War on Carcinogens: Industry Disputes Human Relevance of Chemicals Causing Cancer in Laboratory Animals Based on Unproven Hypotheses, Using Kidney Tumors as an Example

Abstract

Evidence from studies in animals is essential for identifying chemicals likely to cause or contribute to many diseases in humans, including cancers. Yet, to avoid or delay the implementation of protective public health standards, the chemical industry typically denies cancer causation by agents they produce. The spurious arguments put forward to discount human relevance are often based on inadequately tested hypotheses or modes of action that fail to meet Bradford Hill criteria for causation. We term the industry attacks on the relevance of animal cancer findings as the “War on Carcinogens.” Unfortunately, this tactic has been effective in preventing timely and appropriate health protective actions on many economically important yet carcinogenic chemicals, including: arsenic, asbestos, benzene, 1,3-butadiene, formaldehyde, methylene chloride, phthalates, tobacco usage, trichloroethylene [TCE], and others. Recent examples of the “War on Carcinogens” are chemicals causing kidney cancer in animals. Industry consultants argue that kidney tumor findings in rats with exacerbated chronic progressive nephropathy (CPN) are not relevant to humans exposed to these chemicals. We dispute and dismiss this unsubstantiated claim with data and facts, and divulge unprofessional actions from a leading toxicology journal.

Keywords:

• carcinogenicity
• chronic progressive nephropathy
• human relevance
• kidney
• rats
• renal tubule tumors
• risk assessment

Introduction

When a chemical induces tumors in experimental animals, public health agencies evaluate potential risks to humans, while regulatory agencies quantify these health threats allowing risk managers to put procedures into place that would reduce or eliminate human exposure, and thereby minimize numbers of cancer among workers or the general population. In signing the National Cancer Act of 1971, President Nixon’s “War on Cancer” increased federal support for research to find treatments that would cure this dreadful disease(s). A much smaller proportion of federal cancer research funds has been allocated for disease/cancer prevention. One such inaugural program was the National Cancer Institute’s animal bioassay program1, which eventually merged into and was expanded by the National Toxicology Program (NTP). The NTP, created by David Rall in 1978, is headquartered at the National Institute of Environmental Health Sciences.2,3 The major objective of those programs has been to identify occupational and environmental carcinogens that pose carcinogenic risks to humans. While after >40 years the “War on Cancer” has made significant progress, though not yet achieving its ultimate goal, a different but related ‘war’, the “War on Carcinogens”, has been ongoing since the 1950s when tobacco and asbestos companies invested heavily to support the denial of evidence that cigarette smoking or exposures to asbestos was causal of human cancer. Though causal relationships between cigarette smoking and numerous types of human cancer have long been firmly established,4,5 the “War on Carcinogens” continues not only to exist but is thriving.6^–8 While we define the “War on Carcinogens” as efforts by industry with a vested interest in the chemical to dismiss human relevance of animal cancer data, the American Council on Science and Health, an industry-supported group that advocates against chemical regulation, applied this term to chastise state and federal regulatory agencies for banning or limiting human exposures to chemicals that cause cancer in experimental animals.9 In contrast to that position, we commend the efforts of regulatory agencies to reduce human health risks of cancer-causing agents.

Common strategies used to deny reliability or relevancy of tumor data for assessing health risks to humans include: a) claiming that doses/exposures used in animal studies were too high to cause an effect at human exposures (even when no nonlinear processes have been identified), b) claiming that the chemical induced essential precursor [“toxic”] changes in the animal at “high doses” that would not occur at lower doses (even when a consistent causal relationship between the “essential precursor change” and tumor induction has not been demonstrated), c) promoting untested mechanistic hypotheses of tumor induction in animals that are claimed to not occur in humans, d) proclaiming that tumors induced in rodents are not predictive of tumor induction in humans, e) declaring certain tumor sites [e.g., forestomach] are irrelevant because they are not present in humans, or f) discrediting the design, conduct, and interpretations of studies at the laboratory that identified carcinogenic effects. The industry “War on Carcinogens” is being waged to protect financial interests of chemical manufacturers;8,10 while at stake in this war is protection of public health. Thus, it is critical that hypotheses aimed at dismissing human relevance of animal cancer data be rigorously challenged and validated before decisions are made that might permit continued and further widespread human exposure to any carcinogenic agents.11

Rationale and Discussion

We recently experienced another battle in the “War on Carcinogens” when we provided scientific evidence that contradicted the hypothesis that the mechanism of induction of kidney tumors in rats by certain chemicals is not relevant to humans. For about 10–15 years, Gordon Hard and his colleagues have been promoting the hypothesis that enhanced chronic progressive nephropathy (CPN), a relatively common degenerative-regenerative disease of kidney in various strains of aging rats, is a “mode of action” for chemically induced kidney cancers in male rats and that renal tubular cell tumors (RTTs, kidney tumors) induced by chemicals that concomitantly exacerbate CPN are not relevant for assessments of human cancer risk.12^–17 This issue is vitally important from a public health perspective because kidney cancer rates in humans are increasing,18 and, though the causes and mechanisms of kidney cancer in humans are not totally known, exposures to various agents such as trichloroethylene, halogenated hydrocarbons, aromatic amines, other organic solvents, and certain metals have been associated with increased risk of renal cell carcinoma.19^–23 Also known to cause kidney cancers are aristolochic acid, arsenic and inorganic arsenic compounds, cadmium and cadmium compounds, phenacetin and phenacetin containing analgesic mixtures, tobacco smoking, X- and gamma-radiations.24,25

Industry-sponsored reviews of some NTP carcinogenicity studies reported correlations between severe CPN and increased incidences of kidney cancers in male rats;12,14,16,26^–28 however, studies of other agents have noted exposure-related increases in severity of CPN without induction of kidney tumors in male rats.29^–31 Because correlations in a small number of studies often lead to misleading or erroneous conclusions, we conducted a thorough and comprehensive evaluation of possible associations between exacerbated CPN and induction of kidney tumors in carcinogenicity studies conducted by the NTP in F344 rats that had evidence of enhanced CPN.32 Among the 43 studies in which at least one dose group was diagnosed with severe CPN, the overall conclusions by the NTP were that kidney tumor data for only 21 of those 43 chemicals represented some or clear evidence of renal carcinogenicity, i.e., approximately 50%. For some of the 21 chemicals in which increases in kidney tumors in male rats correlated with an increase in CPN severity, increases in kidney cancers were observed in female rats exposed to the same agents without an effect on CPN severity, and for several other chemicals, dose-response data did not support a causal link between CPN and kidney tumors. Thus, in many studies in which a correlation was observed, other mechanisms of carcinogenesis must have been operating. We concluded that the hypothesis promoted by Hard and colleagues that high severity CPN was causal for kidney cancer in rats lacks consistency, is unsupported by the data, and characterization of a mode of action was not possible because mechanisms of chemically exacerbated CPN and kidney tumor development are not known.32 Consequently, there is no scientific basis for dismissing the human relevance of kidney tumors induced by these chemicals in experimental animals, with or without exacerbated CPN.

Six months after our paper was accepted by Toxicological Sciences and published online, the same journal published online an extensive and directed criticism of our work by Hard and other industrial consultants.33 That article, which was offered under the guise of a session summary at the Toxicology Forum’s summer 2012 meeting, should have been published as a letter–to-the-editor (LTE) because it provided no new data on CPN and renal carcinogenicity, but focused almost exclusively on our work, which was only minimally discussed at that industry-sponsored meeting. We were surprised at not being invited by Toxicological Sciences to counterpoint criticisms made by Hard et al.,33 since it is customary and standard practice among reputable scientific journals to offer authors of articles in their journal an opportunity to reply to criticisms of their work. Consequently and necessarily, we took it upon ourselves to prepare and submit a LTE to Toxicological Sciences to clarify misrepresentations by Hard et al.33 of our published evaluation on this topic. Surprisingly, the Editor-in-Chief of Toxicological Sciences rejected our LTE with no real explanation, only this statement:

“Seriously, guys? You published your opinion and you knew there were people who would disagree. This matter is over.”

This obvious attempt to censure our evidence-based disagreement with the industry position was a shock, even coming from Toxicological Sciences (a publication of the Society of Toxicology, an industry-dominated organization). We subsequently contacted the publisher, Oxford University Press, and urged them to look into this unprofessional communication style and the unbalanced policy practiced by one of their journals. The publisher apologized and forwarded out letter to the Society of Toxicology’s (SOT) Board of Publications to resolve this matter. The Board also regretted this unseemly behavior, apologized, and sent our LTE out for peer-review. The reviewers recognized that criticisms by Hard et al.33 were “somewhat excessive for what should be a scientifically focused paper” and “re-interpreted some of the data …. in a distorted fashion,” and supported the publication of our LTE in Toxicological Sciences. However unlike Hard et al. we were permitted only 1000 words for rebuttal, compared with ∼10× that many for Hard et al.33 Though the Editor-in-Chief of that journal has been replaced, an Associate Editor informed us that a subsequent response by Hard et al. to our LTE would be allowed, a most unusual and unprecedented policy. We regarded the SOT Board’s decision to give Hard et al. another opportunity to criticize our evaluation and have the “last word” on our work to be an unfair and biased handling of this situation. A typical sequence for most reputable journals is to publish an Original Article, then LTEs from others raising issues or opinions, followed by a response from Original authors, if necessary. In this case the SDT/ToxSci surprisingly deviated and extended their own guidelines to allow Hard et al. another opportunity to make “excessive” and “distorted” criticisms of our initial analysis and response.

While it is common knowledge that certain journals provide a means for the publication of so called peer-reviewed articles on industry perspectives of animal carcinogenicity data, it is also a disservice to public health when editors of apparently well respected scientific journals of Oxford University Press offer unwarranted opportunities to industrial consultants to promote the dismissal of animal carcinogenicity data in evaluations of human health risk with unproven hypotheses. We remain pointedly concerned that a widely distributed and relied upon journal was and is steered by editors who appear biased. The journal’s outrageous response to our LTE involves direct censure of scientific dialogue and opinions that dissent from industry’s preferred interpretation of scientific data. For these reasons, we opted to withdraw our letter from Toxicological Sciences and with further narrative submit it to the peer-reviewed International Journal of Occupational and Environmental Health (letter shown below). Though the reviewers for Toxicological Sciences recommended that we remove criticisms of that journal, those changes were not made here because they provide relevant perspective for the regulatory and public health communities on how the “War on Carcinogens” is being conducted.

Letter to the Editor (Toxicological Sciences) Cpn: An Invalid Hypothesis for Renal Tumor Induction in Rats

Melnick RL, Burns KM, Ward JM, Huff J.

Note: The Following letter was accepted by Toxicological Sciences but withdrawn by the authors because Hard et al would be invited to again render their rebutted arguments

We are surprised that Toxicological Sciences published (as Advance Access) “Consideration of Rat Chronic Progressive Nephropathy (CPN) in Regulatory Evaluations for Carcinogenicity” by Hard et al.33 as a Forum Article because it provides no new data or information on CPN and renal carcinogenicity and because it includes extensive comments about our paper32 that were not presented by these authors at the Toxicology Forum from which this manuscript was alleged to be a session summary (i.e., it does not comply with ToxSci’s guidelines for Forum Articles). That article should have been a letter-to-the-editor as it largely attempts to respond to our article and circumvent standard journal practices. Because none of us were invited to participate in the Toxicology Forum and because this paper misrepresents our evaluation and perspective on this topic,32 we are compelled to respond and set the record straight.

We strongly disagree that exacerbation of CPN is a “valid MOA” [mode of action] for chemically induced renal tubule tumors (RTTs) in rats. Hard et al.’s proposed MOA is based solely on associations between exacerbated CPN and increased incidence of RTTs for selective chemicals in rats. These authors acknowledge that “the etiology of CPN and underlying basis for disease progression remain unknown;” that is, no precursor events or processes that are essential for RTT induction in rats have been identified. Thus, the declared MOA does not meet the most fundamental requirement for defining a cancer MOA.34 Without this information it is not possible to evaluate how processes linked to CPN might impact human cancer risk.

While Hard et al.33 claim reliance on Bradford Hill criteria for “accepting a plausible MOA”, their proposed MOA does not meet essential criteria used to infer causality from observed associations,35 including consistency, strength of association, biological plausibility, and coherence. For example, without knowledge of the etiology of CPN or RTTs in rats, it is not possible to demonstrate biological plausibility or coherence of the proposed association.

Hard et al.33 misstate our criticisms of their hypothesis. Our conclusions were not based simply on lack of a statistical relationship between severity of CPN and RTT incidence in NTP studies, they were based on 1) lack of knowledge needed to judge biological plausibility, and 2) numerous inconsistencies in the purported relationship between exacerbated CPN and RTT induction. Hard et al.33 noted that in our evaluation of 58 carcinogenicity studies conducted by the National Toxicology Program in male rats, we identified 21 chemicals causing severe CPN and having “some” or “clear” evidence of renal carcinogenicity; however, they neglected to acknowledge that in 22 other studies with severe CPN in male rats there was only equivocal evidence (5) or no evidence (17) of carcinogenic activity in the kidney. Thus, the chance that a chemical induces severe CPN and provides evidence of carcinogenic activity in the male rat kidney is only about 50%, not a consistent association by any objective standard.

Where associations were seen, events other than severe CPN (e.g., genotoxicity or other effects that might lead to tumor induction) may have been causal or contributory to the renal tumor response. Multiple processes are likely involved in RTT development, especially for chemicals that induce tumors in the kidney as well as in other organs. In contrast, the finding that a large number of chemicals induced severe CPN but no increase in RTTs demonstrates that CPN is not a reliable predictor of kidney tumor induction in rats, and underscores our criticism of the validity of the association promoted by Hard et al.33

Hard et al.33 criticized our selection of five chemicals that illustrate how apparent associations between exacerbated CPN and RTT development might be misinterpreted as evidence of causal relationships. Yet, for anthraquinone (male and female rats) and benzofuran (female rats), Hard et al.33 agree with us32 that “other mechanisms of carcinogenesis must be operating.” For tetrafluoroethylene, Hard et al. also agree with us that available evidence contradicts a “CPN-renal tumor association hypothesis.” For benzofuran in male rats, Hard et al. ignored evidence showing exacerbated CPN was not associated with induction of RTTs.

Hard et al.33 claim that their examination of individual animals confirmed the association between end-stage CPN and RTTs for ethyl benzene. However, they ignored our observation that RTTs were also increased in female rats in the absence of severe CPN.32 These results clearly show that advanced CPN is not necessary for induction of RTTs by ethyl benzene in rats. Though t-butyl alcohol induced RTTs in male rats, Hard et al.33 disregarded their own data showing exacerbated CPN in female rats, with similar mean CPN severity as in male rats, but no increase RTTs. Also, in male rats, dose response relationships were dissimilar for CPN and RTT incidence.

The p-values between CPN and RTT incidence that Hard et al.33 present for those five chemicals are meaningless because they ignore the involvement of other processes induced by these chemicals and the significant values for female rats are not based on advanced CPN, which they claim is essential for RTT development.

Hard et al.33 criticized our graphical depiction of relationships between CPN and RTTs because it was limited to dose groups with marked CPN severity. Yet, that is the precursor condition Hard et al.33 claim as a MOA: “the relationship between CPN and renal tumor formation is in the most advanced stages of CPN severity.” Thus, by analyzing the relationship between CPN and RTT incidence for all dose groups, including those with minimal or mild CPN severity, Hard et al.33 contradict a fundamental principle of their own hypothesis. Also, their analysis is subject to their own criticism of “an inappropriate selection of animals” because it includes “chemicals that may enhance renal tumorigenesis by MOAs other than, or in addition to CPN.” Interestingly, based on their analysis of data shown in Figure 1, Hard et al.33 claim a significant linear relationship exists between CPN severity and renal tumor risk, with even mild CPN severity associated with increased RTT incidence.

Consideration of MOAs in regulatory evaluations requires scientific evidence rather than defensive posturing against valid criticisms of unsupported hypotheses. Criteria proposed by Hard et al.33 for concluding that “exacerbation of CPN is a relevant MOA” for renal tumor induction in rats lack credibility because key biological events needed to establish a MOA have not been identified, and available evidence demonstrate numerous inconsistencies between severe CPN and induction of RTTs in rats. Delineation of a MOA for chemically induced cancer requires understanding and characterization of events and processes essential for tumor formation. Criteria proposed by industry consultants for judging and dismissing human relevance of chemically induced RTTs in rats when exacerbated CPN has also occurred may satisfy the interests of chemical producers, but are neither scientifically supported nor in the best interest of public health.

[End of letter]

Comments and Concerns

We remain concerned and determined not to allow skewed science and dubious data interpretations to exist and flourish without objective clarification. Unfortunately, but necessarily, we must remain cognizant and vigilant regarding who supports and funds science that other scientists and regulators may rely on. We further urge and alert research and regulatory scientists and public health communities to be cynically and skeptically aware and cautious of potentially skewed scientific interpretations to avoid producing insufficiently protective health policies. Tomatis36 warned the occupational and environmental health communities that serious public health consequences might follow if decision-making bodies rely on untested mechanistic hypotheses that are later shown experimentally to be incorrect. Thus, when formulating guidelines for acceptance of mechanistic hypotheses, we expect regulatory agencies to require rigorous testing and validation before considering a downgrade in the categorization of chemical carcinogens. The toxicology, risk assessment, and public health communities must demand the elimination of bias and conflicts of interest by publishers, journals, and professional organizations that virtually control the flow of scientific information.11,36^–41