Identification of Tamoxifen− DNA Adducts Formed by α-Sulfate

Chem. Res. Toxicol. 1997, 10, 189-196

189

Identification of Tamoxifen-DNA Adducts Formed by r-Sulfate Tamoxifen and r-Acetoxytamoxifen Lakkaraju Dasaradhi and Shinya Shibutani* Department of Pharmacological Sciences, State University of New York at Stony Brook, Stony Brook, New York 11794-8651 Received July 8, 1996X

R-Sulfate trans-tamoxifen and R-sulfate cis-tamoxifen were synthesized as proposed active metabolites of tamoxifen that react with DNA. R-Acetoxytamoxifen was prepared as a modelactivated form to produce a reactive carbocation. Calf thymus DNA was reacted with R-hydroxytamoxifen or the activated forms of tamoxifen, and tamoxifen-DNA adducts were analyzed by a 32P-postlabeling method. The reactivity of R-sulfate trans-tamoxifen to DNA was much higher than that of R-hydroxytamoxifen. The formation of tamoxifen-DNA adducts induced by R-acetoxytamoxifen and R-sulfate cis-tamoxifen was 1100- and 1600-fold, respectively, higher than that of R-hydroxytamoxifen. Both R-sulfate tamoxifens and R-acetoxytamoxifen were highly reactive to 2′-deoxyguanosine. Four reaction products of dG-tamoxifen were isolated by HPLC and characterized by mass- and proton magnetic resonance spectroscopy. Fractions 1 and 2 that eluted first were identified as the epimers of trans form of dG-N2tamoxifen. Fractions 3 and 4 were identified as the epimers of cis form of dG-N2-tamoxifen. When DNA was reacted with R-acetoxytamoxifen in vitro, three isomers of dG-N2-tamoxifen were detected: fraction 2 was the major adduct while fractions 1 and 3 were minor adducts.

Introduction Tamoxifen [(Z)-1-[4-[2-(dimethylamino)ethoxy]phenyl]1,2-diphenyl-1-butene, Figure 1], an antiestrogen, is widely used in the treatment of breast cancer (1). This drug is being considered as a prophylactic agent for women with a strong family history of breast cancer (2, 3). However, an increased incidence of endometrial cancer has been reported in breast cancer patients treated with tamoxifen (4-6). Tamoxifen has also been shown to be a potent hepatocarcinogen in rats (7-9). A high frequency of p53 mutations were found in hepatocarcinomas induced by tamoxifen treatment (10). Most carcinogens are metabolically activated in cells and bind covalently to DNA (11). DNA damage is an initiating event in human cancer (11). Tamoxifen has been known to be activated in liver microsomes of rats and humans (12-14) and reported to give rise to DNA adducts in the livers of rodents (15, 16). These adducts, if unrepaired (17), may generate mutations that accumulate in genomic DNA and facilitate development of human cancers. R-Hydroxytamoxifen (Figure 1) was identified as a metabolite in plasma of breast cancer patients (18) and in livers of rats (19) and humans (18) undergoing treatment with tamoxifen. However, this metabolite has only a low level of reactivity to DNA in vitro (19). We expected that allylic esters of R-hydroxytamoxifen are highly reactive, forming DNA adducts, as was also observed for benzyl esters of carcinogens such as safrole (20), estragole (21), and 7,12-dimethylbenz[a]anthracene (22, 23). In addition, since the formation of tamoxifen-DNA adducts was inhibited by sulfotransferase inhibitors (24), the R-hydroxy group of tamoxifen may be sulfonated to react with DNA (24, 25). * Author to whom correspondence should be addressed. Telephone: 516-444-8018. Telefax: 516-444-3218. E-mail: Shinya@ pharm.som.sunysb.edu. X Abstract published in Advance ACS Abstracts, January 1, 1997.

S0893-228x(96)00114-2 CCC: $14.00

Figure 1. Structures of tamoxifen-related compounds.

Osborne et al. (25) isolated by HPLC two reaction products from the hydrolysate of DNA treated with R-acetoxytamoxifen: the major product was identified as a trans form of dG-N2-tamoxifen1 while the minor product was not completely characterized. In this paper, R-sulfate tamoxifens and R-acetoxytamoxifen (Figure 1) were synthesized for investigating the reactivities to DNA. We found that R-sulfate tamoxifens and R-acetoxytamoxifen react with dG residue to form four reaction products. These products were identified as diastereoisomers of dG-N2-tamoxifen.

Experimental Procedures Caution. R-Sulfate tamoxifens and R-acetoxytamoxifen are alkylating agents and may be genotoxins. Materials. Organic chemicals used for the chemical synthesis were supplied by Aldrich Chemical (Milwaukee, WI) and 1 Abbreviations used: dN, 2′-deoxynucleoside; dG, 2′-deoxyguanosine; PEI, polyethylenimine; tR, retention time.

© 1997 American Chemical Society

190 Chem. Res. Toxicol., Vol. 10, No. 2, 1997 Fisher Scientific (Pittsburgh, PA). [γ-32P]ATP (specific activity, 3000 Ci/mmol) was obtained from Amesham Corp (Arlington Heights, IL). Polyethylenimine (PEI)-cellulose sheet was purchased from J. T. Baker Inc. (Phillipsburg, NJ). Calf thymus DNA, micrococal nuclease, deoxyribonuclease I, venom phosphodiesterase I (type VII), alkaline phosphatase (type III), and potato apyrase were purchased from Sigma (St. Louis, MO). Spleen phosphodiesterase was obtained from Boehringer Mannheim Corp (Indianapolis, IN). T4 polynucleotide kinase was purchased from Stratagene (La Jolla, CA). HPLC analysis was performed on a Waters 990 HPLC instrument, equipped with a photodiode array detector. Mass spectrometry was performed on a Kratos MS 890, equipped with a Hewlett Packard 5980 detector. NMR spectroscopy was performed on a Bruker 600 MHz NMR instrument. Preparation of r-Sulfate trans- or cis-Tamoxifen and r-Acetoxytamoxifen. A trans or cis form of R-hydroxytamoxifen was synthesized by the established protocol (26). R-Hydroxytamoxifen (65 mg, 0.16 mmol) and SO3‚pyridine (135 mg, 0.83 mmol) were dissolved in 2 mL of dry pyridine and stirred under N2 for 1 h. After 20 mL of ether was added to the reaction mixture, the precipitate was filtered and washed twice with 10 mL of ether. The precipitate was suspended in 3 mL of ethanol, and the insoluble materials were filtered off. R-Sulfate cistamoxifen was obtained as a white crystalline solid with a 40% yield, while R-sulfate trans-tamoxifen was soluble in ethanol and precipitated by adding ether. The precipitate was washed twice with ether and dried under reduced pressure. Since R-sulfate trans-tamoxifen was highly hygroscopic and hydrolyzed when exposed to air, the NMR showed the presence of R-sulfonic acid tamoxifen and R-hydroxytamoxifen. Pure R-sulfate trans-tamoxifen was synthesized by maintaining anhydrous conditions. Pyridinium r-sulfate trans-tamoxifen: IR (KBr) 1256 cm-1, 1H NMR (DMSO) δ 1.1 (3H, d, J ) 5.1 Hz, CH3CH), 2.8 (6H, s, -NMe2), 3.4 (2H, br, -CH2NMe2), 4.1(2H, br, -OCH2CH2), 5.1 (1H, unresolved q, CHOSO3‚Py), 6.6 (2H, d, J ) 8.6, H-3 and 5 of -CC6H4O), 6.8 (2H, d, J ) 8.6, H-2 and 4 of CC6H4O), 7.1- 7.4 (10H, m, phenyl Hs), 8.0-8.7, (5H, m, pyridinium Hs); 13C NMR (DMSO) δ 17.73, 42.86, 55.43, 61.92, 94.69, 113.74, 115.28, 126.79, 127.73, 127.99, 128.12, 128.21, 129.15, 129.23, 129.63, 130.23, 130.85, 131.56, 136.36, 144.50, 146.08, and 156.10. Pyridinium r-sulfate cis-tamoxifen: mp >125 °C dec; IR (KBr) 1256 cm-1; 1H NMR(DMSO) δ 1.2 (3H, d, J ) 6.9 Hz, CH3CH), 2.78 (6H, s, NMe2), 3.53 (2H, br, -CH2NMe2), 4.33 (2H, br, -OCH2CH2), 5.3 (1H, br, CHOSO3‚Py), 6.8-7.3 (14H, m, phenoxy and phenyl Hs), 8.0-8.7, (5H, m, pyridinium Hs); 13C NMR(DMSO) δ 20.00, 42.96, 55.03, 72.02, 114.01, 114.18, 126.59, 126.61, 126.83, 127.25, 129.67, 130.57, 130.73, 131.58, 135.40, 137.24, 138.11, 139.21, 139.56, 140.12, 142.12, and 153.06. Using negative ion FAB-mass spectroscopy, both pyridinium sulfates exhibited at m/z 466 representing the molecular weight (545) minus the pyridine moiety. R-Acetoxytamoxifen was prepared, using a similar protocol previously described by Osborne et al. (25). Digestion of DNA Sample. Aliquots of DNA (4 µg) were incubated with 50 µg of R-hydroxy-trans-tamoxifen, R-sulfate trans- or cis-tamoxifen, or R-acetoxy-trans-tamoxifen in 100 µL of 100 mM Tris-HCl buffer (pH 8.0) at 37 °C for 16 h. The samples were evaporated to dryness, 800 µL of ethanol was added, and mixture was centrifuged to recover the DNA. Each 1.0 µg of the DNA samples was enzymatically digested at 37 °C for 6 h in 10 µL of 17 mM sodium succinate buffer (pH 6.0) containing 8 mM CaCl2, using 0.3 units of micrococcal nuclease and 2.0 units of spleen phosphodiesterase. The samples were dissolved in 100 µL of distilled water and extracted twice with 100 µL of butanol. The butanol fractions were dried and used for analysis of tamoxifen-DNA adducts. Detection of DNA Adducts by 32P-Postlabeling Method. The digests of pooled extracts were incubated with 40 µCi of [γ-32P]ATP (3000 Ci/mmol) and 3 µL of T4 polynucleotide kinase (10 units/µL) (27). The 32P-labeled sample was spotted on a 10 × 10 cm polyethylenimine (PEI)-cellulose thin-layer plate at a location 2 cm from the bottom and 2 cm from the left edge of

Dasaradhi and Shibutani the plate and developed using four different solvents (12, 16). Paper wicks were attached to the top of PEI-cellulose plates except for the second development. The plates were washed twice with distilled water for 30 min at room temperature and air dried. The plates were developed for 16 h in 1.7 M sodium phosphate buffer (pH 6.0) with a paper wick, and subsequently developed in the same direction to the top of the plate with 2.625 M lithium formate and 6.375 M urea (pH 3.5). With a paper wick at the right edge, the plates were developed to 2 cm onto the paper wick at a right angle to the previous direction of development in 0.6 M LiCl, 0.375 M Tris-HCl, and 6.375 M urea (pH 8.0). The plates were subsequently developed in the same direction with 1.7 M sodium phosphate buffer (pH 6.0) to approximately 5 cm onto the paper wick. The position of adducts was established by autoradiography, using Kodak Xomat XAR film. Calculation of Relative Adduct Levels. The radioactive spots on the PEI-cellulose plates were scraped and the radioactivities were measured by scintillation counting. Alternatively, a β phosphoimager (Molecular Dynamics) was used for the measurement. The relative adduct levels were calculated according to Levay et al. (28), using dpm instead of cpm: (total dpm in adducts)/2.02 × 1010 dpm, assuming that 1 µg of DNA was 3.03 × 103 pmol of 3′dNp and the specific activity of the [γ-32P]ATP was 6.66 × 106 dpm/pmol. The specific activity of the [γ-32P]ATP was corrected by calculating the extent of decay. Isolation of dN-Tamoxifens. 2′-Deoxynucleoside (dN, 1.0 mg) was reacted at 37 °C for 4 h with 2.0 mg of R-sulfate tamoxifen or R-acetoxytamoxifen in 500 µL of 100 mM Tris-HCl buffer (pH 8.0). The reaction mixture was centrifuged and the supernatant was extracted twice with 500 µL of butanol. To isolate tamoxifen-modified dNs, the pooled butanol extracts were subjected to a reverse-phase column, µBondapak C18 (0.78 × 30 cm, Waters), eluted over 20 min at a flow rate of 2.0 mL/min with a linear gradient of 0.05 M triethylammonium acetate, pH 7.0, containing 10-50% acetonitrile (29). Using a 100-fold scale of the reaction conditions described above, large amounts of dG-reaction products were collected by HPLC, repurified by the same HPLC system, and characterized by positive ion FAB-mass spectroscopy and Bruker 600 MHz proton magnetic resonance spectroscopy. 1H NMR data of dGreaction products are described in Table 1. Digestion of Tamoxifen-Modified DNA. Calf thymus DNA (400 µg) was incubated at 37 °C for 4 h with 5 mg of R-acetoxytamoxifen in 2.0 mL of 100 mM Tris-HCl buffer (pH 8.0). The sample was evaporated to dryness in a vacuum desiccator, and extracted three times with 1.0 mL of methanol. The recovered DNA (250 µg) was digested at 37 °C for 16 h with 20 units of deoxyribonuclease I and 0.02 units of venom phosphodiestrase I in 500 µL of 50 mM Tris-HCl buffer (pH 7.5), and further incubated at 37 °C for 4 h with 3.0 unit of alkaline phosphatase. The reaction sample was evaporated to dryness in a vacuum desiccator, and extracted three times with 350 µL of methanol. Methanol extracts were evaporated to dryness, and the products were dissolved in 100 µL of distilled water/ ethanol (9:1) and analyzed by HPLC, using a Waters reversephase µBondapak C18 column, as described above.

Results Detection of Tamoxifen-DNA Adducts. The reactivities of R-hydroxytamoxifen, R-sulfate trans- or R-cistamoxifen, and R-acetoxytamoxifen to DNA were examined in vitro. Since >98% of tamoxifen adducts were recovered from the hydrolysate of DNA by butanol extraction (data not shown), tamoxifen-DNA adducts were analyzed by using 32P-postlabeling method combined with butanol extraction. When the DNA was incubated only in the buffer, no DNA adducts were detected (Figure 2A). With R-hydroxytamoxifen, a small amount (5.6 adducts/108 normal nucleotides) of tamoxifen adduct was detected as indicated with an arrow (a)

dG-N2-Tamoxifen DNA Adducts

Chem. Res. Toxicol., Vol. 10, No. 2, 1997 191

Table 1. Proton NMR Spectra of dG-Tamoxifen O 1

N

HN

8

2

HN α 5

α′

CH3

N

(H3C)2N

O

β′ 1 α′

dR

O

β

2

8

2

HN α

5

3

N

HN 3

6

trans

6

(H3C)2N

N

N CH3 β

N dR

cis

2

O

β′

fr-1,2

fr-3,4

δ(ppm)a fr-1,2

fr-3,4

β N(CH3)2 β′ R′ R H3 and 5 H2 and 6 Phenyl

Tamoxifen Fragment 1.26, d,d J ) 6.96 2.12, s 2.48, ndb 3.86, t 4.85, mc 6.55, d, J ) 8.8 6.78, d, J ) 8.8 7.0-7.5, m

1.27, d, J ) 6.96 2.21, s 2.60, t 3.88, t 4.90, mc 7.07, d, J ) 8.4 7.36, d, J ) 8.4 6.9-7.4, m

N2H C8 N1

Guanine Fragment 6.04, d, J ) 7.7 7.99, s 8.90, s

5.97, d, J ) 7.7 7.99, s 9.91, s

2′ 5′ 4′ 3′ 5′OH 3′OH 1′

Sugar Fragment 2.2-2.8, m 3.54, q 3.90, m 4.38, m 4.94, mc 5.34, br 6.19, t

2.4-2.8, m 3.55, q 3.90, m 4.38, m 4.90, mc 5.37, br 6.29, t

a δ, chemical shift from tetramethylsilane. b nd, not detected as it was submerged under DMSO peak. c R and 5′OH were overlapping. d Abbreviations: s, singlet; d, doublet; t, triplet; q, quartet; m, multiplet; br, broad.

(Figure 2B). In contrast, with R-sulfate trans-tamoxifen, several DNA adducts were detected (Figure 2C); the major spot (a) was estimated as 1.0 adduct/105 normal nucleotides. This amount was 180 times higher than that of R-hydroxytamoxifen. R-Sulfate trans-tamoxifen was unstable and rapidly hydrolyzed on exposure to air. If no hydrolysis of R-sulfate trans-tamoxifen has occurred, much higher amounts of DNA adducts would be expected to be obtained. R-Sulfate cis-tamoxifen also produced several DNA adducts (Figure 2D). The major spot (a) was 8.7 adducts/105 normal nucleosides; this amount was 1600-fold higher than that of R-hydroxytamoxifen. With an equal amount of R-acetoxytamoxifen, the same DNA adducts were detected in a similar distribution as observed by R-sulfate tamoxifen (data not shown). The major spot (a) was 6.2 adducts/105 normal nucleosides; this amount was 1100-fold higher than that R-hydroxytamoxifen, but 1.4 times lower than that of R-sulfate cistamoxifen. Isolation of dG-Tamoxifens. Using R-sulfate transor R-cis-tamoxifen or R-acetoxytamoxifen, the reactivities to each deoxynucleosides were examined in vitro. When dG was reacted with R-acetoxytamoxifen, four reaction products were isolated by HPLC (Figure 3A). The retention time of each of the products was as follows: fr1, 32.0 min; fr-2, 32.8 min; fr-3, 36.8 min; fr-4, 38.1 min. On the basis of the amount of unreacted dG, approximately 10.1% of dG was modified when R-acetoxytamoxifen was used. The ratio of fr-1:fr-2:fr-3:fr-4 was 1:1.7: 1.1:0.7. All products have similar UV absorbance (Figure

Figure 2. 32P-postlabeling analysis of DNA reacted with tamoxifen-derivatives. Four micrograms of calf thymus DNA were incubated for 16 h at 37 °C in 100 mM Tris-HCl buffer (pH 8.0), without tamoxifen-derivatives (A), with 50 µg of R-hydroxytamoxifen (B), with 50 µg of the fraction containing R-sulfate trans-tamoxifen (C) or with 50 µg of R-sulfate cistamoxifen (D). One microgram of the recovered DNA was digested enzymatically, extracted with butanol, and labeled with 32P. One-third of the sample was applied to PEI-cellulose thinlayer chromatography and developed using four different buffer solutions, as described under Experimental Procedures. A and B were exposed for 16 h, and C and D were exposed for 1.5 h to X-ray film.

4); the UV maximum (251 nm) of fr-1 and fr-2 was slightly different from that (249 nm) of fr-3 and fr-4. With dA, small amounts of products were detected by HPLC (data not shown), but were insufficient for the determination of UV spectra. With dC or dT, no tamoxifenmodified products were isolated by HPLC (data not shown). Since R-sulfate trans-tamoxifen was unstable and decomposed before reacting with dG, the yield of dGtamoxifens was four times lower than that of R-acetoxytamoxifen (data not shown). However, when R-sulfate cis-tamoxifen was used, the yield of four dG-tamoxifens was 2.3 times higher than that of R-acetoxytamoxifen; the ratio of fr-1:fr-2:fr-3:fr-4 was 1:1.3:1.8:0.8 (Figure 5). Characterization of dG-Tamoxifens. Large amounts of four dG-tamoxifen products were prepared using R-acetoxytamoxifen or R-sulfate cis-tamoxifen and isolated by HPLC. With the use of positive ion FAB-mass spectroscopy, the parent ions of all four products exhibited at m/z 637 (Figure 6), identifying the molecular weight as 636 Da. This molecular ion gave daughter ions at m/z 521 representing 637 minus the deoxyribose moiety, at m/z 370 representing the tamoxifen moiety minus 1, and at m/z 299 representing the tamoxifen moiety minus CH2CH2N(CH3)2 (Figure 6). The peaks of m/z 185, 215, and 279 appeared to be the glycerin matrix used for the measurement of mass spectroscopy. These data suggest that four products are expected to be diastereoisomers of dG-tamoxifen. The structure of each of the isomers of dG-tamoxifens was analyzed by 600 Hz NMR. The spectra were

192 Chem. Res. Toxicol., Vol. 10, No. 2, 1997

Dasaradhi and Shibutani

Figure 3. HPLC separation of dG-tamoxifens and a hydrolysate of DNA reacted with R-acetoxytamoxifen. A: One milligram of dG was reacted with 2.0 mg of R-acetoxytamoxifen in 100 mM Tris-HCl buffer (pH 8.0) for 4 h at 37 °C, and then extracted with butanol, as described under Experimental Procedures. One-tenth of the butanol fraction was dried and subjected to HPLC for the isolation of dG-tamoxifens. A linear gradient of 0.05 M triethylammonium acetate (pH 7.0) containing 10-50% acetonitrile was eluted over 20 min at a flow rate of 2.0 mL/min. B: R-Acetoxytamoxifen-treated DNA (200 µg) was digested with deoxyribonuclease I, venom phosphodiesterase I, and alkaline phosphatase, and extracted with ethanol, as described under Experimental Procedures. Half of the sample was subjected to HPLC. C: A mixture of A and B was subjected to HPLC.

Figure 4. UV spectra of four dG-tamoxifens. During the isolation of dG-tamoxifens by HPLC, UV spectra 220-400 nm were monitored every 2 s using a multiple photodiode array detector (Waters).

recorded in DMSO and the assignment of the peaks was determined by comparing with NMR data of dG and

R-hydroxy trans- or cis-tamoxifens (26) and by D2O exchange for analyzing the exchangable NH and OH protons (Table 1). The two doublet signals in both fr-1 and fr-2 at δ 6.55 and 6.78 corresponding to protons 3,5 and 2,6 of the phenoxy ring indicated that tamoxifen is trans form. In contrast, the two doublet signals in both fr-3 and fr-4 were shifted to δ 7.07 and 7.36 (Figure 7). This confirms that tamoxifen is cis isomer (26). The doublet of the monosubstituted amino group at δ 6.04 for fr-1 and fr-2 and δ 5.97 for fr-3 and fr-4 was coupled to the methine moiety at δ 4.85 or 4.90 proton, and the shift in the signals for the protons on 1 and 8 position of guanine (Table 1 and Figure 7) indicates that the tamoxifen moiety is linked to N2 position of guanine. Thus, fr-1 and fr-2 were identified as the epimers of trans-form of dG-N2-tamoxifen. Fr-3 and fr-4 were identified as the epimers of cis-form of dG-N2-tamoxifen. Analysis of Tamoxifen Adducts in r-Acetoxytamoxifen-Treated DNA. Calf thymus DNA was reacted with R-acetoxytamoxifen and digested enzymatically. The hydrolysate was analyzed by HPLC, as shown in Figure 3B. The four peaks that eluted first represented dC, dG, dT, and dA, respectively. One major reaction product (tR ) 31.9 min) and two minor products (tR ) 32.4 and tR ) 36.5) were detected (Figure 3B). This hydrolysate was

dG-N2-Tamoxifen DNA Adducts

Chem. Res. Toxicol., Vol. 10, No. 2, 1997 193

amount of fr-2 was increased 2-fold and the amounts of fr-1 and fr-3 were increased 11% and 18%, respectively. However, the amount of fr-4 was not increased. Thus, the major reaction product was fr-2 representing a trans form of dG-N2-tamoxifen. The other two minor products were fr-1 and fr-3 representing a trans form and a cis form of dG-N2-tamoxifen, respectively.

Discussion

Figure 5. HPLC separation of dG-tamoxifens formed by R-sulfate cis-tamoxifen. One milligram of dG was reacted with 2.0 mg of R-sulfate cis-tamoxifen in 100 mM Tris-HCl buffer (pH 8.0) for 4 h at 37 °C, and then extracted with butanol, as described under Experimental Procedures. One-tenth of the butanol fraction was dried and subjected to HPLC for the isolation of dG-tamoxifens.

cochromatographed with four stereoisomers of dG-N2tamoxifen (Figure 3A). As shown in Figure 3C, the

Several mechanisms have been proposed to explain the formation of DNA adducts induced by tamoxifen (30). Oxidative species such as 4-hydroxytamoxifen quinone methide may promote the reaction with DNA (31). Tamoxifen 1,2-epoxide, another oxidative species, was initially reported to be reacted with DNA (32). However, the same group recently concluded that pure tamoxifen 1,2-epoxide itself is devoid of DNA-binding activity (25). Therefore, it remains unclear whether a contaminant in the epoxide sample caused the formation of DNA adducts. On the other hand, R-hydroxylation of tamoxifen and its metabolites, tamoxifen N-oxide, N-desmethyltamoxifen, and 4-hydroxytamoxifen, respectively, are also likely candidates, capable of forming DNA adducts with nucleosides (18, 19, 30, 33). Using 32P-postlabeling analysis, we confirmed (Figure 2B) that the reactivity of R-hydroxytamoxifen to DNA in vitro was poor (19). In contrast, R-sulfate trans-tamoxifen was highly reactive to DNA (Figure 2C). R-Sulfate cis-tamoxifen promoted 1600-fold higher amounts of DNA adducts than R-hydroxytamoxifen (Figure 2D); this reactivity was much higher than that of R-acetoxytamoxifen. Thus, the sulfonation of the R-hydroxy moiety of

Figure 6. Mass-spectrum of dG-tamoxifen. Fr-3 shown in Figure 3A was subjected to positive ion FAB mass spectroscopy.

194 Chem. Res. Toxicol., Vol. 10, No. 2, 1997

Dasaradhi and Shibutani

Figure 7. NMR spectra of R-trans- and R-cis-form of dG-N2-tamoxifens.

Figure 8. A proposed mechanism for the formation of tamoxifen-DNA adducts.

Figure 9. A proposed mechanism of conversion between R-trans- and R-cis-form of tamoxifen.

tamoxifen may be involved in the formation of the tamoxifen-DNA adducts in vivo, as shown in Figure 8. During this reaction, a reactive allylic carbocation may be generated as an intermediate to react with DNA (30) (Figure 8). R-Sulfate trans-tamoxifen rapidly hydrolyzed in aqueous solution and on exposure to the atmosphere. Even when R-sulfate trans-tamoxifen was stored in dimethyl

sulfoxide at -20 °C, the half-life was approximately 1.5 days. In contrast, R-sulfate cis-tamoxifen and R-acetoxytamoxifen were much more stable and can be stored at -20 °C for at least 5 months (data not shown). The low stability of R-sulfate trans-tamoxifen may be due to ipsoelectronic effect from the substituent on the phenyl ring (30) and its low reactivity with DNA may be due to the rapid hydrolysis to R-hydroxytamoxifen resulting from

dG-N2-Tamoxifen DNA Adducts

exposure to the atmosphere during the experiment. Three reactive forms of tamoxifen react particularly with dG to produce four diastereoisomers of dG-N2tamoxifen (Figure 3A and 5). A reactive allylic carbocation of the trans and cis form may be converted to that of the cis and trans form respectively through an intermediate, as shown in Figure 9. When R-sulfate transtamoxifen or R-acetoxy-trans-tamoxifen was reacted with dG, the ratio of trans forms:cis forms of dG-N2-tamoxifen was 2.7:1.8 (Figure 3A). The amount of the trans forms was 60% higher than that of the cis forms. In contrast, with R-sulfate cis-tamoxifen, the ratio of trans forms:cis forms was 2.3:2.6 (Figure 5). The amount of the cis forms was slightly higher than that of the trans forms. Thus, in the equilibrium between trans and cis forms, conversion to the trans form appears to be preferred slightly over conversion to the cis form (Figure 9). Osborne et al. (25) isolated two tamoxifen-nucleoside adducts from a hydrolysate of DNA reacted with R-acetoxytamoxifen: the major product was identified as a trans form of dG-N2-tamoxifen, but the minor product has not been completely characterized. With similar experimental conditions, we detected three dG-N2-tamoxifen adducts. The major tamoxifen adduct was identified as fr-2, a trans form of dG-N2-tamoxifen, and two minor adducts were identified as fr-1, a trans form and fr-3, a cis form (Figure 3B). Thus, the major product isolated by Osborne et al. (25) may be a mixture of fr-1 and fr-2; the minor product may be fr-3. The yield of each diastereoisomers of dG-N2-tamoxifen formed in DNA differ from that observed in dG. The conformation of double-stranded DNA may influence the reactivity of the allylic carbocation in each tamoxifen stereoisomer.

Acknowledgment. We thank Mr. Robert Rieger for the measurement by mass spectroscopy. We also thank Dr. F. Johnson for his interest in this program. We are indebted to Dr. A. P. Grollman for his encouragement and for critically reviewing the manuscript. This study was supported by the Emil C. Voll Award, the School of Medicine, SUNY at Stony Brook (to S.S.) and in part by Research Grants CA17395 (to A.P.G.) from the National Institutes of Health.

References (1) Jordan, V. C. (1993) A current view of tamoxifen for the treatment and prevention of breast cancer. Br. J. Pharmacol. 110, 507517. (2) Powles, T. J., Hardy, J. R., Ashley, S. E., Farrington, G. H., Cosgrove, D., Davey, J. R., Dowsett, M., McKinna, J. A., Wash, A. G., Sennet, H. D., Tillyer, C. R., and Treleaven, J. G. (1993) A pilot trial to evaluate the acute toxicology and feasability of tamoxifen for the prevention of breast cancer. Br. J. Cancer 60, 126-131. (3) Nayfield, S. G., Karp, J. E., Ford, L. G., Dorr, F. A., and Kramer, B. S. (1991) Potential role of tamoxifen in prevention of breast cancer. J. Natl. Cancer Inst. 83, 1450-1459. (4) Seoud, M. A.-F., Johnson, J., and Weed, J. C. (1993) Gynecologic tumors in tamoxifen-treated women with breast cancer. Obstet. Gynecol. 82, 165-169. (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) van Leeuwen, F. E., Benraadt, J., Coebergh, J. W. W., Kiemeney, L. A. L. M., Diepenhorst, F. W., van den Belt-Dusebout, A. W., and van Tinteren, H. (1994) Risk of endometrial cancer after tamoxifen treatment of breast cancer. Lancet 343, 448-452. (7) 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.

Chem. Res. Toxicol., Vol. 10, No. 2, 1997 195 (8) Greaves, P., Goonetilleke, R., Nunn, G., Topham, J., and Orton, T. (1993) Two-year carcinogenicity study of tamoxifen in Alderley Park Wister-derived rats. Cancer Res. 53, 3919-3924. (9) 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. (10) Vancutsem, P. M., Lazarus, P., and Williams, G. M. (1994) Frequent and specific mutations of the rat p53 gene in hepatocarcinomas induced by tamoxifen. Cancer Res. 54, 3864-3867. (11) Lawley, P. D. (1994) From fluorescence spectra to mutational spectra, a historical overview of DNA-reactive compounds. In DNA Adducts: Carcinogen & Mutagenic Agents: Chemistry, Identification & Biological Significance (Hemminki, K., et al., Eds.) Publication 125, pp 3-22, IARC Scientific Lyon, France. (12) 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. (13) Pathak, D. N., Pongracz, K., and Bodell, W. J. (1995) Microsomal and peroxidase activation of 4-hydroxy-tamoxifen to form DNA adducts: comparison with DNA adducts formed in SpragueDawley rats treated with tamoxifen. Carcinogenesis 16, 11-15. (14) Hemminki, K., Widlak, P., and Hou, S. M. (1995) DNA adducts caused by tamoxifen and toremifene in human microsomal system and lymphocyte in vitro. Carcinogenesis 16, 1661-1664. (15) Han, X., and Liehr, J. G. (1992) Induction of covalent DNA adducts in rodents by tamoxifen. Cancer Res. 52, 1360-1363. (16) White, I. N. H., de Matteis, F., Davies, A., Smith, L. L., CroftonSleigh, C., Venitt, S., Hewer, A., and Phillips, D. H. (1992) 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. (17) Carthew, P., Martin, E. A., White, I. N. H., DeMatteis, F., Edwards, R. E., Dorman, B. M., Heydon, R. T., and Smith, L. L. (1995) Tamoxifen induces short-term cumulative DNA damage and liver tumors in rats: promotion by phenobarbital. Cancer Res., 55, 544-547. (18) Poon, G. K., Walter, B., Lfnning, P. E., Horton, M. N., and McCague, R. (1995) Identification of tamoxifen metabolites in human Hep G2 cell line, human liver homogenate, and in patients on long-term therapy for breast cancer. Drug Metab. Dispos. 23, 377-382. (19) 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. (20) Boberg, E. W., Miller, E. C., Miller, J. A., Poland, A., and Liem, A. (1983) Strong evidence from studies with brachymorphic mice and pentachlorophenol that I′-sulfooxysafrole is the major ultimate electrophilic and carcinogenic metabolite of I′-hydroxysafrole in mouse liver. Cancer Res. 43, 5163-5173. (21) Phillips, D. H., Miller, J. A., Miller, E. C., and Adams, B. (1981) Structures of the DNA adducts formed in mouse liver after administration of the proximate hepatocarcinogen I′-hydroxyestragole. Cancer Res. 41, 176-186. (22) Watabe, T., Ishizuka, T., Isobe, M., and Ozawa, N. (1982) A 7-hydroxy-methyl sulfate ester as an active metabolite of 7,12dimethyl-benz[a]anthracene. Science 215, 403-405. (23) Watabe, T., Hiratsuka, A., Ogura, K., and Endoh, K. (1985) A reactive hydroxymethyl sulfate ester formed regioselectively from the carcinogen, 7,12-dihydromethyl-benz[a]anthracene, by rat liver sulfotransferase. Biochem. Biophys. Res. Commun. 131, 694-699. (24) Randerath, K., Moorthy, B., Mabon, N., and Sriram, P. (1994) Tamoxifen: evidence by 32P-postlabeling and use of metabolic inhibitors for two distinct pathways leading to mouse hepatic DNA adduct formation and identification of 4-hydroxytamoxifen as a proximate metabolite. Carcinogenesis 15, 2087-2094. (25) 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. (26) Foster, A. B., Jarman, M., Leung, O.-T., McCague, R., Leclercq, G., and Devleeschouwer, N. (1985) Hydroxy derivatives of tamoxifen. J. Med. Chem. 28, 1491-1497. (27) Maniatis, T., Fritsch, E. F., and Sambrook, J. (1982) Molecular Cloning. A Laboratory Manual, Cold Spring Harbor Laboratory Press, Cold Spring Harbor, NY. (28) 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. (29) Shibutani, S., Gentles, R., Johnson, F., and Grollman, A. P. (1991) Isolation and characterization of oligodeoxynucleotides containing

196 Chem. Res. Toxicol., Vol. 10, No. 2, 1997 dG-N2-AAF and oxidation products of dG-C8-AF. Carcinogenesis 12, 813-818. (30) Potter, G. A., McCague, R., and Jarman, M. (1994) A mechanism hypothesis for DNA adduct formation by tamoxifen hepatic oxidative metabolism. Carcinogenesis 15, 439-442. (31) Moorthy, B., Sriram, P., Pathak, D. N., 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 vitro. Cancer Res. 56, 53-57.

Dasaradhi and Shibutani (32) 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. (33) 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.

TX960114H