Key environmental epigenetics paper challenged

http://pubs.acs.org/doi/full/10.1021/es902777p

Rebecca Renner
Environ. Sci. Technol., Article ASAP
DOI: 10.1021/es902777p
Publication Date (Web): September 23, 2009
Copyright © 2009 American Chemical Society

The findings of a key paper (Science 2005, DOI 10.1126/science.1108190) with potentially profound implications for the future of environmental health research and regulation are being challenged by a small group of government and industry scientists who say they cannot replicate its results.

This paper and subsequent publications on epigenetics from molecular biologist Michael Skinner’s lab at Washington State University link exposure to an endocrine-disrupting chemical in the womb with health problems that extend for four generations. “If these findings are reproducible and robust, then the implications of this study are huge,” says epigenetics researcher Bernard Robaire at McGill University (Canada).

“We would have to redefine the mission of our regulatory agencies. That is why it is absolutely essential for these results to be replicated in other labs. The fact that these other researchers do not appear able to replicate the study is very important,” Robaire adds.

Epigenetics refers to heritable changes in gene function that cannot be explained by changes in DNA sequence. DNA methylation and histone modification are two of the mechanisms involved in epigenetic regulation. Numerous studies have shown that nutrition can cause epigenetic changes in animals and people. Other studies of chemical exposure have reported epigenetic effects in the offspring of the exposed parents. Thus far, though, Skinner’s lab is almost alone in linking exposure to environmental chemicals with effects on the epigenetic code that persist for many generations.

The 2005 paper reported the results of exposure of pregnant Sprague−Dawley rats to very high doses of the anti-androgenic fungicide vinclozolin during the critical developmental window when sex determination occurs. The exposure resulted in male reproductive problems that included fewer sperm with less motility than normal. Nearly all males in the four subsequent generations studied were affected. The cause appeared to be an epigenetic effect related to changes in DNA methylation.

“Prior to our observations, most transgenerational effects stopped at the second generation,” says Skinner.

Robaire summarized the implications of the findings, should they prove generally applicable to people, as raising the possibility that your pregnant grandmother’s exposure to chemicals may cause disease in you and your grandchildren (Nat. Med. 2008, DOI 10.1038/nm1108-1186). Others agree. “The impact on human health, should the Skinner data be replicated and expanded in other labs with other chemicals at lower doses, could be enormous,” says Jerry Heindel, a scientific program administrator at the National Institute of Environmental Health Sciences. “This would be a new paradigm that changes everything and puts more emphasis on reducing exposures during pregnancy to lower incidence of disease.”

The Skinner group has expanded upon this original report with nine additional peer-reviewed papers. Some of their further findings link fetal exposure to vinclozolin with a higher incidence of prostate and kidney disease, increased immune-system abnormalities, and tumor development. They report that female offspring show uterine hemorrhage, anemia, or both late in pregnancy, as well as an increase in mammary and other tumors. The lab also has described behavioral effects in which female rats three generations after exposure preferred as mates those males without a history of exposure.

A 2006 paper identifying a series of genes epigenetically altered by vinclozolin exposure (Endocrinology, DOI 10.1210/en.2006-0987) was retracted in June 2009 by the authors. The published retraction points to “the inability to locate and confirm the data” generated by one author and states, “The coauthors were not aware of this falsification of data.” Other researchers in the field do not see this incident as casting doubt on the remainder of the work from Skinner’s lab.

But recent abstracts and presentations by U.S. Environmental Protection Agency (EPA) scientists (Biol. Reprod. 2008, 78: 228. 735) and peer-reviewed papers by industrial labs in Germany (Reprod. Toxicol. 2008, DOI 10.1016/j.reprotox.2008.04.001) and Japan (Toxicol. Appl. Pharmacol. 2009, DOI 10.1016/j.taap.2009.03.004) have not been able to detect any transgenerational effects in similar experiments involving vinclozolin.

Earl Gray, a reproductive toxicologist with EPA, has a long history of studying the anti-androgenic properties of vinclozolin on male rat pups (Toxicol. Appl. Pharmacol. 1994, DOI 10.1006/taap.1994.1117). He has found that high doses of vinclozolin given to pregnant rats demasculinize their male pups when the pregnant mothers are exposed after sexual differentiation has occurred and androgen receptors are active—that is, at gestational day 13−17. The Skinner lab doses pregnant rats earlier, as the gonads are being formed—gestational day 8−15. During this time, natural epigenetic changes affect the germ cells, and it is possible that an environmental agent could interfere with this process and the germ line, Skinner says.

But Gray says that his attempt to replicate the Skinner lab results has been “entirely negative”. His experiment differs from the Skinner lab procedure because Gray gave vinclozolin as an oral dose, rather than as an intraperitoneal (IP) injection, as in the 2005 paper from Skinner’s lab.

There is a good reason for the change, argues Gray. “An IP injection with such a high dose is a terrible way to expose a pregnant rat,” he says. “You’re going to drop the chemical right on top of the uterus and fetus. After necropsy you see particles of the chemical, so there’s no way to quantify the exposure to the fetus.”

Toxicologist Bennard van Ravenzwaay, who works at BASF, vinclozolin’s manufacturer, also believes that IP injection is not an appropriate way to conduct this experiment. Using oral exposure, he has not found multigenerational effects.

Nevertheless, even studies that have used IP injection to replicate the Skinner lab results have failed. In a peer-reviewed paper published this spring, researchers from Sumitomo Chemical Co., Ltd., used IP injection in their attempt to replicate the Science paper results, but failed.

Van Ravenzwaay’s group is also trying to replicate the experiments using IP injection. They presented preliminary results of an IP exposure experiment at the Society of Toxicology meeting in March 2009. “We can certainly state that there is no evidence of a transgenerational effect,” he says. “We did see some minor anti-androgenic effects in the pups exposed as fetuses, but these were not passed on to the next generation.” Van Ravenzwaay adds that he has had no luck in his attempts to correspond with Skinner about the studies.

The type of rat used in the experiments is likely to be a factor, Skinner says. In his lab, he notes, inbred rats do not appear to show as big an epigenetic response as outbred rats, which are more genetically diverse. Skinner believes that outbred rats may have a bigger epigenetic response because they are generally more robust and vigorous than inbred strains. However, Robaire says that his demethylation experiments involving pharmaceuticals achieve the best results when he uses inbred rats. “Using inbred rats with the same genetics increases the signal-to-noise ratio and gives us better results,” he says.

“Our primary focus of the research,” says Skinner, “is not in the toxicology of vinclozolin, but instead its use as a pharmacological agent to induce the transgenerational phenomena to study the mechanisms involved.” His lab has studied another well-known anti-androgenic chemical, flutamide, which does not elicit the same response in rats as vinclozolin. “Right now, we don’t know what it is about vinclozolin that causes this response,” he adds.

Heindel says that the fact that others have not reproduced Skinner’s results may not be so surprising. “The published manuscripts from Mike Skinner’s lab have been shown to be reproducible within his lab. If researchers use different doses or routes of exposure or different strains of animals or different diets, then they are not reproducing the Skinner experiments. So, if they don’t get the same results, that does not affect the validity of the Skinner experiments.”

But Robaire says the verdict is still out. “The Japanese lab repeated exactly the same dose, exactly the same delivery method, and could not find the results. People are still groping to resolve this information. Michael Skinner is a very solid scientist, but it’s possible that this effect may turn out to be unique to one subspecies of rat,” he says.