1H NMR based metabolomic profiling revealed doxorubicin-induced systematic alterations in a rat model

Journal of Pharmaceutical and Biomedical Analysis 118 (2016) 338–348

Contents lists available at ScienceDirect

Journal of Pharmaceutical and Biomedical Analysis journal homepage: www.elsevier.com/locate/jpba

1

H NMR based metabolomic profiling revealed doxorubicin-induced systematic alterations in a rat model Qian-Yun Niu a,b , Zhen-Yu Li a,∗ , Guan-Hua Du a,c , Xue-Mei Qin a,∗ a b c

Modern Research Center for Traditional Chinese Medicine of Shanxi University, No. 92, Wucheng Road, Taiyuan 030006, Shanxi, People’s Republic of China College of Chemistry and Chemical Engineering of Shanxi University, No. 92, Wucheng Road, Taiyuan 030006, Shanxi,People’s Republic of China Institute of Materia Medica, Chinese Academy of Medical Sciences, Beijing 100050, People’s Republic of China

a r t i c l e

i n f o

Article history: Received 13 August 2015 Received in revised form 1 October 2015 Accepted 19 October 2015 Available online 30 October 2015 Keywords: Doxorubicin Toxicity Metabolomic NMR Multivariate data analysis Rats

a b s t r a c t Doxorubicin (DOX) is used as a chemotherapy drug with severe carditoxicity. In this study, an integrated echocardiography along with pathological examination and 1 H NMR analysis of multiple biological matrices (urine, serum, heart, and kidney) was employed to systemically assess the toxicity of DOX. Echocardiographic results showed that impaired left ventricular contractility and degenerative pathology lesions in DOX group, which were in consistent with pathology. The endogenous metabolites in the urine, serum, heart and kidney was identified by comparison with the data from the literature and databases. Multivariate analysis, including PCA and OPLS, revealed 8 metabolites in urine, including succinate, 2-ketoglutarate, citrate, hippurate, methylamine, benzoate, allantion, and acetate were the potential changed biomarkers. In serum, perturbed metabolites include elevation of leucine, ␤-glucose, O-acetyl-glycoprotein, creatine, lysine, glycerin, dimethylglycine, trimethylamine-N-oxide, myo-inositol, and N-acetyl-glycoprotein, together with level decreases of acetone, lipid, lactate, glutamate, phosphocholine, acetoacetate and pyruvate. For heart, DOX exposure caused decline of lipid, lactate, leucine, alanine, glutamate, choline, xanthine, glycerin, carnitine, and fumarate, together with elevation of glutamine, creatine, inosine, taurine and malate. Metabolic changes of kidney were mainly involved in the accumulation of ␣-glucose, lactate, phosphocholine, betaine, threonine, choline, taurine, glycine, urea, hypoxanthine, glutamate, and nicotinamide, coupled with reduction of asparagine, valine, methionine, tyrosine, lysine, alanine, leucine, ornithine, creatine, lipid, and acetate. In addition, alterations of urinary metabolites exhibited a time-dependent manner. Complementary evidences by multiple matrices revealed disturbed pathways concerning energy metabolism, fatty acids oxidation, amino acids and purine metabolism, choline metabolism, and gut microbiota-related metabolism. In addition, the change of endogenous metabolites in rats urine, serum, heart and kidney were correlated with the echocardiography parameters. This integrative study should help to develop a systematic understanding of cardiomyopathy-related diseases and their metabolic events. © 2015 Elsevier B.V. All rights reserved.

1. Introduction Doxorubicin (DOX), a typical antineoplastic antibiotic, is widely used in the treatment of solid tumors and hematopoietic malignancies [1]. However, its clinical use is limited by the severe dose-dependent side effects, especially cardiomyopathy [2]. There are many hypothesis for the underlying mechanisms of DOXinduced cardiotoxicity, such as the formation of reactive oxygen species (ROS) [3], disturbance of cellular and mitochondria Ca2+ [4]. Several biochemical markers, such as cardiac troponin T [5],

∗ Corresponding authors. Fax: +86 351 7011202. E-mail addresses: [email protected] (Z.-Y. Li), [email protected] (X.-M. Qin). http://dx.doi.org/10.1016/j.jpba.2015.10.026 0731-7085/© 2015 Elsevier B.V. All rights reserved.

brain natriuretic peptide and pro-brain natriuretic peptide [6], are clinically utilized for early detection of myocardial injury. However, changes of these proteins only after cardiac damage had occurred. Besides the cardiotoxicity, long-term DOX treatment also caused renal [7], testes, and spleen damage [8]. in animal models by histopathological examination. Metabonomics is defined as “the quantitative measurement of the dynamic multiparametric metabolic responses of living systems to pathophysiological stimuli or genetic modification” [9]. It has been successfully applied in various areas, such as disease diagnosis, nutritional intervention, and drug toxicity [10,11]. Metabolomics, a non-targeted analysis strategy, is able to offer continuous and numerous information rapidly, which makes it a promising approach capable of not only identifying a large num-

Q.-Y. Niu et al. / Journal of Pharmaceutical and Biomedical Analysis 118 (2016) 338–348

ber of differential metabolites, but also revealing the changes of metabolites from a global perspective. Thus, it is extremely suitable for drugs which have various adverse effects like DOX. NMR is widely used in metabonomic analysis because it has advantages of high resolution, non-selectivity, simpleness in sample preparation, and rich structural information [9,12]. Recently, NMR based metabonomic approach has been successfully applied to monocrotaline-induced sinusoidal obstruction syndrome [13], PCBs and TCDD-induced hepatotoxicity [14], melamine-induced acute renal toxicity [15] and so on. Previous studies on the toxicity of DOX have been reported by analyzing the metabolic perturbations in urine [16–18], liver [19], and heart [20,21]. However, studies aimed at comprehensive and holistic understanding of DOX-induced toxicity in multiple biological matrices remains to be achieved, which is important for the elucidating of the pathogenic process and toxicological mechanism of DOX. In this study, we employed NMR based metabolomic profiling technique to investigate the adverse effects of DOX on the metabolic alterations of rat urine, serum, heart and kidney. Echocardiography and histopathology were also applied to find changes of cardiac geometry, systolic function and morphology. The correlations between endogenous metabolites and echocardiography parameters were investigated to find predictive information for DOX-induced toxicity. 2. Materials and methods 2.1. Chemicals Doxorubicin (DOX) was purchased from Haizheng Co., Ltd. (Zhejiang, China). Analytical grade K2 HPO4 ·3H2 O and NaH2 PO4 ·2H2 O were obtained from Guangfu Fine Chemical Research Institute (Tianjin, China) and Zhiyuan Chemical Reagent Co., Ltd. (Tianjin, China), respectively. Sodium 3-trimethlysilyl [2,2,3,3-d4] propionate (TSP) was from Cambridge Isotope Laboratories Inc (Andover, MA, USA). D2 O (99.9%) was procured from Norell (Landisville, Pennsylvania, USA). Phosphate buffer was prepared with K2 HPO4 and NaH2 PO4 (0.1 M, pH 7.4), containing 10% D2 O and 0.01% TSP. DOX was dissolved in sterile saline in dark immediately before use.

339

were observed daily for abnormal signs and weighed every three days during the course of the study. 2.4. Echocardiographic measurements 48 h after the last injection of DOX, rats were anesthetized with 4% chloral hydrate i.p (0.5 ml 100 g−1 ), and maintained on a heated plate. Transthoracic echocardiography was performed using a GE Vivid-7 pro ultrasonograph (GE Vingmed Ultrasound AS N-3191 horten, Norway) with an 10 MHz transducer. The following parameters, including interventricular septum at end-diastole (IVSd) and end-systole (IVSs), left ventricular posterior wall at end-diastole (LVPWd) and end-systole (LVPWs), left ventricular diameters at end diastole (LVIDd) and end-systole (LVIDs) were measured on left ventricular long-axis areas. The left ventricular ejection fraction (EF) and fractional shortening (FS) were calculated from the M-mode recordings. 2.5. Sample collection Following the echocardiographic examination, the blood samples were collected from femoral artery, centrifuged (15,100 × g, 4 ◦ C, 15 min), and the supernatants (serum) were stored at −80 ◦ C prior to use. The heart was immediately excised and weighted. The longitudinal section of the heart was fixed with 10% formalin for pathological examination, and the rest were stored at −80 ◦ C for the followed analysis. The liver, kidney, spleen and lung were also removed and weighted. 2.6. Histopathological assessments Morphological evaluation was conducted for all rats used in the experiment. Formalin-fixed tissues were embedded in paraffin wax and cut transversely into sections of 4–5 ␮m. These sections were then stained with routine hematoxylin-eosin (HE) staining and Masson’ s trichrome staining (Maixin Biotech. Co., Ltd. Fuzhou, China), followed by microscopic assessment. Pathological alterations were evaluated with the following grading scale: no obvious changes, -; minimal alterations, +; moderate alterations, ++; and severe alterations, +++.

2.2. Animals 2.7. Sample preparation for NMR measurements Twenty-six male Sprague Dawley (SD) rats weighing 180–220 g were purchased from Beijing Vital River Laboratories Co., Ltd. (SCXK (Jing) 2011-0012). Animals were acclimated to new environment for 7 days before experiment and feed at a temperature of 23 ± 1.5 ◦ C and a humidity of 45 ± 15% with a 12 h light/dark cycle. The animals had free access to water and food throughout the study. During urine and feces collection periods, rats were housed in metabolic cages. All animal procedures were approved by National Institute of Health guidelines of the Care and Use of Laboratory Animals, and performed strictly in accordance with the guidelines for Animal Experimentation of Shanxi University. 2.3. Experimental protocol Animal experiment was performed as previously described with minor modifications [22]. Briefly, the rats were randomly divided into 2 groups: DOX group (n = 16) and control group (n = 10). For DOX group, DOX was administered intraperitoneally (i.p.) at days 1, 3, 5, 7, 9, 11, 13, and 15 at the dosage of 1, 1, 2, 2, 3, 3, 4, and 4mg kg−1 , respectively. Control group was injected with equal volume of physiological saline the same day as DOX group did. Urine and feces samples were collected at days 0 (before administration), 7 and 15. Urine was centrifuged (4065 g, 4 ◦ C, 15 min), and the supernatants were stored at −80 ◦ C prior to use. The animals

Urine samples were thawed prior to use. A total of 500 ␮L urine of each sample was diluted with 200 ␮L of phosphate buffer, containing D2 O for the purpose of field lock, and TSP as a chemical shift reference. The whole mixture was centrifuged (4 ◦ C, 20 min) to remove the precipitates. The supernatants of 600 ␮L were transferred into 5 mm NMR tubes for 1 H NMR analysis. 450 ␮L of serum samples were mixed with 350 ␮L D2 O, and after centrifugation (15,100 × g, 4 ◦ C, 20 min), 600 ␮L of the supernatants were transferred into 5 mm NMR tubes for 1 H NMR analysis. Organ tissues were extracted as previously described with minor modifications [14]. Briefly, heart and kidney tissues (about 200 mg) were respectively extracted with 1000 ␮L acetonitrile/H2 O (1:1) by a ultrasonic cell homogenizer. After centrifugation (15,100 × g, 4 ◦ C, 20 min), the supernatants were transferred into 5 mL eppendorf (EP) collection tubes and dried under a gentle stream nitrogen. Then the samples were redissolved with 700 ␮L of phosphate buffer. Following final centrifugation (15,100 × g, 4 ◦ C, 20 min), 600 ␮L of supernatants were transferred into 5 mm NMR tubes for 1 H NMR analysis. 2.8. NMR measurement The 1 H NMR spectra of the urine, serum, heart and kidney were recorded at 25 ◦ C on a Bruker 600-MHz NMR spec-

340

Q.-Y. Niu et al. / Journal of Pharmaceutical and Biomedical Analysis 118 (2016) 338–348

Table 1 Effects of DOX treatment on organ weights and organ indexes of rats. Organ

Heart Liver Spleen Lung Kidney

Control

DOX

Organ weight

Organ index/%

1.18 ± 0.05 10.10 ± 0.40 0.75 ± 0.05 1.39 ± 0.06 2.38 ± 0.03

0.38 3.23 0.24 0.45 0.76

± ± ± ± ±

Organ weight

0.01 0.09 0.02 0.01 0.01

0.75 7.77 0.15 1.04 1.82

± ± ± ± ±

Organ index/%

0.05 0.31 0.01 0.04 0.07

0.31 3.17 0.06 0.42 0.74

± ± ± ± ±

0.02* 0.09 0.01* 0.01 0.02

Organ index is a ratio of organ weight to body weight. Values are presented as mean ± SEM.

trometer (600.13 MHz for proton frequency, Bruker, Germany). Serum samples were analyzed using one-dimensional Carr-PurcellMeiboom-Gill (CPMG) NMR spectra, while Nuclear Overhauser Effect Spectroscopy (NOESY, recycle delay (RD)-90◦ –t1 -90◦ –tm 90◦ -acquire) was used for urine, heart, and kidney with water suppression. 2.9. NMR data processing and multivariate statistical data analysis All 1 H NMR spectra of urine, serum, heart and kidney were manually phase- and baseline-corrected using MestReNova software (version 8.0.1, Mestrelab Research, Santiago de Compostella, Spain). The spectral regions of ı 0.5–8.5 for serum, ı 0.5–9.0 for heart and ı 0.5–9.65 for kidney were integrated into equal width of 0.01 ppm, while the spectra of urine were divided and the signal integral computed in 0.04 ppm intervals across the region ı 0.5–9.4. The regions contained residual water signals (ı 4.66–5.06 for urine, ı 4.66–5.10 for serum, ı 4.68–5.06 and ı 4.60–5.18 for heart and kidney tissues, respectively) and urea resonance (ı 5.46–6.24 for urine) were removed prior to data normalization. TSP with a chemical shift at ı 0.00 was used as a spectral reference for urine, heart and kidney, but to the methyl signal of creatine at ı 3.04 for serum samples. The remaining spectra were normalized to the tissue weights, and the total sum of integral area for urine and serum prior to subsequent data analysis. The normalized integral values were then mean-centered and subjected to multivariate pattern recognition analysis using the SIMCA-P 13.0 software package (Umetrics, Umeå, Sweden). The unsupervised principal component analysis (PCA) was used to see the distribution of control and DOX groups and find possible outliers. The supervised partial least squares discriminant analysis (PLS-DA) and orthogonal projections to the latent structures discriminant analysis (OPLS-DA) were then performed to better discrimination between control and DOX groups. The validity of the model was performed with SIMCA-P software by permutation tests (200 permutations), a 7-fold cross-validation method, and CV-ANOVA method. In addition, PLS-loading biplot was applied to probe the relationship between metabolic alterations and cardiomyopathy. All results were presented as mean ± standard error of the mean (SEM). Statistical analysis was performed using One-way Analysis of Variance (ANOVA) model. A calculated p value of less than 0.05 was considered to be statistically significant.

Table 2 Effects of DOX treatment on echocardiography parameters. Parametersa

Control

IVSd (mm) IVSs (mm) LVIDd (mm) LVIDs (mm) LVPWd (mm) LVPWs (mm) EF FS

1.52 2.90 5.36 2.39 2.30 3.15 0.90 0.56

± ± ± ± ± ± ± ±

DOX 0.09 0.11 0.21 0.20 0.13 0.16 0.01 0.02

1.40 2.12 5.54 3.55 1.48 2.27 0.71 0.36

pValue ± ± ± ± ± ± ± ±

0.08 0.11 0.16 0.15 0.07 0.09 0.03 0.02

0.32 6.03 × 10−5 0.51 2.21 × 10−4 4.80 × 10−5 2.76 × 10−4 1.41 × 10−5 7.00 × 10−6

a Parameters: IVSd/IVSs, interventricular septum at end-diastole/end-systole; LVIDd/LVIDs, left ventricular diameters at end diastole/end-systole; LVPWd/LVPWs, left ventricular posterior wall at end-diastole/end-systole; EF, ejection fraction; FS, fractional shortening. Values are presented as mean ± SEM. p < 0.05 represented statistically significant.

increase in control group (about 30%) during the experiment. However, body weights of DOX group were gradually increased and peaked at day 6, then decreased rapidly at days 7–15, suggesting the severe gastrointestinal side effects of DOX. 3.2. Changes in organ index Average organ weights and organ indexes of the heart, liver, spleen, lung and kidney are listed in Table 1. Organ index was computed by a ratio of organ weight to body weight. Severe atrophy of heart and spleen were observed in DOX group. The heart index of DOX group differed significantly from that of the normal controls (p < 0.05) and spleen index was reduced by about 4 fold in DOX group. Although organ weights of liver, lung, and kidney of DOXtreated rats were smaller than the controls, no significant changes were observed for these organ indexes. 3.3. Echocardiography

3. Results

Fig. S2 illustrates the typical two-dimensional and M-mode short-axis of the left ventricle obtained from the two groups, respectively. As shown in Table 2, compared with the controls, LVIDs increased significantly, whereas IVSd, LVPWd, and LVPWs reduced remarkably in DOX group (p < 0.05). And no significant changes were observed for LVIDd and IVSd. In addition, DOX treatment also caused significant reduction of EF and FS (p < 0.05), which are the main indicators of left ventricular systolic function in clinic. These above results indicated the occurrence of ventricular remodeling.

3.1. Mortality and body weight

3.4. Histopathology

Compared with the controls, rats in DOX group exhibited reduced food intake and physical activities (data not shown). Obvious feces alterations suggested DOX induced severe diarrhea. By the end of DOX injection, the mortality was 12.5% in DOX group, while no death was observed in control group. Average body weights of the two groups are shown in Fig. S1. There was a stable body weight

Fig. 1 depicts the results of light microscopy micrographs of HE and Masson’s staining in heart tissue of the two groups. It could be seen from the results of HE staining, morphology of the cardiac tissue was normal in control group, whereas DOX-treated rats showed mild to moderate level of cytoplasmic vacuolization, sarcoplasmic dissolution, inflammatory cell infiltration, and spotty

Q.-Y. Niu et al. / Journal of Pharmaceutical and Biomedical Analysis 118 (2016) 338–348

341

Fig. 1. Histopathological analysis of HE (A and B) and Masson’s trichrome staining (C and D) of rat heart tissue. A and C, control rats; B and D, DOX-treated rats.

necrosis. Masson’s trichrome staining of 5-mm sections showed no abnormality in heart tissue of control rats. However, histopathologic changes related to collagen fiber deposit around small vessels and in myocardial extracellular matrix were observed in the DOX group (Table S1). 3.5. Metabolomics Figs. 2 and 3 the representative 600 MHz 1 H NMR spectra for urine, serum, heart and kidney from control and DOX groups, respectively. NMR signals were identified to specific metabolites based on literature data [23], HMDB (http://www.hmdb.ca/) BMRB (http://www.bmrb.wisc.edu/) and our in house database [24–26]. Identified endogenous metabolites in urine, serum, heart and kidney are listed in Table S2. The detected metabolites in urine, serum, heart and kidney included organic acids, amino acids, glycolysis and tricarboxylic acid cycle (TCA) intermediates, urea cycle intermediates, choline metabolites, organic bases, purines and a number of other metabolites. Visual comparison revealed significant differences between control and DOX groups. For example, the DOX-treated rats showed lower level of urinary citrate, higher level of ␤-glucose in serum, more inosine and urea in heart and renal extracts. In order to extract more details about DOXinduced metabolic changes and identify the potential metabolites contributing to the separation, multivariate data analysis was performed on the metabolic profiles of urine, serum, heart and kidney samples at day 15. 3.6. DOX-induced metabolomic changes in multiple biological matrices Significant differences between control and DOX groups were observed in PCA score plots (Figs. 4A and S3A–S5A ) of urine, serum heart and kidney, indicating the rats metabolic profiles were markedly altered following DOX treatment. Then partial least squares discriminant analysis (PLS-DA) model was further con-

structed and validated using the response of the permutation test through 200 permutations. And the model of which Q2 line intercepts the Y axis at a negative value or the Q2 -values obtained from the permutation model to the left are lower than the original points to the right is deemed to be of great predictive ability and reliability (Figs. 4B and S3B–S5B). The corresponding OPLS-DA was then used to determine the potential biomarkers contributing to the separation (Figs. 4C and S3C–S5C), and the mode quality parameters described by R2 X, R2 Y and Q2 were listed in Table S3. The OPLS-DA S-plot (Figs. 4D and S3D–S5D) combined with variable importance in the projection (VIP) revealed that, compared with the normal controls, the major urinary variations after DOX treatment were decline of succinate, 2-ketoglutarate, citrate, hippurate, methylamine, benzoate, and allantion, as well as elevation of acetate (Table 3). In serum, perturbed metabolites included elevation of leucine, ␤-glucose, O-acetyl-glycoprotein (OAG), creatine, lysine, glycerin, dimethylglycine, trimethylamine-N-oxide (TMAO), myo-inositol, and N-acetyl-glycoprotein (NAG), together with level decreases of acetone, lipid, lactate, glutamate, phosphocholine, acetoacetate and pyruvate. For heart tissue, DOX exposure caused decline of lipid, lactate, leucine, alanine, glutamate, choline, xanthine, glycerin, carnitine, and fumarate, together with elevation of glutamine, creatine, inosine, taurine and malate. Metabolic changes contributing to the separation of kidney were mainly involved in the accumulation of ␣-glucose, lactate, phosphocholine, betaine, threonine, choline, taurine, glycine, urea, hypoxanthine, glutamate, and nicotinamide, coupled with excessive reduction of asparagine, valine, methionine, tyrosine, lysine, alanine, leucine, ornithine, creatine, lipid, and acetate in DOX group. The normalized integral values of metabolites in urine, serum, heart and kidney from DOX and control rats were also compared (Tables S4–S7). Spectral signals, which are overlapped or in a minimal amount, were carefully removed (three in urine). The variations of metabolites levels were in agreement with the results of multivariate statistical analysis.

342

Q.-Y. Niu et al. / Journal of Pharmaceutical and Biomedical Analysis 118 (2016) 338–348

Fig. 2. Typical 600 MHz 1 H NMR spectra of urine from control (UC) and DOX-treated (UD) rats; CPMG NMR spectra of serum from control (SC) and DOX-treated (SD) rats.

Fig. 3. Typical 600 MHz 1 H NMR spectra of heart from control (HC) and DOX-treated (HD) rats; NMR spectra of kidney from control (KC) and DOX-treated (KD) rats.

3.7. Dynamic alterations of urinary metabolites The score plot of unsupervised PCA (Fig. 5A) showed a clear separation of urine samples within the DOX group at day 0, 7 and 15, in which PC1 accounted for 32.0% of the total variance, and PC2 accounted for 13.2% of the total variance. OPLS-DA score plot (Fig. 5B) was used to maximum discrimination among the 3 groups (R2 X = 0.681, R2 Y = 1 and Q2 = 0.833) and CV-ANOVA confirmed the validation the model (p < 0.05). Urine samples at day 0 were clustered together, whereas the samples at day 7 and 15 were

scattered, indicating individual differences to drug effects. From day 0 to day 15, the metabolic state of rats in DOX group gradually moved further away from the normal controls, which suggested the occurrence of metabolic disorder with time and dose increasing. To further assess changes during the development of DOX induced toxicity, urine samples at day 0, 7, and 15 were also compared and visualized as a heatmap image, except for some serious overlap signals (Fig. 6). The warm color (e.g., red) indicated the contents of metabolites were higher than the cold color (e.g., blue). In the dynamic process of myocardial injury, a time-

Q.-Y. Niu et al. / Journal of Pharmaceutical and Biomedical Analysis 118 (2016) 338–348

343

Fig. 4. PCA score plot (A), PLS-DA validation plot (B), OPLS-DA score plot (C) and corresponding S-plot (D) of urine from control and DOX groups at day 15.

Fig. 5. PCA score plot (A) and OPLS-DA score plot (B) of urine within DOX group at day 0 (blue),7 (red) and 15 (black).(For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)

dependent elevation of leucine, isoleucine, alanine, acetate, malate, fucose, U1 and U2, along with reduction of citrate, dimethylamine, fumarate, hippurate, trigonelline, and benzoate were evident in DOX-induced rat urine. The trends of N-acetylglutamate, allantion, lactate, formate, TMAO, 2-hydroxyisobutyrate, butyrate, trimethylamine, and 3-hydroxybutyrate (3-HB) were increased during the onset and decreased during the development of cardiac injury, suggesting that there might be some adaptive response in the evolution of DOX toxicity. In addition, compared with day 0, contents of N-methylnicotinamide and creatinine were increased at day 7, however, no significant changes were observed between day 15 and day 7. 3.8. Correlation analysis between metabolites and myocardial function To further investigate the correlation between echocardiography results and endogenous metabolites, PLS loading-biplot

was performed between eight echocardiography parameters and endogenous metabolites in urine, serum, heart and kidney extracts. The datasets were echocardiography parameters used as y variables and metabolite contents as x variables. As shown in Fig. 7, the loading-biplot for the urine samples suggested that LVIDs were positively correlated with acetate (ı1.91) and alanine (ı1.48) on the first negative component. Citrate (ı2.69), benzoate (ı7.88), hippurate (ı7.55), dimethylamine (ı2.72), fumarate (ı6.53), and trigonelline (ı9.13) showed a positive correlation with the other echocardiography parameters (IVSs, LVPWd, LVPWs, EF, and FS). In Fig. S6, a number of metabolites in serum, including 3-HB (ı1.20), glycine (ı3.56), lipid (ı0.87), glutamate (ı2.35), acetoacetate (ı2.27), and acetone (ı2.22), were clustered together with IVSs, LVPWd, LVPWs, EF, and FS, indicated that they were highly correlated. On the negative side of the first component, LVIDs was closely correlated with creatine (ı3.93). Similar results could also be found in heart and kidney extracts, as illustrated in Figs. S7 and S8. In heart tissue, IVSs, LVPWd, LVPWs, EF, and FS were

344

Q.-Y. Niu et al. / Journal of Pharmaceutical and Biomedical Analysis 118 (2016) 338–348

Table 3 OPLS-DA variable importance in projection (VIP) for perturbed endogenous metabolites in urine, serum, heart, and kidney. ı1 H

Metabolites

Urine

Serum

Heart

Kidney

VIP

VIP

VIP

VIP

Pathway

4.64 5.22 1.32 2.36

␤-Glucose a-Glucose Lactate Pyruvate

– – – –

2.56 – 5.48 1.16

– – 7.50 –

– 1.97 2.61 –

Glycolysis

4.32 6.53 2.41 2.45 2.52

Malate Fumarate Succinate 2-Ketoglutarate Citrate

– – 2.94 3.08 4.89

– – – – –

1.06 1.74 – – –

– – – – –

Tricarboxylic acid cycle

1.91 0.87 3.23 3.65 3.54

Acetate Lipid Carnitine Glycerin myo-Inositol

1.60 – – – –

– 2.06 – 1.47 2.60

– 1.32 4.76 1.72 –

1.55 1.07 – – –

Lipid metabolism

2.61 3.19 3.20 2.91 3.27 3.27

Methylamine Choline Phosphocholine Dimethylglycine Trimethylamine-N-oxide Betaine

1.05 – – – – –

– – 1.49 1.2 1.67 –

– 1.73 – – – –

– 7.46 4.25 – – 3.03

Methylamine metabolism

7.84 7.88

Hippurate Benzoate

1.41 1.22

– –

– –

– –

Gut microbiota-related metabolism

3.03

Creatine

–

1.54

6.54

1.22

Creatine metabolism

3.43 3.56

Taurine Glycine

– –

– –

5.67 –

4.14 5.20

Bile acid metabolism

5.36 7.92 8.20

Allantoin Xanthine Hypoxanthine

1.12 – –

– – –

– 1.02 –

– – 2.07

Purine metabolism

3.04 5.83 2.04

Ornithine Urea N-acetyl-glycoprotein

– – –

– – 2.00

– – –

1.45 2.00 –

Urea cycle

0.96 1.48 2.45 2.35 1.72 7.19 2.63 1.04 2.96 3.59

Leucine Alanine Glutamine Glutamate Lysine Tyrosine Methionine Valine Asparagine Threonine

– – – – – – – – – –

1.02 – – 1.34 1.38 – – – – –

1.02 3.59 3.68 1.32 – – – – – –

2.81 2.61 – 100 2.27 1.01 1.28 1.61 1.09 1.62

Amino acid metabolism

2.22 2.27

Acetone Acetoacetate

– –

1.92 1.58

– –

– –

Ketogenesis

2.13

O-acetyl-glycoprotein

–

3.40

–

–

Glutamine metabolism

7.60

Nicotinamide

–

–

–

1.09

Nicotinamide metabolism

8.35

Inosine

–

–

1.04

–

Nucleic acid metabolism

strongly correlated with carnitine (ı3.23), fumarate (ı6.53), alanine (ı1.48), glycerin (ı3.65), and 3-HB (ı1.20) on the negative side of the second component, while LVIDs was significantly correlated with taurine (ı3.27) on the positive side of the second component in DOX group. For kidney, IVSs, LVPWd, LVPWs, EF, and FS were closely correlated with methionine (ı2.63), lysine (ı1.72), alanine (ı1.48), valine (ı1.04), tryptophan (ı7.29), and leucine (ı0.96) on the first negative component. LVIDs was positively correlated with cytidine (ı6.07) and LVIDd was positively correlated with formate (ı8.45) on the first positive component, respectively. It is worth noting that the changes of above-mentioned endogenous metabolites were correlated with echocardiography parameters in DOX induced cardiotoxicity, and most of them were the same as those of the differential metabolites between control and DOX groups.

4. Discussion During the course of the study, there was nearly no body weight gaining in DOX-treated rats, whereas the average body weight increase was 30% in control rats. One of the common side effects of DOX is gastrointestinal disturbance in cancer patients [27], and DOX-induced gastrointestinal toxicity has also been reported in animal models [8]. Our investigation suggested that gut microbiota functions were affected by DOX administration as well (changes in metabolites levels). The above changes might contribute to body weight decreasement of DOX group. Among various organs examined, an obvious change was severe atrophy in spleen. One of the primary physiological roles of spleen is hematopoiesis function. Significant spleen alterations induced by DOX might be a response to drug-induced hematologic toxicity.

Q.-Y. Niu et al. / Journal of Pharmaceutical and Biomedical Analysis 118 (2016) 338–348

345

Fig. 6. Dynamic alterations of urinary metabolites during the treatment of DOX. The color indicates the contents of metabolites. The warm color (e.g., red) indicates the contents of metabolites were higher than the cold color (e.g., blue).(For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)

Similar adverse effect of DOX on hematopoietic system was also observed in previous studies with decline in red blood cell count, hemoglobin counts, and hematocrit [8]. Systemic abnormalities in metabolites of multiple biological matrices provide some important information about the mechanisms of DOX toxicity. The changes in the metabolic pathway were mainly involved in energy metabolism, fatty acids ␤-oxidation, amino acids and purine metabolism, choline metabolism, and gut microbiota-related metabolism (Fig. 8).

4.1. Cardiotoxicity related metabolic changes The heart is a highly energy-demanded organ, and fatty acids, ketone bodies, glucose, amino acids could all be used to produce ATP to maintain its contractile function. Abnormalities in fatty acids metabolism are important events in the pathogenesis of myocardial cell injury, since fatty acids oxidation-based energy supply provide 60–90% of energy for a healthy heart. Carnitine is of vital importance in the fatty acid ␤-oxidation. Only with the assistance of carnitine, long-chain acetyl-CoA could come into mitochondria for subsequent ␤-oxidation [28]. Decreased level of carnitine demonstrated that the ability of ATP generation is suppressed in myocardia after DOX treatment. Disorders in fatty acids metabolism were also observed in urine, serum and kidney with marked changes of lipid, glycerin, myo-inositol, and acetate (Table 3). Acetate and acetoacetate, products of lipids in liver mitochondria, are mainly utilized in brain and heart tissues. They were reduced in serum in DOX group, providing evidence that ketone bodies metabolism based-energy generation was partly decreased in myocardial tissue. Glucose metabolism-based energy supply is another important

source for the heart pumping. In the present study, decreased levels of lactate, fumarate coupled with increased level of malate in heart tissue suggested that both aerobic and anaerobic were impaired in DOX-treated rats. In addition, a number of glycolysis and TCA related metabolites, such as pyruvate, succinate, and citrate, were also changed in biofluids (urine and serum) and kidney (Table 3). Leucine, glutamate and glutamine are important glucogenic amino acids, which can be converted to intermediates of glucose metabolism for energy production. DOX exposure lead to reduction of alanine, leucine and glutamate together with accumulated glutamine level in cardiac extracts, suggesting that gluconeogenesis was reduced in myocardial pathological state. The above-mentioned evidences indicated that DOX treatment lead to inhibition in energy supply and finally resulted in myocardial lesions, which were confirmed by heart echocardiography and histopathology results in this study. Taurine is a ubiquitous free amino acid present in many tissues which is involved in numerous physiological processes, such as osmoregulation, antioxidant activity, and hepatic detoxification [29]. Taurine is major biosynthesized in liver and is more abundant in cardiac than other tissues [30]. There is evidence that reduction in cardiac taurine level can lead to severe cardiomyopathy [31]. In this study, the level of cardia taurine was increased after DOX treatment, which is in agreement with a previous study [21]. However, it needs further investigations to uncover the possible reasons.

4.2. Nephrotoxicity related metabolic perturbations Renal damage is a common adverse effect of DOX therapy, and many small animal models of nephropathy were induced by DOX

346

Q.-Y. Niu et al. / Journal of Pharmaceutical and Biomedical Analysis 118 (2016) 338–348

Fig. 7. The PLS loading bipiot of between urinary metabolites and echocardiography parameters at day 15. The pc(corr) and t (corr) vectors are displayed for the first and second components.

Fig. 8. Summary of pathway alterarions after DOX treatment in rat urine, serum, heart, and kidney.

administration [32,33]. In the present work, we observed altered renal amino acids profiles in DOX group, including increased levels of glycine and glutamate, as well as decreased levels of leucine, alanine, lysine, tyrosine, methionine, and asparagine. Amino acids are basic units for protein synthesis in organism. Increasing evidence

showed that renal injury is highly associated with abnormality in protein expression and amino acids reabsorption [33]. In this study, the abnormalities in amino acids metabolism might be correlated with DOX induced nephrotoxicity.

Q.-Y. Niu et al. / Journal of Pharmaceutical and Biomedical Analysis 118 (2016) 338–348

Inosine can be transformed into hypoxanthine in rats, which can be further converted into xanthine by xanthine oxidase. Uric acid is formed by oxidation of xanthine and then oxidized to allantion by urate oxidase. The concurrent increasement of renal hypoxanthine and decreasement of urinary allantion and cardia xanthine suggested disturbance of purine metabolism after DOX administration. Allantoin represents as a biomarker for oxidative stress, which is closely related to renal damage [34]. Decreased urine allantoin content in DOX-induced rats implying that renal function is affected. There are three different pathways involved in choline metabolism, including phosphorylated with choline kinase to generate phosphocholine; breaking down to amines (methylamine, dimethylamine, and trimethylamine) by gut microflora; and oxidized to betaine by a two-step reaction and finally lead to the generation of creatine. Choline and phosphocholine are essential elements of cell membrane, and changes in their levels are related to cell membrane damage. Elevation of renal choline and phosphocholine in DOX group indicated that the renal functions were impaired. Betaine acts as an osmoregulatory compound which is highly associated with oxidative stress [34]. Increased renal betaine content revealed that DOX-induced oxidative stress had occurred, which is in agreement with the notion that oxygen free radicals play an important role in renal pathogenic state. Hippurate, formed by the conjugation of benzoate with glycine, is usually detected in urine, and its concentration is related to the microbial activity. Reduced urinary levels of hippurate and benzoate in DOX group suggested alteration of gut microbiota, which might affect the absorption of nutrients, and thus result in limited body weight increase in DOX group. Urinary hippurate has also been regarded as a “window” of renal function [18]. Declined concentration of urinary hippurate in DOX group suggested the injury of kidney, which is in agreement with previous reports [34]. Urinary acetate and citrate are biomarkers of renal injury. There is mounting evidence that contents of acetate and citrate are highly related to renal metabolism [35]. The concurrent elevation of acetate content and depletion of citrate in rats urine suggested that renal function is impaired after DOX treatment. Thus, the above evidence showed that DOX exposure also caused severe nephrotoxicity which resulted in alternations of amino acids metabolism, purine metabolism, choline metabolism, and gut microbiota-related metabolism.

5. Conclusion In this study, an integrated NMR analysis of multiple biological matrices along with pathological examination and echocardiography were employed to systemically assess the toxicity of DOX. To the best of our knowledge, DOX-induced metabolite alternations in rats serum and kidney were studied for the first time. Metabolomic analysis of urine, serum, heart and kidney revealed the complementary evidence of alterations in energy metabolic pathways, fatty acids ␤-oxidation, amino acids and purine metabolism, choline metabolism, and gut microbiota-related metabolism after DOX treatment. These perturbed metabolic changes are helpful in the explanation of mechanism of systematical toxicity induced by DOX. The clinical use of DOX is limited by the severe dose-dependent side effects, and the determination of sensitive biomarkers to predict DOX-induced toxicity is of vital importance. In this study, urinary metabolites such as acetate, citrate and allantion altered in a time dependent manner, which could provide some predictive information of cardiac injury. Furthermore, the changes of endogenous metabolites in the rats urine, serum, heart and kidney were correlated with the echocardiography parameters. Since the urine and serum samples are easy to be obtained, further studies should

347

be conducted to find sensitive and accurate biomarkers for early prediction of DOX-induced toxicity in clinic. In addition, investigations aimed at assessing metabolic changes of other organs should also be conducted to fully understand the systematic toxicity of DOX. Acknowledgements This study was financially supported by Program for the Outstanding Innovative Teams of Higher Learning Institutions of Shanxi (OIT), the National Twelfth Five-year Science and Technology Support Program (No. 2011BA107B01), and Science and technology innovation team of Shanxi Province (No. 2013131015). Appendix A. Supplementary data Supplementary data associated with this article can be found, in the online version, at http://dx.doi.org/10.1016/j.jpba.2015.10.026. References [1] M. Honda, A. Miura, Y. Izumi, T. Kato, T. Ryotokuji, K. Monma, J. Fujiwara, H. Egashira, T. Nemoto, Doxorubicin, cisplatin, and fluorouracil combination therapy for metastatic esophageal squamous cell carcinoma, Dis. Esophagus 23 (2010) 641–645. [2] Y.W. Zhang, J.J. Shi, Y.J. Li, L. Wei, Cardiomyocyte death in doxorubicin-induced cardiotoxicity arch immunol, Arch. Immunol. Ther. Ex. 57 (2009) 435–445. [3] D. Lebrecht, A. Geist, U.P. Ketelsen, J. Haberstroh, B. Setzer, U.A. Walker, Dexrazoxane prevents doxorubicin-induced long-term cardiotoxicity and protects myocardial mitochondria from genetic and functional lesions in rats, Brit. J. Pharmacol. 151 (2007) 771–778. [4] L.E. Solem, T.R. Henry, K.B. Wallace, Disruption of mitochondrial calcium homeostasis following chronic doxorubicin administration, Toxicol. Appl. Pharm. 129 (1994) 214–222. [5] K.B. Wallace, E. Hausner, E. Herman, G.D. Holt, J.T. Macgregor, A.L. Metz, E. Murphy, I.Y. Rosenblum, F.D. Sistare, M.J. York, Serum troponins as biomarkers of drug-induced cardiac toxicity, Toxicol. Pathol. 32 (2004) 106–121. [6] K.J.M. Schimmel, D.J. Richel, R. van den Brink, H.J. Guchelaar, Cardiotoxicity of cytotoxic drugs, Cancer Treat. Rev. 30 (2004) 181–191. [7] Z.H. Ding, L.M. Xu, S.Z. Wang, J.Q. Kou, Y.L. Xu, C.X. Chen, H.P. Yu, Z.H. Qin, Y. Xie, Ameliorating adriamycin-induced chronic kidney disease in rats by orally administrated cardiotoxin from naja naja atra venom, evid-based, Compl. Alt. Med. (2014), http://dx.doi.org/10.1155/2014/621756. [8] V.G. Desai, E.H. Herman, C.L. Moland, W.S. Branham, S.M. Lewis, K.J. Davis, N.I. George, T. Lee, S. Kerr, J.C. Fuscoe, Development of doxorubicin-induced chronic cardiotoxicity in the B6C3F1 mouse model, Toxicol. Appl. Pharm. 266 (2013) 109–121. [9] J.K. Nicholson, J.C. Lindon, E. Holmes, ‘Metabonomics’: understanding the metabolic responses of living systems to pathophysiological stimuli via multivariate statistical analysis of biological NMR spectroscopic data, Xenobiotica 29 (1999) 1181–1189. [10] C.F. Diao, L.C. Zhao, M.M. Guan, Y.Q. Zheng, M.J. Chen, Y.J. Yang, L. Lin, W.J. Chen, H.C. Gao, Systemic and characteristic metabolites in the serum of streptozotocin-induced diabetic rats at different stages as revealed by a 1 H NMR based metabonomic approach, Mol. BioSyst. 10 (2014) 686–693. [11] D.M. Jacobs, N. Deltimple, E. van Velzen, F.A. van Dorsten, M. Bingham, E.E. Vaughan, J. Van Duynhoven, 1 H NMR metabolite profiling of feces as a tool to assess the impact of nutrition on the human microbiome, NMR Biomed. 21 (2008) 615–626. [12] J.C. Lindon, J.K. Nicholson, E. Holmes, J.R. Everett, Metabonomics: metabolic processes studied by NMR spectroscopy of biofluids, Concepts Magn. Reson. 12 (2000) 289–320. [13] R. Conotte, J.M. Colet, A metabonomic evaluation of the monocrotaline-induced sinusoidal obstruction syndrome (SOS) in rats, Toxicol. Appl. Pharm. 276 (2014) 147–156. [14] C. Lu, Y. Wang, Z. Sheng, G. Liu, Z. Fu, J. Zhao, J. Zhao, X. Yan, B. Zhu, S. Peng, NMR-based metabonomic analysis of the hepatotoxicity induced by combined exposure to PCBs and TCDD in rats, Toxical. Appl. Pharm 248 (2010) 178–184. [15] 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, W. Jia, Metabonomic evaluation of melamine-induced acute renal toxicity in rats, J. Proteome Res. 9 (2010) 125–133. [16] K.B. Kim, J.Y. Yang, S.J. Kwack, H.S. Kim, D.H. Ryu, Y.J. Kim, J.Y. Bae, D.S. Lim, S.M. Choi, M.J. Kwon, D.Y. Bang, S.K. Lim, Y.W. Kim, G.S. Hwang, B.M. Lee, Potential metabolomic biomarkers for evaluation of adriamycin efficacy using a urinary 1 H NMR spectroscopy, J. Appl. Toxicol. 33 (2013) 1251–1259. [17] J.C. Park, Y.S. Hong, Y.J. Kim, J.Y. Yang, E.Y. Kim, S.J. Kwack, D.H. Ryu, G.S. Hwang, B.M. Lee, A metabonomic study on the biochemical effects of

348

[18]

[19]

[20]

[21]

[22]

[23] [24]

[25] [26]

Q.-Y. Niu et al. / Journal of Pharmaceutical and Biomedical Analysis 118 (2016) 338–348 doxorubicin in rats using 1 H NMR spectroscopy, J. Toxicol. Environ. Part A 72 (2009) 374–384. S. Halligan, S.J. Byard, A.J. Spencer, T.J. Gray, E.S. Harpur, F.W. Bonner, A study of the nephrotoxicity of three cephalosporins in rabbits using 1 H NMR spectroscopy, Toxicol. Lett. 81 (1995) 15–21. K.E. Price, C.K. Larive, C.E. Lunte, Tissue-targeted metabonomics: biological considerations and application to doxorubicin-induced hepatic oxidative stress, Metabolomics 5 (2009) 219–228. I. Andreadou, M. Papaefthimiou, A. Zira, M. Constantinou, F. Sigala, A.L. Skaltsounis, A. Tsantili-Kakoulidou, E.K. Iliodromitis, D.T. Kemastinos, E. Mikros, Metabonomic identification of novel biomarkers in doxorubicin cardiotoxicity and protective effect of the natural antioxidant oleuropein, NMR Biomed. 22 (2009) 585–592. I. Andreadou, E. Mikros, K. Ioannidis, F. Sigala, N. Katerina, S. Kostidis, D. Farmakis, R. Tenta, N. Kavantzas, S.I. Bibli, E. Gikas, L. Skaltsounis, D.T. Kremastinos, E.K. Iliodromitis, Oleuropein prevents doxorubicin-induced cardiomyopathy interfering with signaling molecules and cardiomyocyte metabolism, J. Mol. Cell. Cardiol. 69 (2014) 4–16. X.M. Li, G.Z. Tian, Y.T. Ou, H.J. Song, Y. Zhao, P. Wang, The establishment of a chronic heart failure model by adrinmycin and the judgement of the indexes, Anat. Res. 27 (2005) 176–178. X.H. Shi, C.N. Xiao, Y.L. Wang, H.R. Tang, Gallic acid intake induces alterations to systems metabolism in rats, J. Proteome Res. 12 (2013) 991–1006. A.P. Li, Z.Y. Li, H.F. Sun, K. Li, X.M. Qin, G.H. Du, Comparison of two different astragali radix by a 1 H NMR-Based metabolomic approach, J. Proteome Res. 14 (2015) 2005–2016. Z.Y. Li, P. He, H.F. Sun, X.M. Qin, G.H. Du, 1 H NMR based metabolomic study of the antifatigue effect of astragali radix, Mol. BioSyst. 10 (2014) 3022–3030. J.S. Tian, B.Y. Shi, H. Xiang, S. Gao, X.M. Qin, G.H. Du, 1 H NMR-based metabonomic studies on the anti-depressant effect of genipin in the chronic unpredictable mild stress rat model, PLoS One 8 (2013) e75721.

[27] S. Boussios, G. Pentheroudakis, K. Katsanos, N. Pavlidis, Systemic treatment-induced gastrointestinal toxicity: incidence, clinical presentation and management, Ann. Gastroenterol. 25 (2012) 106–118. [28] F. Zhang, Q. Zhan, X. Dong, B. Jiang, L.N. Sun, S.H. Gao, Z.Q. He, X. Tao, W.S. Chen, Shengxian decoction in chronic heart failure treatment and synergistic property of platycodonis radix: a metabolomic approach and its application, Mol. BioSyst. 10 (2014) 2055–2063. [29] R.J. Huxtable, Physiological actions of taurine, Physiol. Rev. 72 (1992) 101–163. [30] T. Ito, S. Schaffer, J. Azuma, The effect of taurine on chronic heart failure: actions of taurine against catecholamine and angiotensin II, Amino Acids 46 (2014) 111–119. [31] T. Ito, J. Azuma, Taurine depletion-related cardiomyopathy in animals, Cardiomyopathies-From Basic Res. Clin. Manage. 10 (2012) 537–552. [32] T. Bertani, A. Poggi, R. Pozzoni, F. Delaini, G. Sacchi, Y. Thoua, G. Mecca, G. Remuzzi, M.B. Donati, Adriamycin-induced nephrotic syndrome in rats: sequence of pathologic events, Lab. Invest. 46 (1982) 16–23. [33] C.L. Liang, J.B. Wu, J.M. Lai, S.F. Ye, J. Lin, H. Ouyang, J.Y. Zhan, J.Y. Zhou, Protection effect of Zhen-Wu-Tang on adriamycin-induced nephrotic syndrome via inhibiting oxidative lesions and inflammation damage, Evid-Based. Compl. Alt. (2014), http://dx.doi.org/10.1155/2014/131604. [34] J. Wang, T. Reijmers, L. Chen, R. Van Der Heijden, M. Wang, S. Peng, J. Van Der Greef, Systems toxicology study of doxorubicin on rats using ultra performance liquid chromatography coupled with mass spectrometry based metabolomics, Metabolomics 5 (2009) 407–418. [35] H.F. Wang, J. Bai, G. Chen, W. Li, R.W. Xiang, G.Y. Su, Y.H. Pei, A metabolic profiling analysis of the acute hepatotoxicity and nephrotoxicity of Zhusha Anshen Wan compared with cinnabar in rats using 1 H NMR spectroscopy, J. Ethnopharmacol. 146 (2013) 572–580.