Cancer, Animals and Man – The Case of TCE, Part 2

In the last post on this topic, I introduced how do we figure out whether or not a chemical is a carcinogen, noting that the primary sources of information used in carcinogen identification are epidemiological evidence and results from bioassays with laboratory animals. Epidemiology is an observational science. Epidemiology is a very challenging endeavor because it takes natural settings, which are inherently messy things as it finds them. Because of this, the ability for epidemiology to make inferences about the relationship between patterns of disease and exposure to specific substances can be confounded by different factors. In addition, attributing causation using epidemiology involves filtering study results through a set of guidelines or “viewpoints” (referred to by many as Austin Bradford Hill’s nine criteria, though he reportedly never viewed them as so clear-cut. This reflects a fundamental limitation of empirical science that cause and effect relationships cannot be observed directly or proven true by logic, and therefore must be inferred by inductive reasoning.

Studies with laboratory animals are experiments, which allow more control over the confounding factors, so that an investigator can study in isolation the relationship, “ingestion of substance A in a particular rat species results in the occurrence of adverse effect B”. The drawback with animal studies is that difficulties can arise in extrapolating the results to human exposure situations. In addition, cancer bioassays have their own set of possible confounders. The complexities involved with using animal studies can be seen by reviewing the guidelines EPA uses to evaluate animal studies and apply the results for assessing human health risks. Are you done looking at them? I’m sure that’s cleared things up a bit.

Let’s get down to a few cases. Liver cancer in mice is considered by EPA to be some of the strongest evidence that TCE is carcinogenic in humans. Even the solvents industry acknowledges that TCE is an animal carcinogen; but they will argue that the tumors don’t have relevance to human heath (TCE is thought to cause cancer in other organs – we’ll try and get to those later).

According to EPA’s reassessment, TCE causes liver cancer in mice. It has been suggested that TCE induces liver tumors through its metabolites, such as trichloroacetic acid (TCA), dichloroacetic acid (DCA) and chloral hydrate (CH); all three have been shown in separate studies to cause liver cancer in mice. EPA offers that “[t]he prevailing view of TCE-induced mouse liver carcinogenesis has been that these tumors arise in parallel with peroxisome proliferation** in the liver by TCE metabolites.” At one time, it was thought that the mechanism for liver tumors caused by peroxisome proliferators was due to oxidative damage caused by increases in free radical-generating enzymes and peroxisomal beta-oxidation that initiated carcinogenesis. Under this hypothesis, it was generally believed that because peroxisome proliferation has not been observed in humans, agents that produced this result in rodents would not present a carcinogenic hazard to humans. EPA has noted that a number of public comments on the draft TCE assessment expressed support for this point of view. Take a guess on who some of these commentors might be. At the same time, EPA’s Science Advisory Board suggested giving less weight to peroxisome proliferation as a liver cancer mechanism compared with other hypotheses.

However, in one of a series of white papers provided by EPA to the National Academy of Sciences panel looking into to TCE, more recent data is presented suggesting that chemicals causing peroxisome proliferation cannot be so easily ruled out as human carcinogens. The categorical statement that peroxisome proliferation doesn’t occur in humans, and the corollary that chemicals which are carcinogenic in animals through peroxisome proliferation aren’t going to be human carcinogens, may be a bit premature. Supposedly, peroxisome proliferation doesn’t occur in humans, because the human liver has very low levels of a macromolecule known as peroxisome proliferator-activated receptor-alpha (or PPAR), compared with the animal species that are sensitive to carcinogens acting through peroxisome proliferation. These carcinogens activate PPAR, which causes various metabolic effects, some of which resulting in tumor promotion.

There appears to be sufficient PPAR activity in humans to develop pharmaceuticals (hypolipidemic fibrates used to control triglyceride levels and thiazolidinediones used to reduce insulin resistance and blood glucose levels) that bind to PPAR and produce biological effects. One review from 2001 says:

Hypolipidemic fibrates have been shown to induce hypolipidemia in humans and to modulate gene expression (e.g., genes regulating lipid homeostasis) in human hepatocytes by PPAR activation. Thus, humans are responsive to agents that induce peroxisome proliferation in rats and mice.

A search of PubMed shows that hundreds of papers on PPAR activation have been published in the just the past few years alone. Of course, there’s a wide range of opinions and a lot of argument about how important PPAR activation is to human cancers (EPA’s white paper on the PPAR mechanism of action gives a voice to that range). However, I had previously accepted somewhat uncritically the argument that animal carcinogens acting as peroxisome proliferators weren’t human cancer risks. It’s been interesting to learn that humans are sensitive enough to PPAR activation to design drugs around it, and a message to be cautious about dismissing these kinds of chemicals as potential human carcinogens.

Critics of EPA’s reassessment state that TCE metabolites such as trichloroacetic (TCA) and dichloroacetic acid (DCA) are responsible for liver carcinogenicity in mice (in particular, some would argue that DCA should be viewed as the ultimate carcinogen – the chemical metabolized from TCE that actually causes liver cancer in mice). The argument goes that mice but not humans or rats produce appreciable levels of DCA as a metabolite of TCE; and TCE does not cause liver tumors in rats, so. . . TCE shouldn’t be considered a liver carcinogen in humans. Again, things probably aren’t so clear-cut and simple. Interactions between TCE metabolites and peroxisome proliferation by metabolites, including TCA which is formed in humans, are being offered as rationales for why liver tumors observed in animals exposed to TCE should be a concern for human health. EPA also has written a white paper on the topic of interactions between metabolites. The state of the science paper on mechanisms of TCE toxicity can be found here.

Another argument offered by critics of EPA’s TCE reassessment for why liver tumors in mice are not relevant to human health risks is that the test species used has a high spontaneous liver tumor rate. However, as discussed in EPA’s cancer risk assessment guidelines, animal bioassays are conducted using concurrent controls (a control population during the study) and historical controls (recorded trends in cancer incidence in lab animals across multiple studies) to address this background issue. Also, tumor incidence in a species with a high background incidence shouldn’t be a reason to dismiss the results outright; it provides an avenue for trying to further understand the mode of action in animals and how that might apply to humans. A high background tumor rate might affect how study results are used to quantify the potential cancer risk in humans; but that’s a different matter than using animal bioassays to identify chemicals as human carcinogens.

You can get a sense of the range of opinions on TCE from reading the comments on EPA’s reassessment (these can be found on http://www.regulations.gov/fdmspublic-rel11/component/main, using the advanced search feature and searching for the docket using the key word “trichloroethylene”) and EPA’s white papers.

There are more examples to explore – kidney cancer and non-Hodgkin’s lymphoma, and I’ll be visiting these, as we get closer to the National Academy Sciences committee on TCE’s report (due out sometime the middle of this year).

**A peroxisome is an organelle present in nearly all cells that plays an important role in the metabolism of long chain fatty acids and amino acids, and in synthesis of bile acids and cholesterol that are further used in digestion and hormone synthesis.