Editorial: Estrogens from Plastic—Are We Being Exposed?

The controversy over the potential negative impact on public health of environmental chemicals with estrogenic activity has spilled over from scientific journals (1, 2) into the public domain. News articles on the subject have appeared in multiple sources including Science (3), Journal of the American Medical Association (4), Chemical and Engineering News (5), Science News (6) and Newsweek (7), and in many newspapers, as well as on National Public Radio. The book Our Stolen Future (8) has emphasized the effects of environmental chemicals on wildlife and their long-term effects on the environment. Some stories have hyped the risk of environmental chemicals with provocative terms such as “Ecocancers” (6) and titles such as“ Can environmental estrogens cause breast cancer? (9). Proponents of the position that environmental estrogens are hazardous raise concerns about birth defects due to fetal exposure (8, 10), the increased incidence of breast cancer (9, 11), falling sperm counts and decreased fertility (12), and numerous other conditions (3–12). However, in each case there is controversy over the findings with opponents making strong counter-arguments. In all of these conditions, the risk is hypothetical, with no data yet proving a causal relationship between environmental estrogens and illness or disease in people.

The claim that environmental estrogens are causing a decline in sperm counts has been especially contentious. Investigators in this field have so far failed to agree on whether sperm counts are, in fact, falling. If sperm counts are in decline, they certainly cannot agree on whether environmental estrogens are even remotely involved (12, 13). Similarly, the proposal that environmental estrogens contributed to an increased incidence of breast cancer (11) was quickly refuted (14).

Scientifically sound experiments have documented that various environmental chemicals are capable of acting as endocrine disrupters, either hormone agonists or antagonists, which can potentially alter the hormonal balance in animals and people. A number of studies, including our own (15), have clearly shown the estrogenic activity of some environmental chemicals. What remains controversial and open to serious debate is the level of exposure of the population to these agents and an ascertainment of whether these levels are sufficient to cause harmful effects. The critical question is whether the population is exposed to a high enough concentration of chemicals to cause, in the in vivo setting, the effects that these agents have been demonstrated to initiate in vitro. This question is not easily answered. Although the estrogenic activity of these substances is very much weaker than estradiol, as discussed below, new chemicals with endocrine disrupting potential continue to be discovered, unanticipated pathways of exposure continue to be found, and concern about the cumulative effects of these agents continues to grow. However, the level of exposure is difficult to quantify and the array of potential target organs in the body continues to rise.

In this issue of the journal, Ben-Jonathan and her group extend the accumulated data on the estrogenic activity of Bisphenol-A (BPA), a compound used to manufacture the plastic called polycarbonate. In this article, Steinmetz et al. (16) argue that the pituitary cells that regulate PRL release are a new target for environmental estrogens in general and BPA in particular. Although there are many substances considered to be environmental estrogens, including pesticides, pollutants, and various chemicals, this paper deals with the estrogenicity of BPA, and I will therefore focus my discussion primarily on the estrogenic substances derived from plastics.

BPA is the monomeric unit used to produce the ubiquitous plastic, polycarbonate. This plastic has excellent properties making it tremendously useful in many applications. In 1993, my group reported on the estrogenic activity of BPA that was released from polycarbonate flasks during the autoclaving of media (15). We showed that BPA could act on MCF-7 human breast cancer cells as an estrogen, stimulating cellular proliferation and inducing progesterone receptors. BPA could bind to estrogen receptors, and the estrogenic effects induced by BPA were blocked by the estrogen antagonist tamoxifen, thus supporting the notion that the estrogenic activity of BPA was mediated via the estrogen receptor. Although the chemical structure of BPA is quite similar to diethylstilbestrol (DES), a well known and very potent nonsteroidal estrogenic compound with a bis-phenolic structure, BPA was a much weaker estrogen, approximately 1000- to 2000-fold less potent than estradiol (15). Our results were of interest for two reasons. First, we wanted to alert other investigators about the potential risk of estrogenic artifacts causing spurious effects in laboratory experiments. Second, we wanted to raise for scientific scrutiny the possibility that BPA was a risk to public health. Because polycarbonate is used in the packaging, storing, and preparation of a myriad of foods and beverages, including water jugs, bottled beverages, baby food, and juice containers, we wondered whether BPA might contaminate the food supply and potentially harm the public. However, because limited data were available upon which to draw a conclusion, we cautiously raised the question for study and made no claim that a public health risk existed (15).

In addition to our work, Soto, Sonnenschein, and colleagues at Tufts University showed very similar results with another component of different plastics, p-nonyl-phenol (17). This alkylphenolic substance, used as plastic additive and surfactant, could be released from polyvinyl chloride and polystyrene plastics even without autoclaving. The Tufts group also developed an “E-Screen” to assay the estrogenic activity of unknown substances or mixtures based upon their ability to stimulate MCF-7 cell proliferation (18).

Two subsequent studies further demonstrated the potential public health hazard of BPA. First, Brotons et al. noted that many food cans are lacquer-coated with a plastic lining and that, within some cans, the food contained substantial amounts of BPA (19). Because the food is autoclaved in the cans, the conditions of our laboratory polycarbonate flask experiment are essentially reproduced by the canning industry. While we had found BPA at levels of 2–4 μg in the liquid contents of our plastic flasks, Brutons et al. reported that food cans that contained BPA had contents ranging between 4 and 23 μg. Almost all of the estrogenicity was due to BPA based upon the results of the E-Screen.

A second study found that some routine dental procedures could potentially cause significant amounts of BPA exposure (20). Resin based composites and sealants commonly used in dentistry are made of BPA or BPA-dimethacrylate. Olea et al. found 90–931 μg of BPA in the saliva of patients 1 h after a sealant was applied to their teeth. Unfortunately, we have no data on how long the BPA persisted in the saliva and the total level of exposure after various dental treatments.

In this issue of Endocrinology, Steinmetz et al. (16) confirm by additional methods that BPA is estrogenic, and they estimate that its potency in vitro is 1000- to 5000-fold less active than estradiol. This study raises three additional important issues. First, Steinmetz et al. examined the estrogenic activity of BPA in vivo. Although the administration was from sc implanted SILASTIC (Dow Corning, Midland, MI) brand “capsules” and not the oral route, it is highly significant that the estrogenic effects of BPA were demonstrated in ovariectomized rats. From the data presented on PRL secretion in vivo, it can be surmised that BPA may only be 100- to 500-fold less active than estradiol. Thus, BPA may be an order of magnitude more potent in vivo than assessed by prior in vitro studies. The reasons for this increased potency in vivo are not apparent, but this finding could make estimations of BPA exposure levels based solely on in vitro potency liable to a 10-fold error.

Secondly, Steinmetz et al. raise a new BPA-target organ beyond the obvious ones of breast and uterus. The study shows BPA effects on pituitary cells that secrete PRL and speculates on a possible link between environmental estrogens and PRL secreting pituitary tumors (16).

A third intriguing concept from the paper by Steinmetz et al. is that there are genetic differences in susceptibility to estrogens including BPA; Fischer 344 rats are very sensitive, whereas Sprague-Dawley rats are resistant (16). The mechanism for this differential effect is unclear from the data available. The authors speculate that subpopulations of humans may likewise be more sensitive to BPA and other environmental estrogens than the population at large.

The public health question raised by all of these studies is: are the amounts of ingested BPA from all sources substantial enough to cause significant estrogenic effects in the population? It is difficult to answer this question because there are no data available yet on the estrogenic potency of BPA via the oral route.

Perhaps the most articulate proponent of the view that environmental estrogens are not a public health hazard is Stephen Safe, a pharmacologist/toxicologist who has studied this and related environmental issues for many years. Safe argues that the total amount of environmental estrogens that people are exposed to, especially because of their low potency, is inconsequential (2). He contends that phytoestrogens in our diet far outweigh the estrogenic potency of environmental estrogens. Furthermore, environmental antiestrogens would balance out many of the harmful effects of environmental estrogens (21). Using calculations of “estrogen equivalents,” he concludes that the exposure level to environmental estrogens is trivial in comparison with estrogen levels used in therapeutic settings and even thousands-fold lower than the flavonoid phytoestrogens in food that we routinely consume (2). Safe argues persuasively that there is no evidence proving that there is a problem due to environmental exposure.

However sound these theoretical arguments may appear, I believe, and Safe agrees [personal communication], that a number of considerations have to be taken into account that may modify these estimates of risk: 1) More exposure of the population than anticipated may be occurring due to environmental estrogens from unexpected sources and new, as yet unknown, substances including BPA and other chemicals (22). The effects of BPA and other components of plastic have not yet been accounted for in estimates of estrogen equivalents exposure. 2) Additional mechanisms may be operative that are additive to the negative effects of environmental estrogens. For example, substances that inhibit enzymatic degradation of environmental or natural estrogens or exposure to environmental substances with antiandrogen activity may provide a second negative effect (23). 3) The cumulative effect of many different agents may be additive or synergistic. Although cumulative exposure may be exceedingly important, the added possibility of synergistic interactions among the environmental estrogens, thereby increasing their estrogenic activity, is potentially alarming. This hypothesis was raised by Arnold et al. (24) but has now been refuted by two groups (25, 26) and remains equivocal and unproven. 4) Tissue-specific effects for estrogens are well documented so that some environmental estrogens may be more potent than others in certain tissues, or may be agonists rather than antagonists (27). This tissue specific activity is seen with estrogen antagonists such as tamoxifen and the new“ designer estrogens” such as raloxifene and droloxifene. Thus, calculation of estrogenic potency based on actions using the breast or uterus as models may not be quantitatively predictive of effects in other tissues such as the liver, bone, or brain or in the developing fetus that may be more sensitive to hormonal influences (28). 5) Genetic susceptibility of subpopulations, as raised by Steinmetz et al. (16), may make some groups more susceptible to estrogenic effects. 6) Pharmacokinetic or other in vivo factors may cause the estrogenic effect to be greater than expected based solely on extrapolations from in vitro data, as shown by Steinmetz et al. (16). These factors might include inadequate metabolism of man-made chemicals or metabolism to more active analogs, lack of binding to serum proteins, accumulation in the body and storage in fat, or other conditions causing an increased or prolonged exposure. 7) Finally, all estrogenic substances do not behave in an identical fashion. Some might have additional unique actions, as was seen with DES and vaginal cancer in daughters of treated patients (29). Because BPA is structurally similar to DES and not estradiol, its estrogenic actions might be more predictable from DES than estradiol and might cause unanticipated effects.

Conversely, similar arguments might suggest decreased in vivo activity relative to in vitro potency. Of prime importance is the possibility that ingested BPA would be rapidly metabolized to inactive products and have no estrogenic activity by the oral route. This possibility needs to be tested before the concept that BPA is a public health hazard can be taken seriously.

In conclusion, a healthy skepticism about the risk due to BPA and other environmental estrogens is wise. In my opinion, the data do not yet prove a relationship of environmental estrogens to breast cancer, sperm counts, or any general adverse effect. However, because this potential hazard could occur on a global scale affecting the entire population, this possibility should not be discounted without further study. Although I suspect that the environmental exposure is not adequate to cause serious effects, except in unique situations where a large local contamination occurs (10), the potential hazards are too important, and more investigation is certainly warranted. Especially necessary are in vivo studies examining dose-response effects of orally ingested BPA and other potential environmental estrogens and an examination of their actions at many tissue sites. I agree with Steinmetz et al. (16) that the complex potential problems of additive, cumulative, and persistent effects of these agents must be addressed. This controversy will not go away, nor should it, until adequate experimental results are available to ascertain whether or not the population is being exposed to harmful levels of environmental estrogens.

References