Discriminative power of an assay for automated in vitro screening of teratogens.

Screening for potential teratogenicity of 20 test compounds was performed using a computerised microscope workstation for determination of cytotoxicity, proliferation and morphology of fibroblastoid murine L929-cells. The test compounds, which were divided into four classes according to teratogenicity were: 5-bromo-2(')-deoxyuridine, 6-aminonicotinamide, acrylamide, boric acid, D-(+)-camphor, dimethadione, dimethyl phthalate, diphenhydramine, hydroxyurea, isobutyl-ethyl-valproic acid, lithium chloride, methyl mercury chloride, methotrexate, methoxyacetic acid, penicillin G, all-trans-retinoic acid, pentyl-4-yn-valproic acid, saccharin, salicylic acid and valproic acid. All compounds, with the exception of dimethadione inhibited proliferation in a linear dose-dependent manner, and there were statistically significant compound class-dependent differences between the IC(50)-values for the compounds (p<0.0374), the strongest teratogens being the most potent. Furthermore, the average efficacies (maximum relative change) for 10 parameters describing cell morphology exhibited statistically significant compound class-dependent differences (p<0.0001), the class I and II compounds exhibiting significantly lower efficacies than the class III and IV compounds (p<0.01). Thus, test compounds affected both the proliferation and morphology of L-cells in manner demonstrating a general relationship with the teratogenic potency of the compounds. However, the moderate teratogens dimethadione and lithium chloride only had minor effects on the morphology and proliferation of the cells whereas the non-teratogen diphenhydramine had effects on both proliferation and morphology comparable to the strong teratogens.