Global Metabolomics and Targeted Steroid Profiling Reveal That

Jan 16, 2013 - ... dehydroepiandrosterone (DHEA) sulfate, and p-cresol sulfate levels were significantly decreased following rifampin administration c...
0 downloads 0 Views 1MB Size
Download PDF
Article pubs.acs.org/jpr

Global Metabolomics and Targeted Steroid Profiling Reveal That Rifampin, a Strong Human PXR Activator, Alters Endogenous Urinary Steroid Markers Bora Kim,†,§ Ju-Yeon Moon,‡,§ Man Ho Choi,‡ Hyang Hee Yang,† SeungHwan Lee,† Kyoung Soo Lim,† Seo Hyun Yoon,† Kyung-Sang Yu,† In-Jin Jang,† and Joo-Youn Cho*,† †

Department of Clinical Pharmacology and Therapeutics, Seoul National University College of Medicine and Hospital, Seoul, Republic of Korea ‡ Future Convergence Research Division, Korea Institute of Science and Technology, Seoul, Republic of Korea S Supporting Information *

ABSTRACT: Activation of the pregnane X receptor (PXR) alters the expression of metabolic enzymes and transporters involved in the metabolism of xenobiotics and endobiotics. To identify endogenous biomarkers of PXR activation in humans, rifampin, a strong PXR activator, was administered to 12 healthy male subjects, and their urine was analyzed before and after rifampin administration. Ultraperformance liquid chromatography time-of-flight mass spectrometry (UPLC/ QTOF−MS)-based global metabolomics and gas chromatography− mass spectrometry (GC−MS)-based profiling of 75 steroids were used to screen the urine samples. Global metabolomics revealed that hydroxytestosterone sulfate and glycochenodeoxycholate sulfate levels were significantly increased and that androsterone sulfate, dehydroepiandrosterone (DHEA) sulfate, and p-cresol sulfate levels were significantly decreased following rifampin administration compared with controls. Urinary steroid profiling showed that 16α-OHandrostenedione (16α-OH-A-dione), 16α-OH-DHEA, 7α-DHEA, 7β-DHEA, and 11β-OH-A-dione levels were increased, whereas DHEA, androsterone, etiocholanolone, estrone, β-cortolone, and allo-tetrahydrocortisone levels were decreased in the rifampin group. The analysis of the metabolic pathway and the metabolic ratio of steroids enabled the estimation of the induction of CYP1A/3A/7B/11B/2C and the inhibition of CYP17A/19A in response to PXR activation. These human urinary biomarkers may be useful for predicting the extent of PXR activation, monitoring the activity of DMEs, and anticipating drug−drug interactions in patients administered PXR-activating drugs. KEYWORDS: untargeted and targeted metabolomics, PXR, biomarker, ultraperformance liquid chromatography time-of-flight mass spectrometry (UPLC/QTOF−MS), gas chromatography−mass spectrometry (GC−MS), steroid profiling

■

and metabolizes over 50% of clinical drugs.5 The expression of CYP3A4 is highly inducible and varies widely among individuals.6 In addition to its role as a xenobiotic receptor, PXR regulates the biotransformation and homeostasis of numerous endogenous chemicals, such as bile acids, bilirubin, sterols, and steroid hormones, by activating the same enzyme and transporter systems.2 Accumulating evidence suggests that PXR plays important roles in the maintenance of normal physiology and progression of various diseases, including hepatic steatosis, bile acid homeostasis, steroid hormone homeostasis, and inflammatory bowel diseases.7,8 For example, the excess toxic bile acid, lithocholic acid, is efficiently eliminated through the induction of CYP3A4 and SULT2 via PXR activation.9,10 In

INTRODUCTION Pregnane X receptor belongs to the nuclear receptor superfamily and plays a crucial role in regulating the metabolism of therapeutic drugs.1 Activated PXR forms a heterodimer with the retinoid X receptor (RXR, NR1B); binds to PXR-response elements (PXREs) that are usually located upstream of PXR target gene regions; and induces the expression of drugmetabolizing enzymes (DMEs) (e.g., phase I enzymes, cytochrome P450s (CYPs), phase II enzymes, UDP-glucuronosyltransferases (UGTs) and sulfotransferases (SULTs)), and transporters2 (e.g., multidrug resistance protein 1 (MDR1), multidrug resistance-related protein-3 (MRP3), and organic anion transporting polypeptide-2 (OATP2)). Among these enzymes and transporters, the most significantly induced PXR target genes are the CYP3A and CYP2B subfamilies.1,3,4 Human CYP3A4, the most abundant CYP enzyme expressed in the liver and small intestine, has very broad substrate specificity © 2013 American Chemical Society

Received: October 29, 2012 Published: January 16, 2013 1359

dx.doi.org/10.1021/pr301021p | J. Proteome Res. 2013, 12, 1359−1368

Journal of Proteome Research

Article

analysis. Any other medications and the consumption of alcohol and beverages containing xanthine derivatives were prohibited during the study. Because genetic polymorphisms of PXR and its target CYP3A might affect PXR activation, we measured two well-known single-nucleotide polymorphisms (SNPs), −25385C>T (rs3814055) of NR1I2 (PXR) and CYP3A5*3 (rs776746), as previously reported.19,20

transgenic mice, the activation of PXR by rifampin markedly increased plasma concentrations of corticosterone and aldosterone, which was associated with the induction of adrenal steroidogenic enzymes, including Cyp11a1, Cyp11b1, Cyp11b2, and 3beta-Hsd.11 Furthermore, CYP enzymes are involved in the synthesis and metabolism of endogenous steroids, including corticoids and sex hormones. Despite increasing evidence for the effects of PXR on metabolism, alterations in various endogenous metabolites in human urine have not been well characterized. Metabolomics, the global analysis of thousands of small molecules present in biological fluids, promises to significantly impact both basic biological research and medical practice.12,13 The available analytical platforms, namely, gas chromatography−mass spectrometry (GC−MS), liquid chromatography (LC)−MS, and nuclear magnetic resonance (NMR) spectrometry, are suitable for mapping biochemical signatures in both targeted and untargeted studies.14 In addition, we have reported that urinary steroid profiling analysis using GC−MS generates quantitative results that can explain correlations between enzyme activities.15 In this report, 75 deglucuronidated and desulfated steroids, including 28 androgens, 11 estrogens, 24 corticoids, 9 progestins, and 3 sterols, were validated and quantified in human urine. Recently, a global metabolomic study of PXR activation was performed in a mouse model and revealed that PXR activation attenuated the levels of two vitamin E conjugates.16 However, metabolomic studies of PXR activation by rifampin treatment in humans have not been reported. In this study, using both global metabolomics and validated steroid profiling, we aimed to identify changes in endogenous metabolite levels in human urine that are associated with the activation of PXR. Furthermore, the altered enzyme activities related to metabolic pathways were assessed using the ratios of precursor and product metabolites. Finally, the alteration of steroid metabolism by PXR activation was investigated.

■

Chemicals and Materials

Formic acid, β-glucuronidase (type H-1 from Helix pomatia), and β-sulfatase (type H-1 from Helix pomatia) were purchased from Sigma Aldrich (St. Louis, MO). HPLC-grade solvents (methanol and water) were purchased from J.T. Baker (Center Valley, PA). The reference standards used in this study were obtained from Sigma (St. Louis, CA), Steraloids (Newport, RI), and NARL (Pumble, Australia). The quantified steroids and their abbreviations are listed in Supplementary Table 1. The internal standards used, purchased from NARL and C/D/N isotopes (Pointe-Claire, Quebec, Canada), were as follows: d4androsterone, d3-testosterone, and methyltestosterone for androgens; d4-17β-estradiol for estrogens; d4-cortisol for corticoids; d9-progesterone and d8-17α-hydroxyprogesterone for progestins; and d6-cholesterol for sterols. For solid-phase extraction (SPE), an Oasis HLB cartridge (3 mL, 60 mg; Waters, Milford, MA) was preconditioned with 3 mL of methanol, followed by 3 mL of deionized water. Sodium phosphate monobasic (reagent grade), sodium phosphate dibasic (reagent grade), and L-ascorbic acid (reagent grade) were obtained from Sigma. A 50% glycerol solution of βglucuronidase, extracted from E. coli (140 U/mL), was purchased from Roche Diagnostics GmbH (Mannheim, Germany). The trimethylsilylating (TMS) agents N-methylN-(trimethylsilyl)trifluoroacetamide (MSTFA), ammonium iodide (NH4I), and dithioerythritol (DTE) were purchased from Sigma. All organic solvents were of analytical or HPLC grade and were purchased from Burdick & Jackson (Muskegan, MI). Deionized water was prepared using the Milli-Q purification system (Millipore, Billerica, MA).

EXPERIMENTAL PROCEDURES

Sample Preparation and Instrument Conditions for UPLC/QTOF

Subjects

Healthy Korean male volunteers between the ages of 20 and 50 years, weighing between 50 and 90 kg, and with a body mass index (BMI) in the range of 17 to 28 kg/m2 participated in the study. The volunteers were ascertained to be medically healthy by assessing their medical history and performing a physical examination, clinical laboratory tests, 12-lead ECG, and measurements of vital signs. The purpose and procedures of the study were explained, and a written informed consent form was obtained from all volunteers prior to the study.

Urine was prepared by centrifugation at 15 000 × g for 20 min at 4 °C to remove particles, and the supernatant was diluted with 4 volumes of water. QC samples, which were prepared by pooling equal aliquots from all of the samples, were injected before processing the study samples and throughout the entire batch. The biological standard QC sample was used to assess the quality and reproducibility of the data generated during the analytical workflow (retention time drift, mass precision, and fluctuation of the ion responses over time).21 A 5 μL aliquot of the diluent was injected into a reversephase 2.1 × 50 mm ACQUITY 1.8 μm HSS T3 column (Waters Corp, Milford, MA) using an ACQUITY UPLC system (Waters Corp.) with a gradient mobile phase comprising 0.1% formic acid (A) and methanol containing 0.1% formic acid (B). Each sample was resolved for 20 min at a flow rate of 0.4 mL/min, and the gradient consisted of 95% A for 1 min, 70% A for 8 min, 30% A for 13 min, and 5% A for 14 min maintaining for 2 min. The samples were then equilibrated at 95% A for 3.5 min before the next injection. The QTOF (Waters Corp.) was operated in either the positive ion or negative ion electrospray ionization (ESI+ or ESI−) mode. The capillary and cone voltages were maintained at 3 kV and 40 V, respectively. Nitrogen was used as both a desolvation gas (850

Clinical Study

This open-label study was conducted according to a singlesequence design, in accordance with the Declaration of Helsinki and Korea Good Clinical Practice (KGCP). The study protocol and informed consent form were approved by the institutional review board (IRB) of Seoul National University Hospital (SNUH), Seoul, Korea. On the first day of the study, 12 h urine samples were collected before rifampin administration. The 12 healthy volunteers received oral rifampin (600 mg) once daily from day 2 to day 7, followed by a 12 h urine collection on day 8. The dose and treatment time of rifampin was determined according to many clinical studies for drug−drug interaction.17,18 The urine samples were frozen at −70 °C until 1360

dx.doi.org/10.1021/pr301021p | J. Proteome Res. 2013, 12, 1359−1368

Journal of Proteome Research

Article

Biomarker Deconjugation for Global Metabolomics

L/h) and a cone gas (50 L/h). The source and desolvation temperatures were set at 120 and 350 °C, respectively. The QTOF was calibrated with sodium formate solution (range, m/ z 50−1200) to ensure mass accuracy and was monitored using a lock mass, leucine enkephaline, in real time. To assess the analytical variability of the LC−MS platform, a pooled urine sample was prepared by mixing aliquots of individual samples. Replicates of a pooled urine sample were run in a series of injections, and triplicate individual sample data were obtained by random injection. The data were acquired in centroid mode, from m/z 100 to 1000, using MassLynx (Waters Corp.).

Metabolites that had a fragmentation pattern of glucuronide or sulfate conjugation were deconjugated using enzymes, and the free form of each metabolite was identified by tandem MS fragmentation. The fragmentation patterns of glucuronidated compounds have peaks at m/z 157.02, m/z 113.02, m/z 85.03, and m/z 75.01, while sulfated compounds have fragmentation pattern peaks at m/z 96.96 in the negative ion mode.22 The deconjugation system for urinary metabolites contained the following (in a total volume of 110 μL): 20 μL of urine, 40 U of β-glucuronidase or 10 U of sulfatase, and 100 mM sodium acetate buffer (pH 3.8) for glucuronidase or 200 mM sodium acetate buffer (pH 5.0) for sulfatase. A blank sample containing buffer instead of enzymes was used as a control. The mixtures were incubated for 12 h at 37 °C, and the reactions were terminated by adding 100 μL of cold acetonitrile. The mixtures were then subjected to centrifugation at 13 000 × g for 20 min at 4 °C. Five microliter aliquots were injected into the UPLC/ QTOF system.

Global Metabolomic Data Analysis

To exclude rifampin-derived ions from the chromatographic mass data, MetaboLynx (Waters Corp.) was used to generate a table of the following rifampin metabolites: parent ([M + H]+ 823.4129), desaturation + demethylation − C2H2O2 ([M − C3H6O2]+ 749.3757), desaturation − CH2O ([M − CH4O]+ 791.3863), desaturation − O ([M − H2O]+ 805.3906), hydroxylation ([M − OH]+ 839.41), glucuronide conjugation − C2H2O ([M + C4H6O5]+ 957.4396), glucuronide conjugation ([M + C6H8O6]+ 999.4441), parent ([M − H]− 821.3981), parent − C2H2O ([M − C2H2O]− 779.3868), glucuronide conjugation − C 2 H 2 O ([M + C 4 H 6 O 5 ] − 955.4214), and glucuronide conjugation ([M + C6H8O6]− 997.4276) compounds. Chromatographic mass data were aligned, and the rifampin metabolites listed above were removed using MarkerLynx (Waters Corp.). The parameters of the MarkerLynx method were set as follows: mass tolerance, 0.05 Da; noise elimination level, 6; employment of the full-scan mode in the mass range of 100−900 amu; and initial and final retention times of 0.5 and 10 min, respectively, for data collection. All data were normalized to the sum of the total ion intensity per chromatogram, and the resultant data were imported to EZinfo software (Umetrics) for principal components analysis (PCA), partial least-squares discriminant analysis (PLS-DA), and orthogonal partial least-squares discriminant analysis (OPLS-DA). Prior to PCA, all variables obtained from the LC−MS data sets were Pareto-scaled to increase the importance of low-abundance ions without significant amplification of noise. PCA, an unsupervised multivariate statistical approach, was performed to achieve the natural inter-relationship (grouping, clustering, and outlier detection) among samples and quality controls. To maximize class discrimination, the OPLS-DA results were further analyzed. S-plots were calculated to visualize the relationship between covariance and correlation among the OPLS-DA results. Variables that significantly contributed to the discrimination between groups were considered as potential biomarkers and subjected to further analysis to identify their molecular formulas. From PLS-DA analysis, a variable importance plot (VIP) was created to characterize the responsibility of differentiation.

Quantification of Biomarkers for Global Metabolomics

QuanLynx (Waters Corp.) was used to quantify androsterone sulfate and DHEA sulfate, of which authentic compounds were commercially available. Creatinine was also quantified to normalize to the actual concentrations of each urinary biomarker. One hundred microliters of urine supernatant was diluted with 400 μL of two internal standard mixtures: 100 ng/ mL of d 6 -DHEA sulfate ([M − H] − 373.1953) for androsterone sulfate and DHEA sulfate and 1 μg/mL of Ltryptophan methyl ester ([M + H]+ 219.1134) for creatinine. Calibration curves were constructed from 20 to 5000 ng/mL of androsterone sulfate and DHEA sulfate and from 100 to 10 000 ng/mL of creatinine. The concentration of each biomarker in urine was determined from the calibration curves using linear regression analysis. All determined correlation coefficients were >0.99 for each biomarker, and the resultant concentrations were expressed as μmol/mmol creatinine (normalized). Sample Preparation and Instrument Conditions for GC−MS

The method used for quantitative metabolite profiling of urinary steroids was based on previous reports.15,23 Briefly, urine samples (2 mL), spiked with 20 μL of the 8 internal standards (d3-testosterone and d4-estradiol, 1 μg/mL; d4cortisol and d8-17α-hydroxyprogesterone, 5 μg/mL; d4androsterone, methyltestosterone, d9-progesterone, and d6cholesterol, 10 μg/mL), were extracted with Oasis HLB SPE cartridges coupled to a peristaltic pump. After loading a sample onto a cartridge, the cartridge was washed with 2 mL of water and eluted twice with 2 mL of methanol. The combined methanol eluates were evaporated under a stream of nitrogen, followed by the addition of 1 mL of 0.2 M phosphate buffer (pH 7.2), 100 μL of aqueous 0.2% ascorbic acid, and 50 μL of β-glucuronidase. After incubation at 55 °C for 1 h, the solution was extracted twice with 2.5 mL of ethyl acetate/n-hexane (2:3, v/v). The combined organic solvents were evaporated using a N2 evaporator at 40 °C and further dried in a vacuum desiccator over P2O5−KOH for at least 30 min. Finally, the dried residue was derivatized with MSTFA/NH4I/DTE (40 μL; 500:4:2, v/w/w) at 60 °C for 20 min, and 2 μL of the resulting mixture was subjected to GC−MS in the selected-ion monitoring (SIM) mode. GC−MS was performed using an Agilent 6890 Plus gas chromatograph interfaced with a single-quadruple Agilent 5975 MSD at an electron energy of 70 eV and an ion source

Biomarker Identification for Global Metabolomics

The elemental composition of ions with a high correlation coefficient (pcorr value) was investigated further by MassLynx, tandem MS fragmentation, database searching, deconjugation of urinary metabolites, and confirmation using authentic standards. In addition, the Human Metabolome Database (HMDB, http://www.hmdb.ca/) and Massbank database (http://www.massbank.jp/) were used to perform a massbased identification search. 1361

dx.doi.org/10.1021/pr301021p | J. Proteome Res. 2013, 12, 1359−1368

Journal of Proteome Research

Article

Figure 1. PCA, OPLS-DA score, and loading plots comparing the control and rifampin groups derived from untargeted metabolomic analysis: positive ion mode (a−c) and negative ion mode (d−f). The ions labeled with the red boxes in the OPLS loading S-plot are the selected significant ions (upper-right quadrant, increased ions; lower-left quadrant, decreased ions).

temperature of 230 °C. Each sample (2 μL) was injected in split mode (10:1) at 280 °C and separated through an Ultra-1 capillary column (25 m × 0.2 mm i.d., 0.33 μm film thickness; Agilent Technologies; Palo Alto, CA). The GC oven temperature was initially set at 215 °C, increased to 260 °C at 1 °C/min, and finally increased to 320 °C at 15 °C/min and held for 1 min. The carrier gas, helium, was held at a column head pressure of 210.3 kPa (column flow, 1.0 mL/min at an oven temperature of 215 °C). For quantitative analysis, characteristic ions of each steroid were determined as their TMS derivatives. Peak identifications were achieved by comparing retention times and matching the height ratios of characteristic ions.

PXR and CYP3A5*3, were analyzed individually among the subjects. Four subjects had −25385C/T of PXR, and the other subjects had −25385C/C. In terms of CYP3A5 genotypes, two subjects had CYP3A5*1/*1, four had CYP3A5*1/*3, and six had CYP3A5*3/*3. There were no significant differences in demographic characteristics between the genotypes. Global Metabolomic Analysis

Urine samples from the control and rifampin groups were analyzed by UPLC/QTOF, operated in both positive and negative ionization modes. A large data matrix containing approximately 21 000 ions (positive mode) and 12 000 ions (negative mode) was produced using MarkerLynx. After data normalization, PCA, PLS, and OPLS-DA were performed to classify the metabolic phenotypes and identify significant biomarkers. Unsupervised PCA analysis showed a clear separation between the control and rifampin groups (Figure 1a,d). The QC samples clustered in the middle of the PCA plot, confirming the long-term stability of the instrument and the reproducibility of the data obtained in both the positive and negative modes. The supervised method, OPLS-DA, was used to isolate the variables responsible for the differences observed between the rifampin and control groups (Figure 1b,e). To highlight metabolite differences between the control and rifampin groups, feature selections were performed using Splot, followed by OPLS-DA (Figure 1c,f). After rifampin administration, the ions with significantly increased and decreased concentrations appeared in the upper-right quadrant and lower-left quadrant, respectively. The further away the points were from the center, the higher the VIP value was scored.

Steroid Profiling Data Analysis

Statistical analyses were performed using SAS 9.3 software (SAS Institute Inc.). The selection of both steroids and the ratios of metabolites to precursors (an indicator of enzyme activity) whose mean values were significantly different between the two groups was performed using a paired t-test and Benjamini−Hochberg multiple testing.24 Benjamini−Hochberg multiple testing was controlled by the false discovery rate (FDR), which is the expected proportion of false positives among the tests that is declared significant. An FDR adjusted pvalue < 0.05 was considered to be significant. To visualize differences between the steroid signatures of subjects, heat maps were generated by hierarchical clustering using the unweighted pair group method, using the arithmetic mean (UPGMA) and Pearson’s correlation as similarity measures (Spotfire, Tibco, Palo Alto, CA). Red, gray, and blue indicate high, medium, and low concentrations of metabolites, respectively.

■

Biomarker Identification for Global Metabolomics

RESULTS

To identify the structure of biomarkers, the tandem MS fragmentation pattern and elemental composition were ascertained. Most of the negative ion biomarkers had MS fragmentation patterns of glucuronide or sulfate conjugation, and mass-based HMDB search results also revealed some conjugation forms for which the chemical formulas were

Subjects

The mean and SD of age, height, weight, and BMI of the enrolled 12 subjects were 23.0 ± 4.3 years, 170.0 ± 6.0 cm, 67.4 ± 7.8 kg, and 23.3 ± 2.1 kg/m2. Two SNPs, −25385C>T of 1362

dx.doi.org/10.1021/pr301021p | J. Proteome Res. 2013, 12, 1359−1368

Journal of Proteome Research

Article

Table 1. Identified Urinary Metabolites from the Untargeted Metabolomic Analysis of the Control and Rifampin Groups (n = 12) [M − H]−

RT (min)

VIP

metabolite

mass error (ppm)

regulation

fold change (rifampin/control)

383.1523 383.1526 528.2623 367.1556 369.1729 187.0064

11.07 10.28 13.59 12.15 12.9 4.82

9.97 6.14 7.25 12.45 12.43 41.52

hydroxytestosterone sulfate I hydroxytestosterone sulfate II GCDCA sulfate DHEA sulfate androsterone sulfate p-cresol sulfate

2.6 2.1 2.5 7.8 3.3 3.5

up up up down down down

1.84 2.49 2.59 0.31 0.21 0.08

Figure 2. Box plots of the identified altered metabolites in the control and rifampin groups based on the global metabolomics analysis (*, P < 0.005): a, androsterone sulfate; b, DHEA sulfate; c, GCDCA sulfate; d, hydroxytestosterone sulfate I; e, hydroxytestosterone sulfate II; and f, p-cresol sulfate.

Table 2. Putative Urinary Metabolites from the Untargeted Metabolomics Analysis of the Control and Rifampin Groups (n = 12) [M − H]−

RT (min)

VIP

putative metabolite

mass error (ppm)

regulation

fold change (rifampin/control)

385.1676 479.2285 465.2480 481.2429 369.1735 481.2420

10.49 12.52 12.78 12.91 12.42 11.87

11.69 5.74 4.94 4.13 5.28 4.42

C19H30O3a + sulfate C19H28O3b + glucuronide C19H30O2c + glucuronide C19H30O3a + glucuronide C19H30O2c + sulfate C19H30O3a + glucuronide

2.3 0.8 1.7 1.9 0.3 3.7

up up up up down down

2.32 3.55 2.94 2.61 0.08 0.07

a

C19H30O3: hydroxyandrosterone and androstenetriol. bC19H28O3: hydroxytestosterone, dihydroxyandrostenone, hydroxydehydroisoandrosterone, oxoandrostenediol, hydroxy-DHEA, and ketoetiocholanolone. cC19H30O2: dihydrotestosterone, etiocholanolone, androsterone, and androstenediol.

matched with the highest scoring formulas of the elemental composition results from MassLynx. The chromatograms and subsequent MS/MS spectra of the identified metabolites are indicated in Supplementary Figures 1−5. Hydroxytestosterone sulfate I and II and GCDCA sulfate were identified as follows (Supplementary Figures 1 and 2): (1) a MS/MS fragmentation pattern of sulfate conjugation was present in blank urine; (2) the peak area of the sulfated metabolite decreased, and a desulfated metabolite peak arose in deconjugated urine that was not present in blank urine; and (3) the urinary deconjugated metabolite and the authentic compound had identical MS/MS spectrum patterns. The authenticity of androsterone sulfate and

DHEA sulfate were confirmed by their MS/MS spectrum (Supplementary Figures 3 and 4). For the identification of pcresol sulfate, the MS/MS spectrum pattern of deconjugated urinary p-cresol was compared with the Massbank database and other published results (Supplementary Figure 5).25 Metabolite Changes Identified by Global Metabolomics

The urinary metabolites were affected by rifampin treatment in the following manner: hydroxytestosterone sulfate I and II and GCDCA sulfate were significantly increased by 1.84-, 2.49-, and 2.59-fold, respectively, and DHEA sulfate, androsterone sulfate, and p-cresol sulfate were significantly attenuated by 0.31-, 0.21-, and 0.08-fold, respectively, compared with controls. A list of the 1363

dx.doi.org/10.1021/pr301021p | J. Proteome Res. 2013, 12, 1359−1368

Journal of Proteome Research

Article

Table 3. Summary of Significantly Changed Steroids from GC−MS Steroid Profilesa steroid concentrations (ng/mL) steroids 11β-OH-A-dione 16α-OH-DHEA 16α-OH-A-dione 7β-OH-DHEA 7α-OH-DHEA 16-epi-E3 Epi-DHT 11-dehydroB F E1 Iso-P-one β-cortolone 2-OH-E1 allo-THE An DHEA Etio 2-MeO-E2 β-cortol P-one Allo-P-one E2 P-diol 11β−OH-Etio 2-OH-E3 a

control 28.3 ± 12.7 74.6 ± 18.1 28.0 ± 7.57 177 ± 79.7 46.6 ± 35.6 31.9 ± 9.03 12.8 ± 4.47 63.1 ± 43.9 138 ± 103 35.8 ± 10.4 3.31 ± 1.23 2.61 × 103 ± 16.8 ± 3.57 1.28 × 103 ± 1.29 × 104 ± 226 ± 55.4 9.30 × 103 ± 6.38 ± 1.40 2.74 × 103 ± 194 ± 123 52.8 ± 30.7 9.72 ± 3.38 1.95 × 103 ± 1.44 × 103 ± 5.60 ± 1.51

815 528 1.64 × 103 2.37 × 103 795

1.39 × 103 951

rifampin

regulation

fold change (rifampin/control)

54.5 ± 15.9 188 ± 50.4 51.9 ± 12.7 305 ± 110 125 ± 53.1 55.7 ± 17.2 22.1 ± 5.66 105 ± 50.7 196 ± 85.2 14.5 ± 3.11 0.721 ± 0.576 1.13 × 103 ± 416 10.3 ± 3.06 504 ± 208 9.17 × 103 ± 1.86 × 103 135 ± 33.0 6.47 × 103 ± 1.76 × 103 5.00 ± 1.65 1.76 × 103 ± 602 78.5 ± 24.1 30.8 ± 16.1 6.37 ± 1.48 765 ± 287 562 ± 200 4.31 ± 1.81

up up up up up up up up up down down down down down down down down down down down down down down down down

1.92 2.52 1.85 1.73 2.68 1.74 1.73 1.66 1.42 0.406 0.218 0.431 0.613 0.393 0.713 0.597 0.696 0.784 0.642 0.404 0.583 0.655 0.392 0.390 0.770

FDR (