Pharmacokinetics and cellular uptake of imatinib and its main metabolite CGP74588
- First Online:
- Cite this article as:
- le Coutre, P., Kreuzer, K., Pursche, S. et al. Cancer Chemother Pharmacol (2004) 53: 313. doi:10.1007/s00280-003-0741-6
- 396 Views
Despite the remarkable clinical response rates to imatinib in the treatment of bcr-abl leukemic patients, pharmacokinetic data on this relatively novel substance are needed to improve our understanding of the emergence of resistance, the interindividual variations of clinical response and the clinical and biologic relevance of its main metabolite N-desmethyl-imatinib. We present here pharmacokinetic data obtained with a newly designed HPLC approach in 97 patients with chronic myeloid leukemia or acute lymphatic leukemia (ALL) under treatment with imatinib that allowed us to calculate the AUC (39.5 μg·h/ml for an oral dose of 400 mg daily), the t1/2 (18.2 h) and the peak concentration (1.92 μ/ml for an oral dose of 400 mg daily) of imatinib in plasma. In a subgroup of patients, the same parameters were analyzed for N-desmethyl-imatinib. We also provide data on the imatinib concentration in the cerebrospinal fluid (CSF) of ALL patients and demonstrate that oral administration of imatinib resulted only in a marginal flux across the blood-brain barrier. Finally, in an in vitro setting, we determined cellular concentrations of imatinib in HL-60 cells and showed an over-proportional uptake both in RPMI medium and in human plasma. Using an arithmetical approach combining all parameters obtained in imatinib-treated patients, we finally provide a conclusive approximation of basic pharmacokinetic data for both imatinib and its main metabolite N-desmethyl-imatinib.
The reciprocal t(9;22)(q34;q11) translocation known as the Philadelphia chromosome is the causative chromosomal abnormality in the pathogenesis of more than 95% of all cases of chronic myeloid leukemia (CML) and is also detectable in some cases of acute lymphatic leukemia (ALL). At the molecular level, the Philadelphia chromosome generates the bcr-abl fusion transcript that results in the expression of the oncogenic bcr-abl protein tyrosine kinase. Previous studies have demonstrated that the constitutive tyrosine kinase activity of this protein is the crucial event in the genesis of bcr/abl+ CML and ALL that leads to blocking of apoptosis and ultimately results in the specific leukemic phenotype.
Imatinib mesylate (Glivec, Gleevec, formerly known as STI571) represents a novel treatment modality that is active through selective inhibition of the ATP-binding pocket present in the abl domain of the bcr-abl tyrosine kinase [1, 2]. Consequently, imatinib results in continuous inactivation of bcr-abl tyrosine kinase activity and eradicates the leukemic phenotype by reconstitution of apoptosis in bcr-abl+ cells [3, 4, 5]. Previous clinical trials in patients with chronic-phase disease who have failed interferon-α treatment have demonstrated remarkable hematologic and cytogenetic response rates . A currently ongoing phase III clinical trial has demonstrated the superiority of imatinib over interferon-α in inducing a cytogenetic response within the first 2 years of treatment in patients with chronic-phase disease and in an overall survival advantage in the first year of treatment in combination with fewer grade 3/4 adverse events . However, long-term results are still lacking. Hematologic and cytogenetic remissions are also present in patients with accelerated phase and blast crisis [8, 9].
Chemically, imatinib is 4-[(4-methyl-1-piperazinyl)methyl]-N-[4-methyl-3-[[4-(3-pyridinyl)-2-pyrimidinyl]amino]-phenyl]benzamide methansulfonate (C30H35N7SO4). In the clinical setting, imatinib is administered orally at a dose of 400 mg daily in patients with chronic phase disease and at 600 mg daily in patients with accelerated phase and blast crisis disease including dose modifications according to adverse events and hematologic response. Despite the increasing number of reports on the clinical effects of imatinib, only few focus on the pharmacokinetic characteristics of imatinib in humans. Moreover, no data about cellular concentrations and uptake of imatinib are available.
However, profound pharmacokinetic data will be essential to answer questions related to clinical aspects of imatinib therapy such as passage through the blood-brain barrier, the elimination of imatinib in patients with acute or chronic renal failure as well as interindividual variation of peak plasma levels. Further, previous studies have shown that potential inactivation of imatinib through excessive binding to the plasma protein α1-acid glycoprotein may result in decreased levels of active drug which consequently may result in the induction of resistance [10, 11, 12, 13].
According to the manufacturer, the mean absolute bioavailability of imatinib when administered as capsules is 98%. About 13% is excreted in the urine, but the main site of degradation is the liver. Here, imatinib is a substrate of the cytochrome P 450 isoenzyme cytochrome P3A4 (CYP3A4). The main metabolite of imatinib after metabolization by CYP3A4 is CGP74588 (C29H33N7SO4) also called N-desmethyl-imatinib. N-desmethyl-imatinib, according to the manufacturer, shows comparable biologic activity regarding the bcr-abl, c-abl, PDGF-R and c-kit tyrosine kinases . However, no data are available that compare imatinib and N-desmethyl-imatinib in an in-vivo setting.
Here, we provide data on the pharmacokinetics and cellular uptake of imatinib and its main metabolite obtained by a newly developed high-performance liquid chromatography (HPLC) assay in 39 patients with CML under treatment with 400, 600 and 800 mg daily. Additionally, we analyzed cerebrospinal fluid (CSF) concentrations of imatinib in 17 patients and 24-h urine elimination of imatinib and N-desmethyl-imatinib in another 7 patients. Moreover, in 58 bcr/abl+ ALL patients, we analyzed on a single sample basis imatinib and N-desmethyl-imatinib concentration ratios. In two CML patients, imatinib and N-desmethyl-imatinib were determined after discontinuation of imatinib prior to allogeneic peripheral stem cell transplantation. Data on the cellular uptake of imatinib and N-desmethyl-imatinib investigated in HL-60 cells in vitro are also presented.
Material and methods
Analyses undertaken in the individual patients
Imatinib dose (mg)
Plasma analysis >24 h
Single-day plasma analysis
Analysis at steady-state
The plasma samples from the first subgroup of the study were derived from 32 CML patients who were treated in at least one of our centers as part of an advanced access program with imatinib (CSTI571 113-115). Plasma samples from 12 of these patients were available over a time period of more than 2 days. Six of these patients were in chronic phase and consequently were treated with 400 mg imatinib daily. The other six patients were in either accelerated phase or blast crisis and were treated with 600 mg imatinib daily. From the remaining 20 CML patients at least three samples per patient were available and analyzed in an average calculation of all data points. All patients from this subgroup were analyzed after the first administration of imatinib but not under pseudosteady-state conditions.
Blood samples from this group were taken before oral intake of imatinib and at various time-points (0.5, 1, 3 and 6, 12, 24, 24.5, 25 and 27 h etc. up to 4 days) thereafter. Samples were then centrifuged and plasma was stored at −20°C pending analysis. Since the HPLC analysis at this time was not validated for the measurement of N-desmethyl-imatinib, the results from this patient subgroup are exclusively imatinib pharmacokinetic parameters. After the validation of the HPLC method for analysis of N-desmethyl-imatinib (see below), the analysis of all samples in the remaining patient subgroups included measurement of N-desmethyl-imatinib.
The second subgroup comprised 56 ALL patients who were treated as part of a randomized, multicenter phase II trial in elderly bcr/abl+ ALL patients (GMALL-STI571-Elderly-01/02). In this group 17 CSF samples were analyzed under pseudosteady-state conditions and single point plasma sample analyses were performed in all individuals.
Finally, a third group comprised seven CML patients in whom 24-h urine elimination was determined and the decrease in plasma imatinib and N-desmethyl-imatinib concentrations were determined after discontinuation of imatinib prior to peripheral blood stem cell transplantation.
Cellular uptake of imatinib was investigated in HL-60 cells incubated in RPMI medium or human plasma at concentrations ranging from 150 ng/ml to 50 μg/ml. In total, three independent experiments over a time range of 6 weeks were performed. All measurements were done in triplicate. In addition to the intracellular imatinib analyses, the concentrations in the incubation medium (RPMI and human plasma) were determined at the starting point and at the endpoint in order to validate the analysis. Also all washing solutions were analyzed for imatinib residues. As expected, no metabolism of imatinib by HL-60 cells was found (data not shown). Therefore, the incubation system could be defined as a “closed system” and the sum of the imatinib in all compartments should amount to 100% in relation to the starting concentration. Preliminary experiments focused on the distribution of imatinib within the various compartments of this system. In fact, we found that the sum of all amounts of imatinib in all compartments ranged from 75% to 125%, probably reflecting measurement variations as well as the surface activity of imatinib in aqueous solution (data not shown). We also determined the rediffusion of intracellular imatinib to the washing solution and found that about 10% of the intracellular imatinib was redistributed within 5 min to the surrounding medium (data not shown). Based on this observation the centrifugation time for the washing step was exactly 5 min in each experiment.
For each cellular sample, 2×107 cells were incubated in 5 ml of either RPMI or human plasma at various concentrations of imatinib as indicated above. After an incubation time of 5 h at 37°C in an atmosphere containing 5% CO2, the cells were centrifuged at 1000 g and the resulting pellets were dissolved in 50 ml isotonic NaCl and subsequently again pelleted by centrifugation. After discharging the supernatant the sample pellets were treated for imatinib extraction similar to plasma (see below).
To determine the cellular uptake of N-desmethyl-imatinib we extracted and purified N-desmethyl-imatinib by liquid–liquid extraction from patient urine and subsequently purified it by the newly developed HPLC method. The resulting N-desmethyl-imatinib solution showed a purity of 99% as tested by analytical HPLC. N-desmethyl-imatinib authenticity was proved by the similarity of UV spectra and existed solely in the plasma or urine of imatinib-treated patients. Finally, N-desmethyl-imatinib authenticity was additionally proved by magnetic resonance spectroscopy (data not shown). Due to the restricted availability of N-desmethyl-imatinib, uptake analysis in HL-60 cells was only performed on a smaller incubation concentration scale than imatinib and solely in RPMI medium.
HPLC and sample preparation
The newly developed HPLC method for analyzing imatinib and N-desmethyl-imatinib will be described in full elsewhere. Briefly, the chromatographic system comprised two Knauer 64 analytical HPLC pumps (Knauer Corporation, Berlin, Germany), a Shimadzu UV spectrometric detector SPD-SA (Shimadzu, Duisburg, Germany), a Waters WISP 712 autoinjector and an electric motor-driven autoswitch equipped with a rheodyne valve 7740-001 for online enrichment switching (Besta-HPLC-Technik, Wilhelmsfeld, Germany). Recording, evaluation and quantitation of chromatograms was done by a PC supported the GINA program from Raytest (Straubenhardt, Germany).
A ZirChrom analytical HPLC column (3 μm, PDB-ZrO2, 3% carbon, 50×4×6 mm) with a precolumn of the same solid phase specificity was used. The system was designed as an online enrichment system with another PDB-ZrO2 precolumn as enrichment column. Flow was set at 0.4 ml/min at room temperature for the analytical part and at 2 ml/min at room temperature for the enrichment part. The analytical eluent consisted of 600 ml 0.01 M KH2PO4/0.09 M K2HPO4 + 400 ml methanol, while the enrichment eluent consisted of 450 ml 0.1 M KH2PO4 + 350 ml H2O + 200 ml CH3OH. For quantitation the external standard method was used by regression analysis of six spiked plasma samples with 10 ng/ml, 100 ng/ml, 500 ng/ml, 1000 ng/ml, 10 μg/ml, 25 μg/ml and 50 μg/ml for imatinib and, after validation for the analysis of the main metabolite, with purified N-desmethyl-imatinib, respectively. The analysis was performed by UV detection at 260 nm.
This system had a detection limit of 10 ng/ml for imatinib and N-desmethyl-imatinib using 300 μl plasma, urine or CSF. Within-day variation was 7% for imatinib and 9% for N-desmethyl-imatinib at a concentration of 100 ng/ml. Daily variation determined on ten consecutive days with plasma spiked with 100 ng/ml of imatinib and N-desmethyl-imatinib was 4% and 6%, respectively. At the detection limit, the coefficient of variation was 11% for imatinib and 14% for N-desmethyl-imatinib as proven by ten measurements with spiked plasma samples. Similar results were found for the analysis in urine. For sample preparation, 30 μl concentrated perchloric acid was added to 300 μl plasma or RPMI and rapidly shaken in an autoshaker for 10 min. Afterwards, 200 μl of the enrichment eluent were added to the sample and the mixture was shaken for another 10 min. Proteins were then precipitated by centrifugation for 7 min at 4000 g. Of the resulting supernatant, 200 μl was injected into the HPLC system for analysis. CSF and urine were diluted with enrichment eluent (50/50, v/v) and after 7 min centrifugation at 4000 g a volume of 200 μl was injected without further preparation.
As a weighting function for the measured data 1/y2 was used. The plasma decay curve of N-desmethyl-imatinib was determined independently of the respective data of the parental compound, assuming a first-order rate constant of metabolism. Kinetic parameters as derived from this procedure revealed a close correlation with the results calculated by applying a linear user-defined compartment model in which the measured data of imatinib and N-desmethyl-imatinib were fitted together. Since the amounts of imatinib that had been absorbed and metabolized were unknown, dose-dependent data such as the distribution volume at pseudosteady-state (Vss) and total body clearance (clearancetotal) were approximated by assuming a mean bioavailability of 100% for imatinib and a mean metabolized amount of 20% for N-desmethyl-imatinib, as suggested by published and total urine elimination data.
Imatinib was kindly provided by Novartis Pharma (Basel, Switzerland). The compound was dissolved in sterile water, filtered and stored at −20°C at a concentration of 10 mM.
Peak plasma concentrations, AUC and terminal t1/2 of imatinib
Pharmacokinetic parameters calculated from the individual plasma curves and mean of all data sets including the mean plasma curves calculated proportionately for 600 mg/day for patients 1–32
Mean—all data pointsb
Mean—all data pointsb
All data sets
The AUC showed an over-proportional level of 85.9 μg·h/ml for patients treated with 600 mg compared to 37.5 μg·h/ml for those treated with 400 mg. The discrepancy in these AUC values mainly resulted from patient no. 4 who showed an extreme AUC of 197 μg·h/ml due to a very high plasma concentration of 15.6 μg/ml. In this patient the half-life of imatinib remained in the normal range (17.8 h). Excluding this patient from calculation of the mean of the 600-mg group resulted in a nearly proportional AUC value compared to the 400-mg group (63.6 versus 37.5 μg·h/ml).
CSF analysis and renal elimination of imatinib
Pharmacokinetic data for patients 33–89 treated with 400 or 600 mg/day imatinib (values are means of the indicated patient groups) (n.a. not available, i.e. no sample)
Percent of dose
The mean imatinib concentration of the CSF from all 17 patients analyzed was 38 ng/ml (39 and 34 ng/ml in CSF from patients without and with meningeosis, respectively) indicating that only a small fraction of plasmatic imatinib crosses the blood-brain barrier, regardless of the presence of meningeosis leukemica. N-Desmethyl-imatinib was not clearly detectable in CSF samples, but peak levels were below 10 ng/ml, reflecting a similar imatinib/N-desmethyl-imatinib ratio as found in plasma samples. Since this level was under the evaluated detection limit of the HPLC method, we do not present the N-desmethyl-imatinib levels in CSF as a numerical result.
Imatinib pharmacokinetics after discontinuation of imatinib prior to BMT conditioning therapy in two patients treated with 500 and 800 mg imatinib, respectively. Dose-dependent parameters were calculated for the first dose administered assuming 100% absorption of the imatinib dose
Imatinib in urine (% of dose)
N-desmethyl-imatinib pharmacokinetics after discontinuation of imatinib prior to BMT conditioning therapy in two patients treated with 500 and 800 mg imatinib, respectively. Dose-dependent parameters were calculated for the first dose administered assuming 100% absorption and 20% metabolism of the imatinib dose
N-desmethyl-imatinib in urine (% of dose)
Cellular uptake of imatinib and N-desmethyl-imatinib
Despite the numerous reports of the clinical efficacy of imatinib and the biologic mechanisms involved, only a limited amount of data are presently available on its pharmacokinetic properties both in an in-vitro and in an in-vivo setting. Especially in view of the various mechanisms of resistance to imatinib that have been described so far [16, 17, 18], precise pharmacokinetic data in relation to clinical response are needed, as in a significant number of non-responders none of the underlying mechanisms of resistance can be identified.
Bakhtiar et al.  used liquid chromatography-tandem mass spectrometry to quantify both imatinib and its main metabolite CGP74588 in human plasma. In this study, nominal concentrations of imatinib ranging from 4.0 to 10,000 ng/ml could be detected. However, only limited data on patient plasma are presented and pharmacokinetic parameters as well as inter- and intraindividual variations were not evaluated.
In our series of patients, peak plasma concentrations (Cmax) of imatinib were detected after a mean time of 3.84 and 4.07 h following oral intake of 400 or 600 mg imatinib, respectively. These findings are comparable to those recently reported by Gambacorti et al.  who found a Cmax after 1 and 3 h following oral intake of 400 mg imatinib. Likewise, the parameters mean t1/2, AUC and Cmax were comparable in our group of patients (16.6 h, 37.5 μg/ml·h and 2.02 μg/ml in the 400-mg group and 16.9 h, 85.9 μg/ml·h and 6.76 μg/ml in the 600-mg group) to those reported by Gambacorti et al. (12.5 h, 24.66 μg/ml·h and 2.352 μg/ml in the 400-mg group and 12.5 h, 99.74 μg/ml·h and 7.83 μg/ml in the 600-mg group). Patient no. 4 showed excessive AUC and Cmax values but with a half-life of imatinib in the usual range. Clinically, this patient had a 6.5-year history of CML and was treated with 600 mg imatinib because of the presence of accelerated phase for 32 months with continuous hematologic response over 28 months. Interestingly, in this patient no signs of severe toxicity were observed, despite the relatively high plasma level of imatinib. The assumption of significant alterations caused by interindividual variability of imatinib is further supported by the observation that intraindividual values of AUC are relatively stable.
The calculated half-life of imatinib in this group of patients regardless of dose was 16.6 h (range 10.2–27.3 h), resulting in an approximate time to reach pseudosteady-state plasma levels of 96–120 h. Further, there was a linear correlation between dose of imatinib and the AUC.
According to the manufacturer, N-desmethyl-imatinib has similar inhibitory effects on the tyrosine kinase activity of bcr/abl, c-kit, PDGF-R and c-abl to those of imatinib. However, the plasmatic ratios of imatinib and all other metabolites and their biologic activity have not yet been established. As the oxygen residues in piperazine-N-oxide-imatinib, hydroxymethyl-phenyl-imatinib and pyridine-N-oxide-imatinib will eventually prevent the binding of these molecules to the imatinib binding pocket and because of their relatively low urinary concentrations (data not shown) this work focused on the biologic activity of N-desmethyl-imatinib.
The mean N-desmethyl-imatinib plasma concentrations in our patients (531 ng/ml) under pseudosteady-state conditions, although only 17% of the imatinib plasma concentration, were above the nominal and effective concentrations of imatinib in an in vitro setting (100 ng/ml), indicating the clinical significance of this substance. However, biologically significant competition between imatinib and its main metabolite N-desmethyl-imatinib for the imatinib-binding site still needs to be evaluated.
It is noteworthy that the t1/2 of N-desmethyl-imatinib greatly exceeded the t1/2 of imatinib. Discontinuation of imatinib in two patients resulted in a rapid decrease in plasma imatinib concentrations while plasma N-desmethyl-imatinib concentrations only moderately declined in the same period (>140 h) (Fig. 2). Likewise, urinary levels of imatinib exceeded those of N-desmethyl-imatinib and were characterized by a more rapid decline (Fig. 2). Based on these results, we derived approximate plasma concentration-time curves and cumulative urinary elimination for imatinib and N-desmethyl-imatinib (Fig. 3). This calculation allowed us to estimate the wash-out phase of both imatinib and N-desmethyl-imatinib after discontinuation of treatment below a plasma concentration of 100 ng/ml. Figure 3 shows that despite relatively low concentrations of the metabolite under steady-state conditions, the wash-out phase of N-desmethyl-imatinib significantly exceeded the plasma elimination of imatinib (200 h versus 100 h). This observation may be of clinical significance, as it is still unclear whether continuation of imatinib treatment might increase the intensity of toxicity in myeloablative conditioning therapy in allogeneic stem cell transplantation.
Further, our results confirm a renal elimination of imatinib of 5–10% and a renal clearance of about 10 ml/min, indicating that imatinib pharmacokinetics will not be significantly affected in patients with compensated renal failure. Although there are no data on the issue of hemofiltration and imatinib levels, we assume that because of the high distribution volume and plasma binding, dialysis may not substantially interfere with imatinib plasma levels. Overall, based on these findings we can assume that the main elimination route of imatinib and its main metabolite is via the intestine.
In a subgroup of our patients we additionally compared imatinib levels in CSF samples obtained from ALL patients with or without meningeosis leukemica. As expected, intrathecal concentrations of imatinib were much lower in the CSF than in the plasma, indicating limited penetration of imatinib across the blood-brain barrier. In 14 patients with no signs of meningeal leukemia imatinib concentrations were approximately the same (38 ng/ml) as in three patients with proven meningeosis leukemica (34 ng/ml). Apparently, even in patients in whom the integrity of the blood-brain barrier is impaired by meningeosis leukemica, oral treatment with imatinib will not be a therapeutic option.
Although the involvement of the MDR1 gene (multiple drug resistance 1 gene) in the intracellular pharmacokinetics of imatinib is yet not fully understood, it is of note that p-glycoprotein overexpression has been shown to confer resistance to imatinib in a leukemia cell line model . Further, Dai et al.  have demonstrated that intracellular flux of imatinib can be significantly altered by administration of the p-glycoprotein inhibitor LY335979 and that the brain-to-plasma concentration ratio of imatinib is elevated in mdr1a/b−/− knockout mice. These findings indicate a potential role for p-glycoprotein in the transfer of imatinib to the CSF. We therefore speculate whether the concomitant administration of verapamil, an effective inhibitor of p-glycoprotein, could become a therapeutic option in this group of patients.
As previous studies have demonstrated the emergence of intrinsic resistance in vitro (i.e. amplification of the Philadelphia chromosome) on exposure of cells to increasing concentrations of imatinib, the equilibrium of extra- and intracellular imatinib concentrations certainly plays a critical role in the emergence of resistant clones [16, 17].
The determination of intracellular imatinib concentrations in HL-60 cells demonstrated a dramatic influx of imatinib when compared to the extracellular concentrations as illustrated by the hyperbolic increase in cellular imatinib uptake from both human plasma and RPMI medium (Fig. 4). These results lead to the conclusion that apparently minor alterations in extracellular imatinib concentrations may result in a dramatic shift in the intracellular imatinib concentration. This observation was most significant at extracellular concentrations ranging between 2 and 20 μg/ml, indicating that an active transporter mechanism may be involved in the intracellular uptake of imatinib. However, we admit that clinically relevant concentrations are below this range. Nevertheless, a recent study has demonstrated that incubation of K562 cells with a variety of tyrosine kinase inhibitors including imatinib results in a dose-dependent reduction in intracellular [3H]uridine uptake, indicating that tyrosine kinase inhibitors may eventually affect nucleoside transport .
Additionally, our results add to the observation that in vivo resistance to imatinib in mice may be mediated by increased protein binding through α1-acid glycoprotein [10, 11, 12, 13]. RPMI medium has a lower protein concentration than human plasma and therefore will bind imatinib less effectively. We therefore speculate that the higher cellular influx of imatinib into HL-60 cells incubated in RPMI was a consequence of a relatively low protein content in the medium.
Finally, we also provide data on the cellular uptake of N-desmethyl-imatinib (Fig. 5). Here, we observed an overproportional intracellular increase similar to the cellular uptake of imatinib and a tendency to higher intracellular concentrations, i.e. better uptake compared to imatinib.
In view of the relatively wide range of peak plasma concentrations observed in individual patients, we finally assume that the oral dosages of imatinib administered to patients do not necessarily guarantee effective concentrations in the target cell. Assuming that plasma imatinib concentrations do reflect the intracellular concentration, therapeutic imatinib levels should always exceed a threshold range of 0.1 to 5 μM. Therefore, thorough determination of imatinib levels could be considered essential in patients lacking adequate hematologic and cytogenetic responses.
P. le Coutre, K.-A. Kreuzer, O. Ottmann and G. Ehninger are principal investigators/coinvestigators of clinical trials with imatinib carried out by Novartis Pharmaceuticals.