4-iodotamoxifen aziridine, a new affinity labeling agent for the rapid detection of estrogen receptor isoforms1

PII:

J. Steroid Biochem. Molec. Biol. Vol. 67, No. 2, pp. 95±104, 1998 # 1998 Elsevier Science Ltd. All rights reserved Printed in Great Britain S0960-0760(98)00102-2 0960-0760/98/$19.00+0.00

4-iodotamoxifen Aziridine, a New Af®nity Labeling Agent for the Rapid Detection of Estrogen Receptor Isoforms* Younes Maarou®, Jacques Quivy{, Sunil Trivedi{, Nathalie Gilot and Guy Leclercq Laboratoire J.-C. Heuson de CanceÂrologie Mammaire, Service de MeÂdecine Interne, Institut Jules Bordet, Brussels, Belgium

We describe the simple and fast preparation of a new radioiodinated probe for the detection of the estrogen receptor (ER) and its isoforms. Iodotamoxifen aziridine was labeled with iodine 125 ([125 I]TAZ) in position 4 of the a aromatic ring. The yield was high (>75%), the label was stable and the speci®c activity was near optimal (1900±2170 Ci/mmol). The apparent relative binding af®nity of the probe to a recombinant human ER (hER) was high (RBA = 35 vs estradiol = 100). Electrophoretic studies (SDS±PAGE) with this hER indicated the high potency of [125 I]TAZ at very low concentration (<1 nM) to reveal ER bands after a short exposure time (1±4 days). Competition between this probe and various compounds as well as chemical treatments of the ER with SH-reactive chemicals, demonstrated the labeling speci®city. Analysis of cytosols from a panel of cell lines and various rat reproductive organs displayed characteristic ER bands (67, 50 and 37 kDa) suppressed by unlabeled E2. Detection in nonreproductive organs of 43 kDa E2-nondisplaceable peptide raised the question upon the presence of altered and/or variant ERs in many tissues. Data concerning human breast cancer cytosols were in complete accordance with those established with [3 H]TAZ: high ER polymorphism in most ER-positive samples and peculiar forms (mainly 43 kDa) in ER-negative samples. Hence, [125 I]TAZ appears especially useful for the detection of altered ER or related peptides in breast cancers. # 1998 Elsevier Science Ltd. All rights reserved. J. Steroid Biochem. Molec. Biol., Vol. 67, No. 2, pp. 95±104, 1998

INTRODUCTION

the cytosols of ER variants or proteolytic fragments that might have some clinical relevance [2, 3]. To characterize molecular subspecies, af®nity labeling agents [4, 5] are more appropriate than E2 because they can separate potential ER isoforms under denaturing conditions [6]. Among commercially available agents, much attention has focussed on tritiated tamoxifen aziridine ([3 H]TAZ) [6±8]. Its rather low speci®c activity (20±60 Ci/mmol), however, hampers its use in clinical practice, several weeks of autoradiography usually being required to detect ER and isoforms separated by conventional SDS±polyacrylamide gel electrophoresis (SDS±PAGE). It was for this reason that radioiodinated desethyl tamoxifen aziridine was ®rst prepared [9], but its synthesis was proved unsatisfactory. Labeling and puri®cation of this probe required several steps; labeling ef®ciencies were low (20±45%), as reported for a radioiodinated

The presence of estrogen receptor (ER) in cytosols from breast carcinomas is an important predictor of patients' response to hormone therapy. Currently, ER is determined by measuring the binding of isotopically labeled estradiol (E2) or by the use of ER-speci®c antibodies in immunoenzymatic assays [1]. However, these approaches do not provide any information on the molecular structure of ER, nor on the presence in *Preliminary reports were presented in 1995 at the 18th Annual San Antonio Breast Cancer Symposium (abst. 372) and in 1997 at the 13th Symposium of the Journal of Steroid Biochemistry and Molecular Biology (abst. 138P). {Correspondence to J. Quivy. Tel. +32-2-535-3491; Fax: +32-2534-7328; E-mail: [email protected]. {Present address. Division of Molecular Endocrinology, The Gujarat Cancer and Research Institute, Ahmadabad, 380 016 India. Received 12 Mar. 1998; accepted 29 May 1998. 95

96

Younes Maarou® et al.

tamoxifen (40%) prepared according to a similar method [10]. Moreover, speci®c activities (200± 600 Ci/mmol) of the radioiodinated desethyl tamoxifen aziridine were far from the theoretical maximum (2200 Ci/mmol) and its synthesis produces a mixture of cis and trans isomers [10]. In this paper, we describe the simple, rapid preparation of iodotamoxifen aziridine labeled with high ef®ciency and near maximal speci®c activity with iodine 125 ([125 I]TAZ) in a position Ð position 4 of the a aromatic ring Ð where the iodine atom is stably linked and which avoids isomerization. The ef®cacity of this probe in labeling ER is examined.

MATERIALS AND METHODS

Ligands [3 H]E2 (0100 Ci/mmol) and ORG 2058 (16aethyl-21-hydroxy-19-nor-4-pregnene-3,20-dione) were obtained from Amersham (Bucks, U.K.); ZCMIV (Z-17a-(2-iodovinyl)-11b-chloromethyl estradiol-17b) was purchased from Biocode (LieÁge); E2, bromophenylethylpyrrolidine (BPEP) and unsubstituted triphenylethylene squeleton from Sigma (U.S.A.). Tamoxifen and 4-hydroxytamoxifen were obtained from Zeneca (Maccles®eld, U.K.). RU 39411 (11b-[4-(dimethylaminoethoxy)phenyl]estra1,3,5(10)-triene-3,17b-diol) and RU58668 (11b-[4[5-(4,4,5,5,5-penta¯uoropentyl sulfonyl)pentyloxy] phenyl]estra-1,3,5(10)-triene-3,17b-diol) were from Roussel Uclaf (Romainville, France). Unlabeled 4iodotamoxifen aziridine and tributyltin precursor [11], trans and cis 4-iodotamoxifen and idoxifene were kindly provided by Professor M. Jarman (Institute of Cancer Research, Sutton U.K.). Reagents and antibodies Reagents for SDS±polyacrylamide gel electrophoresis were from Bio±Rad (Richmond, CA); N,Ndimethylformamide, S-methyl methane-thiosulfonate (MMTS), N-ethylmaleimide (NEM) and dimethylsulfoxide (DMSO) from Sigma. H222 and H226 anti-ER monoclonal antibodies were gifts from Dr D. Cotter (Abbott Laboratories, North Chicago). Labeling of trans 4-iodotamoxifen aziridine ((E)-1-[4-[2(aziridyl)-ethoxy]phenyl]-1-(4-iodophenyl)-2-phenyl-1butene) by iodine 125 A mixture of 10 ml of sodium [125 I] iodide solution (1 mCi, IMS30 Amersham) and of 4 ml of 0.1 N HCl was placed in a screw cap vial and vortexed. After addition of 12 ml of tributyltin precursor [11] (1 mg/ml in EtOH) and of 10 ml of Chloramine-T (1 mg/ml in CH3OH), it was vortexed again. The reaction was stopped after 10 min by addition of 10 ml of Na2S (10 mg/ml in H2O). The mixture was injected in an HPLC system (mBondapack C18 3.9  300 mm;

EtOH±H2O, 85:15, v/v; 0.8 ml/min) and puri®ed labeled 4-iodotamoxifen aziridine was eluted at 11 min and collected. The speci®c activity of the probe was calculated by estimation of the mass of eluted material using the HPLC UV trace after comparison with a standard curve. Estrogen receptor (ER) preparations A sample of full length human estrogen receptor (hER) produced in yeast (01 pmol/ml in phosphate buffer; pH 7.5) [12], kindly provided by Dr P. SjoÈholm (Karo Bio, Huddinge, Sweden), was stored at ÿ708C after separation into 50 ml aliquots. On use, the aliquots were diluted with 10 mM Tris±HCl buffer pH 8.0 containing bovine serum albumin (BSA; fraction V, Sigma) at a ®nal protein concentration of 01 mg/ml. Human primary breast cancers and corresponding malignant lymph nodes were minced and homogenized in a phosphate buffer (10 mM potassium phosphate, 1 mM monothioglycerol, 1.5 mM EDTA; pH 7.4) by means of a whole glass homogenizer. Cytosols were obtained by ultracentrifugation (100,000 g, 1 h at 48C). Aliquots were taken for assessing ER content by conventional immunoenzymatic assay (ER±EIA, Abbott Laboratories). Reproductive organs (uterus, ovary) and nonreproductive organs (kidney, liver, muscle, encephalon, intestine, diaphragm) from 25 day-old immature rats (Iffa Credo, France) were processed in a similar fashion. Cytosols from various cell-lines (MCF-7, CHO, CHO±ER, BT-20 and NIH 3T3), maintained in monolayer culture in our laboratory, were prepared as previously described [13]. All cytosols were stored in liquid nitrogen for use less than one week after preparation. Competitive binding assay Samples of 100 ml of hER preparation were mixed with the same volume of buffer (10 mM Tris±HCl; pH 8.0) containing 10 nM [3 H]E2 and increasing amounts of a given unlabeled competitor (six concentrations in the range 10ÿ11±10ÿ6 M, ®nal concentration). After overnight incubation at 08C, [3 H]E2 binding was measured by the dextran-coated charcoal (DCC) adsorption method (0.5% charcoal, 0.05% dextran). Each point was the average of duplicate determinations. A standard with unlabeled E2 was included in each assay. The concentrations of unlabeled E2 and of competitors that reduced [3 H]E2 binding by 50% ([I50]) were determined from the competition curves. The relative binding af®nity of each competitor was given by ([I50]competitor/ [I50]E2)  100. SDS±polyacrylamide gel electrophoresis [6] ER preparations containing 7% dimethylformamide were incubated for 1 h at 0±48C with a 1 nM, or as indicated, concentration of [125 I]TAZ. After removal

4-iodotamoxifen aziridine

of unbound ligand by DCC, the labeled preparation was incubated with H222 anti-ER monoclonal antibody (1 ml per ml of a 1 mg/ml preparation) for 2 h at 0±48C, or with the more appropriate H226 anti-ER MAb to characterize ER isoforms in tumor samples [8]. Immune complexes were adsorbed on anti-rat IgG agarose (100 ml suspension/ml ER preparation, Sigma), centrifuged, washed in high salt buffer [20 mM Tris±HCl, 500 mM KCl; pH 7.5], and resolubilized in a lysis buffer [4% SDS, 20% glycerol, 10% mercaptoethanol, 0.05% bromophenol blue in 500 mM Tris±HCl; pH 6.8] and stored at ÿ208C. Lysates were submitted to electrophoresis on a 10% polyacrylamide gel in buffer [25 mM Tris±HCl, 0.1% SDS, 192 mM glycine; pH 8.2]. Gels were subsequently stained, dried and exposed to Kodak X± OMAT ®lms for 1±3 days, to reveal labeled ER bands. The molecular weights of the labeled peptides were estimated according to the migration of 14± 94 kDa protein standards (Pharmacia). In one set of experiments, the gel was cut into 01.6 mm thick slices and the radioactivity of each slice was measured in a g-detector (Crystal 5400 Series, United Technologies Packard, IL) before and after extraction (30 min) with 0.5 ml DMSO. RESULTS

Synthesis of 4-[125 I]iodotamoxifen aziridine Because radioiodination was carried out in acidic conditions by addition of diluted HCl, a high yield (>75%) of [125 I]TAZ was obtained after only a short 10 min reaction time. The radiochemical ef®ciency based on the radiochromatographic pro®le was 95%. No contaminant cis isomer peak was detected by HPLC. The speci®c activity of the probe was within

97

the range 1900±2170 Ci/mmol and very close to the theoretical maximum of 2200 Ci/mmol. A sample of unlabeled 4-iodotamoxifen aziridine characterized by classical spectroscopic methods eluted with the same retention time as an aliquot of puri®ed [125 I]TAZ. The stability of the ligand after 40 days tested by HPLC was excellent: radioactive pro®le shows only one major peak (297%) of [125 I]TAZ; dehalogenation was very low: less than 3% of free iodine (data not shown). Binding of 4-iodotamoxifen aziridine to ER The apparent relative binding af®nity (RBA) of 4iodotamoxifen aziridine (the radioinert analogue of [125 I]TAZ) for ER under the conditions used for subsequent SDS±PAGE was measured in a competitive binding assay (1 h at 08C) using recombinant human ER (hER). If the RBA of E2 is taken as 100, the RBA of 4-iodotamoxifen aziridine was 35 compared to 199 for 4-hydroxytamoxifen and 4 for tamoxifen. The high RBA value of 4-iodotamoxifen aziridine seems to indicate that iodine linked in position 4 is well tolerated by the receptor. Sensitivity and threshold of ER detection by [125 I]tamoxifen aziridine We examined whether we could detect the receptor complex produced by incubating a low concentration (0.1 nM) of [125 I]TAZ with hER for 1 h. After immunoadsorption with an anti-ER MAb (H222), SDS± PAGE and 2 days autoradiography, we detected a single band of 67 kDa molecular weight (Fig. 1). No band was observed after incubating with a 200-fold excess of unlabeled E2, thus demonstrating the speci®city of the labeling. A lower concentration (0.01 nM) of [125 I]TAZ was just as ef®cient in visua-

Fig. 1. Sensitivity of hER labeling by [125 I]TAZ. hER was incubated with [125 I]TAZ for 1 h at 08C with or without a 200-fold excess of radioinert E2, immunoprecipitated with H222 anti-ER MAb and then analyzed by SDS±PAGE and autoradiography.

98

Younes Maarou® et al.

Fig. 2. Threshold of ER detection by [125 I]TAZ. Various dilutions of an hER preparation (1:1; 750 fmol/mg prot) were labeled with 1 nM [125 I]TAZ before immunoprecipitation with H222 anti-ER MAb and analysis by SDS±PAGE and autoradiography.

lizing receptor as was 1 pM of probe if the gels were exposed for slightly longer (4 days). To determine the minimal ER concentration providing an accurate SDS±PAGE analysis, an hER preparation (750 fmol/mg prot.) was diluted to high (325 and 162 fmol/mg prot.), medium (45 fmol/mg prot.), and low (7.5 and 3.5 fmol/mg prot.) ER levels. Even when the preparation contained less than 10 fmol/mg protein (e.g. dilutions of 1:100 and 1:200), the binding of [125 I]TAZ to the receptor was still detectable after only 1 day of exposure of the gel (Fig. 2). Nonspeci®c binding of [125 I]TAZ, probably to the serum albumin present in the dilution buffer, does occur. This activity (66 kDa), which remained in the

gel after treatment with DMSO or ethanol, could be easily identi®ed because it was not reduced by preincubation with unlabeled E2. Formation of a covalent bond between [125 I]tamoxifen aziridine and ER To ®nd out whether [125 I]TAZ was covalently bound to ER, the radioactivity of a sliced gel was measured before and after extraction with DMSO (Fig. 3). The speci®c 67 kDa hER peak (suppressed by preincubation with an excess of unlabeled E2) could not be extracted by organic solvent, thus con®rming the covalent binding of [125 I]TAZ. To discover whether cysteine residues are involved in the binding of [125 I]TAZ to hER, we used treat-

Fig. 3. Covalent binding of [125 I]TAZ to hER. hER was labeled with 1 nM [125 I]TAZ for 1 h at 08C, in the presence or absence of a 200-fold excess of radioinert E2, immunoprecipitated with H222 anti-ER MAb and then analyzed by SDS±PAGE and autoradiography. The gel was sliced and the radioactivity of each slice was counted before (a) and after (b) treatment with DMSO.

4-iodotamoxifen aziridine

99

Fig. 4. Involvement of cysteine(s) in [125 I]TAZ binding to hER. hER was treated for 30 min at 08C with the sulfhydryl-blocking reagents MMTS and NEM, labeled with 1 nM [125 I]TAZ for 1 h at 08C, immunoprecipitated with H222 anti-ER MAb, and then analyzed by SDS±PAGE and autoradiography.

ment with two sulfhydryl-blocking reagents, MMTS and NEM, for 30 min at 08C. Both reagents strongly suppressed hER electrophoretic band intensity (Fig. 4). Speci®city of [125 I]tamoxifen aziridine in the covalent labeling of hER The speci®city of [125 I]TAZ binding was investigated by labeling hER in the presence of a 200-fold excess of a variety of potential radioinert competitors. hER labeling was blocked by strong estrogens and antiestrogens only (Fig. 5). All the following ligands decreased 67 kDa hER band intensity: E2, 4-iodoTAZ, trans-4-iodo-tamoxifen, trans-4-hydroxytamoxifen, the antioestrogens iodoxifene, RU 39411, and RU 58 668, and Z-CMIV (an estrogen which interacts `quasi-irreversibly' with ER [14]). The cis isomer of 4-iodo-tamoxifen, which has very low af®nity for ER, only displaced [125 I]TAZ slightly whereas the unsubstituted triphenylethylene skeleton, which has an uterotrophic activity 010,000-fold lower than that of E2, was ineffective. All test-compounds devoid of binding af®nity for ER, i.e. progesterone, ORG 2058, hydrocortisone, dexamethasone, and BPEP (an analogue of the side-chain of triphenylethylenic antiestrogens), failed to compete. Ef®ciency of the covalent labeling of cytosolic ER from cell-lines, rat tissues and breast cancers with [125 I]tamoxifen aziridine Cell-lines expressing ER, such as the MCF-7 human breast cancer cell-line and CHO cells transfected with ER cDNA (CHO±ER) [15], exhibited an E2-suppressible 67 kDa ER electrophoretic band, sometimes associated with small amounts of a 50 kDa cleavage product also suppressed by E2 (Fig. 6). As expected, nontransfected control CHO cells and NIH 3T3 cell-lines, which are classi®ed as ER-negative,

failed to show any detectable ER band. Cytosol from the human breast cancer cell-line BT-20 displayed a 67 kDa band that was totally suppressed by an excess of unlabeled E2 as well as a 55 kDa band that was not. This 55 kDa band was present even when the H222 anti-ER MAb was omitted, indicating its nonspeci®c character. The ef®ciency of the covalent labeling of rat organ cytosols by [125 I]TAZ was determined by evaluating the labeling intensity of the ER bands according to a 6-point visual scale: ÿ no band; 2; +; + + ; + + + ; + + + + intense band. Cytosols from rat uterus and ovary showed multiple bands corresponding to native ER (67 kDa) and cleaved forms (50 and 37 kDa) which were all more or less suppressed by an excess of unlabeled E2 (Table 1). Cytosols from nonreproductive organs that contained low ER concentrations according to immunoenzymatic assays (e.g. kidney, liver, . . .) generally also presented a 67 kDa ER band, as well as small amounts of 55, 43 and 35 kDa peptides (Table 2). The patterns highly varied. While some tissues presented all these bands, others displayed only some or none at all. Recognition by H222 MAb also depended upon tissue (Table 2). A small series (n = 8) of primary breast cancers and corresponding invaded lymph nodes differing with respect to ER level, histology and menopausal status of the patients was also tested (Table 3). In the ERpositive primary tumor samples, we identi®ed the characteristic E2-suppressible 67 kDa [125 I]TAZlabeled ER band. Its intensity correlated with receptor content as measured by immunoenzymatic assay. Signi®cant amounts of E2-non supressible 55 and 43 kDa bands were also present. Only one of the four ER-positive lymph nodes had a detectable 67 kDa band, two had a 55 kDa band, and all four presented a rather intense 43 kDa band. In ER-negative tumors

100

Younes Maarou® et al.

Fig. 5. Speci®city of the labeling of hER by [125 I]TAZ. hER was labeled with 1 nM [125 I]TAZ for 1 h at 08C in the absence and presence of a 200-fold excess of various test-compounds. Preparations were immunoprecipitated with H222 anti-ER MAb and then analyzed by SDS±PAGE and autoradiography.

Fig. 6. Labeling of cell-lines with [125 I]TAZ. Cytosols from several cell-lines were labeled with 1 nM [125 I]TAZ for 1 h at 08C in the presence or absence of a 200-fold excess of radioinert E2, immunoprecipitated with H222 anti-ER MAb and then analyzed by SDS±PAGE and autoradiography.

4-iodotamoxifen aziridine

101

Table 1. Tissue speci®city of peptides labeled by [125 I]TAZ in cytosols from rat reproductive organs Molecular forms (kDa) Tissue

ER±EIA (fmol/mg prot)

Uterus

252

Ovary

66

67

50

37

P P + E2 S S + E2 P* P* + E2 S* S* + E2

++++ ÿ ÿ ÿ ++ + ++ +

++ + ÿ ÿ ++ 2 ÿ ÿ

+++ ÿ ÿ ÿ + ÿ ÿ ÿ

P P + E2 S S + E2 P* P* + E2 S* S* + E2

+++ ÿ 2 ÿ + + + +

++ ÿ ÿ ÿ + ÿ ÿ ÿ

+++ ÿ ÿ ÿ + 2 ÿ ÿ

P = H222 immunoprecipitate; S = H222 supernatant. P* = H226 immunoprecipitate; S* = H226 supernatant. E2: 200-fold excess of unlabeled E2. ÿ: not detectable band; + + + + : intense band.

(<10 fmol/mg protein according to ER immunoenzymatic assay), no 67 kDa ER band was found but the E2-nonsuppressible 55 and 43 kDa bands were still

present as reported for [3 H]TAZ [8]. The electrophoretic patterns of the primary tumors and corresponding lymph nodes were virtually superimposable.

Table 2. Tissue speci®city of peptides labeled by [125 I]TAZ in cytosols from rat nonreproductive organs Tissue

ER±EIA (fmol/mg prot)

Molecular forms (kDa) 67

55

43

35

Kidney

58

P P + E2 S S + E2

++ + ÿ ÿ

2 2 ÿ ÿ

+ + 2 2

+ + + +

Liver

34

P P + E2 S S + E2

+ ÿ 2 2

++ ++ ÿ ÿ

2 2 2 2

2 2 2 2

Muscle

4

P P + E2 S S + E2

+ ÿ 2 2

+ 2 ÿ ÿ

+ + ++ ++

++ ++ + +

Encephalon

0

P P + E2 S S + E2

ÿ ÿ ÿ ÿ

+ + + +

ÿ ÿ ÿ ÿ

+ + + +

Intestine

0

P P + E2 S S + E2

2 2 ÿ ÿ

2 2 ÿ ÿ

++ ++ ++ ++

+ + + +

Diaphragm

0

P P + E2 S S + E2

ÿ ÿ ÿ ÿ

ÿ ÿ ÿ ÿ

ÿ ÿ ÿ ÿ

ÿ ÿ ÿ ÿ

P = H222 immunoprecipitate; S = H222 supernatant. E2: 200-fold excess of unlabeled E2. ÿ: not detectable band; + + + + : intense band.

ductal adenocarcinoma poorly differentiated

invasive lobular adenocarcinoma

ductal adenocarcinorna poorly differentiated

ductal adenocarcinoma poorly differentiated

atypical medullary carcinoma

invasive lobular carcinoma

intraÿ epithelial carcinoma moderately differentiated

invasive lobular carcinoma

(1) Post

(2) Peri

(3) Pre

(4) Post

(5) Pre

(6) Peri

(7) Post

(8) Peri

T2

T2

T3

T2

n.a.

T3

T2

T4

3

2

5

2

1

12

1

15

Tumor size Number of invaded LN

P is immunoprecipitate, S supernatant, LN Lymph node, n.a. not available. E2: 200-fold excess of unlabeled E2 ÿ: not detectable band; + + + + : intense band.

Histology

Menopausa status

0

0

0

0

5

13

40

93

ER±EIA (fmol/mg)

ÿ ÿ + +

2

S + E2 P P + E2 S S + E2

ÿ ÿ +

ÿ ÿ ++ ++

ÿ ÿ + +

+ 2 + +

+ ÿ ÿ ÿ

++ 2 ÿ ÿ

+++ 2 ÿ ÿ

P P + E2 S

P P + E2 S S + E2

P P + E2 S S + E2

P P + E2 S S + E2

P P + E2 S S + E2

P P + E2 S S + E2

P P + E2 S S + E2

67

+ + + +

+

ÿ ÿ +

+ + + +

ÿ ÿ + +

++ ++ ++ ++

+ + + +

ÿ ÿ + +

++ ++ ++ ++

55

++ ++ ÿ ÿ

ÿ

+ + ÿ

+ + 2 2

++ ++ 2 2

+++ +++ 2 2

+ + ÿ ÿ

+ + 2 2

ÿ ÿ 2 2

43

Molecular forms (kDa)

Primary tumors

30

0

n.a.

0

0

10

20

84

ER±EIA (fmol/mg)

P P + E2 S S + E2

S + E2

P P + E2 S

P P + E2 S S + E2

P P + E2 S S + E2

P P + E2 S S + E2

P P + E2 S S + E2

P P + E2 S S + E2

P P + E2 S S + E2

+ + + +

++

+ + ++

ÿ ÿ ++ ++

ÿ ÿ ++ ++

ÿ ÿ ++ ++

ÿ ÿ + +

ÿ ÿ + 2

ÿ ÿ + 2

67

2 2 ÿ ÿ

++

++ ++ ++

+ + + +

ÿ ÿ ÿ ÿ

+ + + +

ÿ ÿ + +

ÿ ÿ 2 2

+ + + +

55

++ ++ ÿ ÿ

+

++ ++ ÿ

++ ++ 2 2

++ ++ ÿ ÿ

+++ +++ 2 2

++ ++ ÿ ÿ

++ ++ 2 2

++ ++ ÿ +

43

Molecular forms (kDa)

Invaded lymph nodes

Table 3. ER isoform patterns in breast tumors and axillary lymph nodes as identi®ed by [125 I]TAZ and immunoprecipitation with H226 anti-ER MAb 102 Younes Maarou® et al.

4-iodotamoxifen aziridine DISCUSSION

The above results have established that [125 I]TAZ binds with high af®nity to hER. Binding was observed at probe concentrations as low as one pmol and was detected after exposing gels for just 2 to 4 days. On the other hand, labeling of ER with [3 H]TAZ requires very much higher concentrations (20 nM) and considerably longer exposure times (03 weeks) [7]. [125 I]TAZ binds covalently to ER, probably to the Cys 530 residue identi®ed by Katzenellenbogen et al. for the binding of tritiumlabeled tamoxifen and ketononestrol aziridines [16]. After SDS±PAGE of [125 I]TAZ-labeled hER, we detected a 67 kDa band that was not extracted by DMSO and treatment of hER with two sulfhydrylblocking reagents before electrophoresis strongly depressed hER band intensity. We investigated [125 I]TAZ binding under various circumstances, e.g. in cell-lines, rat organ cytosols and human breast tumors. In receptor-positive celllines and in rat reproductive organs, [125 I]TAZ detected ER isoforms, 67 kDa native ER associated with 50 and 37 kDa cleavage products, that were identical to those already reported for [3 H]TAZlabeled cytosolic receptor in MCF-7 cells and mouse uterus [17, 18]. However, the presence of substantial amounts of nonspeci®c [125 I]TAZ-labeling (E2-nondisplaceable) could not be ruled out because of the serum albumin content of our hER preparation. In breast cancers, [125 I]TAZ detected such conventional ER isoforms with similar labeling speci®city but much higher ef®ciency than [3 H]TAZ. It was also similarly effective in detecting a 43 kDa peptide that apparently does not bind E2. Data from our laboratory revealed that this peptide reacts with ®ve anti-ER antibodies (unpublished data based on four independent experiments). Among such antibodies, the H226 MAb, described here, interacts primarily with activated ER in which DNA-binding domain is accessible [19]. We have suggested that this peptide might be related to biologically aggressive hormoneindependent breast tumors [8] and our present data con®rm its presence in ER-negative (by conventional assays) human breast primary tumors and metastatic lymph nodes. [125 I]TAZ would thus be an ideal tool for the fast routine identi®cation of tumors with this altered (or variant) receptor. [125 I]TAZ af®nity labeling and subsequent immunoprecipitation using H222 MAb, an antibody that recognizes the C-terminal end of the steroid-binding domain, revealed small amounts of conventional 67 kDa ER in normal nonreproductive organs from rats. These ®ndings are in line with those of Echeverria et al. [20] who, after detecting ER in the nucleus of several reproductive and nonreproductive organs by immunohistochemistry with a polyclonal antibody directed against the DNA-binding domain

103

of ER, suggested that E2 exerts its effects through a common nuclear mechanism in the cells of both reproductive and nonreproductive organs. The additional identi®cation by our experimental procedure of 43 and 35 kDa ER peptides that bind TAZ but not E2 in many organs raises many questions. Is this discordant phenotype a methodological artefact? Are these peptides ER variants and, if so, what is their function? Do they act as ligand-independent transcription factors of estrogen-responsive genes? Could they belong to a subclass of the steroid/ thyroid/vitamin superfamily known as orphan receptors [21]? Although some orphan receptors may bind as yet unidenti®ed ligands to function, others may not require any ligand to activate transcription, perhaps through transformation by signaling pathways or by protein±protein interactions. Accordingly, it could be of interest to analyze the potential binding ability of nonsteroidal estrogens by such peptides, to know whether they may or not induced estrogenic response; assessment of antiestrogenic binding would be also informative. [125 I]TAZ also detected another type of band, a 55 kDa band, that was found in BT-20 cells and nonreproductive organs even when the anti-ER MAb was omitted. This property distinguishes the corresponding peptide from all others. BT-20 cells are classi®ed as ER-negative by [3 H]E2 binding assays but do express low levels of wild-type ER transcript as well as a variant ER mRNA which, when expressed in yeast, is translated into a 42 kDa protein unable to bind E2 [22]. The presence of this 55 kDa band cannot be explained by `natural' anti-ER autoantibodies in the secondary antibody preparation (anti-rat serum) [23] because KCl was added before labeling, to prevent any cross-reaction between such autoantibodies and ER. This 55 kDa peptide might therefore be a nonreceptor protein tightly associated with the ER: Greene et al. [24] have described a component that does not bind steroid but that is recognized by the MAb prepared against ER. That ER is needed for the 55 kDa band to be present is con®rmed by its absence in ER-negative CHO cells. Because of the large hydrophobic radionuclide of [125 I]TAZ and its covalent attachment to ER, [125 I]TAZ may prevent the dissociation of a complex between ER and a component that does not bind steroid. Finally, in multidrug resistant cells, [3 H]TAZ has been reported to bind to an overexpressed membrane phosphoglucoprotein, termed P-glucoprotein (Pgp) [25]; this probe also binds to calmodulin [26]. Future studies of the af®nity-labeling properties of [125 I]TAZ could thus be directed not only toward the characterization of ER isoforms and of antiestrogen binding proteins but to the study of these other proteins.

104

Younes Maarou® et al.

AcknowledgementsÐThis work was supported by grants from the Fonds Medic and the Fonds J.-C. Heuson. Y. M. was a recipient of a fellowship of the Rose and Jean Hoguet Foundation.

REFERENCES 1. Leclercq G., Technical pitfalls, methodological improvements and quality controls of steroid hormone receptor assays. Eur. J. Cancer 23 (1987) 453±458. 2. Kute T. E., Heidemann P. and Wittliff J. L., Molecular heterogeneity of cytoplasmic forms of oestrogen receptors from human breast tumours. Cancer 38 (1978) 4307±4313. 3. Murphy L. C., Dotzlaw H., Leygue E., Douglas D., Coutts A. and Watson P. H., Estrogen receptor variants and mutations. J. Steroid Biochem. Mol. Biol. 62 (1997) 363±372. 4. Katzenellenbogen J. A., Carlson K. E., Herman D. F. and Robertson D. W., Ef®cient and highly selective covalent labeling of the estrogen receptor with [3 H]tamoxifen aziridine. J. Biol. Chem. 258 (1983) 3487±3497. 5. Elliston J. F. and Katzenellenbogen B. S., Comparative analysis of estrogen receptors covalently labeled with an estrogen and an antiestrogen in several estrogen target cells as studied by limited proteolysis. J. Steroid Biochem. 29 (1988) 559±569. 6. Piccart M., Muquardt C., Bosman C., Pirotte P., Veenstra S., Grillo F. and Leclercq G., Comparison between tritiated estradiol and tamoxifen aziridine for measuring estrogen receptors in human breast cancer cytosols. J. Natl. Cancer Inst. 83 (1991) 1553±1559. 7. Monsma F. J., Katzenellenbogen B. S., Miller M. A., Ziegler Y. S. and Katzenellenbogen J. A., Characterization of the estrogen receptor and its dynamics in MCF-7 human breast cancer cells using a covalently attaching antiestrogen. Endocrinology 115 (1984) 143±153. 8. Trivedi S., Piccart M., Muquardt C., Gilot N., Hadiy S., Patel D. and Leclercq G., Tamoxifen aziridine labeling of the estrogen receptor: potential utility in detecting biologically aggressive breast tumors. Breast Cancer Res. Treat. 40 (1996) 231± 241. 9. Salituro F. G., Carlson K. E., Elliston J. F., Katzenellenbogen B. S. and Katzenellenbogen J. A., [125 I]iododesethyl tamoxifen aziridine: synthesis and covalent labeling of the estrogen receptor with an iodine-labeled af®nity label. Steroids 48 (1986) 287±313. 10. Tonnesen G. L., Hanson R. N. and Seitz D. E., Positionspeci®c radioiodination utilizing an aryltributylstannyl intermediate. Synthesis of 125 I-iodotamoxifen. Int. J. Appl. Rad. Isot. 32 (1981) 171±173. 11. McCague R. and Potter G. A., Synthesis of 4-stannylated tamoxifen analogues: useful precursors to radiolabelled idoxifene and aziridinyl 4-iodotamoxifen. J. Labeled Compounds Radiopharm. 34 (1994) 297±302. 12. Wooge C. H., Nilsson G. M., Heierson A., McDonnell D. P. and Katzenellenbogen B. S., Structural requirements for high af®nity ligand binding by estrogen receptors: a comparative analysis of truncated and full length estrogen receptors expressed in bacteria, yeast, and mammalian cells. Mol. Endocr. 6 (1992) 861±869.

13. Leclercq G., Legros N. and Piccart M. J., Accumulation of a nonbinding form of estrogen receptor in MCF-7 cells under hydroxytamoxifen treatment. J. Steroid Biochem. Molec. Endocr. 41 (1992) 545±552. 14. Quivy J., Leclercq G., Deblaton M., Henrot P., Velings N., Norberg B., Evrard G. and Zeicher M., Synthesis, structure and biological properties of Z-17a-(2-iodovinyl)-11b-chloromethyl estradiol-17b, a high af®nity ligand for the characterization of estrogen receptor-positive tumors. J. Steroid Biochem. Molec. Biol. 59 (1996) 103±117. 15. Kushner P. J., Hort E., Shine J., Baxter J. D. and Greene G. L., Construction of cell lines that express high levels of the human estrogen receptor and are killed by estrogens. Mol. Endocr. 10 (1990) 1465±1473. 16. Harlow K. W., Smith D. N., Katzenellenbogen J. A., Greene G. L. and Katzenellenbogen B. S., Identi®cation of cysteine 530 as the covalent attachment site of an af®nity-labeling estrogen (ketonenestrol aziridine) and antiestrogen (tamoxifen aziridine) in the human estrogen receptor. J. Biol. Chem. 264 (1989) 17476±17485. 17. Horigome T. F., Ogata T. S., Golding F. and Korach K. S., Estradiol-stimulated proteolytic cleavage of the estrogen receptor in mouse uterus. Endocrinology 123 (1988) 2540±2548. 18. Maarou® Y., Trivedi S. and Leclercq G., Major molecular weight heterogeneity of estrogen receptor from breast cancer is not related to the neoplasia. Cancer Biochem. Biophys. 15 (1995) 67±78. 19. Giambiagi N. and Pasqualini J. R., Interaction of three monoclonal antibodies with the nonactivated and activated forms of the estrogen receptor. Endocrinology 126 (1990) 1403±1409. 20. Echeverria O. M., Maciel A. G., Traish A. M., Wotiz H. H., Ubaldo E. and Vazquez-Nin G. H., Immuno-electron microscopic localization of estradiol receptor in cells of male and female reproductive and nonreproductive organs. Biol. Cell. 81 (1994) 257±265. 21. O'Malley B. W. and Connelly O. M., Orphan receptors: in search of a unifying hypothesis for activation. Mol. Endocrinol. 6 (1992) 1359±1361. 22. Castles C. G., Fuqua S. A. W., Klotz D. M. and Hill S. M., Expression of a constitutively active estrogen receptor variant in the estrogen receptor-negative BT 20 human breast cancer cell line. Cancer Res. 53 (1993) 5934±5939. 23. Borkowsky A., Gyling M., Muquardt C., Body J. J. and Leclercq G., A subpopulation of immunoglobulin G in man selectively interacts in the hormone-binding sites of estrogen receptors. J. Clin. Endocrinol. Metab. 64 (1987) 356±363. 24. Lorincz M. A., Holt J. A. and Greene G. L., Monoclonal antiobody recognition of multiple forms of estrogen receptor tagged with [125 I]methoxy iodovinyl estradiol in ovarian carcinomas. J. Clin. Endocrinol. Metab. 61 (1985) 412±417. 25. Safa A. R., Roberts S., Agresti M. and Fine R. L., Tamoxifen aziridine, a novel af®nity probe for P-glucoprotein in multidrug resistant cells. Biochem. Biophys. Res. Comm. 202 (1994) 606± 612. 26. Lopes M. F. C., Vale M. G. P. and Carvalho A. P., Ca2+dependent binding of tamoxifen to calmodulin isolated from bovine brain. Cancer Res. 50 (1990) 2753±2758.