Metabonomic Investigations into Hydrazine Toxicity in the Rat

Chem. Res. Toxicol. 2001, 14, 975-987

975

Metabonomic Investigations into Hydrazine Toxicity in the Rat Andrew W. Nicholls,*,† Elaine Holmes,† John C. Lindon,† John P. Shockcor,† R. Duncan Farrant,‡ John N. Haselden,§ Stephen J. P. Damment,§,| Catherine J. Waterfield,§ and Jeremy K. Nicholson† Biological Chemistry, Division of Biomedical Sciences, Imperial College of Science, Technology and Medicine, Sir Alexander Fleming Building, South Kensington, London, SW7 2AZ, U.K., Physical Sciences Research Unit, Glaxo Smithkline Medicines Research Centre, Gunnels Wood Road, Stevenage, Herts, SG1 2NY, U.K., and Preclinical Safety Sciences, Glaxo Smithkline Research and Development, Park Road, Ware, Herts., SG12 0DP, U.K. Received November 1, 2000

The systemic biochemical effects of oral hydrazine administration (dosed at 75, 90, and 120 mg/kg) have been investigated in male Han Wistar rats using metabonomic analysis of 1H NMR spectra of urine and plasma, conventional clinical chemistry, and liver histopathology. Plasma samples were collected both pre- and 24 h postdose, while urine was collected predose and daily over a 7 day postdose period. 1H NMR spectra of the biofluids were analyzed visually and via pattern recognition using principal component analysis. The latter showed that there was a dose-dependent biochemical effect of hydrazine treatment on the levels of a range of low molecular weight compounds in urine and plasma, which was correlated with the severity of the hydrazine induced liver lesions. In plasma, increases in the levels of free glycine, alanine, isoleucine, valine, lysine, arginine, tyrosine, citrulline, 3-D-hydroxybutyrate, creatine, histidine, and threonine were observed. Urinary excretion of hippurate, citrate, succinate, 2-oxoglutarate, trimethylamine-N-oxide, fumarate and creatinine were decreased following hydrazine dosing, whereas taurine, creatine, threonine, N-methylnicotinic acid, tyrosine, β-alanine, citrulline, NR-acetylcitrulline and argininosuccinate excretion was increased. Moreover, the most notable effect was the appearance in urine and plasma of 2-aminoadipate, which has previously been shown to lead to neurological effects in rats. High urinary levels of 2-aminoadipate may explain the hitherto poorly understood neurological effects of hydrazine. Metabonomic analysis of highresolution 1H NMR spectra of biofluids has provided a means of monitoring the progression of toxicity and recovery, while also allowing the identification of novel biomarkers of development and regression of the lesion.

Introduction Hydrazine has been used widely as an intermediate in industrial synthetic chemistry and as a rocket propellant (1). It has also been found as a metabolite of a number of pharmacologically important compounds such as the anti-tuberculosis drug isoniazid and the antihypertensive drug hydralazine, both of which are hydrazine derivatives (2). Hydrazine itself has been commonly used as a model hepatotoxin in animal studies (3), although carcinogenicity (4), mutagenicity (4), and neurotoxicity (5) have also been observed. Amenta and Johnston were the first to report that hydrazine caused histopathological alterations in rat liver (6). These changes, evident 24 h after dosing, were characterized by a transient depletion of liver glycogen, with concomitant accumulation of fat in the periportal and mid-zonal hepatocytes. By 72 h post-hydrazine * To whom correspondence should be addressed. Phone: +44 (0)20 7594 3170. Fax: +44 (0)20 7594 3226. E-mail: [email protected]. † Biological Chemistry. ‡ Physical Sciences Research Unit. § Preclinical Safety Sciences. | Current address: Shire Pharmaceuticals, Hampshire International Business Park, Chineham, Basingstoke, Hampshire, RG24 8EP, U.K.

treatment, these effects were no longer apparent (6). More recently, it has been shown that these hepatic lesions were accompanied by an increase in plasma alanine aminotransferase (ALT)1 activity, with little alteration in plasma aspartate aminotransferase (AST) activity, and a reduction in plasma alkaline phosphatase (ALP) activity (3). Serum creatinine, urea, and albumin along with urinary protein and urinary γ-glutamyl transferase (γ-GT) were reported as being normal following hydrazine treatment (3). Urinary taurine levels have been observed to increase up to 48 h postdose, although liver taurine levels showed little or no alteration (3). In addition to its hepatic effects, hydrazine has been reported to have neurotoxic effects leading to disturbances in the central nervous system (CNS) characterized by 1 Abbreviations: TMAO, trimethylamine-N-oxide; ALP, alkaline phosphatase; ALT, alanine aminotransferase; AST, aspartate aminotransferase; γ-GT, γ-glutamyl aminotransferase; CNS, central nervous system; GABA, γ-aminobutyric acid; TSP, 3-trimethylsilyl-2,2,3,3-2H4propionic acid sodium salt; TOCSY, total correlation spectroscopy; FID, free induction decay; 2-AA, 2-aminoadipic acid; 2-AAT, 2-aminoadipate aminotransferase; PLP, pyridoxal 5′-phosphate; 2-OA, 2-oxoadipate; KAT, kynurenine aminotransferase; KYNA, kynurenic acid; KYN, kynurenine; PR, pattern recognition; PCA, principal component analysis; THOPC, 1,4,5,6-tetrahydro-6-oxo-3-pyridazine carboxylic acid; PC principal component; FT, Fourier transformation; CPMG, Carr-PurcellMeiboom-Gill; TFA, trifluoroacetic acid.

10.1021/tx000231j CCC: $20.00 © 2001 American Chemical Society Published on Web 07/03/2001

976

Chem. Res. Toxicol., Vol. 14, No. 8, 2001

seizures (5). These neurotoxic effects have been suggested to arise from alterations in concentrations of the excitatory amino acid γ-aminobutyric acid (GABA) in the brain following hydrazine dosing, although no definitive association has been made (9). Conventional biochemical methods of toxicological analysis are limited, in that only a small number of endogenous markers or key enzymes are monitored routinely. However, 1H NMR spectroscopy of biofluids allows for the simultaneous monitoring of a variety of low molecular weight endogenous metabolites from a range of intermediate metabolic pathways (10, 11). 1H NMR spectroscopy has been used to identify biomarkers indicative of specific organ damage including taurine as a specific marker of hepatotoxicity (7) and creatine as a marker of testicular toxicity (12-14). Previous 1H nuclear magnetic resonance (NMR) spectroscopic studies of rat plasma and urine following hydrazine treatment, indicated an increase in alanine and 2-aminoadipate (2-AA), a decrease of glucose between 6 and 24 h, with a concomitant increase in 3-D-hydroxybutyrate, and a dosedependent increase in lactate, creatine, and taurine (7, 8, 15). High-frequency 1H NMR spectra of biofluids (>500 MHz 1H observation frequency) typically contain several thousand resolvable lines providing structural and quantitative information on potentially hundreds of endogenous metabolites. To examine 1H NMR spectra containing such large quantities of information effectively, statistical data reduction and multivariate analysis techniques, such as principal component analysis (PCA) have been applied (16). PCA is used to calculate a new smaller set of orthogonal variables from a linear combination of a large set of correlated variables, while still maintaining the maximum level of variability from the original data. This permits the simple visualization of separation or clustering between samples, caused by compound-induced metabolic perturbations (16, 17) using two- or three-dimensional plots of the principal components (scores). The weightings (loadings) given to each variable in calculating the principal components (PCs) allow for the identification of those variables of greatest influence to the separation/clustering and hence, the deduction of biomarkers of toxicity or disease states (18). The current NMR pattern recognition (PR) approach to investigating the time-related metabolic effects of drugs and toxins in vivo is described by the term metabonomics (19). This work has used such metabonomic techniques to interpret the NMR spectroscopic data of intact urine and plasma in order to follow the biochemical variation arising from hydrazine toxicity.

Experimental Section Animal Studies. Twenty male Han Wistar rats (200-250 g) were housed individually in metabolism cages (MET 1,20 metabolism cage, Arrowmight-Biosciences, Hereford, U.K.). The animals were allowed to acclimatize for 6 days prior to dosing and placed in the metabolism cages over the urine collection periods specified below. The rats were assigned to four dose groups, viz., control, 75, 90, and 120 mg/kg hydrazine and each animal was administered either hydrazine in sterile water or the dosing vehicle via oral gavage. Urine samples were collected predose (-8 to 0 h), 0-8 h, 8-16 h, 16-24 h, 24-32 h, 48-56 h, 72-80 h, 96-104 h, 120-128 h, and 144-152 h postdose. Urine samples were frozen at -20 °C prior to NMR spectroscopic analysis. Blood samples (0.5 mL) were taken from all animals

Nicholls et al. prior to dosing and at 24 h postdose, into heparin containers. Plasma was separated by centrifugation and retained frozen at -20 °C until analysis by NMR spectroscopy. Prior to autopsy, a blood sample (1 mL) was collected by tail venepuncture into EDTA anticoagulant for subsequent determination of haematological parameters. Animals were killed by exsanguination under isoflurane anaesthesia, 7 days after dosing. As part of the exsanguination procedure, two blood samples (0.6 mL) were retained for measurement of plasma clinical chemistry parameters. The liver, kidneys, and any tissues appearing abnormal at autopsy were fixed in 10% (v/v) neutral buffered formalin, processed to 3 µm wax sections, and were subsequently stained with haematoxylin and eosin for histopathological examination. Clinical Pathology. Standard spectrophotometric methods on a Hitachi 917 analyzer were used for the measurement of the following plasma parameters: ALP, ALT, AST, albumin, bile acids, cholesterol, glucose, R-glutathione-S-transferase, glutamate dehydrogenase, sorbitol dehydrogenase, total protein, and triglycerides. Haematological parameters, i.e., haemoglobin concentration, haematocrit, erythrocyte count, mean cell volume and concentration, total and differential leucocyte count were also measured. Preparation of Samples for 1H NMR Spectroscopy. The urine samples for NMR spectroscopy were made up from 400 µL of urine and 200 µL of buffer solution (0.2 M Na2HPO4/0.2 M NaH2PO4, pH 7.4) which were mixed in a micro-container and the resulting solution left to stand for 10 min. The buffered urine samples were then centrifuged at 13 000 rpm for 10 min to remove any precipitates, and aliquots of the resulting supernatant (500 µL) were placed in 5 mm NMR tubes to which 50 µL of a solution of sodium 3-trimethylsilyl-(2,2,3,3-2H4)-1propionate (TSP) in D2O was added (final concentration ) 1 mM). The D2O plus TSP addition provided both a chemical shift reference (δ 0.0) and a deuterium lock signal for the NMR spectrometer. Plasma samples (100 µL) were placed in 5 mm NMR tubes together with 400 µL of D2O (to provide a field/ frequency lock). Chemical shifts in the plasma spectra were referenced to the lactate doublet at δ 1.33. NMR Spectroscopic Analysis of Plasma and Urine Samples. 1H NMR spectra of the plasma samples were obtained at 500.13 MHz using a Bruker DRX-500 spectrometer. Onedimensional spectra were acquired using a standard presaturation pulse sequence for water suppression (11) with irradiation at the water frequency during the relaxation delay of 3 s and the pulse sequence mixing time of 100 ms. Following four dummy scans, spectra were acquired using 64 scans into 64K points with a spectral width of 10 080 Hz, an acquisition time of 3.25 s and a total pulse recycle time of 6.35 s. The free induction decays (FIDs) were multiplied by an exponential function corresponding to a 0.3 Hz line broadening prior to Fourier transformation (FT). Water suppressed Carr-PurcellMeiboom-Gill (CPMG) spin-echo spectra were acquired with 128 scans after four dummy scans into 64K data points with a total spin-spin relaxation delay (2nτ) of 88 ms and a total pulse recycle delay of 6.25 s. The FIDs were multiplied by an exponential function corresponding to a 0.25 Hz line broadening prior to FT. 1H-1H-total correlation spectroscopy (TOCSY) NMR spectra were acquired for selected samples using 64 scans into 4K data points per increment for 128 increments, with F1 and F2 spectral widths of 7002.8 Hz. The FIDs were multiplied by a shifted sine-bell squared function in both dimensions prior to FT. One-dimensional 1H NMR spectra of urine were measured at 600.13 MHz on a Bruker DRX-600 spectrometer using the same water suppression method (11). Sixty-four FIDs were collected into 64K data points using a spectral width of 7003 Hz, an acquisition time of 4.68 s, and a total pulse recycle time of 7.68 s. The FIDs were multiplied by an exponential weighting function corresponding to a line broadening of 0.3 Hz prior to FT. 1H-1H-TOCSY spectra were acquired for selected samples using 64 scans into 4K data points per increment for 256 increments, with F1 and F2 spectral widths of 7002.8 Hz. The

Metabonomic Studies of Hydrazine Toxicity

Chem. Res. Toxicol., Vol. 14, No. 8, 2001 977

Table 1. Statistically Significant (p < 0.05) Plasma Clinical Chemistry Measurements Following Dosing of Hydrazine at 90 mg/kg groups

white blood cells (109/L)

neutrophils (109/L)

lymphocytes (109/L)

monocytes (109/L)

glucose (mM)

triglycerides (mM)

control (n ) 5) dosed (n ) 4)

7.0 ( 0.9 10.8 ( 1.5

0.9 ( 0.4 2.3 ( 0.4

5.7 ( 0.8 7.9 ( 1.2

0.2 ( 0.05 0.32 ( 0.1

11.2 ( 0.9 9.7 ( 0.4

1.7 ( 0.8 0.8 ( 0.4

FIDs were multiplied by a shifted sine-bell squared function in both dimensions prior to FT. Directly Coupled HPLC NMR Spectroscopic Analysis of Whole Urine. The HPLC system comprised a HewlettPackard 1050 Series pump and a variable-wavelength UV detector (operating at 240 nm). The outlet from the UV detector was connected to the HPLC NMR flow probe via an inert polyether(ether) ketone capillary. The chromatography was controlled using the Bruker Chromstar HPLC data system. Analysis was performed on a YMC ODS-AQ C18 column (4.6 × 250 mm i.d.) packed with ODS, 5 µm. The mobile phase consisted of D2O plus trifluoroacetic acid (TFA) at pH 2:acetonitrile-d3 decreasing from 95.0% D2O at the beginning of the run to 60% D2O after 30 min. The use of acetonitrile-d3 precludes the need for double peak suppression and allowed observation of the resonances from acetylated molecules. The flow rate was 0.5 mL/min. HPLC NMR spectra were acquired on selected samples using a Bruker DMX-500 spectrometer equipped with a 1H flow probe (cell of 3 mm i.d. with a volume of 120 µL). 1H NMR spectra were obtained in the on-flow mode at 500.13 MHz using a standard presaturation pulse sequence for water suppression (11) with irradiation at the water frequency during the relaxation delay of 3 s and the pulse sequence mixing time of 80 ms. FIDs were collected into a pseudo-2D array consisting of 4K data points in each of 174 rows. Sixteen scans per row were acquired using an acquisition time of 0.41 s, a recycle delay of 1.1 s and a spectral width of 5000 Hz. An exponential line-broadening function corresponding to 0.3 Hz was applied prior to FT and these data were zero-filled by a factor of 2. Pattern Recognition Analysis of NMR Spectral Data of Urine and Plasma. All NMR spectra were phased and baseline corrected using XWINNMR (Bruker Analytik, Rheinstetten, Germany) and data-reduced using AMIX (Bruker Analytik, Rheinstetten, Germany) to integrated regions 0.04 ppm wide corresponding to the region δ 10.0 to 0.2. The region δ 6.0-4.5 in the urine spectra was set to zero integral value for the purposes of pattern recognition analysis to remove the variability in presaturation of the water resonance and crossrelaxation effects on the urea signal. For the NMR spectra of plasma, the regions δ 5.1-4.4 were set to zero integral to remove the variation in presaturation of the water resonance. For the urine data, the region δ 2.24-2.08 was excluded from analysis due to interference from resonances of acetylhydrazine and diacetylhydrazine while regions δ 2.86-2.82 and δ 2.58-2.54 were excluded due to the presence of peaks from the hydrazine metabolite 1,4,5,6-tetrahydro-6-oxo-3-pyridazine carboxylic acid. Each data point was normalized to the sum of its row and all variables were mean centered prior to PCA using Pirouette (version 2.7, Infometrix Inc., Woodinville, WA 98072-1528). Scores plots of the principal components were constructed to visualize any inherent separation of the dose groups, and from the values of the PC loadings (which indicated the importance of each variable to the separation) the NMR spectral regions and hence biomarkers were identified. Mean data were calculated for each integrated region for each compound and time period, and analyzed using PCA. Plots of the first two principal components (PC1 and PC2) were mapped to represent a trajectory plot describing the metabolic perturbations with time following dosing.

Results Clinical Observations, Body Weight, and Mortality. No clinical observations were noted for the control

Figure 1. Graph of the % body weight changes following administration of hydrazine for each of the dose groups. ([) Control, (9) 75 mg/kg, (2) 90 mg/kg, (b) 120 mg/kg.

animals and those dosed with hydrazine at 75 mg/kg for the course of the study. Hunched posture, piloerection and subdued behavior were observed in one of the 90 mg/ kg group at 72 h postdose and in all animals from the 120 mg/kg group between 48 and 72 h. There was a doserelated reduction in group mean body weight following dosing with hydrazine (Figure 1). Animals treated at a dose of 90 mg/kg lost 9% of body weight up to 72 h after dosing. Thereafter, all hydrazine-treated animals gained weight, although the mean body weight for the 90 mg/ kg group remained statistically lower (p < 0.05) than the control group up to 144 h after dosing. For the 120 mg/ kg dose group these changes were observed by 48h and, because of the severity of the signs, coupled with an excessive weight loss, the animals were killed for humane reasons at 72 h. One of the animals from the mid-dose group (90 mg/kg) was also noted to show similar body weight change to the high-dose group by 72 h and this animal was also killed at this time point. Clinical Pathology. Shown in Table 1 are the statistically significant (p < 0.05) alterations noted by clinical pathology. Although these alterations were noted to show significant variation following dosing, no other data were observed to suggest a toxicologically significant effect. Histopathology. Macroscopic examination of animals killed in extremis showed that all had pale livers. Microscopic examination (Figure 2, panels A-C) showed varying amounts of midzonal hepatic fat vacuolation (Figure 2B), and in some animals this involved the periportal areas. No obvious differences were seen in the liver of the animal that had received 90 mg/kg hydrazine dose compared to those that received 120 mg/kg. Some animals were noted to show multifocal inflammation which, in one instance, was associated with single cell necrosis (Figure 2C). One of the animals was also noted to have shown localized necrosis with inflammatory cell infiltration. Marked mitotic activity was also observed in one animal from the 120 mg/kg group. No macroscopic or microscopic abnormalities were observed for any of the animals killed 7 days after dosing.

978


Figure 2. (A) Control liver showing central vein and a minor portal region, (B) liver following a 90 mg/kg dose of hydrazine showing midzonal accumulation of lipid droplets, (C) liver from an animal administered a 120 mg/kg dose of hydrazine showing lipid accumulation along with necrosis and macrophage infiltration in the midzonal/centrilobular region.

However, results of a preliminary study (unpublished data) showed that doses of 75 or 90 mg/kg consistently produced fat vacuolation within 48 h of dosing. The lack of any hepatic abnormalities in animals killed at 7 days indicated that the fatty changes were reversible. NMR Spectroscopy of Rat Urine Following Dosing of Hydrazine. Shown in Figure 3 are the 600 MHz 1H NMR spectra of whole rat urine predose (-8 to 0 h) and 0-8 h, 24-32 h, and 144-152 h following administration of hydrazine at 90 mg/kg. The spectral data indicated that substantial alterations in the urinary

Nicholls et al.

metabolite composition had occurred, consistent with a toxicological insult, particularly in the 24-32 h time period. A number of resonances assigned to known metabolites of hydrazine were observed in the 1H NMR spectra. These included acetylhydrazine (δ 2.07, s), diacetylhydrazine (δ 2.03, s), and 1,4,5,6-tetrahydro-6-oxo-3-pyridazine carboxylic acid (THOPC, δ 2.83, t; δ 2.55, t). The resonances for the latter were partially overlapped by the citrate resonances in the 1D NMR spectra and were assigned from the 1H-1H-TOCSY NMR spectrum (Figure 4) and literature data (20). No other metabolites of hydrazine were noted in the NMR spectra. Following administration of hydrazine, a decrease of the intensity of the resonances for hippurate, citrate, succinate, 2-oxoglutarate, trimethylamine-N-oxide (TMAO), fumarate, and creatinine was observed in the NMR spectra of whole urine (NMR assignments are given in Table 2). Conversely, the intensity of the resonances of taurine, creatine, threonine, N-methylnicotinic acid, and β-alanine increased in the urine samples following dosing of hydrazine. These changes to the intensities were identified from comparison with the NMR spectra of urine samples from the control group and predose. Examination of the 1H NMR spectra for the 0-8 h postdose urine samples (not shown) indicated that the 2-oxoglutarate resonances decreased prior to the loss of citrate and succinate. This suggested that the loss of 2-oxoglutarate was due to metabolic routes other than the tricarboxylic acid cycle. Also noted to increase were a number of other resonances, with the largest increase from those at δ 3.77, 2.26, and the region 2.0-1.5 and these were assigned to 2-aminoadipate by comparison to a synthetic standard and literature data (15). From Figure 4, the resonances at δ 3.76, 3.15, 1.88, and 1.55 were assigned to citrulline and the signals at δ 4.27, 3.78, 3.30, 2.88, 2.70, 1.92, and 1.72 were consistent with argininosuccinate based on literature data (21). The resonances at δ 4.13, 3.11, 1.80, 1.66, and 1.52 were proposed to arise from a metabolite of citrulline. To accurately identify this metabolite, directly coupled HPLC NMR spectroscopy was applied to separate the components and to obtain NMR spectra of the relevant substance. HPLC NMR Spectroscopy of Hydrazine-Dosed Rat Urine. The on-flow 500 MHz 1H HPLC NMR chromatogram of whole rat urine obtained 56 h postdosing from a rat dosed with hydrazine at 120 mg/kg is shown in Figure 5A. Four rows corresponding to major observed resonances have been extracted and the 1H HPLC NMR spectra are shown in Figure 5, panels B-E. The corresponding retention times for these spectra were 6.4, 6.9, 10.7, and 13.3 min (Figure 5, panels B-E, respectively). The resonances at δ 1.62, 1.84, 2.37, and 3.61 indicated in Figure 5B were assigned to 2-AA. However, signals from several other coeluting polar species, e.g., taurine, were also observed. Figure 5C shows resonances at δ 1.48, 1.79, 2.36, 3.04, and 3.68 from a structurally similar, but as yet unidentified endogenous metabolite. Figure 5D shows the NMR spectrum of a third endogenous metabolite, creatinine. Figure 5E shows a set of coupled proton resonances at δ 1.56, 1.74, 1.88, 2.03, 3.12, and 4.33. From comparison of these chemical shifts with those observed in Figure 4 and literature data (21) the metabolite was identified as NR-acetylcitrulline.



Figure 3. Stack plot of the 600 MHz 1H NMR spectra of whole rat urine from an animal administered hydrazine at 90 mg/kg. (A) Predose (-8 to 0 h), (B) 0-8h postdose, (C) 24-32 h postdose, (D) 144-152 h postdose. Endogenous and hydrazine metabolites were assigned from literature sources and were as indicated. TMAO ) trimethylamine-N-oxide, THOPC ) 1,4,5,6-tetrahydro-6-oxo-3pyridazine carboxylic acid, DMA ) dimethylamine.

Metabonomic Analysis of Whole Rat Urine. Shown in Figure 6 are the mean scores plots for the first two PCs of the data-reduced 1H NMR spectra of whole urine following dosing of hydrazine. All plots are shown on the same scale and indicate that the three hydrazine groups showed a dose-dependent shift to the left. Both the 75 and 90 mg/kg dose groups were furthest separated from the control region at the 24-32 h time period. Subsequently the two groups recovered back to the control region of the plot indicating a metabolic recovery from hydrazine dosing. The 120 mg/kg group sample points were still moving away from the control region in PC space at 56 h at which point the animals were sacrificed as described earlier. This would suggest that an irreversible toxicity threshold had been crossed at this dose level by the continuation of the metabolic variation. The

loadings plot (Figure 7) for PC1 versus PC2 indicated those regions of the NMR spectra that contributed the most to the position of the samples in Figure 6. From analysis of this plot the biomarkers which were most influential to the separation of the dosed groups from the control were identified. The spectral regions shown to the left of the loadings plot were increased in the hydrazine dosed animals while those to the right of the plot were decreased. From this, it was observed that following hydrazine dosing there was a decrease in the levels of the Krebs cycle intermediates citrate (δ 2.74, 2.70), 2-oxoglutarate (δ 3.02, 2.46), and succinate (δ 2.42) along with a decrease in hippurate (δ 7.86, 7.82, 7.66, 7.62, 7.58, 7.54, 3.98). The actual chemical shifts are given in Table 2. The loadings plot also suggested that the 0.04 ppm region centered at δ 6.06 was important to the separation

980


Nicholls et al.

Figure 4. 1H-1H-TOCSY NMR spectra of whole urine from an animal 24-32 h after administration of hydrazine at 120 mg/kg. The corresponding 1D 1H NMR spectrum is shown at the top. Endogenous metabolites were assigned from literature sources. THOPC ) 1,4,5,6-tetrahydro-6-oxo-3-pyridazine carboxylic acid

in PC1, but on examination of the NMR data this was found to be an error caused by incomplete exclusion of the urea signal in some spectra. Those regions of the NMR spectra noted to increase upon dosing were attributed to taurine (δ 3.42, 3.26), citrulline (δ 3.78-3.74, 3.18-3.14, 1.90, 1.54), NR-acetylcitrulline (δ 3.10, 1.821.78, 1.54-1.50), argininosuccinate (δ 4.26, 3.78, 3.30, 1.94-1.90, 1.74-1.70), and 2-aminoadipate (δ 3.74, 2.26, 1.90, 1.66). NMR Spectroscopy of Rat Plasma Following Dosing of Hydrazine. The 500 MHz 1H NMR spectra of typical pre- and postdose plasma samples (hydrazine, 120 mg/kg) are shown in Figure 8, panels A and B, respectively. The endogenous metabolites were as indicated and were assigned according to previous studies

(11, 22). The broad resonances that are observed in Figure 8 are mainly due to plasma lipoproteins and albumin, on top of which are superimposed the sharp resonances arising from small motionally unconstrained molecules. The NMR spectra showed a number of alterations to the endogenous plasma profile following dosing with hydrazine. These changes were observed from comparison of the pre- and postdose plasma samples for each animal. The broad macromolecular resonance profile indicated a reduction in the plasma lipid level after dosing. This was most apparent in the region of the spectra between δ 1.25-1.3 and 0.85-0.90, i.e., those regions corresponding to the long chain CH2 groups of fatty acids and terminal CH3 groups, respectively. The signals from a variety of low MW compounds were



Table 2. 1H Chemical Shifts and Assignments for the Endogenous Urinary Metabolites Responsible for the Separation of Samples from Control and Hydrazine Dosed Animalsa 1H

endogenous metabolite hippurate citrate succinate 2-oxoglutarate TMAO fumarate creatinine taurine creatine threonine N-methylnicotinic acid β-alanine lactate 2-aminoadipic acid citrulline NR-acetyl-citrulline argininosuccinate

chemical shifts (δ), multiplicity, and assignment 7.73 (d, H2 and H6), 7.64 (t, H4), 7.55 (t, H3 and H5), 3.97 (d, CH2) 2.72 (d, 1/2-CH2), 2.56 (d, 1/2-CH2) 2.42 (s, CH2) 3.01 (t, CH2), 2.45 (t, CH2) 3.26 (s, CH3) 6.53 (s, CH) 4.05 (s, CH2), 3.05 (s, CH3) 3.43 (t, CH2), 3.26 (t, CH2) 3.94 (s, CH2), 3.04 (s, CH3) 1.34 (d, CH3) 9.28 (s, H2), 8.97 (d, H6), 8.90 (d, H4), 8.19 (t, H5) 3.20 (t, CH2), 2.56 (t, CH2) 4.11 (q, CH), 1.33 (d, CH3) 3.77 (t, RCH), 2.26 (t, δCH2), 1.90 (m, βCH2), 1.66 (m, γCH2) 3.76 (t, CH), 3.15 (qu, γCH2), 1.88 (m, βCH2), 1.55 (m, δCH2) 4.13 (m, CH), 3.11 (t, γCH2), 2.03 (s, CH3), 1.80 (m, 1/2-βCH2), 1.66 (m, 1/2- βCH2), 1.52 (m, δCH2) 4.27 (dd, CH), 3.78 (t, CH), 3.30 (t, CH2), 2.88 (ABX, CH2), 2.70, (ABX, CH2), 1.92 (m, CH2), 1.72 (m, CH2)

increase (v) or decrease (V) following hydrazine exposure V V V V V V V v v v v v v v v v v

a Increase or decrease based on loadings for NMR spectral regions. s ) singlet, d ) doublet, t ) triplet, q ) quartet, qu ) quintet, m ) multiplet.

Figure 5. 500 MHz 1H NMR spectra, of endogenous metabolites separated from whole rat urine 56 h postdose hydrazine (120 mg/kg) using on-flow HPLC NMR spectroscopy: (A) on-flow HPLC NMR pseudo-2D NMR spectrum, (B) 2-aminoadipic acid and unidentified coeluting metabolites, (C) unidentified urinary component, (D) creatinine, (E) NR-acetyl-citrulline.

observed to increase; these included glycine, alanine, 3-Dhydroxybutyrate, creatine, valine, isoleucine, histidine, threonine, and tyrosine (chemical shift values are given in Table 3). The aromatic resonances of tyrosine at δ 7.18 and 6.88 could be seen clearly and the level of tyrosine appeared to follow an apparent dose-dependent increase in plasma (data not shown). Figure 9 shows the 1H-1H-

TOCSY NMR spectra of whole plasma predose and following dosing with hydrazine at 120 mg/kg. Unlike the 1D NMR spectra these allowed the identification of a number of species in plasma including 2-aminoadipate, which was only present postdose. Metabonomic Analysis of Whole Rat Plasma. The scores plot of PC1 versus PC2 for predose and 24 h

982


Nicholls et al.

Figure 6. Scores plots of PC1 versus PC2 of the mean data from NMR-based metabonomic analysis of urinary data following dosing of hydrazine for each of the four groups. The % variance of the original data explained in PC1 and PC2 was 68%. Key: C ) control, HZ-L ) low dose (75 mg/kg), HZ-M ) medium dose (90 mg/kg), HZ-H ) high-dose (120 mg/kg). P ) Predose, 8 ) 0-8 h, 16 ) 8-16 h, 24 ) 16-24 h, 32 ) 24-32 h, 56 ) 48-56 h, 80 ) 72-80 h, 104 ) 96-104 h, 128 ) 120-128 h, 152 ) 144-152 h.

postdose samples of plasma following administration of hydrazine is shown in Figure 10A. It was apparent that a dose dependent right-to-left shift could be observed in samples from animals administered with hydrazine. From comparison with the loadings plots of PC1 and PC2 (Figure 10, panels B and C), this separation was attributed to a reduction in the broad resonances from lipoproteins, along with increases in alanine, citrulline, creatine, and tyrosine. An increase in plasma 2-aminoadipate may also have led to the separation in PC1, but this requires further verification. The plasma markers of hydrazine toxicity are tabulated with their chemical shifts in Table 3.

Discussion Recent studies have shown that NMR-based metabonomic methods can provide a rapid analytical tool for the study of biochemical variation in biofluids (15-19). This variation can result from numerous sources, but many studies have principally concentrated on the application to drug toxicity evaluation. The current study has used these methods to show that the dose-dependent variability in urine and plasma following hydrazine toxicity can be visualized easily. Although direct comparison of the NMR spectra themselves can be made, the use of pattern recognition methods to visual the data as metabonomic trajectory plots enables a rapid determination of the state of the animal metabolism following dosing. The regions of the NMR spectra key to the variability following dosing can be identified from the PC loadings. By comparison of these loadings to the trajectory plot the

relative variability in the NMR regions can be determined enabling rapid identification of the metabolic components of interest. In the present work, we have shown a number of substantial changes to the urine of rats experiencing hydrazine-induced hepatotoxicity. Elevation of 2-AA following dosing of hydrazine has been reported previously from both studies involving chronic dosing over 109 days (9) and acute dosing at 90 mg/kg (15). In the latter study, it was identified as a biomarker of hydrazine toxicity up to 48 h and used to develop a model of the metabonomic profile of hydrazine toxicity. In the current study the appearance of 2-AA has been noted up to 48 h, but has also been followed until 7 days after dosing via the use of the metabonomic trajectory plots. This allows for the simple observation of the onset, progression and recovery of the dynamic metabolic perturbations following dosing. This study has also indicated that there was an apparent dose-dependent increase in 2-AA, the largest contributor to separation of the dose groups in Figure 6. 2-AA arises from the breakdown of lysine (23) and the high concentrations excreted here suggests an inhibitory effect on 2-aminoadipate aminotransferase (2-AAT, EC 2.6.1.39), the enzyme responsible for the catalysis of the reversible transamination of 2-AA and 2-oxoglutarate to form 2-oxoadipate and glutamate. Critical to the function of aminotransferases, as well as decarboxylase enzymes such as those involved in the formation of GABA and serotonin, is the cofactor pyridoxal 5′-phosphate (PLP) (23, 24). Hydrazine has been shown to lead to the sequestering of PLP to form pyridoxal phosphate-hydra-



Figure 7. (A) Loadings plot for PC1 versus PC2 from the NMR-based metabonomic analysis of urinary data. Expansions of panel A are shown in panels B and C to better visualize the regions influencing the separation in Figure 6.

zone (25). Although the resulting PLP deficiency is likely to be the cause of the GABA fluctuations previously

reported (9), a reduction in general aminotransferase activity may also occur due to the inability to form the

984


Nicholls et al.

Figure 8. 500 MHz 1H NMR spectra of whole rat plasma predose (A) and postdose (B) from an animal administered hydrazine 120 mg/kg. Endogenous metabolites were assigned from literature sources and were as indicated. Key: VLDL ) very low-density lipoprotein, LDL ) low-density lipoprotein, HDL ) high-density lipoprotein. Table 3. 1H Chemical Shifts and Assignments for the Endogenous Plasma Metabolites Observed to Have the Greatest Influence in Separation of Control and Hydrazine Dosed Samples Based on PC Loadings for NMR Spectral Regions 1H

endogenous metabolites lipoproteins glycine alanine 3-D-hydroxybutyrate creatine valine isoleucine histidine threonine tyrosine citrulline 2-aminoadipic acid a

chemical shifts (δ)

5.32, (-CHdCH-), 2.78 (dHC-CH2-CHd), 2.05 (dHC-CH2-), 1.59 (-CH2-CH2-CO-), 1.30 (-(CH2)n-), 1.25 (-CH3), 0.90 (-CH3), 0.85 (-CH3), 0.69 (-CH3, HDL)a 3.54 (CH2) 1.47 (CH3) 4.19 (CH), 2.36 (CH2), 1.20 (CH3) 3.94 (CH2), 3.04 (CH3) 3.56 (CH), 2.26 (CH), 1.04 (CH3), 0.98 (CH3) 1.27 (CH2), 1.01 (CH3), 0.91 (CH3) 7.73 (H2), 7.02 (H4) 4.21 (CH), 1.31 (CH3) 7.18 (H2 and H6), 6.88 (H3 and H5), 3.49 (1/2-CH2), 3.25 (1/2-CH2) 3.73 (CH), 3.15 (γCH2), 1.87 (βCH2), 1.57 (δCH2) 3.73 (CH), 2.25 (CH2), 1.87 (CH2), 1.64 (CH2)

High-density lipoprotein.

enzyme-PLP Schiff’s base necessary for transamination (26). The sequestration of PLP may lead to an upregulation in the conversion of pyridoxamine 5′-phosphate to PLP via pyridoxamine phosphate aminotransferase. This process is dependent on the conversion of 2-oxoglutarate to glutamate (26) and may explain the initial loss of 2-oxoglutarate from the 0-8 h 1H NMR spectra of whole rat urine. Furthermore, previous studies have shown that hydrazine directly conjugates with 2-oxoglutarate to form the metabolite THOPC (8). Therefore, hydrazine appears to reduce the activity of aminotransferases by removal of two required cofactors, which may explain the elevation of 2-AA observed in this study. One of the greatest concentrations of 2-aminoadipate in the rat occurs in the kidney and this may explain the

large urinary concentration observed in this study. However, the 1H-1H TOCSY NMR spectra (Figure 9) indicated the presence of 2-aminoadipate in whole plasma following hydrazine administration. Therefore, the effects of 2-AA may also need to be considered with regard to hydrazine toxicity. 2-AA has been extensively reported as a glial toxin leading to limbic seizures and convulsions (24, 27). Elevated levels of 2-AA in the hippocampus have been reported to cause a reduction in kynurenic acid (KYNA) concentration (28). KYNA is a broad-spectrum excitatory amino acid receptor antagonist with a high affinity for the glycine co-agonist site of the N-methyl-D-aspartate receptor and has been implicated as possessing neuroprotective and seizure reducing activity (28).



Figure 9. 500 MHz 1H-1H-TOCSY NMR spectra of predose and 24 h postdose hydrazine (120 mg/kg) rat plasma showing the presence of 2-aminoadipate and citrulline. The corresponding 1D 1H NMR spectra are shown above. Endogenous metabolites were assigned from literature sources. Key: VLDL ) very low-density lipoprotein, LDL ) low-density lipoprotein, HDL ) high-density lipoprotein.

Studies have attributed the formation of KYNA from kynurenine (KYN) to the enzymes kynurenine aminotransferase I (KAT I) and kynurenine aminotransferase II (KAT II). KAT I has been shown to be identical to glutamine transaminase K (29, 30), while KAT II has been shown to be identical to 2-AAT (31). Formation of KYNA in the forebrain occurs predominantly via KAT II (32), and thus competition between KYN and 2-AA for KAT II may lead to the observed reduction in KYNA levels and hence an increase in neurological disturbance. The high levels of 2-AA excreted in the urine may also be explained from substrate saturation of KAT II. The conversion of 2-AA to 2-oxoadipate (2-OA) is a reversible process (33), and since KYN is converted to 2-OA via tryptophan degradation, this pathway may also contribute to the formation of 2-AA. Citrulline was observed in both plasma and urine samples from animals which had received a dose of hydrazine. Furthermore, NR-acetylcitrulline and argininosuccinate were also observed in the urine samples indicating that hydrazine had either directly or indirectly affected the urea cycle. Previous studies where hydrazine was dosed over a 4-day period at 35 mg/kg reported the presence of elevated citrulline in the blood (34). This study also reported that the activities of ornithine carbamoyl transferase (EC 2.1.3.3) and argininosuccinate synthase (EC 6.3.4.5) were not affected by hydrazine dosing, but that the activity of argininosuccinase (EC 4.3.2.1) was noted to increase. The increase in urinary argininosuccinate in this study would suggest that the enzyme had been affected either directly or indirectly by

the larger, single dose of hydrazine. Further work is required to identify the mechanism for this apparent inhibitory effect on argininosuccinase. One of the most notable alterations observed in 1H NMR spectra of plasma following dosing of hydrazine was an increase in the level of the free amino acid tyrosine (Figure 8). This increase occurred at all dose levels and displayed a degree of dose dependency. Tyrosinaemia has not been previously noted as a marker of hydrazine hepatotoxicity, but has been observed to occur in previous in vitro studies of hydrazine toxicity of erythrocytes and was suggested to result from haemoglobin proteolysis in red blood cells (35). However, other than the appearance of tyrosine in the plasma, no other molecular species were noted to corroborate the existence of such proteolysis, and a further more specific study is required. The biomarkers noted in previous NMR spectroscopic studies (8) of urine and plasma from hydrazine-treated rats, i.e., increases in urinary alanine, β-alanine, methylamine, an increase in taurine and 2-AA with a decreases in 2-oxoglutarate and citrate, were also observed here. The elevations in amino acids may have resulted from reduced function of the aminotransferases arising from the lack of pyridoxyl phosphate. Furthermore, the subsequent depletion of a carbon source may explain the reduction in the 2-oxoglutrate and citrate. Urinary levels of creatine and taurine have previously been reported to increase following hydrazine dosing (15) and this was confirmed by the current work. Increases in urine levels of both of these components have been previously associated with reduced liver function (7, 15). Whether the rest

986


Nicholls et al.

Figure 10. (A) Scores plot of PC1 versus PC2 of data from NMR spectra of plasma from control and hydrazine dosed rats. The % variance of the original data explained in PC1 and PC2 was 89%. (B) Plot of PC loadings for PC1 and PC2. (C) Expansion of the PC loadings plot in panel B to show the regions of the NMR spectra of increased intensity following hydrazine dosing. Prefix: C ) control, L ) low dose (75 mg/kg), M ) medium dose (90 mg/kg), H ) high dose (120 mg/kg). Suffix: P ) predose, 24 ) 24 h postdose.

of the biomarkers are indicative of hepatotoxicity is currently under further investigation.

Acknowledgment. We thank Glaxo SmithKline Research and Development for the provision of resources


for this work and for financial support (to A.W.N.). We acknowledge the ULIRS Biomedical NMR service at Birkbeck College, University of London for access to the 500 MHz NMR spectrometer.


(18)

References (1) Budavari, S. Merck Index, New Jersey: Merck and Co.; 1996. (2) Blair, I. A., Mansilla-Tinoco, R., Brodie, M. J., Clare, R. A., Dollery, C. T., Timbrell, J. A., and Beaver, I. A. (1985) Plasma hydrazine concentrations in man after isoniazid and hydralazine administration. Hum. Toxicol. 4, 195-202. (3) Waterfield, C. J., Turton, J. A., Scales, M. D. C., and Timbrell, J. A. (1993) Investigations into the effects of various hepatotoxic compounds on urinary and liver taurine levels in rats. Arch. Toxicol. 67, 244-254. (4) Sax, N. I. (1990) Chemical Review: Hydrazine. Dangerous Prop. Ind. Mater. Rep. 10, 21-58. (5) Moloney, S. J., Prough, R. A. (1983) Biochemical toxicology of hydrazines. Rev. Biochem. Toxicol. 5, 313-346. (6) Amenta, J. S., and Johnston, E. H. (1962) Hydrazine-induced alterations in rat liver. A correlation of the chemical and histologic changes in acute hydrazine intoxication. Lab. Invest. 11, 956962. (7) Sanins, S. M., Nicholson, J. K., Elcombe, C., and Timbrell, J. A. (1990) Hepatotoxin-induced hypertaurinuria: a proton NMR study. Arch. Toxicol. 64, 407-411. (8) Sanins, S. M., Timbrell, J. A., Elcombe, C., and Nicholson, J. K. (1992) Proton NMR spectroscopic studies on the metabolism and biochemical effects of hydrazine in vivo. Arch. Toxicol. 66, 489495. (9) Perry, T. L., Kish, S. J., Hansen, S., Wright, J. M., Wall, R. A., Dunn, W. L., and Beilward, G. D. (1981) Elevation of brain GABA content by chronic low-dosage administration of hydrazine, a metabolite of isoniazid. J. Neurochem. 37, 32-39. (10) Nicholson, J. K., and Wilson, I. D. (1987) High-resolution nuclear magnetic resonance spectroscopy of biological samples as an aid to drug development. Prog. Drug Res. 31, 427-429. (11) Nicholson, J. K., Foxall, P. J. D., Spraul, M., Farrant, R. D., and Lindon, J. C. (1995) 750 MHz 1H and 1H-13C NMR spectroscopy of human blood plasma. Anal. Chem. 67, 793-811. (12) Gray, J. A., Nicholson, J. K., and Timbrell, J. A. (1986) Creatinuria as an early indicator of cadmium induced testicular damage Hum. Toxicol. 5, 402-403. (13) Nicholson, J. K., Higham, D. P., Timbrell, J. A., and Sadler, P. J., (1989) Quantitative high-resolution 1H NMR urinalysis studies on the biochemical effects of cadmium in the rat. Mol. Pharmacol. 36, 398-404. (14) Gray, J. A., Nicholson, J. K., Creasy, D. M., and Timbrell, J. A., (1990) Studies on the relationship between testicular toxicity and urinary and plasma creatine concentration. Arch. Toxicol. 64, 443-450. (15) Holmes, E., Nicholls, A. W., Lindon, J. C., Connor, S. C., Connelly, J. C., Haselden, J. N., Damment, S. J. P., Spraul, M., Neidig, P., and Nicholson, J. K. (2000) Chemometric models for toxicity classification based on NMR spectra of biofluids. Chem. Res. Toxicol. 13, 471-478. (16) Holmes, E., Nicholson, J. K, Nicholls, A. W., Lindon, J. C., Connor, S. C., Polley, S., and Connelly, J. (1998) The identification of novel biomarkers of renal toxicity using automatic data reduction techniques and PCA of proton NMR spectra of urine. Chemom. Intell. Lab. Syst. 44, 245-255. (17) Beckwith-Hall B. M., Nicholson, J. K., Nicholls, A. W., Foxall, P. J. D., Lindon, J. C., Connor, S. C., Abdi, M., Connelly, J., and

(19)

(20)

(21)

(22) (23) (24) (25) (26) (27)

(28)

(29)

(30)

(31)

(32)

(33)

(34)

(35)

Holmes, E., (1998) Nuclear magnetic resonance spectroscopic and principal components analysis investigations into biochemical effects of three model hepatotoxins, Chem. Res. Toxicol. 11, 260272. Holmes, E., Nicholls, A. W., Lindon, J. C., Ramos, S., Spraul, M, Neidig, P, Connor, S. C., Connelly, J., Damment, S. J. P., Haselden, J. N., and Nicholson, J. K. (1998) Development of a model for classification of toxin-induced lesion using 1H NMR spectroscopy of urine combined with pattern-recognition; NMR Biomed. 11, 235-244. Nicholson, J. K., Lindon, J. C., and Holmes, E., (1999) ‘Metabonomics’: understanding the metabolic responses of living systems to pathophysiological stimuli via multivariate statistical analysis of biological NMR spectroscopic data. Xenobiotica 29, 1181-1189. Preece, N. E., Nicholson, J. K., and Timbrell, J. A. (1991) Identification of novel hydrazine metabolites in 15N NMR. Biochem. Pharmacol. 41, 1319-1324. Burns, S. P, Woolf, D. A., Leonard, J. V., and Iles, R. A. (1992) Investigation of urea cycle enzyme disorders by 1H NMR spectroscopy. Clin. Chim. Acta 209, 47-60. Lindon, J. C., Nicholson, J. K., and Everett, J. R. (1999) NMR spectroscopy of biofluids. Ann. Rep. NMR Spectrosc. 38, 1-88. Newsholme, E. A., and Leech, A. R. (1995) Biochemistry for the medicinal sciences, John Wiley and Sons, Chichester. Seigel, G. J. (1993) Basic neurochemistry: molecular, cellular and medical aspects, Raven Press, New York. Cornish, H. H. (1969) The role of vitamin B6 in the toxicity of hydrazines. Ann. N. Y. Acad. Sci. 166, 136-145. Stryer, L. (1995) Biochemistry, W. H. Freeman and Company, New York. Olney, J. W., deGubareff, T., and Collins, J. F. (1980) Sterospecificity of the gliotoxic and anti-neurotoxic actions of alphaaminoadipate. Neurosci. Lett. 19, 277-282. Wu, H. Q., Ungerstedt, U., and Schwarcz, R. (1995) L-R-Aminoadipic acid as a regulator of kynurenic acid production in the hippocampus: a microdialysis study in freely moving rats. Euro. J. Pharmacol. 281, 55-61. Alberti-Giani, D., Malherbe, P., Kohler, C., Lang, G., Kiefer, V., Lahm, H.-W., and Cesura, A. M. (1995) Cloning and characterization of a soluble kynurenine aminotransferase from rat brain: identity with kidney cysteine conjugate β-lyase. J. Neurochem. 64, 1448-1455. Mosca, M., Cozzi, L., Breton, J., Speciale, C., Okuno, E., Schwarcz, R., and Benatti, L. (1994) Molecular cloning of rat kynurenine aminotransferase: identity with glutamine transaminase K. FEBS Lett. 353, 21-24. Buchli, R., Alberti-Giani, D., Malherbe, P., Ko¨hler, C., Broger, C., and Cesura, A. M. (1995) Cloning and functional expression of a soluble form of kynurenic/R-aminoadipate aminotransferase from rat kidney. J. Biol. Chem. 270, 29330-29335. Guidetti, P., Okuno, E., and Schwarcz, R. (1997) Characterization of rat brain kynurenine aminotransferases I and II. J. Neurosci. Res. 50, 457-465. Tobes, M. C., and Mason, M. (1975) L-Kynurenine and L-Raminoadipate aminotransferase. I Evidence for identity. Biochem. Biophys. Res. Commun. 62, 390-397. Roberge, A., Gosselin, C., and Charbonneau, R. (1971) Effect of hydrazine on urea cycle enzymes in vitro and in vivo. Biochem. Pharmacol. 20, 2231-2238. Runge-Morris, M. A., Iacob, S., and Novak, R. F. (1988) Characterization of hydrazine-stimulated proteolysis in human erythrocytes. Toxicol. Appl. Pharmacol. 94, 414-426.

TX000231J

Metabonomic Investigations into Hydrazine Toxicity in the Rat

Recommend Documents