The “false-positive” conundrum in the NTP 2-year rodent cancer study database

Abstract

In 1990, Ames and Gold described a conundrum of "too many carcinogens" among chemicals tested in rodent bioassays. Their proposed nongenotoxic carcinogenic mechanism was amplification of the background mutation rate via cytotoxicity induced by high doses of the test chemicals, thereby leading to increases in reparative cellular proliferation rates. Recently, we have statistically and mechanistically analyzed the entire 594-study (470 final reports) NTP 2-year rodent cancer database to better understand the conundrum posed by Ames and Gold. Our analysis provides several lines of evidence that support the contention of Ames and Gold. First, across different routes of administration, relatively phylogenetically similar rats and mice are nonetheless discordant for the development of tumors at similar organ sites. Tumor site concordance across sex within species is higher than tumor site concordance across species. Second, many chemicals negative in the Ames test nonetheless induce tumors in either rats or mice. Third, 11 out of 58 chemicals tested by the inhalation route induce lung tumors in mice and not rats, are negative in the Ames test, and exhibit hyperplasia. In 2017, Tomasetti et al. provided evidence for the clinical relevance in humans of the Ames and Gold mechanism regarding amplification of the background mutation rate by demonstrating that the majority of human tumors result from accumulated mutations due to DNA replication errors.

142 ACGIH Threshold Limit Values® (TLV®s) established from 2008-2018 lack consistency and transparency

“…Statistical analysis is not the basis upon which NCI or NTP studies are usually criticized. 32 21 In total, these 128 chemical listings show a high degree of consistency in terms of their carcinogenicity index rating. For example, a group R rating from NTP is analogous to a group A2 rating from ACGIH and a group 2A rating from IARC.…”

Rodent 2-year cancer bioassays and in vitro and in vivo genotoxicity tests insufficiently predict risk or model development of human carcinomas

“…Our research group has extensively examined the entire NTP database up through 2018. 69,85,[100][101][102][103][104] The correlation between genotoxicity and clastogenicity results and development of tumors in either rats or mice can be calculated from the data in the NTP database. In contrast, the degree of correlation between the genotoxicity test results or rodent tumor data and human cancer risk cannot be determined.…”

“…152 Inflammation can itself be mutagenic through the release of activated oxygen species but can also lead to increases in reparative cellular proliferation rates. 101 In addition to systemic inflammation, obesity is also associated with several clinical conditions characterized by localized chronic inflammation. 153 Acid reflux from gastroesophageal reflux can lead to Barrett's esophagus which is strongly associated with esophageal adenocarcinoma.…”

“…In our previous analysis of the entire NTP database 69,85,[100][101][102][103][104] and of the most commonly employed in silico techniques for the estimation of carcinogenicity from chemical structure, 85,104 we determined that structural elements of carcinogenicity strongly correlated with ordinal ranks of tumorigenicity. 85,104 Each of the 470 chemicals for which final technical reports were available reported results for male rats, female rats, male mice, and female mice.…”

An approximated one-quarter of IARC Group 3 (unclassifiable) chemicals fit more appropriately into IARC Group 4 (probably not carcinogenic)

“…Recently, our group analyzed the entire National Toxicology Program (NTP) 2-year rodent cancer study database. [6][7][8][9][10][11] Of the 482 chemicals tested by NTP that resulted in the writing of a final report or were described in the two additional Report on Carcinogens (RoC) reports, 331 were negative in the Ames Salmonella mutagenicity test. Of the 331 Amesnegative chemicals, 204 chemicals induced tumors in either rats or mice (62%).…”