860
Chem. Res. Toxicol. 1996, 9, 860-865
Oxidative Transformation of 2-Hydroxyestrone. Stability and Reactivity of 2,3-Estrone Quinone and Its Relationship to Estrogen Carcinogenicity Katmerka Tabakovic,† William B. Gleason,†,‡ William H. Ojala,‡ and Yusuf J. Abul-Hajj*,† Department of Medicinal Chemistry, Biomedical Engineering Center, and Department of Laboratory Medicine and Pathology, University of Minnesota, Minneapolis, Minnesota 55455 Received December 8, 1995X
The carcinogenicity of estrogens in rodents and man has been attributed to either alkylation of cellular macromolecules and/or redox-cycling, generation of active radicals, and DNA damage. Metabolic activation of estradiol leading to the formation of catechol estrogens is believed to be a prerequisite for its genotoxic effects. 4-Hydroxyestradiol, although not 2-hydroxyestradiol, is a potent inducer of tumors in hamsters. Previous studies have shown that 3,4-estrone quinone can redox-cycle and is capable of inducing exclusively single strand DNA breaks in MCF-7 breast cancer cells, as well as react with various nucleophiles (thiol, imidazole, amino, phenolate, and acetoxy) to give Michael addition products. These results support the possible involvement of 3,4-catechol/quinone estrogens in estrogen’s carcinogenicity. To explain the decreased carcinogenicity of 2-hydroxyestrogens, the reactions of 2,3-estrone quinone (2,3EQ) with nucleophiles were investigated. Reactions of 4-methylimidazole with 2,3-EQ gave a complex mixture of products leadng to the formation of the catechol, C-O dimerization product, and a 1,6-Michael addition product identified as the 1-(4-methylimidazolo)-2-hydroxyestrone. Reactions of 2,3-EQ under mildly basic conditions with either ethyl phenolate or acetate gave several products which were characterized as the C-O and C-C dimers, catechol, and 3,5dihydroxy-1(10),3-estradiene-2,17-dione. No Michael addition products were detected under these experimental conditions. The same products were also observed during the synthesis of 2,3-EQ, which led us to postulate that the lack of carcinogenicity of 2-hydroxyestrogens may be related to the increased reactivity and decreased stability of the quinone under physiological conditions. These results are contrasted with those obtained with 3,4-EQ which is much more stable and therefore could diffuse from the site of formation to the target tissue. These results along with rapid methylation and clearance may be very likely explanations for the decreased carcinogenicity of 2-hydroxyestrogens.
Introduction Recent studies have shown a connection between excessive exposure to estrogens and the development of cancer in several tissues including breast, endometrium, liver, and kidney in humans and experimental animals (1-7). Although the exact mechanism for carcinogenesis induced by estrogenic compounds is not fully understood, it is generally believed that the estrogen o-quinones produced by the oxidation of catechol estrogens by phenol oxidase (8), prostaglandin H synthase (9), and cytochrome P-450 (10) have the potential to be cytotoxic and/or genotoxic. The molecular mechanisms for o-quinone cytotoxicity have been attributed to the alkylation of cellular macromolecules and/or redox cycling generating reactive oxygen species which oxidize essential cellular components. Previous studies from this laboratory have shown that 3,4-estrone o-quinone (3,4-EQ)1 can redoxcycle leading to the formation of hydrogen peroxide, hydroxyl radical, and the semiquinone of 3,4-EQ (11). Furthermore, we have shown that 3,4-EQ is capable of * Author to whom correspondence should be addressed. † Department of Medicinal Chemistry. ‡ Biomedical Engineering Center and Department of Laboratory Medicine and Pathology. X Abstract published in Advance ACS Abstracts, June 1, 1996. 1 Abbreviations: 2,3-estrone quinone, 2,3-EQ; 3,4-estrone quinone, 3,4-EQ; 2-hydroxyestrone, 2-OHE1; 4-hydroxyestrone, 4-OHE1; 4-hydroxyestradiol, 4-OHE2; triethylamine, Et3N.
S0893-228x(95)00205-0 CCC: $12.00
inducing exclusively single strand DNA breaks in a human breast cancer cell line (12). Studies in our laboratory are continuing to investigate the importance of redox-cycling in the carcinogenicity of estrogen oquinones. Covalent binding of proteins may inhibit essential functions of enzymes or regulatory proteins which may ultimately lead to their toxic effects. Thus, it is quite possible that the carcinogenicity/toxicity of estrogen o-quinones may be, in part, due to their ability to react with proteins. It has been well documented by us and others that thiols react with estrogen o-quinones by Michael addition (13-15). Furthermore, addition reactions between lysine and quinones have been implicated in several protein-quinone interactions. For example, the decrease in lysine content of casein following incubation of caffeoquinone and chlorogenquinone has been attributed to the addition of the -amino group of lysine to the respective quinone moiety (16, 17). Studies in our laboratory showed that lysine can form Michael addition products with 3,4-EQ as well as form the iminoquinone products (18). To further our understanding of the chemistry of conjugate addition to estrogen o-quinones, our laboratory has carried out extensive studies on the reaction of 3,4-EQ with several model compounds representing the amino acid side chains of cysteine, lysine, histidine, aspartic acid, and tyrosine in which catechol © 1996 American Chemical Society
2,3-Estrone Quinone
addition products were obtained (19). On the other hand, amino acid side chain mimics of arginine, serine, tryptophan, and asparagine did not lead to adduct formation. The reactivity and stability of 3,4-EQ have been well documented; however, very little is known about 2,3-EQ even though the 2,3-catechol is the major metabolite of estrogens (20-22). The present study investigates the reactivity of the catechol and the 2,3-estrogen quinone with various amino acid nucleophiles. The results from previous studies and this investigation show that 2,3EQ is considerably less stable than 3,4-EQ and, while it forms addition products with thiols and imidazole, it undergoes extensive transformations under mildly basic conditions to give several products.
Experimental Procedures Chemicals. Estrone was purchased from Steraloids (Wilton, NH). 4-Ethylphenol and 4-methylimidazole were obtained from Aldrich Chemical Co. (Milwaukee, WI). Deuterated solvents were purchased from Aldrich Chemical Co.. The synthesis of the catechol estrogens was carried out as described by Stubenrauch and Knuppen (23). The estrogen o-quinones were synthesized by the oxidation of the catechol estrogens by activated MnO2 (24) or by oxidation of the 2-amino estrone (23). Caution: The estrogen o-quinones are potentially hazardous and were handled in accordance with NIH Guidelines for the Laboratory Use of Chemical Carcinogens (25). Characterization of Compounds. All of the compounds were characterized by nuclear magnetic resonance (1H-NMR), ultraviolet (UV), infrared (IR), and mass spectrometry (MS). Melting points (uncorrected) were taken on a Fisher-Johns apparatus. 1H-NMR spectra were obtained with a GE-300 MHz NMR spectrometer, and the chemical shift data are reported in parts per million (δ) downfield from tetramethylsilane as an internal standard. Nuclear Overhauser effect (NOE) difference spectra were recorded by applying a presaturation pulse with a decoupler on resonance and subtracting the trace from the corresponding reference spectra recorded under identical conditions but with the decoupler off-resonance. Typical spectra were obtained from 64 transients. The UV spectra were obtained using a Beckman DU-70. The IR spectra were obtained on a Nicolet 5 DXC FT-IR spectrometer. Mass spectral data were determined on an AEI MS-30, a VG 7070 E-HF, and a Finnigan MAT 95. X-ray Crystallography. Colorless crystals of 7 were grown in acetone/hexane. All measurements were made on a Rigaku AFC6S X-ray diffractometer with graphite monochromated Cu KR radiation (l ) 1.54178 Å) using the w - 2q scan mode. The crystal temperature was maintained at 173(1) K by using a Molecular Structure Corp. low temperature device. Intensities were corrected for Lorentz and polarization effects. Equivalent reflections were merged, and absorption effects were corrected using the ψ scan technique (transmission factors: 0.96-1.00) as described by North et al. (26). The structure was solved by direct methods using SHELXS86 (27) and refined and finished using the TEXSAN structure analysis package (28). All nonhydrogen atoms were refined anisotropically. Hydrogens bonded to carbon atoms were placed in calculated positions (0.95 Å) and were not refined. Hydrogens bonded to heteroatoms were located in difference Fourier maps, and their positional parameters were refined. Crystal data are given in Table 1. Synthesis of Catechol Estrogens and Estrogen Quinones. Synthesis of the 2-OHE1 and 2,3-EQ was carried out as described by Stubenrauch and Knuppen (23) and by our laboratory (24). 2,3-EQ (1): 1H-NMR (CDCl3) δ 6.25 (d, 1H, H-4, J ) 2.4 Hz), 6.18 (s, 1H, H-1), 0.91 (s, 3H, 18-CH3). 2-OHE1 (2): 1H-NMR (CDCl3) δ 6.81 (s, 1H, H-1), 6.61 (s, 1H, H-4), 2.8 (m, 2H, H-6), 0.86 (s, 3H, 18-CH3). Reaction of 2,3-Estrogen o-Quinone (1) with 4-Methylimidazole. 4-Methylimidazole (160 mg, 1.94 mmol) in CH2Cl2 (50 mL) was added to a solution of freshly prepared (in situ)
Chem. Res. Toxicol., Vol. 9, No. 5, 1996 861 Table 1. Crystal Data for Compound 7 parameter
compound
empirical formula formula wt (amu) crystal color and habit cryst dimensions (mm) cryst system a (Å) b (Å) c (Å) V (Å3) Z space group calcd density (g/cm3) F000 µ(Cu KR) (cm-1) no. of measured reflections total unique no. of obsd refl. (I > 3.00σ(I)) Rint R wR max e- in final difference map min e- in final difference map
C23H30O6 402.49 colorless, needle 0.64 × 0.11 × 0.11 monoclinic 12.573(3) 7.042(2) 12.882(2) 1050.5(3) 2 P21 (#4) 1.272 432 7.47 4234 2017 2967 0.032 0.044 0.045 +0.22 e-/Å3 -0.22 e-/Å3
2,3-EQ (1, 500 mg, 1.76 mmol) in CH2Cl2 (10 mL), and the reaction mixture was stirred at room temperature under nitrogen. The quinone was consumed within 1 h (checked by TLC), after which the products were isolated and purified. Evaporation of the CH2Cl2 showed the presence of five spots which were separated on a silica gel column. Elution of the column with 25% EtOAc in benzene gave compound 2. Further elution of the column yielded a light orange solid of 5: mp 105107 °C dec; Rf ) 0.23 in 25% benzene in EtOAc; UV (CH2Cl2) λmax (nm) 230, 318, 466; 1H-NMR (CDCl3) δ 6.32 (s, 1H, H-1), 6.15 (s, 1H, H-4), 6.06 (s, 1H, H-4′), 5.43 (s, 1H, OH, D2O exchangeable), 1.21 (s, 3H, 18′-CH3), 0.85 (s, 3H, 18-CH3); EI MS m/z 570 (M+ + 2, 18.6), 568 (M+, 26.5), 511 (M+ - C3H5O, 2.3), 483 (M+ - C5H9O, 2.3), 470 (M+ - C6H10O, 1.6), 455 (M+ - C7H13O, 1.5), 444 (M+ - C8H12O, 1.4), 286 (100), in agreement with the proposed structure of 5. Continued elution of the column gave 17.6% yield of a white solid of 3. This compound was found to be unstable; it undergoes oxidation first to the quinone adduct followed by decomposition to 2 and methylimidazole. In a separate experiment, the reaction mixture was acetylated immediately prior to separation and purification over a silica gel column. Acetylation using a 1:1 mixture of pyridine/acetic anhydride followed by the usual workup gave a mixture of the acetates of 2, 3, and 5. Separation on silica gel by elution with a mixture of benzene/EtOAc/MeOH (6.5:2.0:1.5) gave the acetates of 2 and 5 which were not purified any further. Continued elution of the column yielded the 2,3-diacetylated 1-(4methylimidazolo)estrone 4: mp 152 °C dec; Rf ) 0.35 (25% EtOAC in benzene); UV λmax (nm) 220, 240, 318; IR (KBr) cm-1 2931, 2918, 1766, 1740, 1487, 1380, 1372, 1206, 1130, 1015; 1HNMR (CDCl3) δ 7.36 (s, 1H, H-2 imidazole), 6.91 (s, 1H, H-4), 6.19 (s, 1H, H-5 imidazole), 2.89 (m, 2H, H-6), 2.28 (s, 3H, OAc), 2.27 (s, 1H, OAc), 2.17 (s, 3H, CH3-imidazole), 0.91 (s, 3H, 18CH3); ES MS m/z 450 (M+, 8.5). Reaction of 2,3-Estrone Quinone (1) with and without 4-Ethylphenol or Acetic Acid. To a solution of 4-ethylphenol (0.35 mmol) in CH2Cl2 (30 mL) was added three drops of Et3N and 2,3-estrogen o-quinone (1, 100 mg, 0.35 mmol), and the reaction mixture was stirred at room temperature. The quinone was totally consumed within 30 min (checked by TLC). Evaporation of CH2Cl2 left a solid that showed several spots on TLC which were separated on a silica gel column. Elution of the column with 25% EtOAc in benzene yielded compounds 2 and 5 which were characterized as described above. Product 9 was obtained in 5% yield: mp 210 °C dec; Rf ) 0.49 in 25% EtOAc in benzene; UV (CH3OH) λmax (nm) 222, 241, 278, 420; IR (KBr) cm-1 3409, 2961, 2927, 2850, 1739, 1384; 1H-NMR (CDCl3) δ 7.72 (d, 1H, H-6′, J ) 8.3 Hz), 7.34 (s, 1H, H-1′), 7.20 (s, 1H,
862
Chem. Res. Toxicol., Vol. 9, No. 5, 1996
Tabakovic et al.
Figure 1. Products from the reaction of 2,3-estrone o-quinone with 4-methylimidazole. H-4), 7.14 (d, 1H, H-6, J ) 8.3 Hz), 5.92 (bs, 1H, OH, D2O exchangeable), 5.82 (bs, 1H, OH, D2O exchangeable), 1.26 (t, 3H, 18′-CH3), 0.80 (s, 3H, 18-CH3); EI MS m/z 566.8 (M + H)+; FAB MS (glycerol/TFA matrix 2000Res) m/z 567.2 (M + H+), 565.2 (M - H+). Continued elution of the column gave sufficient quantities of the unstable compound 6 for partial characterization: Rf ) 0.26 (25% EtOAc in benzene); 1H-NMR (CDCl3) δ 6.34 (bs, 1H, 3-OH, D2O exchangeable), 6.10 (d, 1H, H-4, J ) 2.8 Hz), 6.06 (s, 1H, H-1), 2.17 (s, 1H, 5-OH, D2O exchangeable), 0.92 (s, 3H, 18-CH3); EI MS m/z 302.2 (M+, 87.6), 286 (M - 16, 55), 284 (M - 18, 38), 163 (18), 145 (46), 139 (100), 107 (22). Compound 7 was identified as the monoacetate: mp 103105 °C; Rf ) 0.23 (25% EtOAc in benzene); UV (CH3OH) λmax (nm) 202, 240; IR (KBr) cm-1 3410, 2990, 2930, 2890, 1771, 1740, 1675, 1650, 1371, 1134; 1H-NMR (CDCl3) δ 6.49 (s, 1H, H-4), 6.02 (s, 1H, H-1), 2.83 (s 1H, OH, D2O exchangeable), 2.26 (s, 3H, CH3CO), 0.88 (s, 3H, 18-CH3); ES MS m/z 344.2 (M+, 10), 367.2 (M+ + Na, 100). Compound 8 had the following characteristics: mp 210-212 °C dec; Rf ) 0.44 (25% EtOAc in benzene); UV (CH3OH) λmax (nm) 206, 237; IR (KBr) cm-1 2980, 2930, 2870, 1773, 1734, 1676, 1652, 1647, 1384, 1201, 1138; 1H-NMR (CDCl3) δ 6.59 (s,1H, H-4), 6.20 (s,1H, H-1), 2.46-2.48 (m, 4H), 2.27 (s, 3H, 3-CH3CO), 2.06 (s, 3H, 5-CH3CO), 0.91 (s, 3H, 18-CH3); ES MS m/z 386 (M+, 70); EI MS m/z 386 (M+, 0.28), 344 (21), 302 (100), 286 (17.7), 43 (19.8).
Results and Discussion Although the underlying mechanisms of estrogen carcinogenesis are as yet unresolved, it has been proposed that metabolic activation of estradiol leading to the formation of catechol estrogens is a prerequisite for its genotoxic activity (7). Whether arylation of macromolecules or redox-cycling of catechols/quinones is involved in estrogen carcinogenesis is the subject of intensive investigations in several laboratories. 2-OHE2 is the predominant metabolite of estrogens in most tissues, yet 4-hydroxyestradiol has been found to be a much more potent inducer of tumors (29). The differences in activities between the two catechols have been ascribed to the rapid methylation and metabolic clearance of 2-OHE2 (29) as well as inhibition of O-methylation of 4-OHE2 by 2-OHE2 (30). Because of these results, our laboratory has conducted extensive studies on the reactions of 3,4-EQ with several nucleophiles to determine the nature of the chemical interactions between the estrogen quinones and macromolecules (13, 18, 19, 31, 32).
Previous studies have shown that thiols react with 2,3EQ by 1,6-Michael addition, leading to the formation of both the 1- and 4-substituted adducts (14), while reactions with amine nucleophiles resulted in 1,2-additions, leading to the formation of the monoiminoquinone and bisiminoquinone products (31). To further our studies on the reactivity of 2,3-EQ toward other nucleophiles and to compare these results with 3,4-EQ, the present investigation reports the results of the reactions of 2,3EQ with methylimidazole, acetate, and ethyl phenolate. In the reaction of 2,3-EQ with 4-methylimidazole, three major products were isolated and purified using a combination of column chromatography and preparative TLC (Figure 1). Spectroscopic data for compound 2 were found to be identical to those of an authentic sample of 2-OHE1 which may be formed by reduction of 2,3-EQ by either the catechol adduct 3 or the catechol precursors to 5. The structure of the dimeric product 5a/5b was deduced essentially from UV, MS, and NMR data. The mass spectral data showed a molecular ion at 568 corresponding to the molecular mass of the proposed structure. The products from oxidative phenolic coupling can be formed through a C-C or C-O bond (33). Confirmation of the presence of the C-O bond comes from the NMR spectrum which shows one singlet at 6.06 ppm attributed to one proton of the o-quinone moiety, and two other singlets at 6.32 and 6.15 ppm which are assigned to the two aromatic protons of the catechol moiety. However, on the basis of NMR spectra, it is not possible to discriminate between two possible isomers which could arise through a C-O coupling reaction. Formation of 5 is postulated to occur either by a oneelectron transfer via a charge transfer complex (34), or by reduction of 2,3-EQ by the Michael addition product 3 leading to the formation of 2, which could Michael add to 2,3-EQ leading to the formation of the catechol precursor to 5 followed by oxidation to 5 by 2,3-EQ, to regenerate 2. It is interesting to note that the C-C coupling products from this reaction were not detected. Compound 3 showed a singlet at 6.88 ppm assigned to an aromatic proton and other peaks corresponding to the structure of 4-methylimidazole, suggesting a Michael addition product. However, full characterization of 3 could not be carried out due to the fact that it undergoes air oxidation to the quinonoid adduct of 3 followed by
2,3-Estrone Quinone
decomposition to 2-OHE1 and methylimidazole as detected on TLC and by NMR. To further characterize the structure of 3, the reaction mixture between 2,3-EQ and methylimidazole was acetylated immediately following evaporation of CH2Cl2 and the acetylated product 4 was obtained as a stable crystalline material. Proton NMR showed the presence of a singlet at 6.91 ppm assigned to the aromatic ring of the steroid, three singlets at 7.36 (1H), 6.19 (1H), and 2.17 (3H) ppm assigned to the methylimidazole moiety, and two singlets at 2.28 and 2.27 ppm assigned to the acetates at C-2 and C-3. Mass spectral data (ES) revealed a molecular ion peak at 450 corresponding to the proposed structure of 4. The assigned regiochemistry of 4 was supported by 1H NOE difference experiments. Irradiation of the resonance at 6.81 ppm in 2 assigned to the C-1 proton in ring A resulted in a significant enhancement of the signals at 2.3 ppm protons for C-11 while irradiation of the singlet at 6.61 assigned to the C-4 proton gave significant enhancement of the signals at 2.8 ppm for C-6 protons. Thus, when the resonance at 6.91 ppm in 4 was irradiated, enhancement of the signal at 2.89 ppm for C-6 was observed. These results strongly suggest that the methylimidazole reacts with 2,3-EQ by 1,6-Michael addition resulting in the formation of the C-1 adduct. It is interesting to note that, under the conditions used in the present study, no C-4 imidazolo addition product was observed. It is quite likely that the C-4 adduct was formed but is sufficiently labile that the reaction may be readily reversible. In contrast, the C-1 adduct is readily isolable and is relatively stable toward retro-Michael reactions, although we ultimately had to stabilize the compound as the diacetate. Alternatively, the C-4 adduct may not be formed due to differences in electrostatic factors leading to increased reactivity at the C-1 relative to the C-4 position even though C-1 is more sterically hindered. At present, we cannot give a satisfactory explanation for these observations. Furthermore, while it is quite possible that either the N-1 or N-3 of methylimidazole adds to the C-1 position of 2,3-EQ, we propose that the N-1 position is involved in bond formation as shown for compound 3. This rationale is based on earlier studies on the reaction of the N-1 methylimidazole position and the C-1 position of 3,4-EQ as fully characterized using X-ray analysis of 1-(4-methylimidazolo)-4hydroxyestrone (19). In the reaction of 2,3-EQ with ethylphenol (Figure 2), a trace amount of triethylamine was used to ionize the phenolate as previously used in our studies with 3,4-EQ (19). Although several products were formed, isolation and identification of the products using spectral methods showed no Michael addition of the phenolate. When acetic acid was reacted with 2,3-EQ in the presence of Et3N, the same products were obtained as in the phenolate reaction. These results suggested that the reactions did not involve the nucleophiles, which led us to investigate further the reactivity and stability of 2,3-EQ and 2,3-catechol estrone under different experimental conditions. Thus, when 1 was treated with trace amounts of Et3N in CH2Cl2, the same reaction products were observed as detected by TLC analysis. During the synthesis of 2,3-EQ from oxidation of 2-aminoestrone as described by Stubenrauch and Knuppen (23) or by oxidation of 2-OHE1 with MnO2 (24) at room temperature, essentially the same reaction products were formed, albeit with different yields. Furthermore, oxidation of 2-OHE1 with MnO2 in extremely dry CH2Cl2 conditions at -70
Chem. Res. Toxicol., Vol. 9, No. 5, 1996 863
Figure 2. Products obtained from the reaction of 2,3-estrone o-quinone with ethylphenol or acetic acid in the presence of Et3N, as well as in the presence or absence of Et3N.
°C gave about 70% of the quinone which underwent transformation within a few minutes to the same products as the temperature increased. Thus, it seems that during the synthesis of 2,3-EQ a considerable amount of chemistry takes place that leads to the formation of several products including 2, 5, 6, and 9. The structures of 2 and 5 were found to be identical to those obtained from the reaction of methylimidazole with 2,3-EQ. Compound 9 was found to be dimeric, and its structure was deduced from UV, MS, and NMR data. The mass spectrum of 9 showed a molecular ion at 566 (ES) corresponding to a molecular mass of the proposed structure and is consistent with 1,4 C-C coupling of estrone quinone. Further support for this was obtained from FAB MS which showed peaks at 567 (FAB+) and 565 (FAB-). In a separate experiment with a crude sample of 9, FAB+ MS showed the presence of an additional small peak at 849 indicating the presence of a trimer, which lends further support to 1,4 C-C coupling of the quinone. Proton NMR of 9 showed the presence of two singlets and two doublets in the aromatic region. The singlets at 7.34 and 7.20 ppm were assigned to the C-1′ and C-4 protons, respectively, whereas the doublets at 7.52 and 7.14 ppm were assigned to the C-6′ and C-6 protons, respectively. The NMR also showed two exchangeable protons at 5.92 and 5.82 ppm assigned to the two phenolic -OH groups. The UV spectrum showed absorption bands at 222, 278, and 420 nm which are consistent with the structure of compound 9. Formation of this compound probably involved 1,4 C-C coupling leading to the quinone intermediate which undergoes isomerization to give ultimately the quinone methide product 9. While previous studies in our laboratory failed to show the formation of quinone methide from either 3,4-EQ or 2,3-EQ, it is quite possible that they may have been formed but were undetected. Recent studies by Bolton’s group (35) have shown the formation of quinone methide intermediates from oxidation of the 2,3-dihydroxy-5,6,7,8-tetrahydronaphthalene, which represent the A/B ring skeleton of 2,3-catechol estrogen. Compound 6 was isolated using a combination of column chromatography and TLC, and even though it was not very stable, sufficient quantities were available to carry out preliminary characterization of the compound. Mass spectral data showed a compound that has
864
Chem. Res. Toxicol., Vol. 9, No. 5, 1996
Figure 3. ORTEP drawing of 7 with crystallographic numbering system. Ellipsoids representing the non-hydrogen atoms are shown at the 50% probability level.
a molecular weight of 302, suggesting the addition of a water molecule to the estrogen quinone. NMR data showed the presence of a singlet at 6.06 ppm and a doublet at 6.10 ppm assigned to the C-1 and C-4 protons, respectively. However, this compound was found to decompose even in the NMR tube. Thus, in order to obtain a pure sample of 6, we carried out a separate study in which the reaction mixture was acetylated immediately using pyridine/acetic anhydride and the mixture worked up in the usual manner. Isolation and purification of the acetylated mixture gave two acetylation products of 6. Compound 7 was obtained as a white solid which was crystallized from acetone/hexane to provide the analytically pure compound which was characterized using UV, IR, 1H-NMR, and mass spectral studies. The proton NMR spectrum of 7 revealed the presence of a quinol product with two singlet peaks in the aromatic region at 6.49 and 6.02 ppm assigned to the C-4 and C-1 protons, respectively. The assigned positions for the C-4 and C-1 protons were supported by 1H NOE difference studies. Irradiation of the resonance at 6.49 ppm assigned to the C-4 proton showed a positive NOE effect on the signal at 2.83 ppm, while irradiation of the resonance at 6.02 ppm assigned to the C-1 proton showed no enhancement of the signal at 2.83 ppm. The NMR also showed an exchangeable proton at 2.83 ppm assigned to an OH group and a singlet at 2.26 ppm assigned to the methyl protons of the acetate group. These results suggested the presence of a secondary acetoxy group, two aromatic protons, and another hydroxyl group which was assigned to the C-5 position. These studies, however, could not differentiate between the R or β orientation of the C-5 OH group. Conclusive evidence for the structure of 7 was obtained from X-ray crystallographic analysis performed on a single crystal of 7. Figure 3 is an ORTEP drawing of 7 which shows the results of 1,4-Michael addition in which the hydroxyl nucleophile attacks the C-5 position of 2,3-EQ from the R-face. The X-ray structural data for 7 showed that this product crystallizes as an acetone solvate, which in fact gave us considerable difficulty in the interpretation of the NMR data of the crystalline product obtained following recrystallization from acetone/hexane. Crystallographic data for 7 are summarized in Table 1 (see Experimental Procedures). Compound 8 was identified as the diacetyl derivative of 6. The presence of the two acetyl moieties was demonstrated by the signals at 2.27 ppm for the C-3 acetyl group and at 2.06 ppm for the C-5 acetyl group. The mass spectral data gave a molecular ion peak at 386 (ES
Tabakovic et al.
MS and EI) which is consistent with the proposed structure of 8. In the present investigation, we used model reactions carried out in CH2Cl2 to delineate possible modes of reaction of nucleophiles with 2,3-EQ. Since it is quite possible that the behavior of o-quinones may be different under physiological conditions, we carried out studies on the reaction of 2,3-EQ with several nucleophiles, including thiol, amine, carboxylate, and phenolate, in a mixture of CH2Cl2/DMF/H2O (1:1:1). The product profiles for these reactions were compared using TLC and were found to be essentially similar to those obtained in reactions in a methylene chloride solution. Furthermore, these observations are consistent with our previous studies on reaction of 3,4-EQ with lysine in an aqueous solution in which no differences were observed between the product profiles in organic and aqueous systems (18). Covalent binding of 2,3-EQ to thiol and imidazole nucleophiles is consistent with 1,6-Michael addition. On the other hand, formation of 3 suggests 1,4-Michael addition of the hydroxyl nucleophile to a methine position. This is a unique reaction especially since the hydroxyl nucleophile could also add to either the C-1 or C-4 positions. To our knowledge, this constitutes the first observation for addition of a nucleophile to a ring junction. Thus, the formation of this intermediate may provide an additional explanation for the transformation of the quinone to the catechol as evidenced by the transformation of 3 to the catechol under mild aqueous conditions. Attempts at showing a similar reaction with 3,4-EQ have not been successful. Whether this observation is unique to 2,3-EQ remains to be elucidated. It is worthwhile here to analyze the differences between the two catechol estrogens. Exogenous 4-hydroxyestrogens are more carcinogenic than 2-hydroxyestrogens in hamsters (29). This observation has been proposed to be not due to differences in the reactivities of the catechol estrogen toward cellular nucleophiles, since their respective reactivities in chemical or biological systems, such as alkylation, oxidation to quinones, or formation of conjugates, appear to be comparable (36), but rather due to reduced bioavailability of 2-hydroxyestrogens to the target tissues as a result of increased metabolic methylation and clearance (29). While these observations may partly explain the differences between the catechol estrogens, other factors obtained from more recent studies may equally play an additional role in these differences. Although previous reports have suggested no differences in reactivities between the catechol estrogens (36), recent results by Bolton’s group (37) showed that 2,3-EQ was found to be 17-fold more reactive than 3,4EQ, thus confirming previous qualitative reports of reactivity differences between the two o-quinones (38). These results as well as those reported in this and previous investigations lead us to suggest that a relationship may exist between the stability and reactivity of the estrogen quinones and their carcinogenic potential. It is tempting to postulate that the lack of carcinogenicity of 2,3-OHE2 may be related to the increased reactivity and decreased stability of the quinone under physiological conditions. In contrast, the greater stability of 3,4-EQ, as evidenced by our ability to obtain this compound in crystalline form (24), may allow more time for diffusion of the reactive intermediate away from the site of formation to interact with vital cellular macromolecules, leading to its high carcinogenicity relative to 2,3-EQ. Thus, a complex interaction among many factors, includ-
2,3-Estrone Quinone
Chem. Res. Toxicol., Vol. 9, No. 5, 1996 865
ing the rate of formation of reactive intermediates, their bioavailability, stability, and inherent reactivity, determines the extent of an estrogen’s cytotoxicity. Consequently, the ability of a reactive metabolite to diffuse could be one important factor leading to its direct carcinogenicity. In conclusion, data have been presented concerning the reactivity of 2-OHE1 and 2,3-EQ leading to the formation of Michael addition adducts in reactions of 2,3-EQ involving imidazole. On the other hand, reactions under mildly basic conditions lead to dimerization and addition of hydroxyl nucleophile to the C-5 position of 2,3-EQ. Our results suggest that the increased reactivity and decreased stability of 2,3-catechol/quinone estrogens may provide an additional explanation for their decreased carcinogenicity.
Acknowledgment. This research was supported in part by the National Cancer Institute, Grant CA57615. Supporting Information Available: The X-ray crystallographic data for compound 7 (24 pages), contained in many libraries on microfiche, immediately follows this article in the microfilm version of the journal, and can be ordered from the ACS (see any current masthead page for ordering information).
References (1) Henderson, B. E., Ross, R., and Bernstein, L. (1988) Estrogens as a cause of human cancer: The Richard and Hinda Rosenthal Foundation award lecture. Cancer Res. 48, 246-253. (2) Harris, J. R., Lippman, M. E., and Veronesi, L. L. (1992) Medical progress. Breast cancer. New Engl. J. Med. 327, 319-328. (3) Palmer, J. R., Rosenberg, L., Kaufman, D. W., Warshner, M. E., Stolley, P., and Shapiro, S. (1989) Oral contraceptive use and liver cancer. Am. J. Epidemiol. 130, 878-882. (4) Tawari, A., Negri, E., Perazzini, F., Franceschi, S., and LaVecchia, C. (1993) Female hormone utilization and risk of hepatocellular carcinoma. Br. J. Cancer 67, 635-637. (5) IARC Group (1979) IARC Monographs on Evaluation of Carcinogenic Risk of Chemicals to Humans, Sex Hormones (II), Vol. 21, pp 173-221, International Agency for Research on Cancer, Lyon, France. (6) Kirkman, H. (1984) Estrogen-induced tumors of the kidney in the Syrian hamster. III. Growth characteristics in the Syrian hamster. Natl. Cancer Inst. Monogr. 1, 1-57. (7) Liehr, J. G. (1990) Genotoxic effects of estrogens. Mutat. Res. 228, 269-276. (8) Kalyanaraman, B., Felix, C. C., and Sealy, R. C. (1985) Semiquinone radicals of catecholamines, catecholestrogens, and their metal ion complexes. Environ. Health Perspect. 64, 185-198. (9) Degen, G. H. (1990) Role of prostaglandin-H synthase in mediating genotoxic and carcinogenic effects of estrogens. Environ. Health Perspect. 88, 217-223. (10) Capdevila, J., Sack, Y., and Falck, J. R. (1984) The mechanistic plurality of cytochrome P-450 and its biological ramifications. Xenobiotica 14, 105-118. (11) Nutter, L. M., Wu, Y.-Y., Ngo, E. O., Sierra, E. E., Gutierrez, P. L., and Abul-Hajj, Y. J. (1994) An o-quinone form of estrogen produces free radicals in human breast cancer cells: Correlation with DNA damage. Chem. Res. Toxicol. 7, 23-28. (12) Nutter, L. M., Ngo, E. O., and Abul-Hajj, Y. J. (1991) Characterization of DNA damage induced by 3,4-estrone-o-quinone in human cells. J. Biol. Chem. 226, 16380-16386. (13) Abul-Hajj, Y. J., Cisek, P. L. (1988) Catechol estrogen adducts. J. Steroid Biochem. 31, 107-110.
(14) Abul-Hajj, Y. J., and Cisek, P. L. (1986) Regioselective reaction of thiols with catechol estrogens and estrogen-o-quinones. J. Steroid Biochem. 25, 245-247. (15) Jellinck, P. H. (1988) 4-Hydroxyestradiol is conjugated with thiols primarily at C-2: Evidence from regio-specific displacement of tritium by cat liver microsomes or tyrosinase. Steroids 51, 395409. (16) Hurrell, R. F., and Finot, P. A. (1984) Nutritional consequences of the reactions between proteins and oxidized polyphenolic acids. Adv. Exp. Med. Biol. 177, 423-434. (17) Hurrell, R. F., Finot, P. A., and Cuq, J. L. (1982) Proteinpolyphenol reactions. I. Nutritional and metabolic consequences of the reactions between oxidized caffeic acid and the lysine residues of casein. Br. J. Nutr. 47, 191-211. (18) Tabakovic, K., and Abul-Hajj, Y. J. (1994) Reaction of L-lysine with estrone-3,4-o-quinone. Chem. Res. Toxicol. 7, 696-701. (19) Abul-Hajj, Y. J., Tabakovic, K., Gleason, W. B., and Ojala, W. H. (1996) Reactions of 3,4-estrone quinone with mimics of amino acid side chains. Chem. Res. Toxicol. 9, 434-438. (20) Purdy, R. H., Moore, P. H., Jr., Williams, M. C., Goldzieher, J. W., and Paul, S. M. (1982) Relative rates of 2- and 4-hydroxyestrogen synthesis are dependent on both substrate and tissue. FEBS Lett. 138, 40-44. (21) Maggs, J. L., Morgan, P., and Park, B. K. (1992) The sexually differentiated metabolism of [6,7-3H] 17β-oestradiol in rats: Malespecific 15R- and male-selective 16R-hydroxylation and femaleselective catechol formation. J. Steroid Biochem. 42, 65-76. (22) Fishman, J. (1963) Role of 2-hydroxyestrone in estrogen metabolism. J. Clin. Endocrinol. Metab. 23, 207-210. (23) Stubenrauch, G., and Knuppen, R. (1976) Convenient large scale preparation of catechol estrogens. Steroids 28, 733-741. (24) Abul-Hajj, Y. J. (1984) Synthesis of 3,4-estrogen-o-quinones. J. Steroid Biochem. 21, 621-622. (25) NIH Guidelines for the Laboratory Use of Chemical Carcinogens (1981) NIH Publication No. 81-2385, U.S. Government Printing Office, Washington, DC. (26) North, A. C. T., Phillips, D. C., and Mathews, F. S. (1968) A semiempirical method of absorption correction. Acta Crystallogr. Sect. A 24, 351-359. (27) SHELXS 86 (1985) Program for the solution of crystal structures, University of Go¨ttingen, Germany. (28) TEXSAN-TEXRAY (1985) Structure Analysis Package, Molecular Structure Corporation, The Woodlands, Texas. (29) Liehr, J. G., Fang, Wan-Fen, Sirbasku, D. A., and Ari-Ulubelen, A. (1986) Carcinogenicity of catechol estrogens in Syrian hamsters. J. Steroid Biochem. 24, 353-356. (30) Roy, D., Weisz, J., and Liehr, J. G. (1990) The O-methylation of 4-hydroxyestradiol is inhibited by 2-hydroxyestradiol: implications for estrogen-induced carcinogenesis. Carcinogenesis 11, 459-462. (31) Khasnis, D., and Abul-Hajj, Y. J. (1994) Estrogen quinones: Reaction with propylamine. Chem. Res. Toxicol. 7, 68-72. (32) Abul-Hajj, Y. J., Tabakovic, K., and Tabakovic, I. (1995) An estrogen-nucleic acid adduct. Electroreductive intermolecular coupling of 3,4-estrone-o-quinone and adenine. J. Am. Chem. Soc. 117, 6144-6145. (33) Ins, A., Meroni, G., and Ferrara, L. (1977) Two dimers, 4:4′- and 2:2′-di(estradiol), obtained by chemical oxidative coupling of estradiol. J. Steroid Biochem. 8, 1259-1261. (34) Becker, H. D., and Turner, A. B. (1988) Quinones as oxidants and dehydrogenation agents. In The Chemistry of Quinonoid Compounds (Patai, S., and Rappaport, Z., Eds.) Wiley. (35) Iverson, S. L., Hu, L. Q., Vukomanovic, V., and Bolton, J. L. (1995) The influence of the p-alkyl substituents on the isomerization of o-quinones to p-quinone methides: potential bioactivation mechanism for catechols. Chem. Res. Toxicol. 8, 537-544. (36) Ball, P., and Knuppen, R. (1980) Catechol estrogens: Chemistry, biogenesis, metabolism, occurrence and physiological significance. Acta Endocr., Copenh. 132, 1-127. (37) Bolton, J. L. (personal communication). (38) Dwivedy, I., Devanesan, P., Cremonesi, P., Rogan, E., and Cavalieri, E. (1992) Synthesis and characterization of estrogen 2,3- and 3,4-quinones. Comparison of DNA adducts formed by the quinones versus horseradish peroxidase-activated catechol estrogens. Chem. Res. Toxicol. 5, 828-833.
TX950205Z
