Abstract
Tamoxifen (TAM) is used as the standard endocrine therapy for breast cancer patients and as a chemopreventive agent for women at high risk for this disease. Unfortunately, treatment of TAM increases the incidence of endometrial cancer; this may be due to the genotoxic damage induced by TAM metabolites. Formation of TAM-DNA adducts in rat liver correlates with the development of hepatocarcinoma. TAM-DNA adducts are proposed to be formed through O-sulfonation and/or O-acetylation of α-hydroxylated TAM and its metabolites. However, the role of O-sulfonation and O-acetylation in the formation of TAM-DNA adducts has not been extensively investigated. Rat or human hydroxysteroid sulfotransferases (HST), acetyltransferases, and liver cytosol were incubated with calf thymus DNA, α-OHTAM, and either 3′-phosphoadenosine 5′-phosphosulfate (PAPS) or acetyl coenzyme A (acetyl-CoA) as a cofactor and analyzed for TAM-DNA adduct formation, using 32P postlableling/polyacrylamide gel electrophoresis analysis. TAM-DNA adduct was formed when PAPS, not acetyl-CoA, was used. No TAM-DNA adducts were produced using human N-acetyltransferase I and II. HST antibody inhibited approximately 90% of TAM-DNA adduct formation generated by the cytosol or HST, suggesting that HST is primarily involved in the formation of TAM-DNA adducts. The formation of TAM-DNA adducts with rat liver cytosol and HST was much higher than that of human liver cytosol and HST. Our results indicate that TAM-DNA adducts are formed via O-sulfonation, not O-acetylation, of α-hydroxylated TAM and its metabolites.
Tamoxifen (TAM) has been widely used for breast cancer therapy and also used as a chemopreventive agent for healthy women at high risk for this disease (Fisher et al., 1998; Osborne, 1998). Besides the significant benefit, long-term TAM treatment to women increases the risk of developing endometrial cancer (van Leeuwen et al., 1994; Fisher et al., 1998). TAM was listed in 1996 as a human carcinogen by the International Agency of Research on Cancer (IARC, 1996).
The cellular mechanism underlying TAM-induced carcinogenesis may be due to its partial estrogenic effect through the estrogen receptors and/or genotoxic damage (reviewed by Kim et al., 2004). Actually, a high level of TAM-DNA adducts produced in the liver of rats (Han and Liehr, 1992; Osborne et al., 1996) initiates the development of hepatocellular carcinomas (Hard et al., 1993). TAM-DNA adducts detected in the endometrium of women treated with TAM (Shibutani et al., 2000; Martin et al., 2003) may also be involved in the development of endometrial cancer.
Several phase I and phase II enzymes are involved in the metabolism of TAM. Cytochrome P450 converts TAM to α-hydroxytamoxifen (α-OHTAM), N-desmethyltamoxifen (N-desTAM), and 4-hydroxytamoxifen (4-OHTAM) (Fig. 1) (reviewed by Kim et al., 2004). Tamoxifen N-oxide (TAM N-oxide) is produced from TAM by flavin-containing monooxygenase. α-Hydroxylated forms of TAM and its metabolites are further metabolized by phase II enzymes and react with cellular DNA, resulting in the formation of TAM-DNA adducts (Phillips et al., 1994; Dasaradhi and Shibutani, 1997).
In general, sulfation effectively decreases the toxicity of xenobiotics; however, in some cases, sulfation increases the toxicity. Because certain sulfate conjugates are unstable, they can form potent electrophilic species. Like the sulfate esters, acetoxy esters generated by acetylation can also be highly reactive electrophilic species. In fact, synthetic TAM α-sulfate and α-acetoxyTAM react rapidly with DNA, resulting in the formation of four diastereoisomers [two trans-forms (fr-1 and 2) and two cis-forms (fr-3 and 4)] of α-(N2-deoxyguanosinyl)tamoxifen (dG-N2-TAM) adducts (Fig. 1) (Osborne et al., 1996; Dasaradhi and Shibutani, 1997). Similar adduct formation has been observed using α-sulfate and/or α-acetyl forms of α-OH-N-desTAM (Gamboa da Costa et al., 2000; Kitagawa et al., 2000) and α-OHTAM N-oxide (Umemoto et al., 1999). Therefore, TAM-DNA adducts are considered to be formed through O-sulfonation and/or O-acetylation of α-hydroxylated TAM and its metabolites.
The formation of TAM-DNA adducts was increased in rat hepatocytes incubated with inorganic sulfate and was decreased in the hepatocytes treated with dehydroisoandrosterone-3-sulfate, an inhibitor of hydroxysteroid sulfotransferase (HST) (Davis et al., 1998). The increased formation of TAM-DNA adducts was observed when cells expressing rat STa were exposed to α-OHTAM, but not with cells expressing human HST; the authors concluded that α-OHTAM is a substrate for STa, but not for human SULT 2A1 (Glatt et al., 1998). In contrast, α-OHTAM can be O-sulfonated by both rat STa and human SULT 2A1 and react with DNA, forming dG-N2-TAM adducts (Shibutani et al., 1998a,b). However, the contribution of O-acetylation in the formation of TAM-DNA adducts in rat or human tissues has not yet been determined. We found in the present study that O-sulfonation, not O-acetylation, of α-hydroxylated TAM and its metabolites contributes primarily to the formation of TAM-DNA adducts in rat and human livers.
Materials and Methods
Materials. TAM, calf thymus DNA, micrococcal nuclease, potato apyrase, 3′-phosphoadenosine 5′-phosphosulfate (PAPS), and acetyl coenzyme A (acetyl-CoA) were purchased from Sigma-Aldrich (St. Louis, MO). Spleen phosphodiesterase was obtained from Worthington Biochemical Corp. (Lakewood, NJ). 3′-Phosphatase-free T4 polynucleotide kinase and nuclease P1 were obtained from Roche Applied Science (Indianapolis, IN). Human arylamine N-acetyltransferase (NAT) I (2400 units/mg; enzyme units are expressed as nanomoles of acetic acid ester product formed from para-aminosalicylic acid per minute), NAT II (220 units/mg; enzyme units are expressed as nanomoles of acetic acid ester product formed from sulfamethazine per minute), and human liver cytosol were purchased from Discovery Labware, BD Biosciences (Bedford, MA). α-OHTAM (Shibutani et al., 2001), α-OH-N-desTAM (Kitagawa et al., 2000), and α-OHTAM N-oxide (Umemoto et al., 1999) were prepared previously. TAM α-sulfate and diastereoisomers (fr-1 and fr-2) of trans-forms and diastereoisomers (fr-3 and fr-4) of cis-forms of dG3′-monophosphate-N2-tamoxifen (dG3′p-N2-TAM) were prepared as described previously (Dasaradhi and Shibutani, 1997; Shibutani et al., 1998a). [γ-32P]ATP (specific activity, >6000 Ci/mmol) was obtained from GE Healthcare (Little Chalfont, Buckinghamshire, UK).
Preparation of Cytosol Fraction. Female Fisher 344 rats (8 weeks) were purchased from Taconic (Germantown, NY). The use of animals was in compliance with the guidelines established by the National Institutes of Health Office of Laboratory Animal Welfare. Animals were housed in temperature (22 ± 2°C)- and humidity (55 ± 5%)-controlled rooms with a 12-h light/dark cycle (light, 7:00 AM-7:00 PM; dark, 7:00 PM-7:00 AM) for 1 week prior to use. Regular laboratory chow and tap water were allowed ad libitum. The rats were euthanized by CO2 asphyxiation and open thoracotomy. Liver was removed quickly, frozen, and stored at –70°C until fractionation of cytosol. The liver was rinsed several times to remove excess blood and fat and minced with scissors in ice-cold homogenization buffer, 0.25 M sucrose containing 10 mM triethanolamine HCl, pH 7.4, and 5 mM mercaptoethanol. All subsequent steps were performed at 0–4°C. The tissue was homogenized in 2 volumes of buffer per gram of wet tissue using a Potter-Elvehjem Tissue Grinder (Wheaton Science Products, Millville, NJ). The homogenates were immediately subjected to centrifugation at 105,000g for 1 h at 4°C, and the resulting supernatant fraction was used as the cytosol fraction. Protein concentration in the cytosol was determined by Bradford method (Bradford, 1976).
O-Sulfonation and O-Acetylation Assay of α-OHTAM.Cytosol. To purify commercially available calf thymus DNA, the DNA was incubated with RNase A, RNase T1, and proteinase K, extracted with phenol/chloroform, and then precipitated by ethanol (Umemoto et al., 1994). These procedures were repeated to remove completely RNA and protein. The concentration of DNA was determined by UV spectroscopy as 50 μg = O.D.260 nm 1.0. For determining O-sulfonation potential, calf thymus DNA (25 μg) was incubated at 37°C for 1 h with cytosol (1 mg of protein), 100 μM α-OHTAM, 200 μM PAPS, and 100 μM EDTA in 0.5 ml of 50 mM potassium phosphate, pH 6.5. For measuring O-acetylation potential, calf thymus DNA (25 μg) was incubated at 37°C for 1 h with cytosol (1 mg of protein), 100 μM α-OHTAM, 1 mM acetyl-CoA, 1 mM dithiothreitol, and 100 μM EDTA in 0.5 ml of 50 mM potassium phosphate, pH 7.4. The DNA was recovered using phenol/chloroform extraction following the ethanol precipitation and used for analysis of TAM-DNA adducts.
Rat and Human Hydroxysteroid Sulfotransferase and Acetyltransferase. Recombinant rat liver STa (42.0 units/mg of protein) (Sheng and Duffel, 2001) and human liver SULT 2A1 (63.0 units/mg of protein) (Apak and Duffel, 2004) were prepared as described previously. The enzyme activity was determined using a methylene blue paired ion extraction method with dehydroepiandrosterone (DHEA) as the substrate (Sheng and Duffel, 2001; Apak and Duffel, 2004). Enzyme units are expressed as nanomoles of sulfuric acid ester product formed from DHEA per minute. DNA (25 μg) was incubated at 37°C for 1 h with STa or SULT 2A1 (5 μg each) or human NAT I or NAT II (5 or 50 μg each) in the buffer described above.
Effect of STa Antibody on the Formation of TAM-DNA Adducts. To determine the inhibitory effect of TAM-DNA adduct formation, anti-STa-serum (0, 10, or 30 μl) prepared previously (Ogura et al., 1990) was preincubated at 37°C with cytosol (1 mg of protein) for 30 min and then added to 0.2 ml of 50 mM potassium phosphate, pH 6.5, containing calf thymus DNA (10 μg), 100 μM α-OHTAM, 200 μM PAPS, and 100 μM EDTA. After the reaction, the DNA was recovered as described above and used for analysis of DNA adducts.
Digestion of DNA Samples. DNA sample (2.5–5.0 μg) was enzymatically digested at 37°C overnight in 100 μl of 17 mM sodium succinate buffer (pH 6.0) containing 8 mM CaCl2, using micrococcal nuclease (30 units) and spleen phosphodiesterase (0.15 unit) (Terashima et al., 2002). The reaction mixture was incubated for another 1 h with nuclease P1 (1 unit). After the incubation, 200 μl of water was added. The reaction samples were then extracted twice with 200 μl of butanol. The butanol fractions were combined, back-extracted with 50 μl of distilled water, and evaporated to dryness.
32P Postlabeling/PAGE Analysis. The DNA digests were incubated at 37°C for 40 min with 10 μCi of [γ-32P]ATP and 3′-phosphatase-free T4 polynucleotide kinase (10 units), and then incubated with apyrase (50 milliunits) for another 30 min, as described previously (Terashima et al., 2002). Known amounts (0.152 pmol, 0.0152 pmol, 0.00152 pmol, or 0.000152 pmol) of dG-N2-TAM-modified oligodeoxynucleotide prepared by a phosphoramidite chemical procedure (Terashima et al., 2002) were mixed with 5 μg of calf thymus DNA (15,200 pmol) and served as a standard (1 adduct/105 nucleotides, 1 adduct/106 nucleotides, 1 adduct/107 nucleotides, or 1 adduct/108 nucleotides). As described previously (Terashima et al., 2002), the amount of TAM adducts detected increased linearly depending on the amounts of oligodeoxynucleotide used. A part of the 32P-labeled sample was electrophoresed for 4 to 5 h on a nondenaturing 30% polyacrylamide gel (35 × 42 × 0.04 cm) with 1400 to 1600 V/20 to 40 mA, as described previously (Terashima et al., 2002). The position of 32P-labeled adducts was established by β-PhosphorImager analysis (GE Healthcare). To quantify the level of 32P-labeled products, integrated values were measured using a β-PhosphorImager and compared with the standards. The detection limit for 5 μg of DNA was approximately 7 adducts/109 nucleotides.
Results
To determine the capability of forming TAM-DNA adducts through O-sulfonation and/or O-acetylation of α-OHTAM, calf thymus DNA was incubated in the mixture of α-OHTAM, rat hepatic cytosol, and either PAPS or acetyl-CoA as a cofactor. Formation of TAM-DNA adducts was analyzed using a 32P postlabeling/PAGE. When PAPS was used, large amounts of TAM-DNA adducts were detected at the level of 94.5 ± 32.9 adducts/107 nucleotides (means ± S.D., n = 3) (Fig. 2). The migration of major and minor TAM-DNA adducts was consistent with standards of dG3′p-N2-TAM. A trans-form (fr-2) was a major adduct at the level of 84.1 ± 28.6 adducts/107 nucleotides; another trans-form (fr-1) and cis-forms (a mixture of fr-3 and fr-4) were minor adducts at the level of 8.0 ± 2.6 adducts/107 nucleotides and 2.5 ± 1.9 adducts/107 nucleotides, respectively. An unknown adduct (5.9 ± 2.3 adducts/107 nucleotides) was observed at the lower position of dG3′p-N2-TAM standards. Only a trace of adduct (0.6 ± 0.2 adducts/107 nucleotides) was detected without PAPS. In contrast, when acetyl-CoA was used as a cofactor, no significant formation of TAM-DNA adducts was observed. The levels of adduct in the reaction condition with and without acetyl-CoA were 0.4 ± 0.1 adducts/107 nucleotides and 0.5 ± 0.1 adducts/107 nucleotides, respectively.
When rat STa (210 mU/5 μg) or human SULT 2A1 (315 mU/5 μg) was used for the reaction condition applied for rat hepatic cytosol and PAPS, TAM-DNA adducts were detected at the level of 70.3 adducts/107 nucleotides and 5.8 adducts/107 nucleotides, respectively (Fig. 3). Comparing with equivalent units of enzyme activity, the amount of TAM-DNA adducts generated by SULT 2A1 was approximately 18 times less than that of STa. Since NATs have O-acetylation activity, in addition to N-acetylation (Land et al., 1989), NAT I and NAT II were used to determine the capability of O-acetylating α-OHTAM in the presence of acetyl-CoA. However, human NAT I (1100 mU/5 μg) and NAT II (12,000 mU/5 μg) did not produce any TAM-DNA adducts (Fig. 3). Although 10-fold amounts of NATs were used, no reactivity to DNA was observed (data not shown).
The formation of TAM-DNA adducts with rat hepatic cytosol and PAPS increased depending on the incubation time (Fig. 4A). The presence of STa in the cytosol was confirmed by Western blot analysis using STa antibody, as described previously (Ogura et al., 1990). The addition of STa antibody prior to the reaction with the cytosol decreased the formation of TAM-DNA adducts to 9.8%, as observed with STa (8.2%) (Fig. 4B). The DHEA sulfation activity was reduced by addition of STa antibody in parallel to the formation of TAM-DNA adducts (data not shown).
To explore the substrate specificity for O-sulfonation, the reactivities of α-OH-N-desTAM and α-OHTAM N-oxide to calf thymus DNA were determined in the reaction mixture containing rat liver cytosol and PAPS and compared with that of α-OHTAM (Table 1). The formation of DNA adducts with α-OH-N-desTAM and α-OHTAM N-oxide was 50 and 2.8%, respectively, than that observed with α-OHTAM.
When liver cytosol (1 mg) from white women (23–71 years old) was incubated with DNA, α-OHTAM, and PAPS, the formation of TAM-DNA adducts was observed at the level of 0.91 ± 0.52 adducts/107 nucleotides (Fig. 5). A liver cytosol pooled from 22 humans (nine white women, 12 white men, and one Hispanic man) also promoted a similar level of TAM-DNA adducts (0.74 adducts/107 nucleotides). In the reaction condition without PAPS, no TAM-DNA adducts was observed. In addition, no adducts were formed when acetyl-CoA was used as a cofactor (data not shown).
Discussion
The capability of forming TAM-DNA adducts from α-OHTAM was investigated with rat and human liver cytosol in the presence of PAPS or acetyl-CoA as a cofactor. TAM-DNA adducts were generated only when PAPS was used. In addition, rat and human hydroxysteroid sulfotransferases (STa and SULT 2A1) generated TAM-DNA adducts from α-OHTAM; however, human NATs did not form any TAM-DNA adducts. Since synthetic α-acetoxyTAM is highly reactive to DNA, forming TAM-DNA adducts (Osborne et al., 1996; Dasaradhi and Shibutani, 1997), our results indicate that α-OHTAM may not be a substrate for acetyltransferases; therefore, no TAM-DNA adducts are formed. Thus, TAM-DNA adducts are formed through O-sulfonation, not O-acetylation, of α-OHTAM.
When α-OHTAM was incubated with DNA and rat liver cytosol in the absence of cofactors, a small amount of TAM-DNA adduct was observed at the level of 0.4 to 0.6 adducts/107 nucleotides. The TAM-DNA adduct formation may be reflected by an intrinsic reactivity of α-OHTAM to DNA (Phillips et al., 1994). However, such adduct formation was not detected under the similar reaction conditions with human liver cytosol. The trace of TAM-DNA adduct formation may be due to endogenous PAPS remaining in the rat liver cytosol, rather than in human liver cytosol, thus serving as a cosubstrate for a very small amount of O-sulfonation.
When a polyclonal antibody to STa was preincubated with rat liver cytosol before reacting with DNA, α-OHTAM, and PAPS, STa antibody inhibited 90% of TAM-DNA adduct formation, as similarly observed with STa. Since this antibody has no cross-reactivity with other rat isoforms, including phenol sulfotransferase and estrogen sulfotransferase (Ogura et al., 1990), our results supported that STa is a principal enzyme involved in O-sulfonation of α-OHTAM. The formation of TAM-DNA adduct was not completely inhibited. The antibody may not completely deny the enzyme active site; therefore, a partial sulfating activity may remain.
Mass spectroscopy and 32P postlabeling/HPLC analyses demonstrated that dG-N2-TAM and α-(N2-deoxyguanosinyl)-N-desmethyltamoxifen (dG-N2-N-desTAM) are the major hepatic DNA adducts of rodents treated with TAM (Rajaniemi et al., 1999; Umemoto et al., 2001). The dG-N2-TAM- and dG-N2-N-desTAM-DNA adducts accounted for over 95% of DNA adducts induced by TAM in rat liver (Umemoto et al., 2001); however, α-(N2-deoxyguanosinyl)tamoxifen N-oxide (dG-N2-TAM N-oxide) was not detected in the livers of rats. Supporting the animal studies, α-OHTAM N-oxide was a poor substrate for rat liver cytosol, as compared with α-OHTAM and α-OH-N-desTAM, for the formation of DNA adducts. The liver cytosol catalyzed α-OHTAM more rapidly than α-OH-N-desTAM, as observed previously with STa (Shibutani et al., 2002). However, the ratio of dG-N2-TAM and dG-N2-N-desTAM adducts observed in rat liver may vary depending on the amounts of α-OHTAM and α-OH-N-desTAM generated by cytochrome P450 enzymes from TAM and N-desTAM, respectively.
The formation of TAM-DNA adduct with human liver cytosol (0.91 adduct/107 nucleotides) was approximately 105 times lower than that when same amount of rat liver cytosol (94.5 adducts/107 nucleotides) was used. As observed with rat STa and human SULT 2A1 having the same enzyme activity for DHEA, the amount of TAM-DNA adducts generated by human SULT 2A1 was approximately 18 times less than that of STa. The lower formation of TAM-DNA adducts using the human cytosol may be due to either a low substrate specificity of SULT 2A1 for α-OHTAM and/or the lower presence of SULT 2A1 in human liver. In an earlier report, no TAM-DNA adducts were detected in livers of women treated with TAM (Martin et al., 1995). However, in the livers of monkeys given 6 times the human equivalent dose of TAM, TAM-DNA adducts including trans-form (fr-2) of dG-N2-TAM were detected at levels of 4.4 adducts/108 nucleotides, using 32P postlabeling/HPLC analysis developed in our laboratory (Shibutani et al., 2003). Similar results were observed for the same monkey DNA samples, using chemiluminescence immunoassay and HPLC electrospray tandem mass spectrometry (Schild et al., 2003). The failure of detecting TAM-DNA adducts in human livers in some studies may be due to differences in the sensitivity of 32P postlabeling analysis used (Shibutani et al., 2002). Since human liver SULT 2A1 is capable of O-sulfonating α-OHTAM, our results indicate the possibility that women receiving TAM may form genotoxic damage in the liver, as observed in monkey (Schild et al., 2003; Shibutani et al., 2003).
The level of each TAM and its metabolites in plasma of patients treated with TAM was 0.1 to 5.0 μM (Etienne et al., 1989). The concentration of TAM and its metabolites in human tissues is approximately 10- to 60-fold higher than in serum (Lien et al., 1991). α-OHTAM is a minor metabolite, accounting for ∼0.1% of the administered TAM dose (Jacolot et al., 1991). In our in vitro experiments, the formation of TAM-DNA adducts (0.63 adduct/107 nucleotides to 128 adducts/107 nucleotides) increased linearly using STa and α-OHTAM between 0.1 and 100 μM (data not shown). Although our in vitro study with relatively high levels of α-OHTAM (100 μM) may not reproduce the physiological conditions found in the tissues of patients with long-term exposure to low levels of α-OHTAM, the ability to form TAM-DNA adducts through O-sulfonation of α-OH-TAM was proven with rat and human liver cytosol.
TAM-DNA adducts were detected in the endometrium of breast cancer patients treated with TAM (Shibutani et al., 2000) and in the cultured human endometrial explants exposed to α-OHTAM (Kim et al., 2005), suggesting that a genotoxic mechanism is involved in the development of endometrial cancer. Although human HST was not detected in the endometrium of women with normal menstrual cycles (Rubin et al., 1999), no determination has been performed with postmenopausal women and women receiving TAM. Human HST may possibly be induced by TAM treatment, as observed in rats (Maiti and Chen, 2003). Other subfamilies of HST, such as SULT 2B1 expressed in human uterus (Geese and Blanchard, 2001), may be involved in O-sulfonation of α-OHTAM. To explore the mechanism of forming TAM-DNA adducts in the endometrium, detailed analysis of human HST expression in the tissue of women exposed to TAM is required. We conclude that α-hydroxylated tamoxifen metabolites are primarily activated via O-sulfonation, not O-acetylation, and then react with DNA, thereby exerting genotoxic effects in target tissues.
Footnotes
-
This research was supported by Grants ES09418 (to S.S.) from the National Institute of Environmental Health Sciences and CA38683 (to M.W.D.) from the National Cancer Institute.
-
Article, publication date, and citation information can be found at http://dmd.aspetjournals.org.
-
doi:10.1124/dmd.105.005330.
-
ABBREVIATIONS: TAM, tamoxifen; α-OHTAM, α-hydroxytamoxifen; fr, fraction(s); dG-N2-TAM, α-(N2-deoxyguanosinyl)tamoxifen; α-OH-N-desTAM, α-hydroxy-N-desmethyltamoxifen; α-OHTAM N-oxide, α-hydroxytamoxifen N-oxide; dG-N2-N-desTAM, α-(N2-deoxyguanosinyl)-N-desmethyltamoxifen; HST, hydroxysteroid sulfotransferase(s); dG3′p-N2-TAM, dG3′p-monophosphate-N2-tamoxifen; PAPS, 3′-phosphoadenosine 5′-phosphosulfate; acetyl-CoA, acetyl coenzyme A; NAT, N-acetyltransferase; DHEA, dehydroepiandrosterone; PAGE, polyacrylamide gel electrophoresis; HPLC, high-performance liquid chromatography.
- Received April 26, 2005.
- Accepted August 11, 2005.
- The American Society for Pharmacology and Experimental Therapeutics