Drug Metabolism and Disposition Fast Forward
First published on May 31, 2007; DOI: 10.1124/dmd.107.014746
0090-9556/07/3509-1554-1563$20.00
DMD 35:1554-1563, 2007
In Vitro Metabolic Study of Temsirolimus: Preparation, Isolation, and Identification of the Metabolites
Ping Cai,
Rushung Tsao, and
Mark E. Ruppen
Chemical and Pharmaceutical Development, Wyeth Research, Pearl River, New York
(Received January 29, 2007;
accepted May 30, 2007)
 |
Abstract
|
|---|
The in vitro metabolism of temsirolimus, (rapamycin-42-[2,2-bis-(hydroxymethyl)]-propionate), an antineoplastic agent, was studied using human liver microsomes as well as recombinant human cytochrome P450s, namely CYP3A4, 1A2, 2A6, 2C8, 2C9, 2C19, and 2E1. Fifteen metabolites were detected by liquid chromatography (LC)-tandem mass spectrometry (MS/MS or MS/MS/MS). CYP3A4 was identified as the main enzyme responsible for the metabolism of the compound. Incubation of temsirolimus with recombinant CYP3A4 produced most of the metabolites detected from incubation with human liver microsomes, which was used for large-scale preparation of the metabolites. By silica gel chromatography followed by semipreparative reverse-phase high-performance liquid chromatography, individual metabolites were separated and purified for structural elucidation and bioactivity studies. The minor metabolites (peaks 1-7) were identified as hydroxylated or desmethylated macrolide ring-opened temsirolimus derivatives by both positive and negative mass spectrometry (MS) and MS/MS spectroscopic methods. Because these compounds were unstable and only present in trace amounts, no further investigations were conducted. Six major metabolites were identified as 36-hydroxyl temsirolimus (M8), 35-hydroxyl temsirolimus (M9), 11-hydroxyl temsirolimus with an opened hemiketal ring (M10 and M11), N- oxide temsirolimus (M12), and 32-O-desmethyl temsirolimus (M13) using combined LC-MS, MS/MS, MS/MS/MS, and NMR techniques. Compared with the parent compound, these metabolites showed dramatically decreased activity against LNCaP cellular proliferation.
Temsirolimus (sirolimus-42-[2,2-bis-(hydroxymethyl)]-propionate) is an ester analog of rapamycin (Fig. 1), a natural macrolide antibiotic with antifungal, antitumor, and immunosuppressive activities (Sehgal et al., 1994
). Temsirolimus has demonstrated significant inhibition of tumor growth both in vitro and in vivo. It binds to the cytoplasmic protein FKBP, forming a complex that antagonizes the mammalian target of rapamycin (mTOR) signaling pathway (Peralba et al., 2003), which consequently inhibits many of the downstream processes affected by mTOR kinase activity, including transcriptional and translational control of important cell cycle regulators, resulting in cell cycle arrest (Sehgal 1995). Temsirolimus is currently in phase III clinical development for the treatment of renal cancer.
View larger version (30K):
[in this window]
[in a new window]
|
FIG. 1. Representative HPLC chromatograms of temsirolimus incubation products in (top) HLM control, (1b) HLM/NADPH-regenerating system, (1c) recombinant 3A4/NADPH-regenerating system; and (1d) recombinant 3A4 control. AU, arbitrary units.
|
|
Although CYP3A4 had been identified as the major enzyme involved in the formation of the main metabolite of rapamycin, 41-O-desmethyl rapamycin (Christiants et al., 1992; Sattler et al., 1992; Wang et al., 1994), little additional information about rapamycin metabolism has been reported because of the complexity of the metabolic profile and the instability of the metabolites. Thus, the objective of this study was to identify the human P450 enzymes involved in temsirolimus metabolism, isolate and characterize the major metabolites, and evaluate their antitumor activities.
|
Materials and Methods
|
|---|
Chemicals and Reagents. Temsirolimus and recombinant cytochrome P450s (3A4, 1A2, 2A6, 2C8, 2C9, and 2E1) were obtained from the Bioprocess Department of Wyeth Research (Pearl River, NY). Human liver microsomes were purchased from BD Biosciences (Woburn, MA). Reagents for the NADPH generation system (disodium salt of NADP, D-glucose 6-phosphate, and glucose-6-phospate dehydrogenase) were purchased from Sigma-Aldrich (St. Louis, MO). Other reagent-grade chemicals and HPLC-grade solvents were purchased from EM Science (Gibbstown, NJ) or J. T. Baker (Phillipsburg, NJ).
Incubation of Temsirolimus with Human Liver Microsomes. Temsirolimus (50 µM) was incubated with 1 mg/ml human liver microsomal proteins in 0.1 M phosphate buffer solution (pH 7.4) containing 1 mM EDTA and 4 mM MgCl2. Reactions were initiated by addition of the NADPH-regenerating system, resulting in a final concentration of 4 mM glucose 6-phosphate, 2 mM NADP, and 1 unit/ml glucose-6-phosphate dehydrogenase. The incubations were carried out at 37°C in a shaking water bath. Control incubations without the NADPH-regenerating system were performed under the same conditions. The reactions were terminated by the addition of cold acetonitrile. Precipitated materials were removed by centrifugation at 5000 rpm for 10 min at 4°C, and the supernatants were collected and dried under vacuum. The dried extracts containing the metabolites were reconstituted in acetonitrile-water (7:3) for HPLC and LC-MS analysis.
Incubation of Temsirolimus with Recombinant Human P450s. Selected recombinant P450s (3A4, 1A2, 2A6, 2C8, 2C9, and 2E1; 0.3 nmol each) were individually incubated with temsirolimus (50 µM). The other components of the reaction mixtures were the same as above for the incubation of human liver microsomes. After incubation for 40 min at 37°C, the reactions were quenched with cold acetonitrile. Subsequent sample treatment was the same as for the human liver microsome samples.
Chemical Inhibition of P450 3A4 Activities. Ketoconazole in ethanol (10 µl) was added to each tube containing the human liver microsomes (1 mg/ml), EDTA (1 mM), MgCl2 (4 mM), temsirolimus (50 µM), and 0.1 M potassium phosphate buffer (pH 7.4), giving a total reaction volume of 900 µl. The final concentration of ketoconazole in the assays ranged from 1 to 100 µM. Duplicate samples were prepared at each concentration. The samples were preincubated for 2 min at 37°C, and then 100 µl of an NADPH-regenerating system (same as described above) was added to initiate the reactions. The reactions were incubated at 37°C for 15 min and then quenched with cold acetonitrile. Subsequent sample processing was the same as described above.
Large-Scale Preparation of Temsirolimus Metabolites. The large-scale incubation was conducted in a 5-liter reactor containing temsirolimus (400 mg in 10 ml of ethanol), recombinant CYP3A4 (1200 nmol), MgCl2 (400 mM x 40 ml), and the NADPH-regenerating system (final concentration: 4 mM glucose 6-phosphate, 1.6 mM NADP, and 0.6 unit/ml glucose-6-phosphate dehydrogenase) in 0.1 M potassium phosphate buffer for a total volume of 4 liters. The incubation was conducted under oxygen (bubbled O2 at 0.30 liter/min) with agitation at 125 rpm. After incubation for 60 min at 37°C, the mixtures were cooled to 25°C and then extracted twice with equivalent volumes of ethyl acetate. The ethyl acetate extracts were combined after the solvent was evaporated under vacuum to yield approximately1gof crude extract.
Isolation of the Metabolites from Large-Scale Preparations. The crude extract was dissolved in 4 ml of acetone and loaded onto a silica gel flash column. The column was sequentially eluted with hexane/acetone and acetone/methanol gradients. A total of 20 fractions were collected (350 ml/fraction) and analyzed by LC-MS. Fractions 1 to 9 consisted mainly of lipids from the CYP3A4 membranes (176 mg), fractions 9 to12 contained unmodified temsirolimus (290 mg), fractions 13 to 17 contained the temsirolimus metabolites (96 mg), and fractions 18 to 20 contained the polar pigments from the membranes (145 mg). Fractions 13 to17 were combined for further separation by semipreparative HPLC on a Supelcosil LC-C18 column (10 x 250 mm, 5 µm). A stepwise gradient consisting of methanol/water with 5 mM ammonium acetate was used at a flow rate of 2.0 ml/min (methanol from 60 to 70% in 30 min and from 70 to 82% in 60 min). Separation was monitored by a UV detector at both 220 and 280 nm. Each metabolite was collected in a separate container on ice as it eluted from the HPLC column. After the organic solvents were removed by vacuum evaporation, the aqueous residues were lyophilized to provide the purified individual metabolites.
HPLC, LC-MS, MS/MS, and NMR Methods. Analytical HPLC was conducted using a Waters Alliance model 2690 HPLC system with a Supelcosil LC-C18 analytical column (4.6 x 250 mm, 5 µm). A gradient of methanol/water with 5 mM ammonium acetate was used as the mobile phase at a flow rate of 0.6 ml/min (methanol from 65 to 85% in 70 min). Semipreparative HPLC was performed on a Supelcosil LC-C18 semipreparative column (10 x 250 mm, 5 µm). A step gradient of water containing 5 mM ammonium acetate/methanol was used as the mobile phase. The gradient was started from 60% methanol and increased linearly to 70% methanol in 30 min, followed by increasing methanol to 82% in 60 min and holding for 5 min. The flow rate of the mobile phase was set at 2 ml/min.
View larger version (15K):
[in this window]
[in a new window]
|
FIG. 2. Expanded semipreparative HPLC chromatograms of temsirolimus incubation products in HLM/NADPH (top) and in recombinant CYP3A4/NADPH (bottom).
|
|
LC-MS analysis was performed on an Agilent 1100 HPLC system connected via 0.005-inch i.d. peek tubing to a MSD1100 mass spectrometer. The above-mentioned analytical HPLC conditions were also applied to LC-MS analysis, and half of the solvent was channeled into the mass spectrometer. Full scan atmospheric pressure chemical ionization and electrospray ionization (ESI) mass spectra were acquired at unit resolution from m/z 200 to 1200 at a step of 0.2. The negative ion mode was recorded.
The MS/MS/CID spectra of the purified metabolites were obtained on an Applied Biosystems-PE Sciex QSTAR PULSAR quadrupole time-of-flight tandem mass spectrometer coupled with a Shimadzu-10Advp HPLC system. Electrospray ionization was conducted in both negative and positive modes. MS/MS/CID experiments were performed using the collision energies from -45 to -50V in the negative mode. For the positive mode, the collision energies were used in a range of 30 to 50 V, depending on the molecular structures. For temsirolimus and analogs, the collision energy was 50 V, but the macrolide ring-opened molecules (seco-temsirolimus type derivatives), the collision energy was 30 V or lower (the molecular ion was not detectable if the CID energy was >30 V). Temsirolimus was used as the internal reference to measure the exact mass of the molecular ion of the metabolite for elemental composition calculation. MS/MS fragment ions and their corresponding elemental compositions were calculated using the measured molecular ion and a common known fragment ion as references.
The LC-NMR data of M8 were acquired on a 600-MHz Bruker DRX spectrometer equipped with a single flow cell LC-NMR probe. The LC separation was performed on an XTerra column (4.6 x 250 mm, 5 µm) with a mobile phase of D2O and acetonitrile-d3 (acetonitrile-d3 from 50 to 90% in 20 min at a flow rate of 0.6 ml/min). Proton spectra were collected with double-solvent suppression. For the other metabolites, the 1H NMR data were acquired on a 500-MHz Bruker DRX 500 spectrometer in acetonitrile-d3.
|
Results
|
|---|
Temsirolimus Metabolism with Human Liver Microsomes and Recombinant CYP3A4. Incubation of temsirolimus with human liver microsomes in the presence of the NADPH-regenerating system indicated that the total metabolite concentration reached a maximum after 40 min. At that time, the concentration of temsirolimus was reduced by 22 ± 3%, which included its conversion to the macrolide ring opened product, seco-temsirolimus (12 ± 3%) (Fig. 1). A representative HPLC chromatogram of the metabolites generated by human liver microsomes is shown in Fig. 1b. The major peak at 25.6 min, which is also observed in the control samples, is seco-temsirolimus identified by direct LC-MS comparison with the authentic standard. Because it was also present in the control samples, this compound was considered to be a nonspecific degradation product that resulted from a hydrolysis of the macrocyclic lactone ring followed by dehydration of C25/C26. The same ring-opening product, seco-rapamycin, had also been detected from the incubation of rapamycin with human liver microsomes and the pooled bile of intravenously dosed rats (Wang et al., 1997). As for seco-rapamycin, seco-temsirolimus did not show any antitumor and FKBP binding activities due to the macrolide ring opening. Compared with the control incubation (Fig. 1a), 15 new peaks appeared in the human liver microsome reaction. Peaks 1 to 7, peak 11', and peak 14 were the minor products, observed at trace levels (each peak area was <0.3% of the total HPLC area of the analyzed sample). Peaks 8 (2.6%), 9 (0.43%), 10 (0.68%), 11 (0.82%), 12 (0.40%), and 13 (0.88%) were the major metabolites designated as M8 through M13 for convenience. The formation of these new peaks was inhibited by bubbling CO through the reaction solution during incubation (data not shown). This observation suggested that biotransformation of temsirolimus required the presence of microsomal P450 monooxygenases.
To identify the enzymes involved in biotransformation, temsirolimus was individually incubated with each of six recombinant human P450 enzymes: 1A2, 2A6, 3A4, 2C8, 2C9, and 2E1. Experimental results indicated that recombinant CYP3A4 generated almost all of the metabolites detected in the human liver microsome studies, with the exception of peak 11', which is shown in Fig. 1c. No significant metabolism was noted in any of the other recombinant P450 reactions.
View larger version (10K):
[in this window]
[in a new window]
|
FIG. 3. Effect of ketoconazole on the formation of temsirolimus metabolites. The data shown are the average of duplicates.
|
|
Further comparison of the metabolic products produced by human liver microsomes and by recombinant CYP3A4 was also performed by LC-MS. After the metabolites were separated and collected from semipreparative HPLC, each individual metabolite peak was analyzed by LC-MS. The semipreparative HPLC chromatograms of the metabolic products from human liver microsomes and from incubation with recombinant CYP3A4 are shown in Fig. 2, and the LC-MS results are summarized in Table 1. The major metabolites (M8-M13) produced in human liver microsomes and in recombinant CYP3A4 incubations displayed closely similar retention times and the same molecular ions.
View this table:
[in this window]
[in a new window]
|
TABLE 1 LC-MS comparison of temsirolimus metabolic products from human liver microsomes and from recombinant CYP3A4
|
|
Inhibition of Temsirolimus Metabolism. Ketoconazole, an antifungal drug, is a potent inhibitor of CYP3A4 (Maurice et al., 1992). The addition of ketoconazole to the reaction inhibited the biotransformation of temsirolimus by human liver microsomes. The effect of ketoconazole on the formation of the major metabolites is shown in Fig. 3. Ketoconazole inhibited the formation of metabolites M8 through M13 with an IC50 < 2 µM. The addition of 5 µM ketoconazole decreased metabolite formation to 10 to 20% of control levels. This result confirmed the fact that CYP3A4 was the main enzyme responsible for the biotransformation of temsirolimus.
Isolation and Structure Elucidation of Temsirolimus Metabolites. To obtain enough material for structural elucidation and antitumor activity studies, a 4-liter reaction was performed by incubating 400 mg of temsirolimus with 1200 nmol of CYP3A4 and the NADPH-regenerating system. After solvent extraction, the crude extract was subjected to silica gel fractionation to eliminate the unreacted starting compound and the impurities from CYP3A4 membranes. The crude metabolite mixture was then separated by semipreparative HPLC to yield purified individual metabolites for structure elucidation.
View larger version (28K):
[in this window]
[in a new window]
|
FIG. 5. a, proposed structure of M8 and the fragmentation assignments (observed fragment ions are listed in Table 2). b, an expanded 1H NMR spectrum of M8 compared with that of temsirolimus.
|
|
The minor components appearing before seco-temsirolimus in the preparative HPLC chromatogram (peaks 1-6) (Fig. 2) were either hydroxylated or desmethylated temsirolimus derivatives, as suggested by their molecular ions listed in Table 1. For example, the peak with [M-H]- at m/z 1060 was a dihydroxylated temsirolimus derivative, the peak with [M-H]- at m/z 1044 was a monohydroxylated temsirolimus derivative, and the peak with [M-H]- at m/z 1046 was either a desmethyl-dihydroxyl temsirolimus or a temsirolimus ester bond hydrolysis product. As for seco-temsirolimus, these components easily fragmented in the positive MS/MS/CID mode (i.e., fragmentation at lower collision energy, 
30 V), suggesting that they were macrolide ring-opened derivatives (open at the O24/C25 bond). Peak 7 from both a human liver microsome incubation and a recombinant CYP3A4 incubation exhibited [M-H]- at m/z 1030 and was identified as a C27 ketone reduced seco-temsirolimus (C=O converted to CH-OH) by its MS/MS and MS/MS/MS spectra. Considering that these compounds were unstable, their abundances were very low, and no biological activity was due to the ring opening, further structural study was not performed on them. Structure elucidation by MS/MS, MS/MS/MS, and NMR focused on the major metabolites: M8, M9, M10, M11, M12, and M13.
MS/MS Fragmentation of Temsirolimus. Temsirolimus exhibited a molecular ion at m/z 1028 in the negative ESI MS spectrum, which gave two fragment ions at m/z 590 and m/z 437 in the MS/MS/CID experiment. These two ions resulted from cleavage of the C31/C32 bond and the O24/C25 ester bond, representing the "southern" portion and the "northern" portion of the molecule, respectively (Fig. 4). These characteristic cleavages were also observed in rapamycin and its analogs. Via loss of neutral species, such as H2O, MeOH, and/or CO2, the northern portion generated the fragment ions of m/z 407, m/z 389, and m/z 371, and the southern portion produced the fragment ions of m/z 546, m/z 528, m/z 514, and m/z 496, which are designated as "group 1" ions in Table 2. Further cleavage of the southern part at C8/C9 and C35/C36 (allylic cleavage) followed by loss of methanol gave rise to the fragment ions at m/z 261, m/z 229, and m/z 147. ß-Cleavage of C33 ketone in the southern part produced the fragment ion of m/z 101. All these fragment ions from the right region of the southern part are classified as "group 2" ions in Table 2. The ion at m/z 252 was formed by C8/C9 cleavage followed by CO2 loss (cleavage at C22/C23) in the left region of the southern portion. The ion at m/z 234 resulted from a further loss of H2O. The additional fragment ions at m/z 168 and m/z 128 arose from cleavages occurring at C13/C15 and C23/O24 and cleavage of the amide bond (C16/N17) in the southern part, respectively (Fig. 4). These ions (m/z 252, m/z 234, m/z 168, and m/z 128) from the left region of the southern part are assigned as "group 3" ions in Table 2. All of the fragmentation assignments described are shown in Fig. 4, which were achieved by accurate mass measurements, MS/MS/MS analysis of the fragment ions at m/z 590 and m/z 261, and comparison with the MS/MS data of rapamycin. These fragment ions were used as the diagnostic criteria for structural elucidation of the metabolites by the mass shift technique. For example, the presence of the m/z 437 fragment ion in a metabolite indicated that the northern portion of the molecule was intact. The disappearance of the group 2 fragment ions at m/z 261, m/z 229, and m/z 147 indicated that biotransformation had occurred in the C1 to C8 region and/or the C36 position of the southern portion. Using this technique, the locations of biotransformation were determined. Table 2 lists the fragment ions observed in the MS/MS/CID spectra of temsirolimus and its metabolites.
Identification of M8. M8 showed a deprotonated molecular ion at m/z 1044 in the ESI MS spectrum, 16 Da higher than that of temsirolimus, indicating a monohydroxylation (or oxidation) product of temsirolimus. The hydroxylation occurred at the southern part of the molecule as suggested by the intact northern fragment ions at m/z 437 and the modified southern fragment ion at m/z 606 in MS/MS spectrum (Table 2). Compared with the parent compound, which showed the group 2 ions at m/z 261, m/z 229, m/z 147, and m/z 101, M8 gave the fragment ions at m/z 277, m/z 245, m/z 163, and m/z 101 (Table 2), suggesting that hydroxylation had occurred on the right side of the southern part, of either C36 or the C1-C8 moiety (Fig. 5a). Because M8 displayed the same UV absorption as temsirolimus, the triene group (C1-C6) should be unchanged, thus limiting the possible biotransformation positions to C7,C8,C36, and two methyl groups. M8 was finally identified as 36-hydroxyl temsirolimus by LC-NMR. In the 1H NMR spectrum of M8 (Fig. 5b), the disappearance of H36 resonance and the significant downfield shifts of H1 (
5.66 ppm) compared with temsirolimus ( 5.46 ppm) suggested a 36-hydroxylation structure for M8. Because hydroxylation would eliminate H36, the C36-methyl proton of M8 would present as the observed singlet at 1.22 ppm instead of a doublet at 0.89 ppm in 1H NMR of temsirolimus, corroborating the proposed structure of M8.
Identification of M9. M9 also showed a deprotonated [M-H]- at m/z 1044 and the fragment ions at m/z 437 and m/z 606 in the negative ESI mass spectrum, suggesting a southern hydroxylation derivative of temsirolimus. In MS/MS/CID experiment, M9 displayed the same group 3 fragment ions (m/z 252, m/z 234, m/z 168, and m/z 128) (Table 2) as the parent compound, indicating that the southern left portion (C9-C23) had been left intact. The appearance of the group 2 ions at m/z 277, m/z 245, m/z 227, and m/z 147 (Table 2) suggested that biotransformation had occurred in the C32 to C35 region (Fig. 6). Like temsirolimus and M8, M9 also showed a fragment ion at m/z 101 due to the cleavage of C34/C35 in the southern part, suggesting that the C32 to C34 region was intact and hydroxylation had occurred at C35. The observation of unique ions at m/z 504, m/z 472, and m/z 175 supported this suggestion (Table 2). On the basis of the accurate mass measurement and elemental composition analysis (Table 3), these ions were assigned and are shown in Fig. 6. The m/z 504 ion resulted from the loss of the C32-C34 moiety of the southern part. Further elimination of MeOH or subsequent cleavage at C8/C9 followed by MeOH loss resulted in the fragment ions at m/z 472 and m/z 175 (Fig. 6), respectively.
Identification of M10 and M11. M10 was also a southern oxygenated derivative of temsirolimus, as suggested by the appearance of the same northern ions as for temsirolimus, but different southern ions in the MS/MS spectrum (Table 2). The same group-2 fragment ions (m/z 261, m/z 229, and m/z 147) seen for temsirolimus were also found in the M10 spectrum, which excluded the possibility of biotransformation at the southern right region from C1 to C8 and C36 to C32 (Fig. 7). The observation of the group 3 ions at m/z 168 (due to the cleavage of C13/C15 and C23/O24) and m/z 128 (due to the cleavage of the amide bond in the southern part) indicated that the upper left region from C15 to O24 of the southern part was also intact. The appearance of unique ions at m/z 196, m/z 240, m/z 289, and m/z 321 (Table 2) indicated that the hemiketal ring opened and the most likely hydroxylation position was C11. The hydroxylation of C11 led to C11/C12 cleavage in the southern portion to form the highest intensity fragment ion at m/z 240 (Fig. 7). Further loss of CO2 gave rise to the fragment at m/z 196. C11 hydroxylation also facilitated the cleavage of the C9/C10 bond, giving the fragment ion at m/z 321, which went on to lose MeOH for a strong peak at m/z 289 (Fig. 7). The same fragmentation pattern had also been observed in the MS/MS/CID spectrum of 11-hydroxyl rapamycin from a previous study we had conducted. For M10, the hemiketal ring opening increased the flexibility of the macrolide. The present of more conformation isomer and tautomer forms in solution led its 1H NMR spectrum to become too complicated to be interpreted. To confirm the structure, accurate mass measurements and elemental composition analysis of the diagnostic fragments (m/z 321, m/z 289, m/z 240, and m/z 196) were performed (Table 4), which provided the support for this structure assignment.
M11 not only showed the same molecular ion and fragmentation ions as M10, but also displayed ion intensities similar to those of M10 in the MS and MS/MS/CID spectra. The presence of the opened hemiketal ring ions (m/z 240 and m/z 196) suggested that M11 has the same structure as M10, and was, most likely, an epimer of M10 with a different stereochemistry at C11 (Fig. 7).
Identification of M12. M12 was also an oxidative metabolite in the southern region, as indicated by its molecular ion signal at m/z 1044 and the southern fragment ions at m/z 606, m/z 588, m/z 562, m/z 544, m/z 530, and m/z 512 (Table 2; Fig. 8a). The presence of the same group 2 ions (m/z 261, m/z 229, m/z 147, and m/z 101) (Table 2) as for temsirolimus in the MS/MS spectra suggested that the right portion of the molecule from C8 to C1 and C36 to C32 was intact (Fig. 8a). Detection of the group 3 fragment ions at m/z 268 and m/z 250 in the MS/MS spectrum of M12 versus m/z 252 and m/z 234 for temsirolimus suggested that biotransformation had occurred in the left side of the southern part, namely the C9 to O24 region (Fig. 8a). Two unique ions at m/z 184 and m/z 144 with elemental compositions of C8H10NO4 and C6H10NO3 (based on accurate mass measurements), respectively, pointed to oxidation at the pipecolinyl ring (Fig. 8a). In the MS/MS/MS spectra, these two ions generated the fragment ions at m/z 168 and m/z 128, respectively, by eliminating oxygen, but not m/z 166 and m/z 126 from losing water, indicating that M12 was an N-oxide derivative of temsirolimus (Fig. 8a). This result was supported by the observation of the downfiled shifts of H22 ( 5.25 ppm) and H18 ( 3.65 ppm) compared with those of the parent compound (H22: 5.12 ppm, overlapping with H25) observed in the 1H NMR spectra (Fig. 8b).
View larger version (26K):
[in this window]
[in a new window]
|
FIG. 8. a, proposed structure of M12 and the fragmentation assignments (observed fragment ions are listed in Table 2). b, an expanded 1H NMR spectrum of M12 compared with that of temsirolimus.
|
|
Identification of M13. In the negative ESI mass spectrum, M13 showed a molecular ion at m/z 1014 and fragmentation ions at m/z 437 and m/z 576, corresponding to a southern desmethylation derivative of temsirolimus (Table 2). The MS/MS spectrum of the molecular ion gave the same group 3 ions, but different group 2 ions compared with temsirolimus (Table 2), suggesting that the C9 to O24 region remained unchanged, and biotransformation had occurred in the lower right region of the southern part. The most likely desmethylation candidates were C7 and C32 (Fig. 9a). The presence of the fragment ion at m/z 147, corresponding to the cleavage of C8/C9 and C35/C36 in the southern part, suggested that the C7-O-methyl group was still intact and that C32-O-desmethylation had occurred. This suggestion was supported by the disappearance of the fragment ion at m/z 101, and the appearance of the ion peak at m/z 87 from C34/C35 cleavage of the southern part. Therefore, the structure of M13 was proposed as C32-O-desmethyl temsirolimus (Fig. 9a). In the 1H NMR spectrum of M13, the disappearance of C32-methoxyl proton signals at 3.24 and the downfield shifting of H31 compared with temsirolimus further confirmed this assigned structure (Fig. 9b).
View larger version (26K):
[in this window]
[in a new window]
|
FIG. 9. a, proposed structure of M13 and the fragmentation assignments (observed fragment ions are listed in Table 2). b, an expanded 1H NMR spectrum of M12 compared with that of temsirolimus.
|
|
Biological Activity of the Major Metabolites. The biological activities of rapamycin and temsirolimus are dependent on the binding of the left-hand portion (C8-C31, the "binding domain") of the molecule to FKBP12 to form a complex that in turn binds the mTOR protein through the remaining portion of the molecule (the "effector domain" (Odagaki et al., 1997; Sedrani et al., 1999). The inhibition of the mTOR pathway by FKBP12-rapamycin (or temsirolimus) blocks multiple downstream signals and leads to a general antiproliferative effect. Therefore, any structural modification of the macrocyclic lactone ring (such as ring opening or hydroxylation of these two domains) could potentially affect the antineoplastic activities of the molecule. According to our proposed structures, the biotransformation locations of these metabolites were either in the binding domain (M10, M11, and M12), or in the effector domain (M8, M9, and M13); therefore, antitumor activities different from those of temsirolimus were expected. The antitumor activity of the metabolites was evaluated in a cellular assay of LNCaP (a prostate carcinoma cell line) proliferation. Compared with temsirolimus, which showed an IC50 of 2 nM, these metabolites showed remarkably decreased activity (IC50 >100 nM) or absolutely no inhibitory activity (IC50 >1000 nM), as seen in Table 5.
|
Discussion
|
|---|
Temsirolimus, an ester derivative of rapamycin, is a selective inhibitor of mTOR that showed clear clinical efficacy and excellent tolerability in the treatment of renal cancer. The in vitro metabolism studies conducted with temsirolimus and human liver microsomes led to the formation of 15 metabolites as detected by LC-MS. As with rapamycin, CYP3A4 was shown to be the major enzyme responsible for biotransformation through an inhibition study and by incubation with individual recombinant P450 enzymes. Such knowledge is of considerable clinical utility with regard to potential drug-drug interactions, as well as interindividual differences in drug-metabolizing capacities stemming from genetic polymorphisms, because CYP3A4 is involved in the biotransformation of approximately 60% of all the xenobiotics on the market today (Waxman et al., 1988; Guengerich, 1989; Thummel and Wilkinson, 1998), including cyclosporine, erythromycin, diazepam, nifedipine, estradiol, paclitaxel, and lovastatin. Coadministration of drugs that are substrates (such as erythromycin, clarithromycin, and lovastatin), inhibitors (such as troleandomycin and human immunodeficiency virus protease inhibitors), or inducers (such as rifampin and carbamazepine) of CYP3A4 may affect the metabolic disposition of temsirolimus in humans. An earlier study on the effects of temsirolimus on patients with recurrent glioblastoma multiforme (Galanis et al., 2005), who were at the same time receiving P450-inducing anticonvulsants (known to increase CYP3A4), showed that the peak concentrations (Cmax) of temsirolimus and rapamycin decreased by approximately 73 and 47%, respectively, compared with those in patients with renal cancer who were not receiving P450-inducing anticonvulsants.
The 15 detectable metabolites were the monohydroxylation, desmethylation, N-oxidation, and dihydoxylation/ring-opening products of temsirolimus. Six major metabolites were identified to be 36-hydroxyl temsirolimus (M8), 35-hydroxyl temsirolimus (M9), 11-hydroxyl temsirolimus with an opened hemiketal ring (M10 and M11), N-oxide temsirolimus (M12), and 32-O-desmethyl temsirolimus (M13) by combined LC-MS, MS/MS, MS/MS/MS, and NMR spectroscopic methods. In humans, temsirolimus converts by hydrolysis to rapamycin, and both temsirolimus and rapamycin are subject to oxidative metabolism (Galanis et al., 2005). However, no rapamycin metabolites were observed in our study, most likely because the formation of rapamycin from temsirolimus in human liver microsomes was not extensive. Temsirolimus metabolites formed primarily via C-oxidation (aliphatic hydroxylation), O-desmethylation, N-oxidation, and ring opening, which was similar to the metabolites found in pooled bile of rats after they were administered rapamycin intravenously (Wang et al., 1997). In addition to seco-rapamycin, nine hydroxylated rapamycin and/or desmethylated rapamycin metabolites were detected in the same pooled rat bile. Although the structures of these metabolites were not identified, the hydroxylation and desmethylation sites of three major biliary metabolites (m2, m10, and m13) were proposed at the regions of C32 to C36 and C1 to C11 on the basis of LC-MS data, resembling temsirolimus metabolites M8, M9, M10, and M11. Two southern portion monohydroxylated rapamycin products were also found as the most abundant metabolites (
73.6% of total metabolites) from the blood of kidney transplantation patients receiving rapamycin (Holt et al., 2003). Whether the temsirolimus metabolites identified in our study share the same biotransformation locations with those in vivo rapamycin metabolites need to be further evaluated.
Although the 41-O-desmethylation product was the predominant metabolite of rapamycin in the incubation of rapamycin with human liver microsomes, this was not the case for temsirolimus. Such metabolic differences may be a consequence of C42-O-esterification of temsirolimus. Introduction of the 2,2-bis-(hydroxymethyl)-propionate side chain at the C42 position resulted in significant steric hindrance, thus blocking CYP3A4 enzyme action on the 41-O-methyl group. It is worth mentioning that the major in vitro rapamycin metabolite, 41-O-desmethyl rapamycin, was not found in rat bile and blood of kidney transplantation patients in significant amounts according to the studies of Wang et al. (1997) and Holt et al. (2003). This result is probably due to the fact that 41-O-desmethyl rapamycin underwent further oxidation and/or ring opening and degraded to various products at very low amounts. It is also possible that rapamycin (or temsirolimus) alters its structure conformation after binding to target proteins in vivo, which changes CYP3A4 action and results in a different biotransformation profile than that found in vitro.
Incubation of temsirolimus with recombinant human CYP3A4 in the presence of a NADPH-regenerating system reproduced almost all of the metabolites detected in human liver microsomes, so this technique was subsequently used to prepare the metabolites on a large scale for isolation and analysis of the metabolites. Six major metabolites (M8-M13) were isolated and purified, enabling unambiguous structural elucidation and bioactivity studies and also providing useful reference standards for future in vivo metabolic studies.
As temsirolimus is a relatively complicated molecule, the structural elucidation of the metabolites was challenging. Combined LC-MS, MS/MS, MS/MS/MS, and accurate mass measurements as well as NMR techniques provided powerful and reliable tools for metabolite structural elucidation. Temsirolimus usually is present in solution as two interchangeable isomer forms, and the most proton signals overlapped in the NMR spectrum. For metabolites, hydroxylation and/or ring opening increases the complexity of the proton NMR spectrum. The characteristic fragmentation pattern of temsirolimus in MS/MS and MS/MS/MS and accurate mass measurements of the fragments can provide more structural information for those metabolites (such as M9, M10, and M11) with heavily overlapping and uninterpretable NMR data.
|
Acknowledgments
|
|---|
We are grateful to Dr. Melissa Lin and Yumin Gong for acquiring NMR data of temsirolimus metabolites.
|
Footnotes
|
|---|
doi:10.1124/dmd.107.014746.
ABBREVIATIONS: temsirolimus, rapamycin-42-[2,2-bis-(hydroxymethyl)]-propionate; HLM, human liver microsomes; mTOR, mammalian target of rapamycin; P450, cytochrome P450; HPLC, high-performance liquid chromatography; LC, liquid chromatography; MS, mass spectrometry; MS/MS or MS/MS/MS, tandem mass spectrometry; NMR, nuclear magnetic resonance; ESI, electrospray ionization; CID, collision induced dissociation.
Address correspondence to: Dr. Ping Cai, Wyeth Research, Building 250/229, 401 N. Middletown Rd., Pearl River, NY 10965. E-mail: caip{at}wyeth.com
|
References
|
|---|
Christians U, Sattler M, Schiebel HM, Kruse C, Radeke HH, Linck A, and Sewing K-FR (1992) Isolation of two immunosuppressive metabolites after in vitro metabolism of rapamycin. Drug Metab Dispos 20: 186-191.[Abstract]
Galanis E, Buckner JC, Maurer MJ, Kreisberg I, Ballman K, Boni J, Peralba JM, Jenkins RB, Dakhil SR, Morton RF, et al. (2005) Phase II trial of temsirolimus (CCI-779) in recurrent glioblastoma multiforme: a North Central Cancer Treatment Group study. J Clin Oncol 23: 5294-5304.[Abstract/Free Full Text]
Guengerich FP (1989) Characterization of human microsomal cytochrome P450 enzymes. Annu Rev Pharmacol Toxicol 29: 241-264.[CrossRef][Medline]
Holt DW, Lee TD, Mckeown DA, Hicks D, Cai P, and Johnston A (2003) Relative proportions of sirolimus (rapamycin) metabolites in blood following kidney transplantation, in Proceedings of the 8th International Congress of Therapeutic Drug Monitoring and Clinical Toxicology; 2003 Sept 7-11; Basel, Switzerland. International Association of Therapeutic Drug Monitoring and Clinical Toxicology, Kingston, ON, Canada.
Maurice M, Pichard L, Daujat M, Fabre I, Joyeux H, Domergue J, and Maurel P (1992) Effects of imidazole derivative on cytochromes P450 from human hepatocytes in primary culture. FASEB J 6: 752-755.[Abstract]
Odagaki Y and Clardy J (1997) Structural basis for peptidomimicry by the effector element of rapamycin. J Am Chem Soc 119: 10253-10254.[CrossRef]
Peralba JM, deGraffenried L, Friedrichs W, Fulcher L, Grunwald V, Weiss G, and Hidalgo M (2003) Pharmacodynamic evaluation of temsirolimus, an inhibitor of mTOR, in cancer patients. Clin Cancer Res 9: 2887-2892.[Abstract/Free Full Text]
Sattler M, Guengerich FP, Yun C, Christians U, and Sewing K-FR (1992) Cytochrome P450 and rat. Drug Metab Dispos 20: 753-761.[Abstract]
Sedrani R, Jones LH, Jutzi-Eme AM, Schuler W, and Cottens S (1999) Cleavage of the cyclohexyl-subunit of rapamycin results in loss of immunosuppressive activity. Bioorg Med Chem Lett 9: 459-462.[CrossRef][Medline]
Sehgal SN (1995) Rapamune (sirolimus, rapamycin): an overview and mechanism of action. Ther Drug Monit 17: 660-665.[Medline]
Sehgal SN, Molnar-Kimber K, Ocain TD, and Weichman BM (1994) Rapamycin: a novel immunosuppressive macrolide. Med Res Rev 14: 1-22.[CrossRef][Medline]
Thummel KE and Wilkinson GR (1998) In vitro and in vivo drug interactions involving human CYP3A. Annu Rev Pharmacol Toxicol 38: 389-430.[CrossRef][Medline]
Wang CP, Chan KW, Schiksnis RA, Scatina JA, and Sisenwine SF (1997) High performance liquid chromatographic isolation and spectroscopic characterization and immunosuppressive activities of two rapamycin degradation products. J Liq Chromatogr 17: 3383-3392.[CrossRef]
Waxman DJ, Attisano C, Guengerich FP, and Lapenson DP (1988) Human liver microsomal steroid hormone 6ß-hydroxylase cytochrome P450 enzyme. Arch Biochem Biophys 263: 424-436.[CrossRef][Medline]
This article has been cited by other articles:
|
|
|
|
 
J. G. Kuhn, S. M. Chang, P. Y. Wen, T. F. Cloughesy, H. Greenberg, D. Schiff, C. Conrad, K. L. Fink, H. I. Robins, M. Mehta, et al.
Pharmacokinetic and Tumor Distribution Characteristics of Temsirolimus in Patients with Recurrent Malignant Glioma
Clin. Cancer Res.,
December 15, 2007;
13(24):
7401 - 7406.
[Abstract]
[Full Text]
[PDF]
|
|
|
![[Feedback]](/icons/banner/feedbackACT.gif)
![[For Subscribers]](/icons/banner/subscriberACT.gif)
![[Archive]](/icons/banner/archiveACT.gif)
![[Search]](/icons/banner/searchACT.gif)
![[Contents]](/icons/banner/tocACT.gif)
Top
Abstract
Materials and Methods
