Urinary Metabolic Fingerprinting for α-naphthylisothiocyanate-induced Intrahepatic Cholestasis in Rats Using Fourier Transform-ion Cyclotron Resonance Mass Spectrometry

Abstract

Urinary metabolic fingerprinting with Fourier transform-ion cyclotron resonance mass spectrometry (FT-ICR MS) was performed to monitor metabolic changes in an α-naphthylisothiocyanate (ANIT)-induced rat model of intrahepatic cholestasis and to investigate the relationships among metabolic changes, histopathology, and blood chemistry. ANIT was administered orally as a single dose of 100 mg/kg. Urine samples were collected predose (–31 to –24 hours) and postdose at 0–7, 7–24, 24–31, 31–48, 48–55, 55–72, and 72–96 hours, and serum samples were collected on days 1, 2, and 4 postdose. Increased levels of alanine aminotransferase, aspartate aminotransferase, alkaline phosphatase, and total bilirubin were seen on day 2. The negative ion profiles for urine samples collected after 7–24, 24–31, 31–48, and 48–55 hours differed from the predose profile based on principal component analysis. Onset of recovery was observed after 24–31 hours, when the urinary composition reverted toward the predose position. In conclusion, it is possible to monitor the progression of and recovery from drug-induced hepatotoxicity by urinary metabolic fingerprinting with FT-ICR MS and to search for potential biomarkers involved in intrahepatic cholestasis.

Introduction

The liver is the first organ to be exposed when drugs are administered perorally or via the portal vein, and the concentration of drugs in the liver is often much higher than the peak plasma concentration. The liver is also the major site for metabolism of xenobiotics, and administration of such compounds can lead to formation of active metabolites. Thus, the liver is an important organ for evaluation in the development of new therapeutic agents. α-naphthylisothiocyanate (ANIT) is a typical hepatotoxin that is used to study intrahepatic cholestasis, since its effects are dose dependent and highly reproducible between studies (Capizzo and Roberts 1971; Goldfarb et al. 1962; Kossor et al. 1998). ANIT induces bile duct epithelial cell necrosis and hepatic parenchymal cell injury, followed by bile duct proliferation, cessation of bile flow, and consequent hyperbilirubinemia (Goldfarb et al. 1962; Plaa and Priestly 1976). ANIT toxicity is mediated through a dithiocarbamyl-linked GSH-ANIT conjugate that is released from hepatic parenchymal cells and directed into the bile, where it dissociates into reduced glutathione and ANIT, thereby exposing the bile duct to a high concentration of ANIT (Carpenter-Deyo et al. 1991; Dahm and Roth 1991; Jean and Roth 1995).

Metabolomics, defined as an attempt to identify and quantify all of metabolites within a cell, tissue, or organism during a genetic modification or physiological stimulus (Fiehn 2002; Oliver et al. 1998; Tweeddale et al. 1998), has recently been developed. The technologies involved in metabolomics can be classified into two categories according to the detector: mass spectrometry (MS) and nuclear magnetic resonance (NMR). The high sensitivity of MS detection makes it an important method for measuring metabolites in complex biosamples. Gas chromatography (GC) and GC-MS methods have been used for quantitative metabolic profiling. GC methods were first employed for disease diagnosis over twenty years ago (Tanaka et al. 1980). GC-MS–based metabolomic applications for the analysis of volatile and thermally stable polar and nonpolar metabolites have grown rapidly. The developments in liquid chromatography (LC)-MS and capillary-electrophoresis-MS have significantly broadened the applicability of MS-based metabolomics. Tandem MS and so-called accurate mass (time of flight) methods are most often used to validate the identities of unknown metabolites. NMR spectroscopy has been applied to the structure analysis of compounds and yields relatively low sensitivity in measurements compared to MS. Nevertheless, NMR-based metabolic profiling can be achieved successfully in many fields, because NMR is highly quantitative and reproducible. The NMR spectrum enables the simultaneous identification and monitoring of a wide range of low molecular weight endogenous metabolites, thus providing a biochemical fingerprint of an organism (Lindon et al. 2001; Lindon, Holmes, Bollard et al. 2004; Lindon, Holmes, and Nicholson 2004: Nicholson et al. 1985).

Metabolic fingerprinting is used to classify samples according to their origin or biological relevance, without aiming to determine the individual level of every metabolite (Fiehn 2002). Metabolic fingerprinting can be used to distinguish between strains of animals (Gavaghan et al. 2000; Holmes et al. 2000) and disease states (Holmes et al. 1997) or to detect pharmacological or toxicological effects obtained following dosing of compounds (Bollard et al. 2005). Recently mass spectrometry has been a major application for metabolic fingerprinting. FT-ICR MS has attracted attention as a technique for metabolic fingerprinting, since it can be used to obtain an ultra-high-resolution (>100,000) mass spectrum in about one second. With FT-ICR MS, separation of metabolites can be achieved solely by ultra-high mass resolution, eliminating the need for chromatography and derivatization. Identification of putative metabolites or classes of metabolites to which they belong can be achieved by determining the elemental composition of the metabolite based on the ultra-high mass accuracy (<1 ppm) (Brown et al. 2005; Marshall et al. 1998; Schmid et al. 2000–2001).

We have shown that urinary metabolic fingerprinting with FT-ICR MS is applicable to evaluation of compounds that may induce phospholipidosis (Hasegawa, Takenaka et al. 2007) and hepatotoxicity (Hasegawa, Ide et al. 2007). In the current study, we performed urinary metabolic fingerprinting with FT-ICR MS in an ANIT-induced rat model of intrahepatic cholestasis to investigate the relationship between histopathology and serum biochemistry and to search for possible biomarkers involved in intrahepatic cholestasis for toxicological evaluation of these compounds.

Materials and Methods

Animal Experiments

Twenty-four 6-week-old male Sprague-Dawley rats (Crl:CD[SD]IGS; Charles River Japan, Kanagawa, Japan) were used in the study. The animals were housed in metal cages in a barrier-system animal room under controlled conditions (temperature: 23°C ± 2°C, humidity: 55% ± 15%, twelve-hour light/dark cycle) and were provided with CRF-1 pellets (Oriental Yeast Co. Ltd., Tokyo, Japan) and tap water ad libitum. During the experimental period, the animals were placed in individual metabolic cages. Twelve rats were given a single oral dose of 100 mg/kg body weight of ANIT (Wako Pure Chemical, Osaka, Japan) dissolved in olive oil at day 0. The other twelve rats received the same volume of the vehicle as a control group. The animal experiments were approved by the Ethical Committee at Tanabe Seiyaku Co. Ltd., and all efforts were made to minimize animal suffering.

Sample Collection and Histopathology

Urine samples were collected on ice, predose (−31 to −24 hours) and 0–7, 7–24, 24–31, 31–48, 48–55, 55–72, and 72–96 hours after dosing; the samples were immediately frozen and stored at −20°C until use. Four rats were sacrificed under ether anesthesia on each of days 1, 2, and 4 postdose. Serum samples were obtained from blood from the abdominal aorta under ether anesthesia and used for assays of alanine aminotransferase (ALT), aspartate aminotransferase (AST), alkaline phosphatase (ALP), and total bilirubin (TBIL) using an automatic analyzer (Hitachi 7180, Tokyo, Japan). Autopsy and serum sampling were similarly performed for control animals. After euthanasia, livers were removed and immediately fixed in 10% phosphate-buffered formalin. Paraffin-embedded tissue sections cut at 5–6 μm were stained with hematoxylin-eosin (HE).

FT-ICR MS Analysis

Mass spectrometry was conducted in the negative ion mode using a FT-ICR MS equipped with a 7.0 T actively shielded superconducting magnet (IonSpec, Lake Forest, CA, USA). Ions were generated using an ESI-ion source (Analytica of Branford, Branford, CT). Samples were diluted to an appropriate concentration (typically a twenty-fold dilution) in 50% (v/v) acetonitrile, 0.1% (v/v) ammonium hydrate, and 49.9% (v/v) water. Sample injection was set to 0.5–0.8 μL/min for electrospray. The needle voltage was set to –3000 V and the gas temperature was 100°C. The ESI-generated ions were accumulated for 5.0–7.0 seconds in the hexapole. MS/MS analysis was performed using sustained off-resonance irradiation collision-induced dissociation (SORI-CID). The ion transfer conditions were optimized using a standard mix of 2,4-dichlorophenoxyacetic acid (2,4-D) (m/z218.96212), ampicillin (m/z348.10235), and reserpine (m/z607.26610). The instrument conditions were optimized for ion intensity.

Data Analysis

For FT-ICR MS measurements, we performed ten successive spectral scans for each sample analysis. Spectral preprocessing was carried out on the acquired data to reduce inter-and intra-experimental variances. For the analysis, all samples were internally calibrated for mass accuracy with the standard mix mentioned above, and we corrected the analytical errors with m/zvalues. FT-ICR MS spectral data were converted into a peak list that included the exact masses and relative intensities of each peak using Omega XP (ver. 8.0.3) software (Ion Spec, Lake Forest, CA, USA). The raw peak list was then exported to a Microsoft Excel worksheet, and the peak intensities in each mass were normalized to those of the internal standards. Thus, metabolome data from a single biological sample consisted of averaged m/zvalues with intensity information from ten pieces of spectral data. The resulting data were analyzed by principal component analysis (PCA) using R (ver. 2.0.1).

Results

Serum and Histopathological Examination

ALT and AST values both increased on day 1 and were further elevated on day 2. The TBIL level was also markedly increased on day 2 and elevated ALP was also observed, but to a lesser extent. These values all decreased on day 4 (Figure 1). Histopathologically, no noticeable hepatic changes were observed in the vehicle-treated group on days 1, 2, and 4 (Figures 2A, 2C, 2E, 2G, and 2I). In the ANIT-treated rats, bile duct degeneration and necrosis, accompanied by peribiliary infiltration and edema of neutrophils, and obstruction of bile ducts by degenerated bile duct epithelium were seen on day 1 (Figures 2B and 2D). Vacuolation and single-cell necrosis of hepatocytes were also observed. On day 2, regenerative bile ducts with edema and inflammation around portal tracts, and bile duct proliferation were observed (Figures 2F and 2H). Degeneration, vacuolation, and single-cell necrosis of hepatocytes were also seen. Bile duct proliferation was seen on day 4 (Figure 2J), and an increase in mitosis of hepatocytes was also observed.

FT-ICR MS Spectra in the Negative Ion Mode

More than 600 ion peaks were detected in the negative ion mode analysis. The FT-ICR MS spectra shown in Figure 3illustrate the changes in the endogenous components of urine observed over the study. The intensity of the ion at m/z178.05168 decreased at 0–48 hours, that at m/z192.06730 decreased at 7–31 hours, and those at m/z512.26845 and 514.28485 decreased at 7–55 hours increased compared to the predose levels.

Principal Component Analysis

To determine the degree of biological variation, PCA, an unbiased method of pattern recognition, was performed for the negative ion data sets of the urine samples. PCA determines an optimal linear transformation for a collection of data points such that the properties of the sample are most clearly displayed along the coordinate (or principal) axes. PCA is particularly useful for identification of differences between two samples by revealing the variables that contribute most to the difference and indicating whether these variables contribute in the same way (i.e., correlated) or independently from each other (i.e., uncorrelated). PCA also quantifies the amount of useful information or signal that is contained in the data. The PCA scores of the data sets presented as mean trajectory plots in each group show that principal components 1 and 2 (PC1 and PC2) provided a clear distinction between the predose and ANIT-treated data (Figure 4A). Both PCs were significant, with PC1 accounting for 88.1% of the total variance and PC2 for 9.3%, and these data allowed classification of the five groups. Hence, the biochemical composition of ANIT-treated rat urine samples at 7–24, 24–31, 31–48, and 48–55 hours differed from the predose composition, and this difference depended upon intrahepatic cholestasis following administration of ANIT. The loading plot revealed the variables (potential biomarkers) responsible for classification in the scores plot, with the further the variable from the origin, the stronger its discriminatory effect. The loading plot of these data sets showed that ions with m/z178.05168, 192.06730, 195.05139, 242.01323, 512.26845, and 514.28485 were potential biomarkers (Figure 4B).

Identification of Potential Biomarkers

To identify the potential biomarkers, an exact mass and formula composition search was performed. The ions at m/z178.05168, 192.06730, 195.05139, 242.01323, 512.26845, and 514.28485 are likely to correspond to hippurate (HA, C₉H₉NO₃, calculated [M–H]⁻=178.05097, Δm/z0.00071), phenylacetylglycine (PAG, C₁₀H₁₁NO₃, calculated [M–H]⁻=192.06662, Δm/z0.00068), [C₆H₁₁O₇]⁻(calculated [M–H]⁻=195.05103, Δm/z0.00036), [C₉H₈NO₅S]⁻(calculated [M–H]⁻=242.01287, Δ m/z0.00036), taurocholic acid (TA) with one double bond (C₂₆H₄₃NO₇S, calculated [M–H]⁻=512.26875, Δm/z–0.00030), and TA (C₂₆H₄₅NO₇S, calculated [M–H]⁻=514.28440, Δm/z0.00045), respectively. Furthermore, MS/MS analysis using SORI-CID from the product ion gave MS/MS spectra of m/z178.05168, 192.06730, 512.26845, and 514.28485 that were identical to HA, PAG, TA with one double bond, and TA, respectively, from commercial sources (data not shown). MS/MS analysis of the ion at m/z242.01323 in ANIT-treated rat urine gave an ion at m/z162.0 that resulted from loss of [SO₃]⁻(–80) (Figure 5). These data suggest that the ion at m/z242.01323 in the ANIT-treated urine was 3-methyldioxyindole sulfate (Table 1).

Discussion

Urinary metabolic fingerprinting with FT-ICR MS was performed to monitor urinary metabolic alterations in a rat model of ANIT-induced intrahepatic cholestasis. More than 600 negative ion peaks were detectable in the rat urine with FT-ICR MS. A plot of PCA scores demonstrated that the biochemical compositions of the ANIT-treated rat urine samples at 7–24, 24–31, 31–48, and 48–55 hours after dosing differed from the predose composition. The composition had almost completely returned to predose levels in samples taken 55–72 and 72–96 hours after dosing. The maximum distance from the predose position was found at 24–31 hours, but the maximal increases of biochemical markers for hepatotoxity (AST and ALT) and cholestasis (ALP and TBIL) occurred on day 2. These results indicated metabolic fingerprinting with FT-ICR MS is sensitive to detect metabolic alternation in ANIT-induced cholestasis.

In the ANIT-treated rat urine, the ion at m/z178.05168 showed decreased intensity at 0–48 hours, whereas increased intensity was observed for the ion at m/z192.06730 at 7–31 hours and for the ions at m/z512.26845 and 514.28485 at 7–55 hours. The ions at m/z178.05168, 192.06730, 195.05139, 242.01323, 512.26845, and 514.28485 were shown to be potential urinary biomarkers for intrahepatic cholestasis on the loading plot, and those at m/z178.05168, 192.06730, 242.01323, 512.26845, and 514.28485 were identified as HA, PAG, 3-methyldioxyindole sulfate, TA with one double bond, and TA, respectively.

TA and TA with one double bond are bile acids, and increases of urinary bile acids in cholestasis have been reported in several studies (Azmi et al. 2005; Beckwith-Hall et al. 1998; La et al. 2005; Lindon et al. 2003; Robertson et al. 2000; Waters et al. 2001). Cholestasis-induced bile duct cell necrosis and subsequent membrane breakdown caused by ANIT and as the result of the serum examinations may account for the elevation of urinary bile acids at 7–55 hours.

PAG has been reported to be a potential biomarker for phospholipidosis (Dieterle et al. 2006; Hasegawa et al. 2007; Espina et al. 2001; Idborg-Bjorkman et al. 2003; Nicholls et al. 2000), and alteration of PAG levels has been found in urinary metabolic fingerprinting for ANIT-induced intrahepatic cholestasis using LC-MS (La et al. 2005). PAG is the end product of phenylalanine metabolism in rodents (James et al. 1972). Further study is required to examine the relation between metabolism of PAG and the effect of ANIT.

A decrease in urinary HA levels has been detected in urinary metabolic fingerprinting for ANIT-induced intrahepatic cholestasis using NMR (Azmi et al. 2005; Robertson et al. 2000), and our results are consistent with this observation. Decreased urinary HA has been proposed as a nonspecific marker of toxicity, reflecting a complex combination of factors such as dietary intake and gut microbial metabolism (Connor et al. 2004; Nicholls et al. 2003).

3-Methyldioxyindole sulfate, which decreased at 0–24 hours and increased at 31–55 hours, is a tryptophan metabolite. These findings showed opposite reactions in the serum examinations. A decrease of an ion at m/z242 has been reported in urinary metabolic fingerprinting with LC-MS in ANIT-induced intra-hepatic cholestasis (La et al. 2005). The exact mass was not reported, but this ion may have been the same as that at m/z242.01323 observed in the current study. These results indicate that a decrease of urinary 3-methyldioxyindole sulfate might correlate with hepatotoxicity. The ion at m/z195.05139 increased at 31–55 hours. This ion may have the empirical formula of [C₆H₁₁O₇]⁻, and possible metabolites with this formula that could be detected in urine include gluconic acid, gulonic acid, and galactonic acid; mass spectrometric analysis cannot distinguish isomers with an identical empirical formula. There are no reports that the alterations of these metabolites are involved in the ANIT-induced cholestasis. Elevated levels of the m/z195.05139 ion and 3-methyldioxyindole sulfate (m/z242.01323) were observed at 31–55 hours. This period corresponds to recovery from ANIT treatment in the PCA plot (Figure 4A) and serum examinations. Therefore, the increased levels of these compounds might be associated with repair of bile ducts and hepatocytes damaged by ANIT. Identification of the specific compound corresponding to m/z195.05139 is required for a better understanding of the mechanism.

In conclusion, it is possible to monitor the progression of and recovery from drug-induced hepatotoxicity and to search for potential biomarkers involved in intrahepatic cholestasis by urinary metabolic fingerprinting with FT-ICR MS.