By Jim Kelly
The first time UTMB professor Cheryl Watson heard of the endocrine disruptor hypothesis, she says, “I thought, now there’s a scary idea.” This was in the early 1980s, and the scary idea was that chemicals in the environment might interfere with the vital hormonal signaling networks that govern animal and human reproduction, development, and behavior.
A small but growing group of wildlife biologists, environmental toxicologists, cancer researchers, and specialists in developmental disorders suggested such interference could be causing a host of problems that recently seemed to have emerged in both animals and humans. Endocrine disruption caused by pollution, they said, was the best explanation for why certain populations of birds near the Great Lakes had mysteriously become unable to produce viable eggs, or in some cases had become completely uninterested in courtship and mating. It might also explain the sudden appearance in England of fish possessing both male and female sex characteristics, and male alligators in a Florida lake with sex organs one-third to one-half normal size. In humans, some asserted, exposure to endocrine disruptors before or just after birth could account for an apparent decline in the average quantity and quality of sperm, as well as an increase in the rates of breast, prostate, and testicular cancer.
We knew that animals were being affected,” Watson says. “You had these classic cases of DDT causing thin-shelled bird eggs and alligators in Lake Apopka in Florida having very small reproductive organs, things like that.” There was no reason, she remembers thinking, to assume that humans were immune to these effects.
And yet, many scientists seemed to be doing just that. Their strongest arguments centered on the lack of an explanation for how the very low doses of endocrine-disrupting chemicals to which most people were exposed could produce any effect at all. Take the largest group of endocrine disruptors, the so-called “estrogen mimics” or “xenoestrogens,” which included such notoriously toxic substances as DDT, dioxins, and PCBs. Certainly they seemed to be able to imitate the body’s own dominant estrogen, estradiol, in some ways—they could make breast cancer cells grow. But at realistic exposure levels they barely triggered any response at all in experiments designed to measure traditional mechanisms of hormone action. They were, at worst, thought to be “weakly estrogenic.”
It was a messy question, and not just scientifically. The issue had become politically controversial as well, as environmentalists and breast-cancer activists called for tighter regulation of dozens of economically important manmade chemicals known to have estrogenic activity—everything from pesticides and herbicides to widely used detergents and plastics.
Some researchers might have just shrugged their shoulders and walked away. But Watson had a different reaction. “I always had a kind of curiosity about how we could see these effects in animals, and yet when we tested mechanistically in the laboratory, we didn’t see any effects,” she says. That curiosity combined with something else—years spent on the leading edge of estrogen signaling research — to produce an idea. Experiments conducted with her frequent collaborator Bahiru Gametchu had convinced her that steroids could act on cells through a “non-traditional pathway,” one that at the time was not yet accepted by the majority of steroid signaling specialists and virtually was unknown outside the field. If a natural estrogen could trigger this mechanism, she reasoned, then it was possible that xenoestrogens could do the same thing. And if she could measure that effect, it might be possible to begin to determine the true threat these chemicals posed.
Now—after years of fine-tuning experiments, struggling against entrenched ideas about how hormones work and submitting and re-submitting papers to journal after journal—she believes she’s succeeded.
To Watson, the results she and her team have produced seem both exciting and deeply disturbing. Watson’s data, published in the National Institute of Environmental Health Sciences-sponsored journal Environmental Health Perspectives, suggest that xenoestrogens can act at concentrations almost too small to imagine, plugging themselves into cellular circuits that biologists are only beginning to understand. “We see xenoestrogens causing the same type of responses that physiological estrogens can, triggering responses at the same low doses for the most part,” Watson says. “They’re doing them in a different timing pattern, but they’re still just as potent. The literature says they’re weak estrogens, and they don’t do anything unless you use high concentrations of them, and we’re saying that’s not true.”
The literature says they’re weak estrogens, and they don’t do anything unless you use high concentrations of them, and we’re saying that’s not true.”
Watson’s journey to uncover a new signaling mechanism for xenoestrogens began with a colleague’s mistake—the kind of serendipitous stumble so often recounted in tales of scientific discovery. That colleague was Gametchu, an immunologist, who in the late 1980s was pushing the limits of the then-new technology of fluorescent antibody labeling. These antibody “fluorescent tags” were designed to attach to only one kind of molecule. Once they got into a cell, examination with a microscope would show exactly where those molecules—in this case glucocorticoid receptors, important players in leukemia therapy—were located. Antibodies are not small enough to slip through a cell’s membrane on their own, so Gametchu had to use a detergent to “punch holes” in his cells to get the fluorescent antibodies in. One day, Watson remembers, Gametchu forgot to use the detergent. By the time he realized his mistake, he was so close to the end of the experiment that he decided to go ahead and take a look through the microscope anyway.
What he saw surprised him—and it surprised Watson, too, when he invited her to come down the hall and look a few minutes later. “It was really weird,” she remembers. “He was trying to determine if the receptor was in the cell nucleus or in the cytoplasm. You would expect to see either a cell with the nucleus all lit up or a cell with an empty nucleus and the cytoplasm all lit up. What I saw was a cell with glowing polka dots all over its surface.”
Both the glucocorticoid and the estrogen signaling molecules that Gametchu and Watson studied were classified as steroid hormones. According to current theory, that meant they slid straight through the plasma membrane surrounding a cell to dock with glucocorticoid and estrogen receptor molecules inside the cell—the only place, according to the conventional wisdom of the time, that such receptors existed. And yet, now Gametchu seemed to have found glucocorticoid receptors on the cell’s outside.
Talking it over with Gametchu a few days later, Watson remembered hearing about a Stanford researcher named Clara Szego who had published data in the 1970s that seemed to show that steroid hormone receptors— estrogen receptors, in fact—existed on the outer membranes of cells. Szego’s claims had been widely disputed, but now it looked as if she might have been right.
Immediately, Watson began thinking about how she might do antibody experiments to look for estrogen receptors on the cell membrane. Estrogen signaling was known to play a critical part in the genesis of breast cancer, and estrogen receptors on the cell membrane might act differently from those inside the cell, activating previously unsuspected biochemical circuits. If they could be found, membrane estrogen receptors could help supply a major piece of the breast-cancer puzzle.
For the next five years, with the help of Gametchu and graduate student Todd Pappas, Watson worked to develop a technique that would enable her to use antibodies to see membrane estrogen receptors. The system she settled on employed a line of rat pituitary gland tumor cells known to respond quickly to estrogen, and a custom-made antibody to the estrogen receptor. (Estrogen receptor antibodies were just beginning to be commercially available, but were prohibitively expensive because they were being marketed for breast-cancer clinical diagnosis, not research.) At the end of a long, difficult process of trial and error, she succeeded: “Using our antibody to the estrogen receptor, we saw essentially the same thing that we had seen with the membrane glucocorticoid receptors.”
Seeing estrogen receptors on the membrane was a great achievement, but getting other people to see them, too, turned out to be much harder than Watson had anticipated. The orthodox view that steroid hormone receptors were found only inside cells was strongly held by the researchers who reviewed articles on this subject for scientific journals. “We spent a lot of energy in the first few years of this effort submitting papers over and over again, re-writing and re-submitting—papers typically would take four and five different submissions to different journals before we could get them accepted,” Watson says.
But in 1995, they managed to break through with a paper in FASEB Journal, a widely respected and read scientific publication. Around the same time, Watson published a paper titled “The Other Estrogen Receptor in the Plasma Membrane: Implications for the Actions of Environmental Estrogens” in Environmental Health Perspectives. (“Environmental estrogens” is a term used for xenoestrogens or estrogen mimics that are environmental contaminants.) In that paper, Watson described her group’s finding that estradiol could interact with membrane estrogen receptors and within minutes trigger inflows of calcium ions that led to the release of large quantities of prolactin, a powerful reproductive hormone whose wide range of effects included the promotion of lactation and maternal behavior.
The speed of this response meant that it had to be what researchers called “non-genomic” or “non-nuclear”—that is, it had to be happening through a different mechanism than the relatively slow process by which scientists had traditionally thought hormones worked. That pathway required the involvement of DNA (the genome, thus—“genomic”) and RNA in the cell’s nucleus to produce new proteins like prolactin, and it took hours or even days, not minutes. The implication was that if membrane estrogen receptors could respond to estradiol in such a rapid, unexpected fashion, they might be doing something similar in response to xenoestrogens—and doing it so much faster that scientists weren’t catching it.
“People were so wed to the idea of a genomic response, and when you’re looking at genomic responses you usually don’t check until twenty-four or forty-eight hours, when you’re sure you can get a good signal,” Watson says. “Well, you could be completely missing a whole mechanism because you chose time points that didn’t let you see it.”
If researchers were “missing a whole mechanism,” Watson recognized, it could shed light on why they’d had so much trouble explaining the apparent xenoestrogen effects observed in wildlife—and, increasingly, in laboratory animals—with test-tube experiments on cells. Those experiments had been designed to detect genomic responses, and they all seemed to show that xenoestrogens only produced such effects at very high concentrations, a thousand or even ten thousand times as high as estradiol. “People look at that and say, that’s a lot,” Watson says. “What are we worried about? Even if a chemical company spills a lot of this stuff it gets quickly diluted below that level. Very rarely would you ever have that kind of quantity around.” But if, like estradiol, xenoestrogens could also work through a membrane receptor, who knew how little might be needed to produce a significant response?
Watson had a hunch the answer would be interesting, and so in 2001 she and postdoctoral researcher Nataliya Bulayeva began developing a series of experiments to look at the responses of rat pituitary tumor cells to a number of known xenoestrogens.
The chemicals studied included pesticides (endosulfan and dieldrin) and their breakdown products (DDE, for example, produced when DDT is metabolized); coumestrol, a xenoestrogen produced by plants and found in alfalfa sprouts and sunflower seeds; and two compounds widely used to make plastics, nonylphenol and bisphenol A. They also tested DES, a synthetic estrogen that had been given to pregnant women from 1940 to the early 1970s in the mistaken belief that it prevented miscarriages. In fact, researchers later found that the daughters of women who had taken DES during pregnancy faced a far greater risk of a rare type of vaginal cancer and malformations of the reproductive tract.
“Our method is more convenient—it gives you more data faster, and that’s the point, the power of numbers,”
Primarily carried out by Bulayeva and research assistant Ann Wozniak, the lab work focused on detecting whether exposure to the different xenoestrogens caused rapid responses in calcium inflow and prolactin release; it also compared the level and time course of any responses to those induced by estradiol. It aimed as well to measure changes in extracellular-regulated kinases (ERKs), enzymes inside the cell that were known to participate in many different signaling processes—some of which led to quick non-genomic responses and others of which influenced complicated processes like cell division.
ERK measurements had been done before, but Bulayeva had adapted a more efficient and precise mechanized process that enabled her to rapidly and accurately scan large numbers of cell preparations. “Our method is more convenient—it gives you more data faster, and that’s the point, the power of numbers,” Bulayeva says. Still, “faster” is a relative term when one is testing the effects of varying concentrations of multiple chemicals at exposure times ranging from three to thirty minutes. Once the ERK analysis system was ready, it took more than six months for Bulayeva to gather the needed data.
The results of the ERK studies and the calcium and prolactin experiments, though, looked like they were worth the wait. “We looked at these data and said these things are just as potent as physiological estrogens like estradiol if you look at these mechanisms,” Watson says.
Take bisphenol A, for example. The chemical is a near-ubiquitous component of the polycarbonate plastics used to make beverage bottles and other food packaging, and it easily leaches into its surroundings, making exposure virtually impossible to avoid; a 2005 Centers for Disease Control study examining American urine samples found that 95 percent contained measurable levels of bisphenol A. The Watson lab’s results showed that cells exposed to less than one part per trillion concentrations of bisphenol A doubled their prolactin output.
Watson’s work was cited in two recent papers by bisphenol A expert Frederick vom Saal of the University of Missouri. One, published in the Proceedings of the National Academy of Sciences, described prostate gland abnormalities found in male mouse fetuses whose mothers had been fed doses of bisphenol A lower than those normally consumed by pregnant women. The other, a commentary published in Environmental Health Perspectives, reviewed the latest test-tube, animal, and epidemiological studies of the low-dose effects of bisphenol A, noted significant differences in results produced by industry- and government-sponsored research, and called on the EPA to re-assess the risks posed by the chemical.
Other data from the Watson group were almost as disturbing. They showed prolactin output activity equal to that prompted by the human hormone estradiol at levels below one part per billion of the pesticide endosulfan and the DDT breakdown product DDE. And all the xenoestrogens tested—except bisphenol A—appeared to be causing strong ERK activation within a matter of minutes at very low concentrations.
“People thought they could ignore such low levels, that in real life unless you bathe in this stuff you don’t have to worry about it,” Watson says. “There are hundreds of articles over the last ten years or so, all saying that you need micromolar quantities of these compounds to get a reaction through the nuclear estrogen receptor gene expression route, and now we come along and say we’ve got nanomolar responses—responses at concentrations a thousand times lower. Well, at the last meeting when I presented this it got people’s attention. They went yikes!”
The Watson group’s results had implications that went beyond simple shock value. One of the most profound was the way their unusually detailed data showed the paradoxical responses of living things to different levels of xenoestrogens at different points in time. When it comes to endocrine signaling effects, the toxicological truism that “the dose makes the poison”—the greater the level of a toxic substance, the greater the response—has recently come to be seen as too simplistic. Reactions to low doses of hormones or hormone mimics can be much greater than reactions to high doses, and such reactions do not simply increase or decrease in a simple fashion over time.
“Here’s the estradiol ERK response,” Watson says, sketching a curve in purple ink on a piece of scrap paper. “Here’s the response in this case to a phytoestrogen, a plant estrogen. Now they both do the trick, but look at the complete difference in phasing.” The two curves don’t match up—one peaks sooner than the other, and then both show second, non-matching peaks. Watson uses the sketch to pose a straightforward question: Do the two add together, cancel each other out, or interact in some other fashion? This opens the door to an incredibly complex problem.
Watson’s results are based on work within a single cell line, testing one xenoestrogen at a time. Even so they seem to show an incredible range of effects. What happens at the tissue level, or at the level of a whole organism that might be exposed at any stage of development? For that matter, what happens to a human—a developing fetus, a newborn baby, a child, an adolescent, an adult—exposed, as we all to some degree have been, to multiple xenoestrogens at the same time?
“That’s a really important question,” Watson says. “We’re at the point of depicting these mechanisms that each compound elicits in a dose and time-dependent way. But when you put them all together, which is undoubtedly how we see them—we know these things are in combinations in the environment—God only knows what these things do together. I think that needs to be studied. We’re really only at the beginning of this.”