The Influence of St. John's Wort on the Pharmacokinetics and Protein Binding of Imatinib Mesylate
Patrick F. Smith, Pharm.D.; Julie M. Bullock, Pharm.D.; Brent M. Booker, Pharm.D.; Curtis E. Haas, Pharm.D.; Charles S. Berenson, M.D.; William J. Jusko, Ph.D. Supported in part by a Research and Education Award from the New York State Council of Health-System Pharmacists, Albany, New York, and grant GM24211 from the National Institutes of Health, Bethesda, Maryland.
Pharmacotherapy 24(11):1508-1514, 2004. © 2004 Pharmacotherapy Publications Posted 12/17/2004
Abstract and Introduction
Study Objective: To determine the effect of St. John's wort
on the pharmacokinetics of imatinib mesylate.
Imatinib mesylate (Gleevec; Novartis Pharmaceuticals, East Hanover, NJ) recently received expedited approval in the United States as an orphan drug for the treatment of chronic myeloid leukemia, and it has rapidly become a cornerstone of therapy. However, this accelerated approval has been accompanied by a relative lack of drug-interaction studies typical of the standard drug-development process. Drug interactions with imatinib are of particular clinical concern, as this agent is metabolized primarily by cytochrome P450 (CYP) 3A4, making it susceptible to significant increases or decreases in systemic exposure by the many agents known to alter CYP3A4 activity. Therapeutic outcomes have been shown to correlate with both dose and drug concentrations; therefore, coadministered agents that alter imatinib metabolism may significantly compromise the clinical efficacy of this important new therapy.
The use of alternative medicine has increased tremendously over the last decade and is particularly common in patients with cancer. Recent surveys have reported that approximately one third to one half of these patients use some type of alternative medicine, such as vitamins and herbal products.[3-5] Because patients may not consider these alternative products to be drugs, they frequently do not report their use of such agents to their physician or pharmacist.
St. John's wort is commonly used for mood elevation in patients with cancer and is known to induce both CYP3A4 and P-glycoprotein.[7, 8] It appears that hyperforin, the active component of St. John's wort, binds and activates the pregnane X receptor, resulting in an increase in messenger RNA expression of CYP3A4. Therefore, a significant drug interaction could result in patients being treated with imatinib and concurrently taking St. John's wort.
The objective of this study was to evaluate the effect of St. John's wort on the pharmacokinetics of imatinib. A brief report of the induction of imatinib metabolism by St. John's wort has been published previously.
The study protocol was approved by the Veteran's Administration Western New York Healthcare System's institutional review board, and written informed consent was obtained from each subject before study participation.
All subjects were judged to be healthy as determined by a medical history, complete physical examination, and laboratory tests. Each subject was required to be 18-55 years of age, with normal renal and hepatic function. Exclusion criteria were pregnancy or lactation (in women), a history of drug or alcohol abuse, use of any investigational agents within 30 days of enrollment in the study, and use of any prescription or nonprescription drugs or supplements known to interact with CYP3A4 metabolism.
This 18-day, open-label, complete crossover, fixed-sequence, pharmacokinetic study evaluated the pharmacokinetics of a single oral dose of imatinib, either when administered alone or after 2 weeks of treatment with St. John's wort (HBC Inc., Los Angeles, CA). After an overnight fast, subjects received imatinib mesylate 400 mg with a standardized low-fat meal. Blood samples were collected at 0, 0.5, 1, 1.5, 2, 3, 4, 6, 8, 12, 24, and 48 hours after dosing. Plasma was harvested immediately after each sample was collected, and was stored at -70°C until analysis.
After collection of the 48-hour blood sample, treatment with St. John's wort 300 mg 3 times/day was started. After 2 weeks, subjects again received a single oral dose of imatinib mesylate 400 mg with a standardized low-fat meal after an overnight fast. Blood collection and processing were the same as earlier in the study. Subjects were contacted periodically to assess adverse effects and compliance with the St. John's wort regimen, and each subject completed a dosing diary.
Plasma concentrations of imatinib were determined by a validated liquid chromatography with tandem mass spectrometry detection (LC-MS-MS) assay (Perkin Elmer SciEx API 3000; Perkin Elmer, Wellesley, MA) with a heated nebulizer interface. The assay was based on a previously described method, adapted to suit our instrumentation. The column used was the 3.5-µm Symmetry Shield TM RP8, 4.6 x 50 mm (Waters Corp., Milford, MA) with a 1-ml/minute flow rate. The assay uses D8-STI571 (Novartis Pharmaceuticals) as an internal standard. Free (unbound) concentrations of imatinib were also determined for all 3- and 24-hour samples by plasma ultrafiltration at 37°C (Centrifree; Amicon, Inc., Beverly, MA) before assay. The LC-MS-MS assay had a lower and upper limit of quantitation of 2 and 3000 ng/ml, respectively. The interday and intraday coefficient of variation were both less than 5%.
Pharmacokinetic parameters were determined by standard noncompartmental methods (WinNonlin 3.3; Pharsight Corp., Cary, NC). The pharmacokinetic parameters for imatinib administered alone versus with St. John's wort were compared by using a paired t test on logarithmically transformed data (SYSTAT 10.0; SPSS Inc., Chicago, IL). A logarithmic transformation was empirically utilized to better approximate a Gaussian distribution function, an underlying assumption of the paired t test.
For the purpose of these statistical analyses, it was assumed that there were no period effects and that 14 days was an adequate washout period between experiments. Bioequivalence of the area under the conentration-time curve (AUC) and maximum observed concentration (Cmax) was evaluated between the two study periods (taking imatinib alone or with St. John's wort) by using 90% confidence intervals of the geometric mean ratios, equivalent to the two one-sided test. In accordance with standard Food and Drug Administration (FDA) guidelines, the 90% confidence intervals of the geometric mean ratios were required to fall between 0.8 and 1.25 in order to declare bioequivalence. Sample-size calculations were performed to have sufficient power to detect a 40% change in AUC for imatinib in the presence or absence of St. John's wort, with use of a paired t test and an intersubject coefficient of variation of 40% (α=0.05, β=0.8). It was determined that 10 subjects would be required to detect a 40% change in AUC. A p value of 0.05 or less was considered to indicate a statistically significant difference.
Seven men and three women, with a mean ± SD age of 43.7 ± 6.4 and 33.3 ± 6.0 years, respectively, were enrolled in and completed the study. The mean body weight was 80.5 ± 6.0 kg. The treatments were well tolerated, with no adverse events reported during the study.
The pharmacokinetics of imatinib were significantly altered after administration of St. John's wort (Figure 1). Pharmacokinetic parameters are summarized in Table 1. The median imatinib AUC from time zero to infinity (AUC0-) was reduced by approximately 32% (28.9 vs 19.7 µg•hr/ml, p=0.0001). This reduction in imatinib exposure appears to be a result of both a decrease in oral absorption and an increase in the rate of elimination, with a 29% reduction in Cmax and a 21% reduction in half-life (p<0.01 for both). The 48-hour concentration (Clast) of imatinib was also significantly reduced by St. John's wort, by approximately 55% (p=0.00002).
Figure 2 illustrates the influence of St. John's wort on imatinib AUC0- in each subject. Subjects with higher AUC0- values at baseline were most susceptible to the enzyme-inducing effects of St. John's wort. The four individuals with a baseline AUC0- of greater than 30 µg•hour/ml had a decrease in AUC0- of 41.7%, 40.8%, 39.6%, and 32.2%, respectively. Reductions of only 20.1% and 11.7% in AUC0- were observed in the two individuals who had a baseline of less than 20 µg•hour/ml. Both AUC0- and Cmax failed the bioequivalence test, with the 90% confidence intervals of the geometric mean ratios falling outside of the required 80-125% (56% [range 46-70%] for AUC0- and 76% [range 59-87%] for Cmax).
Imatinib was approximately 95% protein bound in all subjects, with little intersubject variability (mean 94.9 ± 1.4%, range 90.3-97.7%). Protein binding was concentration independent over the range studied (0.12-2.6 µg/ml) and was not altered by St. John's wort. These protein-binding results are similar to the values reported in imatinib's product labeling.
The primary objective of this study was to evaluate the effect of St. John's wort on the pharmacokinetics of imatinib in healthy volunteers. Compared with imatinib administered alone, imatinib given after 2 weeks of treatment with St. John's wort resulted in a significant reduction in imatinib exposure. The median AUC was reduced by 32%, Cmax was reduced by 29%, and half-life was reduced by 21%. The median apparent oral clearance increased from 13.8 to 20.3 L/hour. Imatinib was highly protein bound (~95%), and protein binding was not affected by the coadministration of St. John's wort.
The clinical significance of this interaction is potentially important. The observed reduction in imatinib exposure, measured by AUC, was approximately 32% on average and as high as 42%. In vitro studies and clinical trials of imatinib have found that clinical responses are correlated with dose and drug concentration, with a dosage of 400 mg/day generally superior to 300 mg/day and a dosage of 600 mg/day superior to 400 mg/day in patients with advanced disease.[2, 14, 15] The 25% difference between a daily dose of 300 versus 400 mg is less than the average reduction observed when imatinib was administered with St. John's wort. When potential metabolic contributions are disregarded, a 400-mg dose of imatinib administered with St. John's wort provided the same exposure as a 272-mg dose given alone.
One patient in a clinical trial failed to respond to imatinib therapy while taking phenytoin, an inducer of CYP3A4. This patient had a very low AUC and experienced a complete response after discontinuing phenytoin and increasing the imatinib dosage. A drug interaction of this magnitude might contribute to imatinib resistance and treatment failure.
Previous clinical studies and case reports have documented that St. John's wort lowers serum drug concentrations in patients taking numerous drugs that are metabolized by CYP3A4, such as cyclosporine, warfarin, oral contraceptives, theophylline, indinavir, amitriptyline, and nortriptyline.[16, 17] These findings led to an FDA public health advisory in February 2000 warning against coadministration of St. John's wort with these agents. Based on the results of this study, St. John's wort also should be avoided in patients treated with imatinib. In cases where St. John's wort is coadministered, a dosage increase or monitoring of imatinib concentrations might be considered.
One earlier drug-interaction study involved imatinib and rifampin in healthy adult volunteers. In that study, the single-dose AUC0- of imatinib 400 mg and its primary active metabolite (CGP74588) were reduced by 74% and 12%, respectively, after administration of rifampin for 8 days. Rifampin is a potent inducer of CYP3A4, CYP2B6, CYP2C9, CYP2C19, and CYP2D6 metabolism, and it appears to be a more potent inducer than St. John's wort. The magnitude of the effect of St. John's Wort on the AUC of imatinib was generally similar to that reported for other CYP3A4 substrates, such as simvastatin, nortriptyline, fexofenadine, and irinotecan; somewhat less than the interaction reported for indinavir; and greater than that reported for amitriptyline and carbamazepine.
In addition to the induction of CYP3A4 metabolism, recent data suggest that St. John's wort is an inducer of P-glycoprotein in the gastrointestinal tract and that imatinib is a substrate for P-glycoprotein. The observed reduction in the Cmax of imatinib may be a result of induction of P-glycoprotein or gastrointestinal CYP3A4, or both, by St. John's wort, leading to a reduction in bioavailability. Although CYP3A4 is the major enzyme responsible for imatinib metabolism, CYP1A2, CYP2D6, CYP2C9, and CYP2C19 could play minor roles.
In addition to the deleterious effect of St. John's wort on the pharmacokinetics of imatinib, the herbal product may affect the activity of imatinib in patients undergoing therapy for chronic leukemia. One of several resistance mechanisms of imatanib appears to be related to P-glycoprotein-mediated drug efflux in leukemia cells.[2, 25] Some cells that are resistant to imatinib overexpress P-glycoprotein,[2, 25, 26] which is reversed through an increase in the intracellular accumulation of imatinib by P-glycoprotein inhibitors such as verapamil and cyclosporine. Thus, there is the potential for St. John's wort to reduce the activity of imatinib against leukemic cells through the induction of P-glycoprotein, leading to an increase in drug efflux. It is unknown whether this effect might be clinically relevant or observed differentially between wild-type cells and resistant cells that overexpress P-glycoprotein. Based on these concerns, there may be a deleterious pharmacodynamic interaction between St. John's wort and imatinib, in addition to the pharmacokinetic interaction.
St. John's wort did not alter the protein binding of imatinib over a wide range of concentrations in vivo. Thus, we are able to confirm that both free and total imatinib concentrations are reduced by St. John's wort. This may be important, as other reports have demonstrated a potential association between α1-acid glycoprotein concentrations and resistance both in vitro and in vivo.[27, 28] Although controversial, these observations suggest that a reduction in free imatinib concentrations, induced by increasing α1-acid glycoprotein concentrations, may lead to imatinib resistance at clinically relevant concentrations.
The primary in vivo circulating metabolite of imatinib is CGP74588, an N-demethylated piperazine derivative that has demonstrated similar in vitro potency as the parent compound. The contribution of this metabolite to the overall activity of imatinib is unknown. Although we did not evaluate the effect of St. John's wort on the pharmacokinetics of CGP74588, the plasma AUC of this metabolite is 16% that of the imatinib AUC. In addition, CGP74588 is a substrate of CYP3A4, which suggests that St. John's wort might induce the metabolism of this compound. In future drug-interaction studies involving imatinib and CYP3A4 inducers or inhibitors, investigators should attempt to further evaluate alterations in CGP74588 pharmacokinetics.
It is also apparent that certain individuals are more susceptible to the metabolic induction effects of St. John's wort. Subjects with the highest imatinib AUC at baseline experienced a greater reduction in imatinib exposure after the administration of St. John's wort. These findings suggest that some individuals have relatively low baseline CYP3A4 activity, which renders them more susceptible to the induction effects of drugs, placing them at higher risk for major drug interactions. This observation has been reported with drug interactions involving antiretroviral agents.
It should be noted that this study did not involve direct observation of subjects ingesting St. John's wort. Compliance with the St. John's wort dosing regimen of 3 times/day, although reported to be excellent by all study participants, cannot be ensured. Hence, it is possible that the full induction capacity of St. John's wort on imatinib was not achieved in all subjects if compliance was less than 100%. It is also possible that patients undergoing imatinib therapy who have a substantially compromised health status compared with that of volunteers may exhibit greater changes than the one third decrease in imatinib exposure noted in this study.
St. John's wort is available as an over-the-counter dietary supplement in several formulations, produced by numerous manufacturers. As dietary supplements are not strictly regulated, the content and bioavailability of the active ingredients -- hyperforin and hypericin -- may vary considerably by product. The formulation of St. John's wort that was used in this study was the same as that used in several National Institutes of Health studies, which demonstrated significant drug-drug interactions. Based on the differences in content and bioavailability of active ingredients in commercial St. John's wort products, the magnitude of this drug-drug interaction with imatinib may vary, making these interactions less predictable or consistent than those solely involving FDA-regulated prescription drugs.
Administration of imatinib in patients taking St. John's wort significantly affected the pharmacokinetics of imatinib. Clinicians should be aware that St. John's wort may reduce imatinib exposure by 30-40% and should take appropriate actions to educate patients receiving imatinib. Awareness of the potential interactions between herbal products and anticancer agents will enable clinicians to take appropriate steps to ensure that such interactions do not unnecessarily compromise therapeutic outcomes.
Presented in part at the annual meeting of the American College of Clinical Pharmacy, Atlanta, Georgia, November 1-4, 2004.
Table 1. Pharmacokinetic Parameters of Imatinib Mesylate Administered Alone and With St. John's Wort
We appreciate the excellent nursing assistance of Denise Cloen and the superb technical contributions of Suzette Mis and Nancy Pyszczynski.
Supported in part by a Research and Education Award from the New York State Council of Health-System Pharmacists, Albany, New York, and grant GM24211 from the National Institutes of Health, Bethesda, Maryland.
Address reprint requests to Patrick Smith, Pharm.D., Roswell Park Cancer Institute, Elm and Carlton Streets, Buffalo, NY 14263.
From the School of Pharmacy and Pharmaceutical Sciences, University of Buffalo (Drs. Smith, Bullock, Booker, Haas, and Jusko); Roswell Park Cancer Institute (Drs. Smith, Bullock, and Booker); and the Clinical Research Center, Veteran's Administration Western New York Health Care System (Drs. Smith, Haas, and Berenson), Buffalo, New York.