I was a member of the Carcinogenic Potency Database (CPDB) Project for about 25 years (1985-2011), longer than any of the other project members except for Lois Gold and Bruce Ames. During that time, I worked at the University of California Berkeley (UCB), Lawrence Berkeley National Laboratory, Children's Hospital Oakland Research Institute, and as an independent consultant. At various times in my career I was involved in just about every aspect of the project. As the person who wrote the HTML code for the website, I apologize for what now appears like rudimentary design: web design was not my forté and it was not much of a priority for the project.

Due in part to Lois' untimely illness and rapid decline in 2012, there was inadequate preparation for archiving of the project's records. As far as I have been able to discern, the entirety of the project's paper records were discarded, and all of the project's computer files were discarded or erased except for the website (formerly located at http://potency.berkeley.edu/, ca. 1996-2012). Lois had arranged for the National Library of Medicine's (NLM) TOXNET site (formerly located at http://toxnet.nlm.nih.gov/cpdb/, 2013-2019) to host the website but it was unceremoniously removed by NLM in 2019. The site was dropped during a transition of TOXNET to other sites, so it could be that NLM decided to do this because it did not fit with the structure of the other databases. The cost for NLM to host the CPDB website was trivial compared to its budget however, so it is unfortunate that the National Institute of Environmental Health Sciences (NLM's parent agency) did not continue to make this information available to the public, since NIEHS was one of several US government funders of the CPDB Project.

As far as significant information that was lost due to the CPDB Project's records not being archived, from my perspective these were:

1. The computer code that generated TD₅₀ values and the published plots. The statistical calculation of the TD₅₀ was documented in the first publication of the plot and in Sawyer and Peto (1984), so this could be reimplemented if someone so desired.
2. An unpublished plot of chronic carcinogenesis bioassays of complex mixtures, what we called “The Food Plot”, that actually included some non-foods. This had been developed over many years and its publication was ever-postponed. In the end, Lois and I decided that there might not be much interest in it, as many of the studies were negative. The studies in "The Food Plot" met our inclusion rules other than the fact of being complex mixtures (though a few complex mixtures were actually published by us, e.g. quillaia extract). Nonetheless, the studies were not always of the best quality (often low numbers of animals were tested, or they were tested for relatively short experiment times). The lower quality of the studies in The Food Plot may have explained the reason for the large number of negative studies.
3. I had discovered the existence of a somewhat large number of carcinogenesis bioassays on synthetic pesticides. These bioassays were conducted by chemical companies and were not externally-peer reviewed. The summary results were published by The Food and Agriculture Organization (FAO), possibly in the journal Pesticide Residues in Food: Toxicological Evaluations (it could have been one of their other journals though - I don't actually remember the name) and were used by the FAO for pesticide registration purposes. The reported pesticides had a very low rate of carcinogenicity. After reviewing the results reported by the FAO, Lois and I decided not to incorporate the information into the CPDB because there was no external peer review of the studies and because there was likely a high degree of selection bias due to the studies being proprietary and conducted by chemical companies (who did not want to go to the expense of testing potential pesticides that might ultimately be unusable because of carcinogenicity). The information from these studies might still be of use to researchers doing chemical toxicity prediction using Quantitative Structure-Activity Relationship (QSAR).

There were a few interesting tangents that our project became involved in. Following the creation of the National Center for Complementary and Integrative Health in 1991 and the passage of the Dietary Supplement Health and Education Act in 1994, our project slowly became more interested in herbal supplements.

In 2002, we became aware of reports of rapid-onset cancer in people who were accidentally given a weight loss treatment at a Belgian spa that included the herb Aristolochia, and that a chemical in Aristolochia, aristolochic acid (AA) was the likely cause (“Aristolochic acid as a probable human cancer hazard in herbal remedies: a review” by V. M. Arlt et al. Mutagenesis 2002;17:265-77). In response, we decided to include a study on AA in the CPDB that had previously been excluded due to the study's having one too few animals in the control group to meet our inclusion rules. As it turned out, this study showed that AA was among the most potent chemicals in the CPDB (See aristolochic acid). My curiosity about this was piqued, and I searched the web to see whether one could buy herbal supplements that included aristolochia on the web, and indeed it was the case. In 2003, we reported our findings first to the Center for Food Safety and Applied Nutrition at the Food and Drug Administration (FDA), and then in a letter summarizing our findings published by the New England Journal of Medicine. As a result, the FDA sent warning letters to the companies that were selling these products. At around this time, Lois and I were also threatened with a lawsuit by one of these companies alleging commercial disparagement. The lawsuit did not eventuate; we were fortunate to have had a UCB lawyer advise us.

Despite there being a ban on importation into the US of Aristolochia and substances suspected of containing AA, at the end of 2018 I discovered that at least one domestically-produced herbal supplement containing Aristolochia is still available for purchase on the web, and is even for sale on Amazon.com. I reported this to the U.S. Department of Health & Human Services' Safety Reporting Portal on January 1, 2019 (FDA is part of HHS). As of this writing, HHS has taken no action and has not responded to my complaint. From this I can only conclude this branch of HHS has been hollowed out since 2017.

For anyone interested, there is a long history of Aristolochia use dating back to antiquity ("Ancient Medicinal Use of Aristolochia: Birthwort's Tradition and Toxicity" by John Scarborough & Andrea Fernandes (2011) Pharmacy in History 53(1):3-21), so it is likely that people have been unknowingly poisoned by Aristolochia for many centuries around the world: various species of Aristolochia have been used medicinally in many societies, most notably in Europe (herbal use as well as Balkan Endemic Nephropathy), China (TCM), and India (Ayurveda). See for example “Local Uses of Aristolochia Species and Content of Nephrotoxic Aristolochic Acid 1 and 2 - a Global Assessment Based on Bibliographic Sources” by Michael Heinrich et al. (2009) J. Ethnopharmacol. 125(1):108-44).

Due to our increasing interest in herbal supplements, we suggested several candidates for bioassays to the National Cancer Institute/National Toxicology Program. Of the suggestions, NTP completed testing on:

• Green tea extract (completed in 2016): no evidence of carcinogenicity in rats or mice
• Ginkgo biloba extract (completed in 2014): some evidence of carcinogenicity in male and female rats (thyroid gland follicular cell neoplasms); clear evidence of carcinogenicity in male mice (hepatocellular carcinoma and hepatoblastoma) and evidence for thyroid gland follicular cell adenoma; clear evidence of carcinogenicity in female mice (hepatocellular adenoma, hepatocellular carcinoma, and hepatoblastoma)
• Aloe barbadensis extract
  • Nondecolorized whole leaf extract of Aloe vera (completed in 2013): clear evidence of carcinogenicity in male and female rats (adenomas and carcinomas of the large intestine), no evidence in mice
  • Photococarcinogenesis (completed in 2010): weak enhancing effect by Aloe on simulated solar light photococarcinogenesis in mice (squamous cell neoplasms).
• Kava kava extract (completed in 2012): equivocal evidence of carcinogenicity in male rats (testicular interstitial cell adenoma), no evidence in female rats, clear evidence of carcinogenicity in male mice (hepatoblastoma), clear evidence of carcinogenicity in female mice (hepatocellular adenoma or carcinoma)
• Ginseng (completed in 2011): no evidence of carcinogenicity
• Milk thistle extract (completed in 2011): no evidence of carcinogenicity
• Goldenseal root powder (completed in 2010): clear evidence of carcinogenicity in male rats (hepatocellular adenoma and hepatocellular adenoma or carcinoma combined); clear evidence of carcinogenicity in female rats (hepatocellular adenoma); some evidence of carcinogenicity in male mice (hepatoblastoma and multiplehepatocellular adenoma); no evidence in female mice

We frequently reported about chemicals in coffee that were rodent carcinogens in comparison to synthetic chemical exposure (e.g. "Rodent carcinogens: Setting priorities" and our Margin of Exposure work). Such information about coffee did not concern us on a personal level, and I know that at least some of the team members, including myself, have been coffee drinkers. Our interest in coffee was due to its widespread consumption, the enormous amount of research on coffee, and the large number of chemicals found in brewed coffee. A reason for the large amount of research might have been due to unsubstantiated anti-coffee claims dating back to the 1500s, but most recently due to the claims of John Harvey Kellogg and C. W. Post in the early 20^th century. In more recent times, coffee has been known to be an important source of dietary antioxidants, so it was not surprising that recent large-scale analyses have found multiple positive effects on health associated with coffee drinking ("Coffee, Caffeine, and Health Outcomes: An Umbrella Review" by Giuseppe Grosso et al. (2017) Annual Review of Nutrition 37:131-156.) as well as increased lifespan ("Association of coffee drinking with total and cause-specific mortality" by N.D. Freedman et al. (2012) New England Journal of Medicine 366:1891-1904.).

• US EPA ("Toxico-Cheminformatics And QSAR Modeling Of The Carcinogenic Potency Database" by Kun Wang et al. Presented at 2007 Society of Toxicology Annual Meeting, Charlotte, NC, March 25 - 29, 2007.
• European Union's CAESAR ("New public QSAR model for carcinogenicity" by Natalja Fjodorova et al. (2010) Chem Cent J. 4(Suppl 1): S3)

• Our own success (sustained database building as well as analyses of the database for its implications such as natural vs. synthetic chemical comparisons and use in risk assessment-type analyses)
• Improvements in QSAR chemical toxicology prediction modeling, reducing the need for bioassays (to which our database partly-contributed)
• The decreasing number of chronic carcinogenesis publications that have been published over recent years (due at least in part due to QSAR modeling)
• The increasing political power of anti-regulation/anti-science/anti-government ideology

Contact