EVALUATION OF CELL VIABILITY, MIXED-FUNCTION OXIDASE ACTIVITY, METALLOTHIONEIN INDUCTION, AND GENOTOXICITY IN RAINBOW-TROUT HEPATOCYTES EXPOSED TO INDUSTRIAL EFFLUENTS .2. VALIDATION OF THE RAINBOW-TROUT HEPATOCYTE MODEL FOR ECOTOXICITY TESTING OF INDUSTRIAL WASTE-WATER

Primary cultures of rainbow trout hepatocyte are proposed as an altern ative to the rainbow trout bioassay for evaluating mortality, hepatic metallothionein and cytochrome P4501A activity inductions, and hepatic DNA damage with industrial effluents. The validation study sought to assess the performance of the hepatocyte model (HM) in terms of specif icity, predictive value, and concordance. In addition, correlation and artificial neural network analysis were used to model the cell system response compared with responses obtained in trout. The HM exhibited an overall sensitivity of 90% for detecting the various effects, sugge sting that most of the time this method was able to detect effluents t hat were toxic to trout. The specificity ranged between 68 and 89%, in dicating that hepatocytes were also able to confirm the absence of eff ects in effluent-exposed trout most of the time, but that in some case s the HM gave false positives, particularly when effects were measured at a concentration greater than 22%. The predictive values showed a s imilar range (i.e., 67-93%), suggesting that the HM was generally pred ictive of fish toxicity. The overall concordance ranged between 79 and 91%, indicating that responses obtained with the HM were consistent w ith the effects measured in effluent-exposed trout. The lower percenta ges obtained for specificity and predictive value can be explained by the fact that with some effluents the HM seemed to be more sensitive t han the trout assay, since it displayed toxic effects even when none w ere detected in trout. In effluents that were concordant, a statistica lly significant linear regression model was derived so that trout toxi city/effect end points could be predicted from those obtained with the cell system. This validation study suggests that the rainbow trout he patocyte model can be used as an alternative testing procedure to the rainbow trout assay. The cell system can be used as a prescreening too l to distinguish effluents that are likely toxic to fish from those th at are not. (C) 1997 by John Wiley & Sons, Inc.