Exploration of the Protective Mechanism of Naringin in the Acetaminophen-Induced Hepatic Injury by Metabolomics

Naringin is a dihydroflavone which was found in citrus fruits. Previous studies have indicated the antiapoptotic, antioxidative stress, and anti-inflammatory effects of naringin. It can improve many common diseases, including fibrosis or hepatotoxicity, cardiovascular disease, and diabetes. Acetaminophen (APAP) is a frequently used painkiller, and hepatotoxic side effects limit its use. The purpose of the current examination is to find the impact of naringin on APAP-induced hepatic injury. Firstly, we pretreated mice model groups with naringin. Then, the liver injury model was established by injecting intraperitoneally into mice with APAP. After the mice were euthanized, we obtained serum and liver tissue samples from the mice. Finally, these samples were analyzed using a metabolomics approach to find the underlying mechanism of the effects of naringin on APAP-induced liver injury and provide a new treatment strategy for APAP-induced liver injury. Our data indicate that naringin significantly improves APAP-induced liver injury in mice and reduces the expression levels of liver injury markers in a dose-dependent manner. Furthermore, analysis of differential metabolites in mice with liver injury showed that naringin reduced APAP-induced hepatotoxicity due to reversing multiple metabolite expression levels and the rescue of energy, amino acid, and purine metabolism.

Some drugs have not been widely used in treating diseases because of severe side effects, including methotrexate (MTX) [4], and cyclophosphamide (CTX) [9]. Therefore, it is of great significance to find natural products that can alleviate the side effects of drugs. Elsawy et al. employed MTX to induce acute liver injury in rats and subsequently treated these rats with naringin. They found that naringin significantly reduced the upregulation of markers of MTXinduced liver injury and reduced oxidative stress, protecting hepatocytes from MTX-induced damage [4]. In another study on cyclophosphamide, naringin also alleviated CTXinduced oxidant production, hepatocyte dysfunction, and inflammation [9].
Acetaminophen (APAP) is frequently used as an analgesic agent whose hepatotoxic side effects are primarily involved in acute liver failure [15][16][17]. At safe doses, once ingested, APAP is quickly absorbed in the small intestine and immediately reaches inside liver cells. Enzymes catalyze the majority (about 80-90%) of APAP from the UDPglucuronic acid transferase (UGT) 1A subfamily or sulfonate transferases. Then, it binds to glucuronic acid, forming nontoxic metabolites and excreting from the body in urine. Finally, this process does not damage liver cells [15,16,18,19]. Whereas once the accumulation of APAP exceeds the tolerable dose for the human body, the glucuronidation and sulfonation pathways tend to saturate, and excessive APAP will be metabolized to form a large number N-acetyl-para-benzoquinoneimine (NAPQI). Unfortunately, NAPQI will deplete glutathione (GSH) stored by liver cells. Subsequently, redundant NAPQI react with biomacromolecules such as proteins, DNA, and unsaturated lipids in the cell, raising downstream events. For instance, reactive oxygen species (ROS) production, mitochondrial dysfunction, cell necrosis, apoptosis, autophagy, etc. cells will die in the end [15,16,[18][19][20]. Hence, we can ameliorate the liver injury caused by APAP by targeting its metabolism.
Several previous studies have shown that naringin protects hepatocytes from liver injury by exerting antioxidant and anti-inflammatory effects [21][22][23][24][25][26]. Therefore, based on APAP metabolism, we are convinced that naringin may also be beneficial in improving APAP metabolism to reduce APAP-induced hepatotoxicity. Herein, a pretreated APAPmediated liver injury mice model was exposed to naringin and then analyzed liver tissue and serum with metabolomic approaches to find whether naringin reduces APAP-induced hepatotoxicity and possible molecular mechanisms involved. 2.4. Animal Treatment. A total of 24 male C57BL/6J mice (age, 6 weeks) were obtained from Guangdong Medical Laboratory Animal Center (Guangzhou, China). Mice were randomly assigned into the following four groups: (1) vehicle control (n = 8), (2) APAP (n = 8), (3) APAP+50 mg/kg/d naringin (n = 8), and (4) APAP+100 mg/kg/d naringin (n = 8). Initially, naringin suspension was prepared from 0.5% carboxymethyl cellulose solution, and APAP was dissolved in normal saline at a concentration of 20 mg/ml and dissolved by heating at 55°C for animal experiment. The control group of these animal experiments was given the same volume of liquid but composed solely of the vehicle solution (0.5% carboxymethylcellulose). Naringin is a single treatment, and the mice were orally given naringin. Then, APAP (300 mg/kg) was injected intraperitoneally at 3 h following naringin treatment. About 48 h post-APAP exposure, mice were anaesthetized with pentobarbital (50 mg/kg, dissolved in saline with a concentration of 2%) via intraperitoneal injection. After that, the mice were sacrificed; tissue and serum samples were collected according to the instructions provided with commercial kits for subsequent experiments.

Materials and Methods
2.5. Serum ALT/AST Activity Analysis. The collected blood samples were placed at room temperature for about 2 h, centrifuge for 15 min at 840xg and separated serum stored at -20°C for later analysis. The serum alanine transaminase (ALT) and aspartate transaminase (AST) assays in each group were performed following the manufacturer's instructions.
2.6. Histopathology. The left liver lobe of the mice was fixed using paraformaldehyde (4%), dehydrated, and embedded in paraffin. The liver sections (5 μm) were prepared and stained using hematoxylin and eosin (H&E), and then, microscopy was done. The cell necrotic boundaries were manually counted in three random fields per sample.

Glutathione Peroxidase and Malondialdehyde
Measurement. The contents of glutathione peroxidase (GSH-Px) and malondialdehyde (MDA) in the liver tissues were assayed using commercial kits following the manual instructions provided by the manufacturers.   3 Oxidative Medicine and Cellular Longevity sodium 3-(trimethylsilyl) propionate-2, 2, 3, 3-d4 (TSP), pH 7.4) and mixed with oscillation. The precipitation was removed by centrifugation. The supernatant was put on the machine for NMR analysis.
2.9. Data Preprocessing and Software. The TopSpin software (Bruker Biospin, Germany) version V3.0 was used to perform Fourier transform (FT), phase adjustment, baseline correction, and calibration. All spectra were multiplied by an exponential window function with a widening factor of 1 Hz when the Fourier transform was performed to improve SNR. The single-peak calibration of TSP at 0.00 ppm was used to remove the residual water peak, and the integral interval was 0.650-4.550 and 5.175-9.700 ppm. The integral interval was 0.015 ppm for data analysis.
The displacement range of finger-recognized compounds was imported into R software to calculate the integral area of each compound, and the fold value was obtained by intergroup comparison. Combined with a oneway analysis of variance, 0.05 was set as the threshold to screen differentially expressed metabolites.
2.10. Statistical Analysis. Data are presented as mean ± SEM. Statistical analysis was performed with the use of SPSS statistic software 15.0. Differences among groups were tested by one-way analysis of variance (ANOVA) with Tukey's post hoc test. In all cases, differences were considered statistically significant with P < 0:05.

Naringin Pretreatment Repaired Liver Injury and
Reversed the Level of Liver Injury Markers in the Samples.
We established a mouse model of liver injury using APAP and measured the extent of liver injury, including the proportion of liver cell necrosis and the production of liver injury markers. The proportion of liver injury and necrosis can be intuitively revealed by H&E staining results, whereas various markers can indirectly reflect APAP-induced endogenous metabolite changes. For this purpose, we selected four classic biomarkers for liver injuries, such as ALT, AST, MDA, and GSH-Px. The design scheme of the animal experiment is well elucidated (Figure 1(a)). Compared with the control, histopathology results of the liver tissue samples of mice in the model group showed that the liver tissue was significantly damaged after intraperitoneal injection of APAP ( Figure 1(b)). In addition, statistics on the proportion of necrotic area showed that about 70% of the tissue was necrotic in the model group. After low-dose (50 mg/kg/d) naringin treatment, the necrotic area was reduced by about half, while high-dose (100 mg/kg/d) naringin treatment showed only about 20% necrotic area ( Figure 1(c), P < 0:05 ). These results indicated that naringin repaired APAPinduced liver tissue damage and showed extremely significant therapeutic effects at high doses.
Furthermore, ALT and AST levels in serum and GSH-Px and MDA in liver tissue were measured. The results showed that naringin pretreatment reversed the levels of these markers of liver injury in the model group. To be specific, the levels of ALT, AST, and MDA in samples of model mice were significantly increased compared to the control group. Nevertheless, following pretreatment with naringin, ALT, AST, and MDA levels downregulated considerably in a dose-dependent fashion (Figures 1(d), 1(e), and 1(g), P < 0:05). Instead, although GSH-Px levels in the liver of model mice were significantly downregulated, naringin treatment reversed them (Figure 1(f), P < 0:05).

Identification of Major Metabolites in the Serum and
Liver of Mice by NMR Spectroscopy. After 1 H, NMR experiments were conducted on the serum and liver samples of the four groups of mice (control group, marked as N; APAP  Oxidative Medicine and Cellular Longevity group, marked as APAP; APAP+low-dose naringin therapy group, marked as YL; APAP+high-dose naringin therapy group, marked as YM), the typical hydrogen spectra as shown in Figure 2 were obtained. The names are identified by querying public metabolomics databases, such as HMDB (http://www.hmdb.ca/) and MMCD (http://mmcd.nmrfam .wisc.edu/), and annotated in legends. 29 and 28 metabolites were identified in extracts of serum ( Figure 2(a)) and liver (Figure 2(b)), respectively. To visualize the identification results, different colors and numbers represent different groups of spectra and identified metabolites, respectively. In addition, based on the low metabolite concentrations in the low-field region (2.175-9.700 ppm), the spectra were magnified by a factor of 15 to align with the high-field region.

Naringin Can Reverse the Level of Small Molecule
Metabolites in Serum and Liver Tissue Induced by APAP in Mice 3.3.1. PCA. We used R software to carry out pattern recognition multivariable analysis for the normalized spectral data. We used the mean centre scaling method to obtain the PCA (principal component analysis) graph, shown in Figure 3, to describe the weight of principal components. These graphics can help us analyze the changes of small molecule metabolites in serum and liver of APAP-induced mice compared to the control group. PCA patterns based on the NMR spectra of mouse serum are shown in Figure 3(a). The APAP-administered group and the control group mostly overlapped, indicating that APAP did not    figure). In the liver, both APAP and drugtreated groups overlapped completely with the control group, indicating that neither APAP nor drugs caused liver metabolite disturbances (Figure 3(b)). However, PCA is an unsupervised analytical method, and to further differentiate metabolic levels between groups, we used the supervised OPLS-DA to filter out irrelevant factors.

OPLS-DA.
The OPLS-DA score chart is shown in Figure 4. Based on the complete separation of APAP and control groups, metabolic disturbances in serum were caused by APAP (Figure 4(a)-B). However, compared with the APAP group, the administration group was closer to the control group, especially the YM group, indicating that the metabolic disturbance caused by APAP was reversed in a dose-dependent manner (Figure 4(a)-A). Consistent with the serum results, the APAP group was completely separated from the control group (Figure 4(b)-B). In contrast, the drug and control groups overlapped completely (Figure 4(a)-A, C, D). The results indicated that APAP caused the disturbance of the overall metabolic level of the liver, but was improved by the drug.

Screening Small Molecule Metabolites with Significant
Differences. To further identify the differential metabolites, the loading value of the first principal component of the OSC-OPLS-DA model was adopted to plot the loading plot, as shown in Figures 5 and 6. Orthogonal signal correction Orthogonal T score [1] Orthogonal T score [1] Orthogonal T score [1] T score [1] T score [1] T score [1] T score [1]  8 Oxidative Medicine and Cellular Longevity (OSC) was performed on the PLS-DA model. OSC could remove the variables that were not meaningful to the grouping to maximize the differences between groups. To find the differential metabolites contributing to group separation, the redder the peak, the more significant difference between the two groups, and the bluer the peak, there is no significant difference. The peak above the loading plot has a higher content in the group on the right side of the OPLS-DA score, while the peak below the loading plot has a higher content on the left side OPLS-DA score. The displacement range of finger-recognized compounds was imported into R software to calculate the integral area of each compound, and the fold value was obtained by intergroup comparison. Combined with a one-way analysis of variance, 0.05 was set as the threshold to screen differentially expressed metabolites. Please refer to Tables 1 and 2 for details. According to the information on differential metabolites, the reduced metabolites (isoleucine, valine, 3-hydroxyisobutyrate, acetate, pyruvate, histamine, inosine, hypoxanthine, and xanthine) can be observed in model groups compared to control groups in serum samples. Metabolites with elevated expression levels include leucine, 3-hydroxybutyrate, alanine, homoserine, glutamine, taurine, glycerol, threonine and 1, 3-dihydroxyacetone. The significant decrease in pyruvate, inosine, hypoxanthine, and xanthine, while 3-hydroxybutyrate, alanine, homoserine, taurine, and threonine is significantly increased. However, although APAP treatment caused this change in metabolites, there is a certain degree of reversal effect after naringin treatment and presented the characteristics of dose-dependence.
Similarly, there were reduced metabolites in the APAP model group in liver tissue samples compared to the control group, including 3-hydroxyisobutyrate, ethanolamine, and ATP. The significantly increased metabolites are leucine, lactate, alanine, glutamate, isocitrate, creatine, choline, and fumarate. Naringin treatment also reversed the expression levels of some metabolites, including ATP, leucine, alanine, and fumarate, in a dose-dependent manner. However, most metabolite changes did not improve with medium doses of naringin, suggesting that further research is needed to determine the therapeutic dose of naringin to achieve the best protective effect.

Summary of Pathway
Analysis. In addition, to reveal metabolic pathways that may be involved in naringin improving liver injury, metabolic pathway analysis of Orthogonal T score [1] Orthogonal T score [1] T score [1] T score [1]    11 Oxidative Medicine and Cellular Longevity differential metabolites of serum and liver between normal and liver-injured mice was performed using MetaboAnalyst 5.0. The results are shown in Figure 7. In general, pathways with an impact value of >0.1 was considered potentially targeted metabolic pathways. As shown in Figure 7, naringin treatment significantly affected the metabolism of various amino acids, TCA cycle, and gluconeogenesis pathways in serum and liver. These data suggest that naringin may play its role in protecting liver cells by targeting these metabolic pathways.

Discussion
Naringin has antioxidant, antiapoptotic, and antiinflammatory effects in several studies. It is found that naringin protects endothelial cells from ox-LDL-induced apoptosis and injury in an in vitro model of atherosclerosis [3].
In this study, we were interested in whether naringin could affect APAP-mediated liver injury, providing new ideas for the treatment or remission of this hepatotoxicity. APAP is an essential factor in the cause of acute liver failure, and this nature has seriously affected the use of this drug. Therefore, people should explore a new substance to alleviate this side effect. There have been some studies on APAP metabolism, but there is no way to avoid this side effect altogether. The challenge is that APAP metabolism involves a wide range of pathways. To this end, we investigated the impact of naringin on APAP-induced hepatotoxicity and the possible molecular mechanisms engaged by analyzing the degree of metabolite changes induced by APAP.
The liver is the chief organ controlling and managing the metabolism in the body, equipped with a variety of meta-bolic enzymes, including AST and ALT, primarily involved in catalyzing the transfer of amino acids between amino acids and ketone acids [27]. However, liver cells damage caused by external stimulation may disturb the integrity of the cellular membranes and enhance the permeability. This increased permeability enhances ALT and AST release into the blood to raise blood serum levels of ALT and AST. Therefore, the detection of blood ALT and AST levels indirectly reflects the degree of liver injury. The liver serves as a major site for biological oxidation; when aerobic cells are metabolized, a burst of reactive oxygen species (ROS) is generated that results in lipid peroxidation of cell membranes and damage to surrounding cells in the liver tissues. ROSmediated lipid peroxidation elevated MDA levels, and fluctuating levels of MDA may reflect the degree of lipid peroxidation in liver cells [28].
On the contrary, the primary function of GSH-Px is to decompose peroxides. Therefore, when liver cells are damaged, the activity of GSH-Px is reduced, and the peroxides cannot be decomposed, and the antioxidant capacity of the liver is reduced [29]. Our results have demonstrated that, compared with normal mouse samples, model mice have significantly increased areas of liver necrosis and significantly increased levels of ALT and AST in serum and MDA in the liver, which naringin treatment successfully reversed in a dose-dependent manner. As expected, APAPinduced liver injury decreased GSH-Px activity, whereas high-dose naringin significantly enhanced GSH-Px activity. On the other hand, R software was used to perform orthogonal partial least squares (OPLS) on normalized data to find the correlation between NMR data (the x variable) and other variables (Y variable, grouping information). Partial least squares discriminant analysis (PLS-DA) uses the data scale scaling method of unit variance scaling. The PLS-DA was used to test the quality of the model with the two-fold cross-validation method, and the R2X and Q2 obtained after the cross-validation were used to evaluate the validity of the model (representing the variables that could be explained by the model and the predictability of the model, respectively).
After that, the model's effectiveness was further tested by changing the order of the classification variable Y randomly several times (n = 2000) to obtain the corresponding random Q2 value. OPLS-DA filtered out irrelevant orthogonal signals to obtain more reliable differential metabolites. PCA graph and OPLS-DA score plots were used to analyze the experimental results of 1 H NMR, which showed that induc-tion of APAP did change the level of small molecule metabolites in the samples, and the expression was more evident in liver samples. However, treatment with naringin saved the levels of metabolites in the model samples, suggesting that naringin has the potential to treat APAP-induced liver injury.
The levels of succinate in serum and fumarate in the liver of the APAP model group were significantly increased, while ATP (Adenosine triphosphate) level was significantly decreased. Both succinate and fumarate are important intermediates of TCA circulation and the respiratory chain. Thus, their increase indicates that the APAP group mice have damaged mitochondria, disrupted TCA circulation, and disrupted the respiratory chain, and reduced efficiency of  13 Oxidative Medicine and Cellular Longevity energy generation [30]. ATP is the main molecule that drives the cellular reaction process. Glucose is converted to pyruvate, and two ATP molecules are produced through glycolysis. Subsequently, when oxygen is sufficient, pyruvate is directed into the tricarboxylic acid cycle, where a reaction involving substrate-level phosphorylation can have a certain amount of ATP. The respiration chains can also phosphorylate the intermediate to produce large amounts of ATP [31].
In contrast, when oxygen is scarce, pyruvate is converted to lactate, a process that produces little ATP. Consistent with this, there was a significant increase in lactate in serum and liver in the APAP group, indicating an enhanced anaerobic glycolysis pathway that ultimately leads to inadequate ATP supply. Therefore, to cope with this energy crisis, the body needs to use other ways to produce energy. Fatty acid oxidation can release a large amount of energy. As a product Color coded according to the logarithmic transformation of fold change (log 2 (FC)), warm color (red) shows the degree of increase in identified metabolites, and blue the decrease in APAP vs. N, YL vs. N, and YM vs. N groups. b P values corrected by Benjamini Hochberg methods were calculated based on a parametric Student's t-test or a nonparametric Mann-Whitney test. * P < 0:05, * * P < 0:01, * * * P < 0:001.

14
Oxidative Medicine and Cellular Longevity of fatty acid oxidation [32], acetate was significantly upregulated in the liver of the APAP model group, indicating that the fatty acid oxidation process was enhanced. Gluconeogenesis is another energy-saving pathway in which enzymes gradually convert pyruvate into glucose, providing the body with more glucose for energy production. A precursor of gluconeogenesis [33], alanine, is a key link between carbohydrate and amino acid metabolism. In this cycle, the ammonia produced by the breakdown of proteins is converted to L-alanine by transamination, which is then transported to the liver. Catalyzed by alanine aminotransferase, the amino group is transferred to α-ketoglutarate to form glutamate and pyruvate, which can be further converted to glucose [34]. However, our data showed that alanine levels in serum and liver of the APAP model group were significantly increased, while pyruvate and glucose levels were not significantly increased, indicating that the gluconeogenesis pathway was impaired. Therefore, fatty acid oxidation may be the leading energy supplement for APAPinduced stimulation in mice in the model group.
Furthermore, liver samples from the APAP model group showed significantly elevated leucine, isoleucine, and valine levels. They are commonly known as branched-chain amino acids essential for protein synthesis [35][36][37]. As previously mentioned, induction of APAP produces ROS, which further destroys proteins in the cytoplasm and membranes of liver cells [38]. The elevated levels of these three essential amino acids indicate enhanced hepatic catabolism and damage to some proteins. Our results suggest that low-dose naringin therapy salvages the increased levels of branched amino acids preventing ROS damage to proteins.
A purine catabolic pathway can be performed to save the body from oxidative damage. In this process, inosine goes through the first step of metabolism to produce hypoxanthine, which is then converted to xanthine by the catalysis of xanthine oxidoreductase and finally converted from xanthine to uric acid, which is an effective free radical scavenger with antioxidant effect [30,39]. Therefore, the significant reduction of inosine, hypoxanthine, and xanthne in the APAP model group may be attributed to a self-protection mechanism of hepatocytes.
In conclusion, our study demonstrates the potential value of naringin in the treatment of APAP-induced liver injury in mice. The underlying mechanisms include antioxidants, salvage, glucose metabolism, regulation of amino acids, and purine metabolism. Our study focused on the changes of endogenous metabolites in mice with liver injury and explored the therapeutic effects of naringin based on them. Using metabolomics analysis based on NMR, we identified small molecule metabolites that changed significantly before and after drug treatment and explored the metabolic pathways involved, providing a new strategy for alleviating the liver toxicity and side effects of APAP drugs.

Data Availability
The original contributions presented in the study are included in the article. Further inquiries can be directed to the corresponding authors.

Conflicts of Interest
The authors declare that they have no conflicts of interest.

Authors' Contributions
ZL is responsible for the investigation, data analysis, and writing-original draft; GW, WG, SZ, ZS, and ML for the resources and investigation; WL for the investigation; YC for the funding acquisition, resources, and supervision; GZ and BC for the writing-review and editing; and CW and TY for the funding acquisition, conceptualization, supervision, resources, and writing-review and editing. All authors have read and agreed to the final manuscript draft. Zihan Lin and Guanzhen Wang contributed equally to this work.