Human Glutathione S-Transferase-Mediated Glutathione Conjugation

In the present study, the structures of both glutathione conjugates of ... Wageningen University/TNO Centre for Food Toxicology. | Nestlé Research Ce...
0 downloads 0 Views 143KB Size
Download PDF
Chem. Res. Toxicol. 2007, 20, 1895–1902

1895

Human Glutathione S-Transferase-Mediated Glutathione Conjugation of Curcumin and Efflux of These Conjugates in Caco-2 Cells Mustafa Usta,*,† Heleen M. Wortelboer,† Jacques Vervoort,‡ Marelle G. Boersma,‡ Ivonne M. C. M. Rietjens,‡,§ Peter J. van Bladeren,§,| and Nicole H. P. Cnubben†,§ TNO Quality of Life, P.O. Box 360, 3700 AJ Zeist, The Netherlands, DiVision of Toxicology, Wageningen UniVersity, P.O. Box 8000, 6700 EA Wageningen, The Netherlands, Wageningen UniVersity/TNO Centre for Food Toxicology, P.O. Box 8000, 6700 EA Wageningen, The Netherlands, and Nestlé Research Centre, P.O. Box 44, CH-1000 Lausanne 26, Switzerland ReceiVed June 21, 2007

Curcumin, an R,β-unsaturated carbonyl compound, reacts with glutathione, leading to the formation of two monoglutathionyl curcumin conjugates. In the present study, the structures of both glutathione conjugates of curcumin were identified by LC-MS and one- and two-dimensional 1H NMR analysis, and their formation in incubations with human intestinal and liver cytosol and purified human glutathione S-transferases and also in human Caco-2 cells was characterized. The results obtained demonstrate the site for glutathione conjugation to be the C1 atom, leading to two diastereoisomeric monoglutathionyl curcumin conjugates (CURSG-1 and CURSG-2). The formation of both glutathionyl conjugates appeared to be reversible. The monoglutathionyl curcumin conjugates decompose with a t1/2 of about 4 h to curcumin and other unidentified degradation products. Both human intestinal and liver cytosol catalyzed curcumin glutathione conjugation. At saturating substrate concentrations, human GSTM1a-1a and GSTA1-1 are shown to be especially active in the formation of CURSG-1, whereas GSTP1-1 and GSTA2-2 have no preference for the formation of CURSG-1 or CURSG-2. GSTT1-1 hardly catalyzes the glutathione conjugation of curcumin. In the Caco-2 human intestinal monolayer transwell model, CURSG-1 and CURSG-2 were formed at a ratio of about 2:1 followed by their excretion, which appeared to be three times higher to the apical (lumen) side than to the basolateral (blood) side. Given that GSTM1a-1a and GSTP1-1 are present in the intestinal epithelial cells, it can be concluded that efficient glutathione conjugation of curcumin may already occur in the enterocytes, followed by an efficient excretion of these glutathione conjugates to the lumen, thereby reducing the bioavailability of (unconjugated) curcumin. In conclusion, the present study identifies the nature of the diastereoisomeric monoglutathionyl curcumin conjugates, CURSG-1 and CURSG-2 formed in biological systems, and reveals that conjugate formation is catalyzed by GSTM1a-1a, GSTA1-1, and/or GSTP1-1 with different stereoselective preference. The formation of glutathione conjugates can already occur during intestinal transport, after which the monoglutathionyl conjugates are efficiently excreted to the intestinal lumen, thereby influencing the bioavailability of curcumin and, as a result, its beneficial biological effects. Introduction Curcumin [1,7-bis(4-hydroxy-3-methoxyphenyl)-1,6-heptadiene-3,5-dione] (CUR)1 is the major yellow pigment extracted from turmuric, a commonly used spice and coloring agent, derived from the rhizome of the herb Curcuma longa. CUR possesses many beneficial biological activities, such as antioxidant, antibactericidal, antiinflammatory, and anticarcinogenic activities (1–3). Because of the beneficial health claims, the use of CUR as a functional food ingredient or as a pharmacological * To whom correspondence should be addressed. Tel: +31-30-6944790. Fax: +31-30-6944989. E-mail: [email protected]. † TNO Quality of Life. ‡ Division of Toxicology, Wageningen University. § Wageningen University/TNO Centre for Food Toxicology. | Nestlé Research Centre. 1 Abbreviations: CUR, curcumin; CURSG, glutathionyl curcumin conjugate; GSTA1-1 and GSTA2-2, R class glutathione S-transferase; GSTM1a1a, µ class glutathione S-transferase; GSTP1-1, π class glutathione S-transferase; GSTT1-1, θ class glutathione S-transferase; GSH, reduced glutathione; HBSS, Hanks’ balanced salt solution; MRP, human multidrug resistance protein; 1H NMR, proton nuclear magnetic resonance.

compound in clinical trials to prevent or even to treat several diseases has gained great interest. CUR is known to minimize oxidative damage partly through free radical scavenging (4, 5). CUR induces glutathione (GSH) levels and glutathione Stransferases (GSTs) and inhibits NFκ-B activation and interleukin 8 release and as such contributes to antiinflammatory and anticancer activities (6, 7). CUR also protects against the development of colon tumors in rats treated with carcinogens such as azoxymethane, due to inhibition of cell proliferation and induction of apoptosis (8–11). Moreover, in cancer-derived cell lines, for example, breast cancer cells, prostate cancer cells, and leukemia cells, CUR inhibited cell proliferation and increased cell apoptosis (12–16). Recently, it was shown that a low dose of CUR could have a beneficial effect in the treatment of Alzheimer’s disease, as CUR blocks the aggregation and fibril formation of the small β-amyloid species (17). CUR has been shown to have a distinct dose-related behavior, as at low doses CUR is known to be an antioxidant and GST inducer, whereas CUR at higher doses acts as a prooxidant and has been shown to suppress GST activities, a virtue that may

10.1021/tx7002245 CCC: $37.00  2007 American Chemical Society Published on Web 11/01/2007

1896 Chem. Res. Toxicol., Vol. 20, No. 12, 2007

Usta et al.

Figure 1. Typical reversed-phase HPLC chromatogram of an incubation of CUR with purified human GSTM1a-1a (65 µg/mL). The experimental conditions are as described in the Material and Methods. Identified peaks are indicated as CURSG-1, CURSG-2, and CUR. M3 is an unidentified metabolite peak, and the peaks with an asterisk (*) indicate the CUR-related peaks, which were also present in blank incubations without GSH.

be of high importance in cell apoptosis and treatment of cancer cells (18, 19). Given the fact that most studies investigated the effects of the unconjugated form of CUR but also that CUR contains two electrophilic R,β-unsaturated carbonyl groups, which can react with GSH via a Michael addition (20–22), it is of importance to better define the metabolism of CUR in mammalian systems. It is known that R,β-unsaturated carbonyl compounds, such as PGA1, PGA2, and ethacrynic acid, are substrates for GST enzymes, and their GSH conjugates were identified (23, 24). For CUR, the reaction kinetics with GSH both nonenzymatically as well as catalyzed by GSTP1-1 was studied by Awasthi et al. (25). The GSH adducts formed were described as mono- and diglutathionyl-CUR adducts as well as cyclic rearrangement products of feruloylmethylketone and feruloylaldehyde based on mass spectrometry data. However, no exact identification of the CUR-GSH products could be made. In a previous study (26), in MDCKII cells, transfected with the human multidrug resistance protein 1 and 2 (MRP1 and MRP2), the CUR-GSH metabolites formed were identified as protonated monoglutathionyl CUR adducts by LC-MS. In this previous study, it was shown that the monoglutathionyl CUR conjugates can be exported from the cells by MRP1 and MRP2. The transporter proteins MRP1 and MRP2 are known to excrete the conjugates of many compounds. MRP1 is localized at the basolateral side, whereas MRP2 is localized at the apical side of the cell. Of the MRPs, especially MRP2 plays an important role in the overall absorption of compounds in the intestine, limiting the bioavailability of MRP substrates by transporting them back into the intestinal lumen. CUR is known as a potential anticancer agent, and in this respect, the compound has been used in clinical trails (27). The bioavailability of CUR, however, appeared to be low as measured in the plasma, although higher levels of CUR were measured in the intestine and liver. In the study of Ireson et al. (28), the bioavailability of CUR was investigated, and the results suggested that CUR is metabolized into CUR glucuronide, CUR sulfate, tetrahydrocurcumin, and hexahydrocurcumin in the intestinal tract. In spite of these previous studies, several characteristic aspects of the GSH conjugation of CUR have not been described, and these include the actual identification of the structure of the GSH conjugates formed and of the human GSTs involved in conjugate formation. Also, the preferential

direction of CUR-GSH conjugate excretion from intestinal cells to either the basolateral (blood) or the apical (lumen) side has so far not been characterized. Therefore, the objectives of the present study were to identify the structure of the two monoglutathionyl CUR conjugates, to investigate the role of human GST in the formation of CUR GSH conjugates, and to study the directional excretion of these GSH conjugates from intestinal Caco-2 cells, to gain further insight into the relation between the various cellular GSH-related processes with respect to the metabolic fate of CUR.

Materials and Methods Materials. CUR (>99% CUR) was obtained from Kordia (Leiden, The Netherlands). [35S]GSH was purchased from Perkin Elmer (Boston, MA). Human intestine cytosols and human hepatic cytosols were obtained from Biopredic International (Rennes, France). Caco-2 HTB-37 was obtained from the American Type Culture Collection (ATCC) (Manassas, VA). HPLC-grade trifluoroacetic acid was obtained from Baker (Deventer, The Netherlands). HPLC-grade methanol was from Merck (Darmstadt, Germany). All other chemicals were of the highest available quality. Synthesis, Isolation, and Stability of Glutathionyl CUR Conjugates (CURSGs). To elucidate the structure of the presumed CURSGs, both conjugates were synthesized as follows: GSH (10 mM) was dissolved in 10 mM HEPES adjusted to pH 8.5. CUR (50 µL of a 10 mM stock solution in DMSO) was added every 30 min (for five times), and the mixture was incubated at 37 °C in the dark. To monitor the reaction mixture, reversed-phase HPLC (Agilent Technologies System) was used. The reverse phase HPLC analysis was performed using an Alltima C18 column (4.6 mm × 150 mm, 5 µm, Alltech, IL) with a flow rate of 1 mL/min using 0.1% (v/v) trifluoroacetic acid in water (solvent A) and acetonitrile (solvent B). The solvent program started isocratically with 0% B for 5 min, followed by a linear gradient to 29% B in 6 min, thereafter an isocratic elution at 29% B for 20 min, followed by a linear gradient to 95% B in 7 min, and equilibration with an isocratic elution at 0% B for 12 min. Detection of metabolites was performed using a Hewlett Packard series 1100 UV detector (200–500 nm). The individual CURSG peaks (Figure 1) with retention times (tR) of 22.6 (CURSG-1) and 23.3 min (CURSG-2) were collected, the eluent was evaporated using a BUCHI RE111 Rotavapor, and the individual CURSGs were analyzed by 1H NMR and LC-MS. The stability of the monoglutathionyl CUR conjugates was determined as follows: Madin Darby canine kidney cells transfected

Glutathione Conjugates of Curcumin with MRP2 transporter protein were exposed to CUR, and the released CUR conjugates (CURSG-1 and CURSG-2) into the medium were collected. In brief, the MDCKII MRP2 cells were cultured to confluence on 24 mm diameter transwells (Costar Corp., Cambridge, MA). Thereafter, cells were exposed to 2 mL of CUR (40 µM) in Hanks balanced salt solution (HBSS). CUR was added to both the apical and the basolateral side for 2 h. Hereafter, cells were washed with HBSS, and fresh HBSS medium (10 mL) was added to both sides of the cells. Aliquots of medium from the apical side were taken after 90 min of dosing and immediately frozen on dry ice. After the aliquots were thawed, the CURSG-containing medium samples were incubated at room temperature in the dark and the CURSG stability was followed over time by analyzing aliquots of the incubations at 0, 1, 2, 4, 6, and 24 h by reversephase HPLC essentially as described above. 1 H NMR Analysis of CURSGs. The CURSGs were characterized by one-dimensional (1D) and two-dimensional (2D) homonuclear (TOCSY) 1H NMR using a Bruker DPX 400 or Bruker AMX500 NMR spectrometer. The NMR spectra were typically collected with 45° pulse angles, 32 K datapoints, 16 ppm spectral width, and presaturation of the DMSO-d6 or HDO signal when necessary. The 2D TOCSY spectra were recorded with the cl-mlev pulse sequence (29) with mixing times varying from 40 to 60 ms. The spectra were recorded at 298 °C. Resonances are reported relative to DMSO-d6 at 2.50 ppm. The total acquisition time of a typical proton spectrum was from about 0.5 h for 1D NMR data and about 10 h for 2D NMR data. Data were analyzed using TOPSPIN 1.3. LC-MS Analysis of CURSGs. LC-MS analysis was performed essentially as described by Wortelboer et al. (26). In short, CURSG peaks were separated on an Alltima C18 column (21mm × 150 mm, 5 µm, Alltech) using an elution flow of 0.2 mL/min and similar elution conditions as described above. Mass spectrometric analysis (Finnigan MAT95, San Jose, CA) was performed in the positive mode with a spray voltage of 4.5 kV and a capillary temperature of 180 °C with nitrogen as an auxiliary gas. Purification of GSTs. Purification of human GSTs A1-1, A2-2, M1a-1a, P1-1, and T1-1 from liver and placenta was performed as described by Bogaards et al. (30). Specific activities of human GSTs A1-1, A2-2, M1a-1a, and P1-1 toward 1-chloro-2,4-dinitrobenzene were 123, 64, 261, and 109 µmol/min/mg, respectively. The purity of GSTs was analyzed by HPLC using a Vydac 218TP5 (250 mm × 4.6 mm) column (The Separations Group, Hesperia, CA) with a flow rate of 1 mL/min. The purity of the GSTs was more than 98%. Incubations with Purified Enzymes. CUR (10–100 µM) was incubated for 5 min at 37 °C with human GSTs (65 µg/mL) in 100 µL incubation mixtures containing 0.1 M potassium phosphate, pH 7.5, and [35S]GSH (1 mM, specific activity 167 MBq/mmol). The reactions were terminated by direct freezing in liquid nitrogen and stored for maximal 8 h at –80 °C. Prior to the HPLC injection, each incubation mixture was thawed at 37 °C and centrifuged for 1.5 min at 14000 rpm. Subsequently, 85 µL was subjected to HPLC analysis. Blanks consisted of incubation mixtures with bovine serum albumin (BSA; 65 µg/mL) instead of GSTs. The catalytic activity of the GST was corrected for the chemically formation of CUR conjugates, which appeared to be less than 20%. HPLC analysis was performed using an Alltima C18 column (4.6 mm × 150 mm, 5 µm, Alltech) as described above. Radioactivity was detected using online radiochemical detection (Canberra Packard A500), and UV detection was performed at 360 nm. Incubations with Human Intestinal and Hepatic Cytosol. CUR (50 and 100 µM) was incubated for 5 min at 37 °C with intestine cytosol (4 mg/mL) (pooled of eleven individuals, with GST activity of 0.135 µmol/min/mg protein) and hepatic cytosol (4 mg/mL) (pooled of men and woman aged between 20 and 74 years, with GST activity of 0.42 µmol/min/mg protein) in 100 µL incubation mixtures containing 0.1 M potassium phosphate, pH 7.5, and [35S]GSH (1 mM, specific activity 167 MBq/mmol). The reactions were terminated by direct freezing in liquid nitrogen and stored for maximal 8 h at -80 °C. The sample preparation, HPLC analysis,

Chem. Res. Toxicol., Vol. 20, No. 12, 2007 1897

Figure 2. Decomposition vs time profile of CURSG-1 (open symbols) and CURSG-2 (closed symbols) in HBSS (pH7.5). Each bar represents the mean ( SD of incubations performed in triplicates.

and detection were performed as described above. A concentration of 4 mg/mL protein was used in the incubation mixtures. Blank incubations consisted of incubation mixtures with BSA (4 mg/mL) instead of cytosol. Efflux of CUR Conjugates from Caco-2 Cells. The efflux of the CUR conjugates (CURSG-1 and CURSG-2) from Caco-2 cells was followed over time. In brief, cells (passage 33) were cultured to confluence on 24 mm diameter transwells (Costar Corp., Cambridge, MA). Thereafter, cells were exposed to 2 mL of CUR (40 µM) in HBSS containing 0.5 mM acivicin (to prevent possible degradation of the GSH conjugates due to γ-glutamyltranspeptidase) and 0.1 µM PSC833 (to prevent efflux of CUR by P-glycoprotein) (26). CUR was added to both the apical and the basolateral side for 2 h. Hereafter, cells were washed with HBSS, and fresh HBSS medium (1 mL) containing 0.5 mM acivicin and 0.1 µM PSC833 was added to both sides of the cells. Aliquots of medium (150 µL) from both sides were taken at three time points (30, 60, and 90 min) and immediately frozen on dry ice. After they were thawed, samples were immediately analyzed by reverse-phase HPLC essentially as described above.

Results Formation, Stability, and Identification of Metabolites. Figure 1 shows a typical HPLC chromatogram (UV 360 nm) of an incubation mixture of CUR incubated with GSH and GSTM1a-1a. The HPLC chromatogram reveals the formation of two major metabolites labeled CURSG1 (tR ) 22.6 min) and CURSG2 (tR ) 23.3 min) and one minor unidentified metabolite with a retention time of 15.1 min (M3). The peaks with tR of 35–38 min are all associated with CUR. In our previous study (26), LC-MS analysis of the two CUR adducts CURSG-1 and CURSG-2 revealed an m/z of 675.8 and 675.9, respectively, which is identical to the mass expected for a protonated monoglutathionyl CUR adduct. In the present study, the metabolites were positively identified as GSH conjugates by incubation experiments with [35S]GSH. The minor metabolite with tR 15.1 min also appeared to be a GSH-related product, possibly a diglutathionyl CUR conjugate, as a radioactive peak in the HPLC chromatogram with similar retention time was observed. The monoglutathionyl conjugates of CUR appeared to be relatively unstable; if the incubation experiments with CUR were continued for longer periods (up to 50 min), a decrease of CURSG-1 and CURSG-2 was observed, whereas other peaks with tR 15, 16, and 18 min appeared (data not shown). The monoglutathionyl CUR conjugates have a t1/2 of about 4 h in physiological environment (pH 7.5 in HBSS) (Figure 2). HPLC analysis revealed that the monoglutathionyl CUR conjugates decompose to CUR, estimated after 2 h to account for 34% of the CURSG loss and several other unidentified degradation products, together amounting to about 60% of the lost CURSG. Also, storage of the incubation mixtures overnight at -20 °C

1898 Chem. Res. Toxicol., Vol. 20, No. 12, 2007

Usta et al.

Figure 3. 1H NMR spectrum of CUR and the 1H NMR spectrum of a mixture of CURSG-1 (minor form) and CURSG-2 (major form). CURSG-1 labeled as a and CURSG-2 labeled as b in δ ppm. The experimental conditions are as described in the Materials and Methods. Numbering of the different H atoms is based on the numbering of the C atoms presented in Figure 3.

resulted in a decrease of CURSG-1 and CURSG-2. For this reason, incubation mixtures were stored in liquid nitrogen, in which no decomposition of CURSG-1 and CURSG-2 was observed. 1 H NMR Analysis of CUR GSH Conjugates. The aromatic region of the 1D 1H NMR spectrum of CUR and of the mixture of CURSG-1 and CURSG-2 obtained from chemical incubations are presented in Figure 3. In Figure 3, the numbering of the various atoms of CUR can be found. Assignment of the resonances to specific protons was based on 1D and 2D 1H NMR spectra of the reference compound CUR, on chemical shifts and splitting patterns (coupling constants), on intensities of the resonances, and on the changes occurring in time in the spectrum and difference spectra obtained. Most characteristic in the 1H NMR spectra of CUR in DMSO-d6 (Figure 3 insert) are the two doublets of the protons H1dH7 and H2dH6 with a coupling constant of 16 Hz, indicating a trans configuration of these two protons (31). Characteristic doublets with coupling constants of 16 Hz can also be observed between 6.5 and 7.5 ppm for the two GSH adducts formed (Figure 3). In Figure 3, these characteristic doublets are present at 6.5 (two mixed doublets) and at 7.3 and 7.4 ppm in the mixture of CURSG-1 and CURSG-2. The NMR analysis also reveals that there are two CUR-SG conjugates present in the collected HPLC mixture at a ratio of about 2:1. These two CUR-SG adducts are not stable in time, as free CUR is formed relatively rapidly in protic solvents. Nevertheless, the assignments of the proton resonances of the two CURSG conjugates could be established for the CUR moiety of the molecules [CURSG-1 labeled as a in Figure 3 δ ppm: H6, 6.52 (d, J ) 16 Hz); H7, 7.31 (d, J ) 16 Hz); H2′, 6.87; H5′, 6.66 (m, AB pattern); H6′, 6.71 (m, AB pattern); H2′′, 7.13; H5′′, 6.79 (d, J ) 7.5 Hz); H6′′, 7.04 (d, J ) 7.5 Hz). Conjugate CURSG-2 labeled as b in Figure 3 δ ppm: H6, 6.50 (d, J ) 16 Hz); H7, 7.37 (d, J ) 16 Hz); H2′, 6.85; H5′, 6.66 (m, AB pattern); H6′, 6.71 (m, AB pattern); H2′′, 7.16; H5′′, 6.79 (d, J ) 7.5 Hz); H6′′, 7.06 (d, J ) 7.5 Hz)]. In 2D NMR spectra

(DQF-COSY and TOCSY), connectivities between specific protons (for example, H6-H7 or H2′-H6′) could be observed. The aliphatic proton resonances of the CUR moiety of the two conjugates could not be assigned, due to extensive overlap with the GSH resonances. It is evident that in both conjugates GSH addition did not occur in the aromatic ring system. The 1H resonances of all aromatic protons, namely, H2′, H5′, H6′, H2′′, H5′′, and H6′′, are still present in both adducts (Figure 3). The doublets of H6 and H7 and the resonances of the aromatic protons have a similar intensity, and this observation together with the number of proton signals observed indicate the loss of symmetry in the molecule. Because for both adducts the doublets for C(1)HdC(2)H are missing, it can be concluded that conjugate formation occurred at the C(1)HdC(2)H moiety. Given the fact that GSH conjugation to R-β-unsaturated carbonyls is known to occur at the β-carbon, it can be concluded that GSH conjugation in both diastereoismeric conjugates is at C1. Figure 4 gives a schematic presentation of the proposed conjugate formation. CUR, in analogy to other β-diketones, may exist mainly in the keto–enol configuration (Figure 4) (5, 32). This is corroborated by the presence of the resonance of H4 at 6.1 ppm in the 1H NMR spectrum of CUR (Figure 3 insert). For the two monoglutathionyl CUR conjugates, this proton resonance is no longer observed, and this points at the diketone configuration as the major form (Figure 4). Incubations with Human Hepatic and Intestinal Cytosols. The formation of CURSG in human hepatic and intestinal cytosol in the presence of CUR (50 and 100 µM) and GSH was determined (Figure 5). Incubations with human hepatic cytosol showed two times higher rate of GSH conjugation for CUR as compared to human intestinal cytosol. At a final concentration of 50 µM for CUR, the rate of enzymatic conjugation in hepatic and intestinal cytosol was about 9.3 and 4.3 times higher than the chemical formation, respectively. The glutathionylcurcumin conjugates, CURSG-1 and CURSG-2, appeared to be formed in the ratio of 2:1 (data not shown).

Glutathione Conjugates of Curcumin

Figure 4. Formation of the diastereoisomers of monoglutathionyl CUR conjugates. The asterisk mark the asymmetric C atom, which underlies the formation of the two diastereoisomers.

Chem. Res. Toxicol., Vol. 20, No. 12, 2007 1899

Formation and Efflux of CUR Conjugates in Caco-2 Cells. To study the formation of possible CUR GSH conjugates and also their directional efflux in human intestinal cells, Caco-2 cells grown in transwell dishes were loaded with CUR (160 nmol/monolayer), followed by studying the efflux of CUR and its GSH conjugates into fresh medium at both the apical and the basolateral side of the cells. After 30 min, a total of 23.1 ( 3.1 nmol nonconjugated CUR/monolayer was excreted to the medium, of which 65% was in the apical compartment and 35% in the basolateral compartment. Figure 6 presents the timedependent efflux of CURGS-1 and CURGS-2 from the cells, indicating the formation of these two conjugates at a ratio of about 2:1 in the human Caco-2 intestinal cells. After 30 min, the excretion of the glutathionyl conjugates CURSG-1 and CURSG-2 was 36 and 16 nmol/monolayer, respectively, of which 89% of total CURSG-1 and 75% of total CURSG-2 was excreted to the apical compartment (Figure 7). Upon increased incubation time, the level of both CURSG’s in the apical and basolateral media did not increase significantly. The typical polarized efflux to the apical side indicates that both CURSGs are efficiently transported by active transport proteins localized in the apical membrane of intestinal epithelial cells.

Discussion

Figure 5. Net formation of CURSG (mixture of CURSG-1 and CURSG-2) in human hepatic and intestinal cytosol incubations with 50 (black bars) and 100 µM (gray bars) CUR. The experimental conditions are described in the Materials and Methods. Each bar represents the mean ( SD of incubations performed in triplicate.

Incubations with Purified GSTs. To investigate the role of individual purified human GSTs in more detail, CUR was incubated with purified human GSTs. The enzymatic activities of human GSTA1-1, GSTA2-2, GSTM1a-1a, GSTP1-1, and GSTT1-1 (65 µg/mL) toward CUR (0–200 µM) are presented in Figure 6. GSTM1a-1a, GSTA1-1, GSTA2-2, and GSTP1-1 catalyzed GSH conjugation of CUR, resulting in the formation of both CURSG-1 and CURSG-2, whereas GSTT1-1 showed no significant catalytic activity toward CUR. GSTA2-2 and GSTP1-1 catalyzed the formation of both conjugates to an almost similar extent (Figure 6b,d). The catalytic activity toward CUR was higher for GSTM1a-1a and GSTA1-1, and both enzymes preferentially catalyzed the formation of CURSG1 over CURSG-2 (Figure 6a,c). The Km and Vmax values for formation of CURSG-1 and CURSG-2 with individual GSTs are shown in Table 1. In general, the Km for GSH conjugation leading to CURSG-1 decreases in the order of GSTM1a-1a > GSTA1-1 > GSTP1-1 > GSTA2-2, while the Km for CURSG-2 formation decreases in the order of GSTP1-1 ) GSTM1a-1a > GSTA1-1 > GSTA2-2. GSTM1a-1a showed the highest Vmax values with respect to the formation of both conjugates. Table 1 also presents the catalytic efficiency of GSTM1a-1a, GSTA1-1, GSTA2-2, and GSTP1-1, calculated as the Vmax/Km. These data reveal that the catalytic efficiencies for formation of CURSG1 and CURSG2 expressed as Vmax/Km do not vary significantly for the four different enzymes, mainly because enzymes with low Vmax values also tend to show low Km values. This implies that at relatively low CUR concentrations (75%). Active

transport of GSH conjugates can be mediated by the MRPs MRP1 and MRP2 (35, 36). MRP1 is localized at the basolateral side, whereas MRP2 is localized at the apical side of the cell. In a previous study, using MRP-transfected cells, it was demonstrated that both CURSG-1 and CURSG-2 can be actively transported by MRP1 and MRP2 (26). The expression level of MRP1 in both Caco-2 cells and in the human intestine is very low, whereas the expression level of MRP2 is high in the small intestine, which is reflected in Caco-2 cells (37, 38). Therefore, it can be concluded that in the human small intestine apical transporter proteins are predominantly involved in the active efflux of the formed monoglutathionyl CUR conjugates. Previous studies already revealed that the apical transporter MRP2

Glutathione Conjugates of Curcumin

Chem. Res. Toxicol., Vol. 20, No. 12, 2007 1901

formation of GSH conjugates can already occur during intestinal transport, after which the monoglutathionyl conjugates are efficiently excreted to the small intestinal lumen, thereby influencing the bioavailability of unconjugated CUR and, as a result, its potential beneficial biological effects. Acknowledgment. We thank J. J. P. Bogaards from TNO Quality of Life (Zeist, The Netherlands) who kindly provided the purified GST isoenzymes. This work was supported by Grant TNOV2000-2169 from the Dutch Cancer Society. Figure 7. Efflux of CUR conjugates CURSG-1 (spherical symbols) and CURSG-2 (triangular symbols) from Caco-2 cells to the apical side (closed symbols) and basolateral side (open symbols).

can be involved in the excretion of CUR conjugates, but this does not exclude a possible role for other apical transport proteins. In a previous study, Samiec et al. (39) also showed the existence of GST activity in the mucus of rat small intestine. The GST (µ and π class) present in the extracellular mucus catalyzes the conjugation of electrophilic compounds and utilizes GSH supplied by the bile and food. These data together indicate that CUR may be conjugated extracellularly in the mucus or intracellularly in the intestinal cells, in the latter case followed by transport of CUR and its conjugates back into the lumen, thereby reducing the bioavailability of unconjugated CUR. In contrast to the small intestine, especially the ileum, the expression levels of the active transporters MRP2, Pgp, and BCRP in the human colon are very low (38), which may result in higher levels of CUR in the colon tissue. As the colon is exposed to both CUR and its metabolites, it is a likely target for the anticarcinogenic activity of these compounds (40). In rats, CUR and metabolites formed in the intestine and liver are predominantly excreted in the feces (28, 41, 42). After oral administration of 400 mg of [3H]CUR to rats, about 60% of the [3H]label of the given dose was absorbed (41). The CUR concentrations observed in rats in the liver and colon mucosa were about 0.8 nmol/g and 1.8 µmol/g, respectively (43). Oral intake in humans at levels up to about 6–8 g CUR per day did not result in adverse effects and resulted in a plasma concentration of about 2 µM (27). Because of low CUR supply by systemic circulation, interaction of CUR with GSTs is more likely to occur in intestinal epithelial cells than in other tissue cells. It is known that GSH conjugation of CUR, like other R,βunsaturated carbonyl compounds, is reversible (44). The reversibility of CURSG in free CUR was also observed in our study, indicating that upon transport of CUR-GSH conjugates back into the intestinal lumen, free CUR may become available for reabsorption as was before suggested (25). In man, marked interindividual differences exist in the expression of GST enzymes. The molecular basis for the variation (polymorphic activities) in µ (GSTM) and θ class (GSTT) GSTs can be attributed to the absence of the GSTM1 gene in 50% of the Caucasian population and deletion of the GSTT1 gene in 10–60% of the population (20, 45). In this case, the absence of GSTM1a-1a may have consequences (such as higher bioavailability of free CUR) for the Caucasian population in the metabolism of CUR. In conclusion, the present study identifies the nature of the diastereoisomeric monoglutathionyl CUR conjugates, CURSG-1 and CURSG-2, formed in biological systems and reveals that conjugate formation is catalyzed by GSTM1a-1a, GSTA1-1, and/or GSTP1-1 with different stereoselective preference. The

References (1) Osawa, T., Sugiyama, Y., Inayoshi, M., and Kawakashi, S. (1995) Antioxidative activity of tetrahydrocurcuminoids. Biosci., Biotechnol., Biochem. 59, 1609–1612. (2) Sugiyama, Y., Kawakishi, S., and Osawa, T. (1996) Involvement of the beta-diketone moiety in the antioxidative mechanism of tetrahydrocurcumin. Biochem. Pharmacol. 52, 519–525. (3) Commandeur, J. N. M., and Vermeulen, N. P. E. (1996) Cytotoxicity and cytoprotective activities of natural compounds. The case of curcumin. Xenobiotica 26, 667–680. (4) Jovanovic, S. V., Steenken, S., Boone, C. W., and Simic, M. G. (1999) H-atom transfer is a preferred antioxidant mechanism of curcumin. J. Am. Chem. Soc. 121, 9677–9681. (5) Masuda, T., Toi, Y., Bando, H., Maekawa, T., Takeda, Y., and Yamaguchi, H. (2002) Structural identification of new curcumin dimers and their contribution to the antioxidant mechanism of curcumin. J. Agric. Food Chem. 50, 2524–2530. (6) Piper, J. T., Singhal, S. S., Salameh, M. S., Torman, R. T., Awasthi, Y. C., and Awasthi, S. (1998) Mechanisms of anti carcinogenic properties of curcumin: The effect of curcumin on glutathione linked detoxification enzymes in rat liver. Int. J. Biochem. Cell Biol. 30, 445– 456. (7) Biswas, S. K., McClure, D., Jimenez, L. A., Megson, I. L., and Rahman, I. (2005) Curcumin induces glutathione biosynthesis and inhibits NF-kappaB activation and interleukin-8 release in alveolar epithelial cells: mechanism of free radical scavenging activity. Antioxid. Redox Signaling 7, 32–41. (8) Huang, M. T., Lou, Y. R., Ma, W., Newmark, H. L., Reuhl, K. R., and Coney, A. H. (1994) Inhibitory effect of dietary curcumin on forestomach, duodenum, and colon carcinogenesis in mice. Cancer Res. 54, 5841–5847. (9) Hanif, R., Qiao, L., Shiff, S. J., and Rigas, B. (1997) Curcumin, a natural plant phenolic food additive, inhibits cell proliferation and induces cell cycle changes in colon adenocarcinoma cell lines by prostaglandin independent pathway. J. Lab. Clin. Med. 130, 576–584. (10) Chen, H., Zhang, Z. S., Zhang, Y. L., and Zhou, D. Y. (1999) Curcumin inhibits cell proliferation by interfering with the cell cycle and inducing apoptosis in colon carcinoma cells. Anticancer Res. 19, 3675–3680. (11) Moraga, L., Jaszweski, R., and Majumdar, A. P. (2001) Curcumin induced modulation of cell cycle and apoptosis in gastric and colon cancer cells. Anticancer Res. 21, 873–878. (12) Deeb, D., Xu, D. X., Jiang, H., Goa, X., Janakiraman, N., Chapman, R. A., and Gautam, S. C. (2003) Curcumin (Diferuloylmethane) enhances tumor necrosis factor related apoptosis inducing ligandinduced apoptosis in LNCaP prostate cancer cells. Mol. Cancer Ther. 2, 95–103. (13) Mehta, K., Pantazis, P., McQueen, T., and Aggarwal, B. B. (1997) Antiproliferative effect of curcumin (Diferuloylmethane) against human breast tumor cell lines. Anticancer Drugs 8, 470–481. (14) Pan, M. H., Chang, W. L., Lin-Shiau, S. Y., Ho, C. T., and Lin, J. K. (2001) Induction of apoptosis by garcinol and curcumin through cytochrome c release and activation of caspases in human leukemia HL-60 cells. J. Agric. Food Chem. 49, 1464–1474. (15) Mukhopadhyay, A., Banerjee, S., Stafford, L. J., Xia, C., Liu, M., and Aggarwal, B. B. (2002) Curcumin-induced suppression of cell proliferation correlates with down- regulation of cyclin D1 expression and CDK4-mediated retinoblastoma protein phosphorylation. Oncogene 21, 8852–8861. (16) Shi, M., Cai, Q., Yao, L., Mao, Y., Ming, Y., and Ouyang, G. (2005) Antiproliferation and apoptosis induced by curcumin in human ovarian cancer cells. Cell. Biol. Int. 30, 221–226. (17) Yang, F., Lim, G. P., Begum, A. N., Ubeda, O. J., Simmons, M. R., Ambegaokar, S. S., Chen, P. P., Kayed, R., Glabe, C. G., Frautschy, S. A., and Cole, G. M. (2005) Curcumin inhibits formation of aβ oligomers and fibrils, binds plaques and reduces amyloid in vivo. J. Biol. Chem. 280, 5892–5901.

1902 Chem. Res. Toxicol., Vol. 20, No. 12, 2007 (18) Kawanishi, S., Oikawa, S., and Murata, M. (2005) Evaluation for safety of antioxidant chemopreventive agents. Antioxid. Redox Signaling 7, 1728–1739. (19) Manson, M. M. (2005) Inhibition of survival signaling by dietary polyphenols and indole-3-carbinol. Eur. J. Cancer 41, 1842–1853. (20) Ketterer, B. (1988) Protective role of glutathione and glutathione transferases in mutagenesis and carcinogenesis. Mutat. Res. 202, 343– 361. (21) Mathews, S., and Rao, M. N. A. (1991) Interaction of curcumin with glutathione. Int. J. Pharm. 76, 257–259. (22) Iersel, M. L. P. S., van, Ploemen, J. H., Lo Bello, M., Federici, G., and van Bladeren, P. J. (1997) Interactions of R,β unsaturated aldehydes and ketones with human glutathione S-transferase P1-1. Chem.-Biol. Interact. 108, 67–78. (23) Ploemen, J. H. T. M., van Ommen, B., and van Bladeren, P. J. (1990) Inhibition of rat and human glutathione S-transferase isoenzymes by ethacrynic acid and its glutathione conjugate. Biochem. Pharmacol. 40, 1631–1635. (24) Bogaards, J. J. P., Venekamp, J. C., and van Bladeren, P. J. (1997) Stereoselective conjugation of prostaglandin A2 and prostaglandin J2 with glutathione, catalyzed by the human glutathione S-transferases A1-1, A2-2, M1a-1a, and P1-1. Chem. Res. Toxicol. 10, 310–317. (25) Awasthi, S., Pandya, U., Singhal, S. S., Lin, J. T., Thiviyananathan, V., Seifert, W. E., Jr., Awasthi, Y. C., and Ansari, G. A. S. (2000) Curcumin-glutathione interactions and the role of human glutathione S-transferase P1-1. Chem.-Biol. Interact. 128, 19–38. (26) Wortelboer, H. M., Usta, M., van der Velde, A. E., Boersma, M. G., Spenkelink, B., van Zanden, J. J., Rietjens, I. M. C. M., van Bladeren, P. J., and Cnubben, N. H. P. (2003) Interplay between MRP-inhibition and metabolism of MRP-inhibitors: The case of curcumin. Chem. Res. Toxicol. 16, 1642–1651. (27) Cheng, A.-L., Hsu, C.-H., Lin, J.-K., Hsu, M.-W., Ho, Y.-F., Shen, T.-S., Ko, J.-Y., Lin, J.-T., Lin, B.-R., Wu, M.-S., Yu, H.-S., Jee, S.-H., Chen, G.-S., Chen, T.-M., Chen, C.-A., Lai, M.-K., Pu, Y.-S., Pan, M.-H., Wang, Y.-J., Tsai, C.-C., and Hsieh, C.-Y. (2001) Phase I clinical trial of curcumin, a chemopreventive agent, in patients with high risk or pre-malignant lesions. Anticancer Res. 21, 2895–2900. (28) Ireson, C. R., Jones, D. J. L., Orr, S., Coughtrie, M. W. H., Boocock, D. J., Williams, M. L., Farmer, P. B., Steward, W. P., and Gescher, A. J. (2002) Metabolism of the cancer chemopreventive agent curcumin in human and rat intestine. Cancer Epidemiol. Biomarkers PreV. 11, 105–111. (29) Griesinger, C., Otting, G., Wuethrich, K., and Ernst, R. R. (1988) Clean TOCSY for h-1 spin system-identification in macromolecules. J. Am. Chem. Soc. 110, 7870–7872. (30) Bogaards, J. J., van Ommen, B., and van Bladeren, P. J. (1992) Purification and characterization of eight glutathione S-transferase isoenzymes of hamster. Comparison of subunit composition of enzymes from liver, kidney, testis, pancreas and trachea. Biochem. J. 286, 383–388. (31) Günther, H. (1995) NMR Spectroscopy: Basic Principles, Concepts and Applications in Chemistry, Wiley, New York.

Usta et al. (32) Mague, J. T., Alworth, J. L., and Payton, F. L. (2004) Curcumin and derivatives. Acta Crystallogr. C60, 608–610. (33) Ogasawara, T., Ohnhaus, E. E., and Hoensch, H. P. (1989) Glutathione and its related enzymes in the small intestinal mucosa of rats: Effect of starvation and diet. Res. Exp. Med. 189, 195–204. (34) Siegers, C. P., Riemann, D., Thies, E., and Younes, M. (1988) Glutathione and GSH-dependent enzymes in the gastrointestinal mucosa of the rat. Cancer Lett. 40, 71–76. (35) Borst, P., Evers, R., Kool, M., and Wijnholds, J. (2000) A family of drug transporters; the multidrug resistance-associated proteins. J. Natl. Cancer Inst. 92, 1295–1302. (36) Leslie, M. E., Deeley, R. G., and Cole, S. P. C. (2005) Multidrug resistance proteins: role of P-glycoprotein, MRP1, MRP2, and BCRP (ABCG2) in tissue defense. Toxicol. Appl. Pharmacol. 204, 216–237. (37) Taipalensuu, J., Tornblom, H., Lindberg, G., Einarsson, C., Sjogvist, F., Melhus, H., Garberg, P., Sjostrom, B., Lundgren, B., and Artursson, P. (2001) Correlation of gene expression of 10 drug efflux proteins of the ATP-binding cassette transporter family in normal human jejunum and in human intestinal epithelial Caco-2 cell monolayers. J. Pharmacol. Exp. Ther. 299, 164–170. (38) Englund, G., Rorsman, F., Ronnblom, A., Karlbom, U., Lazarova, L., Grasjo, J., Kindmark, A., and Artursson, P. (2006) Regional levels of drug transporters along the human intestinal tract: Co-expression of ABC and SLC transporters and comparison with Caco-2 cells. Eur. J. Pharmacol. Sci. 29, 269–277. (39) Samiec, P. S., Dahm, L. J., and Jones, D. P. (1999) Glutathione S-transferase in mucus of rat small intestine. Toxicol. Sci. 54, 52–59. (40) van Erk, M. J., Teuling, E., Staal, Y. C. M., Huybers, S., van Bladeren, P. J., Aarts, J. M. M. J. G., and van Ommen, B. (2004) Time- and dose-dependent effects of curcumin on gene expression in human colon cancer cells. J. Carcinog. 12, 3–8. (41) Ravindranath, V., and Chandrasekhara, N. (1980) Absorption and tissue distribution of curcumin in rats. Toxicology 16, 259–265. (42) Ravindranath, V., and Chandrasekhara, N. (1981) Metabolism of curcumin-studies with [3H]curcumin. Toxicology 22, 337–344. (43) Sharma, R. A., Ireson, C. R., Verschoyle, R. D., Hill, K. A., Williams, M. L., Leuratti, C., Manson, M. M., Marnett, L. J., Steward, W. P., and Gescher, A. (2001) Effects of dietary curcumin on glutathione Stransferases and malondialdehyde-DNA adducts in rat and colon mucosa: relationship with drug levels. Clin. Cancer Res. 7, 1452– 1458. (44) Ploemen, J. H. T. M., Van Schanke, A., van Ommen, B., and van Bladeren, P. J. (1994) Reversible conjugation of ethacrynic acid with glutathione and human glutathione S-transferase P1-1. Cancer Res. 54, 915–919. (45) Hayes, J. D., and Pulford, D. J. (1995) The glutathione S-transferase supergene family: regulation of GST and the contribution of the isoenzymes to cancer chemoprotection and drug resistance. Crit. ReV. Biochem. Mol. Biol. 30, 445–600.

TX7002245