Toward a molecular equivalent dose: Use of the medaka model in comparative risk assessment

Prompt identification of environmental carcinogens is crucial to safeguarding human health. Two-year rodent bioassays are used to determine carcinogenic-risk in humans. Risk estimates are predominantly based on linear extrapolations of high dose tumor data that may not accurately represent risk at environmentally relevant doses. Incorporation of mechanistic data to risk assessment has provided valuable insights into dose-response relationships. Small fish carcinogenicity models are cost-effective alternatives to rodent models. New tools are necessary to aid in the extrapolation of cancer data between species of divergent phylogenies. Demonstration of equivalent molecular and biochemical responses between species will facilitate species-species extrapolations, reduce default assumptions, and validate non-mammalian models in carcinogen risk assessment.;Data provided in this body of work addresses the "molecular equivalent dose" concept in carcinogen risk assessment. In chapter 2, we hypothesized that the molecular dose for rats exposed to the alkylating carcinogen, dimethylnitrosamine (DMN) would be equivalent to the molecular dose for similarly exposed medaka fish. The objective was to determine a dose extrapolation factor using molecular dosimetry data for rats and medaka exposed to the same carcinogen to facilitate future species-species comparisons. For both models, DNA adducts and mutant frequencies (MF) were measured in liver as molecular dose estimates. Adduct concentrations were determined to be similar in magnitude, whereas MFs for DMN-exposed fish were up to 20x higher than that observed for DMN-exposed rats. Although the adduct data suggested a "one-to-one" extrapolation for DMN-exposed medaka and rats, the mutant frequency (MF) data for fish was remarkably higher than that of rats. An extrapolation coefficient derived from DNA adduct data would not have provided an accurate estimate of DMN's mutagenic potential between these two species.;DMN is used in a rodent model of alcoholic cirrhosis. In rats, morphologic changes of DMN-induced cirrhosis are associated with expression of TGF-beta1. In Chapter 3, we hypothesized that DMN-induced hepatopathology in medaka progresses morphologically via similar TGF-beta1-dependent mechanisms as that of rats. Our objectives were (1) to determine if DMN induces a similar extent of fibrosis ("phenotype") in medaka as in the rat cirrhosis model and (2) determine if DMN-induced hepatic fibrosis in medaka has a similar TGF-beta1 dependence to that observed for rats. Histopathology suggested that DMN-induced fibrosis in medaka was associated with TGF-beta1 expression. However, fibrosis in DMN-exposed medaka was significantly less than what occurs in mammalian livers. Despite this difference, medaka may prove useful for studying the underlying molecular mechanisms of DMN-induced hepatotoxicity. Correlation of tumor data with mechanistic data will provide more accurate estimates of risk and demonstration of common biochemical alterations will strengthen the use of fish in risk assessment.;Drinking water disinfection byproducts (DBPs) are an unintended consequence of chemical disinfection with chlorine. Epidemiological studies have suggested an association between cancer in humans and the consumption of DBPs. In chapter 4, we hypothesized that the DBP, dibromonitromethane (DBNM), would be mutagenic in the medaka fish model. Our objectives were to isolate DNA adducts and determine MFs from livers of DBNM-exposed lambda transgenic medaka. 32P-postlabeling did not isolate DBNM adducts from medaka hepatic DNA and MFs in DBNM exposed medaka were statistically no different from that of control fish. These results are contrary to that reported for DBNM in vitro with the Salmonella plate-incorporation assay, which classified DBNM as mutagenic. Despite the lack of significant mutant induct in livers of DBNM exposed medaka, interaction of DBNM with sites of initial contact, such as gills and skin, could not be ruled out.