Estrogenic effects of environmental chemicals: An interspecies comparison

Comparative Biochemistry and Physiology, Part C 141 (2005) 267 – 274 www.elsevier.com/locate/cbpc

Estrogenic effects of environmental chemicals: An interspecies comparison Christel M. Olsen a,*, Elise T.M. Meussen-Elholm a, Jan K. Hongslo a, Jørgen Stenersen b, Knut-Erik Tollefsen c a Norwegian Institute of Public Health, Division of Environmental Medicine, N-0403 Oslo, Norway Division of Molecular Cell Biology, Department of Biology, University of Oslo, N-0316 Oslo, Norway c Norwegian Institute for Water Research (NIVA), N-0411 Oslo, Norway

b

Received 15 February 2005; received in revised form 21 June 2005; accepted 5 July 2005

Abstract The development of various in vitro screening methods has led to identification of novel estrogenic chemicals of natural and anthropogenic origin. In this study, the (anti)estrogenic potential of several environmental chemicals were compared in an array of in vitro test systems comprising: (i) competitive binding to estrogen receptors derived from the human breast cancer cell line MCF-7 (hER) and rainbow trout (Oncorhynchus mykiss) (rtER), (ii) a proliferation assay with MCF-7 cells (E-SCREEN), and iii) induction of vitellogenin (rtVtg) in isolated rainbow trout hepatocytes. The results showed substantial differences in assay sensitivity for potent estrogens like 17hestradiol, diethylstilbestrol and zearalenone (ranking order of sensitivity: E-SCREEN > hER å rtER å rtVtg). Chemicals like 4-n-nonylphenol and bisphenol A had higher relative binding affinity to the hER, whereas 4-t-butylphenol and 4-n-butylphenol showed highest affinity to the rtER. Zearalenone and the novel estrogen 4-t-butylhexanol displayed a considerable higher relative potency in the E-SCREEN than the rtVtg assay, whereas alkylphenols and the novel estrogen mimic 4-t-butyl-nitrobenzene were most potent in fish cells. Correlation analysis of data from the test systems suggest that interspecies differences is largely due to inter-assay variation of the ER-dependent cellular responses, whereas binding to the ER are fairly similar in the two species tested. D 2005 Published by Elsevier Inc. Keywords: Alkylphenol; Estrogen receptor; MCF-7; Rainbow trout; Vitellogenin; Xenoestrogens; Estrogen mimics

1. Introduction Environmental estrogens, or estrogen mimics, have been suspected of modulating the endocrine system through multiple mechanisms of action and may potentially affect natural growth, development and reproduction in wildlife and humans (Colborn et al., 1993). The environmental estrogens, which may be of both natural and anthropogenic

Abbreviations: Relative binding affinity, (RBA); relative estrogenic effect, (REE); relative estrogenic potential, (REP); human estrogen receptor, (hER); rainbow trout estrogen receptors, (rtER). * Corresponding author. Norwegian School of Veterinary Science, Department of Basic Sciences and Aquatic Medicine, P. O. Box 8146 Dep., 0033 Oslo, Norway. Tel.: +47 22 96 46 17; fax: +47 22 59 73 10. E-mail address: [email protected] (C.M. Olsen). 1532-0456/$ - see front matter D 2005 Published by Elsevier Inc. doi:10.1016/j.cca.2005.07.002

origin, have been found in all major compartments of the environment and the potential impact on ecosystem health has elicited considerable concern within the scientific, regulatory and public community. Many of the effects of these environmental estrogens are mediated through activation of the estrogen receptor (ER), which functions as a ligand-dependent transcriptions factor. The ER, has a variable N-terminal region (A/B-domain), a highly conserved central DNA binding region (C-domain), a variable hinge region (D-domain), a moderately conserved ligandbinding region (E-domain) and a poorly conserved short sequence at the carboxy-end (F-domain) (Pakdel et al., 1989). In recent years, presence of ERs have been documented in various vertebrates including fish, amphibians, reptiles, birds, and mammals and multiple forms of ER has been demonstrated in some species like fish and humans

268

C.M. Olsen et al. / Comparative Biochemistry and Physiology, Part C 141 (2005) 267 – 274

(see Hawkins et al., 2000 for details). Although the ER from closely related species exhibits similar binding affinities for endogenous and exogenous estrogens (Tollefsen et al., 2002), large differences in estrogen binding have been demonstrated for the ER from divergent species (Matthews et al., 2000, 2001; Harris et al., 2002). For instance, the ERs derived from rainbow trout (rtER) has a highly divergent amino acid sequence within its ligand binding region (Edomain), with similarity of 60% when compared to the human ERa (hER) (Pakdel et al., 1989). This introduces the possibility of differential ligand-binding preferences and/or affinities for both endogenous and exogenous estrogens in species like fish and humans. Several in vitro systems have been developed and utilised for screening chemicals for their estrogenic properties (Andersen et al., 1999; Gutendorf and Westendorf, 2001). These screening methods, which are derived from native or modified cells systems from various species, have been shown to display an inter-assay variation to estrogen mimics (Andersen et al., 1999; Andersson et al., 1999; Coldham et al., 1997; Fang et al., 2000; Gutendorf and Westendorf, 2001). The aim of this study was to compare the ER affinity and the (anti)estrogenic potencies of a range of estrogen mimics in commonly used test systems derived from fish and humans. Frequently used test protocols were adopted in this study. These systems include (i) competitive binding to estrogen receptors derived from the human breast cancer cell line MCF-7 (hER) and the rainbow trout (Oncorhynchus mykiss) liver (rtER), (ii) a estrogen dependent proliferation assay with MCF-7 cells (E-SCREEN), and iii) induction of vitellogenin (rtVtg) in isolated rainbow trout hepatocytes.

2. Materials and methods 2.1. Chemicals The test chemicals: 17h-estradiol (E2), diethylstilbestrol (DES), bisphenol A (BPA) and zearalenone (ZEN) were all from Sigma (St. Louis, MI, USA), 4-n-butylphenol (C4n) was from TCI (Tokyo, Japan) and 4-t-octylphenol (C8t) was obtained from Aldrich (Milwaukee, WI, USA). 4-n-nonylphenol (C9n), 4-t-butylphenol (C4t), cis/trans 4-t-butylcyclohexanol (C4-Cyclo), 4-n-butyl-chlorobenzene (C4– Cl), 4-t-butylnitrobenzene (C4 – NO2) and 4-n-butylaniline (C4 – NH2) were all supplied by Lancaster, (Morecamble, UK). The antiestrogen ZM 189.154 was kindly supplied by Dr. T. Hutchinson (AstraZeneca, United Kingdom). All the test chemicals had a purity of minimum 97%. Prior to use all chemicals were diluted in analytical grade (purity > 99.8%) methanol, ethanol or DMSO supplied by Sigma. Diluted chemicals were stored in the dark (at 80 -C for the trout studies and 4 -C for the MCF-7 studies) until use. Except for E2, DES, ZEN and radiolabelled estradiol, the same batch of the chemicals was used in all assays. The radio-ligand

2,4,6,7-[3H]estradiol used for hER studies (74 Ci/mmol, NEN Life Science Products, Boston MA, USA) and rtER (85 Ci/mmol, Nycomed Amersham plc, Buckinghamshire, England) had a radio-chemical purity of more than 97%. 2.2. MCF-7 breast cancer cells Estrogen-dependent breast cancer cells, MCF-7, were kindly provided by Dr. A. M. Soto, Tufts University School of Medicine, Boston, MA, USA. The MCF-7 cells were grown for routine maintenance in Dulbecco’s modified Eagle’s medium, DME (GIBCO BRL Life Technologies, Scotland) supplemented with 5% heat-inactivated foetal bovine serum, (GIBCO BRL Life Technologies) at 37 -C with 5% CO2 in air under saturated humidity. 2.3. Fish Immature 200 – 500 g rainbow trout (Oncorhynchus ˚ a fish farm (Rælingen, mykiss) were obtained from the A Norway). The fish were acclimatised (mean T 95% C.I.) as follows: temperature: 12 T 0.1 -C, oxygen saturation: 95 T 7%, pH: 6.5 T 0.2, loading: 10 g fish/l, flow: 5 l/min) in circular fish tanks for minimum 2 weeks prior to the experiments. The tanks received artificial illumination (100 lux) for 12 h a day. Feeding occurred daily with commercial fish pellets (Felleskjøpet, Trondheim, Norway) in amounts corresponding to 1% of total body mass. 2.4. hER and rtER competitive studies Partially purified cell and tissue homogenates containing hER or rtER were obtained from MCF-7 cells or rainbow trout livers essentially as described by Olsen et al. (2002) and Tollefsen et al. (2002), respectively. The cell and tissue homogenates were incubated (hER: 2 h at 4 -C, rtER: 16 h at 4 -C) with 2,4,6,7-[3H]estradiol alone or in combination with unlabelled competitors, 500-fold excess of unlabelled estradiol (non-specific binding) and vehicle (hER: 15% DMSO, rtER: 2% MeOH). Unbound steroids were removed by incubation with dextran-coated charcoal (DCC) solution (hER: 0.5% charcoal / 0.05% dextran for 30 min, rtER: 1.25% charcoal / 0.125% dextran for 5 min) on ice, followed by centrifugation at 1000 g (4 -C). Aliquots of supernatant were removed and radioactivity determined by liquid scintillation spectrometry using standard tritium conditions. The difference in binding between tubes with and without excess of unlabelled E2 represents the specifically bound ligand. Competitive binding was calculated as the binding of [3H]estradiol in the presence of test compounds relative to that of the vehicle (maximum binding) alone. 2.5. MCF-7 cell proliferation assay (E-SCREEN) The MCF-7 cell proliferation assay was performed as previously described (Olsen et al., 2002). In essence, MCF-

C.M. Olsen et al. / Comparative Biochemistry and Physiology, Part C 141 (2005) 267 – 274

269

7 cells (4  104) were seeded in 6-well Falcon plates (Becton Dickingson Labware, Oxnard, CA, USA) in DME with 5% FBS. The following day the medium was changed to phenol red-free DME with 10% DCC treated FBS, and the cells were exposed to the test chemicals diluted in ethanol (final concentration < 0.2% v/v) for 6 days. At the end of the exposure, cell growth was measured (as Ag DNA) by fluorometric quantitation of 1 AM Hoechst 33258 dye (Calbiochem-Boeringer, La Jolla, CA, USA) in 0.02 M EDTA/NaOH at pH 12.5 as described by Brunborg et al. (1988). Proliferative effect (PE) was calculated as the ratio between the cell yield obtained with the test chemical and the cell yield in the hormone-free control. The relative estrogenic effect was calculated as 100  ((PE-1 of the test chemical) / (PE-1 of E2)).

medium was removed and frozen at 80 -C. At convenience, frozen cell medium was thawed and analysed for the estrogenic biomarker vitellogenin (rtVtg) using specific monoclonal salmon antibodies (BN-5, dilution 1 : 2000, Biosense Laboratories AS, Bergen, Norway) essentially as described by Tollefsen et al. (2003). rtVtg production was expressed relative to maximum rtVtg production obtained for E2 (100 nM). Cytotoxic effects on cells were determined at the end of the test by incubation with the viability fluorescent dyes Alamar blue (AB) and 5-carboxyfluorescein diacetate acetoxymethyl ester (CFDA-AM) essentially as described by Ganassin et al. (2000).

2.6. Fish in vitro rtVtg assay

Results in Table 1 are presented as means T the standard deviation (SD). Regression lines and 95% confidence intervals were obtained using Sigmaplot 2000. The curves for cell growth and rtVtg induction were drawn up to the maximal induction point. Most of the chemicals tested gave reduced growth when tested at higher concentrations (results not shown).

2.7. Mathematical and statistical analysis

Hepatocytes were isolated from juvenile rainbow trout, seeded as a mono-layer culture and exposed to test chemicals as described by Tollefsen et al. (2002). Briefly, hepatocytes were isolated by a two-step perfusion method, diluted in serum-free L-15 medium (Biowhittaker Inc., Walkersville, MD, USA) and seeded in 96 wells Primaria\ plates (Becton Dickingson Labware) kept in ambient atmosphere at 15 -C. The cells were then cultured for one day in growth medium prior to replacement of half the initial volume by growth medium containing the vehicle DMSO (final concentration < 0.2% v/v) or different concentrations of test chemicals diluted in DMSO. Half of the medium was changed after two days, and the cells re-exposed for an additional two-day period before the

3. Results 3.1. Binding to the hER and rtER All chemicals tested had detectable affinity to the crudely isolated ERs (Figs. 1 and 2). When comparing IC50 values, E2, DES, ZEN, BPA, C8t and C9n had higher affinity to the

Table 1 Relative Binding Affinity (RBA) for a range of compounds to the human and rainbow trout estrogen receptor (ER) and their relative estrogenic potential (REP) in human breast cancer cells (cell growth of MCF-7) and rainbow trout hepatocytes (vitellogenin production) Test compound

Symbol

Human ER a

IC50 (M) 17h-estradiol Diethylstilbestrol Zearalenone Bisphenol A 4-t-octylphenol 4-n-nonylphenol 4-t-butylphenol 4-n-butylphenol 4-t-butyl cyclohexanol 4-n-butyl chlorobenzene 4-t-butyl nitrobenzene 4-n-butyl aniline a

E2 DES ZEN BPA C8t C9n C4t C4n C4-Cyclo C4 – Cl C4 – NO2 C4 – NH2

Rainbow trout ER b

IC50 (M)

RBA 9

1.8  10 5  10 10 2  10 7 1.8  10 6 3.8  10 5 1.3  10 5 8.7  10 4 2.1 10 4 1.9  10 3 4.1 10 3 wbc 1.2  10 3

100 580 4.1 0.315 6.4  10 2.0  10 2.1 10 1.2  10 8.5  10 3  10 5 – 1.1 10

3 2 4 3 5

4

6.6  10 2.1 10 2.6  10 8.4  10 8.4  10 3.6  10 8.6  10 1.8  10 wbc wbc wbc 7.4  10

a

Cell growth MCF-7 cell e

b

EC50 (M)

RBA 9

9 7 4 4 4 5 5

4

100 390 2.5 5.8  10 7.6  10 9.4  10 4.6  10 2.3  10 – – – 8.0  10

3 3 4 3 2

4

6.1 10 1.8  10 4.1 10 3  10 7 5  10 6 wig 3.2  10 4  10 6 1 10 4 wig wig neh

12 11 10

5

Vitellogenin production EC50 (M)e

f

REP

100 33.8 14.9 2  10 3 1.2  10 – 1.9  10 1.5  10 6.1 10 – – –

4

5 4 6

1.0  10 1.6  10 5.2  10 3.5  10 3.1 10 8.0  10 1.8  10 3.5  10 neh neh 3.7  10 neh

10 9 7 6 6 5 5 6

5

Equilibrium inhibitory concentration (IC50) was calculated as the concentration causing 50% inhibition of [3H]17h-estradiol binding. Relative binding affinity (RBA) was calculated as 100  the ratio between the binding affinity (IC50) of 17h-estradiol and the test compound. c wb: weak binding (<50% inhibition of [3H]17h-estradiol binding). d nb: no binding under conditions employed. e The 50% effect concentration (EC50) was calculated as the concentration causing 50% effect (cell growth or vitellogenin production). f Relative estrogenic potency (REP) is the ratio between 17h-estradiol and xenoestrogen concentration needed to produce EC50  100. g wi: weak inducer (<50% induction of cell growth or vitellogenin production). h ne: no effect under conditions employed. b

REPf 100 6.1 1.9  10 2.9  10 3.2  10 2.0  10 5.6  10 2.9  10 – – 2.7  10 –

2 3 3 4 4 3

4

270

C.M. Olsen et al. / Comparative Biochemistry and Physiology, Part C 141 (2005) 267 – 274

[H3]estradiol binding (%)

hER

rtER E2

100

DES

80

ZEN

60 IC50 40

BPA

20

C8t

0

C9n -12 -11 -10

-9

-8

-7

-6

-5

-4

-3 -12 -11 -10

Log Molarity

-9

-8

-7

-6

-5

-4

-3

Log Molarity

Fig. 1. Displacement of [3H]estradiol from estrogen receptors isolated from MCF-7 cells (hER) and rainbow trout liver cells (rtER). Curves are representative examples, each chemicals was tested as duplicates in three individual experiments. 17h-estradiol (E2); diethylstilbestrol (DES); zearalenone (ZEN); bisphenol A (BPA); 4-n-nonylphenol (C9n); 4-tert-octylphenol (C8t). BPA are partly masked by C8t and C9n in both graphs.

hER than to the rtER (Table 1). However, the alkylphenols C4t and C4n had the highest affinity to the rtER. The relative binding affinity (RBA) values for DES, ZEN and C8t obtained for the two receptor populations were quite similar, whereas RBA values for BPA and C9n were 54 and 21 times higher for hER compared to rtER, respectively (Table 1). The RBA values of C4n and C4t were, on the other hand, ¨20 times higher for the rtER than those obtained for hER. The cyclic hydrocarbon C4-Cyclo was able to completely displace radio labelled E2 from hER, but displaced only 29% from the rtER at the highest tested concentration (Fig. 2). The non-phenolic compounds C4 –Cl, C4 – NO2 and C4 – NH2 had low affinity to both hER and rtER, and needed to be incubated overnight to reach equilibrium in the hER binding assay (Fig. 2). The RBAs for alkylphenols and non-phenolic compounds were variable and ranged from 5000 to 1,000,000 times lower than E2.

3.2. Induction of estrogen-dependent cell responses The E-SCREEN assay (MCF-7 cells) was far more sensitive than the induction of rtVtg in a primary culture of trout hepatocytes (Fig. 3). The EC50 values for E2, DES and ZEN were 16, 88 and 1268 times lower in the ESCREEN compared to the rtVtg-assay (Table 1 and Fig. 3). However, when comparing their relative estrogenic potential (REP), the differences became negligible for DES and BPA, while ZEN still was a more potent inducer of cell growth (784 times) than rtVtg. Interestingly, the EC50 values for the two assays were quite similar for C8t, C4n, C4t, while their REP were higher in the rtVtg-assay than in the E-SCREEN (26, 19 and 29 times, respectively). Some of the non-phenolic compounds were also able to induce weak estrogenic effects in both cell systems, but substantial differences were found for C4-Cyclo and C4 –NO2 between the two assays (Fig. 4). C4-Cyclo induced approximately

[H3]estradiol binding (%)

hER

rtER E2

100

C4t

80

C4n 60

C4-Cyclo

IC50

40

C4-Cl

20

C4-NO2 C4-NH2

0 -12 -11 -10 -9

-8 -7

-6

-5

Log Molarity

-4

-3

-2

-12 -11 -10 -9

-8 -7

-6

-5

-4

-3

-2

Log Molarity

Fig. 2. Displacement of [3H]estradiol from estrogen receptors isolated from MCF-7 cells (hER) and rainbow trout liver cells (rtER). Curves are representative examples, each chemicals was tested as duplicates in three individual experiments. 17h-estradiol (E2); 4-tert-butylphenol (C4t); 4-n-butylphenol (C4n); cis/ trans 4-tert-butylcyclohexanol (C4-Cyclo); 4-n-butylchlorobenzene (C4 – Cl); 4-tert-butylnitrobenzene (C4 – NO2); 4-n butylaniline (C4 – NH2).

C.M. Olsen et al. / Comparative Biochemistry and Physiology, Part C 141 (2005) 267 – 274

Relative estrogenic effect (%)

Cell growth MCF-7 cells

271

rtVtg-induction

120

E2

100

DES

80 60 40

ZEN EC50

BPA C8t

20

C9n

0 -13 -12 -11 -10 -9

-8

-7

-6

-5

-4

-13 -12 -11 -10 -9

Log Molarity

-8

-7

-6

-5

-4

Log Molarity

Fig. 3. Induction of cell growth in MCF-7 cells after 6 days of exposure and induction of vitellogenin (rtVtg) in isolated rainbow trout hepatocytes after 4 days of exposure. Each chemical was tested in duplicates and tested in minimum three individual experiments. Results are mean values. 17h-estradiol (E2); diethylstilbestrol (DES); zearalenone (ZEN); bisphenol A (BPA); 4-n-nonylphenol (C9n); 4-tert-octylphenol (C8t).

50% cell growth at a concentration 106 times higher than E2, but was unable to induce rtVtg production in trout hepatocytes. C4 –NO2 was able to induce more than 50% rtVtg production in the trout hepatocytes, but only marginally stimulate the cell growth (19%) in MCF-7 cells (Fig. 4). C4 –Cl gave weak stimulation of both cell growth and rtVtg induction (less than 20%). C4 –NH2 was the only compound that was negative in both test systems. The estrogen receptor antagonist ZM 189.154 (1 AM) abolished both the rtVtg production and the cell growth induced by the test chemicals (results not shown), thus indicating that the estrogenic response was mediated through interaction with the ER. The concentrations needed to induce half-maximal effect (EC50) for the alkylphenols and the non-phenolic compounds tested were from 106 to 107 times higher than the concentration of E2 (Table 1, Figs. 3 and 4). DES was a complete agonist both in the MCF-7-assay (cell growth) and

Relative estrogenic effect (%)

Cell growth MCF-7 cells

in the trout assay (rtVtg induction) shown by its ability to induce the same level of response as E2. ZEN and BPA were complete agonists in the E-SCREEN, but only partial agonists in the rtVtg-assay. The rest of the compounds tested were partial agonists in both assays, except for C4 – NH2 that had no effect. With the exception of C8t and C4n, which showed cytotoxicity in the AB and CFDA-AM assay at concentrations above 30 AM (results not shown), no toxic effects were seen in the rtVtg-assay for the chosen chemicals and concentrations. 3.3. Comparability between test-systems An interesting aspect was the comparability of estrogenic potencies derived from the different assays. If the chemicals showed the same potency in both of the systems compared, one would expect a slope close to unity. Indeed, when the receptor-binding assay was compared with the estrogen-

rtVtg-induction

100

E2

80

C4t

60

C4n EC EC50 50

C4-Cyclo

40

C4-Cl

20

C4-NO2

0

C4-NH2 -13 -12 -11 -10 -9

-8

-7

Log Molarity

-6

-5

-4

-13 -12 -11 -10 -9

-8

-7

-6

-5

-4

Log Molarity

Fig. 4. Induction of cell growth in MCF-7 cells after 6 days of exposure and induction of vitellogenin (rtVtg) in isolated rainbow trout hepatocytes after 4 days of exposure. Each chemical was tested in duplicates and tested in minimum three individual experiments. Results are mean values. 17h-estradiol (E2); 4-tert-butylphenol (C4t); 4-n-butylphenol (C4n); cis/trans 4-tert-butyl cyclohexanol (C4-Cyclo); 4-n-butylchlorobenzene (C4 – Cl); 4-tert-butylnitrobenzene (C4 – NO2); 4-n butylaniline (C4 – NH2).

C.M. Olsen et al. / Comparative Biochemistry and Physiology, Part C 141 (2005) 267 – 274

responsive-assay in each species, there was a good correlation between RBA and REP in both the MCF-7 and the trout derived assays. The slope was 0.98 and the R 2value 0.86 for the trout assays (Fig. 5C). ZEN, which had approximately 100-fold better affinity to the rtER compared to its REP in the rtVtg induction assay was the only compound outside the 95% confidence interval for the line. When the same comparison was performed with the MCF7-assays R 2-values increased to 0.94, but the slope was somewhat lower (0.8) (Fig. 5D). The comparability of RBA-values was quite good between the two species. When we compared the two species by plotting the RBA-values against each other we got a slope of 0.79 and R 2-value of 0.809 (Fig. 5A). The only compounds outside the 95% confidence interval for the line were BPA and C9n, which had better affinity to the hER compared rtER. However, only a weak correlation was obtained between REP-values for rtVtg-induction and REPvalues from the E-SCREEN, the slope was 0.586 and the R 2-value was 0.75 (Fig. 5B), reflecting the relatively higher sensitivity of the E-SCREEN compared to the rtVtginduction.

2

DES

A

y= -0.2 + 0.79x

E2

Log RBA rtER

r2= 0.809 1

ZEN

0 -1 C4n

-2

C8t

BPA

C4t -3 -4

C9n

4. Discussion Binding of ligand to ER is most often the initial event in a series of activating steps leading to an estrogen specific response. As seen in the present work the ER from humans and fish have the ability to accommodate the binding of several structurally different compounds including E2, DES, ZEN, BPA, alkylphenols and non-phenolic compounds. Some of the well-known estrogenic chemicals like E2, DES and ZEN were highly potent binders in both systems, whereas others like BPA and the alkylphenols were weak binders to the ER. Interestingly, several novel estrogen mimics with structural similarities to alkylphenols were identified. Some of these non-phenolic compounds, which have a functional nitro group (C4– NO2), aniline group (C4– NH2) or a chloride group (C4 –Cl), have weak binding affinity for both the hER and the rtER compared to E2. As seen in this work, even non-aromatic chemicals like C4Cyclo were able to interact with the ER from these two species, thus supporting that the phenol moiety of most estrogen mimics is not a pre-requisite for (anti)estrogenic activity as previously indicated (Ekena et al., 1997).

Log REP rtVtg production

272

C4-NH2

-4

2 1

y= -0.51 + 0.586x

B

E2

r2= 0.75 DES

0 -1 -2 -3

ZEN

C4n C8t

BPA

C4t

-4 -3

-2

-1

0

1

2

-5

-4

Log RBA hER

-3

-2

-1

0

1

2

Log REP cell growth MCF-7

3 y= 0.9 + 0.98x r2= 0.86

C

3

DES

y= 0.64 + 0.8x

E2

2

1

Log RBA hER

Log RBA rtER

2

ZEN

0 -1 C4n

C8t

-2 -3

C4t

BPA

DES E2

1

ZEN

0

BPA

-1 -2

C8t

-3 -4

-4

D

r2= 0.94

C4n C4t

-5 -3

-2

-1

0

Log REP rtVtg

1

2

-5

-4

-3

-2

-1

0

1

2

Log REP cell growth

Fig. 5. (A – D). Correlation and 95% confidence interval between A) relative binding affinity (RBA, E = 100) obtained from hER versus rtER, B) relative estrogen potencies (REP, E2 = 100) obtained in the E-SCREEN and the rtVtg-induction, C) REP obtained in rtVtg-induction and RBA to rtER and D) REP obtained in the E-SCREEN and RBA to hER. 17h-estradiol (E2); diethylstilbestrol (DES); zearalenone (ZEN); bisphenol A (BPA); 4-n-nonylphenol (C9n); 4tert-octylphenol (C8t); 4-n-butylphenol (C4n); 4-tert-butylphenol (C4t); 4-n butylaniline (C4 – NH2).

C.M. Olsen et al. / Comparative Biochemistry and Physiology, Part C 141 (2005) 267 – 274

With the exception of C8t and ZEN, which were equipotent binders in the two assay systems, all chemicals showed an inter-assay difference in relative binding affinity. As also reported by Tabira et al. (1999), binding affinity of alkylphenols increase with the number of alkyl groups on the side chain when assayed in a hER binding-assay, and non-branched analogues bound better than branched. This is in contrast to our findings for rtER, which had higher affinity to the moderate sized alkylphenols like C4t and C4n than long-chained alkylphenols like C8t and C9n. However, a recent study performed on rainbow trout seems to confirm that alkylphenols with moderate sized alkyl groups bound with equal or higher potency than long chained alkylphenols (Knudsen and Pottinger, 1999). Our receptor binding results are generally in agreement with other work on hER and rtER, although some differences are observed. In studies with bacterially expressed glutathione-S-transferase-ER fusion proteins with receptor constructs containing the D, E and F domain of the hER and rtER, DES was shown to have lower affinity to the hER than E2 (Matthews et al., 2000). In the same study BPA and ZEN had a 10-fold higher affinity to rtER compared to hER. This is in contrast to our work with native receptors from cell and tissue extracts where DES had a 10-fold higher affinity than E2 to both hER and rtER, and BPA had higher affinity to hER, while ZEN had approximately the same affinity to both receptors. The difference in affinity to ER observed for some chemicals indicate that there may exist an interspecies distinction in preference for various ligands, although methodological differences such as receptor populations (e.g.. various levels of a-ER, h-ER and g-ER) and solvents used has been proposed as a source for inter-assay variance elsewhere (Tollefsen, 2002; Beresford et al., 2000). Thorough studies to elucidate the effect of any confounding factors on ER binding were not undertaken in this work. However, the fact that most chemicals displayed a similar binding preference measured as relative binding to the ER, suggest that the differences observed are due to interspecies preferences in ligand binding. The inter-assay differences became more pronounced when measuring estrogen-dependent responses. The ESCREEN was far more sensitive than the rtVtg-assay when cells were exposed to the most potent estrogens like E2, DES and ZEN (ranking order for sensitivity E-SCREEN > hER  rtER = rtVtg). Interestingly, the large discrepancy in relative potency of ZEN could not be predicted from the RBA values alone, and similar differences have been observed in hER- and rtER-derived assays elsewhere (Le Guevel and Pakdel, 2001). Furthermore, it is known that there are large variations in sensitivity to the estrogenicity of ZEN between species, which has been ascribed differences in metabolic capacity and presence of various estrogenic metabolites of ZEN (Pompa et al., 1988; Ueno et al., 1983). The fish liver cells, which retain a substantial metabolic capacity in culture (Pesonen and Andersson,

273

1991; Flouriot et al., 1995), may have far better capacity to perform inactivation of ZEN through selective metabolism of the mother compound than the MCF7 cells. Other compounds such as C4 –NO2 did also show diverging responses in the two test systems, which was not reflected in its affinity to the different ERs. The reason for these discrepancies between the chosen in vitro assays is not apparent, but it has been suggested that variations in metabolic capacity may influence the estrogenicity of chemicals (Beresford et al., 2000; Andersen et al., 1999; Andersson et al., 1999; Coldham et al., 1997; Fang et al., 2000). Interestingly, the alkylphenols and non-phenolic compounds tested had approximately the same EC50 values in the E-SCREEN and rtVtg-assay, but because of the very high sensitivity of MCF-7 cells to E2 their calculated REPvalues became much lower than in the rtVtg-assay (Table 1). This large difference in relative potencies may partly be explained by the modulating effects of sex steroid binding proteins. Sex steroid-binding proteins, which are produced and excreted in the cell media from primary cultures of fish hepatocytes (Foucher et al., 1991), are able to specifically bind E2 and lead to reduced cellular uptake of natural sex steroids in comparison to non-steroid chemicals such as xenoestrogens (Nagel et al., 1997; Tollefsen, 2002). The bioavailable concentrations of E2 will thus be higher in MCF7-cells since these cells do not produce SBPs. A question relevant for the use of in vitro screening data is if the potencies derived from different assays are comparable. The best correlation was obtained between the RBA and REP from the same species (Fig. 5C and D). However, there was also a good correlation between the RBA values obtained in receptor binding studies from the two species (Fig. 5A), whereas potencies derived from the biological response to estrogens were poorly correlated. This probably was a result of the relatively higher sensitivity of the E-SCREEN assay for E2 and some of the compounds tested. However, the low correlation between the ESCREEN and the induction of rtVtg also reflect that both assays express indirect responses of ER-binding, and thus probably have several different response-modulating factors in contrast to the RBA which is a measure of a direct measure of ligand interaction. In conclusion, the binding affinities of the various compounds tested in the two test systems were quite similar. The hER had relatively higher affinity to the potent estrogens and longer alkylphenols, while rtER seems to have higher affinity to smaller alkylphenols compared to hER. These variations could possibly be caused by differences in structural requirements for the rtER and hER and possible simultaneous presence of multiple isoforms of the ER, which could be critical for binding affinity. However, for the assays utilising activation of the receptor, divergence in estrogenic potencies derived from the two test systems seem to be more governed by other factors related to the cellular system used. The present results show that the

274

C.M. Olsen et al. / Comparative Biochemistry and Physiology, Part C 141 (2005) 267 – 274

chosen test methods yield different results depending on the chemical in question, thus suggesting use of multiple assays in the screening of chemicals with potential estrogenic properties.

Acknowledgements This study was supported by grant from the Research Council of Norway and Norsk Hydro (Norway). We are grateful to Dr. A. M. Soto for kindly providing us with MCF-7 cells and Dr. T. Hutchinson for providing the antiestrogen ZM189.154.

References Andersen, H.R., Andersson, A.M., Arnold, S.F., Autrup, H., Barfoed, M., Beresford, N.A., Bjerregaard, P., Christiansen, L.B., Gissel, B., Hummel, R., Jørgensen, E.B., Korsgaard, B., Le Guevel, R., Leffers, H., McLachlan, J., Møller, A., Nielsen, J.B., Olea, N., Oles-Karasko, A., Pakdel, F., Pedersen, K.L., Perez, P., Skakkebæk, N.E., Sonnenschein, C., Soto, A.M., Sumpter, J.P., Thorpe, S.M., Grandjean, P., 1999. Comparison of short-term estrogenicity tests for identification of hormone-disrupting chemicals. Environ. Health Perspect. 107 (Suppl 1), 89 – 108. Andersson, P.L., Blom, A., Johannisson, A., Pesonen, M., Tysklind, M., Berg, A.H., Olsson, P.E., Norrgren, L., 1999. Assessment of PCBs and hydroxylated PCBs as potential xenoestrogens: in vitro studies based on MCF-7 cell proliferation and induction of vitellogenin in primary culture of rainbow trout hepatocytes. Arch. Environ. Contam. Toxicol. 37, 145 – 150. Beresford, N., Routledge, E.J., Harris, C.A., Sumpter, J.P., 2000. Issues arising when interpreting results from an in vitro assay for estrogenic activity. Toxicol. Appl. Pharmacol. 162, 22 – 33. Brunborg, G., Holme, J.A., Søderlund, E.J., Omichinski, J.G., Dybing, E., 1988. An automated alkaline elution system: DNA damage induced by 1,2-dibromo-3-chloropropane in vivo and in vitro. Anal. Biochem. 174, 522 – 536. Colborn, T., vom Saal, S.F., Soto, A.M., 1993. Developmental effects of endocrine-disrupting chemicals in wildlife and humans. Environ. Health Perspect. 101, 378 – 384. Coldham, N.G., Dave, M., Sivapathasundaram, S., McDonnell, D.P., Connor, C., Sauer, M.J., 1997. Evaluation of a recombinant yeast cell estrogen screening assay. Environ. Health Perspect. 105, 734 – 742. Ekena, K., Weis, K.E., Katzenellenbogen, J.A., Katzenellenbogen, B.S., 1997. Different residues of the human estrogen receptor are involved in the recognition of structurally diverse estrogens and antiestrogens. J. Biol. Chem. 272, 5069 – 5075. Fang, H., Tong, W., Perkins, R., Soto, A.M., Prechtl, N.V., Sheehan, D.M., 2000. Quantitative comparisons of in vitro assays for estrogenic activities. Environ. Health Perspect. 108, 723 – 729. Flouriot, G., Pakdel, F., Ducouret, B., Valotaire, Y., 1995. Influence of xenobiotics on rainbow trout liver estrogen receptor and vitellogenin gene expression. J. Mol. Endocrinol. 15, 143 – 151. Foucher, J.L., Niu, P.D., Mourot, B., Vaillant, C., Le Gac, F., 1991. In vivo and in vitro studies on sex steroid binding protein (SBP) regulation in rainbow trout (Oncorhynchus mykiss): influence of sex steroid hormones and of factors linked to growth and metabolism. J. Steroid Biochem. Mol. Biol. 39, 975 – 986.

Ganassin, R.C., Schirmer, K., Bols, N.C., 2000. Cell and tissue culture. In: Ostrander, G.K. (Ed.), The Laboratory Fish. Academic Press, San Diego, pp. 631 – 651. Vol. Chapter 38. Gutendorf, B., Westendorf, J., 2001. Comparison of an array of in vitro assays for the assessment of the estrogenic potential of natural and synthetic estrogens, phytoestrogens and xenoestrogens. Toxicology 166, 79 – 89. Harris, H.A., Bapat, A.R., Gonder, D.S., Frail, D.E., 2002. The ligand binding profiles of estrogen receptors alpha and beta are species dependent. Steroids 67, 379 – 384. Hawkins, M.B., Thornton, J.W., Crews, D., Skipper, J.K., Dotte, A., Thomas, P., 2000. Identification of a third distinct estrogen receptor and reclassification of estrogen receptors in teleosts. Proc. Natl. Acad. Sci. U. S. A. 97, 10751 – 10756. Knudsen, F.R., Pottinger, T.G., 1999. Interaction of endocrine disrupting chemicals, singly and in combination, with estrogen-, androgen-, and corticosteroid-binding sites in rainbow trout (Oncorhynchus mykiss). Aquat. Toxicol. 44, 159 – 170 (Amsterdam). Le Guevel, R., Pakdel, F., 2001. Assessment of oestrogenic potency of chemicals used as growth promoter by in-vitro methods. Hum. Reprod. 16, 1030 – 1036. Matthews, J., Celius, T., Halgren, R., Zacharewski, T., 2000. Differential estrogen receptor binding of estrogenic substances: a species comparison. J. Steroid Biochem. Mol. Biol. 74, 223 – 234. Matthews, J.B., Clemons, J.H., Zacharewski, T.R., 2001. Reciprocal mutagenesis between human alpha (L349, M528) and rainbow trout (M317, I496) estrogen receptor residues demonstrates their importance in ligand binding and gene expression at different temperatures. Mol. Cell Endocrinol. 183, 127 – 139. Nagel, S.G., vom Saal, F.S., Thayer, K.A., Dhar, M.G., Boechler, M., Welshons, W.V., 1997. Relative binding affinity-serum modified access (RBA-SMA) assay predicts the relative in vivo bioactivity of the xenoestrogens bisphenol A and octylphenol. Environ. Health Perspect. 105, 70 – 76. Olsen, C.M., Meussen-Elholm, E.T., Holme, J.A., Hongslo, J.K., 2002. Brominated phenols: characterization of estrogen-like activity in the human breast cancer cell-line MCF-7. Toxicol. Lett. 129, 55 – 63. Pakdel, F., Le Guellec, C., Vaillant, C., Le Roux, M.G., Valotaire, Y., 1989. Identification and estrogen induction of two estrogen receptors (ER) messenger ribonucleic acids in the rainbow trout liver: sequence homology with other ERs. Mol. Endocrinol. 3, 44 – 51. Pesonen, M., Andersson, T., 1991. Characterization and induction of xenobiotic metabolizing enzyme activities in a primary culture of rainbow trout hepatocytes. Xenobiotica 21, 461 – 471. Pompa, G., Montesissa, C., Di Lauro, F.M., Fadini, L., Capua, C., 1988. Zearanol metabolism by subcellular fractions from lamb liver. J. Vet. Pharmacol. Ther. 11, 197 – 203. Tabira, Y., Nakai, M., Asai, D., Yakabe, Y., Tahara, Y., Shinmyozu, T., Noguchi, M., Takatsuki, M., Shimohigashi, Y., 1999. Structural requirements of para-alkylphenols to bind to estrogen receptor. Eur. J. Biochem. 262, 240 – 245. Tollefsen, K.-E., 2002. Interaction of estrogen mimics, singly and in combination, with plasma sex steroid-binding proteins in rainbow trout (Oncorhynchus mykiss). Aquat. Toxicol. 56, 215 – 222. Tollefsen, K.-E., Mathisen, R., Stenersen, J., 2002. Estrogen mimics bind with similar affinity and specificity to the hepatic estrogen receptor in Atlantic salmon (Salmo salar) and rainbow trout (Oncorhynchus mykiss). Gen. Comp. Endocrinol. 126, 14 – 22. Tollefsen, K.-E., Mathisen, R., Stenersen, J., 2003. Induction of vitellogenin synthesis in an Atlantic salmon (Salmo salar) hepatocyte culture: a sensitive in vitro bioassay for the oestrogenic and anti-oestrogenic activity of chemicals. Biomarkers 8, 394 – 407. Ueno, Y., Tashiro, F., Kobayashi, T., 1983. Species differences in zearalenone-reductase activity. Food Chem. Toxicol. 21, 167 – 173.