Development of a chiral micellar electrokinetic chromatography-tandem mass spectrometry assay for simultaneous analysis of warfarin and hydroxywarfarin metabolites: application to the analysis of patients serum samples

Abstract

The enantioseparation of warfarin (WAR) along with the five positional and optical isomers is challenging because of the difficulty to simultaneously separate and quantitate these chiral compounds. Currently, no effective chiral CE-MS methods exist for the simultaneous enantioseparation of WAR and all its hydroxylated metabolites in a single run. Polymeric surfactants (aka. molecular micelles) are particularly compatible with micellar electrokinetic chromatography-mass spectrometry (MEKC-MS) because they have a wider elution window for enantioseparation and do not interfere with the MS detection of chiral drugs. Using polysodium N-undecenoyl-L,L-leucylvalinate (poly-L,L-SULV) as a chiral pseudophase in MEKC-MS baseline separation of WAR, its five metabolites along with the internal standard was obtained in 45 min. This is in comparison to 100 min required for separation of the same mixture with packed column CEC-MS using a vancomycin chiral stationary phase. Serum samples were extracted with mixed-mode anion-exchange (MAX) cartridge with recoveries of greater than 85.2% for all WAR and hydroxywarfarin (OH-WAR) metabolites. Utilizing the tandem MS and multiple reaction monitoring mode, the MEKC-MS/MS method was used to simultaneously generate calibration curves over a concentration range from 2 to 5000 ng/mL for R- and S-warfarin, 5 to 1000 ng/mL for R- and S-6-, 7-, 8- and 10-OH-WAR and 10 to 1000 ng/mL for R and S-4'-OH-WAR. For the first time, the limits of detection and quantitation for most WAR metabolites by MEKC-MS/MS were found to be at levels of 2 and 5 ng/mL, respectively. The method was successfully applied for the first time to analyze WAR and its metabolites in plasma samples of 55 patients undergoing WAR therapy, demonstrating the potential of chiral MEKC-MS/MS method to accurately quantitate with high sensitivity.

Figure 1

Structures of (A) WAR and its sites of hydroxylation (indicated by arrow) to generate monohydroxylated WAR metabolites by cytochrome P450s (B) diclofenac (DFC) sodium (used as internal standard). The asterisk indicates stereogenic center.

Figure 2

Effect of pH on simultaneous enantioseparation of WAR and its metabolites. pH 5.0 (A, C) and pH 5.5 (B, D). Conditions: 118 cm long (375 μm O.D., 50 μm I.D.) fused silica capillary. Buffer: 25 mM NH₄OAc/25 mM poly-L,L-SULV. Applied voltage, +30 kV, injection, 5 mbar, 2 s. Spray chamber parameters: nebulizer pressure: 4 psi, drying gas temp.: 200 °C, drying gas flow: 6 L/min; capillary voltage: −3000 V; fragmentor voltage, 91 V; SIM mode; sheath liquid: MeOH/H₂O (80/20, v/v), 5 mM NH₄OAc, pH 6.8 with flow rate of 0.5 mL/min; Sample concentration: 10 μg/mL in ACN/H₂0 (40/60, v/v).

Figure 3

Effect of type of organic solvent on simultaneous enantioseparation of WAR and its hydroxylated metabolites. MeOH (A), ACN (B), EtOH (C) and IPA (D). Conditions: buffer 25 mM NH₄OAc pH 5.0 /25 mM poly-L,L-SULV containing organic solvent. Other MEKC-ESIMS conditions are the same as in Fig. 2.

Figure 4

Optimized enantioseparation of WAR and its hydroxylated metabolites by MEKC-ESIMS (A–C). Conditions: buffer 25 mM NH₄OAc/25 mM poly-L,L-SULV containing 15% (v/v) MeOH at pH 5.0. Other MEKC-ESI-MS conditions are the same as in Fig. 2.

Figure 5

Optimized enantioseparation of WAR and its hydroxylated metabolites by CEC-ESIMS (A–D). Conditions: 40 cm long vancomycin packed column (total long 65 cm, 375 μm O.D. and 75 μm I.D.). Running buffer, ACN / H₂O (45/55, v/v), 10 mM NH₄OAc at pH 4.0 (A) and ACN / MeOH/ H₂O (30/50/20, v/v/v), 10 mM NH₄OAc at pH 4.0 (B). Applied voltage, + 25 kV, injection, 5 kV, 3 s. Spray chamber and sheath liquid conditions are the same as in Fig. 2.

Figure 6

Comparison of the detection of WAR and its hydroxylated metabolites by MEKC-MS (A, C) and MEKC-MS/MS (B, D). For MEKC-MS, all the conditions are the same as in Fig. 4A. MEKC-MS/MS conditions: 120 cm long (375 μm O.D., 50 μm I.D.) fused silica capillary. Buffer: 25 mM NH₄OAc pH 5.0, 25 mM poly-L,L-SULV with 15 % (v/v) MeOH, Injection: 5 mbar for 2 s, Voltage: 30 kV; Spray chamber parameters: drying gas temperature: 200 °C, drying gas flow rate: 8 L/min, nebulizer pressure: 4 psi, capillary voltage: −3000 V, collision energy: 20 eV for all except 5 ev for I.S., fragmentor voltage: 125 V for all except 75 V for I.S.. Sheath liquid: MeOH/H₂O (80/20, v/v) with 5mM NH₄OAc, sheath liquid flow rate: 5 μL/min.

Figure 7

The effect of injection time on S/N ratio (A) and chiral resolution (B) of WAR and OHWAR metabolites. All other conditions are the same as Fig. 6B. Analyte: 0.1 μg/mL of WAR, 0.2 μg/mL of each OH-WAR, and 10.0 μg/mL of I.S. in ACN/H₂O (40/60, v/v).

Figure 8

Extracted ion electrochromatograms of a standard mixture of WAR and OH-WAR under optimized injection conditions of 5 mbar for 300 s. All other conditions are the same as Fig. 6B. Analyte: 0.1 μg/mL of WAR, 0.2 μg/mL of each OH-WAR, and 10.0 μg/mL of I.S. in ACN/H₂O (40/60, v/v). For WAR and each OH-WAR, the R-enantiomer eluted first followed by the S-enantiomer.

Figure 9

Extracted ion electrochromatograms of subject 11 and subject 5 (with mutant gene CYP2C9*2 or *3). The MEKC-MS/MS conditions are the same as in Fig. 8.