E. Genschow et al., Development of prediction models for three in vitro embryotoxicity tests which are evaluated in an ECVAM validation study, ALTEX-AL TI, 16(2), 1999, pp. 73-83
Development of prediction models for three in vitro embryotoxicity tests wh
ich are evaluated in an ECVAM validation study
In 1997 ZEBET started the co-ordination of a study funded by the European C
entre for the Validation of Alternative Methods (ECVAM) with the aim of pre
validation and validation of three in vitro embryotoxicity tests. These tes
ts employ the cultivation of postimplantation whole mt embryos (WEC test),
cultures of primary limb bud cells of mt embryos (micromass test, MM-Test),
and cultures of a pluripotent mouse embryonic stem cell line (embryonic st
em cell test, EST). In the current Validation Study each of the tests is ev
aluated in four laboratories under blind conditions. In an initial phase of
the validation study six out of 20 test chemicals comprising different emb
ryotoxic potential (non, weak and strong embryotoxic) were tested. The resu
lts were used to improve the prediction models (PM) for the WEC test and th
e MM-Test in order to identify the embryotoxic potential of test chemicals.
In addition, the existing PM for the EST was evaluated using the results f
rom resting of the initial six chemicals. The PM for the EST was developed
using the results of a previous prevalidation study (Scholz et al., 1999),
in which 94% of the learning sample were classified correctly. Applying thi
s PM to the results of the initial phase of the current validation study, 8
0% of the experiments were classified correctly according to the embryotoxi
c potential of the tested chemicals in vivo. Applying the PM for the MM-Tes
t and the WEC test that were developed during the current validation study
in both tests, 79% correct classifications were achieved, Since the PM of t
he WEC-Test took into account only parameters of growth and development, bu
t not cytotoxicity data, a second PM (PM2)for the WEC was developed that wa
s improved by incorporating cytotoxicity data of the differentia red mouse
fibroblast cell line 3T3 derived from the EST This approach, which has prev
iously never been used as an adjunct to the WEC test, resulted in an increa
se of correct classifications to 96%.