![[Feedback]](/icons/banner/feedbackACT.gif){kind=icon}
![[For Subscribers]](/icons/banner/subscriberACT.gif){kind=icon}
![[Archive]](/icons/banner/archiveACT.gif){kind=icon}
![[Search]](/icons/banner/searchACT.gif){kind=icon}
![[Contents]](/icons/banner/tocACT.gif){kind=icon}
|
|
|
|
|
||
| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Institut National de la Santé et de la Recherche Médicale U640, Centre National de la Recherche Scientifique UMR8151, Université Paris Descartes, Faculté de Pharmacie, Laboratoire de Pharmacologie Chimique et Génétique (M.H.P., D.S., G.G.C.) and Plateau Spectrométrie de Masse (IFR 71), Service de Chimie Analytique, Paris, France (N.A., A.R., A.D., M.-C.M.); AP-HP, Hôpital Européen Georges Pompidou, Département de Pharmacologie et de Toxicologie, Paris, France (M.H.P.); Ecole Nationale Supérieure de Chimie, Paris, France (D.S.); and Centre National de la Recherche Scientifique, UMR176, Institut Curie, Centre de Recherche, Paris, France (D.D.).
(Received June 5, 2007; Accepted July 27, 2007)
 |
Abstract |
|---|
 Top Abstract Materials and MethodsResultsDiscussionReferences |
|---|
). Flavonoids have also been reported to exert several anticancer activities that are linked to either cancer prevention and/or therapy (López-Lázaro, 2002).
Among the flavonoids that have been tested in cancer therapy, flavone-8-acetic acid (FAA, 2-phenyl-8-(carboxymethyl)benzopyran-4-one) (Fig. 1) emerged as an interesting lead because it has demonstrated potent in vivo antitumor activity in mouse solid tumors (Corbett et al., 1986; Plowman et al., 1986; Bibby et al., 1988; Pratesi et al., 1988) and also in human tumors transplanted in nude mice (Giavazzi et al., 1988). However, contrary to its potent antitumor activity reported in mice, FAA has not shown anticancer activity in humans (Weiss et al., 1988; Kerr et al., 1989; de Forni et al., 1995). This interspecies difference in FAA anticancer activity has been hypothesized to be, at least partially, due to a possible metabolic activation of this flavonoid in vivo in mice (Capolongo et al., 1987; Chabot et al., 1989).
|
; Capolongo et al., 1987; Schroyens et al., 1987), which are otherwise responsive to FAA when these tumor cells are implanted in mice, several indirect mechanisms of anticancer action have therefore been suggested: augmentation of natural killer activity and induction of interferon (Hornung et al., 1988); antiangiogenic properties (Lindsay et al., 1996); and tumor blood flow shut down and induction of extensive hemorrhagic necrosis (Corbett et al., 1986; Evelhoch et al., 1988; Bibby et al., 1989; Hill et al., 1989; Zwi et al., 1989). At the molecular level, FAA has been reported to induce DNA damage in Glasgow osteogenic sarcoma after in vivo administration, which could indicate the in vivo formation of reactive metabolites in mice (Bissery et al., 1988).
In view of the poor cytotoxic activity of FAA in vitro and its potent anticancer activity in vivo in mice, the requirement for FAA metabolic activation has therefore been hypothesized (Capolongo et al., 1987; Chabot et al., 1989). The aims of the present study were therefore to 1) identify the principal FAA phase 1 metabolites formed by mouse microsomes and 2) to compare the metabolic activity of mouse and human microsomes. We report here the identification of six new FAA metabolites formed by mouse microsomes. In addition we show that human microsomes do not metabolize FAA as efficiently as mouse microsomes. These data indicate a marked interspecies difference in FAA metabolism that could be involved in the differential activity of this molecule observed between the two species. The importance of the newly identified metabolites for the in vivo anticancer activity of FAA remains to be determined.
|
Materials and Methods |
|---|
|
TopAbstractMaterials and MethodsResultsDiscussionReferences |
|---|
). The abbreviations used for the compounds throughout this report and their code names are as follows: flavone-8-acetic acid (FAA, LM975, NSC347512); 2'-hydroxy-FAA (2'-OH-FAA, EMD107818, LM2795); 3'-hydroxy-FAA (3'-OH-FAA, EMD107510, LM2728); 4'-hydroxy-FAA (4'-OH-FAA, EMD107433, LM2711; 3-hydroxy-FAA (3-OH-FAA, EMD106896, LM2594); 5-hydroxy-FAA (5-OH-FAA, LM2798); and 7-hydroxy-FAA (7-OH-FAA, LM2771). Solvents and common chemicals were obtained from commercial sources and were of the highest purity available.
Synthesis of 6-Hydroxyflavone-8-Acetic Acid. The so far unknown 6-OH-FAA (Fig. 1) was prepared according to our previously described methodology using appropriate starting materials (Dauzonne and Demerseman, 1990). The multistep synthesis of this compound is detailed in the Supplemental Data (Dauzonne et al., 1997; Bauvois et al., 2003).
Synthesis of FAA Methyl Ester. Esterification of FAA was accomplished as described previously (Chabot and Gouyette, 1991). Briefly, 400 mg of FAA dissolved in 95% ethanol was added to a solution of potassium hydroxide in water (5 g/8 ml). Diazomethane generated from p-toluenesulfonylmethylnitrosamide (Diazald, Aldrich Chemical Co., Milwaukee, WI) was added as an ethereal solution (21.4 g/130 ml of diethyl ether). The molecular weight of the FAA methyl ester was 294 as determined by fast atom bombardment mass spectrometry (glycerol matrix). The product was analytically pure based on HPLC.
Synthesis of Methyl Ester Derivatives of FAA Monohydroxylated Standards. Three microliters of concentrated sulfuric acid (96%) was added to 100 µl of a methanolic solution of either FAA, 2'-OH-FAA, 3'-OH-FAA, or 4'-OH-FAA at 1.5 µg/ml. The methanolic solution was refluxed for 2 h, cooled, added with 200 µl of water, saturated with NaCl, and extracted with 500 µl of ethyl acetate under agitation for 20 min. The ethyl acetate phase was harvested, washed with 100 µl of 5% NaHCO3, decanted, dried over anhydrous MgSO4, and evaporated under a nitrogen stream. The dry residue was reconstituted with 100 µl of mobile phase (28% methanol, 15% acetonitrile, and 57% aqueous 0.1% TFA, v/v/v), and 50 µl was injected onto the HPLC-UV-MS system.
|
Preparation of Aroclor 1254-Induced Microsomes. Aroclor 1254-induced microsomes were prepared as described previously (Breinholt et al., 2002). Aroclor 1254 was used because this inducer is a potent inducer of several drug-metabolizing enzymes including the major cytochrome P450 families (Easterbrook et al., 2001). In brief, 26 female C57BL/6 mice, 8 weeks of age were purchased from Janvier (Le Genest-St-Isle, France) and acclimated for 1 week in our animal facility and kept on a 12-h light/dark cycle with free access to food. Mice were injected i.p. with Aroclor 1254 (500 mg/kg b.wt., dissolved in corn oil) on day 0. Five days later the mice were sacrificed by cervical dislocation after a 24-h fasting period. All animal experiments complied with the French regulations concerning the protection of animals used for experimental and other scientific purposes (D2001-486), and with the European Commission regulations (OJ of ECL358 Dec. 18, 1986). The hepatic microsomes were prepared as follows: the livers were excised and rinsed in 1.15% ice-cold KCl solution and cut into small pieces with scissors; 4 volumes of 0.1 M phosphate buffer (pH 7.4) were added, and the livers were homogenized using a Teflon potter at a speed of 50 rpm with 10 strokes; the liver homogenate was centrifuged at 15,000g for 20 min, and, after aspiration of the upper lipidic layer, the supernatant was ultracentrifuged at 100,000g during 60 min, at 4°C; the pellets containing the microsomes were collected, covered with a phosphate buffer containing 20% glycerol, snap-frozen on dry ice, and stored at -80°C until use. Total microsomal protein was determined using the Pierce BCA Protein Assay Reagent Kit (Pierce Chemical, Rockford, IL).
Optimization of Incubation Conditions. To optimize the incubation conditions for FAA metabolism, we first used a protocol described previously (Breinholt et al., 2002). However, the various incubation parameters had to be optimized with regard to FAA and microsomal protein concentrations, NADPH alone or a NADPH-regenerating system (BD Biosciences, San Jose, CA), incubation time, buffer type, shaking speed, opened or closed vials, and presence or absence of bovine serum albumin, vitamin C, or MgCl2. After a 1-h incubation time at 37°C the mixture (300 µl) was extracted twice with 800 µl of ethyl acetate containing 2% acetic acid (v/v). The organic phase was dried under a gentle nitrogen steam, the dry residue was reconstituted with 300 µl of the HPLC mobile phase (described below), and 100 µl was injected onto the HPLC-UV system described thereafter. It was found that the optimal conditions yielding the maximum formation of metabolites were as follows: FAA concentration 8 µg/ml, microsomal protein 0.5 mg/ml, an NADPH-generating system corresponding to 1.3 mM, in 300 µl of 0.1 M Tris-HCl, pH 7.4, containing 1.6 mg/ml bovine serum albumin, and 5 mM MgCl2. The incubation was performed in opened 2-ml polypropylene tubes in a shaking water bath (150 oscillations/min) at 37°C. The reaction mixture was prewarmed at 37°C for 3 min before the NADPH-generating system was added, and the incubation was extended for an additional 60 min.
Epoxide Hydrolase Incubation. The presence of possible FAA epoxides was assessed by using recombinant epoxide hydrolase from Aspergillus niger (Sigma-Aldrich, St. Louis, MO) at 0.5 mg of protein/ml (0.85 U) in the optimal incubation conditions described above with Aroclor-induced microsomes and FAA at 8 µg/ml. After the 1-h incubation time, the mixture was extracted with ethyl acetate, dried under nitrogen, and reconstituted with 300 µl of the HPLC mobile phase (26% methanol, 19% acetonitrile, and 55% aqueous acetic acid 2%, v/v/v), and 100 µl was injected onto the Shimadzu HPLC-UV system described below.
Acidic Treatment of Potential Epoxides. To determine the exact position of the oxygen on the putative arene oxide, the metabolite was treated under soft acidic conditions, which are known to yield the corresponding monohydroxylated derivative(s) after aromatization, as described previously (Coombs et al., 1981). In brief, the suspected epoxide metabolite was incubated with TFA (final concentration 0.5 M) at 37°C for 30 min, and the resulting products were analyzed by HPLC-UV-MS and compared with available monohydroxylated standards of FAA.
Dehydration of Potential Dihydrodiol. To determine the presence of a dihydrodiol, the suspected metabolite fraction was dehydrated as described previously (Coombs et al., 1981). In brief, the dry residue corresponding to 25 microsomal incubations was dissolved in 100 µl of methanol, added with 100 µl of 2.5 M sulfuric acid, and heated at 80°C for 2 h. After cooling, 200 µl of water were added, and the mixture was saturated with NaCl, and extracted with 500 µl of ethyl acetate under continuous agitation for 20 min. The organic phase was washed with 100 µl of 5% NaHCO3, decanted, and evaporated to dryness under a nitrogen stream. The dry residue was reconstituted in 70 µl of mobile phase (28% methanol, 15% acetonitrile, and 57% aqueous 0.1% TFA, v/v/v), and 50 µl was injected onto the HPLC-UV-MS system described below. In these dehydration conditions, two monohydroxylated compounds were identified as methyl ester derivatives on the carboxylic acid function. For identification purposes, FAA and its monohydroxylated derivatives on the 2'-, 3'-, or 4'-positions were also submitted to identical strong acidic conditions to yield their corresponding methyl ester derivatives.
HPLC-UV and HPLC-UV-MS Conditions. Two HPLC systems were used during this study. The first HPLC-UV system included Shimadzu CLASS-VP (software version 5.3) and was equipped with a C18 column (Beckman Ultrasphere ODS, 5 µm; 4.6 x 25 mm) thermostated at 20°C and a UV detector set at 300 nm. The mobile phase consisted of a solution of 26% methanol, 19% acetonitrile, and 55% aqueous acetic acid (2%), v/v/v, at a flow rate of 1 ml/min. This HPLC system was used to collect the metabolic fractions. To compare the quantitative yield of metabolites from different microsomal preparations, we first determined the area under the curve of each new metabolite peak expressed in microvolts per second. In our conditions, the limit of detection corresponding to 3 times the baseline noise, was calculated as a peak area 
360 µV x s.
To further characterize the collected metabolic fractions, a second HPLC system was used. This system was composed of a Surveyor high-performance liquid chromatograph coupled with a UV diode array detector and a mass spectrometer (LCQ-Advantage; Thermo Electron Corporation, Waltham, MA). The mass spectrometer consisted of an electrospray ionization system used in positive ion polarity mode and an ion trap mass analyzer. The operating parameters were as follows: the spray needle voltage was set at 4.0 kV and the spray was stabilized with a nitrogen sheath gas. Electrospray capillary voltage was 4.0 V, nitrogen was used as the auxiliary gas, and the capillary temperature was 250°C. To obtain ion masses and to propose a plausible scheme of fragmentation for each compound studied, two scan modes were used. The MS full scan mode was used to obtain the parent ion masses, and the MS/MS scan mode in full scan was used to fragment the parent ions into their product ions. The fragmentation was accomplished by collision-induced dissociation with a collision energy between 30 and 40% of maximum energy, and helium was used as the buffer gas. A C18 reversed-phase column (EC 250/2 Nucleodur Sphinx RP, 5 µm, 250 mm x 2.1 mm; Macherey-Nagel, Hoerdt, France) thermostated at 30°C was used, and the mobile phase was composed of 28% methanol, 15% acetonitrile, and 57% aqueous 0.1% TFA, v/v/v, at a flow rate of 0.2 ml/min. The initial mobile phase composition was maintained for 14 min and then was changed to 35% methanol, 19% acetonitrile, and 46% aqueous 0.1% TFA, during the next 2 min, and maintained in these conditions for an additional 5 min. Xcalibur software was used for data acquisition and analysis.
Data Presentation and Statistics. Results are presented as means ± S.E.M. The total metabolite areas (microvolts x second) are arbitrary values, because each metabolite has a different UV extinction coefficient. To compare the metabolic activity of the various microsomal preparations, the sum of the metabolite areas was used and compared with the sum of metabolites of uninduced mouse microsomes using Dunnett's t test. Comparison between groups was considered significant if p < 0.05.
|
Results |
|---|
|
TopAbstractMaterials and MethodsResultsDiscussionReferences |
|---|
|
|
|
, and references therein). The main FAA fragmentation patterns are presented in Fig. 3C. The MS analysis of all possible monohydroxylated FAA derivatives indicated that the three ions at m/z 179, 161, and 133 were indicative of monohydroxylation on the B or the C ring only (i.e., 2'-, 3'-, 4'-, and 3-OH-FAA). For the FAA derivatives hydroxylated on the A ring (i.e., 5-, 6-, and 7-OH-FAA), the ions at m/z 195, 177, and 149, (i.e., +16 amu) were observed (data not shown).
|
|
). M4 presented a molecular ion m/z 297 [M + H]+ (Fig. 4C) indicating the presence of an oxygen atom (+16 amu) on the FAA molecule (FAA m/z 281 [M + H]+). The MS fragmentation pattern of the M4 molecular ion m/z 297 presented two product ions at m/z 251 and 149 (Fig. 4C). The standard 6-OH-FAA also presented a molecular ion at m/z 297 with a MS spectrum exhibiting product ions m/z 251 and 149 corresponding to [M + H - H2CO2]+ and [1,3A+-H2CO2], respectively (Fig. 4D). As stated above, the ion m/z 149 is specifically observed with FAA derivatives hydroxylated on the A ring. Because of its Rt and UV and MS spectra similar to those of standard 6-OH-FAA, metabolite M4 was therefore assigned to the structure of the 6-OH-FAA (Fig. 4).
|
Identification of Metabolite 3B (M3B). The M3B peak was also identified by comparison with the authentic standard 3'-OH-FAA by comparing Rts (data not shown), UV spectra (Fig. 6, A and B), and MS fragmentation patterns (Fig. 6, C and D).
Identification of Metabolite 3C (M3C). Metabolite M3C eluted with a Rt of 11.2 min (Fig. 2B). This metabolite had a UV spectrum with two maxima at 315 nm (band I) and 250 nm (band II) (Fig. 7A). The positions of bands I and II were similar to those observed above for FAA derivatives hydroxylated on the B ring, which could suggest a modification on the same ring. The MS analysis of M3C yielded a molecular ion m/z 297 [M + H]+, indicating the presence of an oxygen atom on the FAA molecule (i.e., +16 amu). In addition, the MS fragmentation pattern of the M3C molecular ion yielded three product ions at m/z 179, 161, and 133, indicating that the oxygen atom was not on the A ring (Fig. 7B).
Because this mono-oxygenated metabolite did not coelute with any of the monohydroxylated standards of FAA, an epoxidation of FAA was therefore hypothesized, probably on the B ring, as suggested by the UV spectrum and the MS fragmentation pattern. To verify the presence of an epoxide, this metabolite was then submitted to recombinant epoxide hydrolase from A. niger. As shown in Fig. 8, this enzyme caused the almost complete disappearance of the M3C peak (Fig. 8). As a matter of fact, the M3C peak area was reduced by 96% in presence of the enzyme within a 1-h incubation period (compare the M3C peak on Fig. 8, A and B). To determine the exact position of the epoxidic oxygen, we next submitted the metabolite M3C to soft acidic conditions (TFA), because arene oxide exposure to such conditions is known to yield the corresponding phenolic compounds by protonation (Coombs et al., 1981). In those conditions, M3C was readily transformed into two hydroxylated derivatives of FAA that were identified as the 3'-OH-FAA (23%) and the 4'-OH-FAA (77%), as determined by coelution with authentic standards and by comparing their UV and MS spectra (data not shown). On the basis of M3C metabolite UV and MS data, its aromatization to 3'-OH-FAA and 4'-OH-FAA under soft acidic conditions, and its sensitivity to epoxide hydrolase, this metabolite was therefore assigned to the structure of the 3',4'-epoxy-FAA (Fig. 7C).
|
|
We next submitted M2 to hydrolysis by A. niger recombinant epoxide hydrolase, and it was found that the M2 peak area was decreased by 55% within a 1-h incubation time (Fig. 8, A and B), thus indicating that it was an FAA epoxide. To further determine the exact position of the epoxidic oxygen on M2, we submitted this metabolite to soft acidic conditions (TFA) to protonate the epoxide and to induce aromatization into the corresponding phenolic compounds (Coombs et al., 1981). Under such acidic conditions, metabolite M2 yielded two compounds that presented identical Rts to the standards 5-OH-FAA (16%) and 6-OH-FAA (84%) (data not shown). In addition, the UV and MS spectra of the M2 products after acidification were superimposable to the ones obtained with standards of 5-OH-FAA (data not shown) and 6-OH-FAA (cf. Fig. 4B and 4D). On the basis of its UV and MS spectra, its acidic aromatization to 5-OH-FAA and 6-OH-FAA, and its sensitivity to epoxide hydrolase, M2 was therefore assigned to the structure of the 5,6-epoxy-FAA (Fig. 9C).
Identification of Metabolite 1 (M1). This metabolite was the most hydrophilic and eluted with a short Rt of 4.1 min (Fig. 2B). The UV spectrum of M1 exhibited a band I at 315 nm and a more intense band II at 245 nm, suggesting a B ring modification of FAA (Fig. 10A). M1 did not match any B ring monohydroxylated standards of FAA (i.e., 2'-, 3'-, or 4'-OH-FAA) in terms of Rts or UV and MS spectra. The MS analysis of M1 presented a molecular ion m/z 315 [M + H]+; i.e., a gain of 34 amu over the parent compound FAA, which was suggestive of the addition of two hydrogen atoms and two oxygen atoms. The fragmentation pattern of the molecular ion m/z 315 yielded six product ions (Fig. 10B). The first three ions at m/z 297, 279, and 251 corresponded to the loss of one water molecule [M + H - H2O]+, two water molecules [M + H - 2H2O]+, and the loss of a water molecule plus a formic acid molecule [M + H - H2O- H2CO2]+, respectively. Loss of two water molecules (-36 amu) substantiated the hypothesis that M1 was a dihydrodiol. For the ions smaller than m/z 251, i.e., m/z 179, 161, and 133, we found the typical fragmentation pattern indicating that the A ring was not modified, as suspected above by the UV spectrum bands. Therefore, because of its UV and MS spectra, metabolite M1 was most likely a dihydrodiol of FAA on the B ring.
|
). Because it was found that these dehydration conditions also allowed the formation of the methyl ester of the carboxylic function of FAA in the presence of methanol, we therefore compared the results of M1 dehydration with the methyl esters of the B ring monohydroxylated FAA derivatives (i.e., methyl esters of 2'-, 3'-, and 4'-OH-FAA). In these conditions, it was observed that the dehydration of M1 yielded two compounds with Rts and UV and MS spectra identical to the methyl ester derivatives of 3'- and 4'-OH-FAA. On the basis of these results, metabolite M1 was therefore assigned to the structure of the 3',4'-dihydrodiol-FAA (Fig. 10C). FAA Metabolism by Human Microsomes. After the identification of the principal FAA metabolites formed with mouse microsomes, we were interested in investigating the metabolism of FAA with human microsomes. To do so, we tested two different pools of human microsomes and also 15 microsomal preparations from individual donors. For comparison purposes, the average metabolite peak areas formed by uninduced and Aroclor-induced mouse microsomes are also presented in Table 1. For the human microsomes, one pool composed of 29 donors (pool A) could not metabolize FAA (Table 1), whereas another pool of 20 donors (pool B) yielded only a small metabolite peak corresponding to the Rt of the 5,6-epoxy-FAA (Table 1). For comparison purposes, it was noteworthy that uninduced mouse microsomes formed more metabolites than either pool of human microsomes with a total metabolite area of 22,698 µVxsec, whereas pool A did not yield any detectable metabolite, and pool B yielded total areas of only 1154, which were 20-fold lower than those of uninduced mouse microsomes (Table 1).
Because only one metabolite of FAA was barely detected using the above human microsomal pools, we were next interested in testing human microsomes from individual donors to assess interindividual variation. As presented in Table 1, 6 of 15 individual preparations did not yield any detectable metabolite peak (samples 1-6). Seven microsomal preparations (samples 7-13) could produce a small peak corresponding to the Rt of the 5,6-epoxy-FAA. Only one human microsome could yield a small peak corresponding to the Rt of the 4'-OH-FAA (sample 14). Finally, only one microsomal preparation (sample 15) could clearly produce three metabolites corresponding to the Rt of the 3',4'-dihydrodiol-FAA, the 4'-OH-FAA, and the 3',4'-epoxy-FAA.
With regard to the quantitative analysis of these results, it can be noted that the individual human microsomes that did produce some metabolite(s) were doing so in very small yields (Table 1) compared with uninduced mouse microsomes. In the same experimental conditions, uninduced mouse microsomes yielded a total area of metabolite production that was 12- to 32-fold higher than that of human microsomes, with the exception of one microsomal preparation (sample 15) that could produce a meaningful metabolite yield that was, however, 3-fold lower than that for the uninduced mouse microsomes.
|
Discussion |
|---|
|
TopAbstractMaterials and MethodsResultsDiscussionReferences |
|---|
; Chabot et al., 1989), the FAA metabolic pathways in this species have not yet been elucidated. The only mouse metabolite identified so far is the acyl glucuronide of FAA (Cummings et al., 1989; Chabot and Gouyette, 1991), a metabolite that has also been identified in humans, but has not shown cytotoxic activity in vitro (Cummings et al., 1989). In the present study, we were interested in elucidating the principal phase 1 metabolic pathways of FAA using mouse microsomes. Because the metabolic profile obtained with uninduced mouse microsomes was identical to that obtained with Aroclor-induced microsomes, we therefore used the latter preparation to obtain a sufficient quantity of metabolites to allow their identification. Aroclor 1254 is a known broad inducer of several different cytochromes P450 including the subfamilies CYP1A, CYP2B, and CYP3A, and can increase their metabolic capacity toward planar compounds such as the polycyclic aromatic hydrocarbons and flavones (Nielsen et al., 1998; Easterbrook et al., 2001). In this study, we have shown that FAA can indeed be metabolized in vitro using Aroclor-induced mouse microsomes into six previously undescribed metabolites. In addition, we have shown that FAA metabolism is qualitatively and quantitatively more important in murine microsomes compared with human microsomes.
|
). It is however interesting that beside the 3',4'-dihydrodiol-FAA possessing a saturated 3',4' bond, we did not observe the formation of a 3',4'-dihydroxylation on the intact B ring (catechol), as has been observed with some B ring-unsubstituted flavones (Nielsen et al., 1998; Nikolic and Van Breemen, 2004).
The detection and identification of two epoxides, one on the B ring (3',4'-epoxy-FAA, M3C) and another one on the A ring of FAA (5,6-epoxy-FAA, M2) was rather surprising, given the fact that such metabolites are often unstable and readily undergo chemical degradation during the incubation period and/or in the process of extraction. In fact, these epoxides were relatively stable because they could not only sustain a 1-h incubation time at physiological conditions (pH 7.4, 37°C) but also resist the extraction procedures at room temperature, both conditions not considered favorable for epoxide stability. To our knowledge, such stable epoxides have rarely been described for other flavones, perhaps with the exception of the 5,6-epoxy-
-naphthoflavone which is also chemically stable, apparently because of the adjacent C-4 carbonyl group (Vyas et al., 1983).
Although the sensitivity of the two epoxides to epoxide hydrolase can be considered as a rather convincing proof of the epoxidic character of these metabolites (Guengerich, 2003), a significant difference in their respective sensitivity to this enzyme was noted, as the 3',4'-epoxy-FAA was a better substrate than the 5,6-epoxy-FAA. It is noteworthy that another flavone epoxide at the same 5,6-position, i.e., the 5,6-epoxy--naphthoflavone, has also been found to be quite resistant to epoxide hydrolase, probably because of the proximate carbonyl group at the C-4 position of the molecule which could cause steric hindrance of the epoxidic function to the enzyme catalytic site (Vyas et al., 1983).
Another intriguing observation about the two FAA epoxides is that we did not see the formation of their corresponding dihydrodiols after treatment with A. niger epoxide hydrolase, in conditions under which the epoxides were obviously hydrolyzed. This lack of formation could be due to the fact that epoxide hydrolase converts their substrates via the intermediate formation of a covalent enzyme-substrate complex, and it has been shown that the formation of the intermediate can proceed much faster than the subsequent hydrolysis, i.e., the formation of the terminal dihydrodiol products. Thus, the enzyme would act like a molecular sponge by binding and inactivating the epoxide very fast, whereas the subsequent release of the dihydrodiol product is considerably slower (Oesch et al., 2001). However, it should be noted that the observed formation of the 3',4'-dihydrodiol-FAA (M1) during the incubation with mouse microsomes indicates that this preparation contained mouse epoxide hydrolase that could release the 3',4'-dihydrodiol-FAA. In contrast, we did not detect 5,6-dihydrodiol-FAA probably because the mouse epoxide hydrolase could not hydrolyze 5,6-epoxy-FAA or could not release the product of the reaction.
Although metabolism frequently results in the formation of more hydrophilic metabolites usually possessing a lower biological activity than that of the parent compound, there are several examples in which metabolism can lead to the formation of more active compounds (Breinholt et al., 2002). The FAA phase 1 metabolites identified in this study were all more hydrophilic than the parent compound and are therefore expected to be more easily excreted. However, concerning the potential biological activity of the identified FAA metabolites, it is difficult to predict which one(s) could be involved (if any) in the in vivo activity of this compound. In the flavonoid family, it is often observed that minor differences in structure can profoundly influence biological activity (Nikolic and Van Breemen, 2004). It is also of interest that several flavonoid metabolites have been reported to exert antiproliferative activity in vitro (Meng et al., 2006), thereby underlining the importance of the identification of these metabolites. Concerning the two relatively stable epoxides identified in this study, it is possible that these compounds could be involved in the antitumor activity, considering that DNA strand breaks were observed in tumors after administration of FAA in mice (Bissery et al., 1988). Indeed, it has been reported that epoxidic compounds can cause DNA damage through covalent binding (Xue and Warshawsky, 2005, and references therein).
In addition to substantial differences in FAA phase 1 metabolism between mice and humans, shown in this study, differences in phase 2 metabolism is also possible. FAA metabolism has already been looked at in humans and in mice, and the only phase 2 metabolite identified so far in both species is the acyl glucuronide of FAA (Cummings et al., 1989; Chabot and Gouyette, 1991). The six new FAA phase 1 metabolites identified in this study will undoubtedly facilitate the identification of other phase 2 metabolites potentially formed in mice. The identification of these metabolites is of importance, because they could be involved in the remarkable anticancer activity observed in mice and perhaps explain the lack of activity in humans.
In conclusion, we have identified six new FAA metabolites using mouse microsomes, the species in which this compound was shown to be a potent anticancer agent. We also have shown that human microsomes could not metabolize FAA as efficiently as mouse microsomes. This marked interspecies difference in metabolism could be involved, as least in part, in the differential anticancer activity observed between the two species. The implications of the these new metabolites, either unconjugated or conjugated (Williamson et al., 2005), in FAA anticancer activity remains to be investigated.
|
Footnotes |
|---|
M.H.P. and N.A. contributed equally to this work.
ABBREVIATIONS: FAA, flavone-8-acetic acid; 2'-OH-FAA, 2'-hydroxy-FAA; 3'-OH-FAA, 3'-hydroxy-FAA; 4'-OH-FAA, 4'-hydroxy-FAA; 3-OH-FAA, 3-hydroxy-FAA; 5-OH-FAA, 5-hydroxy-FAA; 7-OH-FAA, 7-hydroxy-FAA; 6-OH-FAA, 6-hydroxy-FAA; 3',4'-epoxy-FAA, 3',4'-dihydro-3',4'-epoxy-FAA; 3',4'-dihydrodiol-FAA, 3',4'-dihydro-3',4'-dihydroxy-FAA; 5,6-epoxy-FAA, 5,6-dihydro-5,6-epoxy-FAA; 5,6-dihydrodiol-FAA, 5,6-dihydro-5,6-dihydroxy-FAA; TFA, trifluoroacetic acid; HPLC, high-performance liquid chromatography; MS, mass spectrometry; NADPH, β-nicotinamide adenine dinucleotide 2'-phosphate reduced; MS/MS, mass fragmentation pattern of molecular ion; Rt, retention time; RP, reversed-phase; amu, atomic mass units.
The online version of this article (available at http://dmd.aspetjournals.org) contains supplemental material. 
Address correspondence to: Dr. Guy G. Chabot, Laboratoire de pharmacologie chimique et génétique, INSERM U640 CNRS UMR8151, Faculté de pharmacie, Université Paris Descartes, 4 avenue de l'Observatoire, Paris, F-75006 France. E-mail: guy.chabot{at}univ-paris5.fr
|
References |
|---|
|
TopAbstractMaterials and MethodsResultsDiscussionReferences |
|---|
-H2AX. Cancer Res 66: 9656-9664.-naphthoflavone and β-naphthoflavone by rat liver microsomes and highly purified reconstituted cytochrome P-450 systems. J Biol Chem 258: 5649-5659.
| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
 |
 |
 |
|
 |
 |
 |
