Commercial polychlorinated biphenyls (PCBs) and environmental extracts contain complex mixtures of congeners that can be unequivocally identified and quantitated. Some PCB mixtures elicit a spectrum of biochemical and toxic responses in humans and laboratory animals and many of these effects resemble those caused by 2,3,7,8-tetrachlorodibenzo-p-dioxin (TCDD) and related halogenated aromatic hydrocarbons, which act through the aryl hydrocarbon (Ah)-receptor signal transduction pathway. Structure-activity relationships developed for PCB congeners and metabolites have demonstrated that several structural classes of compounds exhibit diverse biochemical and toxic responses. Structure-toxicity studies suggest that the coplanar PCBs, namely, 3,3',4,4'-tetrachlorobiphenyl (tetraCB), 3,3',4,4',5-pentaCB, 3,3',4,4',5,5'-hexaCB, and their monoortho analogs are Ah-receptor agonists and contribute significantly to the toxicity of the PCB mixtures. Previous studies with TCDD and structurally related compounds have utilized a toxic equivalency factor (TEF) approach for the hazard and risk assessment of polychlorinated dibenzo-p-dioxin (PCDD) and polychlorinated dibenzofuran (PCDF) congeners in which the TCDD or toxic

TEQ = Σ([PCDF_i x TEF_i]_n) + Σ([PCDD_i x TEF_n]_n]

equivalent (TEQ) of a mixture is related to the TEFs and concentrations of the individual (i) congeners as indicated in the equation (note: n = the number of congeners). Based on the results of quantitative structure-activity studies, the following TEF values have been estimated by making use of the data available for the coplanar and monoortho coplanar PCBs: 3,3',4,4',5-pentaCB, 0.1; 3,3',4,4',5,5'-hexaCB, 0.05; 3,3',4,4'-tetraCB, 0.01; 2,3,3',4,4'-pentaCB, 0.001; 2,3',4,4',5-pentaCB, 0.0001; 2,3,3',4,4',5-hexaCB, 0.0003; 2,3,3',4,4',5'-hexaCB, 0.0003; 2',3,4,4',5-pentaCB, 0.00005; and 2,3,4,4',5-pentaCB, 0.0002. Application of the TEF approach for the risk assessment of PCBs must be used with considerable caution. Analysis of the results of laboratory animal and wildlife studies suggests that the predictive value of TEQs for PCBs may be both species- and response-dependent because both additive and nonadditive (antagonistic) interactions have been observed with PCB mixtures. In the latter case, the TEF approach would significantly overestimate the toxicity of a PCB mixture. Analysis of the rodent carcinogenicity data for Aroclor 1260 using the TEF approach suggests that this response is primarily Ah-receptor-independent. Thus, risk assessment of PCB mixtures that uses cancer as the endpoint cannot solely utilize a TEF approach and requires more quantitative information on the individual congeners contributing to the tumor-promoter activity of PCB mixtures.