Developing a Marker of Exposure to gull embryos exposed in ovo to DDT and other pesticides (4). Similarly, the decreased Xenoestrogen Mixtures in Human Serum reproductive success of alligators and tur- tles in Lake Apopka, Florida, was linked to Ana M. Soto, Mariana F. Fernandez, Maria F. Luizzi, a spill of kelthane, a pesticide formulation Anita S. Oles Karasko, and Carlos Sonnenschein containing DDE (5). Other evidence that certain chemicals acted as endocrine dis- Tufts University School of Medicine, Department of Anatomy ruptors was revealed when various occupa- and Cellular Biology, Boston, Massachusetts tional exposures in humans were reported. In 1949, aviation crop dusters handling It has been hypothesized that environmental estrogens may play a role in the increasing DDT were found to have reduced sperm incidence of breast cancer, testicular cancer, and other problems of the reproductive system. counts (6). While a single causal agent can While a single causal agent can be identified in cases in which humans have had occupational be identified in cases in which humans exposures, wildlife showing signs of reproductive damage have usually been exposed to a have had occupational exposures, wildlife combination of endocrine disruptors that may act cumulatively. The development of appropriate showing signs of reproductive damage are biomarkers of cumulative exposure, and their measurement at developmental points where usually exposed to a combination of exposure is critical, are required to test the environmental estrogen hypothesis. Measuring levels endocrine disruptors. Additionally, some of of each of the xenoestrogens in blood is a better approximation of real exposure at the target these chemicals may traverse the human organ level than inferring cumulative exposure by estimating from mass balance of dietary levels. placenta and have the potential to However, the cumulative estrogenicity of mixtures cannot be directly concluded from individual adversely affect the developing fetus (7). xenoestrogen plasma levels. Two approaches may be used to assess total load: a) the development Although some of these effects may be of methods to study mixtures of these xenoestrogens, to quantify their cumulative effects, and to obvious upon birth, others may not mani- begin to understand their interactions (i.e., additivity, synergy, antagonism, or independent maturity has action), so that plasma concentrations may be translated into units of activity such as "estradiol fest themselves until sexual equivalents"; and b) the development of methods to separate xenoestrogens from ovarian been reached. For example, boys exposed estrogens in blood and to directly measure the estrogenic activity of the xenoestrogen extract in utero to PCBs appeared normal during using a bioassay. The cumulative activity may be used as a marker of exposure to xenoestrogens. childhood, but at puberty their penises This article reports the development of a method to extract and separate xenoestrogens from were significantly smaller than those of ovarian estrogens using human serum as a source, followed by using a bioassay for determination nonexposed controls (7). of the cumulative xenoestrogen load as "estradiol equivalents." Environ Health Perspect It has been hypothesized that environ- 1 05(Suppl 3):647-654 (1997) mental estrogens may play a role in the declining quantity and quality of human Key words: xenoestrogens, estradiol equivalents, environmental exposure, p,p'-DDE, DDD, semen during the last 50 years, as well as in BBP, BPA, PCBs the increased incidence of testicular cancer and cryptorchidism in males and breast cancer incidence in both females and Introduction males in the industrialized world (8-10). Regarding breast cancer, the main risk fac- Xenoestrogens include several lipophilic, biphenyl (PCB) congeners. More recently, tor for its development is exposure to estro- persistent compounds to which humans nonchlorinated compounds used as antiox- gens throughout an entire lifetime (11). and wildlife have been exposed. Among idants and plasticizers were found to be Moreover, increased plasma levels of "bio- these synthetic chemicals are a number of estrogenic (1-3). The detrimental effects available" ovarian estrogens (not bound to chlorinated organics, such as the insecti- ofenvironmental exposure to xenoestrogens sex hormone-binding globulin [SHBG]) in cides kepone, dieldrin, DDT and its are evident when observing the abnormal postmenopausal women correlate positively metabolites, and some polychlorinated development of the reproductive system in with breast cancer (12). Among the estrogenic xenobiotics, PCBs and DDT were considered suitable This paper was presented in part at the Workshop on Hormones, Hormone Metabolism, Environment, and markers of exposure for breast cancer Breast Cancer held 28-29 September 1995 in New Orleans, Louisiana. Manuscript received at EHP 6 June 1996; manuscript accepted 8 August 1996. because they were released massively into This work was partially supported by grants from the W. Alton Jones Foundation, U.S. Environmental the environment beginning approximately Protection Agency (CR 820301), National Institutes of Health (CA-13410), National Science Foundation (DCB- 50 years ago and they are persistent; their 9105594), the Silent Spring Institute, and the Massachusetts Department of Public Health (DPH 79005-214 Hi 1). The assistance of The Center for Reproductive Research at Tufts University (P30 HD 28897) is gratefully presence in serum may represent cumula- acknowledged. We are also grateful to C. Michaelson for her skillful technical assistance. tive exposure during a lifetime. Three Address correspondence to Dr. A.M. Soto, Tufts University School of Medicine, Department of Anatomy recent studies showed a correlation between and Cellular Biology, Center for Reproductive Research, 136 Harrison Avenue, Boston, MA. Telephone: (617) the occurrence ofbreast cancer and levels of 636-6954. Fax: (617) 636-6536. E-mail: [email protected] Abbreviations used: BBP, benzylbutylphthalate; BPA, bisphenol-A; CD, charcoal-dextran; CDHuS, CD human xenoestrogens. Wolff et al. (13) found that serum in phenol red-free DMEM; DDD, dichlorodiphenyldichloroethane; DDE, dichlorodiphenyltrichloroethyl- serum DDE levels correlated with breast ene; DDT, dichlorodiphenyltrichloroethane; DES, diethylstilbestrol; DMEM, Dulbecco's modified Eagle's cancer incidence in a study of 58 breast medium; DMSO, dimethyl sulfoxide; E2, 17,Bestradiol; EEqs, estradiol equivalents; HCB, hexachlorobiphenyl; HPLC, high-performance liquid chromatography; Pa, pascals; PCBs, polychlorinated biphenyls; PgR, proges- cancer patients and 171 controls that were terone receptor; RPE, relative proliferative effect; SHBG, sex hormone-binding globulin; tD, doubling time. well matched for risk factors and age.
Environmental Health Perspectives * Vol 105, Supplement 3 * April 1997 647 SOTO ET AL.
Another study documented that estrogen- determine the effect of xenoestrogen mix- immediately before use. A 5% charcoal- receptor positive breast cancer correlated tures by means of the E-SCREEN bioassay are 0.5% dextran T70 (Pharmacia-LKB, with higher concentrations of DDE in their presented. This bioassay measures the pro- Uppsala, Sweden) suspension was pre- tissues (14). Krieger et al. (15) studied 150 liferative effect of estrogens on their target pared. CD suspension aliquots of a volume women with breast cancer and 150 controls; cells. The proliferative effect of xenoestro- similar to that of serum aliquots to be each set consisted of 50 African-American, gens is expressed as "estradiol equivalents." processed were centrifuged at 100xg for 10 50 Caucasian, and 50 Asian-American min. Supernatants were aspirated and women. When the data from all ethnic Materials and Methods serum aliquots were mixed with the char- groups were pooled, no significant correla- Chemicals coal pellets. This charcoal-serum mixture tion was observed between plasma levels of was maintained in suspension by rolling at DDE and breast cancer (15). However, Estradiol-17P (E2) was obtained from 4 cycles/min at 370C for 1 hr. This suspen- when the cases and matching controls were Calbiochem (Richmond, CA). DDT sion was centrifuged at 2000xgfor 20 min. evaluated separately, according to their eth- (technical grade), o,p'-DDT, p,p'-DDT, The supernatant was then filtered through nic group, high serum DDE levels were cor- o,p'-DDD, p,p'-DDD, p,p'-DDE, a PCB a Nalgene filter with a pore size of 0.45 related with breast cancer incidence in congener (2,2'3,3'6,6'-hexachlorobiphenyl pm. More than 99% of serum sex steroids Caucasian and African-American women [2,2',3,3',6,6'-HCB]), o-phenylphenol, were removed by this treatment, as mea- but there was no significant correlation in m-phenylphenol, p-phenylphenol, and ben- sured by removal of 3H-estradiol (17); Asian-American women. Evidence of a link zylbutylphthalate were from Ultra Scientific estradiol concentrations after CD treat- between exposure to PCBs and breast cancer (North Kingstown, RI). Bisphenol-A was ment were less than 0.01 pg/ml as mea- incidence is equivocal (16). However, these purchased from Aldrich Chemical Co. sured by radioimmunoassay. CD-treated studies correlated exposure to total PCBs, (Milwaukee, WI). Each of these chemicals sera were stored at -20°C until needed. rather than to the levels ofspecific congeners. was dissolved in ethanol or dimethyl sulf- Samples kept in the freezer for 1 year Epidemiological studies should be oxide (DMSO) before being tested in the maintained their inhibitory properties on conducted involving larger sample sizes to various assay systems. The final solvent con- the proliferation of human estrogen-sensi- clarify the relevance of these findings. It centration in culture medium did not tive breast tumor MCF-7 cells; plasma- and would be impractical to test only for the exceed 0.1%; this did not affect cell yields. blood-derived sera were equally effective. presence of DDT metabolites and PCBs The organic solvents, hexane, methanol, since not all the PCB congeners are estro- and 2-propanol were all of high-perfor- Bioaay for Measuring Estogenicity genic, and many other environmental estro- mance liquid chromatography (HPLC) Theoretical underpinnings of the E-SCREEN gens may also play a role. The newly grade and were purchased from Fisher assay have been published elsewhere identified estrogens may be less persistent Scientific (Pittsburgh, PA). Ethanol and (17,18). Cloned MCF-7 cells were trypsi- than PCBs and DDT metabolites; however, ethyl ether were from Sigma Chemical Co. nized and plated into 12-well plates they are widely used and exposure may (St. Louis, MO) and petroleum ether was (Falcon, Lincoln Park, NJ) at initial con- occur steadily due to their presence in foods from Aldrich Chemical Co. (Milwaukee, centrations of 30,000 cells per well (19,20). (3). Measuring the levels of each of the WI). All other chemicals were of the highest Cells were allowed to attach for 24 hr; then xenoestrogens in blood is a better approxi- quality available from commercial sources. the seeding medium (5% fetal bovine mation of real exposure at the target organ serum in Dulbecco's modified Eagle's level than inferring cumulative exposure Plasma-derived and Blood-derived medium [DMEM]) was replaced with the from estimation of mass balance of dietary Human Serum experimental medium [5% CD-treated levels. However, cumulative estrogenicity of Plasma-derived human serum was prepared human serum supplemented to phenol red- mixtures cannot be directly deduced from from outdated plasma supplied by the New free DMEM (CDHuS)]. A range of con- individual xenoestrogen plasma levels. Two England Medical Center Blood Bank. centrations of the test compounds was approaches may be used to assess total load: Calcium chloride was added to a final con- added to this medium. The bioassay was a) the development of methods to study centration of 30 mmol/liter to facilitate terminated on day 6 (late exponential mixtures of these xenoestrogens in order to clot formation. Blood-derived serum was phase) by removing the media from the quantify their cumulative effects and to from blood samples obtained from healthy wells, adding a cell-lysing solution (10% begin to understand their interactions (i.e., adult volunteers according to a protocol ethylhexadecyl-dimethylammonium bro- additivity, synergy, antagonism, or inde- approved by the Human Investigation mide [Eastman Kodak Co., Rochester, NY] pendent action), so that plasma concentra- Review Committee at Tufts-New England in 0.5% Triton X-100, 2 mM MgCI2, 15 tions may be translated into units ofactivity Medical Center. After donation, blood was mM NaCl, 5 mM phosphate buffer, pH such as "estradiol equivalents," and b) the allowed to clot in glass centrifuge tubes for 7.4) and counting the nuclei in a Model development of methods to separate xenoe- 2 to 4 hr to obtain serum. Plasma- and ZM Coulter Counter Apparatus (Coulter strogens from ovarian estrogens in blood blood-derived sera were clarified by cen- Electronics, Hialeah, FL). and directly measure the estrogenic activity trifugation (2000xg for 10 min), stripped The best estimate of the proliferative of the xenoestrogen extract using a bioassay. with charcoal-dextran (CD), and stored in behavior of a cell population is doubling This cumulative activity may be used as a glass tubes at -20°C until use. time (tD); the time interval in which an marker ofexposure to xenoestrogens. exponentially growing culture doubles its This article describes the extraction and Removal ofSex Steroids by Charcoal- cell number. Determining tD requires mea- separation of xenoestrogens from ovarian Dextran Treatment ofSerum suring cell yields at several time intervals estrogens and phytoestrogens using human Charcoal (Norit A, acid washed; Sigma) during the exponential proliferation phase. serum as a source. Next, methodologies to was washed twice with cold, sterile water A less cumbersome alternative to measuring
648 Environmental Health Perspectives * Vol 105, Supplement 3 * April 1997 XENOESTROGENS IN HUMAN SERUM
proliferation rates is to compare the cell yield p,p'-DDE, and 2,2',3,3'6,6'-HCB. Each of serum, mixed by vortexing, and then 5 ml achieved by similar cell inocula harvested these xenoestrogens was tested for prolifera- of hexane: ethyl ether (1: 1, by vol) was simultaneously during the late exponential tive response at a range of concentrations added to extract the mixture. The mixture phase of proliferation. The proliferative from 10 pM to 100 pM. From these dose- was first agitated on a rotary mixer for 15 effect is measured as the ratio of the highest response curves, a minimal proliferative min and then centrifuged at 2000xg for 5 cell yield obtained with the test chemical to concentration was determined for each min. The organic phase was collected, and the hormone-free control. Under these chemical. The xenobiotics, in three- and the aqueous phase was extracted twice more. experimental conditions, cell yield repre- four-component mixtures, were combined The organic phases were pooled and subse- sents a reliable estimate of the relative pro- at the minimal effective dose that produced quently concentrated to 1 ml by evaporation liferation rate achieved by similar inocula a proliferative effect. Proliferative yield under a nitrogen stream. exposed to different proliferation regulators. results, obtained from the mean cell num- Acid Cleanup ofthe Organic Phase In our experimental design, MCF-7 cell ber, were normalized to the steroid-free Prior to HPLC. Concentrated sulfuric yields were measured 6 days after to; how- control to correct for differences in the ini- acid (0.5 ml) was added to the concen- ever, significant differences between control tial plating density. These results were then trated 1-ml sample of organic phase. The and estrogen-treated cultures are already converted to estradiol equivalents (EEqs), sample was briefly vortex-mixed and then apparent after 4 days (21). an activity unit representing xenoestrogen centrifuged for 5 min at 2000xg. The The estrogenic activity of xenobiotics proliferative effect. The normalized mean organic phase was collected and the aque- was assessed by determining relative prolif- cell number was used to interpolate EEqs ous phase was extracted twice more with erative potency (RPP); this is 10Ox the from a normalized estradiol response curve. 1 ml of hexane. The organic phases were ratio between the minimal concentration The minimal detectable EEq from the pooled and dried completely under nitro- ofestradiol needed for maximal cell yield at E-SCREEN bioassay was determined by gen. The sample was resuspended in 6 days and the minimal dose of the test t-test. A p value < 0.05 was regarded as sig- hexane and then injected into HPLC. compound needed to achieve a similar nificant. The measured mixture EEqs were HPLCAnalysis/Sample Preparation effect, and by measuring relative prolifera- compared with calculated EEqs. Calculated for Measuring Estrogenic Activity by tive effect (RPE); this is 10Ox the ratio EEqs for the mixtures were determined by Bioassay. Separation was performed by the between the highest cell yield obtained adding the measured EEqs ofthe individual method described by Medina and Sherman with the chemical and with estradiol. The xenobiotics. The calculated EEqs for the (24), modified by Sonnenschein et al. (1) RPE indicates whether or not the com- mixtures represent the expected proliferative in a Kratos chromatographer (solvent deliv- pound being tested induces a proliferative effect in the case ofthe three or four xenobi- ery system: Spectroflow 400; gradient con- response quantitatively similar to the one otic combinations interacting in an additive troller and absorbance detector: Spectroflow obtained with E2, that is, a full agonist manner. Differences between the measured 783; injector/mixer: Spectroflow 491; all (RPE = 100), or a proliferative yield signifi- and calculated EEq values were assessed from ABI Analytical, Inc, Santa Clara, CA). cantly lower than the one obtained with using chi-square analysis. A p value < 0.05 Aliquots of 500 pll were injected into a 4 x E2, that is, a partial agonist. When com- was regarded as significant. The results 220 mm Spheri 5 silica column (Brownlee bined with estradiol, antagonists signifi- were expressed as EEqs ± SD. The cumula- Labs, ABI) equilibrated with 100% n- cantly lower the proliferative yield obtained tive effect experiments were conducted in hexane. Xenoestrogens were eluted by a gra- with estradiol alone. duplicated wells and repeated 4 times. dient of two mobile phases: n-hexane (phase A), and n-hexane:methanol:2-propanol, Antagonistic Effect Extration ofXenoestrogens (40:45:15 by vol) (phase B) at a flow rate of The E-SCREEN bioassay was also used to from Plasma-derived Serum 1.5 ml/min and a pressure of 5 x 10- pas- evaluate possible antagonistic activity of and Extrac Cleanup cals (Pa). The gradient was developed as xenobiotics used in the cumulative effect Extraction of xenoestrogens from serum follows: after injection, 100% phase A experiments. Bisphenol-A (BPA) and p,p'- was performed as described by Burse et al. flowed for 2 min; then, the concentration DDE were tested at a range of concentra- (23), with modifications (1). To develop of phase B increased from 0 to 400 ml/liter tions with and without the addition of 100 the extraction procedure, CD-serum was in 15 min and from 400 to 1000 ml/liter in pM estradiol. The proliferative yield was used for spiking and as a "blank" prepara- the next 10 min. The elution profile was compared to the steroid-free control. The tion. The extracts from the "blank" prepa- monitored at 280 nm with a Kratos differences between the estradiol-treated rations are devoid of absorbance at 280 Spectroflow 783 detector. All tested xenobi- groups were assessed by a posteriori Shaffe's nm, and therefore do not interfere with the otics eluted during the first 10 min (1); test (22). A p value of < 0.05 was regarded quantification of added xenoestrogens. In they were evaporated completely and resus- as significant. Results were expressed as addition, CD-treated serum is lacking in pended in ethanol. In the pilot experiment mean ratio of proliferative yield to steroid- estrogenic activity; this also holds for quan- reported herein, the HPLC fractions from free control ± SE. Benzylbutylphthalate titation of the spiked estrogen by means of two 4-ml extracts were combined and resus- (BBP) and 2,2',3,3',6,6'-HCB were also the E-SCREEN bioassay. To assess the recov- pended in 140 pl. This was done to allow tested using the same method. ery of xenoestrogens from serum, 4-ml for monitoring recoveries of spiked chemi- aliquots of CD-treated serum were spiked cals and proliferative activities in replicates. A Method to Determine Cumulaive with a known amount of standard xenoe- The efficiency of the step to solubilize xeno- Effecs ofXenoestrogens strogen and allowed to equilibrate for 2 hr estrogens eluted from the HPLC column The E-SCREEN bioassay was used to at room temperature. was tested by HPLC (percent recovery). evaluate the cumulative effect of four xeno- Sample Extraction. Methanol (2 ml) After these steps had been performed, biotics: benzylbutylphthalate, bisphenol-A, was added to each of two 4-ml aliquots of 100 pl of the remaining resuspended eluate
Environmental Health Perspectives * Vol 105, Supplement 3 * April 1997 649 SOTO ET AL.
Table 1. Estrogenic effect of industrial chemicals mea- 1200 was pipetted into 9.9 ml of the CD-treated sured by the E-SCREEN assay. culture medium (5% CD-treated human serum). This sample is designated "undi- RPE, RPP, luted preparation," and from it, the serial Compound Concentrationa %b %C 1000 2-fold dilutions were made; 1:1, 1:4, 1:8, Estradiol 100 pM 100 100 1:16, and 1:32. After equilibration at room 4-Phenylphenol 10 pM 88 0.001 temperature for 20 min, the preparation 3-Phenylphenol 50 pM 60 0.0002 -800- 2-Phenylphenol 100 pM 30 0.0001 0 was rendered sterile by filtration through a Benzylbutylphthalate 10 pM 90 0.001 0.22 pm (pore size) membrane. The dilu- Bisphenol-A 1 pM 82 0.01 tions were assayed by the E-SCREEN test. A -o 600- 2,2',3,3',6,6'-HCB 10 pM 62 0.001 E "blank" was prepared from CD-treated DDTd 10 pM 80 0.001 human serum by following the fractiona- o,p'-DDT 10 pM 86 0.001 CDI tion protocol that was described for the p,p'-DDT 10 pM 71 0.001 400- amounts p,p'-DDE 10 pM 25 0.001 preparations spiked with known o,p'-DDD 10 pM 73 0.001 ofxenoestrogens. p,p'-DDD 10 pM 19 0.001 Quantafication. To obtain calibration 200- curves, chemicals to be analyzed were alndicates the lowest concentration needed for maxi- injected at five to seven different doses, mal cell yield. bRelative proliferative effect (RPE) is cal- culated as 100 x (PE-1I of the test compound/(PE-1) qf ranging from 0 to 500 nmol. The correla- E2; a value quantitatively similar to E2 (RPE= 100) indi- 0 - tion coefficient for each curve was 0.99. cates that the compound tested is a full agonist, a 0 0.01 0.1 1 10 value of 0 indicates that the compound lacks estro- Concentration, jM Results genicity at the doses tested, and a value significantly Identification ofXenoestrogens lower than E2 (RPE= 100) suggests that the xenobiotic Figure 1. Antagonistic effects of p,p'-DDE and the is a partial agonist. cRelative proliferative potency bisphenol-A. Cells were harvested after 6 days of by E-SCREEN Assay (RPP) is the ratio between E2 and xenobiotic doses exposure. Ordinate: cell number expressed as the The three isomers of phenylphenol, DDT needed to produce maximal cell yields xlOO. All com- treated/control ratio x100; abscissa: concentration. (technical grade), and DDT metabolites pounds designated as full or partial agonists (RPE>15) Data points represent the mean of four experiments; increased cell yields significantly over the hormoneless error bars represent SE. Bisphenol-A was tested with were tested for estrogenicity (Table 1). control (p<0.05). dTechnical grade. (U) and without (E) 100 pM E2. Shaffe's test showed Among the three isomers of phenylphenol, no significant differences between the two treatment p-phenylphenol was a full agonist and the groups. p,p'-DDE was tested with (A) and without (A) most potent (RPE = 88, RPP = 0.00 1); control (p< 0.05 as determined by the 100 pM E2. Shaffe's test revealed significant differ- o-phenylphenol was the least potent and a t test). Therefore, 1 fmol of estradiol/well ences between the two treatment groups (p< 0.05). partial agonist (RPE = 30, RPP = 0.000 1). was determined as the lowest detectable The meta isomer was also a partial agonist; estrogen amount in this assay. Mixtures of its potency fell within the other two xenoestrogens were prepared to assess their -50 (RPE = 60, RPP = 0.0002). cumulative effect. Each xenobiotic was set 0600- o,p'-DDT, o,p'-DDD, and p,p'-DDT at a dose which would yield the minimal were full agonists; meanwhile p,p'-DDE, proliferative effect (the equivalent to 1 400- and p,p'-DDD were partial agonists. DDT fmol estradiol). The xenobiotics BBP, 0 (technical grade) was a full agonist; this 2,2',3,3',6,6'-HCB, BPA, and p,p'-DDE 200- was expected since it consists of 20% of were combined in three and four compo- o,p'-DDT and 80% ofp,p'-DDT. nent mixtures. Cell numbers were converted 0.1 1 10 100 1,000 10,000 In testing for antagonistic effects, to estradiol equivalents by interpolation on Femtomol/well BPA (a full agonist) did not alter the pro- the estradiol dose-response curve (Table 2) liferative effect of estradiol (Figure 1). (measured cumulative effect). The effect of Figure 2. Dose-response curve of estradiol. Cells in were harvested after 6 days of exposure. Ordinate: cell BBP and 2,2',3,3',6,6'-HCB were also each xenoestrogen the mixture was esti- number expressed as the treated/control ratio x100; tested; they had no antagonistic effect mated; cell yields were converted to estra- abscissa: concentration. Data points represent the (data not shown). The partial agonist diol equivalents. These values were added mean of four experiments; error bars represent SD. p,p'-DDE showed statistically significant, to calculate the expected estrogenic effect t-test revealed significant differences between 1 fmol partial antagonistic effects on MCF-7 cell of the mixture. Table 2 shows that the of E2 and the control (p< 0.05). proliferation (Figure 1). measured cumulative effects are signifi- cantly different from the expected values Development ofa Method to Reveal (p < 0.05 as determined by chi-square estradiol (Figure 4A) and o,p'-DDT Cumulative Effects ofXenoestrogens analysis). This suggests that the cumulative (Figure 4B) were performed simultaneously The first step toward this end was to evaluate effect of these xenoestrogens is synergistic with the E-SCREEN test of the extracts. The the sensitivity of the E-SCREEN test by rather than additive (Figure 3). various dilutions of blank extracts obtained means of a detailed estradiol dose-response from CD-treated serum were assayed to curve (Figure 2). At any concentration Testing the Estrogenic Activity assess whether toxic or spurious estrogenic below 1 pM (1 fmol in 1 ml culture Extracted from Plasma-derived Serum activity was acquired during sample prepa- medium), the mean cell number did not To measure the estrogenic effect of the ration. The cell yields were similar to those differ significantly from the steroid-free spiked extracts, dose-response curves to obtained when similar inocula were grown
650 Environmental Health Perspectives * Vol 105, Supplement 3 - April 1997 XENOESTROGENS IN HUMAN SERUM
Table 2. Cumulative effects of a mixture of four xenobiotics. and humans. The lack of deleterious effects BBP, BPA, pp'-DDE, 2,2',3,3',6,6'-HCB, Mean cell #/ Calculated Measured of phytoestrogens on human health may be PM PM PM pM % control ± SD EEq EEq explained by adaptive phenomena that O 0 0 0 100 ± lOa - 0 developed slowly during the coevolution of 2 0 0 0 127±49 - 1.0 plants and mammals. Sudden exposure of 0 0.05 0 0 122 ± 18 - 1.0 organisms to phytoestrogens that are for- 0 0 2 0 125± 12 - 1.0 eign to their natural ecosystems may result 0 0 0 0.5 108 ± 10 - 1.0 in deleterious health effects (25). The role 2 0.05 2 0 271 ±49 3.0 8.0 2 0.05 0 0.5 290 ± 68 3.0 9.0 of the different estrogens (endogenous, 2 0 2 0.5 256 ± 46 3.0 7.3 synthetic, and phytoestrogens) in the etiol- 0 0.05 2 0.5 242 ± 30 3.0 6.6 ogy of cancer incidence and reproductive 2 0.05 2 0.5 329±49 4.0 12.0 disorders awaits further, rigorous epidemi- Cells were harvested after 5 days of exposure. BBP at 2 pM, BPA at 0.05 pM, p,p'-DDE at 2 pM, and 2,2',3,3',6,6'- ological studies that would correlate expo- HCB at 0.5 pM induce the minimal proliferative effect of 1 fmol of EEqs. EEqs of the xenobiotics were interpolated sure to each type of estrogen with their from the E2 dose-response curve (Figure 2) using normalized mean cell number of four experiments ± SD. subsequent health effects. Calculated EEqs were determined by adding the EEqs of the mixture components. Measured EEqs were interpo- In this article, we evaluate the concept lated from E2 dose-response curve (Figure 2) using the normalized mean cell number of four experiments ± SD. of measuring xenoestrogen exposure in The calculated and measured EEqs for each mixture were significantly different using chi square analysis (p< 0.05). humans by separating xenoestrogens from "Negative control: cell yield in 50 ml/liter CD-treated serum alone. endogenous estrogens in human plasma or serum, and by measuring the combined 157 estrogenicity of these xenoestrogens by means of a bioassay using estrogen-target I cells in culture. Evaluation ofBioassays to 10 T Measure Estrogenic Activity LU. The hallmark of estrogen action is consid- E T T ered to be the ability of estrogens to induce 0 proliferation of their target organs (26). LLE Thus, a bioassay that measures cell prolifer- 5. ation as the end point is the most reliable assay to assess estrogenicity. The exposure of male birds, amphibians, and fish to nongonadal estrogens is well documented since the egg yolk protein vitellogenin is a induced by estrogen in the liver, and its 0 A B C D ABC ABD ACD BCD ABCD presence can be detected in blood. The Figure 3. Cumulative effects of a mixture of four xenobiotics. Cells were harvested after 6 days of exposure. Data presence of vitellogenin in the plasma of points represent the mean of four experiments; error bars represent SD. BBP (A) at 2 pM, BPA (B) alt 0.05 pM, males or females whose reproductive tracts p,p'-DDE (C) at 2 pM, and 2,2',3,3',6,6'-HCB (D) at 0.5 pM induce the minimal proliferative effect of 1 fmol of EEqs. should be dormant is indicative of expo- EEqs of the xenobiotics were interpolated from the E2 dose-response curve (Figure 2) using the mean clell number sure. However, this marker cannot be used of four experiments. Calculated EEqs (solid bar) were determined by adding the EEqs of the mixture co)mponents. in reproductively active females since ovar- Measured EEqs (shaded bar) were interpolated from E2 dose-response curve (Figure 2) using the mean c:ell number ian estrogens will induce vitellogenin and of four experiments. The calculated and measured EEqs for each mixture were significantly differentt using chi thus obscure the contribution ofxenoestro- square analysis (p<0.05). gens to vitellogenin plasma levels. An equivalent marker of exposure has not yet in the presence of 5% CD-treated serum, CD-treated human serum supplemented been found in mammals. indicating the absence ofestrogenic activity medium in Figure 4B. Lack of toxic activ- Uterotropic assays measure increases of (Figure 5A). Microscopic observation did ity in this extract is demonstrated by maxi- uterine wet weight, which are due to the not reveal any toxic effects. Additionally, a mal estrogenic activity when each dilution additive effect of cell proliferation, hyper- full estrogenic response was obtained when of the spiked extract was tested in the trophy, and water imbibition. While cell 100 pM estradiol was added to this prepa- presence of 100 pM estradiol (Figure 5B). proliferation of the genital tract is specifi- ration (Figure SA). The spiked serum cally induced by estrogens, different end extract was also tested at various dilutions. Discussion points are affected by hormones other than The recovery ofspiked o,p'-DDT was 84% In addition to being exposed to their own estrogens (3). Most estrogens produce sim- (83 and 85%). Thus, the most concen- steroidal estrogens, humans and wildlife are ilar effects in different species. However, trated dilution ("undiluted") was approxi- also exposed to synthetic xenoestrogens tamoxifen is a full agonist in mice and a par- mately 12 pM. As shown in Figure SB, (pesticides, plasticizers, antioxidants, etc.) tial agonist and antagonist in rats and in the dilution curve of the spiked preparation and natural ones, such as phytoestrogens. human target cells (27). This may explain is in agreement with the dose-response Exposure to synthetic estrogens, such as the discrepancy regarding the designation of curve to o,p'-DDT added directly to 5% DDT, produces adverse effects in wildlife some DDT metabolites as being estrogenic.
Environmental Health Perspectives * Vol 105, Supplement 3 * April 1997 651 SOTO ET AL.
A B target-organ level, since no false positives 1,000 1,000] or negatives have been observed among the estrogens and nonestrogens tested thus far. 800 800 Novel Xenoestogens Identified by the E-SCREEN Assay a) 600- 600 Many chemicals that are present in our C) environment are estrogenic (19,33). 0) a) Phenylphenol is one of the top 10 active 400- CD, 400 ingredients in household pesticides and is the second most widely used ingredient for 200 200 indoor applications (34). We reported pre- viously that p-phenylphenol is a full estro-
gen agonist (20); o- and m-phenylphenol 0 are 1 1,000 10,000 partial agonists (Table 1). o-Phenyl- 0.1 10 100 phenol is used in the rubber industry, in Concentration, pM Concentration, LM agricultural fungicides, and in disinfec- tants; is used in the rubber Figure 4. Dose-response curves of (A) estradiol and (B) o,p'-DDT on MCIF7 cells in culture. Cells were harvested p-phenylphenol after 6 days of exposure. Ordinate: cell number; abscissa: concentration. Data points represent the mean for a set industry and in the manufacture of resins. of four experiments; error bars represent SD. (o) Negative control: cell yielId in 50 ml/liter CD-treated serum alone. There has been controversy regarding which DDT metabolites are estrogenic, and also their relevance to breast cancer inci- A E dence (28). The E-SCREEN assay revealed 1000 - 1000 that all the assayed metabolites are estro- genic. Additionally, full agonists are devoid of antagonistic effects (Figure 1); partial 800 800 agonists are also partial antagonists (Figure 1). The antagonistic activity of these com- a) 600- 600 pounds, although statistically significant, is -3:C) C,Z; biologically marginal (Figure 1).
C-, 400- 400 Cumulative Effects Environmental exposures are very different 200 200 \S11& >~ from most occupational exposures in which one chemical is found to be the causative agent. In environmental exposures, wildlife r 0 1- 1 '| and humans are concurrently or sequen- Undiluted 1:1 1:4 1:8 1:16 1:32 Undiluted 1:1 1:4 1:8 1:16 1:32 tially exposed to a wide array of chemicals Dilutions Dilutions from various sources. Simultaneous expo- sure to mixtures result in Figure 5. Bioactivity of serum extracts on MCF-7 cell proliferation: dose-response curves of serial dillutions of (A) may additivity, blank extract from CD-treated serum and (B) CD-treated serum spiked with o,p'-DDT. The ordinate re-presents cell synergism, antagonism, or independent yield after 6 days in culture. Data points represent the mean for a set of four experiments; error bars represent SD. action. Toxicologists use a toxic equivalency (o) Blank extract. (0) Blank extract supplemented with 100 pmol/liter estradiol. (A) Extract from CD-
652 Environmental Health Perspectives * Vol 105, Supplement 3 * April 1997 XENOESTROGENS IN HUMAN SERUM
This is of concern when deciding how to (Waters Corporation, Milford, MA), since exposed to DES or mycoestrogens. Further measure the cumulative exposure to estro- it results in coelution of unknown moieties optimization of the HPLC separation gens in plasma, fat, or other biological sam- somewhat toxic to MCF-7 cells (1). When protocols may allow the separation of ples. Measuring individual xenoestrogens the Sep Pak step was eliminated, the toxic phytoestrogens from ovarian estrogens. and calculating cumulative exposure by effect disappeared (Figure 5). This method The data and analysis presented here deriving it from the relative estrogenic resulted in 83 to 85% recovery of the represents an attempt to cope with intensi- potency of each chemical may underesti- spiked op'-DDT. Moreover, the biological fying demands by the public and regulatory mate the estrogen load by ruling out syner- activity accurately paralleled the dose- agencies to rigorously evaluate the role of gistic effects. This problem may be obviated response curve to this chemical. Therefore, xenoestrogens in the increased incidence of by directly measuring the effect of the mix- it is now feasible to use this method to male genital tract anomalies, testicular ture extracted from the biological specimen measure, in serum, the cumulative effect of cancer, and breast cancer in human popula- through a bioassay. the xenoestrogens tested thus far, since tions. In this article, we use novel extraction they were separated from ovarian estrogens. protocols and HPLC to measure the cumu- A Method for Measuring We propose to measure the combined lative exposure to xenoestrogens. The full the Cumulative Effect of effect of the compounds eluted during the development of this concept into a reliable Xenoestoens in Hman Serum first 10 min of our HPLC separation as a method to measure exposure in human The extraction/clean-up methodology of marker of exposure to xenoestrogens. populations will require further validation. pesticides from human serum, described by Estrogenic pesticides, PCBs, hydroxylated Additional techniques, aimed at identifying Burse et al. (23) and modified as described PCBs, phenolic antioxidants, and plasticiz- all of the individual estrogenic xenobiotics by Sonnenschein et al. (1), also allows ers elute from the HPLC column during present in serum samples of people sus- quantitative extraction of more polar the first 10 min, whereas ovarian estrogens, pected of being exposed, will help in corre- estrogens (from alkylphenols to estradiol). phytoestrogens, diethylstilbestrol, and lating the relative contribution of each The method was further modified by mycoestrogens are retained for more than xenoestrogen with the observed health eliminating the step that entailed passing 12 min (1). Populations that will be effect in exposed individuals. the sample through a Sep Pak cartridge investigated are not expected to have been
REFERENCES 1. Sonnenschein C, Soto AM, Fernandez MF, Olea N, Olea- 12. Toniolo PG, Levitz M, Zeleniuch-Jacquotte A, Banerjee S, Serrano MF, Ruiz-Lopez MD. Development of a marker of Koenig KL, Shore RE, Strax P, Pasternack BS. A prospective estrogenic exposure in human serum. Clin Chem study of endogenous estrogens and breast cancer in post- 41:1888-1895 (1995). menopausal women. J Natl Cancer Inst 87:190-197 (1995). 2. Krishnan AV, Starhis P, Permuth SF, Tokes L, Feldman D. 13. Wolff MS, Paolo G, Toniolo PG, Lee EW, Rivera M, Dubin Bisphenol-A: an estrogenic substance is released from polycar- N. Blood levels of organochlorine residues and risk of breast bonate flasks during autoclaving. Endocrinology cancer. J Natl Cancer Inst 85:648-652 (1993). 132:2279-2286 (1993). 14. Dewailly E, Dodin S, Verreault R, Ayotte P, Sauve L, Morin J, 3. Soto AM, Sonnenschein C, Chung KL, Fernandez MF, Olea Brisson J. High organochlorine body burden in women with N, Olea-Serrano MF. The E-SCREEN assay as a tool to identify estrogen receptor positive breast cancer. J Natl Cancer Inst estrogens: an update on estrogenic environmental pollutants. 86:232-234 (1994). Environ Health Perspect 103:113-122 (1995). 15. Krieger N, Wolff MS, Hiatt RA, Rivera M, Vogelman J, 4. Fry DM. Sex ratio -kew and breeding patterns of gulls: demo- Orentreich N. Breast cancer and serum organochlorines: graphic and toxicological considerations. Stud Avian Biol a prospective study among white, black, and Asian women. 10:26-43 (1987). J Natl Cancer Inst 86:589-599 (1994). 5. Guillette LJ, Gross TS, Masson GR, Matter JM, Percival HF, 16. Wolff MS, Toniolo PG. Environmental organochlorine expo- Woodward AR. Developmental abnormalities of the gonad and sure as a potential etiologic factor in breast cancer. Environ abnormal sex hormone concentrations in juvenile alligators Health Perspect 103:141-145 (1995). from contaminated and control lakes in Florida. Environ 17. Soto AM, Sonnenschein C. The role of estrogens on the prolif- Health Perspect 102:680-688 (1994). eration of human breast tumor cells (MCF-7). J Steroid 6. Singer PL. Occupational oligospermia. JAMA 140:1249 Biochem 23:87-94 (1985). (1949). 18. Soto AM, Sonnenschein C. Mechanism of estrogen action on 7. Yu M, Hsu C, Guo Y, Lai T, Chen S, Luo J. Disordered cellular proliferation: evidence for indirect and negative control behavior in the early-borne Taiwan Yucheng children. on cloned breast tumor cells. Biochem Biophys Res Commun Chemosphere 29:2413-2422 (1994). 122:1097-1103 (1984). 8. Sharpe RM, Skakkebaek NE. Are oestrogens involved in falling 19. Soto AM, Justicia H, Wray JW, Sonnenschein C. p-Nonyl- sperm count and disorders of the mate reproductive tract? phenol: an estrogenic xenobiotic released from "modified" Lancet 341:1392-1395 (1993). polystyrene. Environ Health Perspect 92:167-173 (1991). 9. Giwercman A, Carlsen E, Keiding N, Skakkebaek NE. Evidence 20. Soto AM, Lin T-M, Justicia H, Silvia RM, Sonnenschein C. for increasing incidence of abnormalities of the human testis: a An "in culture" bioassay to assess the estrogenicity of xenobi- review. Environ Health Perspect 101:65-71 (1993). otics. In: Chemically-Induced Alterations in Sexual 10. Davis D. Medical hypothesis: xenoestrogens as preventable Development: The Wildlife/Human Connection (Colborn T, causes of breast cancer. Environ Health Perspect 101:372-377 Clement C, eds). Princeton NJ:Princeton Scientific Publishing, (1993). 1992;295-309. 11. Kelsey JL, Berkowitz GS. Breast cancer epidemiology. (Review) 21. Soto AM, Silvia RM, Sonnenschein C. A plasma-borne specific Cancer Res 48:5615-5623 (1988). inhibitor of the proliferation ofhuman estrogen-sensitive breast
Environmental Health Perspectives - Vol 105, Supplement 3 * April 1997 653 SOTO ET AL
tumor cells (estrocolyone-I). J Steroid Biochem Mol Biol 29. Soto AM, Sonnenschein C. Cell proliferation of estrogen-sensi- 43:703-712 (1992). tive cells: the case for negative control. Endocr Rev 8:44-52 22. Winer BJ. Design and analysis of single-factor experiments. In: (1987). Statistical Principles in Experimental Design. 2nd ed. New 30. Katzenellenbogen BS, Kendra KL, Norman MJ, Berthois Y. York:McGraw-Hill, 1962:149-260. Proliferation, hormonal responsiveness, and estrogen receptor 23. Burse VW, Head SL, Korver MP, McClure PC, Donahue JF. content of MCF-7 human breast cancer cells grown in the Determination of selected organochlorine pesticides and poly- short-term and long-term absence of estrogens. Cancer Res chlorinated biphenyls in human serum. J Anal Toxicol 47:4355-4360 (1987). 14:137-142 (1990). 31. Reiner GCA, Katzenellenbogen BS. Characterization of estro- 24. Medina MB, Sherman JT. HPLC separation of anabolic gen and progesterone receptors and the dissociated regulation oestrogens and ultraviolet detection of 17-0 oestradiol, zeranol, of growth and progesterone receptor stimulation by estrogen in diethyTstilboestrol or zearalenone in avian muscle tissue extract. MDA-MB-134 human breast cancer cells. Cancer Res Food Addit Contam 3:263-272 (1986). 46:1124-1 131 (1986). 25. Shutt DA. The effects of plant estrogens on animal reproduc- 32. Zysk JR, Johnson B, Ozenberger BA, Bingham B, Gorski J. tion. Endeavour 35:110-113 (1976). Selective uptake of estrogenic compounds by Saccharomyces 26. Hertz R. The estrogen problem-retrospect and prospect. In: cerevisiae: a mechanism or antiestrogen resistance in yeast Estrogens in the Environment. II: Influences on Development expressing the mammalian estrogen receptor. Endocrinology (McLachlan JA, ed). New York:Elsevier, 1985;1-1 1. 136:1323-1326 (1995). 27. Martin L. Estrogens, antiestrogens and the regulation of cell 33. Soto AM, Chung KL, Sonnenschein C. The pesticides endosul- proliferation in the female reproductive tract in vivo. In: fan, toxaphene, and dieldrin have estrogenic effects on human Estrogens in the Environment (McLachlan JA, ed). New estrogen sensitive cells. Environ Health Perspect 102:380-383 York:Elsevier/North-Holland, 1980;103-130. (1994). 28. Safe S. Environmental and dietary estrogens and human health: is 34. Grossman J. Dangers of household pesticides. Environ Health there a problem? Environ Health Perspect 103:346-351 (1995). Perspect 103:550-554 (1995).
654 Environmental Health Perspectives * Vol 105, Supplement 3 * April 1997