ELECTROPHILICITY AS MEASURED BY K(E) - MOLECULAR DETERMINANTS, RELATIONSHIP WITH OTHER PHYSICAL-CHEMICAL AND QUANTUM-MECHANICAL PARAMETERS, AND ABILITY TO PREDICT RODENT CARCINOGENICITY

This paper analyzes electrophilicity data as measured by the K(e) system for 205 chemicals including both rodent carcinogens and non-carcinogens. Multivariate statistical methods were used. The analysis identified atoms and substructures contributing to electrophilicity, and permitted to establish a theoretical method by which the K(e) value (electrophilicity) of chemicals can be easily estimated. In a subset of chemicals, the K(e) parameter was compared with other physical-chemical and quantum mechanical properties: K(e) appeared to be mostly correlated with the energy of the lowest unoccupied molecular orbital and with the absolute electro-negativity. The role of K(e) in structure-activity studies was also investigated; in particular, a comparative analysis of the performance of K(e), Salmonella typhimurium and Ashby's structural alerts in predicting carcinogenicity was carried out. The K(e) system performed better than the other systems. However, because of the many different mechanisms underlying carcinogenesis, the K(e) system cannot predict the potential carcinogenicity of all kinds of chemicals. It is concluded that the main role of K(e) in risk assessment consists in producing a probabilistic estimate of the rodent carcinogenicity of the chemicals: e.g. a chemical with K(e) higher than 3.0 x 10(12) M-1s-1 has nearly 80% probability of being a carcinogen. Such a probability estimate can be used to rank the chemicals in a priority scale for subsequent and more detailed studies, either theoretical or experimental. In view of this, the role of our method for estimating K(e) is particularly important: it gives rapidly and at no cost a chemical classification for risk assessment and priority setting.