Anal. Chem. 2008, 80, 3404-3411
Hydrophilic Interaction Liquid Chromatography-Tandem Mass Spectrometry Determination of Estrogen Conjugates in Human Urine Feng Qin,† Yuan-Yuan Zhao,† Michael B. Sawyer,‡ and Xing-Fang Li*,†
Division of Analytical and Environmental Toxicology, Department of Laboratory Medicine and Pathology, University of Alberta, 10-102 Clinical Sciences Building, Edmonton, AB, Canada T6G 2G3, and Department of Oncology, Cross Cancer Institute, 11560 University Avenue, Edmonton, AB, Canada T6G 1Z2
We report a hydrophilic interaction liquid chromatography (HILIC) separation with tandem mass spectrometry (MS) detection method for analysis of seven urinary estrogen conjugates. HILIC separation employing a mobile phase with high organic solvent content resulted in enhanced electrospray ionization efficiency and MS sensitivity compared with reversed-phase (RP) LC-MS methods. Solidphase extraction (SPE) was used to further improve the limit of detection and to eliminate interferences for the analysis of urine samples. No hydrolysis or derivatization was required in the sample pretreatment. This SPE/ HILIC-MS/MS method provided limits of quantification (LOQs at S/N ) 10) for the seven conjugates ranging from 2 to 1000 pg/mL with only 1 mL of urine sample, representing an improvement of 1 order of magnitude over the RPLC tandem MS methods previously reported. This method provided a linear dynamic range of 3 orders of magnitude, recovery of 92-109%, intraday accuracy of 84-109%, intraday precision of 1-14%, interday accuracy of 80-111%, and interday precision of 1-22%. We have successfully applied this technique to determine the seven estrogen conjugates in urine samples of a pregnant woman and found unique concentration changes of six estrogen conjugates at different stages of pregnancy while the concentration of estriol-3-glucuronide (E3-3G) remained constant. We further studied the profiles of individual estrogen conjugates in breast cancer patients before and after treatment and found patient-dependent effects of aromatase inhibitor treatment on estrogen phase-II metabolism, which have not been reported previously. This study demonstrates the potential clinical application of the HILIC-MS/MS technique for sensitive monitoring of the changes of urinary estrogen conjugates in a clinical setting. Estrogens play important roles in the development and maintenance of secondary sexual characteristics, pregnancy, and * To whom correspondence should be addressed. Phone: 780-492-5094. Fax: 780-492-7800. E-mail:
[email protected]. † University of Alberta. ‡ Cross Cancer Institute.
3404 Analytical Chemistry, Vol. 80, No. 9, May 1, 2008
long-bone maturation. They also have been found to be associated with the development of breast cancer. Conjugation to sulfate and glucuronide is one of the major metabolizing pathways of the estrogens, and these conjugates are excreted mainly through urine. Free estrogens in urine usually occur at extremely low levels and are not detectable. Estrogen metabolism varies significantly from person to person,1 suggesting that individual profiles of estrogens and their metabolites may provide information on variations in estrogen metabolism and cancer risk. The accurate measurement of these estrogen conjugates may further the study of the roles they play in estrogen-related physiological processes. A number of analytical methods have been used to determine estrogens. Immunoassays have been widely used for the determination of estrogens in biological matrixes because of their sensitivity and high-throughput capability for clinical analysis.2-4 However, these methods suffer from cross-reactions due to the similar structures of the estrogens. Gas chromatography/mass spectrometry (GC/MS) could determine individual estrogens with sensitivity and selectivity.5-8 However, because estrogen metabolites are polar and have low volatility, multiple steps including hydrolysis and derivatization are required for GC/MS analysis. These procedures are time-consuming, labor-intensive, and may introduce artifacts and thus are not preferred in clinical analysis. High-performance liquid chromatography (HPLC) has been used for the direct determination of estrogens and their metabolites. HPLC with electrochemical detection (HPLC-ECD) has been used to determine estrogens and their metabolites in animal urine9 and (1) Raftogianis, R.; Creveling, C.; Weinshilboum, R.; Weisz, J. J. Natl. Cancer Inst. Monogr. 2000, 27, 113-124. (2) Giese, R. W. J. Chromatogr., A 2003, 1000, 401-412. (3) England, B. G.; Parsons, G. H.; Possley, R. M.; McConnell, D. S.; Midgley, A. R. Clin. Chem. 2002, 48, 1584-1586. (4) Taieb, J.; Benattar, C.; Birr, A. S.; Lindenbaum, A. Clin. Chem. 2002, 48, 583-585. (5) Ding, W. H.; Chiang, C. C. Rapid Commun. Mass Spectrom. 2003, 17, 5663. (6) Gibson, R.; Tyler, C. R.; Hill, E. M. J. Chromatogr., A 2005, 1066, 33-40. (7) Zuo, Y. G.; Zhang, K.; Lin, Y. J. J. Chromatogr., A 2007, 1148, 211-218. (8) Quintana, J. B.; Carpinteiro, J.; Rodriguez, I.; Lorenzo, R. A.; Carro, A. M.; Cela, R. J. Chromatogr., A 2004, 1024, 177-185. (9) Nakagomi, M.; Suzuki, E. Chem. Res. Toxicol. 2000, 13, 1208-1213. 10.1021/ac702613k CCC: $40.75
© 2008 American Chemical Society Published on Web 03/19/2008
human tissues.10 However, ECD could not provide enough sensitivity for clinical analysis and usually suffered from significant background interference due to the complexity of the biological samples. Recently, HPLC-MS methods have been developed for analysis of estrogens and/or their metabolites in biological samples.11-17 These studies reported direct analysis of estrogens11 or analysis after derivatization with toluenesulfonhydrazide12,13 or dansyl chloride.14-17 Total estrogens were often measured after hydrolysis of the conjugates,12,13,16,17 and total conjugates were obtained by subtracting the free estrogens from the total.17 This approach could not determine the levels and changes of individual conjugates and positional isomers.12,13,16,17 For example, estriol-3glucuronide, estriol-16-glucuronide, and estriol-3-sulfate would all be recognized as estriol after hydrolysis. In addition, the hydrolysis of some estrogen metabolites may be incomplete and nonspecific.18 Several studies have developed reversed-phase (RP) LC-MS/ MS methods for the direct determination of estrogen conjugates.19-24 Zhang and Henion20 reported RPLC-MS/MS determination of three endogenous estrogen sulfates and two synthetic estrogen conjugates in human urine, and a few other studies reported several estrogen glucuronides and sulfates in urine19,22 or environmental samples.21-24 To the best of our knowledge, all these methods used RPLC separation with a mobile phase containing a small amount of organic solvent in water in order to retain and separate the highly polar estrogen conjugates. To obtain high-efficiency electrospray ionization (ESI) in the LC-MS, a mobile phase containing a high proportion of organic solvent is preferred. Hydrophilic interaction liquid chromatography (HILIC) uses a high-organic-low-aqueous mobile phase and a polar stationary phase to retain and separate polar compounds. HILIC offers several advantages, such as enhanced sensitivity in ESI-MS,25,26 higher speed due to the lower viscosity of the mobile (10) Rogan, E. G.; Badawi, A. F.; Devanesan, P. D.; Meza, J. L.; Edney, J. A.; West, W. W.; Higginbotham, S. M.; Cavalieri, E. L. Carcinogenesis 2003, 24, 697-702. (11) Draisci, R.; Palleschi, L.; Ferretti, E.; Marchiafava, C.; Lucentini, L.; Cammarata, P. Analyst 1998, 123, 2605-2609. (12) Xu, X.; Ziegler, R. G.; Waterhouse, D. J.; Saavedra, J. E.; Keefer, L. K. J. Chromatogr., B 2002, 780, 315-330. (13) Xu, X.; Keefer, L. K.; Waterhouse, D. J.; Saavedra, J. E.; Veenstra, T. D.; Ziegler, R. G. Anal. Chem. 2004, 76, 5829-5836. (14) Nelson, R. E.; Grebe, S. K.; O’Kane, D. J.; Singh, R. J. Clin. Chem. 2004, 50, 373-384. (15) Xia, Y. Q.; Chang, S. W.; Patel, S.; Bakhtiar, R.; Karanam, B.; Evans, D. C. Rapid Commun. Mass Spectrom. 2004, 18, 1621-1628. (16) Xu, X.; Veenstra, T. D.; Fox, S. D.; Roman, J. M.; Issaq, H. J.; Falk, R.; Saavedra, J. E.; Keefer, L. K.; Ziegler, R. G. Anal. Chem. 2005, 77, 66466654. (17) Xu, X.; Roman, J. M.; Issaq, H. J.; Keefer, L. K.; Veenstra, T. D.; Zieger, R. G. Anal. Chem. 2007, 79, 7813-7821. (18) Huang, C. H.; Sedlak, D. L. Environ. Toxicol. Chem. 2001, 20, 133-139. (19) Volmer, D. A.; Hui, J. P. M. Rapid Commun. Mass Spectrom. 1997, 11, 1926-1934. (20) Zhang, H. W.; Henion, J. Anal. Chem. 1999, 71, 3955-3964. (21) Isobe, T.; Shiraishi, H.; Yasuda, M.; Shinoda, A.; Suzuki, H.; Morita, M. J. Chromatogr., A 2003, 984, 195-202. (22) D’Ascenzo, G.; Di Corcia, A.; Gentili, A.; Mancini, R.; Mastropasqua, R.; Nazzari, M.; Samperi, R. Sci. Total Environ. 2003, 302, 199-209. (23) Gomes, R. L.; Birkett, J. W.; Scrimshaw, M. D.; Lester, J. N. Int. J. Environ. Anal. Chem. 2005, 85, 1-14. (24) Reddy, S.; Iden, C. R.; Brownawell, B. J. Anal. Chem. 2005, 77, 70327038. (25) Grumbach, E. S.; Wagrowski-Diehl, D. M.; Mazzeo, J. R.; Alden, B.; Iraneta, P. C. LCGC North Am. 2004, 22, 1010-1023.
phase compared to that of standard RP,27 and direct injection of the solid-phase extraction (SPE) elute onto the separation column without reconstitution.28 The objective of this study is to develop a method that can directly determine urinary estrogen metabolites, mainly estrogen sulfate and glucuronide conjugates, without hydrolysis or derivatization. We combine the advantages of HILIC separation with ESI-MS/MS detection to achieve rapid and highly sensitive analysis of urinary estrogen conjugates. EXPERIMENTAL SECTION Materials. Estrone-3-sulfate (E1-3S), estradiol-3-sulfate (E23S), estrone-3-glucuronide (E1-3G), estradiol-3-glucuronide (E23G), and estriol-16-glucuronide (E3-16G) were purchased from Sigma (St. Louis, MO); estriol-3-sulfate (E3-3S) and estriol-3glucuronide (E3-3G) were purchased from Steraloids (Newport, RI). The chemical structures of these analytes are shown in Figure 1. The internal standard (IS), androsterone glucuronide-d2, was obtained from CDN Isotopes (Pointe-Claire, QC, Canada). Optimagrade water, HPLC-grade acetonitrile and methanol, analyticalgrade ammonium acetate, ammonium hydroxide, acetic acid, and phosphoric acid were obtained from Fisher Scientific (Fair Lawn, NJ). Estrogen-free urine was prepared from human urine (Gemini Bio-Products, West Sacramento, CA) by shaking with activated charcoal for 30 min at room temperature. Preparation of Stock and Calibration Standards. Stock solutions of the analytes and IS (200 µg/mL) were prepared with 1 mg of each compound in 5 mL of methanol in amber vials; the solutions were stored at -20 °C. Working standard solutions of the analytes were prepared at 5, 1, 0.1, and 0.01 µg/mL by diluting the stock solutions with methanol. The IS of 0.5 µg/mL was used. Calibration solutions consisted of appropriate amounts of the analytes and IS in estrogen-free urine. For E1-3S, E2-3S, E3-3S, E1-3G, and E2-3G, the calibration solutions ranged from 0.002 to 10 ng/mL, and three levels (low, medium, and high) of the quality control (QC) samples were 0.07 (QC L), 0.7 (QC M), and 9.0 ng/ mL (QC H). For E3-16G, calibration standards ranged from 0.2 to 300 ng/mL, and QC samples were 0.4 (QC L), 30 (QC M), and 240 ng/mL (QC H). For E3-3G, calibration standards ranged from 1.0 to 1500 ng/mL, and QC samples were 2.0 (QC L), 150 (QC M), and 1200 ng/mL (QC H). Urine Sample Collection. Seven urine samples were collected in this study under informed consent. Three samples were collected from a pregnant woman (age 29) in the early (second month), middle (fourth month), and late (seventh month) stages of pregnancy. The other four samples were collected from two breast cancer patients (patient 1 at age 62; patient 2 at age 53). Two samples were collected from each of the patients; one was collected before treatment and the other after the patient had been treated with exemestane (a steroidal compound) for 1 month. All the samples were stored at -80 °C prior to analysis. The concentration of creatinine in each sample was measured based on Jaffe´’s reaction: creatinine interacts with alkaline picrate to (26) Garbis, S. D.; Melse-Boonstra, A.; West, C. E.; van Breemen, R. B. Anal. Chem. 2001, 73, 5358-5364. (27) Shou, W. Z.; Chen, Y. L.; Eerkes, A.; Tang, Y. Q.; Magis, L.; Jiang, X. Y.; Weng, N. D. Rapid Commun. Mass Spectrom. 2002, 16, 1613-1621. (28) Weng, N. D.; Shou, W. Z.; Addison, T.; Maleki, S.; Jiang, X. Y. Rapid Commun. Mass Spectrom. 2002, 16, 1965-1975.
Analytical Chemistry, Vol. 80, No. 9, May 1, 2008
3405
Figure 1. Structures of the seven estrogen conjugates.
form red addition products.29 This method is automated and routinely performed using a Roche Hitachi 917 analyzer in clinical laboratories (University of Alberta Hospital). The urinary creatinine concentrations were used to normalize the concentrations of the estrogen metabolites to control the variation of the volume of urine because the daily excretion of creatinine is constant in the same person. Sample Preparation. Urine samples were extracted using Waters Oasis HLB cartridges (1 mL, 30 mg per cartridge; Milford, MA) with a VISIPREP SPE manifold (Supelco, Bellefonte, PA). Before extraction, each HLB cartridge was preconditioned with 1 mL of methanol followed by 1 mL of water and 1 mL of aqueous phosphoric acid (0.3%, v/v). To a 1.0 mL aliquot of urine, 40 µL of the IS working standard (0.5 µg/mL) was added, followed by 1.0 mL of aqueous phosphoric acid (0.3%, v/v). The acidified urine was loaded onto a Waters Oasis HLB cartridge. After sample loading, the cartridge was washed with 1.0 mL of water followed by 1.0 mL of methanol/ water/acetic acid (60/40/2, v/v/v). The analytes were then eluted with 1.0 mL of methanol containing 2% ammonium hydroxide. The eluted fraction was evaporated to dryness under a stream of nitrogen gas at 40 °C. The residue was dissolved in 100 µL of acetonitrile/methanol/water (75/10/15, v/v/v). Calibration standards, QCs, and unknown samples were all extracted in the same way. LC-MS/MS Method. Separations were performed with an Agilent 1100 series LC system equipped with a binary pump and an autosampler (Agilent, Waldbronn, Germany). HILIC separation of the analytes was performed on a TSKgel Amide-80 (2.0 mm × (29) Kaplan, A.; Jack, R.; Opheim, K. E.; Toivola, B.; Lyon, A. W. Clinical Chemistry: Interpretation and Techniques, 4th ed.; Williams &Wilkins: Baltimore, MD, 1995; pp 172-180.
3406
Analytical Chemistry, Vol. 80, No. 9, May 1, 2008
150 mm, 5 µm, 80 Å; Tosoh Bioscience, Montgomeryville, PA) with a guard cartridge (2.0 mm × 10 mm) at room temperature. Isocratic elution was used, and the mobile phase was acetonitrile/ aqueous ammonium acetate (5 mM, pH 6.80) (85/15, v/v). The flow rate of the mobile phase was 200 µL/min, and the injection volume was 20 µL. All the eluent from the LC column was directly transferred into the ion source of the mass spectrometer without postcolumn splitting. RPLC separation of the analytes was performed on a Luna C18(2) column (100 mm × 2.0 mm i.d., 3 µm; Phenomenex, Torrance, CA) with a step gradient elution. Mobile phase A was acetonitrile/ aqueous ammonium acetate (5 mM, pH 6.80) (5/95, v/v), and mobile phase B was acetonitrile/aqueous ammonium acetate (5 mM, pH 6.80) (95/5, v/v). The gradient was as follows: 0-2 min, 20% B; 2-10 min, 30% B; 10-18 min, 80% B; 18-25 min, 20% B. The flow rate of the mobile phase was 150 µL/min, and the injection volume was 20 µL. A hybrid quadrupole/linear ion trap 4000 Q TRAP system (Applied Biosystems/MDS Sciex, Concord, ON, Canada) equipped with a Turbo ion spray source was employed to determine the estrogen conjugates. The electrospray ionization was operated in negative mode. Quantification of the analytes was performed by multiple reaction monitoring (MRM). RESULTS AND DISCUSSION Mass Detection and HILIC Separation of the Estrogen Metabolites. The electrospray ionization and MS detection of the estrogen conjugates were examined using direct infusion. The transition ions of the seven analytes were selected based on the highest intensities of the ions in the ESI-MS spectra we obtained, which were consistent with those reported elsewhere.20-22 Two pairs of transition ions for each analyte were used in this study.
Table 1. MRM Conditions Used for HILIC-MS/MS Analysis of the Estrogen Conjugates and the Internal Standard (ESI, Negative Mode) analytes E1-3S
precursor ion (m/z)
product ion (m/z)
DP
EP
CE
CXP
349
269a
-60
-10
-54
-13
-65
-10
-50
-13
-65
-10
-50
-15
-75
-10
-50
-15
-75
-10
-50
-15
-75
-10
-50
-15
-75
-10
-50
-15
-80
-10
-50
-15
E2-3S
351
E3-3S
367
E1-3G
445
E2-3G
447
E3-16G
463
E3-3G
463
IS
467
a
145 271a 80 287a 80 269a 113 271a 113 287a 113 287a 113 289a 291
The more abundant product ion was used for quantitative analysis.
The paired ions with the higher intensity were used for quantification, but both pairs of ions were used for confirmation. Several compound-dependent parameters were optimized by direct infusion of individual standards (100 ng/mL) at a flow rate of 20 µL/ min using a syringe pump. The compound-dependent parameters, including the declustering potential (DP), entrance potential (EP), collision energy (CE), and collision cell exit potential (CXP), were optimized and are summarized in Table 1. The source-dependent parameters were optimized at the following values: ionspray voltage, -4500 V; ionspray source (gas 2) temperature, 475 °C; curtain gas, 10 (arbitrary unit); gas 1, 50 (arbitrary unit); gas 2, 60 (arbitrary unit). The determination of estrogen conjugates in urine samples requires separation in order to eliminate matrix interference. To improve sensitivity, HILIC separation was examined as an alternative strategy to the commonly used RPLC separation. The column selected was the TSKgel Amide-80, in which the carbamoyl groups were attached onto the surface of silica through an aliphatic carbon chain.30 This column showed HILIC characteristics when separating peptides30 and some small polar compounds31 under highorganic-low-aqueous mobile phases. The initial mobile phase used in our experiment was acetonitrile/aqueous ammonium acetate (1 mM, pH 6.80) (90/10, v/v). The retention and separation of the glucuronides on the column exhibited typical HILIC characteristics. Decreasing the acetonitrile content from 90% to 80%, or increasing the salt concentration from 1 to 10 mM increased significantly the retention and resolution of the estrogen glucuronides. Varying the pH of the mobile phase from 3.0 to 6.8 increased slightly the retention of the four glucuronides, which may be due to increased hydrophilicity of the conjugates at higher pH (pKa of the glucuronide conjugate was ∼3.68). Changing the mobile phase conditions described above had little effect on the retention behavior of the three sulfate conjugates on the HILIC column. (30) Yoshida, T. Anal. Chem. 1997, 69, 3038-3043. (31) Guo, Y.; Gaiki, S. J. Chromatogr., A 2005, 1074, 71-80.
Figure 2. Chromatograms of the estrogen conjugates in the extract of spiked estrogen-free urine. Chromatographic conditions are described in the Experimental Section.
The composition of the mobile phase is known to affect electrospray ionization efficiency, which can, in turn, affect the sensitivity of the MS detection. In this study, HILIC separation uses a higher proportion of acetonitrile in the mobile phase, providing desirable conditions for ESI-MS. To demonstrate the enhanced sensitivity of HILIC-MS/MS, standard solutions of the seven conjugates were analyzed using both HILIC-MS/MS and RPLC-MS/MS with the same instrumentation (see the Experimental Section). The sensitivity of HILIC-MS/MS for the seven conjugates was 2-10-fold higher than that of RPLC-MS/MS, showing that HILIC separation provides a significant improvement in sensitivity over RPLC. In order to obtain both reproducible separation and high sensitivity within a reasonable separation time (
