Quick viewing(Text Mode)

Targeted Metabolomics Identifies Differential Serum and Liver Amino Acids Biomarkers in Rats with Alcoholic Liver Disease

Targeted Metabolomics Identifies Differential Serum and Liver Amino Acids Biomarkers in Rats with Alcoholic Liver Disease

J Nutr Sci Vitaminol, 66, 536–544, 2020

Targeted Metabolomics Identifies Differential and Amino Acids Biomarkers in Rats with Alcoholic

Chenze Shi1, Lei Wang2, Kejun Zhou3, Mingmei Shao1, Yifei Lu1 and Tao Wu1,*

1 Institute of Interdisciplinary Integrative Medicine Research, Shanghai University of Traditional Chinese Medicine, Shanghai 201203, China 2 Department of Hepatology, Longhua Hospital, Shanghai University of Traditional Chinese Medicine, Shanghai 200032, China 3 Shanghai Key Laboratory of Pediatric Gastroenterology and Nutrition, Shanghai Institute for Pediatric Research, Shanghai 200072, China (Received November 21, 2019)

Summary To investigate changes in serum and hepatic levels of amino acids in ALD and to provide novel evidence and approaches for the prevention and treatment of ALD. Twenty specific pathogen-free SD male rats were devided into two groups, ten for the control group, and ten for the model group. Serum biochemical markers, including aminotransfer- ase, aspartate aminotransferase, laminin and hyaluronidase were measured. Histological analysis of liver tissues was performed. Serum and liver amino acids levels were quantita- tively determined by ultra-high-performance liquid chromatography-tandem quadrupole mass spectrometry (UPLC-TQMS)-based targeted metabolomics. Compared with the normal group, ALD rats showed an obvious increase in the levels of b-alanine, alanine, , orni- thine, and the tyrosine ratio, while there was a decrease in levels, the BTR ratio and Fischer’s ratio in serum. Additionally, ALD rats exhibited a significant increase in the levels of and putrescine, while there was a decrease in sarcosine, b-alanine, serine, , , , ornithine, , , tyrosine, symmetric dimethy- larginine, , and methionine-sulfoxide levels in liver tissues compared with the normal group. The serum and liver amino acids showed significant changes in ALD rats and can be considered as potential specific diagnostic biomarkers for ALD. Key Words ALD, , targeted metabolomics, alanine, arginine

Alcoholic liver disease (ALD) is a chronic liver disease ble factors are oxidative stress and ROS, since the liver is that is directly associated with long-term overdose alco- the main site of alcohol . Alcohol dehydro- hol consumption, including hepatic steatosis, alcoholic genase and cytochrome P-450 (CYP2E1) are the two steatohepatitis, , hepatocellular carcinoma and main pathways of alcohol metabolism in the liver (7). liver failure (1). Alcohol dehydrogenase converts alcohol to acetalde- Alcohol abuse has been as one of the top five risk fac- hyde, which is a cytosolic . Acetate tors for death and disability globally and results in 2.5 is subsequently metabolized via the mitochondrial million deaths and 69.4 million annual disability ad- enzyme acetaldehyde dehydrogenase by acetaldehyde. justed life years (2). As a result, ALD has been a world- It causes the reduction of nicotinamide adenine dinu- wide disease (2, 3) since its increasing incidence has cleotide (NAD) to NADH. Via the inhibition of glucone- made it a great threat to human health and social ogenesis and fatty acid oxidation, the ratio of NAD/ development. Approximately 60% to 90% of individuals NADH is altered and may promote fatty liver develop- who drink more than 60 g of alcohol per day have been ment. CYP2E1 also converts alcohol to acetaldehyde. It shown to have hepatic steatosis, which means a high is upregulated in chronic alcohol use, and free radicals incidence of ALD because of the consumption of alco- are generated through the oxidation of nicotinamide hol (4). ALD can adversely affect multiple organ sys- adenine dinucleotide phosphate (NADPH) to NADP. tems such as gastrointestinal system, central nervous Hepatic macrophages are activated due to chronic alco- system, hematologic system, the cardiovascular system hol abuse, and the activated macrophages produce and renal system (5). tumor necrosis factor-a (TNF-a). TNF-a induces mito- Many factors contribute to the development of ALD, chondria to increase the production of ROS. Oxidative such as oxidative stress, generation of reactive oxygen stress may promote hepatocyte necrosis and apoptosis. species (ROS) during alcohol metabolism, adipokines Free radicals initiate lipid peroxidation, which causes from visceral adipose tissue, and endotoxin derived from inflammation and fibrosis. These two pathways are con- the gut, etc. (6). Of these, the most common and possi- sidered the main ALD pathogenesis pathways. Although the pathogenesis of ALD has been widely investigated, * To whom correspondence should be addressed. the precise mechanisms remain to be elucidated. Thus, E-mail: [email protected] biomarkers for the early detection of ALD are urgently

536 Differential Serum and Liver Amino Acids Biomarkers in ALD 537

Fig. 1. Serum levels and histological analysis of liver tissue in rats with ALD. (A) serum ALT and AST levels in groups. ** p,0.01 vs Normal control group. (B) Liver histopathology was detected by hematoxylin and eosin staining and light microscopy (3200 magnification). N: Normal control group; M: Model ALD group. necessary. systematically detect the amino acid levels using quan- Metabolomics is a quantitative analytical technique, titative methods. that measures all endogenous metabolites occurring in On the basis of these applications of metabolomics on biosystems (cells, tissues and organisms) after exci- ALD, our research on differential serum and liver amino tations or interpretations from the external environ- acids in rats with ALD was performed based on ultra- ment (8). It is a powerful technology that allows the high-performance liquid chromatography coupled with assessment of global low-molecular-weight metabolites tandem quadrupole mass chromatography (UPLC-​ in biological systems and shows great potential in bio- TQMS)-based quantitative targeted metabolomics anal- marker discovery (9). As is well known, the liver is the ysis for detecting the pathogenesis of ALD and for main location of alcoholic metabolism, and it is the searching for new approaches to prevent and treat ALD. most injuried target organ (10). During the process of MATERIALS AND METHODS alcohol-induced liver injury, changes in hepatic meta- bolic pathways and metabolite levels may occur. By Animal experiments. Twenty specific pathogen-free applying metabolomics to the research of ALD, the lev- male SD rats were purchased from SLAC Laboratory els of differential metabolites and their effects on the Animal Center, Inc. (Shanghai, China). They were development of ALD can be revealed. randomly classified into two groups, ten in the control Amino acids, as one of the most important nutrients group, and ten in the model group. A rat model of in the metabolism of organisms, are the basic materials chronic ALD was established using a mixture (500 mL/L to make up proteins and are essential substances for or- alcohol, 8 mL/kg per day; corn oil, 2 mL/kg per day; ganic nutrition (11). The main site of the catabolism of pyrazole, 24 mg/kg per day) once a day and intraperi­ amino acids is the liver. In alcohol-induced liver toneal injections of 0.25 mL/kg of a 25% solution of disease, the level of the and some specific CCl4 in olive oil twice a week for 12 wk according to our amino acids may show significant changes. In 2010, previous study (15). After 12 wk, the rats were sacri- Mukherjee et al. (12) analyzed the patterns of changes ficed. Blood samples were obtained, and serum was col- in plasma amino-acid concentrations due to ALD. In lected by centrifugation (12,000 rpm, 4˚C, 10 min) and 2015, Liang et al. (13) performed a metabolomics ap- then stored at 280˚C. Liver pieces were fixed in 10% proach in a group of 206 ALD patients. A total of six neutral buffered formalin or snap frozen in liquid nitro- differential urinary metabolites that contributed to ALD gen for further analysis. The protocols of all animal progress were identified, and more importantly, they experiments were undertaken in accordance with the discovered three of them with an accuracy of more National Institute of Health Guide for the Care and Use than 95%. In 2016, Harada et al. (14) performed a pop- of Laboratory Animals, with the approval of the Local ulation-based, cross-sectional study to identify potential Ethics Committee for Animal Research Studies at the biomarkers of alcohol intake and alcohol-induced liver Shanghai University of Traditional Chinese Medicine. injury by metabolomics profiling using capillary electro- Serum biochemical assay. Serum biochemical assays, phoresis-mass spectrometry. Taken together, metabolo- including measurements of alanine transaminase (ALT), mics has been applied and has provided a new approach (AST), laminin and hyaluroni- to the study of ALD. However, the above studies did not dase, were performed with an automatic biochemistry 538 Shi C et al.

Table 1. Altered serum amino acid profiles in rats with ALD.

AAs (mM) Normal Model p-value

Creatinine 40.975 (37.172–42.372) 46.969 (43.244–52.738) 0.274 867.868 (814.128–993.445) 1081.062 (1022.977–1138.628) 0.147 Sarcosine 5.140 (4.759–5.926) 4.88 (3.962–5.643) 0.660 b-Alanine 8.528 (7.794–10.513) 13.906 (11.872–16.305) 0.039 Alanine 857.348 (845.415–883.310) 989.483 (956.161–996.962) 0.030 Serine 406.141 (397.772–416.530) 548.262 (450.421–562.177) 0.011 Proline 367.940 (330.332–389.071) 369.197 (354.909–415.875) 0.578 Valine 230.819 (224.673–241.771) 230.093 (223.003–241.592) 0.834 Cysteine 1.934 (1.500–2.526) 1.474 (1.259–1.804) 0.489 Threonine 455.176 (441.033–462.343) 498.252 (475.911–514.262) 0.069 Taurine 599.283 (556.882–646.830) 643.545 (593.089–684.036) 0.834 OH-proline 0.034 (0.029–0.044) 0.053 (0.045–0.066) 0.572 122.856 (118.437–129.463) 124.440 (116.374–146.317) 0.836 Ornithine 163.341 (136.844–178.724) 613.048 (550.619–629.335) 0.003 708.343 (678.051–759.150) 709.055 (626.957–758.605) 0.834 1752.128 (1708.983–1830.497) 1777.516 (1716.758–1850.909) 0.587 Lysine 816.093 (765.229–864.076) 902.931 (854.177–923.006) 0.089 Histidine 110.807 (96.695–114.937) 118.380 (97.5833–146.558) 0.749 Arginine 474.166 (442.695–507.337) 21.233 (7.322–78.512) 0.002 Citrulline 112.358 (106.455–121.955) 121.672 (111.824–131.515) 0.431 Tyrosine 111.196 (108.525–134.475) 168.779 (148.328–191.966) 0.029 Asymmetric dimethylarginine 1.006 (0.944–1.275) 1.529 (1.487–1.556) 0.802 Symmetric dimethylarginine 0.357 (0.323–0.538) 0.482 (0.389–0.712) 0.247 Methionine 108.547 (96.135–110.221) 117.618 (95.158–131.064) 0.431 Histamine 0.132 (0.109–0.183) 0.171 (0.124–0.184) 0.834 Isoleucine 166.525 (160.191–184.241) 182.791 (171.185–195.988) 0.147 207.885 (205.778–225.112) 218.058 (209.707–227.843) 0.587 Phenylalamine 134.985 (121.653–139.395) 160.681 (143.379–175.820) 0.117 147.844 (139.454–154.123) 150.171 (144.287–156.301) 0.836 Kynurenine 2.568 (2.468–3.331) 2.861 (2.057–3.833) 1.000 3-Nitrotyrosine 0.007 (0.006–0.014) 0.008 (0.006–0.009) 0.748 Methionine-sulfoxide 8.263 (6.935–8.601) 6.704 (5.573–8.340) 0.326 Serotonine 7.615 (6.551–8.491) 7.301 (6.348–9.502) 0.834 Putrescine 1.682 (1.609–1.804) 2.729 (1.679–4.414) 0.415 Phenethylamine 0.031 (0.020–0.057) 0.063 (0.051–0.084) 0.230 Aspartic aicd 115.316 (100.544–133.551) 100.064 (81.480–123.055) 0.287 Tryptophan ratio 0.168 (0.160–0.174) 0.154 (0.147–0.162) 0.834 Tyrosine ratio 0.125 (0.121–0.142) 0.175 (0.164–0.197) 0.022 BCAA/AAA ratio 1.520 (1.4245–1.6156) 1.344 (1.199–1.439) 0.069 BTR ratio 5.469 (4.788–5.709) 3.848 (3.307–4.215) 0.016 Fischer’s ratio 2.430 (2.247–2.561) 1.917 (1.751–2.238) 0.039 Kynurenine/tryptophan ratio 0.019 (0.018–0.021) 0.018 (0.014–0.027) 1.000 Serotonin/tryptophan ratio 0.051 (0.047–0.054) 0.049 (0.043–0.067) 0.674 Glutamate/glutamine ratio 0.406 (0.386–0.422) 0.383 (0.346–0.409) 0.574 /tyrosine ratio 1.206 (1.016–1.219) 0.924 (0.828–0.988) 0.069

p value means when compared with Normal group. Tryptophan ratio: tryptophan/(phenylalanine1tyrosine1valine1leucine1isoleucine); Tyrosine ratio: tyrosine/(phenylala- nine1​tryptophan1​valine1​leucine1​isoleucine); BCAA/AAA ratio: (valine1​leucine1​isoleucine)/(tyrosine1​phenylala- nine1​tryptophan); BTR (BCAAs/tyrosine ratio, BTR) ratio: (valine1​leucine1​isoleucine)/tyrosine; Fischer’s ratio: (valine1​ leucine1​isoleucine)/(tyrosine1​phenylalanine). analyzer (Hitachi Ltd., Tokyo, Japan). amino acids (16) were used as internal standards. Liver histology and morphometry. Liver tissue fixed in Sample preparation: Samples were prepared accord- formalin were then paraffin-embedded, and sectioned at ing to a previously reported protocol (17). Rat serum 5 mm. For standard histology, liver sections were stained samples were thawed at 4˚C, and then 10 mL of each with Hematoxylin-Eosin (H&E). The stained sections serum sample was transferred into an eppendorf tube. were observed and photographed under a light micro- Then, 10 mL of ddWater was added into each tube, and scope (with 2003​magnification). 5 mL of the internal standard mixture was added. Next, Serum and liver amino acid analysis using UPLC-TQMS 40 mL of cold isopropanol (with 1% formic acid, v/v) based targeted metabolomics. was added to precipitate proteins and then was vortex Preparation of internal standard: Isotopically labeled mixed. All EP tubes were maintained at 220˚C for Differential Serum and Liver Amino Acids Biomarkers in ALD 539

Fig. 2. 3D PLS-DA score plots show clear separation between the normal and model groups based on serum amino acids. (A) PLS-DA scores plot: R2X50.536, R2Y50.962, Q2Y50.762. (B) Permutation analysis. Class 1, the normal group; Class 2, the model group.

20 min, and then, serum samples were centrifuged at Table 2. VIP value from PLS-DA score plots of differen- 13,000 3g for 10 min. Then, 10 mL of the supernatant tial amino acids in serum and liver from ALD rats. was transferred to a glass HPLC vial for derivatization. The details are shown in Supplemental Online Material, Serum VIP Liver VIP Fig. S1. Serum and liver amino acid analysis using UPLC-TQMS. b-Alanine 1.185 Sarcosine 1.296 UPLC-TQMS analysis was performed using an Acquity Alanine 1.224 b-Alanine 1.099 Serine 1.357 Serine 1.121 UPLC binary solvent manager, sampler manager, and Ornithine 1.709 Proline 1.108 column manager interfaced with a Xevo TQ-S tandem Arginine 2.070 Valine 1.139 quadrupole mass spectrometer. MS/MS detection was Tyrosine 1.250 Cysteine 1.233 via electrospray ionization (ESI) in positive ion mode Threonine 1.049 using multiple reaction monitoring (MRM) for the Ornithine 1.320 quantification of each compound. was used as Lysine 1.262 the desolvation gas, and argon was as the collision gas. Histidine 1.131 The following source conditions were used: capillary Tyrosine 1.096 voltage, 1.5 kV; source offset, 50 V; desolvation tem- Symmetric 1.184 perature, 600˚C; source temperature, 150˚C; desolva- dimethylarginine tion gas flow, 1,000 L/h; cone gas flow, 150 L/h; nebu- Methionine 1.207 lizer gas, 7.0 bar; and collision gas, 0.15 mL/min. The Isoleucine 1.233 chromatographic separation used reversed-phase gradi- Methionine-sulfoxide 1.567 ent chromatography on an HSS T3 2.13​150 mm, 1.8 Putrescine 1.215 mm column. The mobile phase was composed of 0.1% VIP values were from PLS-DA score plots based on serum formic acid in water (v/v) (A) and 0.1% formic acid in amino acids or liver amino acids. acetonitrile (v/v) (B). The column temperature was maintained at 45˚C, and linear gradient elution was performed at 0.6 mL/min starting at 4% B and held for cessed by the TargetLynx application package within 0.5 min before increasing to 10% over 2 min, then to MassLynx software. The raw data was smoothed, and 28% over 2.5 min, and finally increasing to 95% for peak integration was performed using ApexTrak algo- 1 min, before returning to 4% B (1.3 min) for re-equili- rithm. Multivariate models were constructed using par- bration. The weak wash was 95 : 5 water/acetonitrile tial least squares discriminant analysis (PLS-DA) using (v/v), and the strong wash was 100% isopropanol, re- SIMCA-P 11.5 (Umetrics, Sweden). The models were spectively. The QC sample contained all the standards validated using 7-fold cross-validation, and the valid and internal standards. After the injection of a further and robust models were assessed using internal cross- double blank and a single blank, the analysis was validation (R2X, R2Y, Q2Y values) and permutation started with injections of the calibration curve followed analysis. Mann Whitney U test was applied to deter- by a double blank injection. The QC standards were mine if differences observed in concentrations between interspersed evenly throughout the study samples. normal rats and model rats with ALD were statistically Statistical analysis. The raw LC-MS data were pro- significant using SPSS 19.0 software (SPSS, Inc.). The 540 Shi C et al.

Table 3. Altered liver amino acid profiles in rats with ALD.

AAs (mM) Normal Model p

Creatinine 5.839 (5.410–6.176) 6.07 (5.631–6.889) 0.660 Glycine 1768.179 (1754.462–1817.798) 1712.627 (1637.497–1740.441) 0.209 Sarcosine 19.968 (17.1852–24.681) 8.969 (5.568–12.304) 0.046 b-Alanine 52.050 (49.635–63.102) 41.419 (37.298–44.235) 0.016 Alanine 1220.280 (1156.064–1247.212) 1170.644 (1111.875–1191.793) 0.413 Serine 586.737 (555.389–618.892) 494.354 (471.180–513.514) 0.023 Proline 267.594 (249.460–315.611) 225.857 (212.306–235.247) 0.046 Valine 253.386 (246.083–260.989) 216.593 (204.854–224.021) 0.023 Cysteine 0.262 (0.217–0.343) 16.349 (12.668–28.292) 0.004 Threonine 423.203 (384.590–435.981) 337.760 (313.861–355.863) 0.039 Taurine 742.475 (564.617–1259.081) 562.752 (520.654–679.991) 0.413 OH-proline 0.019 (0.000–0.054) 0.000 (0–0.068) 0.822 Asparagine 102.198 (84.263–120.196) 101.783 (71.449–124.305) 0.916 Ornithine 310.036 (280.520–325.742) 206.327 (196.873–229.145) 0.008 Glutamic acid 1082.657 (1028.406–1109.162) 1067.484 (1016.631–1075.221) 1.000 Glutamine 1387.327 (1147.828–1694.527) 1768.887 (1723.294–1827.352) 0.247 Lysine 433.323 (400.901–470.047) 297.070 (251.444–325.135) 0.011 Histidine 392.128 (363.054–417.371) 312.905 (292.961–324.186) 0.033 Arginine 6.184 (4.776–6.880) 6.399 (5.013–7.255) 0.916 Citrulline 8.493 (6.955–9.243) 8.941 (8.078–10.973) 0.587 Tyrosine 205.272 (179.247–224.935) 160.172 (143.473–170.710) 0.043 Asymmetric dimethylarginine 9.372 (8.709–12.032) 8.757 (6.815–14.172) 0.834 Symmetric dimethylarginine 0.745 (0.591–0.803) 0.426 (0.385–0.515) 0.021 Methionine 118.906 (110.031–124.845) 88.405 (83.891–100.300) 0.046 Histamine 0.257 (0.215–0.320) 0.434 (0.341–0.505) 0.674 Isoleucine 170.898 (161.832–181.933) 133.523 (120.610–136.119) 0.029 Leucine 266.0935 (249.9697–268.8068) 237.091 (229.7412–244.8012) 0.137 Phenylalamine 160.340 (142.263–174.960) 140.408 (126.748–149.324) 0.430 Tryptophan 32.637 (29.851–34.680) 28.598 (27.722–30.371) 0.668 Kynurenine 1.479 (1.294–2.309) 0.955 (0.760–1.070) 0.063 3-Nitrotyrosine 0.032 (0.021–0.038) 0.029 (0.0205–0.035) 0.875 Methionine-sulfoxide 6.709 (5.704–6.939) 3.242 (2.994–3.546) 0.002 Serotonine 0.071 (0.054–0.084) 0.063 (0.047–0.138) 0.842 Putrescine 1.009 (0.817–1.631) 2.955 (2.315–5.704) 0.011 Phenethylamine 0.028 (0.020–0.039) 0.018 (0.005–0.022) 0.495 Aspartic aicd 85.618 (68.980–118.820) 118.343 (71.413–159.974) 0.836 Tryptophan ratio 0.031 (0.029–0.032) 0.033 (0.031–0.034) 0.354 Tyrosine ratio 0.230 (0.216–0.252) 0.209 (0.197–0.228) 0.519 BCAA/AAA ratio 1.755 (1.647–1.849) 1.840 (1.692–1.911) 0.749 BTR ratio 3.348 (3.096–3.684) 3.770 (3.353–4.008) 0.574 Fischer’s ratio 1.910 (1.784–2.019) 2.015 (1.853–2.124) 0.620 Kynurenine/tryptophan ratio 0.054 (0.039–0.068) 0.032 (0.029–0.036) 0.079 Serotonin/tryptophan ratio 0.002 (0.002–0.003) 0.003 (0.002–0.005) 0.963 Glutamate/glutamine ratio 0.747 (0.647–0.984) 0.604 (0.586–0.615) 0.091 Phenylalanine/tyrosine ratio 0.752 (0.730–0.906) 0.864 (0.830–0.898) 0.856

p value means when compared with Normal group. Tryptophan ratio: tryptophan/(phenylalanine1​tyrosine1​valine1​leucine1​isoleucine); Tyrosine ratio: tyrosine/(phenylala- nine1​tryptophan1​valine1​leucine1​isoleucine); BCAA/AAA ratio: (valine1​leucine1​isoleucine)/(tyrosine1​phenylala- nine1​tryptophan); BTR (BCAAs/tyrosine ratio, BTR) ratio: (valine1​leucine1​isoleucine)/tyrosine; Fischer’s ratio: (valine1​ leucine1​isoleucine)/(tyrosine1​phenylalanine). data were all described as the mean6​standard devia- the prolongation of modeling time. Compared with the tion (SD) and were analyzed by using nonparametric normal group, serum levels of alanine transaminase test. p,​0.05 was considered statistically significant. (ALT) and aspartate transaminase (AST) in the model group increased significantly (Fig. 1A). Combined with RESULTS pathology changes observed with H&E stain (Fig. 1B), it Liver injury appeared in ALD rats could be determined that the model ALD rats were The rats in the normal group were more active than established successfully. The purpose of adding CCl4 in those in the model group. The model group rats only the model is to make the ALD model more stable and to showed temporary excitement along with daily drink- reduce the experimental period (18). ing, drunk and fallen asleep, and were even worse with Differential Serum and Liver Amino Acids Biomarkers in ALD 541

Fig. 3. 3D PLS-DA score plots show clear separation between the normal and model groups based on liver amino acids. (A) PLS-DA scores plot: R2X50.604, R2Y50.898, Q2Y50.804. (B) Permutation analysis. Class 1, the normal group; Class 2, the model group.

Altered amino acid profiles in rats with ALD ornithine, lysine, histidine, tyrosine, symmetric dimeth- Serum amino acid profiles. A total of 39 amino acids ylarginine, methionine, isoleucine and methionine-sulf- and 9 ratios in normal and ALD rats were determined in oxide were lower in ALD (p,​0.05) compared with the the present study. The detailed serum amino acid pro- control group. According to the VIP value, urine methi- files in rats with ALD are shown in Table 1. It was onine-sulfoxide and ornithine (VIP.​1.3) may be the shown that there were distinct differences between the most important and possible differential metabolites normal and model groups in the 3D PLS-DA scores plot between the two groups. based on serum amino acids (R2X5​0.536, R2Y5​ DISCUSSION 0.962, and Q2Y5​0.762) (Fig. 2A). Permutation analy- sis showed the reliability of the 3D PLS-DA model (Fig. ALD is a toxic liver disease caused by excessive or 2B). The VIP values from the PLS-DA score plots of dif- long-term drinking, and it is characterized by a rapid ferential amino acids in serum from ALD rats are shown decline in liver function following a catastrophic insult in Table 2. Differential metabolites were obtained under to the liver (16). The liver plays a major role in amino the principle of VIP value.1 from PLS-DA model and acid metabolism and is central to the regulation of met- p,​0.05 from the nonparametric test. Serum b-alanine, abolic pathways. It is also responsible for the metabo- alanine, serine, ornithine, tyrosine and tyrosine ratio lism of hormones that affect protein, carbohydrate and were significantly higher in ALD (p,​0.05), while argi- lipid metabolism. Chronic and acute liver diseases can nine, branched chain amino acids (BCAAs)/tyrosine profoundly alter the nutritional status and amino acid ratio (BTR) ratio and Fischer’s ratio were lower in ALD metabolism of patients with ALD (19). (p,​0.05) compared with the normal control group. In the present study, we have demonstrated the According to the VIP value, serum ornithine, arginine changes of serum and liver levels of 39 amino acids and and serine (VIP.​1.3) may be the most important and 9 amino acid ratios in normal rats and ALD rats which possible differential metabolites between the two groups. were quantitatively determined. Liver amino acid profiles. The detailed liver amino It is realized that the immunity mechanism is related acid profiles in rats with ALD are shown in Table 3. It to the pathogenesis and progression of ALD, while the was shown that there were distinct differences between amino acids constitute to the basic structural sub- the normal and model groups in the 3D PLS-DA scores stances of the immune system. When alcohol is con- plot based on liver amino acids, and the related parame- sumed in large quantities, the immune system is ters were R2X5​0.604, R2Y5​0.898, and Q2Y5​0.804 destroyed and the levels of amino acids show corre- respectively (Fig. 3A). Permutation analysis showed the sponding changes. reliability of the 3D PLS-DA model (Fig. 3B). The VIP In the detection of serum amino acids, the levels of values from PLS-DA score plots of differential amino b-alanine, alanine, serine, ornithine, arginine, tyrosine, acids in liver from ALD rats are shown in Table 2, and tyrosine ratio, BTR ratio and Fischer’s ratio were signifi- differential metabolites were obtained under the princi- cantly different between the two groups. In the detec- ple of VIP value.1 in the PLS-DA model and p,​0.05 tion of liver amino acids, the levels of sarcosine, b-ala- from the nonparametric test. Liver cysteine and putres- nine, serine, proline, valine, cysteine, threonine, orni- cine were significantly higher in ALD (p,​0.05), while thine, lysine, histidine, tyrosine, methionine, isoleucine, sarcosine, b-alanine, serine, proline, valine, threonine, symmetric dimethylarginine, methionine-sulfoxide and 542 Shi C et al.

Fig. 4. Summary of the alteration of serum and liver amino acids in ALD rats. putrescine were significantly different between the two diseases progress, the imbalance of amino acids tends groups (Fig. 4). Among these, four serum amino acids to become more marked, and aminograms are useful for had levels that were contrary to the levels observed in assessing the prognosis of cirrhotic patients with or liver tissue; these were b-alanine, serine, ornithine and without hepatocellular carcinoma (HCC). tyrosine. The levels were significantly higher in serum, Differential liver amino acid in ALD rats while lower in liver in ALD model group compared with ALD rats showed a significant decrease in the levels normal control group. However, the detailed mecha- of isoleucine and valine in liver than normal rats. These nism for this difference is unclear. two BCAAs are among the nine essential amino acids. Differential serum amino acid in ALD rats According to D’Antona et al. (24), the increase of Amino acids are metabolized to provide energy and BCAAs may prolong the lifespan of male mice, which are also used to synthesize proteins, glucose, and/or was related to enhancing mitochondrial biogenesis other bioactive molecules (20). The serum level of ala- and reducing ROS production by upregulating the ex­ nine in ALD rats was significantly increased than that pression of peroxisome proliferator-activated receptor of normal rats, while the serum level of arginine was (PPAR) g coactivator-1a (PGC-1a). Thus, it is supposed markedly decreased. Arginine can promote the orni- that the pathogenesis of ALD may be linked to ROS tis- thine cycle to convert more blood ammonia to urea, sue damage. BCAAs were shown to be able to reduce which can expedite the elimination and metabolism of hepatic apoptosis, promote hepatocyte regeneration blood ammonia (21). The decrease of arginine is a clear and stimulate the production of hepatocyte growth fac- indicator of liver disease. tor. As a result, we found that BCAAs supplement may The tyrosine ratio is the ratio of serum tyrosine to be a possible treatment approach to delay the progres- valine, leucine, isoleucine and phenylalanine, which sion of chronic liver injury. On the other hand, as two exhibited a slight increase in ALD rats. The BTR ratio is of the nine essential amino acids, they have stimulatory the ratio of serum valine, leucine and isoleucine to tyro- effects on hepatic protein synthesis and an inhibitory sine, and it was significantly decreased in ALD rats com- effect on proteolysis. The decrease of BCAAs in ALD pared to the control rats. Fischer’s ratio indicates the also indicates that alcohol may cause a deficiency of ratio of serum valine, leucine and isoleucine to tyrosine nutrient intake (19). Zhao et al. (25) found that BCAAs and phenylalanine, which showed a decrease in ALD exacerbated obesity-related hepatic glucose and lipid rats compared to control rats (22). A low Fischer’s ratio metabolic disorders through attenuating Akt2 signal- has been shown associated with hepatic encephalopa- ing. Tedesco et al. (26) found that enriched in BCAAs thy. All the above three ratios are considered good indi- reverted these molecular defects and mitochondrial dys- cators of the severity of hepatic injury (23). In liver dis- function, suggesting that the mitochondrial integrity ease, the depletion of the BCAAs, including valine, leu- obtained with the amino acid supplementation could be cine, and isoleucine, and increased concentrations of mediated through a Sirt1-eNOS-mTOR pathway. tyrosine, may affect the regulation of metabolic path- Moreover, amino acids are the major nitrogen source ways. The perturbation of the urea cycle and a limited for glutamine synthesis in muscles. The synthesis of capacity for detoxification of ammonia with glutamine glutamine is activated during critical illnesses, such as are associated with and, in particular, cancer and trauma. However, the body demands for with hepatic encephalopathy, although the mechanism glutamine in certain physiologic conditions are enor- and prognostic significance remains unclear. As liver mous, and increased utilization of glutamine often Differential Serum and Liver Amino Acids Biomarkers in ALD 543 exceeds its synthesis, which finally results in its defi- TQMS-based method. Tedesco et al. (26) also used ciency in plasma and muscles. The needs of BCAAs for UPLC-MS (AB Sciex, Milan, Italy) to detect amino acids synthesis of glutamine are connected with the break- in rodents, but they could only detect 15 amino acids down of muscle proteins which results in muscle-pro- simultaneously, including threonine, asparagine, tyro- tein wasting. It is reported that glutamine synthesis sine, serine, glycine, alanine, leucine, isoleucine, valine, relies on in the cytosol of pericen- proline, histidine, methionine, , glutamine, tral , where it ensures the clearance of phenylalanine, glutamic acid, lysine, arginine, and tryp- ammonium and then controls blood ammonium con- tophan. However, our methods could simultaneously centration (27, 28). The decrease in the levels of liver quantitatively detect the levels of 39 amino acids and 9 amino acids may lead to the lack of apolipoprotein syn- amino acid ratios using UPLC-TQMS based targeted thesis, so that excretion may be disordered, metabolomics. Furthermore, our findings demonstrate which was the possible cause of ALD. that determination of certain amino acids which can be A significant decrease of sarcosine levels occurred in used to detect the diagnosis of ALD. the of rats with ALD. Sarcosine is an acidic prod- However, there are still several limitations in the pres- uct of the liver that can provide energy for muscular ent study. First, the number of the rats in each group is cells (https://nootriment.com/sarcosine/). relatively small. Second, amino acid metabolomics anal- In the present study, glycine was significantly lower, ysis from patients with ALD would be better suited to while cysteine was higher in the livers of ALD rats than illustrate the abnormal amino acid in ALD in normal control rats. The exact reason for such phe- than rat models. Furthermore, the detailed understand- nomena is yet to be found. Cysteine and glycine are two ing of the mechanistic roles of specific amino acids in amino acids that are required to synthesize hepatic glu- the pathogenesis of ALD needs to be evaluated in future tathione (GSH). GSH is the most important antioxidant research. In the future we will perform further research molecule in the human body, and it can be used to com- to understand the mechanistic roles of specific amino bat free radical damage (29). Alcohol can produce enor- acids in the pathogenesis of ALD. We will further mous free radicals during its metabolism in microsomes explore the related enzymes which involved in the through cytochrome P4502E1 and P2E1 (30). Oxida- metabolism of serum ornithine, arginine and serine, tive stress damage is the most important pathogenic urine methionine-sulfoxide and ornithine, according to mechanism of ALD. However, the experimental results the information from KEGG pathway. As for ornithine, it did not support this pathogenic pathway, which remains is shown that four enzymes including ornithine race- to be evaluated in further studies. mase, amino acid racemase, d-ornithine 4,5-aminomu- Methionine was lower in the livers of ALD rats com- tase subunit beta, d-arginase are involved in the synth- pared with normal rats, which may also be caused by sis of ornithine, another three enzymes including ala- oxidative stress damage. Alcohol and its metabolic pro- nine transaminase, d-ornithine/d-lysine decarboxylase, duction can produce oxidative free radicals that damage d-amino-acid oxidase are involved in the metabolism of the activity of methionine synthetase (MA) and S-ade- ornithine. We can further perform RNA sequence to nosyl methionine synthetase (MAT), which influences find differential molecules, and then use modern tech- the synthesis of methionine and S-adenosyl methionine nologies such as CRISPR/CAS9, targeted metabolomics, (SAM). Therefore, the disorder of liver methionine is a adenovirus transfection through animal and cell mod- probable contributor to the pathogenesis of ALD, and els by focusing on related molecules which are involved the significant decrease in the level of methionine is a in those differential metabolites to further explore the potential diagnostic for ALD. deep mechanisms. Ornithine showed a marked decrease in the liver of In conclusion, the application of targeted UPLC-MS/ ALD rats compared to the level in normal controls. It is MS analysis provides a useful approach to study the key one of the major amino acids in the ornithine cycle, effect of amino acids in the pathogenesis of ALD. Amino which is mainly located in the liver. The liver can elimi- acids can be considered as specific diagnostic biomark- nate the toxicity of ammonia by synthesizing urea ers for ALD. The metabolic patterns of amino acids can through the ornithine cycle. In that case, a decrease of provide theoretical evidence for pathogenic pathways in ornithine means a decrease in the synthesis of urea, ALD and can offer new ideas and targets for the preven- which may mean an impaired detoxifying function of tion and treatment of ALD. the liver. Thus, a decrease in ornithine may be a possible diagnostic biomarker for ALD. Authorship Putrescine showed a slight increase in the liver of T.W. initiated and supervised the whole manuscript; ALD rats compared to that of normal rats. Putrescine is C.Z.S., L.W. and K.J.Z. performed the experiment; K.J.Z. synthesized in small quantities by healthy living cells by and T.W. analyzed the data; M.M.S. and Y.F.L. partici- the action of . The slight pated in the data collection; C.Z.S. and L.W. wrote the increase in putrescine could reflect prophase, which manuscript; K.J.Z. and T.W. revised the manuscript. All may contribute to the occurrence of ALD. authors have read and approved the final manuscript. According to Prystupa et al. (19) in 2015, the amino C.Z.S and L.W. are co-first authors. The corresponding acids were determined by automated ion-exchange authors had full access to all the data in the study and chromatography, which is different from our UPLC- had final responsibility for the decision to submit for 544 Shi C et al. publication. 16) Gray N, Zia R, King A, Patel VC, Wendon J, McPhail MJ, Coen M, Plumb RS, Wilson ID, Nicholson JK. 2017. Disclosure of state of COI High-speed quantitative UPLC-MS analysis of multiple The authors have no conflicts of interest to declare. amines in human plasma and serum via precolumn derivatization with 6-aminoquinolyl-N-hydroxysuccin- imidyl carbamate: Application to acetaminophen-in- Acknowledgments duced liver failure. Anal Chem 89: 2478–2487. This work was supported by the National Natural Sci- 17) Zhou K, Xie G, Wen J, Wang J, Pan W, Zhou Y, Xiao Y, ence Foundation of China (81873076, 81973725), the Wang Y, Jia W, Cai W. 2016. Histamine is correlated Shanghai Rising-Star Project (15QA1403500) and the with liver fibrosis in biliary atresia. Dig Liver Dis J 48: Shanghai Talents Development Fund Project in China 921–926. (2017090). 18) Ji G, Wang L, Zhang SH, Liu JW, Zheng PY, Liu T. 2006. Effect of Chinese medicine Qinggan Huoxuefang on Supporting information inducing HSC apoptosis in alcoholic liver fibrosis rats. Supplemental online material is available on J-STAGE. World J Gastroenterol 12: 2047–2052. 19) Prystupa A, Szpetnar M, Boguszewskaczubara A, Grzy- REFERENCES bowski A, Sak J, Załuska W. 2015. Activity of MMP1 1) Liu GT, Zhu YC, Zhang T, Wang J, Yan T, Lv W, Zhou DH. and MMP13 and amino acid metabolism in patients 2017. Advances in research of . with alcoholic liver cirrhosis. Med Sci Moni Intern Med J World Chinese Journal of Digestology 25(15): 1382. Exp Clin Res 21: 1008–1014. 2) Torruellas C, French SW, Medici V. 2014. Diagnosis of 20) Adeva MM, Souto G, Blanco N, Donapetry C. 2012. alcoholic liver disease. World J Gastroenterol 20: 11684– Ammonium metabolism in humans. Metabolism 61: 11699. 1495–1511. 3) Yin G, Kuang JE. 2016. Research progress on pathogen- 21) Andrew PJ, Mayer B. 1999. Enzymatic function of nitric esis of alcoholic liver disease. J Changsha Med Univ 30: oxide synthases. Cardiovasc Res 43: 521–531. 160–163. 22) Morgan MY, Milsom JP, Sherlock S. 1978. Plasma ratio 4) O’Shea RS, Dasarathy S, McCullough AJ. 1995. Alco- of valine, leucine and isoleucine to phenylalanine and holic liver disease. Hepatology 51: 227. tyrosine in liver disease. Gut 19: 1068–1073. 5) Karsan HA, Parekh S. 2012. Management of alcoholic 23) Gao J, Guo J, Yu C. 2004. Changes of amino acid metab- : Current concepts. World J Hepatol 4: 335–341. olism in alcoholic fatty liver rats. Journal of Preventive 6) Beier JI, Arteel GE, McClain CJ. 2011. Advances in alco- Medicine of Chinese People’s Liberation Army 22(3): 167– holic liver disease. Curr Gastroenterol Rep 13: 56–64. 170. 7) Mounika N, Vijayalaximi T, Qadrie ZL. 2018. Compre- 24) D’Antona G, Ragni M, Cardile A, Tedesco L, Dossena M, hensive review of alcoholic liver disease. International Bruttini F, Caliaro F, Corsetti G, Bottinelli R, Carruba Journal of Pharmacy & Pharmaceutical Research. Human MO. 2010. Branched-chain amino acid supplementa- 12(1): 358–373. tion promotes survival and supports cardiac and skele- 8) Wang X, Luo J, Zha X. 2014. Application of proteomics tal muscle mitochondrial biogenesis in middle-aged and metabonomics in studies on alcoholic liver disease. mice. Cell Metabolism 12: 362–372. Journal of Clinical Hepatology 030(002): 168–173. 25) Zhao H, Zhang F, Sun D, Wang X, Zhang X, Zhang J, 9) Zhao YY, Cheng XL, Vaziri ND, Liu S, Lin RC. 2014. Yan F, Huang C, Xie H, Lin C, Liu Y, Fan M, Yan W, Chen UPLC-based metabonomic applications for discovering Y, Lian K, Li Y, Zhang L, Wang S, Tao L. 2020. Branched- biomarkers of diseases in . Clin Bio- chain amino acids exacerbate obesity-related hepatic chem 47: 16–26. glucose and lipid metabolic disorders via attenuating 10) Crawford JM. 2012. Histologic findings in alcoholic liver Akt2 signaling. 69(6): 1164–1177. disease. Clin Liver Dis 16: 699–716. 26) Tedesco L, Corsetti G, Ruocco C, Ragni M, Rossi F, Car- 11) Wu G. 2009. Amino acids: metabolism, functions, and ruba MO, Valerio A, Nisoli E. 2018. A specific amino nutrition. Amino Acids 37: 1–17. acid formula prevents alcoholic liver disease in rodents. 12) Mukherjee S, Vaidyanathan K, Vasudevan DM, Das SK. Am J Physiol Gastrointest Liver Physiol 314(5): G566– 2010. Role of plasma amino acids and gaba in alcoholic G582. and non-alcoholic —a pilot study. 27) Morris SM Jr. 2002. Regulation of enzymes of the urea Indian J Clin Biochem 25: 37–42. cycle and arginine metabolism. Annu Rev Nutr 22: 13) Liang Q, Wang C, Li B, Zhang AH. 2015. Metabolomics 87–105. of alcoholic liver disease: A clinical discovery study. Rsc 28) Massafra V, Milona A, Vos HR, Ramos RJJ, Gerrits J, Wil- Advances 5: 80381–80387. lemsen ECL, Ramos Pittol JM, Ijssennagger N, Houwel- 14) Harada S, Takebayashi T, Kurihara A, Akiyama M, ing M, Prinsen H, Verhoeven-Duif NM, Burgering BMT, Suzuki A, Hatakeyama Y, Sugiyama D, Kuwabara K, van Mil SWC. 2017. Farnesoid X receptor activation Takeuchi A, Okamura T. 2016. Erratum to: Metabolo- promotes hepatic amino acid catabolism and ammo- mic profiling reveals novel biomarkers of alcohol intake nium clearance in mice. Gastroenterology 152: 1462– and alcohol-induced liver injury in community-dwelling 1476 e10. men. Environ Health Prev 21: 18–26. 29) Paolicchi A. 2003. The changing faces of , a 15) Wu T, Chen JM, Xiao TG, Shu XB, Xu HC, Yang LL, Xing cellular protagonist. Biochem Pharmacol 66: 1499–1503. LJ, Zheng PY, Ji G. 2016. Qinggan Huoxue Recipe sup- 30) Seth D, Haber PS, Syn WK, Diehl AM, Day CP. 2011. presses epithelial-to-mesenchymal transition in alco- Pathogenesis of alcohol-induced liver disease: classical holic liver fibrosis through TGF-b1/Smad signaling concepts and recent advances. J Gastroenterol Hepatol pathway. World J Gastroenterol 22: 4695. 26: 1089–1105.

Top View