Tamoxifen−DNA Adducts Detected in the Endometrium of Women

Women treated for breast cancer with tamoxifen are at increased risk of developing endometrial cancer. This carcinogenic effect has been attributed to...
0 downloads 0 Views 98KB Size
Download PDF
646

Chem. Res. Toxicol. 1999, 12, 646-653

Tamoxifen-DNA Adducts Detected in the Endometrium of Women Treated with Tamoxifen† Shinya Shibutani,*,‡ Naomi Suzuki,‡ Isamu Terashima,‡ Steven M. Sugarman,§ Arthur P. Grollman,‡,§ and Michael L. Pearl| Department of Pharmacological Sciences, Department of Medicine, and Department of Obstetrics and Gynecology, School of Medicine, State University of New York at Stony Brook, Stony Brook, New York 11794 Received February 26, 1999

Women treated for breast cancer with tamoxifen are at increased risk of developing endometrial cancer. This carcinogenic effect has been attributed to estrogenic stimulation and/ or to a genotoxic effect of this drug. To examine genotoxicity, we developed a 32P-postlabeling TLCL/HPLC procedure for quantitative analysis of tamoxifen-DNA adducts in endometrial tissue. This assay is several orders of magnitude more sensitive than those previously used for this purpose; with it, we can detect five tamoxifen-DNA adducts in 1011 bases. Endometrial tissue was obtained from women undergoing tamoxifen therapy and from untreated control subjects. DNA adducts, identified as trans and cis epimers of R-(N2-deoxyguanosinyl)tamoxifen, were detected in six of thirteen patients in the tamoxifen-treated group. Levels of trans and cis adducts ranged from 0.5 to 8.3 and from 0.4 to 4.8 adducts/108 nucleotides, respectively. Tamoxifen-DNA adducts were not detected in endometrial tissue obtained from the control subjects. We conclude from this study that one or more tamoxifen metabolites react with endometrial DNA to form covalent adducts, establishing the potential genotoxicity of this drug for women and suggesting the use of TAM-DNA adducts as biomarkers for investigations of tamoxifen-induced endometrial cancer.

Introduction More than 500 000 women with breast cancer in the United States are currently being treated with tamoxifen (TAM)1 (1). In addition, approximately one-third of the 180 000 new cases of breast cancer diagnosed this year will require antiestrogen therapy (2, 3). Recently, TAM was approved for use in women considered to be at high risk of developing breast cancer, raising hope that emergence of this disease, which affects one of nine women in the United States, might be prevented or delayed (4). However, the potential benefits obtained when TAM is used as a chemopreventive agent in otherwise healthy women must be weighed against the increased risk of thromboembolic events and endometrial cancer associated with the use of this drug (5-11). The association between TAM therapy and endometrial cancer was first recognized in 1985 (9). After similar reports appeared, a combined analysis of all randomized trials of adjuvant TAM therapy concluded that the incidence of endometrial cancer in women receiving † Preliminary reports of this work were presented before the American Association for Cancer Research, March 1998, in New Orleans, LA, and at the 21st Annual San Antonio Breast Cancer Symposium, December 1998, in San Antonio, TX. * To whom correspondence should be addressed. Phone: (516) 4448018. Fax: (516) 444-3218. E-mail: [email protected]. ‡ Department of Pharmacological Sciences. § Department of Medicine. | Department of Obstetrics and Gynecology. 1 Abbreviations: dG, 2′-deoxyguanosine; dG , 2′-deoxyguanosine 3′P 3′-monophosphate; TAM, tamoxifen; R-OHTAM, R-hydroxytamoxifen; 2 2 dG-N -TAM, R-(N -deoxyguanosinyl)tamoxifen; dG3′P-N2-TAM, dG 3′monophosphate-N2-TAM; fr-1 and fr-2, epimers of the trans form of dG3′P-N2-TAM; PEI, poly(ethyleneimine).

tamoxifen doubled in trials of one or two years and quadrupled over a period of five years (10). The average increase in endometrial cancer for 37 000 women treated with tamoxifen for early breast cancer was 2.58 (2p < 0.000 001) (10). An increased incidence of endometrial cancer also was observed in women enrolled in a TAM chemoprevention trial (4). TAM-induced endometrial cancer has been attributed to the partial agonist (estrogenic) effects of this drug on the human uterus (6, 12). TAM also could induce endometrial cancer via its genotoxic effects (6, 13). The latter view is supported by observations that TAM is a potent hepatocarcinogen in rats (14, 15) and that an activated form of the drug (16-18) forms covalent adducts with DNA (19-21). So far, attempts to detect DNA adducts in women treated with tamoxifen have resulted in conflicting reports. Using standard 32P-labeling techniques, Phillips et al. failed to detect TAM-DNA adducts in human leucocytes (22) or in endometrial tissue (23). Using a 32Ppostlabeling HPLC method, Hemminki et al. reported low levels of a TAM adduct in human leucocytes (5.5 adducts/ 109 nucleotides) (24) and in endometrial tissue (2.7 adducts/109 nucleotides) (25). This analysis did not utilize an authentic standard, and the level of radioactivity associated with the putative adduct was only twice that of background, leading one group to question the significance of this report (26). To help resolve this issue and to establish the mechanism(s) involved, we developed an ultrasensitive, highly specific procedure for quantifying TAM adducts in endometrial DNA. This method employs a 32P-postlabeling

10.1021/tx990033w CCC: $18.00 © 1999 American Chemical Society Published on Web 06/16/1999

Tamoxifen-DNA Adducts in Human Endometrium

Chem. Res. Toxicol., Vol. 12, No. 7, 1999 647 additionally incubated for 30 min with 1.5 mg of proteinase K. Reaction mixtures were extracted sequentially with phenol, 1:1 (v/v) phenol/chloroform, and chloroform. Following ethanol precipitation, the DNA was washed twice with 500 µL of 70% cold ethanol. These procedures were repeated to minimize contamination, including RNA (30). The purified DNA was dissolved in 1 mL of 0.01× SSC (1× SSC is 0.15 M NaCl/0.015 M sodium citrate)/1 mM EDTA. The concentration of endometrial DNA was estimated with the relation 50 µg ) 1.0 OD260nm. Approximately 100-250 µg of endometrial DNA was obtained using this procedure.

Figure 1. Mechanism for the formation of cis and trans dGN2-TAM-DNA adducts.

TLC/HPLC technique combined with a butanol extraction procedure. Epimers of the trans (fr-1 and fr-2) and cis (fr-3 and fr-4) forms of 32P-labeled R-(N2-deoxyguanosinyl)tamoxifen 3′-monophosphate (dG3′P-N2-TAM) (Figure 1) were used as standards. The assay can detect five TAM-DNA adducts in 1011 bases and is several orders of magnitude more sensitive and more specific than other methods used to detect this lesion (22-25). Using it, trans and cis TAM-DNA adducts were detected in endometrial tissue of six of thirteen women treated with the drug. These lesions were not detected in endometrial DNA obtained from untreated control subjects. We conclude that one or more TAM metabolite(s) can form covalent adducts with endometrial DNA. These adducts are known to miscode (27) and must be associated with mutagenic events (28, 29).

Materials and Methods Chemicals. [γ-32P]ATP (specific activity of 6000 Ci/mmol) was obtained from Amersham Corp. (Arlington Heights, IL). PEI-cellulose plates were purchased from Machery-Nagel (Duren, Germany). Proteinase K, and potato apyrase and nuclease P1, were purchased from Sigma (St. Louis, MO), and Boehringer Mannheim (Indianapolis, IN), respectively. RNase A, RNase T1, micrococcal nuclease, and spleen phosphodiesterase were obtained from Worthington Biochemical Co. (Freehold, NJ). DNA Extraction from Endometrial Tissues. Endometrial tissue was collected by endometrial biopsy utilizing a Pipelle needle and by hysterectomy; samples were stored at -80 °C until the analysis was performed. This clinical research study was approved by the Institutional Review Board of the State University of New York at Stony Brook (97-2578 and 98-2578); informed consent was obtained from all patients. Endometrial tissue (50-125 mg) suspended in 3 mL of 1.0% SDS/10 mM EDTA/20 mM Tris-HCl (pH 7.4) was homogenized for 30 s at 0 °C using a Polytoron homogenizer. Following an established protocol (30), the homogenate was incubated at 37 °C for 30 min with 600 µg of RNase A and 102 units of RNase T1, and then

Isolation of DNA Adducts. Samples containing 25 µg of DNA were digested at 37 °C for 2 h in 30 µL of 17 mM sodium succinate buffer (pH 6.0) containing 8 mM CaCl2, using 3.0 units of micrococcal nuclease and 0.3 unit of spleen phosphodiesterase. Afterward, 2.0 units of nuclease P1 were added, and the reaction mixture was incubated at 37 °C for 1 h. Samples were dissolved in 100 µL of distilled water and extracted twice with 200 µL of butanol. The butanol fraction was back-extracted with 50 µL of distilled water, dried, and then used for analysis of TAM-DNA adducts. As reported previously, after the reaction of calf thymus DNA with tamoxifen R-sulfate (31), approximately 95% of the TAM adducts can be recovered by butanol extraction. The pooled extracts were incubated at 37 °C for 40 min with 4 µL of [γ-32P]ATP (approximately 10 µCi/µL), 3 µL of T4 polynucleotide kinase (10 units/µL), and 23 µL of 70 mM Tris-HCl buffer (pH 7.6) containing 10 mM MgCl2 and 5 mM dithiothreitol, as described previously (32). The reaction mixture was further incubated at 37 °C for 30 min with 1 µL of potato apyrase (80 × 10-3 unit/ µL). Labeling efficiencies of the standard trans and cis forms of dG3′P-N2-TAM were 65 and 56%, respectively. The 32P-labeled sample was developed for 16 h on a 10 cm × 10 cm PEI-cellulose thin-layer plate using 1.7 M sodium phosphate buffer (pH 6.0) (D1 buffer), using a paper wick. 32P-labeled products remaining on the TLC plate were recovered, using 4 M pyridinium formate (pH 4.3), and evaporated to dryness. Recovery of 32P-labeled material was approximately 84%. Part of the 32P-labeled products was again developed on a 10 cm × 10 cm PEI-cellulose TLC plate using three different solvents (32). Plates were developed for 16 h in 1.7 M sodium phosphate buffer (pH 6.0) (D1 buffer) with a paper wick, and subsequently developed in the same direction with 1.1 M lithium formate and 2.7 M urea (pH 3.5) (D2 buffer). The plates were further developed at a right angle to the previous direction of development in 0.48 M LiCl, 0.3 M Tris-HCl, and 5.1 M urea (pH 8.0) (D3 buffer). The position of adducts was established by β-phosphorimaging (Molecular Dynamics Inc.) or by autoradiography, using Kodak Xomat XAR film. Radioactive spots on PEI-cellulose were scraped from the plates, and the radioactivity was measured by scintillation counting, as described previously (32). Relative adduct levels were calculated according to the method of Levay et al. (33), using disintegrations per minute instead of counts per minute (total disintegrations per minute in adducts)/1.01 × 1012 dpm, assuming that 25 µg of DNA represented 7.58 × 104 pmol of dN3′P and that the specific activity of the [γ-32P]ATP was 1.33 × 107 dpm/pmol. The detection limit was approximately 5.0 × 10-11 adduct. The specific activity of [γ-32P]ATP was corrected for the extent of decay. Assessment of 32P-Labeled DNA Adducts by HPLC. Standard stereoisomers of dG3′P-N2-TAM were prepared as described previously (32) and labeled with 32P (34). A portion of the 32P-labeled products recovered from PEI-cellulose after development with 1.7 M sodium phosphate buffer (pH 6.0) was subjected to a Supelcosil LC-18S column (0.46 cm × 25 cm, Supelco), eluted over the course of 30 min at a flow rate of 1.0 mL/min with a linear gradient of 0.2 M ammonium formate (pH 4.2), containing 10 to 70% methanol. Fractions (1.0 mL) were collected, and the radioactivity was measured by scintillation counting. HPLC analysis was performed on a Waters 990 HPLC instrument.

648 Chem. Res. Toxicol., Vol. 12, No. 7, 1999

Shibutani et al.

Table 1. Patient Information

patient

age (years)

tamoxifen dose (mg/day)

C1 C2

33 37

N/A N/A

N/A N/A

C3

41

N/A

N/A

C4 C5

41 48

N/A N/A

N/A N/A

C6

38

N/A

N/A

C7 C8

44 44

N/A N/A

N/A N/A

dysmenorrhea cervical carcinoma in situ leiomyomata and adenomyoma fibroids fibroids and adenomyosis cervical carcinoma in situ prolapse fibroids

C9 C10 T1 T2 T3

43 37 46 44 68

N/A N/A 20 20 20

N/A N/A 4 34 59

fibroids fibroids breast cancer breast cancer breast cancer

T4 T5

45 58

20 20

37 3

T6 T7 T8

29 49 75

20 20 20

1 40 34

T9 T10 T11

46 38 73

20 20 20

7 7 72

T12

76

20

15

T13

71

20

23

breast cancer adenomyosis, leiomyoma, and prolapse breast cancer fibroids endometrial adenocarcinoma fibroids breast cancer breast and endometrium cancer breast cancer, adenomatous colonic polyps, and benign endmetrial polyps breast cancer

duration (months)

diagnosis

Figure 2. TLC analysis of 32P-labeled dG3′P-N2-TAM. An authentic standard of (A) trans or (B) cis forms of 32P-labeled dG3′P-N2-TAM was developed on a PEI-cellulose plate as described in Materials and Methods.

Results Endometrial tissue was collected from 13 women undergoing treatment with TAM (20 mg/day) and from 10 women not receiving TAM or another form of hormonal therapy. The clinical diagnosis for each patient and duration of TAM therapy are listed in Table 1. Endometrial DNA was prepared for adduct analysis as described in Materials and Methods. Standard trans and cis forms of 32P-labeled dG3′P-N2TAM were placed on a PEI-cellulose TLC plate and developed with three different buffers (Figure 2). Using this system, trans and cis forms can be resolved. Two major adducts (a and b), denoted by the arrows (Figure 3), were detected in patients T1 and T11. The position of

other drug used

alcohol/tobacco

Synthroid, Propulsid, Prilosec none

none/1 pack per year none/none

Imatrex, Naprosyn, Butalbital/APAP/Caffeine iron none

social/none none/none none/none

none

none/yes

none Mylanta

none/none 1 glass of wine per day/none none/none none/none social/none social/none none/15 packs per year, none in 14 years social/none none/none

Feosol none none Nortriptyline Calan, Dyazide, Melacor, Welbutrin, Relafin none Levobunolol, potassium citrate, Coumadin none none Metoprolol, Lasix, Ancef, Maalox

none/none none/none none/none

none insulin Axid, Xanax

none/none none/none none/none

Coumadin, Cardiazam, Vasotec, Synthroid, K-dur, vitamin B12, Digoxin, Lopressor, Lasix, Cholchicine, Pepcid, Fosamax

none/none

Atenolol, Zestril

social/none

adducts a and b on the TLC corresponded to those of the trans and cis forms of dG3′P-N2-TAM, respectively (Figure 2), which was confirmed by cochromatography with authentic standards (data not shown). The trans form of dG-N2-TAM was the major adduct in T7 and T10 (Figure 3); the cis form was the major adduct in T4 and T8. Minor adducts were observed, prominently in T1 and T11. TAM adducts were not detected in T2, T5, and T12 (Figure 3), or in T3, T6, T9, and T13 (data not shown). TAM-DNA adducts also were not detected in the endometrial tissue of patients not treated with the drug (C3, C6, and C8 in Figure 3; for other control subjects, data are not shown). The levels of trans and cis forms of dG3′P-N2-TAM were determined in the six adduct-containing samples (Table 2). Levels of the trans adduct ranged from 0.5 to 8.3 adducts/108 nucleotides and levels of the cis adduct from 0.4 to 4.8 adducts/108 nucleotides. Authentic trans (fr-1 or fr-2) and cis forms of dG3′PN2-TAM were labeled with [γ-32P]ATP and subjected to HPLC (32). The trans form of dG3′P-N2-TAM was separated from the cis form; retention times of the fr-1 and fr-2 trans forms were 35 (Figure 4A) and 36 min (Figure 4B), respectively. The retention time of the cis form was 41 min (Figure 4C). When a portion of the 32P-labeled products of T1 was subjected to HPLC, retention times for material eluted in the two primary peaks were 35 and 41 min (Figure 4). These values are consistent with retention times observed for the trans (fr-1) and cis forms of dG3′P-N2-TAM, respectively. The first peak exhibits a

Tamoxifen-DNA Adducts in Human Endometrium

Chem. Res. Toxicol., Vol. 12, No. 7, 1999 649

Figure 3. Detection of TAM-DNA adducts in endometrial tissue. Endometrial DNA (25 µg) obtained from TAM-treated patients with breast cancer or from untreated patients was digested enzymatically and extracted by butanol, as described in Materials and Methods. The butanol fraction was evaporated to dryness and then labeled with 32P. 32P-labeled samples were developed on a PEIcellulose TLC plate using 1.7 M sodium phosphate buffer (pH 6.0) with a paper wick. 32P-labeled products remaining on the plate were eluted with 4 M pyridinium formate (pH 4.3). One-fifth of the 32P-labeled products was again developed on a PEI-cellulose TLC plate using three different buffer solutions, as described in Materials and Methods. The position on the plates was compared with standard 32P-labeled dG3′P-N2-TAM. T1, T4, T7, T8, T10, T11, T2, T5, and T12 denote endometrial samples obtained from patients treated with TAM. C3, C6, and C8 denote samples from untreated control subjects. The X-ray film was exposed for 14 h with samples from C3, C6, C8, T1, T4, T11, T2, T5, and T12, for 8 h with samples from T7 and T8, and for 24 h with the sample from T10.

shoulder at 36 min and may contain the other trans isomer (fr-2). When the major adducts of T1 (a and b) were recovered from the TLC plate (Figure 3) and subjected to HPLC (data not shown), their retention times were identical to those of the trans and cis forms

of dG3′P-N2-TAM, respectively. Both the trans and cis forms of dG3′P-N2-TAM were detected in T4, T7, and T11. The trans form was detected only in T10, while in T8, the cis form was the major product (data not shown). TAM-DNA adducts were not detected in endometrial

650 Chem. Res. Toxicol., Vol. 12, No. 7, 1999

Shibutani et al.

Table 2. Levels of dG-N2-TAM-DNA Adducts in TAM-Treated Endometrial Samplesa duration of trans form cis form total therapy (adducts/108 (adducts/108 (adducts/108 sample (months) nucleotides) nucleotides) nucleotides) T1 T4 T7 T8 T10 T11

4 37 40 34 7 72

5.5 1.3 4.6 0.46 1.5 8.3

4.6 3.2 0.37 4.2 ND 4.8

10.1 4.5 5.0 4.7 1.5 13.1

a Data taken from the TLC analysis shown in Figure 3. The reported adduct level takes into account recovery and postlabeling efficiency. ND means not detectable.

tissue obtained from seven TAM-treated women (for example, T2 in Figure 4) or from 10 untreated control subjects (for example, C3 in Figure 4). To evaluate the efficiency of butanol extraction in the analysis of TAM-DNA adducts and to establish the degree to which free 32P interferes with the detection of TAM-DNA adducts, a sample known to contain TAM adducts (T11) was digested with micrococcal nuclease, spleen phosphodiesterase, and nuclease P1. Half the digest was labeled with 32P; the remainder was extracted twice with butanol, evaporated to dryness, and then labeled with 32P. The step involving elution with D1 buffer was omitted. 32P-labeled samples were developed on a PEI-cellulose plate using the three different buffers. The trans form (a) and the cis form (b) of dG3′P-N2-TAM adducts were detected in the butanol-extracted sample (Figure 5B). When the butanol extraction step was omitted (Figure 5A), the trans form TAM adduct (a) was only faintly observed. Detection was not improved by exposing the film for longer periods of time. In contrast, when 32P-labeled products were eluted from the TLC plate with D1 buffer (Figure 3, see T11), background did not interfere with adduct detection, even when the film was exposed for 14 h.

Discussion By identifying TAM-DNA adducts in endometrial tissue, this study establishes a potential genotoxic effect of this drug. Central to this finding was the development of a sensitive, reproducible assay for TAM-induced lesions. TAM-DNA adducts can be labeled efficiently with 32P. To minimize background, free 32P was removed by development on the TLC plates. To increase the sensitivity and specificity of postlabeling, a butanol extraction step was introduced. This procedure efficiently extracts TAM-DNA adducts following enzymatic digestion, separating them from unmodified nucleosides (31, 32, 35). When the butanol extraction step was omitted, TAMDNA adducts were detected only in trace amounts, accounting possibly for the failure of Carmichael et al. to detect TAM-DNA adducts in human endometrial tissue (23). Via the demonstration that TAM-DNA adducts comigrated with authentic standards and were absent in endometrial tissue obtained from untreated control subjects, the objections raised by Orton et al. (26) to the report of Hemminki et al. (25) do not pertain to the study presented here. Approximately one DNA adduct in 108-109 nucleotides can be detected by conventional 32P-postlabeling methods; the 32P-postlabeling system used in this paper increases the limit of detection for TAM-DNA adducts to five adducts in 1011 bases. The

butanol extraction step and removal of free 32P by TLC are essential to achieving this level of sensitivity. Reactive TAM species are formed by oxidation (3639). Products include R-OHTAM and its principal metabolites: tamoxifen N-oxide, N-desmethyltamoxifen, and 4-hydroxytamoxifen (40-43). Sulfation of R-OHTAM is catalyzed by rat and human hydroxysteroid sulfotransferases (31, 32), leading to formation of four trans and cis diastereoisomers of R-(N2-deoxyguanosinyl)tamoxifen (dG-N2-TAM) (Figure 1) (35). These isomers interconvert via a short-lived carbocation intermediate (44). Trans and cis forms of dG-N2-TAM are identified here as the major DNA adducts in six endometrial samples obtained from women treated with TAM. This observation is consistent with the report that the major TAM adduct detected in the liver of rats treated with TAM or R-OHTAM was dGN2-TAM (45). Adducts corresponding to 4-hydroxytamoxifen or tamoxifen 1,2-epoxide were not detected (45), supporting our proposal that the primary route of TAM metabolism is by sulfation of R-OHTAM (31, 32, 35). The several unidentified minor adducts may be induced by R-hydroxylation of TAM metabolites, such as R-hydroxytamoxifen N-oxide and N-desmethy-R-hydroxytamoxifen. Steady-state levels of TAM-DNA adducts in endometrial tissue are determined by the rates of formation and repair of these lesions. Trans and cis forms of dG-N2TAM block DNA synthesis in vitro (27) and, as bulky lesions, most likely are excised by nucleoside excision repair. When the trans form of R-OHTAM was incubated with DNA in the presence of human hydroxysteroid sulfotransferase (hHST) and 3′-phosphoadenosine 5′phosphosulfate, the ratio of trans to cis adducts formed was 8:1 (32). In two endometrial samples, the trans form of dG-N2-TAM adduct was dominant; in four other samples, the cis form was exclusively present, the ratio of trans to cis forms ranging between 0.1 and 1.7. This striking variation among individuals may reflect differential adduct repair. If the cis adduct was repaired more slowly than the trans adduct, over time, it could become the dominant TAM-DNA adduct in endometrial tissue. TAM-DNA adducts were not detected in endometrial samples obtained from several of the patients treated with TAM. Sulfation of R-OHTAM is required for the formation of dG-N2-TAM adducts (31), and the activity of hHST among individuals may vary. Furthermore, R-OHTAM analogues may be a substrate for gluthathione S-transferases (46). Thus, if R-OHTAM is detoxified rapidly, TAM-DNA adducts might not be detected in endometrial tissue. Two patients (T1 and T11) exhibited a relatively high level of TAM adducts (1.0-1.3 adducts/ 107 nucleotides). Subject T1 was treated with TAM for only 4 months. TAM-DNA adducts were not detected in patients T6 and T5, treated for 1 and 3 months, respectively. It appears that some women accumulate TAMDNA adducts more rapidly than others, possibly due to variations in nucleotide excision repair (47) or to differing rates of R-OH-TAM metabolism. Further studies, including measurement of activities and/or polymorphisms for gluthathione S-transferases, sulfotranferases, and nucleotide excision repair in human endometrium, are required. Such studies, performed in conjunction with analyses of TAM-DNA adducts, would be useful in monitoring human exposure to potential toxicity of this widely used drug (48-50).

Tamoxifen-DNA Adducts in Human Endometrium

Chem. Res. Toxicol., Vol. 12, No. 7, 1999 651

Figure 4. Assessment of TAM-DNA adducts in endometrial tissue using HPLC. A standard of the fr-1 (A) and fr-2 (B) trans forms or cis form (C) of 32P-labeled dG3′P-N2-TAM recovered from a PEI-cellulose TLC plate, as described in Materials and Methods, was subjected to a Supelcosil LC-18S column (0.46 cm × 25 cm, Supelco). A linear gradient of 0.2 M ammonium formate (pH 4.2) containing 10 to 70% methanol was eluted over the course of 30 min at a flow rate of 1.0 mL/min. The eluate was collected in 1.0 mL fractions and radioactivity measured using a scintillation counter. One-fifth of the 32P-labeled products (C3, T1, and T2) recovered from the TLC plates in Figure 3 were subjected to HPLC.

Figure 5. Effect of butanol extraction on the assessment of TAM-DNA adducts. Five micrograms of endometrial DNA (T11) was digested enzymatically, as described in Materials and Methods. (A) Half of the digest was directly labeled with 32P. (B) The other half was dissolved in 100 µL of distilled water and extracted twice with 200 µL of butanol. The butanol fraction was back-extracted with 50 µL of distilled water. Butanol fractions were evaporated to dryness and labeled with 32P. Both 32P-labeled samples were subjected to TLC on a PEI-cellulose plate and developed using three different buffers as described in Materials and Methods. The X-ray film was exposed for 1.5 h.

The level of TAM-DNA adducts detected in endometrial tissue was lower than that reported (8 adducts/107 nucleotides) for liver DNA of rats treated with a dose of tamoxifen that generates 50% hepatocarcinomas (48); however, it is similar to the level of 7-alkyl-2′-dG adducts (0.1-4.7 adducts/107 nucleotides) detected in autopsy samples obtained from human lung tissue (51) and much higher than the level of benzo[a]pyrene diol epoxide (BPDE) adducts (0.45 adduct/107 nucleotides) detected in lung tissue of smokers (52). dG-N2-TAM adducts

exhibit a higher mutagenic potential in simian kidney cells (29) than DNA adducts derived from the established chemical carcinogen, acetylaminofluorene (53). Thus, individuals with TAM-DNA adducts are at risk for developing mutations that could initiate endometrial cancer. The antiestrogenic effects of TAM account for its effectiveness in the treatment and prevention of breast cancer; there is reason to believe that the therapeutic and chemopreventive properties of this drug may be separated from its carcinogenic effects. We base this proposal on structure-activity relationships of antiestrogens (54, 55) and the observation that metabolism to a DNA reactive form requires R-hydroxylation of the ethyl side chain of TAM (32, 35). In principle, R-hydroxylation could be minimized or abolished by modifying the structure of tamoxifen while retaining the antiestrogenic properties of this molecule. Such structures exist in the form of raloxifene and toremifene. Raloxifene (55), a selective estrogen response modifier, was shown to reduce the incidence of breast cancer in women at high risk of developing this disease (56). Raloxifene binds to the estrogen receptor; however, its metabolic products are unlikely to react with DNA due to the absence of the ethyl moiety. Unlike TAM, raloxifene does not demonstrate proliferative effects on the uterus of postmenopausal women (57). Toremifene (58) has estrogenic effects on the human endometrium, but under conditions where TAM readily forms DNA adducts, toremifene reacts only weakly with DNA (59-61). Both raloxifene and toremifene lack the potent hepatocarcinogenicity of TAM in rodents.

652 Chem. Res. Toxicol., Vol. 12, No. 7, 1999

Both these FDA-approved drugs are likely to have less endometrial genotoxicity than tamoxifen, an important consideration in selecting antiestrogens for the chemoprevention of breast cancer.

Acknowledgment. We are grateful to Dr. Eva Chalas for her advice and assistance throughout this investigation and to Ms. Susan Rigby for preparing the manuscript. This research was supported in part by NIH Grants ES04068, ES09418, and CA17395, a generous award from the Norman M. Morris Foundation, and a pilot study grant from the School of Medicine of the State University of New York.

References (1) Peto, R. (1996) Five years of tamoxifensor more? J. Natl. Cancer Inst. 88, 1791-1793. (2) Ries, L. A. G., Kosary, C. L., Hankey, B. F., et al. (1997) SEER Cancer Statistics Review, 1973-1994: Tables and Graphs. NIH Publication 97-2789, National Cancer Institute, Bethesda, MD. (3) Jordan, V. C. (1995) Tamoxifen for breast cancer prevention. Proc. Soc. Exp. Biol. Med. 208, 144-149. (4) Fisher, B., Costantino, J. P., Wickerham, L., Redmond, C. K., Kavanah, M., Cronin, W. M., Vogel, V., Robidoux, A., Dimitrov, N., Atkins, J., Daly, M., Wieand, S., Tan-Chiu, E., Ford, L., Wolmark, N., et al. (1998) Tamoxifen for prevention of breast cancer: report of the national surgical adjuvant breast and bowel project P-1 study. J. Natl. Cancer Inst. 90, 1371-1388. (5) Fischer, B., Costantino, J. P., Redmond, C. K., Fisher, E. R., Wickerham, D. L., and Cronin, W. M. (1994) Endometrial cancer in tamoxifen-treated breast cancer patients: Findings from the National Surgical Adjuvant Breast and Bowel Project (NSABP, B-14). J. Natl. Cancer Inst. 86, 527-537. (6) Stearns, V., and Gelman, E. P. (1998) Does tamoxifen cause cancer in humans? J. Clin. Oncol. 16, 779-792. (7) Grady, D., and Ernster, V. L. (1996) Endometrial cancer. In Cancer Epidemiology & Prevention (Schottenfeld, D., and Framemi, J. F., Eds.) 2nd ed., pp 1058-1089, Oxford University Press, New York. (8) Mignotte, H., et al. (1998) Iatrogenic risks of endometrial carcinoma after treatment for breast cancer in a large French case-controlled study. Int. J. Cancer 76, 325-330. (9) Killackey, M., Hakes, T. B., and Pierce, V. K. (1985) Endometrial adenocarcinoma in breast cancer patients receiving antiestrogens. Cancer Treat. Rep. 69, 237-238. (10) Clarke, M., Cillins, R., Davies, C., Godwin, J., Gray, R., and Peto, R. (1998) Tamoxifen for early breast cancer: An overview of the randomized trials. Lancet 351, 1451-1467. (11) Barakat, R. R. (1996) Endometrial cancer and tamoxifen. In Tamoxifen: a Guide for Clinicans and Patients (Jordan, V. C., Ed.) pp 101-117, PRR, Inc., Huntington, NY. (12) Kedar, R. P., Bourne, T. H., Powles, T. J., Collins, W. P., Ashley, S. E., Cosgrove, D. O., and Campbell, S. (1994) Effects of tamoxifen on uterus and ovaries of postmenopausal women in a randomised breast cancer prevention trial. Lancet 343, 13181321. (13) Wogan, G. N. (1997) Review of the toxicology of tamoxifen. Semin. Oncol. 24 (Suppl. 1), 87-97. (14) Williams, G. M., Iatropoulos, M. J., Djordjevic, M. V., and Kaltenberg, O. P. (1993) The triphenylethylene drug tamoxifen is a strong liver carcinogen in the rat. Carcinogenesis 14, 315317. (15) Greaves, P., Goonetilleke, R., Nunn, G., Topham, J., and Orton, T. (1993) Two-year carcinogenicity study of tamoxifen in Alderley Park Wistar-derived rats. Cancer Res. 53, 3919-3924. (16) Pathak, D. N., and Bodell, W. J. (1994) DNA adduct formation by tamoxifen with rat and human liver microsomal activation systems. Carcinogenesis 15, 529-532. (17) Pathak, D. N., Pongracz, K., and Bodell, W. J. (1995) Microsomal and peroxidase activation of 4-hydroxytamoxifen to form DNA adducts: comparison with DNA adducts formed in SpragueDawley rats treated with tamoxifen. Carcinogenesis 16, 11-15. (18) Hemminki, K., Widlak, P., and Hou, S. M. (1995) DNA adducts caused by tamoxifen and toremifene in human microsomal system and lymphocytes in vitro. Carcinogenesis 16, 1661-1664. (19) Han, X., and Liehr, J. G. (1992) Induction of covalent DNA adducts in rodents by tamoxifen. Cancer Res. 52, 1360-1363. (20) White, I. N. H., de Matteis, F., Davies, A., Smith, L. L., CroftonSleigh, C., Venitt, S., Hewer, A., and Phillips, D. H. (1992)

Shibutani et al.

(21)

(22)

(23)

(24)

(25)

(26)

(27)

(28)

(29)

(30)

(31)

(32)

(33)

(34)

(35)

(36)

(37)

(38)

(39)

(40)

(41)

Genotoxic potential of tamoxifen and analogues in female Fischer F344/n rats, DBA/2 and C57B1/6 mice and in human MCL-5 cells. Carcinogenesis 13, 2197-2203. Osborne, M. R., Hewer, A., Hardcastle, I. R., Carmichael, P. L., and Phillips, D. H. (1996) Identification of the major tamoxifendeoxyguanosine adduct formed in the liver DNA of rats treated with tamoxifen. Cancer Res. 56, 66-71. Phillips, D. H., Hewer, A., Grover, P. L., Poon, G. K., and Carmichael, P. L. (1996) Tamoxifen does not form detectable adducts in white blood cells of breast cancer patients. Carcinogenesis 17, 1149-1152. Carmichael, P. L., Ugwumadu, A. H. N., Neven, P., Hewer, A. J., Poon, G. K., and Phillips, D. H. (1996) Lack of genotoxicity of tamoxifen in human endometrium. Cancer Res. 56, 1475-1479. Hemminki, K., Rajaniemi, H., Koskinen, M., and Hansson, J. (1997) Tamoxifen-induced DNA adducts in leucocytes of breast cancer patients. Carcinogenesis 18, 9-13. Hemminki, K., Rajaniemi, H., Lindahl, B., and Moberger, B. (1996) Tamoxifen-induced DNA adducts in endometrial samples from breast cancer patients. Cancer Res. 56, 4374-4377. Orton, T. C., Topham, J. C., and Park, A. (1997) Correspondence re: K. Hemminki et al., tamoxifen-induced DNA adducts in endometrial samples from breast cancer patients. Cancer Res. 57, 4148. Shibutani, S., and Dasaradhi, L. (1997) Miscoding potential of tamoxifen-derived DNA adducts: R-(N2-deoxyguanosinyl)tamoxifen. Biochemistry 36, 13010-13017. Davies, R., Oreffo, V. I. C, Martin, E. A., Festing, M. F. W., White, I. N. H., Smith, L. L., and Styles, J. A. (1997) Tamoxifen causes gene mutations in the livers of λ/lacI transgenic rats. Cancer Res. 57, 1288-1293. Terashima, I., Suzuki, N., and Shibutani, S. (1999) Mutagenic potential of R-(N2-deoxyguanosiny)tamoxifen lesions, the major DNA adducts detected in endometrial tissues of patients treated with tamoxifen. Cancer Res. 59, 2091-2095. Umemoto, A., Kajikawa, A., Tanaka, M., Hamada, K., Seraj, M. J., Kubota, A., Nakayama, M., Kinouchi, T., Ohnishi, Y., Yamashita, K., and Monden, Y. (1994) Presence of mucosa-specific DNA adduct in human colon: possible implication for colorectal cancer. Carcinogenesis 15, 901-905. Shibutani, S., Shaw, P., Suzuki, N., Dasaradhi, L., Duffel, M. W., and Terashima, I. (1998) Sulfation of R-hydroxytamoxifen catalyzed by human hydroxysteroid sulfotransferase results in tamoxifen DNA adducts. Carcinogenesis 19, 2007-2011. Shibutani, S., Dasaradhi, L., Terashima, I., Banoglu, E., and Duffel, M. W. (1998) R-Hydroxytamoxifen is a substrate of hydroxysteroid (alcohol) sulfotransferase, resulting in tamoxifen DNA adducts. Cancer Res. 58, 647-653. Levay, G., Pongracz, K., and Bodell, W. J. (1991) Detection of DNA adducts in HL-60 cells treated with hydroquinone and p-benzoquinone by 32P-postlabeling. Carcinogenesis 12, 1181-1186. Maniatis, T., Fritsch, E. F., and Sambrook, J. (1982) Molecular Cloning. A Laboratory Manual, Cold Spring Harbor Laboratory Press, Cold Spring Harbor, NY. Dasaradhi, L., and Shibutani, S. (1997) Identification of tamoxifen-DNA adducts formed by R-sulfate tamoxifen and R-acetoxytamoxifen. Chem. Res. Toxicol. 10, 189-196. Moorthy, B., Sriram, P., Pathak, D. S., Bodell, W. J., and Randerath, K. (1996) Tamoxifen metabolic activation: comparison of DNA adducts formed by microsomal and chemical activation of tamoxifen and 4-hydroxytamoxifen with DNA adducts formed in vivo. Cancer Res. 56, 53-57. Marques, M. M., and Beland, F. A. (1997) Identification of tamoxifen-DNA adducts formed by 4-hydroxytamoxifen quinone methide. Carcinogenesis 18, 1949-1954. Phillips, D. H., Hewer, A., White, I. N. H., and Farmer, P. B. (1994) Co-chromatography of a tamoxifen epoxide-deoxyguanylic acid adduct with a major DNA adduct formed in the livers of tamoxifen-treated rats. Carcinogenesis 15, 793-795. Phillips, D. H., Carmichael, P. L., Hewer, A., Cole, K. J., Hardcastle, I. R., Poon, G. K., Keogh, A., and Strain, A. J. (1996) Activation of tamoxifen and its metabolite R-hydroxytamoxifen to DNA-binding products: comparisons between human, rat and mouse hepatocytes. Carcinogenesis 17, 89-94. Potter, G. A., McCague, R., and Jarman, M. (1994) A mechanistic hypothesis for DNA adduct formation by tamoxifen following hepatic oxidative metabolism. Carcinogenesis 15, 439-442. Phillips, D. H., Potter, G. A., Horton, M. N., Hewer, A., CroftonSleigh, C., Jarman, M., and Venitt, S. (1994) Reduced genotoxicity of [D5-ethyl]tamoxifen implicates R-hydroxylation of the ethyl group as a major pathway of tamoxifen activation to a liver carcinogen. Carcinogenesis 15, 1487-1492.

Tamoxifen-DNA Adducts in Human Endometrium (42) Phillips, D. H., Carmichael, P. L., Hewer, A., Cole, K. J., and Poon, G. K. (1994) R-Hydroxytamoxifen, a metabolite of tamoxifen with exceptionally high DNA-binding activity in rat hepatocytes. Cancer Res. 54, 5518-5522. (43) Jarman, M., Poon, G. K., Rowlands, G., Grimshaw, R. M., Horton, M. N., Potter, G. A., and McCague, R. (1995) The deuterium isotope effect for the R-hydroxylation of tamoxifen by rat liver microsomes accounts for the reduced genotoxicity of [D5-ethyl]tamoxifen. Carcinogenesis 16, 683-688. (44) Sanchez, C., Shibutani, S., Dasaradhi, L., Bolton, J. L., Fan, P. F., and McClelland, R. A. (1998) Lifetime and reactivity of the ultimate tamoxifen carcinogen: the tamoxifen carbocation. J. Am. Chem. Soc. 120, 13513-13514. (45) Martin, E. A., Heydon, R. T., Brown, K., Brown, J. E. B., Lim, C. K., White, I. N. H., and Smith, L. L. (1998) Evaluation of tamoxifen and R-hydroxytamoxifen 32P-post-labelled DNA adducts by the development of a novel automated on-line solid-phase extraction HPLC method. Carcinogenesis 19, 1061-1069. (46) Ramakrishna, K. V., Fan, P. W., Boyer, C. S., Dalvie, D., and Bolton, J. L. (1997) Oxo substituents markedly alter the phase II metabolism of R-hydroxybutenylbenzenes: models probing the bioactivation mechanism of tamoxifen. Chem. Res. Toxicol. 10, 887-894. (47) Mohrenweiser, H. W., and Jones, I. M. (1998) Variation in DNA repair is a factor in cancer susceptibility: a pradigm for the promise and perils of individual and population risk estimation? Mutat. Res. 400, 15-24. (48) Ottender, M., and Lutz, S. K. (1999) Correlation of DNA adduct levels with tumor incidence: Carcinogenic potency of DNA adducts. Mutat. Res. 424, 237-247. (49) Wogan, G. N. (1992) Molecular epidemiology in cancer risk assessment and prevention: recent progress and avenues for future research. Environ. Health Perspect. 98, 167-178. (50) Kato, S., Petruzzelli, S., Bowman, E. D., Turteltaub, K. W., Blomeke, B., Weston, A., and Shields, P. G. (1993) 7-Alkyldeoxyguanosine adduct detection by two-step HPLC and the 32Ppostlabeling assay. Carcinogenesis 14, 545-550. (51) Blomeke, B., Greenblatt, M. J., Doan, V. D., Bowman, E. D., Murphy, S. E., Chen, C. C., Kato, S., and Shields, P. G. (1996) Distribution of 7-alkyl-2′-deoxyguanosine adduct levels in human lung. Carcinogenesis 17, 741-748.

Chem. Res. Toxicol., Vol. 12, No. 7, 1999 653 (52) Lodovici, M., Akpan, V., Giovannini, L., Migliani, F., and Dolara, P. (1998) Benzo[a]pyrene diol-epoxide DNA adducts and levels of polycyclic aromatic hydrocarbons in autopsy samples from human lungs. Chem.-Biol. Interact. 116, 199-212. (53) Shibutani, S., Suzuki, N., and Grollman, A. P. (1998) Mutagenic specificity of (acetylamino)fluorene-derived DNA adducts in mammalian cells. Biochemistry 37, 12034-12041. (54) Parker, M. (1998) Structure and function of estrogen receptors. Vitam. and Horm. (San Diego) 51, 267-287. (55) Jordan, V. C. (1998) Antiestrogen action of raloxifene and tamoxifen: today and tomorrow. J. Natl. Cancer Inst. 90, 967971. (56) Jordan, V. C., Glusman, J. E., Eckert, S., Lippman, M., Powles, T., Costa, A., Morrow, M., and Norton, L. (1998) Raloxifene reduces incident primary breast cancers: integrated data from multicenter double-blind, placebo-controlled, randomized trials in postmenopausal women. Breast Cancer Res. Treat. 50, 227. (57) Paech, K., Webb, P., Kuiper, G. G. J. M., Nilsson, S., Gustafsson, J.-A., Kushner, P. J., and Scanlan, T. S. (1997) Differential ligand activation of estrogen receptors ERR and ERβ at AP1 sites. Science 277, 1508-1510. (58) Maenpaa, J. U., and Ala-Fossi, S. L. (1997) Toremifene in postmenopausal breast cancer. Efficacy, safety and cost. Drugs Aging 11, 261-270. (59) Hard, G. C., Iatropoulos, M. J., Jordan, K., Radi, L., Kaltenberg, O. P., Imondi, A. R., and Williams, G. M. (1993) Major difference in the hepatocarcinogenicity and DNA adduct forming ability between toremifene and tamoxifen in female Crl:CD(BR) rats. Cancer Res. 53, 4534-4541. (60) Montandon, F., and Williams, G. M. (1994) Comparison of DNA reactivity of the polyphenylethylene hormonal agents diethylstibestrol, tamoxifen and toremifene in rat and hamster liver. Arch. Toxicol. 68, 272-275. (61) Rajaniemi, H., Ma¨ntyla, E., and Hemminki, K. (1998) DNA adduct formation by tamoxifen and structurally-related compounds in rat liver. Chem.-Biol. Interact. 113, 145-159.

TX990033W