Ellipticine is an antineoplastic agent, the mode of action of which is considered to be based on DNA intercalation and inhibition of topoisomerase II. We found that ellipticine also forms the cytochrome P450 (CYP)-mediated covalent DNA adducts. We now identified the ellipticine metabolites formed by human CYPs and elucidated the metabolites responsible for DNA binding. The 7-hydroxyellipticine, 9-hydroxyellipticine, 12-hydroxyellipticine, 13-hydroxyellipticine, and ellipticine N2-oxide are generated by hepatic microsomes from eight human donors. The role of specific CYPs in the oxidation of ellipticine and the role of the ellipticine metabolites in the formation of DNA adducts were investigated by correlating the levels of metabolites formed in each microsomal sample with CYP activities and with the levels of the ellipticine-derived deoxyguanosine adducts in DNA. On the basis of this analysis, formation of 9-hydroxyellipticine and 7-hydroxyellipticine was attributable to CYP1A1/2, whereas production of 13-hydroxyellipticine and ellipticine N2-oxide, the metabolites responsible for formation of two major DNA adducts, was attributable to CYP3A4. Using recombinant human enzymes, oxidation of ellipticine to 9-hydroxyellipticine and 7-hydroxyellipticine by CYP1A1/2 and to 13-hydroxyellipticine and N2-oxide by CYP3A4 was corroborated. Homologue modeling and docking of ellipticine to the CYP3A4 active center was used to explain the predominance of ellipticine oxidation by CYP3A4 to 13-hydroxyellipticine and N2-oxide.

Ellipticine (5,11-dimethyl-6H-pyrido[4,3-b]carbazole), an alkaloid isolated from Apocyanaceae plants, and several of its more soluble derivatives exhibit significant antitumor and anti-HIV activities (1). The main reason for the interest in ellipticine and its derivatives for clinical purposes is their high efficiencies against several types of cancer and their rather limited toxic side effects (2). Nevertheless, ellipticine is a potent mutagen (for an overview, see ref. 1).

Ellipticine is an antineoplastic agent, the mode of action of which was considered to be based mainly on DNA intercalation and/or inhibition of topoisomerase II (1, 2). We showed that ellipticine also covalently binds to DNA after being enzymatically activated (1, 3, 4, 5). Using human recombinant cytochrome P450 (CYP) enzymes, CYP3A4, 1A1, and 1B1–those enzymes expressed in tumors sensitive to ellipticine (i.e., breast cancer; ref. 6)—were found to be the most efficient CYPs activating ellipticine to form covalent DNA adducts (1). Deoxyguanosine was identified to be the target for CYP-mediated ellipticine binding (4, 5). The formation of these adducts was also detected in V79 lung fibroblast cells transfected with human CYP3A4, 1A1, and 1A2 (3), in human breast adenocarcinoma MCF-7 cells (7), and in vivo in rats exposed to ellipticine (5). On the basis of these data, ellipticine might be considered a drug for which pharmacological efficiency and/or genotoxic side effects are dependent on its enzymatic activation in target tissues (1, 3, 4, 5, 7).

Although the target deoxynucleotide for ellipticine binding to DNA has already been identified, the metabolites responsible for such binding have not yet been characterized. In rats in vivo and in microsomes, two ellipticine metabolites, 9-hydroxyellipticine and 7-hydroxyellipticine, were found (2). Rat CYP1A1 and CYP1A2 are assumed to be the most active enzymes oxidizing ellipticine to these metabolites (2, 8). Only 9-hydroxyellipticine was formed in human liver microsomes (8).

To date, it is not known which CYPs oxidize ellipticine in humans. Nevertheless, this knowledge is important to evaluate the pharmacological efficacy and/or the genotoxic side effects of ellipticine in humans. Therefore, the present study was undertaken to resolve four principal problems: (1) to determine the capability of human CYP enzymes to oxidize ellipticine; (2) to identify which human CYPs are involved in this oxidation; (3) to characterize ellipticine metabolites formed by CYP; and (4) to identify which ellipticine reactive species form deoxyguanosine adducts in DNA. This knowledge will be useful to predict human susceptibility to ellipticine.

Chemicals and Reagents.

Ellipticine, NADP+ and NADPH were from Sigma Chemical Co. (St. Louis, MO), and 9-hydroxyellipticine was from Calbiochem (San Diego, CA). The 7-hydroxyellipticine and ellipticine N2-oxide were synthesized as described previously (9, 10). Enzymes and chemicals for the 32P-postlabeling assay were obtained from sources described previously (1, 3, 4, 5). Supersomes, microsomes isolated from insect cells transfected with Baculovirus constructs containing cDNA of human CYPs and expressing NADPH:CYP reductase were from Gentest Corp. (Woburn, MA). Microsomes from livers of eight human donors who died after traffic accidents were isolated and characterized as described previously (4, 11).

Incubations.

Incubations used for study of the ellipticine metabolism contained in a final volume of 500 μL [50 mmol/L potassium phosphate buffer (pH 7.4), 1 mmol/L NADP+, 10 mmol/L d-glucose 6-phosphate, 1 unit/mL d-glucose 6-phosphate dehydrogenase, 10 mmol/L MgCl2, microsomes containing 0.4 μmol/L CYP, and 10 μmol/L ellipticine dissolved in 5 μL methanol]. Incubations, in which the efficiencies of Supersomal CYPs were tested, were the same except that 5 μmol/L ellipticine and 0.02 to 0.1 μmol/L CYP were used. After incubation (37°C, 20 minutes) the reaction was stopped by adding 100 μL of 2 mol/L NaOH. Five microliters of 1 mmol/L phenacetine in methanol was added as an internal standard, and the ellipticine metabolites were extracted with ethyl acetate. The evaporated extracts were dissolved in 50 μL of methanol, and ellipticine metabolites were separated by high-performance liquid chromatography (HPLC). The column used was a 5-μm Ultrasphere ODS (4.6 × 250 mm; Beckman, Fullerton, CA), the eluent was 64% methanol plus 36% of 5 mmol/L heptane sulfonic acid containing 32 mmol/L acetic acid in water with flow rate of 0.8 mL/minute, and detection was at 296 nm. Five ellipticine metabolites with the retention times of 6.3, 6.9, 7.8, 8.5, and 11.2 minutes were separated. Fractions containing these metabolites were collected from multiple HPLC runs, concentrated, and analyzed by mass spectroscopy and/or nuclear magnetic resonance (NMR).

Incubations in which ellipticine-DNA adducts were analyzed contained in a final volume of 750 μL [50 mmol/L potassium phosphate buffer (pH 7.4), 1 mmol/L NADPH, microsomes containing 0.1 to 1 nmol CYP, 50 to 100 μmol/L ellipticine dissolved in 7.5 μL methanol, and 1 mg DNA]. Incubations in which 13-hydroxyellipticine and ellipticine N2- oxide collected from multiple HPLC runs were used instead of ellipticine contained in a final volume of 500 μL [50 mmol/L potassium phosphate buffer (pH 7.4) and 5 to 10 μmol/L 13-hydroxyellipticine or ellipticine N2-oxide dissolved in 5 μL of methanol and 1 mg DNA]. After 60 minutes (37°C), incubations were extracted with ethyl acetate. DNA was isolated by the phenol/chloroform extraction (1, 4) and used for evaluation of DNA adduct formation by the 32P-postlabeling assay. Liver DNA modified by ellipticine in vivo, by treating rats with ellipticine (5), was also used. The 32P-postlabeling analyses were done with nuclease P1 enrichment, the procedure found to be appropriate to detect and quantify ellipticine-derived DNA adducts (1, 3, 4, 5, 7).

High-Performance Liquid Chromatography Analysis of 32P-Labeled Adducts.

Adduct spots detected by the 32P-postlabeling assay in DNA incubated with 13-hydroxyellipticine and N2-oxide of ellipticine and those in DNA from liver of rats treated with ellipticine (5) were excised from thin-layer plates and extracted (3, 4, 5). The dried extracts were dissolved in 100 μL of methanol/phosphate buffer (pH 3.5) 1:1 (v/v) and analyzed on a phenyl-modified reversed-phase column (3, 4, 5, 7).

Mass and Nuclear Magnetic Resonance Spectrometry.

Electron spray ionization mass spectra were recorded on a Bruker Esquire quadrupole ion trap mass spectrometers (Bruker GmbH, Bremen, Germany). NMR spectra were measured on Varian UNITY Inova 400 MHz instrument (400 MHz for 1H and 100.58 MHz for 13C). Samples were measured in deuterated DMSO or methanol at 25°C and 35°C. Shigemi microtubes were used for the measurement of small quantities of metabolites (150 μL). As an internal standard the signal of the solvent was used (DMSO, δ = 2.50 for 1H, δ = 39.5 for 13C-NMR spectra; and methanol, δ = 3.31 for 1H, δ = 49.1 for 13C-NMR spectra). Chemical shifts (δ, ppm) and coupling constants (Hz) were obtained by the first order analysis. COSY, NOESY, gHSQC, and gHMBC spectra were recorded for all of the samples. COSY spectra were measured in absolute value mode with standard two-pulse sequence. NOESY spectra with mixing time of 0.3 seconds were taken as phase sensitive with standard three-pulse sequence. Connectivities C-H were obtained from HSQC and HMBC experiments, which were done as gradient experiments. All of the two-dimensional experiments were measured in spectral windows 5,000 Hz for proton and 25,000 Hz for carbon resonances.

Molecular Modeling.

The procedure consisted of the following steps: (1) homology modeling of the CYP3A4 enzyme; (2) additional improvement of the model with molecular dynamics; (3) calculation of ellipticine structure and atomic charges; and (4) docking of ellipticine into enzyme model.

The homology modeling was started by sequence alignment, with the CP34_HUMAN sequence (accession number P08684, Swiss-Prot Release 42.0; refs. 12, 13). (This sequence was truncated and renumbered; and 28 NH2-terminal and 6 COOH-terminal residues were omitted.) In additional work, the CYP2B4 structure known from the X-ray crystallography (Protein Data Bank code 1PO5; refs. 14, 15) was used as the template.

Alignment of the CYP3A4 and template was done in CLUSTALX (16) with the Gonnet 250 scoring matrix, and value of the gap-opening penalty was set to 10.00 of the gap-extension penalty to 0.20. Comparative modeling was then done with MODELLER (17) with the predefined routine “model.” Ten models were generated. Quality of the resulting models was checked by PROCHECK (18). The model with the lowest number of residues in disallowed regions, lowest value of probability density function, and low G-factor value was chosen. Finally, the heme was inserted into this model structure (heme moiety from the CYP2B4 template, overlaid with the oxoferryl heme from CYP101, PDB code 1DZ9).

The structure was improved with force field computations. The protonated structure was built in CHARMM (19, 20) and solvated in SYBYL 6.9.1 with the XFIT and Molecular Silverware algorithms. The resulting system (protein in a water droplet, overall hydration shell thickness 12 Å) consisting of 81,587 atoms (24,622 water molecules, 7,721 atoms in the protein) was energy-minimized in CHARMM (adopted basis Newton-Raphson method). In this and the subsequent steps, the following constraints were applied: the outer 6 Å of water hydration shell, the oxoferryl heme (porphyrine ring, iron, and oxygen atoms) were fixed, soft restrains were applied to the distances from thiolate to the NH-backbone groups of the three residues after the coordinating cysteine [values according to Shaik et al. (21)].

The molecular dynamics simulation (372 ps) consisted of heating from 0 to 310 K (24 ps), followed by equilibration (140 ps) at 310 K, new heating (20 ps) to 500 K, 20 ps of equilibration at 500 K, and simulated annealing (168 ps) from 500 K to 190 K. The calculation was done as simple Newtonian constant energy dynamics. The annealed structure was again energy-minimized (adopted basis Newton-Raphson method) and its stereochemical parameters evaluated by PROCHECK.

The ellipticine structure was optimized with the DFT/B3LYP/6–31G* calculation (in HyperChem) and docked onto the enzyme active site with the AutoDock (22). The aromatic ring atoms were hold fixed (root part), whereas the methyl groups were allowed to rotate (branches). The explicit water molecules were deleted from the annealed model, polar hydrogens added, and partial charges assigned. This structure was used for docking (number of grid points in all directions 120, spacing 0.18 Å, grid center 4.8 Å above the heme iron atom), the resulting docked structures were ranked in clusters (ligand position within 1.0 Å root-mean-square-deviation).

Statistical Analyses.

The correlation coefficients were based on a sample size of 8 and calculated with version 6.12. Statistical Analysis System software. All Ps are two-tailed and considered significant at the 0.05 level.

Oxidation of Ellipticine by Human Hepatic Microsomes.

When ellipticine was incubated with human hepatic microsomes, five metabolites were separated by HPLC. On the basis of cochromatography with synthetic standards, mass and NMR spectroscopy, the structures of all five metabolites were identified. In the positive-ion electrospray mass-spectrum, all of the metabolites showed the protonated molecule at m/z 263.0, indicating the molecular mass of hydroxylated (oxygenated) derivatives of ellipticine. Products eluting with retention times of 6.3 and 8.5 minutes were identified by cochromatography with synthetic standards to be 9-hydroxyellipticine and 7-hydroxyellipticine, respectively. The two ellipticine metabolites eluting at retention times of 6.9 and 7.8 minutes were identified by NMR spectroscopy to be the derivatives of ellipticine with a hydroxylated methyl group, 12-hydroxyellipticine and 13-hydroxyellipticine, respectively. Structure assignments of 12-hydroxyellipticine were based on 1H NMR, COSY, and poor gHMBC spectra. Signals of all seven aromatic protons are present in spectra measured in both solvents (methanol and DMSO) from which it is evident that this metabolite has no phenolic hydroxyl group. Broad singlet in 1H NMR spectra at δ 5.60 is indicative of hydroxymethyl group. Two and three bond C-H connectivities are visible in gHMBC spectrum only for protons of methyl and hydroxymethyl groups at δ 2.83 [to carbons at δ 110.0 (C-5), 132.0 (C-4a), and 140.3 (C-5a)] and 5.60 [to carbons at δ 121.4 (C-11a), 123.7 (C-10b), and 130.1 (C-11)].

The structure of 13-hydroxyellipticine was again based on 1H NMR, COSY, and poor gHSQC and gHMBC spectra. Signals of all seven aromatic protons are visible in both solvents. Signal of one methyl group was not found; on the other hand, a new signal at δ 5.41 with one-bond connectivity to the carbon C-13 at δ 57.2 is present. These signals are characteristic for hydroxymethyl group on the aromatic ring. In gHMBC spectrum, there are clearly visible connectivities from hydrogens in 12-methyl group to carbons at δ 123 (C-11a), 126 (C-10b), and 132.5 (C-11) and from methylene hydrogens of hydroxymethyl group to carbons at δ 112 (C-5), 136.5 (C-4a), and 143 (C-5a). All of these data are in agreement only with the structure of 13-hydroxyellipticine.

The fifth metabolite eluting after 11.2 minutes exhibited the same chromatographic properties as ellipticine N2-oxide, and NMR data of this metabolite confirmed this structure. As in the case of other metabolites, the structure of this metabolite was based on 1H NMR and COSY spectra because of a very small quantity of the sample. The assignment of H-2 and H-3 follows from coupling constants: 1H-1H coupling constant 4J through nitrogen atom is for N-oxides 1.4 to 1.8 Hz (in our case 1.8 Hz) and 5J about 0.4 Hz. The only substantial changes in the chemical shifts are those for protons in the positions 1 and 3, on carbons adjacent to the nitrogen atom, caused by changes of electron distribution in the aromatic ring. Both protons are shifted upfield, which is together with observed coupling constants indicative for formation of the salt or N-oxide in the position 2.

To resolve which CYP enzymes are responsible for the oxidation of ellipticine, two approaches were used: (1) correlation of CYP-linked activities in human hepatic microsomes with the amounts of ellipticine metabolites generated by the same microsomes; and (2) use of heterologous baculovirus expression systems of human CYPs (Supersomes).

Correlation of CYP-Linked Enzyme Activities in Human Microsomes with Ellipticine Oxidation.

Human hepatic microsomes from eight donors used in the study catalyzed reactions known to be associated with specific CYPs (Table 1). All microsomal samples oxidized also ellipticine (Table 1).

The formation of 9-hydroxyellipticine was highly correlated with ethoxyresorufin O-deethylation, a marker for CYP1A (r = 0.822, P < 0.01), and a correlation tendency was seen between this CYP1A activity and formation of 7-hydroxyellipticine (r = 0.642).

The formation of 13-hydroxyellipticine was highly correlated with 6β-hydroxylation of testosterone, a marker for CYP3A4 (r = 0.910, P < 0.001). A significant correlation was also seen between production of ellipticine N2-oxide and 6β-hydroxylation of testosterone (r = 0.802, P < 0.05).

Oxidation of Ellipticine by Human Recombinant CYP Enzymes.

Using Supersomes containing recombinant human CYPs and NADPH:CYP reductase corroborated the results obtained with human hepatic microsomes, except for very effective N2-oxidation by CYP2D6 in this system (Fig. 1). Human recombinant CYP1A1 and 1A2 were the major enzymes oxidizing ellipticine to 9-hydroxyellipticine and 7-hydroxyellipticine. CYP1B1 and 2D6 were also efficient to catalyze these reactions but to a lower extent. The 12-hydroxyellipticine was generated by CYP3A4 and 2C9 but only in very low levels (data not shown). The major enzyme oxidizing ellipticine to 13-hydroxyellipticine is human recombinant CYP3A4, followed by CYP1A2, 2D6, and 2C9. CYP3A4 also forms ellipticine N2-oxide, but human recombinant CYP2D6 is much more effective in generating this product (Fig. 1). Because CYP3A4 is the most abundant CYP in human liver (∼30%; ref. 23), oxidation of ellipticine to 13-hydroxyellipticine and N2-oxide should be the major metabolic pathway of the drug in human livers. Indeed, the predominant ellipticine metabolite formed by human hepatic microsomes is 13-hydroxyellipticine followed by N2-oxide (Table 1). Whereas only low levels of 12-hydroxyellipticine were formed by human recombinant CYPs, the levels of this metabolite produced by microsomes are in the range of 9-hydroxyellipticine (Table 1). Testing the stability of N2-oxide under the experimental conditions used for oxidation of ellipticine with human hepatic microsomes, we found it to be unstable and to decompose to produce 12-hydroxyellipticine. Therefore, 12-hydroxyellipticine found in human hepatic microsomes is formed also nonenzymatically. A Polonovsky-type rearrangement (24) of ellipticine N2-oxide might be the mechanism of this reaction.

Molecular Modeling.

To examine the molecular basis of the ellipticine oxidation by the most prominent CYP enzyme in human liver, CYP3A4, computer modeling was used. By docking ellipticine into the model structure of oxo-ferryl CYP3A4 optimized by molecular-dynamics we found four clusters of populated binding modes (Fig. 2). The most populated cluster A does not have any atom in closer proximity to the oxoferryl iron, and thus the ellipticine molecule bound in this manner cannot be hydroxylated. From the point of view of the enzyme reaction, the binding mode cluster B is the most substantial. Here the hydrogens of the C-13 methyl atom and of the nitrogen N-6 are located in a close distance and represent a likely target for hydroxylation. One of these positions in the ellipticine molecule is really hydroxylated by CYP3A4; 13-hydroxyellipticine is the metabolite formed by CYP3A4 (Table 1; Fig. 1). The binding modes C and D are considerably less populated, and they bring the oxygen close to the hydrogens of C-12 methyl atom, followed by C-10 or C-3 and the nitrogen N-2, respectively. Two of these positions were found to be oxidized by CYP3A4, N-2, and C-12, because ellipticine N2-oxide and 12-hydroxyellipticine were generated by CYP3A4 (Fig. 1; Table 1). The docked structure B is the only one showing the ellipticine molecule oriented roughly parallel to the heme (see Fig. 3 details).

For a comparison, ellipticine was docked also to CYP2B4 (the template molecule used for homology modeling). The most populated binding modes for this enzyme suggest that hydroxylation by CYP2B4 might occur predominantly in positions N-2 and C-3 and also to a lesser extent on carbon atoms in positions 9 and 8 (Fig. 2).

Formation of Ellipticine-DNA Adducts Is Mediated by 13-Hydroxyellipticine and Ellipticine N2-Oxide.

The used human hepatic microsomes also formed ellipticine-derived DNA adducts (Fig. 4). The formation of 13-hydroxyellipticine was highly correlated with levels of the major deoxyguanosine-ellipticine adduct 1 in DNA formed in human hepatic microsomes (r = 0.942, P < 0.001). Lower but significant correlation was also found between formation of this metabolite and ellipticine-deoxyguanosine adduct 2 (r = 0.912, P < 0.01). The formation of this adduct highly significantly correlated with amounts of ellipticine N2-oxide produced by human microsomes (r = 0.920, P < 0.01), whereas a lower correlation was observed between levels of ellipticine-DNA adduct 1 and formation of this metabolite (r = 0.810, P < 0.05). These results suggest that 13-hydroxyellipticine and ellipticine N2-oxide, either themselves or intermediates, generated during the CYP3A4-mediated oxidation of ellipticine might be responsible for the formation of the ellipticine-DNA adducts.

To resolve this problem, 13-hydroxyellipticine and ellipticine N2-oxide were isolated from incubations and used for further studies. Both metabolites generated ellipticine-DNA adducts. As shown in Fig. 4,C, incubations of DNA with 13-hydroxyellipticine produced one major adduct migrating similarly to deoxyguanosine-ellipticine adduct 1 formed in DNA by ellipticine activated with human hepatic microsomes (Fig. 4,B) and/or in liver DNA of rats treated with ellipticine (Fig. 4,A). Ellipticine N2-oxide formed a major adduct behaving chromatographically similar to the ellipticine-DNA adduct 2 (Fig. 4 D).

The major adduct spots formed by 13-hydroxyellipticine and ellipticine N2-oxide (spots 1 and 2 in Fig. 4, C and D, respectively) and the major DNA adducts generated by ellipticine in rat liver (Fig. 4,A) were excised, extracted, and analyzed by cochromatography on reversed-phase HPLC. These experiments showed that both major adducts formed in DNA by 13-hydroxyellipticine and ellipticine N2-oxide were indistinguishable from those obtained in the in vivo experiments. The adduct formed in DNA by 13-hydroxyellipticine (spot 1 in Fig. 4,C) eluted with a retention time of 11.77 minutes, corresponding to the retention time of 11.78 minutes of deoxyguanosine adduct spot 1 in DNA formed in rat liver. When equal amounts of radioactivity of these adduct spots were mixed before analysis, a single peak was found. Adduct spot 2 of Fig. 4 D generated by ellipticine N2-oxide produced a major radioactive peak (retention time of 8.88 minutes), corresponding to adduct spot 2 formed by ellipticine in rat liver, eluting with a retention time of 8.88 minutes.

Here we present for the first time the detailed characterization of ellipticine metabolites produced by individual human CYPs present in human hepatic microsomes. The study also identifies the ellipticine metabolites responsible for the formation of ellipticine-deoxyguanosine adducts in DNA (Fig. 5).

Beside 9-hydroxyellipticine, which was found previously to be formed in human hepatic microsomes (2, 8), we identified four additional ellipticine metabolites generated in this human enzymatic system; although 13-hydroxyellipticine, ellipticine N2-oxide, and 12-hydroxyellipticine are the major metabolites, 7-hydroxyellipticine is a minor metabolite formed in human liver. The CYP enzyme predominantly expressed in human livers, CYP3A4, oxidizes ellipticine mainly to 13-hydroxyellipticine and ellipticine N2-oxide. The importance of CYP3A4 for the generation of these metabolites was also confirmed with human recombinant CYP3A4. Molecular modeling where ellipticine was docked to the active site of the human CYP3A4 model explains partly the formation of these ellipticine metabolites by CYP3A4. Human recombinant CYP1A1 and 1A2 oxidize ellipticine to 9-hydroxyellipticine and 7-hydroxyellipticine, whereas 13-hydroxyellipticine is a minor metabolite produced by CYP1A2. Whereas recombinant CYP2D6 efficiently oxidized ellipticine to the N2-oxide, it is not important for N2-oxidation in human hepatic microsomes. Low levels of CYP2D6 expression in human livers (23) might be the reason for these results.

One of the most important results of our present work is the identification of the ellipticine metabolites responsible for the formation of DNA adducts. The present study shows that the major ellipticine-deoxyguanosine adduct 1 in DNA is generated from 13-hydroxyellipticine, whereas the adduct 2 is formed from N2-oxide. Therefore, the CYP enzymes generating these metabolites should also be responsible for the formation of the ellipticine-derived DNA adducts. Indeed, we showed that the enzymes oxidizing ellipticine to these metabolites (CYP3A4 and CYP1A) activate ellipticine to form DNA adducts (1, 3, 4, 5). Their expression in human tissues might therefore determine the pharmacological efficiencies or the genotoxic side effects of ellipticine.

Although the ellipticine derivatives responsible for covalent modification of DNA by ellipticine were clearly established, the exact reactive species as well as the positions in guanine where these species are bound remain to be elucidated. Now we can only speculate on the structure of the adduct generated by the ellipticine metabolite, 13-hydroxyellipticine, with deoxyguanosine in DNA. This ellipticine derivative might, depending on the environment, decompose spontaneously to the reactive carbenium ion, which reacts with one of the nucleophilic centers in deoxyguanosine in DNA.

It is not possible to show if the antitumor, cytostatic, and/or genotoxic activities of ellipticine are related to only one or several of the DNA damage effects (1, 2, 3). Nevertheless, for ellipticine antitumor activity to cancer cells, mutagenicity caused by DNA adducts, might be relevant. Rekha and Sladek (25) showed that antineoplastic activity of ellipticine to MCF-7 cells depends on the levels of CYP enzymes activating ellipticine to DNA-binding species. These authors showed that MCF-7 cells treated with 3-methylcholanthrene transiently expressed elevated levels of CYP1A, and cells were transiently much more sensitive to ellipticine. CYP1A also activate ellipticine to species binding to DNA (1, 3, 4, 5), and in an earlier study, we found the typical ellipticine-DNA adducts in these cells (7). The CYP-dependent higher sensitivity of MCF-7 cells to ellipticine observed by these authors might, therefore, be explained by ellipticine-DNA adduct formation.

Taken together, the activities and expression levels of CYPs, which activate ellipticine to metabolites forming DNA adducts, may be important factors in the specificity of ellipticine for breast cancer. Nevertheless, to confirm this suggestion, formation of ellipticine-DNA adducts in breast cancer tissues in vivo remains to be evaluated. Preliminary results indicate that DNA adducts are detectable not only in healthy organs of rats exposed to ellipticine (5) but also in the target tissue for the treatment, mammary tumors.

Our study can form the basis for additional chemical characterization of the ellipticine-DNA adducts. In addition, it is the prerequisite for biomonitoring studies in humans.

Fig. 1.

Oxidation of ellipticine by human recombinant CYPs. Ten to 50 pmol human recombinant CYP/incubation and 10 μmol/L ellipticine were used in all of the experiments. Values of ellipticine metabolites are averages of triplicate incubations. SDs were ≤10%.

Fig. 1.

Oxidation of ellipticine by human recombinant CYPs. Ten to 50 pmol human recombinant CYP/incubation and 10 μmol/L ellipticine were used in all of the experiments. Values of ellipticine metabolites are averages of triplicate incubations. SDs were ≤10%.

Close modal
Fig. 2.

Representation of the most populated docked binding modes of ellipticine to CYP3A4 (clusters A–D) and CYP2B4 (clusters 1–5). For each structure, the substrate and the heme moiety are shown (note the different orientations). Structures are arranged according to decreasing population.

Fig. 2.

Representation of the most populated docked binding modes of ellipticine to CYP3A4 (clusters A–D) and CYP2B4 (clusters 1–5). For each structure, the substrate and the heme moiety are shown (note the different orientations). Structures are arranged according to decreasing population.

Close modal
Fig. 3.

Active site of CYP3A4 with ellipticine docked in the binding mode B. Shown are the substrate, heme, and the side chains of Thr308, Glu307, Ile368, and Leu372 (clockwise from the heme; numbering of the whole CYP3A4 sequence).

Fig. 3.

Active site of CYP3A4 with ellipticine docked in the binding mode B. Shown are the substrate, heme, and the side chains of Thr308, Glu307, Ile368, and Leu372 (clockwise from the heme; numbering of the whole CYP3A4 sequence).

Close modal
Fig. 4.

Autoradiographs of polyethyleneimine-cellulose TLC maps of 32P-labeled digests of DNA isolated from liver of rats treated with ellipticine (A), calf thymus DNA reacted with ellipticine, NADPH and human liver microsomes containing 1 nmol CYP (sample H5; B), with 13-hydroxyellipticine (C), and ellipticine N2-oxide (D). Film exposure was 1 hour for A and 30 minutes for B–D at −80°C.

Fig. 4.

Autoradiographs of polyethyleneimine-cellulose TLC maps of 32P-labeled digests of DNA isolated from liver of rats treated with ellipticine (A), calf thymus DNA reacted with ellipticine, NADPH and human liver microsomes containing 1 nmol CYP (sample H5; B), with 13-hydroxyellipticine (C), and ellipticine N2-oxide (D). Film exposure was 1 hour for A and 30 minutes for B–D at −80°C.

Close modal
Fig. 5.

Metabolism of ellipticine by human CYPs showing the metabolites and those leading to DNA adducts.

Fig. 5.

Metabolism of ellipticine by human CYPs showing the metabolites and those leading to DNA adducts.

Close modal

Grant support: Grant Agency of the Czech Republic (Grant 203/05/2186) and the Ministry of Industry of the Czech Republic (Grant FD-K/096).

The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

Requests for reprints: Marie Stiborová, Department of Biochemistry, Faculty of Science, Charles University, Albertov 2030, 128 40 Prague 2, Czech Republic. Phone: 420-2-2195-1285; Fax: 420-2-2195-1283; E-mail: stiborov@natur.cuni.cz

Table 1

CYP-dependent catalytic activities, amounts of ellipticine metabolites and DNA adducts formed by ellipticine in human hepatic microsomes

Human hepatic microsomal samplespmol CYP per mg proteinEROD (CYP1A1/2) *Coumarin 7-hydroxylation (CYP2A6) *Tolbutamide methyl hydroxylation (CYP2C9) *Bufuralol 1′-hydroxylation (CYP2D6) *Chlorzoxazone 6-hydroxylation (CYP2E1) *Testosterone 6β-hydroxylation (CYP3A4) *
H1 80 5.65 0.93 0.76 2.46 1.80 8.44 
H2 90 4.09 1.42 0.06 0.34 0.69 2.22 
H3 270 6.34 2.03 0.18 5.55 2.09 6.31 
H4 60 7.19 0.40 0.06 3.71 1.83 14.39 
H5 220 10.82 0.84 0.17 2.53 2.45 10.19 
H6 140 11.73 0.98 0.26 0.92 2.61 6.94 
H7 460 6.86 0.43 0.11 4.11 1.37 3.96 
H8 400 12.08 0.58 0.15 3.13 1.51 6.86 
Average value § 215 (141.8) 8.095 (2.83) 0.95 (0.51) 0.219 (0.214) 2.84 (1.58) 1.79 (0.577) 7.41 (3.50) 
Human hepatic microsomal samplespmol CYP per mg proteinEROD (CYP1A1/2) *Coumarin 7-hydroxylation (CYP2A6) *Tolbutamide methyl hydroxylation (CYP2C9) *Bufuralol 1′-hydroxylation (CYP2D6) *Chlorzoxazone 6-hydroxylation (CYP2E1) *Testosterone 6β-hydroxylation (CYP3A4) *
H1 80 5.65 0.93 0.76 2.46 1.80 8.44 
H2 90 4.09 1.42 0.06 0.34 0.69 2.22 
H3 270 6.34 2.03 0.18 5.55 2.09 6.31 
H4 60 7.19 0.40 0.06 3.71 1.83 14.39 
H5 220 10.82 0.84 0.17 2.53 2.45 10.19 
H6 140 11.73 0.98 0.26 0.92 2.61 6.94 
H7 460 6.86 0.43 0.11 4.11 1.37 3.96 
H8 400 12.08 0.58 0.15 3.13 1.51 6.86 
Average value § 215 (141.8) 8.095 (2.83) 0.95 (0.51) 0.219 (0.214) 2.84 (1.58) 1.79 (0.577) 7.41 (3.50) 

NOTE. All results are presented as means of duplicate experiments. Assays for CYP activities were carried out as described previously (11).

Abbreviations: EROD, 7-ethoxyresorufin O-deethylation; n.d., not detectable (<0.001 pmol ellipticine metabolites/min/nmol CYP).

*

CYP activities in nmol/min/nmol CYP, except for EROD activity, which is in pmol/min/nmol CYP.

The pmol ellipticine metabolites/min/nmol CYP.

Relative adduct labeling (RAL)/107nucleotides per nmol CYP.

§

Arithmetic means for eight hepatic microsomal samples (H1-H8), values in parentheses are SDs, representing the interindividual variability.

Table 1A

Continued

Ellipticine metabolites (total) 9-OH-ellipticine 12-OH-ellipticine 13-OH-ellipticine 7-OH-ellipticine N2-oxide of ellipticine Ellipticine-DNA adduct 1 Ellipticine-DNA adduct 2
0.310 0.02 0.026 0.150 0.011 0.101 6.35 0.47 
0.261 0.041 0.108 0.092 0.02 n.d. § 3.91 0.14 
0.355 0.053 0.047 0.120 0.03 0.105 6.33 0.25 
1.236 0.083 0.113 0.681 0.042 0.317 19.1 1.10 
0.50 0.062 0.072 0.296 0.03 0.04 12.2 0.27 
0.501 0.114 0.099 0.155 0.04 0.093 9.68 0.08 
0.237 0.035 0.072 0.099 n.d. 0.031 3.39 0.013 
0.445 0.140 0.073 0.123 0.07 0.039 4.96 0.07 
0.481 (0.301) 0.069 (0.038) 0.076 (0.028) 0.215 (0.186) 0.030 (0.020) 0.091 (0.093) 8.24 (4.95) 0.299 (0.332) 
Ellipticine metabolites (total) 9-OH-ellipticine 12-OH-ellipticine 13-OH-ellipticine 7-OH-ellipticine N2-oxide of ellipticine Ellipticine-DNA adduct 1 Ellipticine-DNA adduct 2
0.310 0.02 0.026 0.150 0.011 0.101 6.35 0.47 
0.261 0.041 0.108 0.092 0.02 n.d. § 3.91 0.14 
0.355 0.053 0.047 0.120 0.03 0.105 6.33 0.25 
1.236 0.083 0.113 0.681 0.042 0.317 19.1 1.10 
0.50 0.062 0.072 0.296 0.03 0.04 12.2 0.27 
0.501 0.114 0.099 0.155 0.04 0.093 9.68 0.08 
0.237 0.035 0.072 0.099 n.d. 0.031 3.39 0.013 
0.445 0.140 0.073 0.123 0.07 0.039 4.96 0.07 
0.481 (0.301) 0.069 (0.038) 0.076 (0.028) 0.215 (0.186) 0.030 (0.020) 0.091 (0.093) 8.24 (4.95) 0.299 (0.332) 
1
Stiborová M, Bieler CA, Wiessler M, Frei E The anticancer agent ellipticine on activation by cytochrome P450 forms covalent DNA adducts.
Biochem Pharmacol
2001
;
62
:
1675
-84.
2
Auclair C Multimodal action of antitumor agents on DNA: The ellipticine series.
Arch Biochem Biophys
1987
;
259
:
1
-14.
3
Frei E, Bieler CA, Arlt VM, Wiessler M, Stiborová M Covalent binding of the anticancer drug ellipticine to DNA in V79 cells transfected with human cytochrome P450 enzymes.
Biochem Pharmacol
2002
;
64
:
289
-95.
4
Stiborová M, Stiborová-Rupertová M, Bořek-Dohalská L, Wiessler M, Frei E Rat microsomes activating the anticancer drug ellipticine to species covalently binding to deoxyguanosine in DNA are a suitable model mimicking ellipticine bioactivation in humans.
Chem Res Toxicol
2003
;
16
:
38
-47.
5
Stiborová M, Breuer A, Aimová D, et al DNA adduct formation by the anticancer drug ellipticine in rats determined by 32P-postlabeling.
Int J Cancer
2003
;
107
:
885
-90.
6
Murray GI, Weaver RJ, Paterson PJ, et al Expression of xenobiotic metabolizing enzymes in breast cancer.
J Pathol
1993
;
169
:
347
-53.
7
Bořek-Dohalská L, Frei E, Stiborová M DNA adduct formation by the anticancer drug ellipticine and its hydroxy derivatives in human breast adenocarcinoma MCF-7 cells.
Collect Czech Chem Commun
2004
;
69
:
603
-15.
8
Lesca P, Beaune P, Monsarrat B Ellipticines and human liver microsomes: spectral interaction with cytochrome P-450 and hydroxylation. Inhibition of arylhydrocarbon metabolism and mutagenicity.
Chem-Biol Interact
1981
;
36
:
299
-309.
9
Wijsmuller WFA, Wanner MJ, Koonen GI, Pandit UK Pyridocarbazole alkaloids. Synthesis of olivacine and ellipticine analogues.
Heterocycles
1986
;
24
:
1795
-7.
10
Boogaard AT, Pandit UK, Koomen GJ Ring-D modifications of ellipticine. 2. Chlorination of ellipticine via its N-oxide and synthesis and selective acetylation of 5,6,11-trimethyl-5H benzo[B]carbazole.
Tetrahedron
1994
;
50
:
4811
-28.
11
Stiborová M, Martínek V, Rýdlová H, Hodek P, Frei E Sudan I is a potential carcinogen for humans: evidence for its metabolic activation and detoxication by human recombinant cytochrome P450 1A1 and liver microsomes.
Cancer Res
2002
;
62
:
5678
-84.
12
Boeckmann B, Bairoch A, Apweiler R, et al The SWISS-PROT protein knowledgebase and its supplement TrEMBL in 2003.
Nucleic Acids Res
2003
;
31
:
365
-70.
13
Schwede T, Kopp J, Guex N, Peitsch MC SWISS-MODEL: an automated protein homology-modeling server.
Nucleic Acids Res
2993
;
31
:
3381
-5.
14
Berman HM, Westbrook J, Feng Z, et al The protein data bank.
Nucleic Acids Res
2000
;
28
:
235
-42.
15
Scott EE, He YA, Wester MR, et al An open conformation of mammalian cytochrome P450 2B4 at 1.6 A resolution.
Proc Nat Acad Sci USA
2003
;
100
:
13196
-201.
16
Thompson JD, Gibson TJ, Plewniak F, Jeanmougin F, Higgins DG The CLUSTAL_X windows interface: flexible strategies for multiple sequence alignment aided by quality analysis tools.
Nucleic Acids Res
1997
;
25
:
4876
-82.
17
Sali A, Blundell TL Comparative protein modeling by satisfaction of spatial restraints.
J Mol Biol
1993
;
234
:
779
-815.
18
Laskowski RA, MacArthur MW, Moss DS, Thornton JM PROCHECK: a program to check the stereochemical quality of protein structures.
J Appl Cryst
1993
;
26
:
283
-91.
19
Brooks BR, Burccoleri RE, Olafson BD, States DJ, Karplus M CHARMM: a program for macromolecular energy, minimization, and dynamics calculations.
J Comput Chem
1983
;
4
:
187
-217.
20
MacKerell AD, Jr, Bashford D, Bellott M, et al All-atom empirical potential for molecular modeling and dynamics studies of proteins.
J Phys Chem B
1998
;
102
:
3586
-616.
21
Schoneboom JC, Lin H, Reuter N, et al The elusive oxidant species of cytochrome P450 enzymes: characterization by combined quantum mechanical/molecular mechanical (QM/MM) calculations.
J Am Chem Soc
2002
;
124
:
8142
-51.
22
Morris GM, Goodsell DS, Halliday RS, et al Automated docking using a Lamarckian genetic algorithm and an empirical binding free energy function.
J Comput Chem
1998
;
19
:
1639
-62.
23
Rendic S, DiCarlo FJ Human cytochrome P450 enzymes: a status report summarizing their reactions, substrates, inducers, and inhibitors.
Drug Metab Rev
1997
;
29
:
413
-80.
24
Hofle G, Glase N, Leibold T, Sefkow M Epothilone A-D and their thiazole-modified analogs as novel anticancer agents.
Pure Appl Chem
1999
;
71
:
2019
-24.
25
Rekha GK, Sladek NE Multienzyme-mediated stable and transient multidrug resistance and collateral sensitivity induced by xenobiotics.
Cancer Chemother Pharmacol
1997
;
40
:
215
-24.