Relative Potency of Xenobiotic Estrogens in an Acute in Vivo Mammalian Assay S.R. Milligan, A. V. Balasubramanian, and J.C. Kalita Physiology, Biomedical Sciences Division, King's College, London, United Kingdom based on the assumption that estimates of The in vii effcts ofxenoestrogens are ofinterest in relation to their potential health risk and/or potency from early responses would mini- beneficial efficts on humans and animals. However, the apparent in yim potency ofthe examined mize complications arising from metabolism, response can be confounded by a short half-ife, and the metabolism ofestogens is very dependent dearance, rapid dissociation from receptors, on the nature of conversion and/or inactivation. To minimize such variables, we examined the etc. The early response chosen was the rapid estrogenic potency of a range of xenoestrogens in an acute in vivo assay-the stimulation of increase in uterine vascular permeability increased uterine vascular permeability in ovariectomized mice 4 hr after subcutaneous administra. [apparent within 15-30 min, with very large tion. While estradiol (E2) and estriol (F3; a relatilyweak natural estrogen) readily induced vascu- responses (three- to fourfold increase) within lar responses [median effiv dose (ED50) 10-9 mol], much higher amounts of xenoestrogens 3-4 hr] of estrogen exposure (5,6). This were required. Bisphenol A was about 10,000-fold less potent than E4 and E, and octylphenol response has the advantage that it can be and nonylphenol were about 100,000-fold less potent; dioctyl phthalat, beyl buty ptalate, readily quantified by monitoring the extrava- dibutyl phtbalate, and trichlorinated biphenol produced no effect. Coumestrol was the most active sation of intravenously administered [1251] phytoestrogen, with an ED50 between 10-6 and 10-7 mol; genistein wa about 10-fold less potent labeled human serum albumin (7). than coumestrol, and neither daidzein nor formononetin produced any marked efflect, even at doses up to 1O5mol. All increases in vascular permeability could be blocked by the pure anti- Methods estrogen ICI 182,780. There was no evidence that any ofthe compounds could act as an aniesro- Animals. Female Swiss albino mice (A. Tuck gen in his assay or that they could exert ynegistc ef in.combion. These results indicate & Son Ltd., Battlesbridge, Essex, U.K.), that even short-term exposure to most of the xenobiotic estogens can induce typical estrogenic approximately 3 months ofage and weighing efFecs in vivo, but their esuogenic potency is very weak even when assessed in an acute response. 25-35 g, were maintained under constant Key wori: environmental estrogens, phytoestrogens, uteru, vasular permeability, xenobiotic. conditions of lighting (lights on from 0600 Environ Health Prect 106:23-26 (1998). [Online 2 January 1998] hr to 1800 hr) and temperature (21 ± PC) hatt:p//ehpnatl.niehs.nib.gol/docs/I998/106p23-26milgan/abseract.btmI and fed on a pelleted diet (Economy Rodent Maintenance, Essex, UK) ad libitum. All experiments were performed on ovariec- There is considerable public and scientific on the complex in vivo pharmacokinetics tomized animals. Ovariectomies were per- interest in the potential effects of the expo- associated with uptake, distribution, formed under tribromoethanol anesthesia at sure of humans and other animals to envi- metabolism, and excretion. least 2 weeks before the start of each experi- ronmental estrogens. These nonsteroidal There have been virtually no systemat- ment. All test compounds were administered estrogenic compounds essentially fall into ic, detailed studies of the nature and time by sc injection in a volume of0.1 ml. two categories: they are derived from either course of the estrogenic effects of these Chemicals. 17,-Estradiol (E2), genistein dietary sources or industrial sources. The chemicals in vivo in mammals. As a first (4',5,7-trihydroxyisoflavone), and daidzein recent surge ofinterest in this area has large- stage to assessing the in vivo estrogenic (4',7-dihydroxyisoflavone) were obtained ly come from the use of very sensitive in activity, this study was undertaken to from Sigma Chemical Co. Ltd. (Dorset, vitro assays, which have allowed the identifi- establish the relative potency of the com- U.K.). Coumestrol (3,9-dihydroxy-6H-benzo- cation of weak estrogenic activity in many pounds in an acute in vivo mammalian furo-[3,2-c]-[l]benzopyran-6-one) and for- industrially derived chemicals (1). This has assay. In vivo, the complex pattern of estro- mononetin (7-hydroxy4'-methoxyisoflavone) been coupled with natural and experimental genic action on the uterus can be dissociat- were obtained from Acros Organics observations of estrogenic effects in aquatic ed into a group of early responses (e.g., (Springfield, NJ) and Extrasynthese (Genay, animals (e.g., fish) living under conditions increased vascular permeability, uterine France), respectively. 4-Nonylphenol, 4-tert- of continuous exposure to the compounds edema, uterine eosinophilia) and a group octylphenol, bisphenol A [2,2-bis(4 hydrox- (2). The possible health benefits ofdiets rich of later responses (a host of biochemical yphenyl)propane, dioctyl phthalate, benzyl in phytoestrogens has provided an addition- changes leading to cell division, differentia- butyl phthalate, and dibutyl phthalate were al and alternative focus ofattention (3). tion, and uterine growth) (5). obtained from Aldrich Chemical Co. Ltd. While the existence of dietary and envi- Estimates of estrogenic potency are (Gillingham, U.K.). Trichlorinated biphenol ronmental compounds with estrogenic markedly affected depending on which of (TCB; 3,4,3',4'-tetra-chlorobiphenyl) was a activity is clear, speculation linking them to these two groups of responses are used as gift from P. Darbre (University of Reading, epidemiological data on fertility, reproduc- the baseline. Thus while estriol and estradiol Reading, U.K.). Stock solutions ofthe chemi- tive abnormalities, and disease is severely are relatively similar when assessed in terms cals were normally dissolved in ethyl alcohol hampered by our very poor understanding of the ability of single subcutaneous (sc) or DMSO. The injection vehicle was either of the in vivo biological activity of these injections to induce early uterine responses, arachis oil (phytoestrogens) or saline (others) compounds. In vitro assays have established such single exposures to estriol are consider- that all the compounds are relatively weak ably less effective than estradiol in produc- Address correspondence to S.R. Milligan, estrogens, with both binding affinity to the ing sustained effects due to the rapid disso- Physiology, Biomedical Sciences Division, King's its College, Strand, London, WC2R 2LS, United estrogen receptor and in vitro biological ciation ofestriol from receptor (5). Kingdom. activity usually <1/1,000 of estradiol (4). The rationale behind this study was to We are grateful to the Medical Research Council for However, it is difficult to extrapolate from use one of the early specific estrogen-stim- support. J.C.K., a Commonwealth scholar from such in vitro data to the in vivo situation in ulated responses as the basis for comparing India, was supported by the British Council. which the biological activity is dependent relative estrogenic potencies. This was Received 10 June 1997; accepted 1 1 September 1997. Environmental Health Perspectives . Volume 106, Number 1, January 1998 23 Articles * Milligan et al. containing 25% DMSO or alcohol. The pure |* Control * Estriol A Octylphenol | Control M Esiriol A Genistein antiestrogen ICI 182,780 (a gift from A. A 17,B-Estradiol * BisphenolA 0 Nonylphenol IA* 17p-Estradiol Coumestrol 0 Daidzein Wakeling, Zeneca Pharmaceuticals, 30 30 Macclesfield, U.K.) was dissolved in alcohol to give a stock solution of 10 mg/ml ethanol cc 25 Q 25 and administered in an arachis oil vehide. LU .20 Vascular permeability changes in 20 :zU response to test treatment. A quantitive 00 index of the permeability of the uterine vas- L. 15 IC. 15 culature 4 hr after sc injection of the test U compound was obtained from the leakage of =0 10 o 10 radiolabeled albumin from the circulation .E L. U (7,8). In brief, 3.5 hr after the injection of CD the test substance, 50 pl 0.5 ,Ci [1251]- U 5 A- 5 labeled human serum albumin was injected 3 into the jugular vein of mice anesthetised 0 E~MMARW" with tribromoethanol. Thirty minutes later, -14 -13 -12 -11 -10 -9 -8 -7 -6 -5 -4 -3 -2 -14 -13 -12 -11 -10 -9 -8 -7 -6 -5 .4 a blood sample was drawn from the subor- Log of injected doses (mol) Loo of injected doses (mol) bital canthal sinus with a heparinized capil- Figure 1. The dose-response effects of 1713-estradi- Figure 2 The dose-response effects of 17f3-estradi- lary pipette and the animals were killed by ol, estriol, and the environmental estrogens bisphe- ol, estriol, and the phytoestrogens coumestrol, cervical dislocation. The blood sample was nol A, octylphenol, and nonylphenol on uterine vas- genistein, and daidzein on uterine vascular perme- centrifuged to provide a 100 pi plasma sam- cular permeability 4 hr after subcutaneous adminis- ability 4 hr after subcutaneous administration in ple. The uteri and a sample of thigh musde tration in ovariectomized mice. Saline controls were ovariectomized mice. Saline controls were included were removed, briefly washed in saline, and included for each experimental data point Vascular for each experimental data point. Vascular perme- weighed. The radioactivity in the uterus, permeability was determined by the extravascular ability was determined by the extravascular accu- accumulation of [12511-albumin and was expressed mulation of [12511-albumin and was expressed as plasma sample, and musde was determined. as extravascular albumin volume (EAV; see extravascular albumin volume (EAV; see Methods). Previous studies on the uterus (8,9,10) have Methods). The data are expressed as mean ± stan- The data are expressed as mean ± standard error suggested that the blood volume of the dard error (n = 6-12). Where no error bars are (n = 6-12).