Materials

JQ1, [tert-butyl (S)-2-(4-(4-chlorophenyl)-2,3,9-trimethyl-6Hthieno[3,2-f][1,2,4]triazolo[4,3-a][1,4]diazepin-6-yl)acetate], ketoconazole (KCZ), NADPH (β-nicotinamide adenine dinucleotide 2′-phosphate reduced tetrasodium salt hydrate) and formic acid were obtained from Sigma-Aldrich (St. Louis, MO). Nootkatone was purchased form Cayman Chemical Company (Ann Arbor, MI). Human and mouse liver microsomes and recombinant human cytochrome P450s (EasyCYP Bactosomes) were purchased from XenoTech (Lenexa, KS). All solvents for liquid chromatography and mass spectrometry were of the highest grade commercially available.

Metabolism of JQ1 in human and mouse liver microsomes and recombinant cytochrome P450s

Incubations were conducted in 1× phosphate-buffered saline (1 × PBS, pH 7.4) containing 20 µM JQ1 and 0.2 mg human or mouse liver microsomes or 2 pmol of each cDNA-expressed P450 enzyme (control, CYP1A2, 2A6, 2B6, 2C8, 2C9, 2C19, 2D6, 2E1 and 3A4) in a final volume of 190 µl. After a 5 min pre-incubation at 37 °C, 10 µl of 20 mM NADPH was added (final concentration 1.0 mM), and incubation was continued for 30 min with gentle shaking. Incubations without NADPH were used as controls. Coincubations of JQ1 (20 µM) and KCZ (CYP3A4 inhibitor, 2.0 µM) or nootkatone (CYP2C19 inhibitor, 10 µM) in human and mouse liver microsomes were performed to determine the role of CYP3A4 and CYP2C19 in the formation of JQ1 metabolites. Reactions were terminated by adding 200 µl of ice-cold methanol and vortexing for 30 s and then centrifuged at rcf 15 000 for 15 min. Each supernatant was transferred to an autosampler vial, and 3.0 µl was injected onto UHPLC-Q Exactive MS system for analysis. Incubations were performed in duplicate for cDNA-expressed enzymes and in triplicate for liver microsome experiments.

Metabolism of JQ1 in S9 fraction from Cyp3a-null and wild-type mice

Liver tissue (150 mg) (Cyp3a-null or WT) was homogenized in 600 ml of 1× PBS at 0 °C. The homogenate was centrifuged at 9000 × g for 20 min at 4 °C, and the resulting supernatant was designated the enzyme fraction (S9 fraction). JQ1 (25 µM) or MDZ (30 µM) was incubated in the 1× PBS in the presence of 60 ml of S9 fraction freshly prepared, a final volume of 190 ml. After 5 min of pre-incubation at 37 °C, the reaction was initiated by adding 10 ml of 20 mM NADPH (final concentration 1.0 mM) and continued for 30 min for JQ1 and 15 min for MDZ with gentle shaking. Reactions were stopped by adding 200 ml of ice-cold methanol and vortexing for 30 s. The reaction mixtures were centrifuged at 15 000 × g for 20 min. Each supernatant was transferred to an autosampler vial for metabolite analysis. Incubations were performed in triplicate.

UHPLC-Q Exactive Orbitrap MS analyses

JQ1 and its metabolites were resolved, identified and quantified (relatively) using UHPLC coupled with Q Exactive Orbitrap MS (Thermo Fisher Scientific, San Jose, CA) equipped with 100 mm × 2.1 mm column (BEH C-18, Agilent Technologies, Santa Clara, CA). The column temperature was maintained at 40 °C. The flow rate was 0.3 ml/min with a gradient ranging from 2 to 95% aqueous acetonitrile containing 0.1% formic acid in a 15 min run. Q Exactive MS was operated in positive mode with electrospray ionization. Ultrapure nitrogen was applied as the sheath (45 arbitrary unit), auxiliary (10 arbitrary unit), sweep (1.0 arbitrary unit) and collision gas. The capillary gas temperature was set at 275 °C and the capillary voltage was set at 3.7 kV. MS data were acquired from 80 to 1200 Da in profile mode; reference ions at m/z 371.1012 and 445.1200 in the positive mode were used as lock masses during acquisition. MS/MS of JQ1 metabolites was performed in targeted mode with an isolation width of 2 m/z with ramp collision energy being set at 15, 20 and 35 eV.

Data analysis

Mass chromatograms and mass spectra were acquired and processed using the Xcalibur software (Thermo Fisher Scientific, San Jose, CA) in profile formats from m/z 50 to 750. The acquired data were processed by Compound Discoverer 3.0 software (Thermo Fisher Scientific, San Jose, CA) to generate a multivariate data matrix. Data matrices were exported into SIMCA14 (Umetrics, Kinnelon, NJ) for multivariate data analysis [28]. Orthogonal projection to latent structure discriminant analysis (OPLS-DA) was conducted on Pareto-scaled data [29]. For chemometric analysis, matrix data were processed from m/z 100 to 750. Statistical analysis was conducted using Student's independent t-test. Experimental data are presented as mean ± SEM.

Profiling JQ1 metabolism in human liver microsomes using metabolomic approaches

The results of metabolomic analysis on the ions produced by UHPLC-QTOFMS of control and JQ1 group are shown in Figure 1. Metabolomic analysis revealed two clusters (Figure 1A) corresponding to the control and JQ1 group in the score plots, which indicates that the chemical components are different between control and JQ1 groups. The human liver microsome data S-plot (Figure 1B) generated from OPLS-DA displays ions that contribute to group separation. Many of the top-ranking ions were identified as JQ1 and its metabolites, as marked in the S-plots (Figure 1B). All nine distinguishable JQ1 metabolites were present in both human and mouse liver microsomes (Figure 1C and Table 1), including three monohydroxylated JQ1 (M1–M3), two acids (M4 and M8), two dihydroxylated JQ1 (M5 and M6), one monohydroxylated/dehydrogenated JQ1 (M7) and one dihydroxylated/dehydrogenated JQ1 (M9). The detailed information associated with JQ1 and its metabolites are summarized in Table 1. The relative abundance of ions for these metabolites in human and mouse liver microsomes is presented in Figure 1C. The putative structures of JQ1 metabolites were identified based on the exact mass, MS/MS fragments, and predicted formula, as the standards of these metabolites are not available.

Identification of monohydroxylated-JQ1 metabolites (M1–M3)

Three monohydroxylated JQ1 metabolites (M1–M3) whose formation is NADPH-dependent (Figure 3B) were identified and characterized in human and mouse liver microsomes. M1 is the dominant metabolite in both human and mouse liver microsomes. Metabolite M1 eluted at 10.16 min (Figure 2A) and was detected as a protonated molecule at m/z 473.1410, 16 Da higher than parent JQ1 (Figure 2C and D). The MS/MS of M1 produces major fragmental ions at m/z 455 (loss of H₂O), 417 (loss of tert-butyl), 399 (loss of tert-butyl and H₂O), 357 and 290. Comparison of the fragmental ions at m/z 357 and 290 with the fragmental ions of JQ1 (Figure 2C) indicates that the oxidation occurs on the fused three-ring moiety (boxed in Figure 2D). The M1 fragmental ions are interpreted in the inlaid structural diagram (Figure 2D). Metabolite M2 eluted at 10.37 min (Figure 2A) and was detected as a protonated molecule at m/z 284.1272 (Figure 2E). MS/MS of M2 produced major fragmental ions at m/z 455, 417, 399, 357 and 290, matching the pattern of M1 and indicating that the oxidation occurs on the same fused three-ring moiety as in M1 (Figure 2E), though the different elution positions of M1 and M2 mean different oxidation positions within that moiety. The M2 fragmental ions are interpreted in the inlaid structural diagram (Figure 2E). Metabolite M3 eluted at 10.38 min (Figure 2A) and was detected as a protonated molecule at m/z 473.1410 (Figure 2F). MS/MS of M3 produces major fragmental ions at m/z 455, 417, 399, 357 and 274 (Figure 2F). The fragmental ions at m/z 357 and 274 compared with the fragmental ions of JQ1 indicate that the oxidation in this metabolite occurs on the 4-chlorophenyl moiety (Figure 2C), as interpreted in the inlaid structural diagram (Figure 2F). The exact oxidation sites of metabolites were not defined because standards are not available for comparison by LC-MS.

Identification of JQ1 metabolites (M4–M9)

JQ1 metabolites (M4–M9) were detected in both human and mouse liver microsomes. The formation of M5–M9 is completely NADPH-dependent. However, low amounts of M4 were detected without addition of NADPH to human and mouse liver microsomes (1.7 and 3.5%, respectively, of the M4 levels seen with addition of NADPH; data not shown). The formation M4 in the control group is not likely due to slow ester hydrolysis in the original JQ1 stock, because did not observe the M4 in the JQ1 stock solution. The small amount of M4 formation may be mediated by esterases in liver microsomes. M4 eluted at 9.32 min, and was detected as a protonated molecule at m/z 401.0833 (Table 1). MS/MS of M4 produced major fragmental ions at m/z 383, 341, 289, and 274 (Figure 3A), which are interpreted in the inlaid structural diagram. These data identify M4 to be de-tert-butylated JQ1. Two doubly hydroxylated-JQ1 metabolites (M5 and M6) were identified in both human and mouse liver microsomes, though much more M5 is made by human liver microsomes than by mouse liver microsomes (Figure 1C). M5 eluted at 8.28 min, and was detected as a protonated molecule at m/z 489.1359 (Table 1), 32 Da higher than that of JQ1. MS/MS of M5 produced major fragmental ions at m/z 489, 471, 399, 357 and 290 (Figure 3B). The fragmental ions at m/z 399 and 290 (Figure 3B) indicate that one oxidation occurred on the tert-butyl group. The ions at m/z 357 and 290 suggested that the second oxidation occurred on the fused three-ring moiety. Fragmental ions from M5 are interpreted in the inlaid structural diagram in Figure 3B. M6 eluted at 9.74 min and was detected as a protonated molecule at m/z 489.1359 (Table 1). MS/MS of M6 produced major fragmental ions at m/z 489, 433, 415, 373 and 290 (Figure 3C). The fragmental ion at m/z 433 indicates that no oxidation occurred on the tert-butyl group. The ions at m/z 373 and 290 suggested that one oxidation occurred on the 4-chlorophenyl group and the other one occurred on the fused three-ring moiety. Fragmental ions from M6 are interpreted in the inlaid structural diagram in Figure 3C.

Figure 1.

Metabolomic analysis of control and JQ1 groups in human liver microsomes and relative abundance of metabolites of JQ1. The incubation conditions of JQ1 were detailed in experimental procedures. All the samples were analysed using UHPLC-Q Exactive MS. (A) Separation of control and JQ1 group in OPLS-DA score plot. The t[1] and to[1] values represent the score of each sample in principal components 1 and 2, respectively. (B) Loading S-plot generated by OPLS-DA analysis. The x-axis is a measure of the relative abundance of ions, and the y-axis is a measure of the correlation of each ion to the model. The top-ranking ions are labelled. The number of ions (metabolite identification) shown was accordant with Table 1. (C) Relative abundance of metabolites of JQ1in human and mouse liver microsomes. Incubation conditions were detailed in Methods. The relative quantification was conducted based on the peak area. The data are expressed as mean ± SEM (n = 3). ^*P < 0.05, ^**P < 0.01.

Table 1.

Summary of metabolites of JQ1 in HLM and MLM.

Figure 2.

Identification of monohydroxylated-JQ1 metabolites (M1, M2 and M3). The incubation conditions of JQ1 were as described in Figure 1C. All the samples were analysed using UHPLC-Q Exactive MS. Structural elucidations were performed based on accurate mass (mass errors less than 5 ppm) and MS/MS fragmentation. MS/MS was performed with collision energy ramping from 10 to 35 V. The major fragmental ions are interpreted in the insets. (A) Extracted ion chromatograms of JQ1 metabolites M1, M2 and M3. (B) Representative trend plot of JQ1 metabolite M1. (C) MS/MS of JQ1. (D) MS/MS of M1. (E). MS/MS of M2. (F). MS/MS of M3.

Figure 3.

Identifying JQ1 metabolites (M4–M9). The incubation conditions of JQ1 were as described in Figure 1C. All the samples were analysed using UHPLC-Q Exactive MS. The metabolite elucidation conditions are as described in Figure 3 legend. (A) MS/MS of M4. (B) MS/MS of M5. (C) MS/MS of M6. (D) MS/MS of M7. (E) MS/MS of M8. (F) MS/MS of M9.

M7 eluted at 11.07 min and was detected as a protonated molecule at m/z 471.1255 (Table 1). MS/MS of M7 produced major fragmental ions at m/z 415, 397, 355 and 288 (Figure 3D). Compared to the fragments of monohydroxylated JQ1 (M1), the ions at m/z 355 and 288 indicate that monohydroxylation and dehydrogenation occurred on the fused three-ring moiety of JQ1; fragmental ions are interpreted in the inlaid structural diagram (Figure 3D). M8 eluted at 7.48 min and was detected as a protonated molecule at m/z 417.0782 (Table 1). MS/MS of M8 produced major fragmental ions at m/z 399, 357 and 290 (Figure 3E). Compared to the fragments of de-tert-butyl JQ1 (M4), the ions at m/z 357 and 290 indicated that hydroxylation occurred on the 4-chlorophenyl moiety. Fragmental ions from M8 are interpreted in the inlaid structural diagram in Figure 4E. M9 was observed at a retention time of 10.32 min, having a protonated molecular ion at m/z 487.1202 (Table 1). The MS/MS of M9 produced major fragmental ions at m/z 431, 413, 371 and 304 (Figure 3F). Compared to the fragments of dihydroxylated JQ1 (M5), ions at m/z 371 and 304 indicate that dihydroxylation and dehydrogenation all occurred on the fused three-ring moiety of JQ1. Fragmental ions from M9 are interpreted in the inlaid structural diagram in Figure 3F.

Role of cytochrome P450s in JQ1 metabolism

The roles of individual cytochrome P450s in the metabolism of JQ1 were determined by incubating JQ1 with nine human cDNA-expressed P450s (control, CYP1A2, 2A6, 2B6, 2C8, 2C9, 2C19, 2D6, 2E1 and CYP3A4), by chemical inhibitory experiments in human and mouse liver microsomes and by incubation with liver S9 fractions of wild-type and Cyp3a-null mice. Incubating JQ1 with human cDNA-expressed P450s revealed that, of these enzymes, only CYP3A4 is able to generate metabolites M3–M6, M8 and M9 (Table 2) and both CYP3A4 and CYP2C19 can form M7. Six of the tested enzymes can form monohydroxylated metabolite M1, with strong activity demonstrated by CYP3A4, CYP2C8 and CYP2C19. None of the nine human cDNA-expressed P450s tested can generate the monohydroxylated metabolite M2.

Figure 4.

Roles of CYP3A4 in the formation of JQ1 metabolites in human and mouse liver microsomes. KCZ (2.0 µM, CYP3A4 inhibitor) was used in the inhibitory tests in human and mouse liver microsomes. The incubation conditions of JQ1 in human or mouse liver microsomes were detailed in experimental procedures. All samples were analysed by UHPLC-Q Exactive MS. (A) Effects of KCZ (2.0 µM) on the formations of M1–M9 in human liver microsomes. The relative abundance of JQ1 metabolites from the incubation with human liver microsomes without KCZ was set as 100%. (B) Effects of KCZ (2.0 µM) on the formations of the formations of M1–M9 in mouse liver microsomes. The relative abundance of JQ1 metabolite from the incubation in mouse liver microsomes in the absence of KCZ was set as 100%. All data are expressed as mean ± SEM (n = 3). ^*P < 0.05, ^**P < 0.01. HLM, human liver microsomes; MLM, mouse liver microsomes; KCZ, ketoconazole.

The dependence of metabolite formation on CYP3A4 was verified by co-incubation of JQ1 with human or mouse liver microsomes and with ketoconazole (KCZ), a potent CYP3A4 inhibitor. In human liver microsomes, the formation of metabolites M1–M9 was suppressed by 79, 91, 99, 97, 99, 100, 98, 100 and 99%, respectively, by KCZ at 2.0 µM (Figure 4A). In mouse liver microsomes, the formation of M1–M9 was suppressed by 85, 87, 93, 80, 99, 100, 98, 100 and 99%, respectively, by 2.0 µM KCZ (Figure 4B). Incubation of cDNA-expressed P450 CYP2C19 with JQ1 shows that this enzyme is capable of forming both M1 and M7 (Table 2); the CYP2C19 inhibitor nootkatone (10 µM) has no effect on M7 formation in human liver microsomes (not shown) but suppresses M1 formation by 9%. Quercetin (30 µM), a CYP2C8 inhibitor, inhibited M1 formation in human liver microsomes by 78% (data not shown).

Because CYP3A4 is by far the major enzyme shown to contribute to JQ1 phase I metabolism (Table 2) and the CYP3A4 inhibitor ketoconazole greatly diminishes the generation of all detectable metabolites in human and mouse liver microsomes, we anticipated that Cyp3a-null mice might show greatly decreased metabolism of JQ1. Before undertaking long and expensive in vivo pharmacokinetic analyses, we tested the relative abilities of S9 fractions from wild-type and Cyp3a-null mice to generate the detected JQ1 metabolites. Most JQ1 metabolites are produced at lower levels by Cyp3a-null S9 fraction than by wild type (Figure 5), but the Cyp3a-null S9 fraction makes 2.67-fold more M4 than wild type and generates major metabolite M1 at much higher levels (65% of wild type) than would be expected from the ketoconazole inhibition data (Figure 4).