Among the different types of liver tumor, hepatocellular neoplasms predominate by far in both animals and man. Consequently, preneoplastic foci of altered hepatocytes (FAH), preceding both hepatocellular adenomas and carcinomas, represent the most prevalent form of hepatic preneoplasia observed in animals for a long time, and identified in human chronic liver diseases associated with, or predisposing to, hepatocellular carcinomas more recently. Morphological, microbiochemical, and molecular biological approaches in situ revealed striking similarities in specific changes of the cellular phenotype of preneoplastic FAH developing in experimental and human hepatocarcinogenesis, irrespective of whether this was elicited by chemicals, hormones, viruses or radiation. The advantage of using FAH for risk identification (aiming at primary cancer prevention) in long-term and medium-term carcinogenesis bioassays has been well documented, but quantitative morphometric approaches appear to be indispensable for an appropriate evaluation of both bioassays. The detection of phenotypically similar FAH in various animal models and in humans prone to develop or bearing hepatocellular carcinomas favors the extrapolation from data obtained in animals to humans. Moreover, the recently reported frequent finding of FAH in fine-needle biopsies of patients suffering from chronic liver diseases opens new perspectives for secondary prevention of human hepatocellular carcinoma.

No abstract

A rat liver foci bioassay (RLFB) based on an initiation-promotion protocol employing preneoplastic foci of altered hepatocytes (FAH) as an endpoint, was prevalidated in 5 different laboratories. FAH were identified by immunohistochemical demonstration of glutathione-S-transferase (placental form, GSTP) and by staining with hematoxilin/eosin (H&E), and their area fraction was quantified morphometrically. The four model hepatocarcinogens N-nitrosomorpholine, 2-acetylaminofluoren, phenobarbital, and clofibrate were selected according to characteristic differences in their presumed mode of action, and tested in a total of 1,600 male and female rats at 2 different dose levels. The chemicals were found to differ characteristically in their potency and dose-response relationship to induce FAH when given alone or when administered following initiation with diethylnitrosamine. The interlaboratory variation was small for results obtained with the GSTP-stain and somewhat larger with respect to H&E. The assessment of the carcinogenic potential of the four chemicals by the different laboratories was in the same range and the nature of their dose-response relationships did not differ essentially between laboratories. Our results suggest that this RLFB is a sensitive bioassay, providing potentially valuable information for risk assessment including the classification of carcinogenic chemicals according to their mode of action.