Sections 2.1-2.3 summarize the weight of evidence that links specific structural fragments (toxicophores) to adverse outcomes. Literature evidence linking overall molecular properties and adverse outcomes is less prevalent. One approach is to examine the link between overall physicochemical parameters and toxic endpoints and has been the subject of an excellent recent review which discusses the physico-chemical parameters commonly associated with toxicity endpoints . Lipophili-city , usually described by the octanol-water partition coefficient, (log P) or the calculated value, (clog P) appears to be the most general requirement in quantitative structure-toxicity relationships (QSTRs) reflective of the link between lipophilicity and access to tissue and cell compartments through membrane permeability and lipophilicity driven non-specific binding to hydrophobic protein sites or membranes. This paper also highlights the literature evidence linking lipophili-city to metabolism, including binding to and inhibition of CYP450 enzymes. Other physicochemical parameters linked to toxicity outcomes include bond dissociation energies (e.g. the homolytic cleavage of a phenolic O-H bond) to produce free radicals [62,63], HOMO-LUMO energy gap and phototoxicity [62,64], the link between oxidation potential and aromatic amine carcinogenicity [62,65] and the association of lipophilic basic compounds with phospholipidosis . Struc-ture-toxicity relationships associated with mitochondrial toxins have been recently reviewed. Mitochondrial uncoupling mechanisms are common toxic pathways: for example, a correlation of mitochondrial toxicity with pKa and clog P for phenolic compounds is observed . A recent paper discusses QSTRs of a series of nonsteroidal anti-inflammatory agents (NSAIDs); a clear correlation between lipo-philicity and cytotoxicity in rat hepatocytes was established. In addition, for the benzoic acid class of NSAID, a correlation of HOMO-LUMO gap with cytotox-icity was also noted, probably reflective of the ease of oxidation of the diphenyla-mine template of many compounds in this class . This paper highlights the use of QSTR analysis on subsets of compounds within a class to establish parameters correlating with cytotoxic outcomes and subsequent use of this information to link to possible chemical mechanisms of toxicity.
Was this article helpful?