In vitro cytotoxicity testing: Biological and statistical significance

This study was designed to determine the potential of an in vitro model for predicting acute human chemical toxicity. Rat lung epithelial cells (L2) were tested for their ability to incorporate radiolabelled amino acids into newly synthesized proteins, in the absence or presence of increasing doses of the test chemical, during a 24-hr incubation. The MTT assay was also performed as a parallel measure of toxicity. IC(10), IC(50) and IC(75) values (10%, 50% and 75% inhibitory concentrations, respectively) were extrapolated from dose-response curves after linear regression analysis. The biological significance of the results of testing 30 chemicals shows that the experimental IC(50) values were more accurate predictors of human toxicity than equivalent toxic blood concentrations derived from rodent LD(50)s. Overall, the 24-hr protein synthesis experiments were at least as sensitive as the MTT protocol for detecting cytotoxicity. Individually, the toxicity of eight of 15 chemicals was underestimated with the MTT assay. In addition to calculating the correlation coefficient, the hypothesis test for B = 0 (zero slope) was computed for each experiment. This test, which is based on the slope of the sample regression equation, is used to determine the statistical significance of dose-response curves, yet it has not been routinely incorporated into cytotoxicity testing studies. It is anticipated that this procedure, together with a related battery of tests, may supplement or replace currently used animal protocols for human risk assessment.