MEIC EVALUATION OF ACUTE SYSTEMIC TOXICITY - PART III - IN-VITRO RESULTS FROM 16 ADDITIONAL METHODS USED TO TEST THE FIRST 30 REFERENCE CHEMICALS AND A COMPARATIVE CYTOTOXICITY ANALYSIS

Results from tests on the first 30 MEIC reference chemicals in 16 diff erent systems are presented as a prerequisite to the subsequent in vit ro/in vivo comparisons of acute toxicity data, i.e. the Final MEIC eva luation of all test results of the study. The study is a supplement to the previously published results from 68 methods (including methods 4 5B and 46B [old numbers]) used to test the same set of chemicals. The strategies and methods of the preceding paper were employed to enable a comparative cytotoxicity analysis of the results from these 68 metho ds and from the 16 new methods to be made. Principal components analys is (PCA) of 82 assays demonstrated a dominating first component which described as much as 83% of the variance in the toxicity data. This re markable similarity of all toxicity data was the main finding of the p resent study, and confirmed the results of the previous study with a l ess-extensive database. Also, the influence on the general variability of results of several key methodological factors was evaluated by ana lysis of selected sets of data, including linear regression of the res ults of pairs of methods, which were similar in all respects except fo r the factor under analysis. This analysis of the same 82 assays as be fore also confirmed previous results from the 68 assay database: a) th e toxicities of a third of the chemicals increased considerably with e xposure time; b) in general, cytotoxicity for human cells was well pre dicted by cytotoxicity tests with animal cells; c) this prediction was poor for two chemicals, i.e. digoxin and malathion; d) prediction of human cytotoxicity by ecotoxicological tests was only fairly good; e) 25 comparisons of similar assays employing different cell lines showed strikingly similar toxicities (mean R-2 = 0.86); f) 22 comparisons of similar pairs of assays employing different primary cultures and cell lines also revealed similar toxicities (mean R-2 = 0.79); and g) 15 c omparisons of similar assays with different growth/viability endpoint measurements demonstrated strikingly similar toxicities (mean R-2 = 0. 89). Results b, e, f and g must be the main causes of the general simi larity of results, while results a, c and d, together with other facto rs, could explain the 20% dissimilarity. These findings support the ba sal cytotoxicity concept and may assist in guiding and refining in vit ro toxicity testing in the future.