Abstract
Pharmacokinetics of trans-resveratrol in its aglycone (RESAGL) and glucuronide (RESGLU) forms were studied following intravenous (15 mg/kg i.v.) and oral (50 mg/kg p.o.) administration of trans-resveratrol in a solution of β-cyclodextrin to intact rats. In addition, the enterohepatic recirculation of RESAGL and RESGLU was assessed in a linked-rat model. Multiple plasma and urine samples were collected and concentrations of RESAGL and RESGLU were determined using an electrospray ionization-liquid chromatography/tandem mass spectrometry method. After i.v. administration, plasma concentrations of RESAGL declined with a rapid elimination half-life (T1/2, 0.13 h), followed by sudden increases in plasma concentrations 4 to 8 h after drug administration. These plasma concentrations resulted in a significant prolongation of the terminal elimination half-life of RESAGL(T1/2TER, 1.31 h). RESAGLand RESGLU also displayed sudden increases in plasma concentrations 4 to 8 h after oral administration, withT1/2TER of 1.48 and 1.58 h, respectively. RESAGL bioavailability was 38% and its exposure was approximately 46-fold lower than that of RESGLU (AUCinf, 7.1 versus 324.7 μmol·h/l). Enterohepatic recirculation was confirmed in the linked-rat model since significant plasma concentrations of RESAGL and RESGLU were observed in bile-recipient rats at 4 to 8 h. The percentages of the exposures of RESAGL and RESGLU that were due to enterohepatic recirculation were 24.7 and 24.0%, respectively. The fraction of drug excreted in the urine over a period of 12 h was negligible. These results confirm that RESAGL is bioavailable and undergoes extensive first-pass glucuronidation, and that enterohepatic recirculation contributes significantly to the exposure of RESAGL and RESGLU in rats.
Interest in the study of phenolic compounds present in red wine has grown since epidemiological studies have shown an inverse correlation between red wine consumption and the incidence of cardiovascular diseases (Nanji and French, 1986; Hegsted and Ausman, 1988). Resveratrol (3,5,4′-trihydroxystilbene), a molecule from the viniferin family of polymers, was identified as a biologically active compound in red wine in 1992 (Siemann and Creasy, 1992). Since then, numerous in vivo and in vitro studies have assessed the ability of resveratrol in preventing multiple pathophysiological processes. Resveratrol has the ability to inhibit the peroxidation of lipid membranes (Fauconneau et al., 1997), to decrease the concentration of low- and very-low-density lipoproteins (Frankel et al., 1993), and to inhibit platelet aggregation (Kimura et al., 1985), three conditions that help prevent cardiovascular diseases. Although significant estrogenic-like activity of resveratrol has been demonstrated in vitro (Gehm et al., 1997; Bhat et al., 2001), this was not proven in vivo in rats (Turner et al., 1999). Finally, trans-resveratrol was shown to have cancer chemoprotective properties and to induce apoptosis in leukemia and human breast carcinoma (Jang et al., 1997; Mgbonyebi et al., 1998; Lu and Serrero, 1999). Other potential benefits oftrans-resveratrol are related to its anti-inflammatory properties since it inhibits cyclooxygenase-1 and hydroxyperoxidase activities (Kimura et al., 1985).
Information on the pharmacokinetics of trans-resveratrol remains scarce despite the vast amount of research published on its potential efficacy. Bertelli et al. (1996) showed that significant concentrations of trans-resveratrol are seen in plasma and other tissues after either short-term or prolonged administration of red wine to rats. Soleas et al. (2001) also showed measurable concentrations of trans-resveratrol in serum and blood after intragastric administration of trans-resveratrol in rats. However, the radioactivity levels following intragastric administration of tritiated trans-resveratrol in serum declined far more slowly than those of the parent compound, suggesting the presence of radioactive metabolites (Soleas et al., 2001).
The absorption, metabolism, and disposition oftrans-resveratrol must be determined before any conclusion on the benefits of dietary or commercially available resveratrol can be drawn. The present investigation was conducted to determine the pharmacokinetics of trans-resveratrol in its aglycone (RESAGL) and glucuronide (RESGLU) forms following i.v. and p.o. administration to rats. RESAGL was administered orally to bile-donor rats, and their bile flowed directly into the duodenum of bile-recipient rats via surgically implanted catheters so that the contribution of enterohepatic recirculation to the overall disposition would be determined.
Materials and Methods
A total of 18 male Sprague-Dawley rats (Charles River Canada, St-Constant, QC, Canada) were used in this study with a mean ± S.D. body weight of 335 ± 36 g. Environmental conditions were monitored (temperature, 21 ± 3°C; humidity, 30–70%; and photoperiod, 12 h light/12 h dark) during the acclimation period (1 week) and the conduct of the study. Rats were fed with a standard certified commercial laboratory diet (Teklad 7012; Harlan, Indianapolis, IN), and reverse osmosis UV-treated water was available ad libitum, except before dosing when food was withheld for 12 h. The Institutional Animal Care and Use Committee approved the experimental protocol before the study was conducted. After anesthesia with 70 mg/kg ketamine (Wyeth-Ayerst Canada Inc., St-Laurent, QC, Canada) and 10 mg/kg xylazine (Miles Canada, Etobicoke, ON, Canada), the bile duct and duodenum of rats were cannulated. Surgical procedures pertaining to the linked-rat model were performed according to the method of Waynforth and Flecknell (1992). Briefly, one catheter (PE-10) was inserted into the proximal portion of the bile duct toward the liver, and a second catheter (PE-50) was inserted directly into the duodenum. Both catheters were secured in place by ligation with surgical sutures and exteriorized subcutaneously at the back of the animal prior to closure of the abdomen. Finally, the catheters were connected via a dual swivel device to allow free flow of the bile in the animal. During the recovery period (1 week), bile duct and duodenum catheters remained connected together such that normal bile circulation within the animal was not interrupted. Approximately 2 h before dosing, rats were paired so that the bile cannula of one rat was connected to the duodenal cannula of a second rat.
Resveratrol is only commercially available as thetrans-isomer, the most stable and pharmacologically active form of resveratrol (Sigma-Aldrich, St. Louis, MO). RESAGL was prepared at a concentration of 6 mg/ml in a saline solution (0.9% NaCl) of 20% hydroxypropyl β-cyclodextrin (American Maize-Products Co., Hammond, IN). Rats were weighed and RESAGL was administered intravenously (15 mg/kg, n = 6) and orally (50 mg/kg,n = 6) to intact rats. RESAGL was also administered orally to bile-donor rats (50 mg/kg,n = 3) paired to bile-recipient rats (n= 3) in a linked-rat model. Blood samples were collected from the jugular vein prior to dosing and at 0.08, 0.25, 0.5, 1, 2, 4, 6, 8, and 12 h after dosing. Urine samples were collected every 4 h in bile-donor and bile-recipient rats only. Plasma and urine samples were maintained on ice before being centrifuged (4°C for 10 min at 3200g), and aliquots were stored at −80°C pending the assay.
Concentrations of RESAGL were determined using an ESI-LC/MS/MS method developed at MDS Pharma Services. Total glucuronide metabolites of resveratrol (RESGLU) were obtained after incubation with purified β-glucuronidase (type X-A; Sigma-Aldrich). Briefly, plasma and urine samples were divided into two 50-μl samples, and then 10 μl of β-glucuronidase enzyme solution (3.17 mg/ml in enzyme buffer) was added to the first sample, whereas 10 μl of enzyme buffer (0.2 M sodium acetate, pH 5.0) was added to the second sample. The samples were vortexed and incubated at 40°C for 1 h. After the incubation, 150 μl of the internal standard working solution (50 ng/ml naringenin in methanol) was added to each sample, and the tubes were vortexed. The samples were centrifuged at 13,000 rpm for 15 min, and their supernatants were aliquoted into vials. The samples were then analyzed using an online column switching setup consisting of a Waters 717 autosampler (Waters, Milford, MA), a Shimadzu LC-10AD VP pump system (Shimadzu, Columbia, MD) for sample loading and washing, a PE Series 200 micro pump system for sample elution (PerkinElmer Instruments, Norwalk, CT), and an external six-port valve (Valco Instruments Co., Inc., Houston, TX) for switching between sample loading and elution. Each sample was injected (20 μl) and trapped on a Zorbax SB-C18 guard column (12.5 × 4.6 mm diameter, 5-μm pore size; Agilent Technologies, Inc., Palo Alto, CA) with a flow rate of 1.0 ml/min and a loading phase consisting of 25 mM ammonium acetate in water/methanol (70:30, v/v). After 1.0 min, the guard column was switched online and the sample was eluted on a Zorbax SB-C18 analytical column (75 × 4.6 mm, 3.5 μm) with a flow rate of 1.0 ml/min and an elution phase consisting of 25 mM ammonium acetate in water/methanol (35:65, v/v). The flow from the column was split 1:3 into a PE Sciex API3000 triple quadrupole mass spectrometer (PerkinElmerSciex, Concord, ON, Canada) equipped with a TurboIonspray source operating at 450°C. Resveratrol and naringenin were monitored in negative mode under MS/MS conditions. The retention times for resveratrol and naringenin were 2.1 and 2.5 min, respectively. RESAGL was determined from the sample incubated with the buffer solution, whereas RESGLU was determined by subtracting the concentrations of the sample incubated with buffer from the concentrations measured in the sample incubated with β-glucuronidase. Appropriate dilutions were performed when concentrations fell outside the analytical range (5–5000 μg/l). Linearity was assessed by plotting area ratios versus standard concentrations and using a linear regression weighted 1/x. The correlation coefficients ranged from 0.9967 to 0.9990 for the four batches. Intraday and interday precision and accuracy of the assay were assessed with three levels of quality control samples (4000, 2000, and 15 μg/l) in quadruplicate. The intraday CV% ranged from 2.0 to 16.5% and the interday CV% ranged from 2.5 to 9.2% for all three quality control samples. Intraday accuracy ranged from 93.0 to 112.7% and interday accuracy ranged from 99.6 to 106.3% for all three quality control samples. Concentrations of RESAGL and RESGLU were adjusted for the molar weight of resveratrol (228.2 g/mol) and presented as micromoles per liter.
Pharmacokinetic parameters of RESAGL and RESGLU were calculated using noncompartmental methods (Rowland and Towzer, 1995). The area under the curve from time 0 to the last measurable plasma concentration (AUC0–t) was calculated using the linear trapezoidal rule. After i.v. administration, a rate constant of elimination (kel) was calculated using the first three plasma concentrations, and the elimination half-life (T1/2) was calculated using 0.693/kel. When enterohepatic recirculation was thought to occur, a terminal rate constant of elimination (kel TER) was calculated using the last three measurable plasma concentrations of the profile, and a terminal elimination half-life (T1/2TER) was calculated using 0.693/kel TER. The area under the curve extrapolated to infinity (AUCinf) was calculated using AUC0–t +Clast/kel TERwhere Clast was the last measurable plasma concentration. Clearance (CL), oral clearance (CL/F), and RESGLU apparent clearance (CL/Fm) were calculated by dividing the dose by the appropriate AUCinf. The mean residence time (MRT) was obtained after i.v. administration by dividing the area under the first moment-time curve (AUMCinf) by the AUCinf, and the total volume of distribution (Vss) was calculated using CL × MRT. The apparent volumes of distribution of RESAGL and RESGLU after oral dosing (Varea/F andVarea/Fm, respectively) were calculated using dose/(AUCinf ×kel TER). The bioavailability (F) of RESAGL was calculated using (AUCinf p.o./dosep.o.)/(AUCinf i.v./dosei.v.). The percentage of the exposure to RESAGL and RESGLU due to enterohepatic recirculation (%ER) was assessed using the ratio of AUC0–t of bile-recipient rats relative to the sum of AUC0–t found in both bile-donor and bile-recipient rats.
Differences between T1/2 andT1/2TER after i.v. administration were assessed using paired t tests. Statistical analyses were performed using SYSTAT version 8.0 for Windows (SPSS Inc., Cary, NC; 1998), and the level of statistical significance was set a priori atp < 0.05.
Results
Mean (±S.D.) plasma concentrations of RESAGL and RESGLUfollowing the i.v. administration of 15 mg/kg are depicted in Fig.1, and pharmacokinetic parameters are presented in Table 1. The plasma concentrations of RESAGL declined rapidly over the first 2 h with a mean elimination half-life (T1/2) of 0.13 ± 0.02 h and then increased abruptly over the 4- to 8-h time period. When enterohepatic recirculation was thought to occur, mean terminal elimination half-life (T1/2TER) of RESAGL was prolonged to 1.31 ± 0.27 h (p < 0.05). RESGLU displayed similar increases in plasma concentrations over the 4- to 8-h time period with a mean T1/2TER of 1.52 ± 0.47 h. RESAGL clearance was extensive (CL, 11.7 ± 1.0 l/h/kg) and markedly higher than that of RESGLU (CL/Fm, 1.73 ± 0.27 l/h/kg). Consequently, the systemic exposure (AUCinf) of RESAGL was approximately 7-fold lower than that of RESGLU. Mean (±S.D.) plasma concentration profiles of RESAGL and RESGLU following oral administration are depicted in Fig. 2, and pharmacokinetic parameters are presented in Table 2. Concentration profiles of RESAGL and RESGLU displayed similar increases in plasma concentrations over the 4- to 8-h time periods withT1/2TER of 1.48 ± 0.44 h and 1.55 ± 0.42 h, respectively. RES clearance (CL/F, 32.4 ± 7.5 l/h/kg) after oral administration was markedly higher than that of RESGLU (CL/Fm, 0.70 ± 0.15 l/h/kg). The bioavailability of RESAGLadministered in a solution of hydroxypropyl β-cyclodextrin was 38.1 ± 13.5%.
Mean (±S.D.) plasma concentrations (μM) of RESAGL (●) and RESGLU (○) after an intravenous dose of 15 mg/kg resveratrol in intact rats (n = 6).
Mean ± S.D. pharmacokinetic parameters of RESAGL and RESGLU following intravenous administration of 15 mg/kg to intact rats (n = 6)
Mean (±S.D.) plasma concentrations (μM) of RESAGL (●) and RESGLU (○) after an oral dose of 50 mg/kg resveratrol in intact rats (n = 6).
Mean ± S.D. pharmacokinetic parameters of RESAGL and RESGLU following oral administration of 50 mg/kg to intact rats (n = 6)
Mean (±S.D.) plasma concentration profiles of RESAGL and RESGLU in bile-donor and bile-recipient rats are depicted in Fig.3, and pharmacokinetic parameters are presented Table 3. Plasma concentrations of RESAGL and RESGLU in bile-donor rats did not display sudden peaks like those observed in intact rats receiving either i.v. or p.o. doses. Bile-recipient rats had significant peak plasma concentrations of RESAGL(Cmax, 2.9 ± 0.3 μmol) and RESGLU (Cmax, 9.1 ± 2.1 μmol) at 6.0 h. The terminal elimination half-life (T1/2TER) of RESAGL and RESGLU in bile-recipient rats could not be calculated adequately since only a few data points above the limit of quantitation were available. Based on the observed AUC0–t of bile-donor and bile-recipient rats, the percentages of the exposure that were due to enterohepatic recirculation (%ER) were calculated to be 24.7 ± 15.1% and 24.0 ± 8.5% for RESAGL and RESGLU, respectively. The cumulative amount of drug excreted in urine (Ae0–12) as RESAGL was lower than that of RESGLU in both bile-donor and bile-recipient rats. The greatest elimination of RESGLU in urine was found in bile-recipient rats (0.158 ± 0.087 μmol), and the value corresponded to approximately 0.2% of the administered dose.
Mean (±S.D.) plasma concentrations (μM) of RESAGL (●) and RESGLU (○) after an oral dose of 50 mg/kg resveratrol in bile-donor (n = 3) rats with their bile cannula connected into the duodenum of bile-recipient rats (n = 3).
Mean ± S.D. pharmacokinetic parameters of RESAGL and RESGLU following oral administration of 50 mg/kg to bile-donor rats (n = 3) rats with their bile cannula connected into the duodenum of bile-recipient rats (n = 3)
The fraction of drug excreted as RESAGL and RESGLU in urine over the 0- to 4-, 4- to 8-, and 8- to 12-h time intervals in bile-donor and bile-recipient rats is presented in Fig. 4. The fraction of drug excreted as RESAGL and RESGLU in urine was lower than 0.15% in both cases, and the interval associated with the greatest elimination in bile-recipient rats was the 4- to 8-h interval.
Mean (±S.D.) fraction of drug excreted as RESAGL (black bars) or RESGLU (open bars) in urine over different time intervals following an oral dose of 50 mg/kg resveratrol in bile-donor (n = 3) rats with their bile cannula connected into the duodenum of bile-recipient rats (n = 3).
Discussion
Plasma concentrations of RESAGL declined in a monoexponential manner in the initial elimination phase following i.v. administration. Plasma concentrations then increased abruptly due to enterohepatic recirculation over the 4 to 8 h time period and resulted in a significant prolongation in the terminal elimination half-life of RESAGL. The exposure of RESAGL was approximately 7-fold lower than that of RESGLU, confirming the importance of the glucuronidation elimination pathway of RESAGLafter i.v. administration. RESAGL was found to be 38% bioavailable after its oral administration in a solution of hydroxypropyl β-cyclodextrin. Following the initial absorption phase, plasma concentrations of RESAGL and RESGLU displayed sudden peaks over the 4 to 8 h time interval due to enterohepatic recirculation. The clearance of RESAGL after p.o. administration was markedly higher than that of RESGLU, resulting in a systemic exposure of approximately 46-fold lower than that of RESGLU.
Recently, the enterohepatic recirculation of methylergometrine and doxorubicin (Adriamycin) was assessed by diverting the bile cannula from a bile-donor rat into the duodenum of a bile-recipient rat (Bredberg and Paalzow, 1997; Behnia and Boroujerdi, 1998). Using similar methods, we administered RESAGL orally (50 mg/kg) in a linked-rat model, and multiple plasma and urine samples were collected to assess the pharmacokinetics of RESAGL and RESGLU in both bile-donor and bile-recipient rats. The plasma concentrations of RESAGL and RESGLU in bile-donor rats declined with no sudden increases in plasma concentrations after the initial absorption phase. This is most likely due to the interruption of the recirculatory pathway in bile-donor rats. Enterohepatic recirculation was confirmed by the presence of significant plasma concentrations of RESAGL and RESGLU in bile-recipient rats over the 4- to 8-h time period. These observations suggest that RESGLU is most likely excreted in the bile of bile-donor rats and reabsorbed in the intestine of bile-recipient rats in its aglycone and/or glucuronide forms. Plasma concentrations of RESAGL and RESGLU in bile-recipient rats coincided with the sudden peaks in plasma concentrations observed in intact rats receiving i.v. or p.o. doses. In addition, the fraction of the drug excreted in urine reached a maximum value during the 4- to 8-h time interval in bile-recipient rats, coinciding with their respective time to maximum plasma concentrations.
Enterohepatic recirculation in rats has been shown to be governed by the transit time of a drug to reach the cecum after it has been released from the bile (Walsh and Levine, 1975). There, the glucuronide metabolites may undergo enzymatic cleavage by the β-glucuronidase enzyme, and reabsorption of the aglycone parent compound may occur. The times to maximum plasma concentration of RESAGLand RESGLU in bile-recipient rats are in agreement with those of other compounds undergoing enterohepatic recirculation such as morphine-3-glucuronide (Ouellet and Pollack, 1995), valproic acid (Pollack and Brouwer, 1991), and phenolphthalein (Colburn et al., 1979).
Previous studies on the absorption and metabolism oftrans-resveratrol using an isolated rat small intestine model revealed that trans-resveratrol is most likely to be absorbed in the form of a glucuronide after crossing the small intestine (Andlauer et al., 2000; Kuhnle et al., 2000). In our study, RESGLU exposures were approximately 7- and 46-fold higher than those of RESAGL after intravenous and oral administration, respectively. This supports the working hypothesis that the intestine plays an important role in the presystemic glucuronidation of resveratrol. Some glucuronides have been associated with pharmacological activity. For example, morphine-6-β,d-glucuronide has been shown to be a significant contributor to the overall pharmacological activity of morphine (Hanks and Wand, 1989). These observations support investigating the pharmacological activity of RESGLU further, since its systemic exposure is so much greater than that of RESAGL after both routes of administration.
To our knowledge, this is the first study assessing the metabolism and disposition of trans-resveratrol in vivo. We have shown that RESAGL is bioavailable at 38% when administered in a solution of hydroxypropyl β-cyclodextrin and undergoes extensive first-pass glucuronidation, and that enterohepatic recirculation contributes to the overall systemic exposures of RESAGL and RESGLU in rats. The fraction of RESAGL and RESGLU excreted in urine appears to be minimal compared with the biliary elimination pathways. Whether or not enterohepatic recirculation of RESAGL and RESGLU contributes significantly to the overall pharmacological activity remains to be determined.
Acknowledgments
We thank Isabelle Ramier and Vivian Beausoleil from MDS Pharma Services for excellent technical assistance.
Footnotes
-
DOI: 10.1124/jpet.102.033340
- Abbreviations:
- RESAGL
- resveratrol in its aglycone form
- RESGLU
- total glucuronide metabolites of resveratrol
- Ae0–12
- cumulative amount of drug excreted in urine from time 0 to 12 h
- AUC0–t
- area under the curve from time 0 to the last measurable plasma concentration
- AUCinf
- area under the curve extrapolated to infinity
- CL
- clearance
- CL/F
- oral clearance
- CL/Fm
- apparent clearance of metabolite
- Cmax
- maximum observed plasma concentration
- %ER
- percentage of the exposure due to enterohepatic recirculation
- ESI-LC/MS/MS
- electrospray ionization liquid chromatography tandem mass spectrometry
- MRT
- mean residence time
- T1/2
- apparent elimination half-life before enterohepatic recirculation was thought to occur
- T1/2TER
- terminal elimination half-life with enterohepatic recirculation
- Tmax
- time of maximum observed plasma concentration
- Varea/F
- apparent volume of distribution after oral administration
- Varea/Fm
- apparent volume of distribution of metabolite
- Vss
- total volume of distribution
- Received January 18, 2002.
- Accepted March 27, 2002.
- The American Society for Pharmacology and Experimental Therapeutics