Environmental Tobacco Smoke
I read with interest the article from Bermudez et al., "Environmental Tobacco Smoke Is Just as Damaging to DNA as Mainstream Smoke" (EHP 102: 870-874). Environmental tobacco smoke (ETS) is a complex mixture of chemicals resulting from dilution in a confined environment of tobacco smoke. ETS has three forms: 1) sidestream smoke (SS) is produced by a cigarette during the puff intervals, 2) mainstream smoke (MS) is released by the smoker after smoke inhalation, and 3) residual mainstream smoke (RMS), which is a minimal proportion, slowly seeps from the mouth end of a cigarette during the puff intervals. Thus, ETS cannot be identified by one of these components alone. Furthermore, SS or MS determinations cannot be used as a predictor of the concentration of compounds in the ambient air because the composition and the chemical nature of ETS changes dramatically as it ages and is diluted in the environment (the same can be said regarding the prediction of ETS's effects in terms of public health). Therefore, I was rather surprised to read that ETS is equivalent to sidestream smoke (see the Introduction and Material and Methods), so the particulate matter trapped on a Cambridge filter is equivalent to ETS "tar." This statement is obviously untrue and deliberately disregards the evidence that ETS is a dilute system compared to MS and/or SS;
The in vitro tests used to monitor the adverse effects of SS-derived tar trapped on a Cambridge filter consisted of 1) rat alveolar macrophages for the measurement of the electron spin resonance (ESR) to detect the presence of a persistent radical after incubation with the tar solution, 2) isolated rat thymocytes incubated with the tar solution that were then submitted to fluorescence analysis of DNA unwinding to determine DNA damage. Both these assays gave positive results in terms of an effect of the test material employed. After having obtained these results, Bermudez et al. concluded: "to our knowledge, this is the first report of the DNA nicking capability of tar from ETS" (p. 873). I cannot agree for at least two reasons: tar was collected from SS and not from ETS, and a genotoxic effect of SS tar has been known for a while (1,2).
In the article, Bermudez et al. cite work by Hammond et al. (3) indicating macromolecular adduction in people exposed to ETS. Hammond et al. examined the relationship between quantitative measurements of 4-aminobiphenyl-hemoglobin adducts (4-ABP-Hb) in nonsmoking, pregnant women. Surprisingly, only one blood sample was collected at delivery, and a relationship was found between women exposed to ETS (monitored during the third trimester of pregancy by a questionnaire and by wearing a monitor which sampled nicotine by passive diffusion to a filter treated with sodium bisulfate) and the level of 4-ABP-Hb adducts found at the time of delivery. The conclusion of these authors was that the increase in the levels of 4-ABP-Hb was not dramatic and that the public health significance was unclear. Bermudez et al. failed mention a number of studies aimed at detecting increased levels of DNA and hemoglobin adducts in people exposed to ETS, all with questionable or frankly negative outcomes (4-10).
I would suggest repeating the alveolar macrophage study using cells obtained by the same technique (bronchoalveolar lavage) from rats exposed to a real ETS environment, controlling certain parameters: particle concentration, particle size, and carbon monoxide. This would produce much more meaningful information. Alternatively, repeat the alveolar macrophage study using trapped particulate matter carried by persons exposed to an ETS environment and compare it with the material trapped by the filters obtained from devices carried in a smoke-free environment. Phillips et al. (11) were able to prove that, in a confined environment where smoking was permitted, only a median of 2.5% of the particulate matter trapped by portable monitors was of ETS origin.
Angelo Cerioli
Castelleone, Italy
References
1. Ong TM, Stewart J, Whong WZ. A simple in situ mutagenicity system for detection of mutagenic air pollutants. Mutat Res 139:177-182 (1984).
2. Mohtahamipur E, Norpoth K, Archer MC. Urinary excretion of frameshift mutagens in rats caused by passive smoking. J Cancer Res Clin Oncol 108:296-301 (1994).
3. Hammond SK, Coghlin J, Gann PH, Taghizadeh PMK, Skipper PL, Tannenbaum SR. Relationship between envrionmental tobacco smoke exposure and carcinogen-hemoglobin adducts levels in nonsmokers. J Natl Cancer Inst 85:474-478 (1993).
4. Perera FP, Santella RM, Brenner D, Poirier MC, Nunshi AA, Fishman HK, Ban Tyzin J. DNA adducts, protein adducts and sister chromatid exchange in cigarette smokers and nonsmokers. J Natl Cancer Inst 79:449-457 (1987).
5. Bartsch H, Caporaso N, Coda M, Kadlubar F, Malaveille C, Skipper P, Talaska G, Tannenbaum SR, Vieneis P. Carcinogen hemoglobin adducts, urinatry mutagenicty and metabolic phenotype in active and passive cigarette smokers. J Natl Cancer Inst 82: 1826-1831 (1990).
6. Maclure M, Katz RB, Bryant MS, Skipper PL, Tannenabum SR. Elevated blood levels of carcinogens in passive smokers. Am J Public Health 79:1381-1384 (1989).
7. Holz O, Krause T, Schere G, Schmidt-Preuss U, Ruediger HW. 32P-postlabelling anaysis of DAN adducts in monocytes of smokers and passive smokers. Int Arch Occup Environ Health 62:299-303 (1990).
8. Perera F, Mayer J, Jaretzki A, Hearne S, Brenner D, Young TL, Gischman HK, Grimes M, Grantham S, Tang MX, Tsai WY, Santella TM. Comparison of DNA adducts and sister chromatid exchange in lung cancer cases and controls. Cancer Res 49:4446-4451 (1989).
9. Lewtas J, Gallagher J. Comparison of bioindicators of exposure to genotoxic indoor air pollutants. Report no. 600/D-90/166. Washington, DC:U.S. Environmental Protection Agency, 1990.
10. Schere G, Daube H, Adlkofer F. DNA adducts as biomarkers for exposure to genotoxic substances in tobacco smoke. Presented at the Symposium Arbeitsgemeindsc. Molekular und Zytogentetic der Deutschenhaft Krebsgeselleschaft, 24-25 September 1992, Heidelberg.
11. Phillips K, Howard DA, Browne D, Lewsley M. Assessment of personal exposure to environmental tobacco smoke in British nonsmokers. Environ Int 20:693-712 (1994).
Assessing Chemicals for Estrogenic/Hormone-Disrupting Properties: Lessons from Carcinogenicity Assessment
Recent articles and letters in EHP (1-9) have highlighted the growing interest in chemicals that have the potential to mimic estrogens or in other ways disrupt endocrine hormone balances. The specific concerns were listed succinctly in the Wingspread consensus statement of 1991 (10). Any such "new" area of toxicology poses particular problems for those charged with assessing the safety of industrial or other environmental chemicals--all chemicals concomitantly come under suspicion, but the screening assays necessary to assess this toxic potential are usually only in the early stages of development. As a consequence, assay method development and chemical evaluations proceed in parallel, with many potential mishaps along the way. Thus, at this moment, chemical companies and commercial testing laboratories around the world face an apparent toxicological problem of undefined dimensions, but in the absence of agreed techniques by which to assess or solve it. In this situation, valuable parallels are already evident between estrogenicity testing and carcinogenicity prediction.
The field of environmental carcinogenesis was underpinned from the start by data on approximately 50 discrete chemical or environmental exposure situations in which a firm link between chemical exposure and the induction of cancer in humans was established. This reference point of stability is missing with environmental estrogens. In its place are a range of suspected associations with reduced human sperm counts and increases in the incidences of human testicular, prostate, or breast cancer. Thus, the reality or otherwise of a human problem will have to be evaluated concurrently with the development of methods and approaches to reduce, or solve, that assumed problem. This poses unique difficulties that are best recognized at the outset. In contrast to the human situation, LeBlanc (6) described local situations where chemical pollution has been more convincingly associated with endocrine-mediated changes in fish and bird populations. Such effects will be capable of rectification by local cleanup measures, as much as enhanced industrial hygiene can remedy local instances of occupational carcinogenesis. The real concern to address is the validity of the implied extrapolation from local ecological effects to global effects on human populations.
In common with environmental carcinogens, environmental estrogens will be capable of prediction/study using both in vitro and in vivo assays. The initial proliferation of in vitro techniques for the prediction of carcinogenicity, and their subsequent culling to a few useful assays, is well known to all. That such a trend is happening with hormone-disrupting agents is already evident. Thus, McLachlan (2) has described a panel of at least nine chemical receptors that can be linked to reporter genes and developed into in vitro screening tests. In addition, the use of one or more of the available subclones of MCF-7 cells is already being considered for screening purposes, as discussed in EHP by Villalobos et al. (3). In fact, Villalobos et al. have made an early and critical contribution to the field by establishing the problems intrinsic to some of those clones. Such studies were delayed by a decade, to general disadvantage, in the field of environmental carcinogen prediction. Obviously, a period of assay development will be critical to this new field, but a harmonized approach to testing, including the recognition and unanimous rejection of unreliable assays, and early agreement on criteria for activity in the favored assays, will be to the common good. In the field of carcinogen/ mutagen prediction, such harmonization is being attempted at present--probably a decade too late.
Another important generic point concerns the different roles to be played by in vitro and in vivo assays for hormone-disrupting activities. In the field of mutagenesis/carcinogenesis prediction, the Salmonella mutation assay and its analogues rapidly replaced both the existing and the concomitantly developed rodent mutation assays. Eventually, however, these rodent assays came back into use as a means to distinguish which chemicals, from among the myriad in vitro genotoxins, were likely to pose a significant (actual) hazard to humans. Thus, it will be useful to accept that the primary evidence for an estrogenic hazard to humans should derive from functional experiments conducted in rodents. Villalobos et al. (3) have noted that one such in vivo assay [the uterine weight assay (7,8)] is not suitable for large-scale screening, and they used that conclusion to introduce their in vitro studies with MCF-7 cells. Such is acceptable so long as a critical role is retained for functional in vivo assays as the final arbiters of a possible human hazard. Specifically, agents showing estrogenic properties in vitro should be confirmed as being capable of eliciting similar effects in rodents before they are classified as potential estrogen mimics in humans. An additional reason for the early consideration of in vivo functional assays is that some chemicals may be capable of eliciting potentially significant hormone-disrupting effects in rodents in the absence of an ability to elicit estrogenic effects in in vitro systems. Such a situation may arise by the chemical altering the metabolism of natural hormones, by it disrupting intratissue hormonal control mechanisms, or as the result of it being uniquely metabolized to an estrogen mimic in vivo. Such effects will be difficult or impossible to simulate in isolated cell or tissue systems. In that sense an analogy may be drawn between pure estrogen receptor agonists/antagonists and DNA-reactive carcinogens and between rodent-specific hormone-disrupting chemicals and the large variety of mechanistically distinct nongenotoxic rodent carcinogens. If so, valuable lessons could be learned from the maturing field of carcinogen prediction/ assessment regarding human risk estimation and interspecies extrapolation of test data.
The greatest current problem faced by those with the safety of chemicals in their care is that few useful structure-activity relationships (SAR) have yet been discerned for estrogenic/hormone-disrupting agents, as noted earlier by McLachlan (2). Thus, although appropriate modeling may enable the structure of DDT or nonylphenol to be fitted to the estrogen receptor, the very fact that chemicals as remotely related as kepone and nonyl phenol can jointly be referred to as estrogen mimics brings temporary insecurity for essentially all organic chemicals, until, in fact, each is established as inactive in these respects. And again, this situation is strongly reminiscent of SAR in carcinogenesis. There, a subfamily of electrophiles/pro-electrophiles can be recognized, as will probably eventually develop for pure estrogen receptor agonists/ antagonists. However, information on electrophilicity does not alert to the nongenotoxic rodent carcinogenicity of, for example, saccharin or limonene. Further, the carcinogenicity SAR of derivatives of saccharin has nothing to offer the carcinogenicity SAR of derivatives of limonene, and vice versa. Therefore, a precautionary recognition that no single SAR will dominate the toxicology of estrogenicity/hormone disruption will probably aid the development of useful subgroup SARs. Within such a framework, the conclusion of McLachlan (2) that study of the functional properties of chemicals will be a more productive exercise than isolated consideration of their chemical structure/ physicochemical properties is perhaps overly pessimistic. Thus, while it would be unwise to seek or to rely upon simple global structural fragments empirically associated with estrogenic activity, detailed and model-based SAR and quantitative SAR should prove invaluable within structurally coherent series of chemicals. The latter is elegantly illustrated by the resolving power and differential specificity of the quantitative SAR study of estrogenic chlorohydroxybiphenyls recently reported by Waller et al. (9). The guiding principles of predictive SAR developed for carcinogenesis/mutagenesis have been definitively reviewed by Richard (11), and most of those principles and warnings will apply equally to the study of estrogenic/hormone-disrupting chemicals.
If some environmental chemicals are genuinely affecting human sexual development or endocrine function, then they must be rapidly identified and regulated. That will be achieved most effectively if the basic tenets of toxicology are accepted to apply to this new endeavor. These tenets are that activity observed in vitro is only indicative of activity in vivo, that with rodent studies the route of chemical exposure and the frequency of administration of the test agent are often crucial to the outcome of the experiment, and that toxic potency usually varies over many orders of magnitude. Finally, it is important to note that synthetic estrogen mimics currently occupy most attention, with only passing reference being made to their naturally occuring analogues. Thus, in the Wingspread statement (10,12) the balance is set at "a large number of man-made chemicals and a few natural ones." In the field of carcinogen assessment, it took over a decade to reverse a similar preliminary assumption (13). In summary, while attempting to respond effectively to this new toxicological concern, we should guard from the outset against the eventual publication of an analogue of that seminal paper by Ames and Gold (13), this time entitled "Too Many Rodent Estrogens."
John Ashby
Zeneca Central Toxicology Laboratory
Alderley Park, Cheshire, UK
References
1. Colborn T, vom Saal FS, Soto A.M. Developmental effect of endocrine disrupting chemicals in wildlife and human. Environ Health Perspect 101:378-385 (1993).
2. McLachlan JA. Functional toxicology: a new approach to detect biologically active xenobiotics. Environ Health Perspect 101:386-387 (1993).
3. Villalobos M, Olea N, Brotons JA, Olea-Serrano MP, Ruiz de Almodovar JM, Pedraza V. The E-screen assay: a comparison of different MCF-7 cell stocks. Environ Health Perspect 103:844-845 (1995).
4. Wolff MS. Environmental estrogens (letter). Environ Health Perspect 103:784 (1995).
5. Safe SH. Response [to Wolff] (letter). Environ Health Perspect 103:784-785 (1995).
6. LeBlanc G. Are environmental sentinels signaling? Environ Health Perspect 103: 888-890 (1995).
7. Hertz R. The estrogens problem-retrospect and prospect: In: Estrogens in environment II. Influences on development (McLachan JA, ed). New York:Elsevier, 1985; 1-11.
8. Jansen HT, Cooke PS, Porcelli J, Liu TC, Hansen LG. Estrogenic and antiestrogenic actions of PCBs in the female rat: in vitro and in vivo studies. Reprod Toxicol 7:237-248 (1993).
9. Waller CL, Minor DL, McKinney JD. Using three-dimentional quantitative SAR to examine estrogen receptor binding affinities of polychlorinated hydroxybiphenyls. Environ Health Perspect 103: 702-707 (1995).
10. Colburn T, Clement C, eds. Chemically-induced alterations in sexual and functional development: the wildlife/human connection. Princeton, NJ:Princeton Scientific Publishing, 1992.
11. Richard AM. Application of SAR methods to non-congeneric databases: issues and approaches. Mutat Res 305:73-98 (1994).
12. vom Saal FS. Environmental estrogenic chemicals: their impact on embryonic development. Hum Ecol Risk Assess 1:3-15 (1995).
13. Ames BN, Gold LS. Too many rodent carcinogens. Science 249:970-971 (1990).
Radon Risks
My co-authors and I were delighted to have our article, "Effects of Residential Mobility on Individual versus Population Risk of Radon-Related Lung Cancer," published in the December issue of EHP (103:1144-1149). In light of two problems, however, we thought you might appreciate the following feedback.
The first problem is relatively minor: we found two typos in equations. In the second equation on p. 1145, the subscript on Pop in the second term after the equals sign should be i, not j. Also, all-capital letters for variable names were changed to lowercase letters. Thus, our "LOG" became "log" except for the last term in the first equation on p. 1145, where it is "Log." (We doubt the latter will cause any confusion.)
The second problem we consider more serious. The summary of our article on the "In This Issue" page (p. 1076)--never shown to us before publication--is factually incorrect. Our article reports that although the population risk of radon is likely to be as previously reported, the risk faced by individuals currently living at high radon exposures is much less than implied by the work of the EPA due to the effects of residential mobility. We took great pains to explain this quite clearly in the article. The summary states, however, that "Warner et al. report that estimates of radon-related lung cancer risks are lower than originally thought when residential mobility is taken into account," not distinguishing between population and individual risk. The summary continues, incorrectly, that "Because most people move about 10 times during their lives, potential exposure in the 7% of homes with elevated radon is actually well below levels that would result in elevated risks for lung cancer." The exposure in those homes is precisely what the EPA says it is, and the cumulative population risk of lung cancer associated with people living at those homes is, collectively, exactly what the EPA estimates (assuming the BEIR IV model is correct, as we do). The point is that individuals currently living in such homes will have a lower risk because they will move frequently throughought their lives and hence will live at lower levels of exposure most or all their years. As a consequence, as we explain the the paper, the distribution of individuals' lifetime exposures is much more tightly concentrated about the mean than is the distribution of exposures in homes per se.
We suggest that the "In This Issue" summaries be approved and edited, as needed, by the authors. The summary of our paper is wrong and misleading. We haven't yet heard from anyone confused by this, but we are disappointed and concerned.
Kenneth E. Warner
Department of Health Management and Policy
University of Michigan
Ann Arbor, Michigan
Last Update: May 6, 1997
