Metabolomics research on the hepatoprotective effect of Angelica sinensis polysaccharides through gas chromatography–mass spectrometry

Journal of Chromatography B, 973 (2014) 45–54

Contents lists available at ScienceDirect

Journal of Chromatography B journal homepage: www.elsevier.com/locate/chromb

Metabolomics research on the hepatoprotective effect of Angelica sinensis polysaccharides through gas chromatography–mass spectrometry Peng Ji, Yanming Wei ∗ , Hongguo Sun, Wenxin Xue, Yongli Hua, Pengling Li, Wenquan Zhang, Ling Zhang, Haifu Zhao, Jinxia Li Institute of Traditional Chinese Veterinary Medicine, College of Veterinary Medicine, Gansu Agricultural University, Lanzhou 730070, PR China

a r t i c l e

i n f o

Article history: Received 3 April 2014 Accepted 6 October 2014 Available online 18 October 2014 Keywords: Metabolomics Angelica sinensis polysaccharide Hepatoprotective effect Gas chromatography–mass spectrometry Carbon tetrachloride

a b s t r a c t Angelica sinensis polysaccharides (ASP) have an established hepatoprotective effect, but the mechanism for this effect remains unclear. A novel approach using biochemical parameters coupled with metabolomics based on gas chromatography–mass spectrometry (GC–MS) and chemometrics was established in this study to explain the hepatoprotective effect mechanism of ASP. The superoxide dismutase activity, malonaldehyde content, alanine aminotransferase, aspartate aminotransferase, and ␥-glutamyl transpeptidase in plasma were measured. Pathological changes in the liver were observed. Plasma and liver homogenate obtained from mice were analyzed using GC–MS. Distinct changes in metabolite patterns in the plasma and liver homogenate after being induced by carbon tetrachloride and drug intervention were observed using principal component analysis (PCA) and partial least squares-discriminate analysis (PLS-DA). Potential biomarkers were found using PLS-DA and T-test. The results of the pathological changes observed in the liver, the biochemical parameters in plasma, and the metabolomics of the plasma and liver homogenate all showed that liver injury was successfully reproduced, ASP exhibited hepatoprotective effect, and the medium dose of ASP exhibited the best. Nine endogenous metabolites in the liver homogenate and ten endogenous metabolites in the plasma were all considered as potential biomarkers. They were considered to be in response to hepatoprotective effects of ASP involved in the amino acids metabolism, energy metabolism, and lipids metabolism. Therefore metabolomics is a valuable tool in measuring the efficacy and mechanisms of action of traditional Chinese medicines. © 2014 Elsevier B.V. All rights reserved.

1. Introduction Angelica sinensis (AS) is the root of A. sinensis (Oliv.) Diels, one of the most widely used materials in traditional Chinese medicine

Abbreviations: AS, Angelica sinensis; GC–MS, gas chromatography–mass spectrometry; PCA, principal component analysis; PLS-DA, partial least squaresdiscriminant analysis; ASP, Angelica sinensis polysaccharides; LASPG, low dose group of ASP; MASPG, middle dose group of ASP; HASPG, high dose group of ASP; CCl4 , carbon tetrachloride; HE, hematoxylin and eosin; MDA, malonaldehyde; SOD, superoxide dismutase; ALT, alanine aminotransferase; AST, aspartate aminotransferase; ␥-GT, ␥-glutamyl transpeptidase; MSTFA, N-methy-N-(trimethylsilyl) trifluoroacetamide; TMCS, trimethylchlorosilane; VIP, variable importance in the projection; VLDL, very low density lipoproteins; NMR, nuclear magnetic resonance; LC–MS, liquid chromatography–mass spectrometry; CSV, comma-separated value; TCM, traditional Chinese medicines. ∗ Corresponding author. Tel.: +86 931 7631954; fax: +86 931 7631109. E-mail address: [email protected] (Y. Wei). http://dx.doi.org/10.1016/j.jchromb.2014.10.009 1570-0232/© 2014 Elsevier B.V. All rights reserved.

(TCM) [1]. It is mainly distributed in Gansu province in China [2]. Pharmacological tests have revealed that AS can be used to treat irregular menstruation [3], amenorrhea, dysmenorrhea, anemia, gastrointestinal disease, cardiovascular disease, chronic bronchitis, asthma, rheumatism, hypertension, and other diseases in females [4]. Over 70 kinds of compounds have been identified from AS, including polysaccharides, essential oils, organic acids, and ester [5,6]. Among these compounds, A. sinensis polysaccharides (ASP) are an important group of pharmacologically active substances against liver injury [7]. Superoxide dismutase (SOD) and malonaldehyde (MDA) were the two important biochemical parameters to evaluate the hepatoprotective effect indirectly [8]. In the CCl4 -induced liver damage model, when the reactive oxygen species (ROS) in liver cells significantly increased, the SOD would be consumed excessively, and MDA accumulation would be in large quantities. Treated by ASP [9], SOD and MDA would be restored to normal. Alanine aminotransferase (ALT), aspartate aminotransferase (AST), and ␥-glutamyl transpeptidase (␥-GT)

46

P. Ji et al. / J. Chromatogr. B 973 (2014) 45–54

in the plasma were the three important biochemical parameters to evaluate the hepatoprotective effect directly [10]. In the CCl4 -induced liver damage model, membrane disintegration of hepatocytes with subsequent release of ALT, AST, and ␥-GT marker enzymes of hepatotoxicity, centrilobular necrosis and steatosis are some of the consequences of CCl4 -induced lipid peroxidation [11]. Treated by Lycium barbarum polysaccharides [12], ALT, AST, and ␥GT would be also restored to normal. So the above 5 biochemical parameters in plasma have been used to evaluate the hepatoprotective effect of polysaccharides. However, the exact hepatoprotective effect mechanism of ASP remains unknown. Metabolomics is a new discipline in systems biology. According to the “holism” philosophy, metabolomics techniques can provide important information for TCM research [13,14]. Multifarious metabolic characteristics of normal, pathological, or drug-treated subjects can be revealed by metabolomics, and this method has been used to explore the therapeutic effect mechanism of TCMs [15,16]. Metabolomics have been studied through numerous methods. Compared with NMR and LC–MS-based metabolomics methods, GC–MS-based metabolomics [17] has numerous advantages, such as the availability of many structure databases, higher sensitivity, better ability of material separation, and easier identification of metabolites. PCA and PLS-DA are the indispensable pattern recognition methods in the metabolomics research. PCA is an unsupervised multivariable statistical method. It is firstly carried out to investigate whether two groups can be separated and to find out their metabolic distinction. PLS-DA is a supervised multivariable statistical method. It is used to sharpen an already established (weak) separation between groups of observations plotted in PCA [18,19]. In this work, the hepatoprotective effects of ASP are studied by evaluating superoxide dismutase (SOD) activity, malonaldehyde (MDA) content, alanine aminotransferase (ALT), aspartate aminotransferase (AST), and ␥-glutamyl transpeptidase (␥-GT) in the plasma and pathological changes. Moreover, a metabolomics strategy based on GC–MS is employed to access the metabolic response of ASP in mice with CCl4 -induced liver injury. 2. Experimental 2.1. Chemicals A. sinensis was purchased from Minxian County, Gansu Province, China and authenticated by Dr. Yanming Wei (School of Veterinary Medicine, Gansu Agriculture University, Lanzhou, China). Soybean oil, O-methyl hydroxyllmine hydrochloride, N-methyl-N(trimethylsilyl) trifluoroacetamie (MSTFA), Trimethylchlorosilane (TMCS), and docosane (used as internal standard) were all purchased from Sigma–Aldrich (St. Louis, MO, USA). Assay kits for MDA, SOD, ALT, AST and ␥-GT were purchased from Nanjing Jiancheng Bioengineering Institute (Nanjing, China). Deionized water was purified by the Milli-Q system (Millipore, Bedford, MA, USA). All the reagents were analytical grade or chromate-graphic grade.

following procedure: protein was then removed using papain (mass fraction = 1%) and Sevage reagent [chloroform: n-butanol, 4:1 (v/v)] (volume fraction = 20%), while pigments were removed using hydrogen peroxide with the concentration of 30% (volume fraction = 2.5%). In the end, purified polysaccharide fraction ASP was obtained.

2.3. Animal studies 50 six-week-old male Kunming mice (20.0 ± 2.0 g) (SCXKZ (Gan) 2009-0004) were purchased from Experimental Animal Center of Lanzhou University (China). All animals were kept in a barrier system with controlled conditions of 22 ± 0.5 ◦ C, 50 ± 2.0% RH, and on a 12/12-h light/dark cycle. The mice were also allowed free access to basal pellet diet and tap water. After one week of feeding, the mice were randomly divided into five groups (10 mice for each group): control group, liver injury group, low dose group of ASP (LASPG), middle dose group of ASP (MASPG), and high dose group of ASP (HASPG). ASP was dissolved in normal saline for use. Mice in the control and liver injury groups were given normal saline; whereas those in the LASPG, MASPG, and HASPG groups were given ASP (60, 120, and 240 mg/kg/day, respectively) [20,21]. Drugs and normal saline were orally administered once daily for three successive days. Starting from the 4th day, liver injury was induced using 0.1% CCl4 in soybean oil at 20 mL/kg through intraperitoneal injection in mice from the liver injury group, LASPG, MASPG, and HASPG, whereas the mice in the control group was only injected with soybean oil. After liver injury was established, mice in LASPG, MASPG, and HASPG were administered with ASP at the dosage of 60, 120, and 240 mg/kg/day, respectively; normal saline was used for the control and liver injury groups. Each mouse was administered with an orally accurate volume of 1 mL/100 g.

2.4. Sample preparation All mice were sacrificed after 36 h of being induced with CCl4 . Blood were collected into heparinized tubes, and then centrifuged at 3000 rpm and 4 ◦ C for 10 min. The plasma was collected. One part of the plasma was used to detect SOD activities, MDA content, ALT, AST, and ␥-GT according to the instructions; the other part was frozen at –80 ◦ C until GC–MS analysis. One part of the left lobe of the liver tissue was fixed in 10% formalin for observations of pathological changes. The remaining liver tissue was made into liver homogenate with twice the weight of normal saline and was frozen at –80 ◦ C until GC–MS analysis. Animal welfare and experimental procedures were always performed according to the guide for the Care and Use of Laboratory Animals and were also approved by the Animal Ethics Committee of Gansu Agricultural University.

2.5. Determination of biochemical parameters in plasma 2.2. Extraction of ASP Crude polysaccharide was extracted according to the following optimized procedure: 100 g of every AS sample was prepared, and 838 mL water was put in; 229 mL solution was obtained after concentrated in a rotary evaporator under reduced pressure; according to the formula: alcohol density (%) = Vanhydrous alcohol /(Vanhydrous alcohol + Vsolution ), the alcohol density was 65.80%, and the reflux extraction time was set at 120 min. Then crude polysaccharide was purified according to the

MDA content, SOD activity, ALT, AST, and ␥-GT in plasma were measured using commercially available kits.

2.6. Pathological changes After 72 h of being fixed in 10% phosphate-buffered formalin, the liver blocks were embedded in paraffin, cut into 5 ␮m sections, and stained with hematoxylin and eosin (H&E staining).

P. Ji et al. / J. Chromatogr. B 973 (2014) 45–54

2.7. GC–MS analysis of plasma and liver homogenate samples 2.7.1. Plasma and liver homogenate samples preparation The liver homogenate and plasma samples were thawed at room temperature prior to analysis. 250 ␮L of acetonitrile was added into 100 ␮L of the sample in 1.5 mL centrifuge tube to precipitate the protein. The mixture was ultrasonically extracted for 10 min, followed by centrifugation (4000 rpm) for another 10 min. 400 ␮L of supernatant was then transferred to the centrifuge tube and evaporated to dryness under a stream of nitrogen gas. Approximately 100 ␮L of methoxyamine pyridine solutions (15 mg/mL) was then added to the vial. Methoxymation was performed at 70 ◦ C for 1 h, and then 50 ␮L of MSTFA with 1% TMCS was added to the vial. After the silylation was performed at 70 ◦ C for 1 h, 150 ␮L of n-heptane, with docosane (0.10 mg/mL) as internal standard, was added to the mixture. The mixture was then centrifuged for 10 min, and the supernatant was transferred to the GC microvial for GC–MS analysis. 2.7.2. Chromatography GC–MS was applied to analyze the liver homogenate and plasma samples. Chromatographic analysis was performed in an Agilent 6890N/5973N series GC–MS (Agilent Corporation, USA) equipped with an OV-1701 capillary column (30 m × 0.15 ␮m × 15 mm). 2.7.3. Analytical procedure The initial temperature (85 ◦ C) was held for 3 min and then raised to 280 ◦ C at a rate of 10 ◦ C/min. All samples were injected in split mode. The injection temperature was 270 ◦ C. The mass spectrometer was operated in EI mode (positive ion, 70 eV), and the quadrupole was 150 ◦ C. Mass spectra was acquired in full scan mode with repetitive scanning from 60 m/z to 600 m/z in 1 s. Ion source temperature was 230 ◦ C. 2.8. Data analysis The biochemical parameters values were expressed as mean ± standard deviation (SD). All statistical analyses were carried out by SPSS for Windows (Version 13.0). Lots of GC–MS data was exported in the format of commaseparated value (CSV) and then processed by Agilent chromatographic work software. Firstly, original data was processed by filtering, processing retention time correction, proofreading the peaks and integrating peaks. Then Pattern recognition methods based on principal component analysis (PCA) and partial least squares-discriminate analysis (PLS-DA) were accomplished by SIMCA-P V11.5 (Umetrics, Sweden). 3. Results and discussion 3.1. Plasma biochemical parameters ALT, AST, and ␥-GT are important enzymes in liver that are closely related when the organ experiences damage [22,23]. The concentrations of these enzymes in plasma were determined, and the results were listed in Table 1. Compared with the control group, a highly significant (p < 0.01) increase in plasma ALT, AST, and ␥-GT levels was observed in the liver injury group. After treated with middle dose ASP (MASP), concentrations of ALT, AST, and ␥-GT in plasma exhibited a highly significant (p < 0.01) decline, coming close to the level observed in the control group. After treated with low dose ASP (LASP) or high dose ASP (HASP), concentrations of ALT, AST, and ␥-GT in plasma significantly (p < 0.05) declined. MDA content and SOD activity reflect the severity of liver injury caused by free radicals, and the ability of clearing these free radicals,

47

respectively [24]. Oxidative stress is related with the pathogenesis of liver injury. Induced with CCl4 , the liver was damaged, the free radicals in the liver accumulated, and the antioxidant ability of the liver decreased. This study showed that plasma SOD activity in the liver injury group was significantly (p < 0.01) lower than that in the control group, and the plasma MDA content was significantly (p < 0.01) increased (Table 1). Compared with the liver injury group, plasma SOD activity significantly (p < 0.01) increased in MASPG, as well as in HASPG and LASPG (p < 0.05). Plasma MDA content significantly (p < 0.01) declined in MASPG, as well as in HASPG and LASPG (p < 0.05). These results showed that ASP had good antioxidant ability, and may protect liver-injured mice from oxidative stress [25]. These results were in accordance with the study that L. barbarum polysaccharides could effectively reduce serum ALT level induced by CCl4 , remarkably inhibit the expression of cytochrome P450 2E1, and restore the expression levels of antioxidant enzymes [10]. 3.2. Pathological changes observation By macroscopic observation, the liver surface membrane of mice in the control group was smooth and elastic with dark red color, whereas that of mice in liver injury group was nonluminous and brittle with slightly yellow color. The observation of the liver surface membrane of mice in all the ASP treatment groups was between the two groups mentioned above; liver membrane surface of mice in MASPG was similar to that in the control group. The results of the pathological changes in the liver tissue of mice in the different groups were illustrated in Fig. 1. In the control group, hepatocyte of mice was normal, nuclear structure was clear, hepatic cords were visible clearly, and perisinusoidal space was also obvious (Fig. 1B). In liver injury group, the hepatocyte of mice had a large area with hydropic degeneration, hepatic cords were disorganized, perisinusoidal space disappeared, inflammatory cell infiltration was present, some cells was found with cytolysis, the cell membrane was ruptured, and nuclear disappeared, and was concentrated (Fig. 1A). In LASPG, hepatocyte was normal, hydropic degeneration was absent, hepatic cords were visibly clear with the arrangement slightly crowded, and perisinusoidal space was stenosis (Fig. 1C). In HASPG, hepatocyte and hepatocyte around the central vein had degeneration, perisinusoidal space turned stenosis, the structure of small hepatocyte was indistinct, and cytoplasm got dissolved. However, the nuclear structure was intact, and nuclear membrane nucleoli were dyed normally (Fig. 1E). In MASPG, hepatocyte was mainly normal, hydropic degeneration was absent, and hepatic cords were clearly, visible with the arrangement a slightly crowded (Fig. 1D). According to the above results, all the treatment groups of ASP had certain hepatoprotective effect, wherein the MASP treatment group showed the best hepatoprotective effect. This result was in accordance with the results of biochemical parameters (Table 1). 3.3. GC–MS analysis The representative chromatograms of the liver homogenate and plasma samples analyzed by GC–MS were shown in Fig. 2A and B. A total of 23 compounds have been identified both in the liver homogenate and in the plasma. Most of these compounds were fatty and amino acids involved in lipid peroxidation, energy metabolism, and amino acid metabolism. All data containing the retention time, exact mass, and peak intensity were recorded for multiple statistical analyses, including PCA and PLSDA. These analyses were chosen because of their ability to cope with highly multivariate, noisy, collinear, and possibly incomplete data.

48

P. Ji et al. / J. Chromatogr. B 973 (2014) 45–54

Table 1 Effects of ASP on MDA content, SOD activity, ALT, AST and ␥-GT in different groups. Group

Dosage (mg/kg)

MDA (nmol/mg)

Control group Liver injury group LASPG MASPG HASPG

– – 60 120 240

7.77 9.59 7.90 7.58 8.24

± ± ± ± ±

1.89 2.33** 0.66 0.75 1.87

SOD (U/mg) 317.98 269.09 302.93 335.41 308.56

± ± ± ± ±

33.30 28.90** 20.60 16.97 29.92

ALT (U/L) 55.52 122.17 98.46 62.45 99.43

± ± ± ± ±

␥-GT (U/L)

AST (U/L) 3.77 9.98** 7.22 8.54 8.99

90.21 195.41 165.45 104.75 155.02

± ± ± ± ±

11.25 21.74** 28.43 12.9 16.63

0.82 1.25 1.01 0.88 1.04

± ± ± ± ±

0.08 0.16** 0.08 0.08 0.08

Data were presented as mean ± SD (n = 10) and analyzed by one-way ANOVA, followed by Duncan’s multiple-range tests. Compared with the control group, ** denotes that the difference is highly significant (p < 0.01). Compared with liver injury group,  denotes that the difference is significant (p < 0.05);  denotes that the difference is highly significant (p < 0.01).

3.4. Comparison of metabolic profiling between the control and the liver injury groups The GC–MS data is displayed as “score plots” by PCA. “Score plots” represent the distribution of samples in multivariate space. PCA score plots were obtained from the GC–MS data of the control group and liver injury group. The PCA score plots (Fig. 3A

and B) showed that the liver injury group exhibited a tendency to be away from the control group, which revealed the visible perturbation of the liver homogenate and plasma metabolic profiles in the liver injury group. Fig. 3A illustrated the separations by the model parameter R2 X = 0.935 for GC–MS data of the liver homogenate. Fig. 3B illustrated the separations by the model parameter R2 X = 0.877 for GC–MS data of plasma. R2 X shows the

Fig. 1. Liver pathological changes (200×) of mice in different groups. (A) Liver injury group; (B) the control group; (C) LASPG; (D) MASPG; (E) HASPG.

P. Ji et al. / J. Chromatogr. B 973 (2014) 45–54

49

Fig. 2. Total ion current chromatograms for the metabonomics analysis of liver homogenate (A) and plasma (B) in mice. Note: 1, fumaric acid; 2, acetamide; 3, propanoic acid; 4, acetic acid; 5, valine; 6, butanoic acid; 7, citric acid; 8, glycine; 9, succinic acid; 10, malic acid; 11, phosphoric acid; 12, gluconic acid; 13, fructose; 14, d-glucose; 15, alanine; 16, galactose; 17, palmitelaidic acid; 18, hexadecanoic acid; 19, inositol; 20, docosane; 21, linoleic acid; 22, octadecanoic acid; 23, arachidonic acid.

explanative ability of the model, indicating that the data can be highly elucidated by these two models. Thus, the liver injury model was successful. The model induced by CCl4 is an established experimental model in studying the liver injury mechanism and the hepatoprotective effect of drugs [26], and can thus be used in this study to explore the hepatoprotective mechanism of ASP. 3.5. Intervention of LASP, MASP, and HASP on the metabolic profiling PLS-DA was conducted to determine whether LASP, MASP, and HASP influenced the metabolic pattern of liver injury, and further refine the metabolites. The score plots (Fig. 4A and B) of the PLS-DA model showed a clear separation among the control group, liver injury group, LASPG, MASPG, and HASPG, wherein LV1 and LV2 explained 48.71%, 18.19% (Fig. 4A) and 56.73%, 19.18% (Fig. 4B) of the variance in the GC–MS data, respectively. All samples fell inside the 95% confidence interval, which was represented by an ellipse in Fig. 4. LASPG, MASPG, and HASPG were all located at the opposite direction along LV1 compared with the liver injury group. Moreover, these three groups showed a trend to be close to the control group. MASPG and the control group were at the same side of LV2, but opposite to the liver injury group, indicating that MASP provided the best hepatoprotective effect on liver injury of mice. This result was in accordance with the results of biochemical parameters (Table 1) and pathological changes observation (Fig. 1). They all showed that the hepatoprotective effects of LASP and HASP were not better than that of MASP. Firstly, that was because the

effective concentration of ASP in mice after given LASP (60 mg/kg/day) was lower than MASP (120 mg/kg/day). The effect was in accordance with the doses between 60 and 120 mg/kg/day. Yu [27] also found that the hepatoprotective effect of ASP with 120 mg/kg/day was better than ASP with 60 mg/kg/day. Secondly, with the dose increasing, the hepatoprotective effects of HASP were not better than that of MASP. The effect was not in accordance with the doses between 120 and 240 mg/kg/day. That was because such high dose of ASP (240 mg/kg/day) might have a toxic effect on mice and have slightly injured the mice liver [28–31]. 3.6. Identification of potential biomarkers In PLS-DA model, the loading plots and variable importance in the projection (VIP) value of all the candidate metabolites were employed to identify the features contributing to group separation. PLS-DA loading plots of liver homogenate (A) and plasma (B) were shown in Fig. 5A and B. The candidate metabolites, with VIP value above 1.0 and p-value below 0.05, were considered as potential biomarkers. Following the threshold above, nine endogenous metabolites in the liver homogenate (Table 2) and ten endogenous metabolites in the plasma (Table 3) were considered as potential biomarkers that correlated with how ASP influenced the liver injure mice. Information on the mass spectra of the metabolites of GC–MS analysis was based on NIST 2005. The candidate metabolites were confirmed based on the metabolite detection process by the EI mass spectra of debris and the information stored in the database of

50

P. Ji et al. / J. Chromatogr. B 973 (2014) 45–54

Fig. 3. PCA score plots of liver homogenate (A) and plasma (B) in the control and the liver injury groups.

the standard substance of MS. The potential biomarkers mainly included amino acids, organic acids, and fatty acids, as shown in Tables 2 and 3.

liver-injured mice were significantly affected by LASP, MASP, and HASP treatments. Interestingly, the levels of potential biomarkers in liver homogenate and plasma of mice in MASPG changed parallel with those in the control group, implying that MASP could restore the altered levels of these metabolites in liver-injured mice.

3.7. Alteration in relative content of potential biomarkers The relative content of all the potential biomarkers were calculated according to the internal standard method; the results were listed in Tables 2 and 3. Compared with the control group, all potential biomarkers in the liver injury group were significantly affected by CCl4 . The relative content of the potential biomarkers in the

3.8. Biochemical explanation in metabolic pathway The metabolic profiling strategy of plasma and liver tissues is able to reveal the multi-pathway metabolic perturbation associated with CCl4 and the ASP treatment, as summarized in Fig. 6.

Table 2 Comparison of relative content of biomarker in liver homogenate (mean ± SD). Metabolite Malic acid Fumaric acid Glycine Succinic acid Hexadecanoic acid Octadecanoic acid Valine Linoleic acid Arachidonic acid

Retention time/min 19.80 10.49 16.95 17.69 28.51 33.29 15.67 31.52 37.29

VIP value 2.162 2.001 1.953 1.861 1.632 1.505 1.411 1.265 1.191

Control group 0.051 0.196 0.187 0.098 0.716 0.601 0.436 0.512 0.915

± ± ± ± ± ± ± ± ±

0.015 0.048 0.051 0.025 0.142 0.123 0.085 0.184 0.259

Liver injury group 0.085 0.284 0.453 0.231 0.983 0.389 0.131 0.812 0.467

± ± ± ± ± ± ± ± ±



0.014↑ 0.056↑ 0.078↑ 0.031↑ 0.111↑ 0.124↓ 0.025↓ 0.234↑ 0.211↓

HASPG 0.071 0.246 0.396 0.191 0.883 0.483 0.232 0.711 0.553

± ± ± ± ± ± ± ± ±

MASPG 

0.015↓ 0.051↓ 0.578↓ 0.032↓ 0.174↓ 0.121↑* 0.047↑* 0.178↓ 0.223↑*

0.061 0.213 0.253 0.121 0.742 0.576 0.321 0.592 0.688

± ± ± ± ± ± ± ± ±

LASPG 

0.013↓ 0.047↓ 0.964↓ 0.021↓ 0.154↓ 0.122↑** 0.052↑** 0.111↓ 0.319↑**

0.072 0.243 0.336 0.148 0.832 0.492 0.273 0.692 0.569

± ± ± ± ± ± ± ± ±

0.012↓ 0.088↓ 0.549↓ 0.026↓ 0.144↓ 0.131↑* 0.036↑* 0.251↓ 0.229↑*

Note: ↑ and ↓ denote increasing and declining tendency, respectively. Compared with the control group,  denotes that the difference is highly significant (p < 0.01); compared with liver injury group, * denotes that the difference is significant (p < 0.05), ** denotes that the difference is highly significant (p < 0.01).  denotes that the difference is significant (p < 0.05).  denotes that the difference is highly significant (p < 0.01).

P. Ji et al. / J. Chromatogr. B 973 (2014) 45–54

51

Fig. 4. PLS-DA score plots of liver homogenate (A) and plasma (B) in each group.

3.8.1. Amino acid metabolism perturbation The liver is the hub of amino acids metabolism, and any hepatic injury might induce amino acids metabolic disturbances [32]. The metabolomic results showed that compared with the control group, valine decreased in the liver homogenate and the plasma of mice in the liver injury group (p < 0.01), but glycine increased in the liver homogenate and the plasma of mice in the liver injury group (p < 0.01). Alanine was increased in the plasma of mice in the

liver injury group (p < 0.01). These results suggested that glycine, alanine, and valine metabolism were affected by CCl4 . It was in accordance with the previous research. Recent research has indicated that glycine significantly decreases liver injury via a direct effect on hepatocytes [33]. Valine is also an important amino acid for removing excess nitrogen (potential toxins) in the liver [34]. Generally, the impact of acute liver injury on amino acids concentration results from three factors: protein synthesis and

Table 3 Comparison of relative content of biomarker in plasma (mean ± SD). Metabolites

Retention time/min

VIP value

The control group

Fumaric acid Alanine Malic acid Glycine Citric acid Arachidonic acid Octadecanoic acid Hexadecanoic acid Linoleic acid Valine

10.49 25.59 19.81 16.95 16.86 37.28 33.29 28.51 31.53 15.67

1.801 1.741 1.664 1.451 1.348 1.321 1.305 1.232 1.194 1.151

0.206 0.615 0.059 0.203 0.411 0.695 0.561 0.826 0.523 0.449

± ± ± ± ± ± ± ± ± ±

0.043 0.113 0.016 0.064 0.072 0.223 0.123 0.143 0.194 0.098

Liver injury group 0.295 0.766 0.098 0.497 0.589 0.991 0.819 0.683 0.234 0.128

± ± ± ± ± ± ± ± ± ±

0.056↑ 0.119↑ 0.023↑ 0.076↑ 0.096↑ 0.213↑ 0.121↑ 0.114↓ 0.143↓ 0.031↓

HASPG 0.276 0.669 0.075 0.422 0.480 0.897 0.753 0.788 0.309 0.192

± ± ± ± ± ± ± ± ± ±

MSPG 0.050↓ 0.131↓ 0.034↓ 0.058↓ 0.086↓ 0.223↓ 0.141↓ 0.171↑* 0.168↑ 0.037↑*

0.219 0.629 0.063 0.285 0.428 0.751 0.598 0.846 0.487 0.334

LASPG ± ± ± ± ± ± ± ± ± ±

0.040↓ 0.105↓ 0.024↓ 0.038↓ 0.087↓ 0.321↓ 0.187↓ 0.152↑** 0.129↑** 0.046↑**

0.243 0.657 0.069 0.334 0.472 0.871 0.712 0.802 0.321 0.232

± ± ± ± ± ± ± ± ± ±

0.067↓ 0.129↓ 0.026↓ 0.049↓ 0.084↓ 0.221↓ 0.206↓ 0.141↑* 0.172↑* 0.041↑*

Note: ↑ and ↓ denotes increasing and declining tendency respectively. Compared with the control group,  denotes the difference was highly significant (p < 0.01); compared with liver injury group, * denotes the difference was significant (p < 0.05), ** denotes the difference was highly significant (p < 0.01);  denotes the difference was significant (p < 0.05),  denotes the difference was highly significant (p < 0.01).

52

P. Ji et al. / J. Chromatogr. B 973 (2014) 45–54

Fig. 5. PLS-DA loading plots of liver homogenate (A) and plasma (B).

catabolism, BCAAs (branched chain amino acids) metabolism, and hepatic amino acid clearance. When liver injure occurs, catabolism promoting hormones, such as glucocorticoids and catecholamines, increase in secretion and decrease in deactivation, resulting in increased protein catabolism and increased levels of a number of amino acids in systemic circulation. Many amino acids are metabolized by liver enzymes and hence directly associated with the activity of hepatic enzymes. ALT and AST, two important enzymes in liver, were detected as well. BCAAs, such as valine, glycine, and alanine are essential amino acids typically involved in stress, energy and muscle metabolism [35]. With the increasing protein catabolism, valine decreased in the liver homogenate and the plasma of mice in the liver injury group. An increase of alanine in the plasma of mice in the liver injury group may be associated with an outflow of ALT from hepatocellular mitochondrion and an indication of impaired hepatic regulating function [36]. Fig. 6. Possible metabolic pathways. Note: words in red denote the 11 endogenous metabolites used in explaining the hepatoprotective mechanism. Words in black were the two kinds of metabolites not detected in our experiment. Words in blue denote the names of the possible metabolic pathways. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)

3.8.2. Lipid metabolism perturbation It has been reported that CCl4 derived free radicals may attack polyunsaturated fatty acids (PUFA) in cell membranes, forming fatty acid free radicals, which initiate an autocatalytic lipid

P. Ji et al. / J. Chromatogr. B 973 (2014) 45–54

53

peroxidation process and generate more lipid hydroperoxides and reactive hydroxyalkenals during membrane disruption. CCl4 can also promote the production of fatty acids and triglyceride inside the liver, accelerate the lipid esterification and cholesterols synthesis, and thus, reduce the content of partially unsaturated fatty acids. In this study, arachidonic acid and octadecanoic acid in the liver homogenate of mice in the liver injury group were significantly reduced, whereas hexadecanoic acid and linoleic acid significantly increased, compared with the control group. Arachidonic acid and octadecanoic acid in the plasma of mice in the liver injury group significantly increased, whereas hexadecanoic acid and linoleic acid significantly decreased. On one hand, CCl4 can accelerate the rate of lipid esterification and the synthesis of cholesterol, and reduce the content of some unsaturated fatty acid in liver [37]. On the other hand, if ␤-oxidation of fatty acids in the liver was damaged, the metabolism of some saturated fatty acids was disturbed, and saturated fatty acids accumulated in the liver [38]. This phenomenon can also be explained as the inhibition of lipoprotein synthesis by hepatic enzyme dysfunction, especially the synthesis of very low density lipoproteins (VLDL) [39]. VLDL is a critical hepatic protein required for the transport and secretion of lipids from the liver to plasma.

The results showed that citric acid, malic acid, fumaric acid and succinic acid were all restored to normal levels after ASP administration, in accordance with the biochemical metabolism. Glycine and alanine were all restored to the normal state after ASP administration, in accordance with the variation tendency of ALT in each group, as reported previously [43,44]. In conclusion, the hepatoprotective effect of ASP on liver injury may involve in regulating the dysfunction of lipid peroxidation, energy metabolism, and amino acid metabolism.

3.8.3. Energy metabolism perturbation Citric acid, malic acid, fumaric acid and succinic acid, which are belonging to fatty acids, can be decomposed by ␤-oxidation to acetyl coenzyme A to participate in the energy supply for the body. And they are all metabolites in tricarboxylic acid cycle (TCA), the most important energy source in mammals [40]. Malic acid, fumaric acid and succinic acid in the liver homogenate, as well as citric acid, malic acid and fumaric acid in the plasma of mice in the liver injury group significantly increased compared with the control group (p < 0.01). These results indicated that energy metabolism in TCA was slowed down after the liver was damaged by CCl4 . Under oxidative stress, TCA cycle is slowed down in cellular regulation to reduce the natural production of ROS. So we infer that the augment of the 4 kinds of fatty acids mentioned above is due to the dysfunction of mitochondria and the block of natural energy production by CCl4 intoxication. Numerous key enzymes, including dehydrogenase, gathered in the intracellular mitochondria. Hence, the alteration in TCA also indicated that the function in the intracellular mitochondria was disturbed [41].

Conflict of interest statement

3.8.4. Supposition of hepatoprotective effect mechanism of ASP CCl4 is widely used in animal models to induce liver injury. It is generally believed that the toxicity of CCl4 results from its reductive dehalogenation by the cytochrome P450 enzyme system into the highly reactive free radical trichloromethyl radical. In addition, it has been shown that CCl4 induced toxicity may stimulate endogenous reactive oxygen and nitrogen species which have also been suggested to play an important role in the pathogenesis of hepatotoxicity. The results (Tables 2 and 3) showed that the relative content of biomarkers significantly recovered after ASP administration, indicating that the hepatoprotective effect mechanism was in relation to the metabolic pathway wherein these biomarkers participated in (Fig. 6). Hexadecanoic acid, linoleic acid and octadecanoic acid, which are significantly related to lipid peroxidation, returned to normal levels after ASP administration; the effect of MASP (120 mg/kg/day) was the best (Tables 2 and 3). The results showed that the protective effect of ASP on liver injury may be related with lipid peroxidation, which is consistent with the variation tendency of MDA and SOD in each group, as previously reported previously [42].

4. Conclusion A novel approach using biochemical parameters coupled with metabolomics based on GC–MS and chemometrics was established to explain the hepatoprotective effect mechanism of ASP. Pathological change, SOD activity, MDA content, ALT, AST, and ␥-GT in plasma all had a certain degree of recovery after ASP intervention, with the best dose of 120 mg/kg/day. Nine potential biomarkers in the liver homogenate and ten potential biomarkers in the plasma were found to explain the hepatoprotective mechanism of ASP. The hepatoprotective effect of ASP on liver injury may involve in regulating the dysfunction of lipid peroxidation, energy metabolism, and amino acid metabolism.

We declare that we have no conflicts of interest to this paper submitted. Acknowledgements We are grateful to all other staff in the Institute of Traditional Chinese Veterinary Medicine of Gansu Agricultural University for their assistance in the experiments. The project is supported by National Natural Science Foundation of China (30972210, 31272600). References [1] Y.L. Sun, S.W. Cui, J. Tang, X.H. Gu, Carbohydr. Polym. 80 (2010) 544–550. [2] S. Zhang, B. He, J.B. Ge, H.B. Li, X.Y. Luo, H. Zhang, Y.H. Li, C.L. Zhai, P.G. Liu, X. Liu, X.T. Fei, Int. J. Biol. Macromol. 47 (2010) 546–550. [3] J. Jiang, Y.J. Guo, A.J. Niu, Carbohydr. Polym. 77 (2009) 384–388. [4] W. Cao, X.Q. Li, L. Liu, T.H. Yang, C. Li, H.T. Fan, M. Jia, Z.G. Lu, Q.B. Mei, Carbohydr. Polym. 66 (2006) 149–159. [5] L. Yi, Y. Liang, H. Wu, D. Yuan, J. Chromatogr. A 1216 (2009) 1991–2001. [6] G.H. Lu, K. Chan, K. Leung, C.L. Chan, Z.Z. Zhao, Z.H. Jiang, J. Chromatogr. A 1068 (2005) 209–219. [7] Y.L. Sun, J. Tang, X.H. Gu, D. Li, Int. J. Biol. Macromol. 36 (2005) 283–289. [8] S. Sreelatha, P.R. Padma, M. Umadevi, Food Chem. Toxicol. 47 (2009) 702–708. [9] F. Yu, H. Li, Y. Meng, D. Yang, Carbohydr. Polym. 94 (2013) 114–119. [10] H.K. Lim, H.S. Kim, H.S. Choi, S. Oh, J. Choi, J. Ethnopharmacol. 72 (2000) 469–474. [11] N. Singh, V. Kamath, K. Narasimhamurthy, P.S. Rajini, Environ. Toxicol. Pharmacol. 26 (2008) 146–241. [12] G.U. Sai, H. Jiang, Chongqing Med. 36 (2007) 60–62. [13] A. Zhang, H. Sun, Z. Wang, W. Sun, P. Wang, X. Wang, Planta Med. 76 (2010) 2026–2035. [14] L.L. Geng, H.Y. Sun, Y. Yuan, Z. Liu, Y. Cui, K. Bi, X. Chen, Fitoterapia 84 (2013) 286–294. [15] C. Chen, K.W. Krausz, Y.M. Shah, J.R. Idle, F.J. Gonzalez, Chem. Res. Toxicol. 22 (2009) 699–707. [16] F. Kaplan, J. Kopka, D.W. Haskell, W. Zhao, K.C. Schiller, N. Gatzke, D.Y. Sung, C.L. Guy, Plant Physiol. 136 (2004) 4159–4168. [17] N.C. Posecion, E.M. Ostrea, D.M. Bielawski, J. Chromatogr. B 862 (2008) 93–99. [18] X.Z. Li, S.N. Zhang, L. Fang, C.F. Liu, Y. Wang, Y. Bai, N. Wang, S.M. Liu, Phytomedicine 20 (2013) 1219–1229. [19] Y. Liu, Z.B. Lin, G.G. Tan, Z.Y. Chu, Z.Y. Lou, J.P. Zhang, Z.Y. Hong, Y.F. Chai, Metabolomics 9 (2013) 1082–1095. [20] Y.N. Ye, E.S.L. Liu, Y. Li, H.L. So, C.C.M. Cho, H.P. Sheng, S.S. Lee, C.H. Cho, Life Sci. 69 (2001) 637–646. [21] R. Nie, J. Wuhan Polytech. Univ. 27 (2008) 23–25, 105. [22] M. Kanter, O. Coskun, M. Budancamanak, World J. Gastroenterol. 11 (2005) 6684.

54

P. Ji et al. / J. Chromatogr. B 973 (2014) 45–54

[23] H. Özbek, S. U˘gras¸, H. Dülger, I. Bayram, I. Tuncer, G. Öztürk, A. Öztürk, Fitoterapia 74 (2003) 317–319. [24] K.A. Shenoy, S.N. Somayaji, K.L. Bairy, Indian J. Pharmacol. 33 (2001) 260–266. [25] G.M.M. Rao, C.V. Rao, P. Pushpangadan, A. Shirwaikar, J. Ethnopharmacol. 103 (2006) 484–490. [26] J. Xiao, E.C. Liong, Y.P. Ching, R.C.C. Chang, K.F. So, M.L. Fung, G.L. Tipoe, J. Ethnopharmacol. 139 (2012) 462–470. [27] F. Yu, J. Chin. Med. Mater. 32 (2009) 1580–1583. [28] H. Ding, R.X. Peng, J.P. Yu, Chin. J. Hepatol. 9 (2001) 50–52. [29] L.P. Fan, S.D. Ding, L.Z. Ai, K.Q. Deng, Carbohydr. Polym. 90 (2012) 870–874. [30] X.Q. Cheng, H. Li, X.L. Yue, J.Y. Xie, Y.Y. Zhang, H.Y. Di, D.F. Chen, J. Ethnopharmacol. 130 (2010) 363–368. [31] F. Li, Q.P. Yuan, F. Rashid, Carbohydr. Polym. 78 (2009) 80–88. [32] D.Y. Kwon, Y.S. Jung, S.J. Kim, H.K. Park, J.H. Park, Y.C. Kim, J. Nutr. 139 (2009) 63–68. [33] M. Froh, Z. Zhong, P. Walbrun, M. Lehnert, S. Netter, R. Wiest, L. Conzelmann, E. Gabele, C. Hellerbrand, J. Scholmerich, R.G. Thurman, World J. Gastroenterol. 14 (2008) 5996–6003. [34] R.F. Butterworth, Liver Int. 23 (2003) 5–9.

[35] R. Bruck, J. Wardi, H. Aeed, Y. Avni, H. Shirin, I. Avinoach, M. Shahmurov, R. Hershkoviz, Liver Int. 23 (2003) 276–282. [36] Y.L. Hua, W.X. Xue, M. Zhang, Y.M. Wei, P. Ji, J. Ethnopharmacol. 151 (2014) 1090–1099. [37] M.Y. Wang, G. Anderson, D. Nowicki, J. Jensen, Plant Foods Hum. Nutr. 63 (2008) 141–145. [38] O. Hussein, M. Grosovski, E. Lasri, S. Svalb, U. Ravid, N. Assy, World J. Gastroenterol. 13 (2007) 361. [39] D. Amalraj, C. Parkavi, K. Murugaiah, T.S. Dhanaraj, Asian J. Pharm. Technol. 2 (2012) 51–53. [40] L.M. Hinder, A. Vivekanandan-Giri, L.L. McLean, S. Pennathur, E.L. Feldman, J. Endocrinol. 216 (2013) 1–11. [41] R.D. Brinton, Trends Neurosci. 31 (2008) 529–537. [42] M. Zhao, Y.Q. Du, L. Yuan, N.N. Wang, Am. J. Chin. Med. 38 (2010) 241–249. [43] Z. Ac¸ıkgöz, H. Bayraktar, Ö. Altan, S.T. Akhisaroglu, F. Kırkpınar, Z. Altun, J. Sci. Food Agric. 91 (2011) 1277–1282. [44] T. Chen, X. Zhang, Y. Long, H. Yu, X. Ran, Y. Gao, H. Lu, X. Xie, X. Chen, Y. Ren, J. Shi, H. Tian, J. Endocrinol. Invest. 35 (2012) 772–775.