The preparation of anthraquinone used in the National Toxicology Program cancer bioassay was contaminated with the mutagen 9-nitroanthracene.

Commercial anthraquinone (AQ) (9,10-anthracenedione) is produced by at least three different production methods worldwide: oxidation of anthracene (AQ-OX), Friedel-Crafts technology (AQ-FC) and by Diels-Alder chemistry (AQ-DA), with the final product varying in color and purity. AQ-OX begins with anthracene produced from coal tar and different lots can contain various contaminants, particularly the mutagenic isomers of nitroanthracene. AQ has been reported to be negative in a variety of genotoxicity tests including numerous Ames Salmonella mutagenicity assays. In addition, we report that AQ-DA is negative in the Salmonella-Escherichia coli reverse mutation assays, the L5178Y mouse lymphoma forward mutation assay, for inducing chromosomal aberrations, polyploidy or endoreduplication in Chinese hamster ovary cells, and in the in vivo mouse micronucleus assay. Further, a previous 18 month bioassay conducted with AQ administered to male and female B6C3F(1) and (C57BL/6xAKR)F(1) mice reported no induction of cancer. Thus, it was somewhat unexpected that in a long-term study conducted by the National Toxicology Program (NTP) AQ-OX induced a weak to modest increase in tumors in the kidney and bladder of male and female F344/N rats and a strong increase in the livers of male and female B6C3F(1) mice. In the studies reported here, a sample of the AQ-OX used in the NTP bioassay was shown to be mutagenic in the Ames tester strains TA98, TA100 and TA1537. Addition of an S9 metabolic activation system decreased or eliminated the mutagenic activity. In contrast, the purified NTP AQ-OX as well as the technical grade samples AQ-FC and AQ-DA were not mutagenic in the Ames test. The chemical structure of AQ does not suggest that the parent compound would be DNA reactive. Therefore, a mutagenic contaminant was present in the NTP bioassay sample that is either directly mutagenic or can be activated by bacterial metabolism. Analytical studies showed that the primary contaminant 9-nitroanthracene (9-NA) was present in the NTP AQ-OX at a concentration of 1200 p.p.m., but not in the purified material. The 9-NA and any other contaminants that might have been present in the NTP AQ-OX induced measurable mutagenicity at 9-NA concentrations as low as 0.15 microg/plate in tester strain TA98, indicating potent mutagenic activity. On the basis of revertants per microgram, 9-NA was more potent than benzo[a]pyrene (B[a]P) and was about equally as potent as the 2-nitrofluorene run concurrently as positive controls. TD(50) quantitative carcinogenicity potency estimates indicate that a carcinogen of a potency in the range between B[a]P and dimethylnitrosamine would be required to produce the observed carcinogenic response at the levels of the contaminants found in the test sample. While recognizing that there are limitations in extrapolating mutagenic potency to potential carcinogenic potency, these estimates do indicate that it is plausible that the 9-NA contaminant might have been responsible for all of the tumor induction observed in the NTP study. In fact, in the absence of reliable cancer data, the genetic toxicology profile indicates that AQ would not be a genotoxic carcinogen. Thus, no conclusion as to the carcinogenic activity of AQ can be made at this time.