NMR based metabolomic approach revealed cyclophosphamide-induced systematic alterations in a rat model

Abstract

Cyclophosphamide (CY) is the most commonly used antineoplastic drug, but various side effects are often observed. In this work, a ¹H NMR based metabolomic approach combined with biochemical assay and histopathological inspection were employed to study the toxic effects of CY on rats. The endogenous metabolites in liver, kidney, and heart were identified. Multivariate analysis revealed that the 18 metabolites in the liver, 14 metabolites in the kidney, and 6 metabolites in the heart were significantly altered after CY treatment. Metabolic pathway analysis showed that the metabolic alterations were related to amino acids metabolism, energy metabolism, choline metabolism, nucleotide metabolism and oxidative stress-related metabolism. This integrative study should be useful for the clinical safety evaluation of CY, and shows the power of a ¹H NMR-based metabolomic approach to evaluate the toxicity of drugs systematically.

---

1. Introduction

Cyclophosphamide (CY) is the most commonly used antineoplastic drug for multiple types of tumors, especially malignant lymphoma. Despite its pharmacological efficacy, the major limitation of CY chemotherapy is its injury to normal tissue including immunosuppression, myelosuppression and genotoxicity.^1–3 There are many hypotheses for the underlying mechanisms of CY-induced toxicity. Immunotoxicity mechanism was connected with the modulation of tryptophan metabolism, phospholipid metabolism, energy metabolism, amino acids metabolism, oxidative stress, and choline metabolism.^4,5 Hepatotoxicity mechanism was involved with the formation of reactive oxygen species (ROS), inflammatory, disturbance of mitochondria and telomerase, and related to voltage-dependent anion channel (VDAC) in mitochondrial outer membrane of liver.^6–9 Cardiotoxicity mechanism was related to gene expression of carnitine palmitoyl transferase I, heart fatty acid binding protein, NF-κB pathway, and cytochrome P450.^7,10,11 Nephrotoxicity mechanism was relevant to oxidant stress and cell apoptosis.^12,13

Metabolomics is a complementary technique to genomics, transcriptomics and proteomics. It is a comprehensive technique for the assessment and simultaneous profiling of endogenous metabolic changes in living systems.^14,15 Nuclear magnetic resonance (NMR) has been one of the most widely utilized approaches in metabolomic analyses, and has unique advantages, such as rapidity, nonselectivity, high reproducibility, and great stability.¹⁶ This is likely to be particularly useful when the mechanisms of toxic action of a xenobiotic on an organism is not known or is not fully understood. Recently, ¹H NMR based metabolomic approach have been successfully applied to a series of toxicity studies, such as lipopolysaccharide-induced liver and kidney toxicity,¹⁷ melamine-induced acute renal toxicity,¹⁸ and doxorubicin-induced heart toxicity.¹⁹

Previous metabolomic studies on the toxicity of CY were mainly focused on its immunotoxicity, by analyzing the metabolic perturbations in urine, thymus, spleen, and plasma.^4,20,21 The immunomodulating mechanism of Huangqi injection on leucopenia mice induced by CY have been investigated in our previous study,⁵ and the side effects to other organs in addition to bone marrow in leucopenia animal models were also observed. However, comprehensive and holistic understanding of CY-induced toxicity remains to be achieved, which is important for elucidating the toxicological mechanism of CY.

In this study, the toxicity of CY was investigated by an NMR-based metabolomic approach. Rats were intraperitoneally injected with CY once a day for five consecutive days. Then liver, kidney and heart samples were collected and their ¹H NMR spectra were recorded, which were then analyzed by multivariate techniques. Biochemical assays and histopathological examination were also employed to study the toxic effects of CY on rats.

2. Materials and methods

2.1. Chemicals

CY was purchased from Shanxi Pude Pharmaceutical Co., Ltd (Batch No. 15092125, Shanxi, China). After dissolved in sterilized physiological saline to 8.0 mg mL⁻¹, CY was stored at 4 °C before use. Sodium 3-trimethylsilyl[2,2,3,3-d₄]propionate (TSP) was obtained from Cambridge Isotope Laboratories Inc (Andover, MA, USA). D₂O was bought from Norell (Landisville, NJ, USA). Analytical grade K₂HPO₄·3H₂O and NaH₂PO₄·2H₂O were obtained from Guangfu Fine Chemical Research Institute (Tianjin, China) and Zhiyuan Chemical Reagent Co., Ltd (Tianjin, China), respectively. Phosphate buffer was prepared with K₂HPO₄·3H₂O and KH₂PO₄·2H₂O (0.2 M, pH 7.4), containing 10% D₂O and 0.01% TSP. Aspartate amino transferase (AST), alanine aminotransferase (ALT), urea nitrogen (BUN), and creatinine (CREA) kits were supplied by Shanghai Fuxing Changzheng Medical Science Co., Ltd (Shanghai, China). Superoxide dismutase (SOD), catalase (CAT), citrate synthase (CS), and pyruvate kinase (PK) kits were supplied by Sangon Biotech (Shanghai, China).

2.2. Animals and sample collection

Sixteen male Sprague Dawley (SD) rats (180–220 g) were purchased from Beijing Vital River Laboratories Co., Ltd. (SCXK (Jing) 2011-0012, Beijing, China). All rats were acclimated for one week under the follow conditions, temperature of 20–24 °C, relative humidity 65 ± 10% and a 12 h dark–light cycle, with free access to food and water.

Animal experiment was performed as previously described with minor modifications.²² After one week of environmental adaptation, the rats were randomly divided into 2 groups: CY group (n = 8) and control group (n = 8). For CY group, rats were intraperitoneally injected (i.p.) with CY for 5 days at a dose of 40 mg per kg per day. For the control group, rats received equal volume of physiological saline. Seven days after the last injection, all rats were sacrificed after collecting blood from ophthalmic venous plexus. 20 μL blood was collected by retro-orbital bleed into heparin, and determined of white blood cell (WBC), lymphocyte (LY), and monocytes (MO) by HEMAVET950FS automatic animal blood analyzer. The remaining blood was centrifuged at 13 000 rpm for 10 min to obtain serum for the determination of AST, ALT, BUN, and CREA activities. Liver, kidney, and heart were also collected, each one was split into two parts, where one part was fixed in 10% formalin solution, and the other part was snap-frozen in liquid nitrogen and stored at −80 °C for further analysis.

2.3. DNA content of bone marrow cells

Rats femur were collected and the cells were flushed by 10 mL of CaCl₂ solution (5 mmol L⁻¹) to centrifuge tube. After centrifugation for 15 min at 2500 rpm, 5 mL of HClO₄ solution (0.2 mol L⁻¹) was added to the precipitate and the mixture was heated for 15 min in water bath (90 °C), the supernatant was used to determine the ultraviolet absorbance value at 268 nm after centrifuging for 10 min at 3500 rpm.^23,24

2.4. Biochemical assays

The levels of ALT, AST, BUN, CREA in serum, SOD, CAT in liver, CS, PK in kidney and heart were assayed according to the instructions of commercial kits.

2.5. Histopathological assessments

Histopathological evaluation was conducted for all rats. Liver, kidney, and heart from 10% formalin solution were embedded in paraffin blocks, thin-sliced to 4–5 μm thickness, and then stained with hematoxylin–eosin (HE) staining. The thin-session descriptions were carried out using a bright-field Carl Zeiss(r) optical microscope, and then photographed at 200× magnification to compare the tissue structures of the organs.

2.6. Sample preparation for NMR measurements

Organ tissues were extracted as previously described with minor modifications.²⁵ Liver, kidney and heart tissues (about 200 mg) were extracted with 900 μL of methanol/water (2 : 1, v/v) using an ultrasonic cell crusher (Ningbo Scientz Biotechnology Co., Ltd, Ningbo, China). After centrifugation (13 000 rpm, 4 °C, 15 min), each supernatant was dried under nitrogen flow. The dried sample was redissolved in 700 μL of phosphate buffer (0.2 M K₂HPO₄–0.2 M KH₂PO₄, pH 7.4) containing 0.01% TSP and 10% D₂O. Following final centrifugation (13 000 rpm, 4 °C, 10 min), 600 μL supernatant was transferred into a 5 mm NMR tube for ¹H NMR analysis.

2.7. NMR measurements for liver, kidney, and heart

¹H NMR and 2D NMR spectral data were acquired on a Bruker 600 MHz AVANCE III NMR spectrometer (Bruker, Germany) operated at 600.13 MHz. ¹H NMR used noesypr1d pulse sequence for all tissue extracts. Each ¹H NMR spectrum was consisted of 64 scans, and each scan required 2.654 s acquisition time with the following parameters: spectral width of 12 345.7 Hz, spectral size of 65 536 points, and a relaxation delay (RD) of 1.0 s.

Each ¹H–¹H correlated spectroscopy (COSY) spectrum was consisted of 17 scans with the following parameters: spectral widths of 6602.1 Hz in both dimensions, and relaxation delay (RD) of 1.5 s. Each heteronuclear single quantum coherence (HSQC) spectrum was consisted of 55 scans with the following parameters: spectral widths of 6602.1 Hz in the ¹H dimension and 36 220.3 Hz in the ¹³C dimension, and relaxation delay (RD) of 1.2 s. Each J-resolved (JRES) spectrum was consisted of 18 scans with the following parameters: spectral width of 6602.1 Hz in F2 and 40 Hz in F1, and relaxation delay (RD) of 1.0 s. All 2D spectra were calibrated to TSP at 0.00 ppm.

2.8. NMR data processing

All spectra were manually phased and baseline-corrected using MestReNova software (version 8.0.1, Mestrelab Research, Santiago de Compostela, Spain). TSP with a chemical shift at δ 0.00 was used as spectral reference for all tissue extracts. A linear interpolation method provided in MestReNova was used to align the spectra. The regions contained residual water and methanol signals (δ 4.70–5.00 and δ 3.35–3.38 for liver, kidney, and heart, respectively) were removed prior to data normalization. All spectra were then segmented at 0.01 ppm intervals across δ 0.50–9.50 for liver, kidney, and heart extract samples, respectively. Normalization to a total sum of all integrals for tissue extracts was conducted before multivariate analysis.

2.9. Multivariate data analysis

Multivariate data analysis was performed using Simca-P 13.0 software (Umetrics, Sweden). Principal component analysis (PCA) was first performed on mean-centered data to identify outliers between the control and CY groups. Partial least-squares discriminant analysis (PLS-DA) was applied to discriminate metabolic profiles between samples groups. Orthogonal-projection to latent structure-discriminate analysis (OPLS-DA) is a supervised method to identify metabolites that contributing to the group separation. The validity of the models was assessed by 200 permutation tests, 7-fold cross-validation and CV-ANOVA method.

Relative amounts of metabolites were calculated on the basis of the integrated regions (buckets) from the least overlapping NMR signals of metabolites. Experimental values were expressed as mean ± standard deviation (SD). Results of body weight, WBC, LY, MO, DNA, ALT, AST, BUN, CREA, SOD, CAT, CS, PK, and the peak areas (buckets) of the differential metabolites were further compared by t-test using SPSS16.0 software, and p-values less than 0.05 were considered to be significant.

3. Results

3.1. Body weight

Rats were injected with CY on days 1 to 5. By the end of experiment (day 12), the mortality was 12.5% in CY group, and no death was observed in control group. The body weight of rats in CY and control groups were shown in Fig. 1. It was obvious that the rat body weight consistently increased in the control group, however, the body weight of CY group was decreased. Significant differences of the body weight between the control and CY groups occurred on the 5th day (p < 0.01). For the CY treated rats, the body weight showed no significant change from day 1 to day 8, then decreased rapidly at day 12, suggesting that severe toxicity of CY appeared on the 7th day after the last injection.

Fig. 1 Changes in body weight of control and CY groups. Compared with control group, ^#p < 0.05, ^##p < 0.01.

3.2. Blood parameters and DNA content of bone marrow cells

The levels of WBC, LY, MO, and DNA content of rats in the CY and control groups were depicted in Fig. 2A. Compared with the control group, the rats in CY group showed lower levels of WBC, LY, and MO (p < 0.01), which was in agreement with the previous studies.^5,26,27 In addition, CY treatment caused the decrease of DNA content of bone marrow cells (p < 0.05), indicating that CY had serious bone marrow suppression effect.

Fig. 2 Scatter figures of periphery blood parameters and DNA content of bone marrow cells (A), serum biochemistry (B) in control and CY groups. Compared with control group, ^#p < 0.05, ^##p < 0.01.

3.3. Serum biochemistry

ALT and AST are normally localized in the liver cytoplasm and released into the circulation with liver cellular damages.²⁸ Fig. 2B showed that ALT and AST were increased significantly in the CY group (p < 0.05). In addition, CY treatment also caused significant changes of CREA and BUN (p < 0.05), which provided an indication of ‘renal status’ in the animals.

The levels change of SOD, CAT, CS and PK in the CY and control groups were shown in Fig. 3. Compared with the control group, SOD and CAT were lower in the liver of CY group (p < 0.01), which was in agreement with the previous report.²¹ In addition, the decrease of CS and PK in kidney and heart (p < 0.01) were also observed in the CY group.

Fig. 3 Scatter figures of SOD and CAT in liver (A), CS and PK in kidney (B) and heart (C) in control and CY groups. Compared with control group, ^#p < 0.05, ^##p < 0.01.

3.4. Liver, kidney, and heart histopathology

Fig. 4 depicted the results of light microscopy micrographs of HE staining of liver, kidney, and heart tissues of control and CY groups. Morphology of the liver tissue was normal in control group. However, liver sections from CY treated rats showed extensive liver injuries, characterized by dot necrosis, mild degeneration and eosinophilic change in liver cell, and mild level of proliferation in bile duct. Compared with the control group, kidney tissue of CY group was injured severely, showing tubular epithelial cell degeneration (edema), mild to moderate level of expansion in the lumen, atrophy in malpighian cell, and largen of cisternae in glomerular. Histopathological examination revealed no significant change of heart for control group, whereas CY-treated rats showed mild to moderate level of atrophy in cardiac muscle fibers, and mild degeneration in vacuoles.

Fig. 4 Histopathological photomicrographs of rat liver (A), kidney (B), and heart (C) (A, B and C ×200). LC, KC, and HC mean control group rats, LCY, KCY, and HCY mean CY group rats.

3.5. Metabolomics

3.5.1 Metabolite assignment. Fig. 5 showed typical ¹H NMR spectra of liver, kidney, and heart extracts of the control and CY groups. NMR signals were assigned to specific metabolites (Table S1†) based on the literature^29,30 and further confirmed by NMR databases such as HMDB (http://www.hmdb.ca/), and BMRB (http://www.bmrb.wisc.edu/). In addition, 2D NMR spectra including ¹H–¹H COSY, HSQC and JRES were also used in the metabolite identification (Fig. S1–S3†). The detected metabolites in liver, kidney, and heart included 4 organic acids (taurine, 3-hydroxybutyrate, acetate, and formate), 13 amino acids (valine, leucine, isoleucine, alanine, lysine, arginine, ornithine, glutamate, methionine, glycine, threonine, tyrosine, and phenylalanine), 2 glycolysis products (lactate and pyruvate), 3 tricarboxylic acid cycle (TCA) intermediates (fumarate, malate, and succinate), 3 choline metabolites (choline, phosphocholine, and glycerophosphocholine), 2 organic bases (dimethylamine and trimethylamine), 11 nitrogen compounds (creatine, betaine, myo-inositol, inosine, uracil, uridine, cytidine, histidine, xanthine, hypoxanthine, and nicotinamide), and a number of other metabolites. Endogenous metabolites identified in liver, kidney, and heart, were listed in Table S1.†

Fig. 5 NMR spectra of liver (A), kidney (B) and heart (C) from control (LC, KC, HC) and CY-treated (LCY, KCY, HCY) rats.

3.5.2 CY-induced metabolomic changes in multiple biological matrices. Visual inspection of ¹H NMR spectra of rats liver, kidney, and heart revealed significant differences between control and CY groups. To extract more details about CY-induced metabolic changes, multivariate analysis was further performed.

In the PCA score plots (Fig. 6A, S4A and S5A†), which were generated by PC1 (37.5%) and PC2 (18.3%) for liver, PC1 (29.6%) and PC2 (20.1%) for kidney, and PC1 (28.1%) and PC2 (21.7%) for heart, obvious separation were observed between control and CY groups. The PLS-DA model was further constructed and validated using 200 permutation tests, in which all R² and Q² values were lower than the original ones, suggesting the validity of the discriminant model (Fig. 6B, S4B and S5B†). The corresponding OPLS-DA was used to determine the potential biomarkers contributing to the separation (Fig. 6C, S4C and S5C†), and the parameters R² (0.926), Q² (0.871), p value (1.88 × 10⁻⁴) for liver, R² (0.907), Q² (0.753), p value (7.71 × 10⁻³) for kidney, and R² (0.991), Q² (0.793), p value (4.80 × 10⁻²) for heart, indicated the interpretability, predictability, as well as the validity of these established models. The OPLS-DA loading plot (Fig. 6D, S4D and S5D†) combined with variable importance in the projection (VIP > 1.0) was applied to find metabolites contributing to the separation.

The major metabolites variations in liver after CY treatment were the decline of 3-hydroxybutyrate (3-HB), acetate, glutamate, glutamine, oxidized glutathione (GSSG), phosphocholine (PC), choline, glycerin, glycerophosphocholine (GPC), cytidine, uracil, inosine, formate, and tyrosine, as well as the elevation of lactate, malate, aspartate, and taurine (Table 1). For kidney tissue, CY exposure caused level decrease of isoleucine, leucine, 3-HB, acetate, methionine, glutamine, pyruvate, inosine monophosphate, and hypoxanthine, and level elevation of lactate, creatine, choline, GPC, and carnitine (Table S2†). Metabolic changes contributing to the separation of heart were mainly involved in the accumulation of acetate, PC, taurine, and xanthine, as well as excessive reduction of lactate, and malate in CY group (Table S3†). The level of a metabolite can be measured by the peak areas (buckets), and was further compared by t-test. Results showed that all the perturbed metabolites showed significant difference (p < 0.05) between control and CY groups (Tables 1, S2 and S3†).

Fig. 6 PCA score plot (A), PLS-DA permutation test (B), OPLS-DA score plot (C), and loading plot (D) of rats liver between control and CY group.

Table 1 Comparison of normalized integral levels of metabolites in liver (as some of the values are too small, all of the data were magnified 100 times)^a

δ^1H	Metabolites	Control group	CY group	VIP
a Note: 3-HB: 3-hydroxybutyrate, GSSG: oxidized glutathione, PC: phosphocholine, GPC: glycerophosphocholine. ^#p < 0.05 and ^##p < 0.01 vs. control rats.
1.21	3-HB	23.58 ± 5.98	13.64 ± 4.26^##	1.52
1.34	Lactate	175.21 ± 40.79	321.16 ± 91.17^##	6.00
1.93	Acetate	45.21 ± 5.81	29.46 ± 9.93^##	1.94
2.08	Glutamate	39.32 ± 6.73	28.55 ± 7.45^#	1.49
2.15	Glutamine	63.26 ± 8.72	44.86 ± 13.20^#	2.01
2.17	GSSG	68.80 ± 10.74	44.63 ± 5.28^##	2.57
2.66	Malate	4.25 ± 1.33	15.96 ± 9.12^#	1.65
2.67	Aspartate	14.09 ± 3.68	32.09 ± 17.17^#	1.92
3.21	PC	118.61 ± 40.56	49.65 ± 23.17^##	4.09
3.22	Choline	103.12 ± 49.59	27.98 ± 11.11^##	4.26
3.42	Taurine	87.23 ± 43.50	212.49 ± 45.25^##	5.86
3.66	Glycerin	48.83 ± 8.60	26.44 ± 3.13^##	2.53
3.68	GPC	44.61 ± 5.88	27.17 ± 1.65^##	2.27
6.12	Cytidine	13.19 ± 6.13	6.07 ± 2.93^#	1.20
6.92	Tyrosine	6.37 ± 0.95	4.85 ± 1.56^#	1.07
7.57	Uracil	3.77 ± 0.78	1.99 ± 0.92^##	1.27
8.37	Inosine	32.28 ± 5.14	17.72 ± 8.10^#	1.60
8.47	Formate	11.83 ± 7.69	3.71 ± 1.61^#	1.27

---

3.6. Correlation analysis of blood parameters, DNA content, serum biochemistry and metabolites

To investigate the relationship between blood parameters, DNA content, serum biochemistry and perturbed metabolites in liver, kidney and heart extracts, correlation matrixes were generated by calculating the Pearson's correlation coefficient (Fig. 7 and S6†). As shown in Fig. 7A, ALT and AST showed positive correlations with malate, aspartate, and taurine, and negative correlations with 3-HB, GSSG, PC, choline, glycerin, GPC, cytidine, uracil, inosine, and formate in liver extracts. In addition, BUN showed a positive correlation with lactate, choline, and GPC and negative correlation with leucine, acetate, methionine, glutamine, pyruvate, and hypoxanthine in kidney extracts, while CREA showed opposite results with BUN (Fig. 7B). Moreover, WBC, LY, MO and DNA showed obviously positive correlation with GSSH, PC, choline, glycerin, GPC, cytidine, formate, acetate, and hypoxanthine, as well as negative correlation with lactate, malate, aspartate, taurine, creatine, carnitine, and xanthine in organ extracts (Fig. S6†).

Fig. 7 Correlation analysis of ALT, AST, BUN, CREA, and metabolites. ((A) Pearson's correlations of ALT, AST and quantities of the metabolites determined from liver samples; (B) Pearson's correlations of BUN, CREA and quantities of the metabolites determined from kidney samples). Red and blue represent positive and negative correlations, respectively, the colour scale represents Pearson's correlation coefficients.

4. Discussion

CY, a prodrug, must be catalyzed by the hepatic cytochrome P450 in liver to 4-hydroxy-cyclophosphamide and later to phosphoramide mustard and acrolein.²⁷ Acrolein is linked with the side effects of CY.³¹ Toxicities of CY include the suppression of white blood cells, nausea, vomiting, gonadal atrophy, as well as liver, renal, and bladder injury. Studies showed that heart toxicity was one of the major side effects of CY and contributed to a high rate of morbidity and mortality,^7,10 thus in our study we chose liver, kidney, and heart to investigate the CY induced toxicity on rats.

In this study, immune-suppression in CY-treated rats was revealed by the level decrease of WBC, LY, MO and DNA, indicating that CY showed obvious bone marrow suppression. Meanwhile, heptatotoxic, nephrotoxicity, and cardiotoxicity were also observed, which were consistent with the previous reports.^7,10,32,33 The result of metabonomic analysis showed that CY induced disturbances in amino acids metabolism, energy metabolism, choline metabolism, nucleotide metabolism and oxidative stress-related metabolism. A schematic diagram of the perturbed metabolic pathways was shown in Fig. 8.

Fig. 8 Summary of pathway alterations after CY treatment in rat liver, kidney, and heart.

4.1. Hepatotoxicity related metabolic changes

Based on previous studies, oxidative stress is one of the major causes of CY-induced hepatotoxicity and is mediated by the secondary metabolites of CY.¹⁹ Acrolein is a highly reactive metabolite of CY, and could readily react with glutathione (GSH). When GSH was exhausted, it could react with cellular nucleophiles, leading to the loss of protein function, and inducing the oxidative stress, which finally give rise to the disastrous effects on hepatocyte.^21,34,35 GSH, is the major natural antioxidant, which combats oxidative injury. Glutamine (via glutamate), cysteine and glycine are the precursor amino acids for the synthesis of GSH.³⁶ Disorders of glutamate and glutamine might be a consequence of an inhibited GSH synthesis, and further lead to increased level of reactive oxygen species (ROS).¹⁶ The ROS could disrupt both the structure and function of membranes, eventually resulting in the rupture of cell and organelles.³⁷ Taurine, a free amino acids and synthesized from cysteine in the liver,³⁸ may cause an enhancement in GSH level by directing cysteine into the GSH synthesis pathway.³⁹ In the present study, decreased levels of glutamate, glutamine and elevated levels of taurine were observed in the CY group. It might be a consequence of their over consumption to counteract ROS generation, and was correlated with CY induced hepatotoxicity, which were confirmed by the significantly reduced level of SOD and CAT in CY group. Previous study on hepatotoxicity showed similar results for glutamate and glutamine.³⁹

Inosine, a purine nucleoside containing the base hypoxanthine and the sugar ribose, is formed during the breakdown of adenosine by adenosine deaminase.⁴⁰ Variation in nucleotide synthesis in the liver is particularly important, as the liver provides the nucleotides necessary to maintain the functions of other organs.⁴¹ Compared with control group, a significant decrease of inosine in liver was found in CY group, suggesting that the disturbance of nucleotide metabolism was induced. The decreased cytidine and uracil in CY treated rats suggested that the hepatotoxicity of CY may be also correlated with disturbance of nucleotide metabolism. Furthermore, catabolism of ATP produced adenosine, which was further decomposed into inosine to produce extra energy.⁴² The decreased inosine revealed that the disturbance of energy metabolism was induced after CY treatment. The marked increase of lactate and malate, and decrease of 3-HB and acetate also verified the disturbance of energy metabolism by CY.

4.2. Nephrotoxicity related metabolic perturbations

Renal damage is a common adverse effect of CY therapy. In our work, altered amino acids profiles were observed in the kidney tissue of CY rats, including decreased levels of isoleucine, leucine, methionine, and glutamine. Amino acids are basic units for protein synthesis in organism, previous study showed that metabolic disorder of amino acids is commonly used as the criteria for diagnosis of kidney morbidity.⁴³ Thus, the abnormalities in amino acids metabolism might be correlated with CY induced nephrotoxicity, which was in agreement with previous report.⁴³

Pyruvate, an important intermediate product of glycolysis, can be used to produce acetyl-CoA by pyruvate dehydrogenase complex. Acetyl-CoA could enter into the TCA cycle, and plays a key role in glucose aerobic oxidation and energy production.⁴⁴ Carnitine is important in the fatty acid β-oxidation. Only with the assistance of carnitine, acetyl-CoA can go into mitochondria for subsequent β-oxidation.⁴⁵ 3-HB is a metabolic intermediate of fatty acids β-oxidation in the mitochondrial, which is a form of the output energy.⁴⁶ Leucine and isoleucine are important ketogenic amino acids, which can be converted to intermediate product of 3-HB for energy production.⁴⁷ The decreased level of pyruvate, 3-HB, leucine, isoleucine and increased level of carnitine might indicate the inhibition of energy metabolism after CY treatment, which was in agreement with the significantly lowered expressions of CS and PK in kidney in CY group.

Choline, an essential element of cell membrane, can be oxidized to betaine by a two-step reaction and finally lead to the generation of creatine.¹⁷ Creatine is regarded as a marker of kidney function, especially for chronic kidney disease.^48,49 In this study, the elevated choline and creatine indicated that CY treatment lead to inhibition of choline metabolisms, and finally resulted in nephrotoxicity. CREA is the non-enzymatic degradation product of creatine and phosphocreatine, and plays a fundamental role in the transfer of energy from mitochondria to cytosol.⁵⁰ The increased creatine and decreased CREA in CY group were also evidences of nephrotoxicity induced by CY.

Hypoxanthine can be converted into xanthine by xanthine oxidase. Allantion, a metabolite closely related with renal damage, can be synthesized by xanthine through urate oxidase.⁵¹ Decrease of hypoxanthine suggested disturbance of purine metabolism after CY administration, and implied that the renal functions were impaired.

4.3. Cardiotoxicity related metabolic changes

Fatty acids, ketone bodies, glucose, amino acids could be used to produce ATP to maintain the contractile function of heart.¹⁹ Malate is an important intermediate of the TCA cycle, the decreased levels in heart tissue might suggest an inhibition of the TCA cycle.⁵² Alanine is an important glucogenic amino acid, and can be converted to intermediates of glucose metabolism for energy production. CY exposure lead to reduction of alanine level in cardiac extracts, suggesting that gluconeogenesis was reduced in myocardial pathological state. The above-mentioned evidences indicated that CY treatment lead to inhibition in energy supply and finally resulted in myocardial lesions, which were confirmed by histopathology results and the significantly lowered expressions of CS and PK in CY group.

5. Conclusion

A ¹H NMR based metabolomic approach combined with biochemical assay and histopathological examination had been employed to study the toxic effects of CY on rats. The disturbance of metabolic profiles of the liver, kidney, and heart extracts were analyzed by multivariate statistical analyses. The results revealed by NMR based metabolomic approach suggested that the toxicity of CY was related with the change of a series of endogenous metabolites, but not a single specific metabolite. The CY induced metabolic alterations were related with amino acids metabolism, energy metabolism, choline metabolism, nucleotide metabolism and oxidative stress-related metabolism.

CY-induced hepatotoxicity, nephrotoxicity and cardiotoxicity were investigated by metabolomic approach for the first time. Hepatotoxicity was involved with oxidative stress-related metabolism and energy metabolism, which was in agreement with previous reports.^6,8 Nephrotoxicity was related with amino acids metabolism, energy metabolism, choline metabolism, and nucleotide metabolism, while cardiotoxicity was related to energy metabolism. The results presented in this study explained the mechanism of nephrotoxicity and cardiotoxicity from a metabolomic point of view. This study should be useful for the clinical safety evaluation of CY, and shows the power of a ¹H NMR-based metabolomic approach to evaluate the toxicity of drugs systematically. In order to further investigate the toxicity of CY, HPLC-MS based metabolomic approach will be used to explore whether the specific metabolites related with the toxicity of CY. In addition, proteins that play critic roles in CY induced toxicity in different organs will be investigated by proteomic approach, to further understand the toxicities induced by CY.

Conflict of interest

We confirm that this manuscript has not been published by another journal. All authors have approved the manuscript and agree with its submission to your journal. The authors declare that there is no conflict of interests regarding the publication of this paper.

Ethics statement

All experiments in this study were in accordance with NIH Guide for the Care and Use of Laboratory Animals, and approved by institutional ethical committee of Shanxi University. Maximum efforts were made to minimize animal suffering and the number of animals necessary for the capture of reliable data.

Acknowledgements

This study was financially supported by the National Natural Science Foundation of China (No. 31570346), Science and Technology Innovation Team of Shanxi Province (No. 201605D131045-18), Institutions of Higher Learning Innovative Talents Support Program of Shanxi Province.

References

1. U. B. Das, M. Mallick, J. M. Debnath and D. Ghosh, Protective effect of ascorbic acid on cyclophosphamide-induced testicular gametogenic and androgenic disorders in male rats, Asian J. Androl., 2002, 4, 201–207.
2. S. J. Hosseinimehr, M. Azadbakht and A. J. Abadi, Protective effect of hawthorn extract against genotoxicity induced by cyclophosphamide in mouse bone marrow cells, Environ. Toxicol. Pharmacol., 2008, 25, 51–56.
3. X. H. Huyan, Y. P. Lin, T. Gao, R. Y. Chen and Y. M. Fan, Immunosuppressive effect of cyclophosphamide on white blood cells and lymphocyte subpopulations from peripheral blood of Balb/c mice, Int. Immunopharmacol., 2011, 11, 1293–1297.
4. J. Li, W. S. Lin, W. W. Lin, P. Peng, J. M. Zhang, H. S. Yang and X. M. Ling, Metabolic profiling study on potential toxicity and immunotoxicity-biomarker discovery in rats treated with cyclophosphamide using HPLC-ESI-IT-TOF-MS, Biomed. Chromatogr., 2015, 29, 768–776.
5. T. L. Qu, Z. Y. Li, S. J. Zhao, A. P. Li and X. M. Qin, A metabonomic analysis reveals novel regulatory mechanism of Huangqi injection on leucopenia mice, Immunopharmacol. Immunotoxicol., 2016, 38, 113–123.
6. S. S. Nair, J. J. Manalil, S. K. Ramavarma, I. M. Suseela, A. Thekkepatt and A. C. Raghavamenon, Virgin coconut oil supplementation ameliorates cyclophosphamide-induced systemic toxicity in mice, Hum. Exp. Toxicol., 2016, 35, 205–212.
7. Y. F. Song, C. Y. Zhang, C. X. Wang, L. Zhao, Z. Wang, Z. J. Dai, S. Lin, H. F. Kang and X. B. Ma, Ferulic acid against cyclophosphamide-induced heart toxicity in mice by inhibiting NF-κB pathway, Evidence-Based Complementary and Alternative Medicine, 2016, 2016, 1–8.
8. Z. T. Deng, Study the pharmacology and toxicology of cyclophosphamide, Tianjin Pharmacy, 2009, 21, 42–44.
9. B. V. Shoshan, The voltage-dependent an ion channel-characterization, modulation, and role in mitochondrial function in cell life and death, Cell Biochem. Biophys., 2003, 39, 279–292.
10. M. M. Sayed-Ahmed, M. L. Aldelemy, O. A. Al-Shabanah, M. M. Hafez, K. A. Al-Hosaini, N. O. Al-Harbi, S. D. Al-Sharary and M. M. Al-Harbi, Inhibition of gene expression of carnitine palmitoyl transferase I and heart fatty acid binding protein in cyclophosphamide and ifosfamide-induced acute cardiotoxic rat models, Cardiovasc. Toxicol., 2014, 14, 232–242.
11. M. Colvin and J. Hilton, Pharmacology of cyclophosphamide and metabolites, Cancer Treat. Rep., 1980, 65, 89–95.
12. O. Sinanoglu, A. N. Yener, S. Ekici, A. Midi and F. B. Aksungar, The protective effects of spirulina in cyclophosphamide induced nephrotoxicity and urotoxicity in rats, Urology, 2012, 80, 1392. e1–1392. e6.
13. M. U. Rehman, M. Tahir, F. Ali, W. Qamar, A. Lateef, R. Khan, A. Quaiyoom, O. O. Hamiza and S. Sultana, Cyclophosphamide-induced nephrotoxicity, genotoxicity, and damage in kidney genomic DNA of Swiss albino mice: the protective effect of Ellagic acid, Mol. Cell. Biochem., 2012, 365, 119–127.
14. X. J. Wang, A. H. Zhang, Y. Han, P. Wang, H. Sun, G. C. Song, T. W. Dong, Y. Yuan, X. X. Yuan, M. Zhang, N. Xie, H. Zhang, H. Dong and W. Dong, Urine metabolomics analysis for biomarker discovery and detection of jaundice syndrome in patients with liver disease, Mol. Cell. Proteomics, 2012, 11, 370–380.
15. X. J. Wang, B. Yang, H. Sun and A. H. Zhang, Pattern recognition approaches and computational systems tools for ultra performance liquid chromatography-mass spectrometry-based comprehensive metabolomic profiling and pathways analysis of biological data sets, Anal. Chem., 2012, 84, 428–439.
16. P. P. Guo, D. D. Wei, J. S. Wang, G. Dong, Q. Zhang, M. H. Yang and L. Y. Kong, Chronic toxicity of crude ricinine in rats assessed by ¹H NMR metabolomics analysis, RSC Adv., 2015, 5, 27018–27028.
17. S. T. Liao, P. Li, J. S. Wang, Q. Zhang, D. Q. Xu, M. H. Yang and L. Y. Kong, Protection of baicalin against lipopolysaccharide induced liver and kidney injuries based on ¹H NMR metabolomic profiling, Toxicol. Res., 2016, 5, 1148–1159.
18. G. X. Xie, X. J. Zheng, X. Qi, Y. Cao, Y. Chi, M. M. Su, Y. Ni, Y. P. Qiu, Y. M. Liu, H. K. Li, A. H. Zhao and W. Jia, Metabonomic evaluation of melamine-induced acute renal toxicity in rats, J. Proteome Res., 2010, 9, 125–133.
19. Q. Y. Niu, Z. Y. Li, G. H. Du and X. M. Qin, ¹H NMR based metabolomic profiling revealed doxorubicin-induced systematic alterations in a rat model, J. Pharm. Biomed. Anal., 2016, 118, 338–348.
20. F. Li, A. D. Patterson, C. C. Höfer, K. W. Krausz, F. J. Gonzalez and J. R. Idle, Comparative metabolism of cyclophosphamide and ifosfamide in the mouse using UPLC-ESI-QTOFMS-based metabolomics, Biochem. Pharmacol., 2010, 80, 1063–1074.
21. H. Zhu, M. H. Long, J. Wu, M. M. Wang, X. Y. Li, H. Shen, J. D. Xu, L. Zhou, Z. J. Fang, Y. Luo and S. L. Li, Ginseng alleviates cyclophosphamide-induced hepatotoxicity via reversing disordered homeostasis of glutathione and bile acid, Sci. Rep., 2015, 5, 1–14.
22. C. Jin and Y. Jie, Effect of moxibustion or acupuncture at Geshu acupoint on the granulocyte-macrophage colony stimulating factor of cyclophosphamide induced leukopenic rats, Zhongguo Zuzhi Gongcheng Yanjiu Yu Linchuang Kangfu, 2007, 11, 5473–5476.
23. L. M. Dai, Y. A. Xiong, G. Y. Yang, Y. Y. Lin and M. Yang, Protective effect of tannins from Sanguisorba officinalis on cyclophosphamide-induced myelosuppression in mice, Nat. Prod. Res. Dev., 2014, 26, 499–503.
24. J. Ma, X. G. Liu, P. Zhang, X. X. Yi and Z. W. Wang, Study on the influence of Shengbai capsules on cyclophosphamide-induced leucopenia in mice, Journal of Gansu College of Traditional Chinese Medicine, 2010, 27, 20–22.
25. C. F. Lu, Y. M. Wang, Z. G. Sheng, G. Liu, Z. Fu, J. Zhao, J. Zhao, X. Z. Yan, B. Z. Zhu and S. Q. Peng, NMR-based metabonomic analysis of the hepatotoxicity induced by combined exposure to PCBs and TCDD in rats, Toxicol. Appl. Pharmacol., 2010, 248, 178–184.
26. J. X. Wang, X. Tong, P. B. Li, H. Cao and W. W. Su, Immuno-enhancement effects of Shenqi Fuzheng injection on cyclophosphamide-induced immunosuppression in Balb/c mice, J. Ethnopharmacol., 2012, 139, 788–795.
27. J. X. Wang, X. Tong, P. B. Li, M. H. Liu, W. Peng, H. Cao and W. W. Su, Bioactive components on immuno-enhancement effects in the traditional Chinese medicine Shenqi Fuzheng injection based on relevance analysis between chemical HPLC fingerprints and in vivo biological effects, J. Ethnopharmacol., 2014, 155, 405–415.
28. K. J. Lee, J. H. Choi, H. G. Kim, E. H. Han, Y. P. Hwang, Y. C. Lee, Y. C. Chung and H. G. Jeong, Protective effect of saponins derived from the roots of Platycodon grandiflorum against carbon tetrachloride induced hepatotoxicity in mice, Food Chem. Toxicol., 2008, 46, 1778–1785.
29. X. H. Shi, C. N. Xiao, Y. L. Wang and H. R. Tang, Gallic acid intake induces alterations to systems metabolism in rats, J. Proteome Res., 2013, 12, 991–1006.
30. A. P. Li, Z. Y. Li, H. F. Sun, K. Li, X. M. Qin and G. H. Du, Comparison of two different Astragali radix by a ¹H NMR-based metabolomic approach, J. Proteome Res., 2015, 14, 2005–2016.
31. H. X. Sun and X. Y. Peng, Protective effect of triterpenoid fractions from the rhizomes of Astilbe chinensis on cyclophosphamide-induced toxicity in tumorbearing mice, J. Ethnopharmacol., 2008, 119, 312–317.
32. G. Cuce, S. Cetinkaya, T. Koc, H. H. Esen, C. Limandal, T. Balcl, S. Kalkan and M. Akoz, Chemoprotective effect of vitamin E in cyclophosphamide-induced hepatotoxicity in rats, Chem.-Biol. Interact., 2015, 232, 7–11.
33. M. O. Germoush and A. M. Mahmoud, Berberine mitigates cyclophosphamide-induced hepatotoxicity by modulating antioxidant status and inflammatory cytokines, J. Cancer Res. Clin. Oncol., 2014, 140, 1103–1109.
34. C. R. Myers and J. M. Myers, The effects of acrolein on peroxiredoxins, thioredoxins, and thioredoxin reductase in human bronchial epithelial cells, Toxicology, 2009, 257, 95–104.
35. C. R. Myers, J. M. Myers, T. D. Kufahl, R. Forbes and A. Szadkowski, The effects of acrolein on the thioredoxin system: implications for redox-sensitive signaling, Mol. Nutr. Food Res., 2011, 55, 1361–1374.
36. E. Roth, R. Oehler, N. Manhart, R. Exner, B. Wessner, E. Strasser and A. Spittler, Regulative potential of glutamine-relation to glutathione metabolism, Nutrition, 2002, 18, 217–221.
37. M. Valko, D. Leibfritz, J. Moncol, M. T. Cronin, M. Mazur and J. Telser, Free radicals and antioxidants in normal physiological functions and human disease, Int. J. Biochem. Cell Biol., 2007, 39, 44–84.
38. P. R. Wang, J. S. Wang, M. H. Yang and L. Y. Kong, Neuroprotective effects of Huang-Lian-Jie-Du-Decoction on ischemic stroke rats revealed by ¹H NMR metabolomics approach, J. Pharm. Biomed. Anal., 2014, 88, 106–116.
39. S. W. Schaffer, K. C. Ramila, C. Ju Jong, A. Shetewy, K. Shimada, T. Ito, J. Azuma and E. Cioffi, Does taurine prolong lifespan by improving heart function, Adv. Exp. Med. Biol., 2015, 555–570.
40. J. Barankiewicz and A. Cohen, Purine nucleotide metabolism in resident and activated rat macrophages in vitro, Eur. J. Immunol., 1985, 15, 627–631.
41. K. Fujisawa, T. Takami, Y. Kimoto, T. Matsumoto, N. Yamamoto, S. Terai and I. Sakaida, Circadian variations in the liver metabolites of medaka (Oryzias latipes), Sci. Rep., 2016, 6, 1–10.
42. C. Y. Lin, F. P. Huang, Y. S. Ling, H. J. Liang, S. H. Lee, M. Y. Hu and P. N. Tsao, Use of nuclear magnetic resonance-based metabolomics to characterize the biochemical effects of naphthalene on various organs of tolerant mice, PLoS One, 2015, 10, e0120429.
43. Y. Ni, M. M. Su, Y. P. Qiu, M. J. Chen, Y. M. Liu, A. H. Zhao and W. Jia, Metabolic profiling using combined GC-MS and LC-MS provides a systems understanding of aristolochic acid-induced nephrotoxicity in rat, FEBS Lett., 2007, 581, 707–711.
44. X. J. Zhao, C. Y. Huang, H. H. Lei, X. Nie, H. R. Tang and Y. L. Wang, Dynamic metabolic response of mice to acute mequindox exposure, J. Proteome Res., 2011, 10, 5183–5190.
45. F. Zhang, Q. Zhan, X. Dong, B. Jiang, L. N. Sun, S. H. Gao, Z. Q. He, X. Tao and W. S. Chen, Shengxian decoction in chronic heart failure treatment and synergistic property of platycodonis radix: a metabolomic approach and its application, Mol. BioSyst., 2014, 10, 2055–2063.
46. B. Plecko, S. Stoeckler-Ipsiroglu, E. Schober, G. Harrer, V. Mlynarik, S. Gruber, E. Moser, D. Moeslinger, H. Silgoner and O. Ipsiroglu, Oral β-hydroxybutyrate supplementation in two patients with hyperinsulinemic hypoglycemia: Monitoring of β-hydroxybutyrate levels in blood and cerebrospinal fluid, and in the brain by in vivo magnetic resonance spectroscopy, Pediatr. Res., 2002, 52, 301–306.
47. X. H. Shi, C. N. Xiao, Y. L. Wang and H. R. Tang, Gallic acid intake induces alterations to systems metabolism in rats, J. Proteome Res., 2013, 12, 991–1006.
48. V. R. Dharnidharka, C. Kwon and G. Stevens, Serum cystatin C is superior to serum creatinine as a marker of kidney function: A meta-analysis, Am. J. Kidney Dis., 2002, 40, 221–226.
49. A. D. Rule, T. S. Larson, E. J. Bergstralh, J. M. Slezak, S. J. Jacobsen and F. G. Cosio, Using serum creatinine to estimate glomerular filtration rate: accuracy in good health and in chronic kidney disease ann, Ann. Intern. Med., 2004, 141, 929–937.
50. M. Xiao, G. K. Du, G. B. Zhong, D. J. Yan, H. Z. Zeng and W. W. Cai, Gas chromatography/mass spectrometry-based metabolomic profiling reveals alterations in mouse plasma and liver in response to fava beans, PLoS One, 2016, 11, e0151103.
51. J. S. Wang, T. Reijmers, L. J. Chen, R. Van Der Heijden, M. Wang, S. Q. Peng, T. Hankemeier, G. W. Xu and J. Van Der Greef, Systems toxicology study of doxorubicin on rats using ultraperformance liquid chromatography coupled with mass spectrometry based metabolomics, Metabolomics, 2009, 5, 407–418.
52. X. L. Liu, L. B. Zhang, L. P. You, J. B. Yu, J. M. Zhao, L. Z. Li, Q. Wang, F. Li, C. H. Li, D. Y. Liu and H. F. Wu, Differential toxicological effects induced by mercury in gills from three pedigrees of Manila clam Ruditapes philippinarum by NMR-based metabolomics, Ecotoxicology, 2011, 20, 177–186.

---

Footnote

† Electronic supplementary information (ESI) available. See DOI: 10.1039/c6ra18600a

---