification based nano-SAR that can be used to identify decision boundaries with respect to hazard ranking of nanoparticles. In addition, self-organizing maps (SOM) analysis was performed to identify clusters corresponding to sub-lethal pro-inflammatory responses associated with ROS generation, lethal genotoxic responses, and cytotoxicity. Finally, complex network theory methods were applied to identify relationships between cell responses (pathway activation and cytotoxicity measures) as well as physicochemical properties of nanoparticles. For example, the analysis based on the above cell lines revealed a hierarchical activity-activity pattern where sub-lethal effects such as ROS generation and the intracellular Ca2+ flux are strongly related to lethal effects (e.g. cell membrane damage and cell death) and that eNM primary size, aggregation and its dissolution tendency were critical parameters affecting the observed toxicity behavior.