Effect Measure

Bisphenol A. What’s all the noise about?

You’re in a crowded bar near the airport and your co-worker is trying to tell you something important. She wants you to do something before you drive her car to the garage for her. She is heading out of town. But you can’t hear her over the din from the crowd. It’s too noisy, too much cross talk. Later you discover she was telling you the gas gauge is broken and the tank almost empty. But you know that. After you ran out of gas on the freeway. Now imagine you are a developing fetus. Genes in your nervous system are turning on and off in a precise sequence in response to what’s going on in your developing brain. Your neurons are growing, making new connections, responding to the cues from other parts of the system that are also developing. The signals that coordinate this involve very tiny amounts of chemicals coursing through the blood stream. Hormones, like the the estrogens. But there’s a lot of noise from artificial chemicals that also stimulate cells, but not in response to a coordinated development plan. Chemical noise from the environment.

That’s the concern about environmental endocrine disruptors. They are like the din in the crowded bar and we worry that important signals will be missed or misinterpreted or the noise will be mistaken for a message. One of the most ubiquitous of the endocrine-like noises comes from a chemical found in plastics, bisphenol-A (BPA):

It’s hard to avoid bisphenol A. One of the highest-volume chemicals in commercial production, it’s the starting material used to make polycarbonate plastics. Those are the hard, clear plastics used in baby bottles, flatware, watercooler bottles, and the work bowls of food processors. Bisphenol A (BPA) also serves as an essential ingredient of epoxy resins used to line food and beverage cans and even to seal cavity-prone teeth.

But BPA doesn’t stay put. It inevitably leaches into foods and people’s mouths, such that traces of the chemical now show up in everyone’s body. (Janet Raloff, Science News)

Scientists on two different panels of the NationalToxicology Program of the National Institute of Environmental Health Sciences have looked at the large scientific literature on BPA and come up with slightly different conclusions. One panel did not consider animal studies involving injected BPA on the grounds this was not the usual route of exposure. It bypasses metabolism by the liver. The other panel included those studies (some 200 of them, for a total of 700). The first panel was less concerned, although it did make and exception for concerns “that exposure to the chemical might perturb neural development in the womb or shortly afterward.” Quite an exception. The second panel was more alarmed, pointing to animal data associating BPA and effects that “alter rodents’ and other lab animals’ sex-specific behaviors, perturb developmentally important hormones, boost fat cell numbers and their accumulation of lipids, foster precancerous changes in cells, and induce insulin resistance, a harbinger of diabetes.” The reports of the two NTP panels will be merged and a final assessment to be issued later. Good luck.

Meanwhile more reports are coming out about BPA effects on animals at doses comparable to those to which people are exposed. The range of reported effects is broad and hard to integrate. It is likely they all reflect one or a few basic biological mechanisms, but what that might be we don’t know as yet. While it might also be true that the risk to most people is small, the ubiquity of exposure — almost everyone — is such that even rare effects can come up fairly often. Thus an effect of one in a million per year is still 300 people a year in a country with a population of 300 million. So the risk may be appreciable to the population without being particularly high for any specific individual. But there aren’t any human epidemiological data as yet.

So pretty much we don’t know about the risk to people. But the word is getting around and the market will likely be a more powerful lever than the absent government regulators. My daughter has already ditched all the plastic baby bottles for her new baby (I had nothing to do with it; she read about BPA on one of the many “mommy blogs”). She ordered glass ones through Amazon because the store was all sold out.

One of the iconic scenes in recent American film is when the young Dustin Hoffman in The Graduate is taken aside by the family friend at his graduation party for some friendly career advice: “I just have one word for you,” the friend whispers in his ear. “Plastics.”

Indeed.

Comments

  1. #1 Gerard Harbison
    October 4, 2007

    OK, so the binding constants of bisphenol-A to the two estrogen receptors are both around 1 uM. Typical levels of circulating bisphenol-A in the blood are 2 ppb, or somewhat under 10 nM. Explain to me how an inhibitor with a K_I of 1 uM inhibits its receptor at 0.009 uM concentration.

    If you like, I can supply you with the relevant competitive binding equation.

  2. #2 Mark Powell
    October 4, 2007

    Nicely done! Great analogy at the beginning, great discussion of a problematic but inscrutable problem. Great solution discussion at the end. Golly, that Revere sure knows what he’s doing.

  3. #3 revere
    October 4, 2007

    Mr. Harbison: You’re right. Nothing designed like a bumblee could possibly fly. But it does. The literature showing effects at environmental levels is getting quite large. But you’re safe. It’s impossible! LOL.

  4. #4 hibob
    October 4, 2007

    revere: big golf clap for your response to Harbison.
    Harbison: there are two families of estrogen receptors (that I know of), but each family can vary in the splicing of segments that are involved in hormone binding.
    Would it make sense if 1, there are different estrogen receptors expressed transiently during fetal development for which bisphenol has a lower IC50?, or 2, there are cells/tissues that accumulate bisphenol enough to make it an effective estrogen disruptor despite its high IC50 (looks like the fat:blood partition is about 3.3, so it would have to be an active mechanism).

  5. #5 Gerard Harbison
    October 4, 2007

    Mr. Revere

    The epidemiological literature showing effects of EM fields on health is also quite large. Nonetheless, physicists discount them, because of lack of a feasible mechanism. You, on the other hand, have postulated a mechanism, and that mechanism is disruption of the endocrine system. In fact, the people who are pushing bisphenol A as an ED have volunteered the estrogen receptor for the job. The biochemistry of binding of small molecules to hormone receptors is well known. And the numbers don’t add up. Unless you can provide me with a receptor with a dissociation constant for bisphenol-A in the neighborhood of 10 nM, you’re in the same category as the EMF crowd. The antagonism of estrogens will be insignificant, and there will be no physical basis for the effect.

    It’s a shame you don’t take the scientific method a little more seriously, and approach scientific challenge so defensively.

    hibob: thanks for actually arguing the science. The study I read (JBC, circa 2000, can provide an exact ref. if you want) suggested both alpha and beta receptors were in the same neighborhood of Kd. It is possible, but implausible, a splice might have a lower dissociation constant than either.

  6. #6 revere
    October 4, 2007

    Mr. Harbison: So evidence of biological effects at those doses isn’t scientific evidence? BTW, physicists are now trying to come up with mechanisms for non ionizing biological effects because the effects have been shown in vitro, not to mention the epidemiological evidence. Maybe epidemiology isn’t science in your book? Only physics counts? You might wish to look at Physics Today for one example. I would say your riposte was more than a little defensive but that’s not really relevant.

    Here’s the point: you say your receptor kinetics argument means it couldn’t happen. Yet there are a number of well conducted experimental articles showing biological effects at the levels you say there can’t be any effects. So who isn’t taking scientific evidence seriously?

    Maybe you should tell me what counts as scientific evidence and what doesn’t. If two results conflict, that means one is right and one is wrong? Or both are wrong? Or both are right once you understand what you are looking at? What?

    The demarcation problem (look it up) is deep. Not that you’d know it from your glib rendition. Not only glib but a rhetorical tactic, to which you are entitled (like me you are a blogger, after all) but when you get on your high horse I’ll call you on it.

  7. #7 ac
    October 4, 2007

    Gerald: Models need to match the evidence, not the other way around. If your biochemical model of choice can’t explain the measured effects of BA (or EMF for that matter), then your model is lacking.

  8. #8 Gerard Harbison, FCD
    October 4, 2007

    Gerald: Models need to match the evidence, not the other way around. If your biochemical model of choice can’t explain the measured effects of BA (or EMF for that matter), then your model is lacking.

    It’s Gerard.

    The evidence is that bisphenol-A at its current level of occurrence in biological tissue does not bind to estrogen receptors sufficiently to cause the effects it’s postulated to have. The prevalent hypothesis is two orders of magnitude away from working. Given that, I think good alternative hypotheses are investigator incompetence and ideological bias.

    But people are always prone to attribute strange effects to physical phenomena they don’t understand. Magnetism is a particular favorite.

  9. #9 Gerard Harbison, FCD
    October 4, 2007

    You might wish to look at Physics Today for one example. I would say your riposte was more than a little defensive but that’s not really relevant.

    Yeah. Too bad they didn’t have a control group.

    They claim fields of 1 – 2 V/cm have effects on cell division. What would you say the field across a typical cell membrane is? Let’s see, a couple of hundred millivolts potential, a distance of 5 nm…

  10. #10 Mark Powell
    October 5, 2007

    Hey Revere, can you take on the latest seafood/mercury dustup? See http://blogfishx.blogspot.com/2007/10/tainted-science-about-mercury-in-fish.html for starters. And would you get an email address for your blog? This is akward. How about effectmeasure@gmail.com? It’s easy to set up.

  11. #11 bar
    October 5, 2007

    Revere: Thank you for this page. There are a lot of “new” pollutants in our environment, about which I am far more concerned than those known. (e.g. CO2.)

    More like this please.

    Gerard: I suppose you also have proof that there is no so far undetected catalytic possibility?

  12. #12 marquer
    October 5, 2007


    Mr. Harbison: So evidence of biological effects at those doses isn’t scientific evidence?

    It is quite possible both that the effects Revere refers to are real, and that Mr. Harbison’s critique is correct.

    Let us bear in mind that what superficially appear to be estrogenic effects due to estrogen receptor binding may in fact be due to interferences with other mechanisms that have an indirect effect upon estrogenically responsive sites and systems.

  13. #13 revere
    October 5, 2007

    Gerard: Yes, that bumblebee can’t fly. Everyone is wrong but you. Some scientist. Just disregard the evidence it if doesn’t match your preconceived notions. As for ideological bias, what’s the name of your blog again? The Right Wing Professor? (as Gerard knows, this is not snark; that’s actually the name of his blog). BTW, the way a real scientist would phrase the point he makes is this: “given some numbers I have on binding constants (which estrogen receptor, which model system, under what conditions?) what is the explanation for the observed biological effects?” Instead Gerard says, “I have these numbers and they imply that the biological effects being seen in studies which I haven’t bothered to critique but whose reliability I question a priori because they disagree with my libertarian bias not to interfere with the vaunted market can’t possibly be true.” Some scientist.

  14. #14 revere
    October 5, 2007

    Mark: The blog address is Revere’s contact address: phrevere@gmail.com. The Reveres don’t always get a chance to return all emails but we try our best. Depends on what is going on in our “real” lives.

  15. #15 Gerard Harbison, FCD
    October 5, 2007

    Gerard: Yes, that bumblebee can’t fly. Everyone is wrong but you. Some scientist. Just disregard the evidence it if doesn’t match your preconceived notions.

    You haven’t cited any biological studies. Go ahead, and I’ll critique them.

    Nature about 15 years ago published a paper suggesting biological effects of a substance diluted to the point where there was no substantial statistical chance of there being one molecule of the substance in the sample. (It was in support of homeopathy). They proffered some sort of outlandish idea that the substance was imprinting its shape on the water. I, and every other scientifically literate physicist and chemist in the country, discounted it immediately, because we know water doesn’t ‘remember’ solutes on a time scale of much longer than a picosecond. It enabled us to say that whatever was happening (and it turned out to be probably fraud), it wasn’t what the authors thought.

    Using a familiarity with the laws of physics and chemistry to discard what is known to be impossible is, certainly, using preconceived notions, but its exactly how science works, and if you don’t realize that, then you truly don’t understand science. Science is interpreted in terms of what we previously know. We stand on the shoulders of giants.

    Bisphenol A may have biological effects – I look forward to your citing some work you feel is convincing – but it does not have biological effects at 10 nM concentration by competitively inhibiting the estrogen receptor(s), which is the hypothesis you’re pushing.

    As for your other example: there’s a paper every year or so reporting biochemical effects of very low EM fields. Most often, those papers are consigned to the usual fate of bad science. They’re ignored. Occasionally we find out that there was a real effect, but it’s due to something entirely different. Fo example, cisplatin was discovered when a group looking for electric field effects on bacteria accidentally did some electrochemical reactions with their platinum electrodes, and created a new compound that inhibited cell division. A potential difference of 10 V across a 10 cm cell can’t possibly directly perturb cellular contents in any significant way, but it sure as heck can do some electrochemistry. But you have to know some physics and chemistry to figure that out.

    There is no ‘progressive science’, or ‘right wing science’. There is good science and bad science. ‘Progressive science’ and ‘right-wing science’ both fall into the latter category.

  16. #16 revere
    October 5, 2007

    Gerard: Where to start? I am not going to argue with you about right wing and progressive science. I agree there is “just” science. But exactly what it is and what the scientific method is is a far from settled question. It is much more subtle and deeper than most scientists take it to be. Causal necessity, for example, is not an empirical fact. You cannot find it with an experiment but you can infer it from scientific evidence. Exactly what it is is a matter of debate. It is a strange beast. I’m pretty well aware of the fate of scientific evidence having spent several years in a history of science department at one point in my life and having written articles on the subject in peer reviewed science journals. I am also editor in chief of a journal so I have an appreciation for what peer review means and how it works. I also know the literature on EMFs quite well and have published on the subject in an IEEE journal. So I am not just blowing smoke. Maybe you know this field thoroughly and maybe you don’t. You haven’t given me any evidence to think you do except to parrot arguments from nay sayers like Ken Foster. But that’s a side issue. Back to BPA.

    Here’s one way to start. As a recent news article in JAMA notes, older data that BPA is a weak estrogen mimic was based on estrogen receptors in the nucleus but there are estrogen receptors on the cell membrane, too, and estrogen mimics like BPA are equipotent with estrogen there. See Zsarnovszky A et al. Endocrinol. 2005;146[12]5388-5396, Wozniak AL et al. Environ Health Perspect. 2005;113:431-439). There are also two estrogen receptors, ERa and ERb (something you don’t seem to realize). Also see the current paper by Vandenberg et al in Reproductive Toxicology – 1 pM of BPA is able to cause changes in cell function. Erb has about 10x higher affinity than ERa and depending on the co-regulatory protein present, the response at Erb can be greatly enhanced and not predicted just based on the Kd. The NTP panel’s report is in Reproductive Toxicology (vom Saal F et al. Reproductive Toxicology). You’ll find all the cites there you want — and which you need. In the words of a colleague who is one of the world’s experts on this, “These are comments from someone who hasn’t taken the time to study the issue.” Maybe someone with an ideological bias? Someone who accuses others of ideological bias but then gets on his high horse and says there is no right wing or progressive science? Could be.

    Let me guess. Your response will be to change the subject back to EMF or some other topic.

  17. #17 wenchacha
    October 5, 2007

    My daughter has already ditched all the plastic baby bottles for her new baby (I had nothing to do with it; she read about BPA on one of the many “mommy blogs”). She ordered glass ones through Amazon because the store was all sold out.

    Of course there is always the original container.

    Just the obvious snark in a public health forum. I used the Playtex “bottle liner” system to give water or expressed milk. I’ve no idea what dangers lurked in the plastic of those little baggies. And the breast milk itself probably had some measurable amount of environmental pollutants.

    My grown kids are healthy for now, anyway! I still have my moments of guilt when I remember how much my baby daughter loved chewing on curly phone cord. It was a favorite teething toy. How much lead was in that beige plastic back there in the 80s?

  18. #18 Gerard Harbison
    October 5, 2007

    Revere:

    No my response is I’ll have to get those papers and read them. But thanks for posting the reference. SInce I explicitly mentioned the two types of estrogen receptors in an earlier post, the claim I’m not aware they exist is just a cheap shot. I don’t even know who Ken Foster is, let alone parroting him; but I have been working on the magnetic resonance of biological systems for a quarter century, as well as doing measurements and computations of electromagnetic fields in proteins, so I’m quite secure in my ability to think about such systems.

  19. #19 Gerard Harbison
    October 5, 2007

    So I took a look at the raw results in two of the papers from the Galveston group. They’re measuring what they say is excess calcium release in cells by a pretty standard fluorescence assay. What’s notable is the minuscule level of their reported responses; in the Environmental Health Perspectives, 113 : 431-439 (2005) paper (1) their excess fluorescence, said to represent BPA stiimulated Ca2+ release, at 1 pM BPA, is approximately 1.2% over basal. It then decreases to 1% at 100 pM; increases to a whopping 4% at 1 nM, and then decreases again to about 2.5% at 10 nM. So what we have, in effect is a ‘response’ that fluctuates wildly over a very small range of 1% to 4% over 4 orders of magnitude increase in BPA concentration. DES and E2, on the other hand, show small but at least monotonic increases with concentration. Minute noisy responses, no persuasive dependence on concentration over a huge range of concentrations, from a material which (as you note) is a pretty ubiquitous contaminant?!?
    Please! (How the heck do you make sure there isn’t picomolar BPA in your solutions?)

    The prolactin release data, obtained at the highest (10 nM) BPA concentration, are less noisy, and they claim a doubling of release over the control. However, they were able to show no time resolution in the response. These are also prolactinoma cell lines, selected for high membrane ER alpha receptor levels.

    The Steroids 72:124134 ( 2007 ) (2) paper is even more worrisome, because it casts doubt on the prolactin data. As I said, the previous paper noted a doubling of prolactin release at 10 nM BPA. Trouble is, they saw the same doubling of release at 1 pM BPA, while the release went down to near normal over the entire range between 10 pM and 1 nM! That’s a heck of a bimodal curve. The calcium data is just as bad as previously (actually it looks like the same data, tut tut!). They report no effect at all of BPA on ERK phopsphorylation.

    This is lousy data in any field, and making public policy on the basis of it is inconceivable. If I saw these data in a paper, without some attempt to explain the wild fluctuations, I’d reject it; it looks like noise.

    I’ll take a look at some of the other work you cited as soon as it come in to our library.

  20. #20 Tony
    October 6, 2007

    “estrogenic effects due to estrogen receptor binding may in fact be due to interferences with other mechanisms that have an indirect effect upon estrogenically responsive sites and systems”

    Great point! I do recall reports of Aryl Hydrocarbon Receptor (AHR) activation leading to induction/suppression of androgen/estrogen responsive genes through various mechanisms totally unrelated to the classical ligand-receptor model.

    There was a recent paper in Science discussing how dioxin can hijack AHR into forming an atypical ubiquitin E3 ligase that recognizes estrogen receptors, causing their ubiquitination/degradation. So, it would seem that endocrine disruption is not as simple as inappropriate activation/inhibition of hormonal receptors.

  21. #21 Ian Musgrave
    January 21, 2008

    Gerard Harbison raised a number of issues about the GH3 paper; unfortunately, he misunderstands the science in this paper. As I have had a fair bit of experience with measuring intracellular calcium, and have measured intracellular calcium in endocrine cells (eg Ann N Y Acad Sci. 1995 Jul 12;763:272-82. Endocrine. 1998 Aug;9(1):71-7. J Immunol. 2006 Jun 15;176(12):7489-94.), so I would like to comment if I may.

    Firstly, endocrine cells such as the GH3 ones used here are a bit of a challenge to measure intracellular calcium ([Ca2+]i) accurately, as they are excitable cells and produce spontaneous oscillations when activated. This necessarily means that the data from activated cells will be somewhat noisy. However, it can quite clearly be seen in the original traces that activation only occurs when drugs (either estrogen, bisphenol A, nonylphenol etc.) are added. Before drug addition, cells are quite stable (see traces in figure 1). Furthermore, when the drugs (including estrogen) are placed on membrane associated estrogen-receptor deficient cells, no activation is seen (figure 3). As the receptor deficient cells still respond to other activators, we can conclude that the effect is specific to membrane bound estrogen receptors. Furthermore, when calcium is omitted from the medium, or in the presence of a calcium channel blocker, the response goes away (eg see figure 6, while the response is not modified by the calcium store releaser thapsigargin). All these results strongly suggest that the effect of BPA (and nonylphenol and endosulfan) are, like estrogen, produced through the membrane associated estrogen receptor.

    Part of Gerard Harbison’s critique rests on the apparent size of the response. Firstly, he re-interprets the ratios as percentages. You can’t do that as the response is strongly non-linear, a 0.04 increase in fluorescence ratio is not a 4% increase in [Ca2+]i. It has to be interpreted in relation to calibration stimuli such as the 20mM K+ stimuli and the thapsigargin stimulus. In resting GH3 cells, basal [Ca2+]i is roughly 80 nM, the thapsigargin data indicate that estrogen is producing around a 200 nM increase in [Ca2+]i, and a quick back of the envelope calculation for the 20mM K+ stimulus gives roughly the same value. This would mean that the increase produced by BPA is roughly 80-100 nM [Ca2+]i. This is a substantial and physiologically relevant increase.

    Another part of the critique rests on the supposed “unphysiological” changes in the graph of BPA. However, biological data is noisy, and we almost never see ideal concentration-response curves in real experiments. The curve for BPA is perfectly compatible with standard biological and experimental variation (and what about nonylphenol, which has pretty much a standard concentration response curve). Also, on the curve, using log values is not “weaselling”, it is standard and accepted practice for displaying concentration response curves (in part because of the equations governing receptor occupancy follow log distributions). Pick up any copy of the British Journal of Pharmacology and you will see almost every article using the same approach.

    The time course experiments are perfectly ok as well. The smallest time point used was 1 minute (which was the fastest practical under the circumstances), but free diffusion and mixing of ligands with receptors happens on the order of seconds, and the response, especially for excitable cells like GH3 or PC-12 cells (the ones I have worked most with), and indeed be maximal or near maximal within 1 minute (even in slow poke tissues like ileal smooth muscle the response to muscarinic agonists is pretty well maximal by 1 minute)
    In summary, the data displays exactly what we would expect if BPA and nonylphenol are acting via estrogen receptors. The increases in fluorescence ratio are consistent with physiological changes in [Ca2+]i, that the “wobblyness” of graphs is with experience of biological and measurement variation.

    Thus it can be concluded that BPA (and nonylphenol and the other xenoestrogens) act on GH3 cells at low (nanomolar) concentrations that are similar to those of estrogen itself, and act on the membrane bound estrogen alpha receptor via a calcium dependent mechanism.

  22. #22 revere
    January 21, 2008

    Ian: Many thanks for your additional comments. I discussed the matter with Fred vom Saal (he chaired one of the two BPA review groups) and his comments were similar to yours, indicating that Harbison was unfamiliar with the field. There is a vast literature on this which he admits he is not familiar with. Since his motives are not scientific but political I am not surprised. He accuses me of ad hominem attacks but you will note that is often (not always) his own modus operandi, usually starting with an accusation of a leftist bias. Yes, I am a (proud) person of the Left. I am also a scientist. I don’t believe my political values are advanced by intellectual dishonesty and I do my best to think as clearly as I can. That doesn’t mean I don’t state positions or opinions, only that I try to make them grounded in reality. Harbison has come here and other scienceblogs to comment (often), sometimes substantively and sometimes in a curmudgeonly fashion. He is welcome. I note that his “worst of science blogs” does not permit comments and he moderates them on his own blog.

  23. #23 Patrick
    June 16, 2008

    As somebody who doesn’t have a dog in this fight (and a physician who was thoroughly lost on the science pretty early in this string–sorry), I just wanted to defend Mr. Harbison. Revere, your tone is just nasty, and apparently without provocation. Clearly you are disagreed with, but surely you realize that is not an “ad hominem” attack? Mr. Harbison may be completely wrong on all counts, but his arguments were presented in a (more or less) scholarly fashion. Your responses all began with a put-down and focused more on a perceived attack than on science–with the exception of the post where you did, in fact, present some studies. Though even that post included the snarky line “I suppose your response will be…”

    What is this, 3rd grade? Geez…

  24. #24 revere
    June 16, 2008

    I don’t know Patrick. If you want to see 3rd grade (maybe 2nd grade) visit Mr. Harbison’s blog:

    http://homepage.mac.com/gerardharbison/blog/RWP_blog.html

    That’s what I was dealing with. And I knew it.