Discovery of metabolic signatures for predicting whole organism toxicology.
Hines A, Staff FJ, Widdows J, Compton RM, Falciani F, Viant MR.
Toxicol Sci. 2010 Jan 11. [Epub ahead of print]
Toxicol Sci. 2010 Jan 11. [Epub ahead of print]
School of Biosciences, University of Birmingham, Edgbaston, Birmingham, B15 2TT, UK.
Toxicological studies in sentinel organisms frequently use biomarkers to assess biological effect. Development of 'omic' technologies has enhanced biomarker discovery at the molecular level, providing signatures unique to toxicant mode-of-action (MOA). However these signatures often lack relevance to organismal responses, such as growth or reproduction, limiting their value for environmental monitoring. Our primary objective was to discover metabolic signatures in chemically-exposed organisms that can predict physiological toxicity. Marine mussels (Mytilus edulis) were exposed for 7 days to 12 and 50 mug.l(-1) copper and 50 and 350 mug.l(-1) pentachlorophenol, toxicants with unique MOAs. Physiological responses comprised an established measure of organism energetic fitness, scope for growth (SFG). Metabolic fingerprints were measured in the same individuals using nuclear magnetic resonance-based metabolomics. Metabolic signatures predictive of SFG were sought using optimal variable selection strategies and multivariate regression, and then tested upon independently field-sampled mussels from rural and industrialised sites. Copper and pentachlorophenol induced rational metabolic and physiological changes. Measured and predicted SFG were highly correlated for copper (r(2)=0.55, p=2.82x10(-7)) and pentachlorophenol (r(2)=0.66, p=3.20x10(-6)). Predictive metabolites included methionine and arginine/phosphoarginine for copper, and allantoin, valine and methionine for pentachlorophenol. When tested on field-sampled animals, metabolic signatures predicted considerably reduced fitness of mussels from the contaminated (SFG=6.0 J.h(-1).g(-1)) versus rural (SFG=15.2 J.h(-1).g(-1)) site. We report the first successful discovery of metabolic signatures in chemically-exposed environmental organisms that inform on molecular MOA and that can predict physiological toxicity. This could have far-reaching implications for monitoring impacts on environmental health.
PMID: 20064833 [PubMed - as supplied by publisher]
