Gehua Jiejiu Dizhi decoction (葛花解酒涤脂汤) ameliorates alcoholic fatty liver in mice by regulating lipid and bile acid metabolism and with exertion of antioxidant stress based on 4DLabel-free quantitative proteomic study

Abstract

OBJECTIVE:

To analyze the effect and molecular mechanism of Gehua Jiejiu Dizhi decoction (葛花解酒涤脂汤, GJDD) on alcoholic fatty live disease (AFLD) by using proteomic methods.

METHODS:

The male C57BL/6J mouse were randomly divided into four groups: control group, model group, GJDD group and resveratrol group. After the AFLD model was successfully prepared by intragastric administration of alcohol once on the basis of the Lieber-DeCarli classical method, the GJDD group and resveratrol group were intragastrically administered with GJDD (4900 mg/kg) and resveratrol (400 mg/kg) respectively, once a day for 9 d. The fat deposition of liver tissue was observed and evaluated by oil red O (ORO) staining. 4DLabel-free quantitative proteome method was used to determine and quantify the protein expression in liver tissue of each experimental group. The differentially expressed proteins were screened according to protein expression differential multiples, and then analyzed by Gene ontology classification and Kyoto Encyclopedia of Genes and Genomes pathway enrichment. Finally, expression validation of the differentially co-expressed proteins from control group, model group and GJDD group were verified by targeted proteomics quantification techniques.

RESULTS:

In semiquantitative analyses of ORO, all kinds of steatosis (ToS, MaS, and MiS) were evaluated higher in AFLD mice compared to those in GJDD or resveratrol-treated mice. 4DLabel-free proteomics analysis results showed that a total of 4513 proteins were identified, of which 3763 proteins were quantified and 946 differentially expressed proteins were screened. Compared with the control group, 145 proteins were up-regulated and 148 proteins were down-regulated in the liver tissue of model group. In addition, compared with the model group, 92 proteins were up-regulated and 135 proteins were down-regulated in the liver tissue of the GJDD group. 15 differentially co-expressed proteins were found between every two groups (model group vs control group, GJDD group vs model group and GJDD group vs control group), which were involved in many biological processes. Among them, 11 differentially co-expressed key proteins (Aox3, H1-5, Fabp5, Ces3a, Nudt7, Serpinb1a, Fkbp11, Rpl22l1, Keg1, Acss2 and Slco1a1) were further identified by targeted proteomic quantitative technology and their expression patterns were consistent with the results of 4D label-free proteomic analysis.

CONCLUSIONS:

Our study provided proteomics-based evidence that GJDD alleviated AFLD by modulating liver protein expression, likely through the modulation of lipid metabolism, bile acid metabolism and with exertion of antioxidant stress.

1. INTRODUCTION

Alcoholic fatty live disease (AFLD) is the key pathological premise for the progressive development and treatment of alcoholic liver disease (ALD).^1,2 Despite extensive research on the pathophysiology of alcohol induced fatty liver, there is still no targeted treatment over the last 50 years, except for temperance, as well as symptomatic and nutritional support.^3,4 In view of this important clinical problem, based on the understanding of AFLD of Traditional Chinese Medicine (TCM), the Gehua Jiejiu Dizhi decoction (葛花解酒涤脂汤, GJDD) was employed by our research group, which was composed of 9 kinds of Traditional Chinese Herbal Medicine, such as Gehua (Flos Puerariae Lobatae), Fuling (Poria), Mianyinchen (Herba Artemisia Capillaris), Xiaoqingpi (Fructus Citri Reticulatae Immaturus), Gegen (Radix Puerariae Lobatae), Juemingzi (Semen Cassiae Obtusifoliae), Shanzha (Fructus Crataegus Pinnatifidae), Baifan (Alumen) and Gancao (Radix Glycyrrhizae). In this prescription, various drugs cooperate to play the anti-AFLD effects of strengthening the spleen and infiltrating dampness, clearing heat and detoxification, soothing the liver and resolving depression, eliminating food and regulating Qi, removing blood stasis and dredging collaterals, harmonizing the stomach and reducing turbidity, generating saliva and relieving thirst, and clearing fat. In order to explore the effect and possible mechanism of GJDD in the treatment of AFLD, we first analyzed and found the potential active compounds and possible target of GJDD in the treatment of AFLD using the network pharmacology analysis method.⁵ More importantly, our previous studies showed that the GJDD could significantly ameliorate alcohol-induced hepatic steatosis in C57BL/6J mice, restore the levels of serum free cholesterol, triglyceride and alanine aminotransferase to normal levels, and significantly up-regulate the mRNA and protein expression levels of liver X receptor alpha (LXRα), cytochrome P450 3A25 (CYP3A25), pregnane X receptor (PXR) in AFLD mice liver tissue.^6,⇓-8 However, the potential AFLD protective mechanisms of GJDD remain to be further clarified to support its clinical application.

As we all know, diseases are often accompanied by protein dysfunction at multiple levels.^9,10 Therefore, it is of great significance to study the occurrence and prevention of diseases from the protein level. Liver is the main organ of alcohol metabolism and damage, and the damage induced by alcohol intake is accompanied by distinct changes of the protein expression profile.¹¹ Using a variety of protein mass spectrometry techniques, such as 2-DE + matrix-assisted laser desorption lonization-time of flight (MALDI-TOF)/TOF, 2-DE + MALDI-mass spectrometry (MS), 2-DE + liquid chromatography-mass spectrometry (LC-MS)/MS,^12,⇓-14 separately revealed the protein expression changes in rat or mouse liver after different degrees of alcohol intake at the tissue level. However, there is still a lack of understanding about the changes of liver protein expression profile in alcoholic fatty liver disease, an important pathological initial stage of ALD.¹⁵ The progress of proteomics technology makes it possible to screen differentially expressed proteins related to pathogenesis, clinical diagnosis, treatment, and prognosis at the same time.¹⁶ Song's research group analyzed the effect of Cili (Rosa Roxburghii) intake on protein expression in liver tissue of hyperlipidemia mice by proteomic method, and found 15 proteins related to lipid metabolism.¹⁷ Similarly, using quantitative chemoproteomic profiling, baicalin extracted from Huangqin (Radix Scutellariae Baicalensis) activated hepatic carnitine palmityl transferase 1 to achieve the intervention effect on metabolic diseases such as liver fatty degeneration and obesity.¹⁸ Using protein quantitative proteomics, Du et al ¹⁹ found that the lysine malonylation could be closely related to type 2 diabetes treatment. Using protein post-translational modification techniques, Liu et al ²⁰ revealed that USP14 played a role in nonalcoholic fatty liver by stabilizing its substrate fatty acid synthase. More and more literatures have confirmed that proteomics technology provides a promising and powerful tool for the study of disease pathogenesis and drug intervention mechanism.

Therefore, in this study, based on our previous findings that GJDD can significantly improve liver steatosis in AFLD mice and the application of proteomics in the study of multiple metabolic diseases, we intend to use 4DLabel-free quantitative proteomics to understand the proteins associated with AFLD and explore the therapeutic mechanism mediated by GJDD.

2. MATERIALS AND METHODS

2.1. Drug preparation

GJDD, a TCM compound, is composed of Gehua (Flos Puerariae Lobatae) 18 g, Fuling (Poria) 19 g, Mianyinchen (Artemisia capillaris) 15 g, Xiaoqingpi (Citri Reticulatae Pericarpium Viride) 15 g, Gegen (Radix Puerariae) 12 g, Juemingzi (Cassiae Semen) 12 g, Shanzha (Fructus Cartaegi) 12 g, Baifan (Alumen) 4 g and Gancao (Radix Glycyrrhizae) 9 g. It is provided by the pharmacy room of the Second Affiliated Hospital of Guizhou University of TCM. Its application methods could be consulted in previous published papers.⁸ Resveratrol (Sigma, St. Louis, MO, USA, R5010) was prepared with high-pressure sterilized 30% koliphor HS15 (BASF, 30944788Q0).

2.2. Animal experiment design

Thirty-two, 9-week-old specific pathogen free (SPF) C57BL/6J male mice were provided by Viewsolid Biotech Co., Ltd., with experimental animal production license No. SCXK (Beijing) 2016-0009. They were divided into control group (C), AFLD group (M), GJDD group (P) and resveratrol group (R) according to the random number table method, with suitable feeding temperature, humidity and clock. Control group mice were fed with SPF feed, and those in the other groups were first used to prepare the alcoholic fatty liver model with reference to the National Institute on Alcohol Abuse and Alcoholism (NIAAA) method.²¹ This method is based on the classical Lieber-DeCarli method. The process includes 5 d of adaptation period, 10 d of model preparation period and 1 time of 95% ethanol gavage on the 16th day. Lieber-DeCarli alcohol liquid feed TP4030D and its control liquid feed TP4030C were purchased from Nantong Teluofei feed Technology Co., Ltd. (Nantong, China). After the successful preparation of the AFLD model, the mice in the intervention groups were orally administered GJDD (4900 mg/kg per day) and resveratrol (400 mg/kg per day) respectively once a day for a total of 9 d, while the control group mice were given distilled water by gavage. Resveratrol, as a sirtuins1 (SIRT1) agonist, is the most widely studied and promising dietary polyphenol in the treatment of alcoholic liver disease^22,23 and play a role in liver protection after alcohol intake through various ways, such as antioxidant stress, activation of liver SIRT1 signal, inhibition of apoptosis, induction of autophagy and so on.^24,⇓-26 There is still no other recommended specific treatment drug for AFLD over the last 50 years.^3,4 So, resveratrol was chosen as the positive drug. Twenty-four hours after the final administration of medication, liver tissues were collected from each group of mice for the preparation of frozen sections, with the remaining tissues stored at －80℃. Institutional and national guidelines for the care and use of animals were followed and all experimental procedures involving animals were approved by the institutional animal ethical committee of Guizhou Medcial University (Permit Number: SYXK2015 -0001).

2.3. Analysis of liver tissue steatosis in mice

3-μm frozen slices of mice liver tissue were made and then oil red O (ORO) staining was used to detect liver lipid deposition. According to the literature,^27,⇓-29 the degrees of steatosis in each group mice liver tissue were compared by semi-quantitative analysis of macro-(MaS)/micro-(MiS) and total steatosis (ToS), which was derived by two experienced pathologists (XIANG Yining and TONG Pingzhen) separately to determine the percentage of steatosis to total hepatocytes based on MaS, MiS and ToS. A single lipid droplet in hepatocytes that is equal or larger than to the nucleus and causes the nucleus to the periphery was defined as MaS. Those in hepatocytes that were smaller than the nucleus and at least one was considered as Mis. ToS is the sum of MaS and MiS. Among them, MaS is evaluated according to the three degrees of mild, moderate, and severe respectively. The number of affected hepatocytes less than 30% was defined as mild, between 30%-60% as moderate, and more than 60% as severe MaS.

2.4. 4DLa bel-free quantitative proteomic detection and analysis of protein expression profile in mouse liver tissue

Three mouse liver tissue samples were randomly taken from each group, and a total of 12 samples were subjected to 4DLabel-free proteomics analysis. The detection and analysis processes are as follows: protein preparation, trypsin digestion, high performance liquid chromatography (HPLC) fractionation, LC-MS/MS analysis, and bioinformatics analysis.

2.4.1. Protein preparation and trypsin hydrolysis

Each liver tissue samples (50 mg) was ground into powder with liquid nitrogen, and then 4 volumes cracking buffer (8M urea, 1% protease inhibitor, 3 μM Trichostatin A, 50 mM Nam) were added for high-intensity ultrasonic cracking. After centrifugation at 4 ℃for 10 min at 12000 × g, the supernatant was collected for protein concentration determination by bicinchoninic acid method (Biyuntian, Shanghai, China). For digestion, took an equal amount of protein sample and an appropriate amount of standard protein mixture, added trichloroacetic acid from Sigma (St. Louis, MO, USA). with a final concentration of 20% v/v, fully mixed and precipitated at 4 ℃ for 2 h. After centrifugation at 500×g for 5 min, the samples were washed and precipitated with precooled acetone. Ultrasonic dispersion was conducted in the final concentration of 200 mM triethylammonium bicarbonate (TEAB) solution, then added trypsin according to the ratio of protease: protein = 1∶50. DTT from Sigma (St. Louis, MO, USA) was added with a final concentration of 5 mM after overnight enzymatic hydrolysis and then reduced at 56 ℃ for 30 min. Finally, iodoacetamide (IAA) was added with a final concentration of 11 mM.

2.4.2. Peptide separation and detection of LC-MS/MS

The tryptic peptides were dissolvent in 0.1% formic acid from Fluka (St. Louis, MO, USA) and 2% acetonitrile (solvent A), directly loaded onto a home-made reversed-phase analytical column. Then using a nanotube ultra-high-performance liquid chromatography (UHPLC) System with 0.1% formic acid and 100% acetonitrile (solvent B) gradient increases conducted as follows: Set the flow rate to 300 NL/min, and performed it according to 2%-5% B,0-1 min; 5%-27% B 1-76 min; 27%-35% B, 76-82 min and 35%-85% B, 82-86 min. The Peptides were subjected to capillary HPLC to concentrate them followed by mass spectrometry using a Tims TOF Pro system. The separated peptides were ionized by capillary ion source with a voltage of 1.75kv, and then analyzed by primary and secondary (with a scan range of 100-1700) tandem mass spectrometry, which dynamic exclusion time scan was 30 s.

2.4.3. Database search

The resulting MS/MS data was retrieved using Maxquant (v1.6.15.0). With protein sequences in the us_musculus_10090_SP_20201214. Tandem mass spectra were searched against the Swissprot Mouse database (17063 sequences), the secondary spectrum generated by mass spectrometry was searched and compared with the theoretical secondary spectrum map, and the matching theoretical peptide sequences were obtained by algorithmic scoring and filtering to complete the sample protein identification. The protein information could be recognized by the identified protein-specific peptides. The false discovery rate for proteins and peptides was adjusted to 1%.

2.4.4. Differential expression protein analysis

The mean ratio of repeated quantitative values of each protein in the three sample pairs of AFLD group/control group (M/C), GJDD group/AFLD group (P/M), and GJDD group/control group (P/C), respectively, was used as the fold change of difference (FC). The relative quantitative value of each protein in the two sample pairs was analyzed by t-test after log2 logarithmic conversion. The screening of differentially expressed proteins with significant up-regulation or down-regulation must meet the two conditions of P-value ≤ 0.05 and the FC over 2 or below 0.5. First, the differentially expressed proteins in each sample pair were screened. Second, the proteins significantly differentially expressed in the three sample pairs were used as differentially co-expressed proteins.

2.4.5. Bioinformatics analysis

We conducted Gene Ontology (GO) Consortium-based vocabulary of terms. Subcellular biological process, cellular compartment and molecular function, were all performed using GO terms and Kyoto Encyclopedia of Genes and Genomes (KEGG) pathways analysis of differentially expressed proteins in each comparison group. Histogram and bubble polt were used to show whether these differentially expressed proteins had a significant enrichment trend in some functional types (P ≤ 0.05, Fisher's exact test).

2.5. Identification of differentially expressed proteins based on targeted proteome quantitative assay

Targeted proteome quantification based on LC-MS/MS was performed for quantitative confirmation of the co-expressed proteins from the above screen. Mobile phase A: 0.1% (V / V) formic acid aqueous solution, Mobile phase B: 0.1% formic acid and 90% acetonitrile aqueous solution. The above peptides for proteomic analysis were dissolved in mobile phase A and separated by an EASY-nLC 1000 ultra-high performance liquid system with a flow rate of 500 l/min. Liquid phase gradient setting: 7%-25% B, 0-16 min; 25%-35%, 16-22 min B; 35%-80% B, 22-26 min; 80% B, 26-30 min. Subsequently, the mass spectrometry detection of peptides was carried out with the procedures and parameters of reference.³⁰ The next data acquisition mode uses data independent scanning (DIA) program, and the final data processing is carried out according to the corresponding conventional methods.¹⁷

2.6. Statistical analyses

All data were analysed with SPSS 26.0 (IBM Corp., Armonk, NY, USA). Quantitative data are expressed as the mean ± standard deviation ($\bar{x}±s$). Wilcoxon signed-rank sum test was used to analyze ORO semiquantitative data. Student's t test was used to analyze the proteomic data, and the P ≤ 0.05 was considered to have significant difference.

3. RESULTS

3.1. Semiquantitative evaluation of GJDD on hepatic steatosis in mice

The results of ORO staining under light microscope showed that diffuse and red dropwise lipid vacuoles were found in the liver tissue of AFLD mice, but significantly decreased in GJDD and resveratrol-treated mice, in the form of small drops and no medium or above lipid droplets were found, especially in the GJDD group, the red small lipid droplets were scattered (Figure 1). Similarly, the results of ORO semiquantitative analysis of ToS, MaS, and MiS in liver tissue of AFLD mice were also significantly higher than those in GJDD and resveratrol-treated mice. Mild MaS was seen in 32.2% of AFLD mice, but no MaS was seen in the drug intervention group, only a small amount of MiS was seen (Table 1).

Figure 1. Hepatic steatosis of liver in 4 groups of mice by ORO staining (× 100 magnification).

A: control group:treated only withfree drinking water; B: AFLD model group:treated with ethanol and drinking water (2.6 g/kg); C: GJDD intervention group: treated with GJDD of 4.9 g/kg; D: resveratrol intervention group: treated with resveratrol of 4 g/kg. Diffuse and red dropwise lipid vacuoles were found in the AFLD mice liver tissue, but they were significantly reduced both in GJDD and resveratrol-treated mice, accompanied by small drops, especially in the GJDD group. Semiquantitative analysis of fatty degeneration showed that AFLD mice were mainly MiS, and the moderate MaS was observed also in 32.2% of mice. Correspondingly, GJDD group and resveratrol group showed a small amount of MiS and no MaS. Black arrow indicates large steatosis in model group, but not in other groups (× 200 magnification). ORO: oil red o; MaS: macrovesicular steatosis; MiS: microvesicular steatosis.

Table 1.

Results of semiquantitative analysis of liver steatosis by ORO staining in each group ($\bar{x}±s$)

Group	n	MaS (%)			MiS (%)	ToS (%)
		Mild	Moderate	Severe
Control	8	0	0	0	2.0±1.2	2.0±1.2
Model	8	10.0±2.5a	32.4±6.5a	0	36.3±7.6a	68.8±10.4a
GJDD	8	0.9±1.2	0.3±6.0	0	5.0±2.6b	7.2±2.7b
Resveratrol	8	0.9±2.3	0.3±4.0	0	10.3±2.7b	18.6±3.2b

Notes: control group: treated only with free drinking water; AFLD model group: treated with ethanol and drinking water (2.6 g/kg); GJDD intervention group: treated with GJDD of 4.9 g/kg; resveratrol intervention group: treated with resveratrol of 4 g/kg. ORO: oil red o; MaS: macrovesicular steatosis; MiS: microvesicular steatosis; ToS: total steatosis. Quantitative data were expressed as mean ± standard deviation. Multiple group comparisons were conducted using one-way analysis of variance, with prior assessments for homogeneity of variance and normality. In cases of homogeneity of variance and adherence to normal distribution, the least significant difference and Student-Newman-Keuls tests were employed. Compared with control group, ^aP ≤ 0.05; compared with AFLD group, ^bP ≤ 0.05.

3.2. Results of 4DLabel-free quantitative proteomic analysis of liver tissue in mice

Based on a series of 4DLabel-free quantitative proteomics techniques, this paper obtained the protein expression profiles of liver tissue samples of AFLD mice and drug intervention mice, then the differentially expressed proteins were screened and analyzed by GO classification and KEGG pathway enrichment analysis.

3.2.1. Protein identification and quantitative analysis

A total of 4513 proteins were identified, including 3763 quantitative proteins. The ratio between groups with the same protein expression above 2 or below 0.5 was considered to have significant difference. A total of 946 proteins with significant difference in expression level were found by pairwise comparison between groups. Compared with the control group, 145, 68 and 15 proteins were up-regulated expression in the liver tissues of AFLD mice, GJDD- treated mice and resveratrol-treated mice. On the contrary, 148, 73 and 14 proteins were down-regulated expression, respectively. On the other hand, compared with the AFLD mice, the expression levels of 92, 69 proteins in the liver tissues of GJDD-treated mice and resveratrol-treated mice increased significantly, and 135, 94 proteins expression decreased significantly, respectively. Among them, GJDD intervention could correspondingly down-regulate 72 proteins with increased expression and up-regulate 63 proteins with decreased expression in AFLD mice. A total of 95 proteins were found to be involved in the differential expression in M/C, P/M and R/M at the same time, of which, both GJDD and resveratrol treatment could significantly down-regulate or up-regulate 51 proteins with increased expression and 44 proteins with decreased expression in AFLD mice respectively after treatment. The Volcano Plot was used to analyze the differences of protein expression levels among different treatment groups (M/C, P/M, and R/M). It was considered to have significant difference when P ≤ 0.05 and protein quantitative ratio is greater than 2 or lower than 0.5 (Figure 2).

Figure 2. Volcano plots showed the differentially protein expression in every two groups (M/C, P/M, R/M).

A: AFLD group/Control group (M/C); B: GJDD group/AFLD group (P/M); C: resveratrol group/AFLD group (R/M). Control group: treated only with free drinking water; AFLD model group: treated with ethanol and drinking water (2.6 g/kg); GJDD intervention group: treated with GJDD of 4.9 g/kg; resveratrol intervention group: treated with resveratrol of 4 g/kg. AFLD: alcoholic fatty live disease; GJDD: Gehua Jiejiu Dizhi decoction. The ratio of the mean value of repeated quantitative value of protein in the comparison sample pair was taken as the fold change. The relative quantitative value of protein in the comparison sample pair was tested by t-test. The specified thresholds include log2 fold change and adjusted P-value after log conversion, which were used to measure the fold and significance of protein expression differences respectively. In the figure, positive and red dots indicated that the difference was up-regulated. On the contrary, negative and green dots indicated that the difference was down-regulated.

3.2.2. Go enrichment and KEGG pathway analysis of proteins with differential expression levels

The biological process, cell composition and molecular function of 683 differentially expressed proteins from three comparison groups (M/C, P/M and R/M) were annotated in GO to explain the biological role of proteins from different angles. In the three comparison groups, the results of GO enrichment analysis showed that the first four biological processes (cellular process, metallic process, biological regulation and response to stimulus), the first two cell compositions (cell and organelle) and the first two molecular functions (binding and catalytic activity) involved in the differential proteins of M/C, P/M and R/M are the same (supplementary Figure 1A-1C). Using KEGG enrichment analysis method, the top 10 enrichment pathways of differentially expressed proteins in each comparison group were respectively displayed, including three enrichment pathways, such as retinol metabolism, pyruvate metabolism and steroid hormone biosynthesis, which co-existed in each comparison group (Figure 3A-3C).

Figure 3. Bubble plot of KEGG pathway enrichment analysis of proteins with differential expression level.

A: AFLD group/Control group; B: GJDD group/AFLD group; C: resveratrol group/AFLD group. Control group: treated only with free drinking water; AFLD model group: treated with ethanol and drinking water (2.6 g/kg); GJDD intervention group: treated with GJDD of 4.9 g/kg; resveratrol intervention group: treated with resveratrol of 4 g/kg. X-axis represents enrichment factor. Y-axis represents pathway name. The darker the mapping color, the smaller the P-value and the more obvious the significance, the larger the mapping value and the more enrichment. The same KEGG enrichment pathway involved in the differential proteins of M/C, P/M, and R/M comparison groups are boxed in three colors respectively, in which red is “retinol metabolism”, purple is “pyruvate metabolism”, and blue is “steroid hormone biosynthesis”. KEGG: Kyoto Encyclopedia of Genes and Genomes; AFLD: alcoholic fatty live disease; GJDD: Gehua Jiejiu Dizhi decoction.

3.2.3. Screening of co-expressed differential proteins and their Go enrichment analysis results

Fifteen differentially co-expressed proteins were screened between every two groups (M/C, P/C, P/M), 3 differentially co-expressed proteins were screened between every two groups (M/C, R/C, R/M), but only 1 differentially co-expressed protein was screened between every two groups (M/C, P/C, R/C, P/M, R/M). Among them, GJDD could correspondingly up-regulate or down-regulate the 10 down-regulated or 5 up-regulated proteins in the AFLD mice to a level similar to that of the control group (Table 2). Except for the three co-expressed proteins without biological process (Ces3a, Keg1, H1-1), the other 12 proteins involved multiple biological process, such as transcriptional regulation, metabolic process, stimulus response, signal transduction, precursor metabolites and energy production and immune system process. In particular, seven proteins such as Aox3, Fabp5, Stub1, Acss2, Slco1a1, Mup1, and Mup2 involve at least two biological process at the same time. They also involve cellular components such as cell, organelle, membrane, and molecular functions such as catalytic activity, binding, molecular function regulator, structural molecule activity and transporter activity et al.

Table 2.

Differential co-expressed proteins in groups of C, M and P

Proteins accession	Protein description/ Gene	MW (kDa)	Coverage (%)	Unique peptides	Ratio
					M/C	P/C	P/M
P43275	Histone H1.1/ H1-1	21.79	31.9	4	6.54	2.81	0.43
P43276	Histone H1.5/ H1-5	22.58	25.1	4	6.48	2.37	0.37
Q9WUD1	STIP1 homology and U box-containing protein/ Stub1	34.91	34.2	9	5.66	2.55	0.45
Q9D7S7	60S ribosomal protein L22-like 1/ Rpl22l1	14.47	45.1	4	5.08	2.30	0.45
Q9D1M7	Peptidyl-prolyl cis-trans isomerase/ Fkbp11	22.14	41.3	5	4.91	1.85	0.38
G3X982	Aldehyde oxidase 3/ Aox3	146.9	55.4	52	0.46	3.60	7.9
Q99P30	Peroxisomal coenzyme A/ Nudt7	26.86	51.7	14	0.45	3.11	6.9
Q9DCY0	Glycine N-acyltransfera se-like protein/ Keg1	33.72	55.3	13	0.40	2.53	6.28
Q63880	Carboxylesterase 3A/Ces3a	63.32	52.7	11	0.20	2.68	13.17
Q05816	Fattyacid-bindingprotein 5/ Fabp5	15.14	72.6	9	0.19	0.50	2.66
Q9QXG4	Acetyl-coenzyme A synthe tase/ Acss2	78.86	43.2	25	0.19	0.48	2.48
P11589	Major urinary protein 2/ Mup2	20.66	78.3	3	0.14	2.94	21.74
Q9QXZ6	Solute carrier organican ion transporter family member/ Slco1a1	74.40	32.1	16	0.11	5.21	45.54
P11588	Major urinary protein 1/ Mup1	20.65	81.1	3	0.09	5.46	63.4
Q9D154	Leukocyte elastase inhibitor A/ Serpinb1a	42.58	35.4	12	0.08	0.26	3.11

Notes: Ces3a: carboxylesterase 3a; Fabp5: fatty acid-binding protein 5; Serpinb1a: seukocyte elastase inhibitor 1a; Acss2: acetyl-coenzyme a synthetase2; Slco1a1: solute carrier organic anion transporter family member 1a1; Keg1: glycine n-acyltransferase-like protein1; Aox3: aldehyde oxidase 3; Nudt7: nudix hydrolase 7; Rpl22l1: ribosomal protein l22 like 1; H1-5: histone h1.5; Fkbp11: peptidyl-prolyl cis-trans isomerase11; Mup2: major urinary protein 1; Mup1: major urinary protein 2; Stub1: stip1 homology and U box-containing protein1; H1-1: histone h1.1. MW: molecular weight; M/C: AFLD group/Control group; P/C: GJDD group/Control group; P/M: GJDD group/AFLD group; Control group(C): treated only with free drinking water; AFLD model group (M): treated with ethanol and drinking water (2.6 g/kg); GJDD intervention group (P): treated with GJDD of 4.9 g/kg; AFLD: alcoholic fatty live disease; GJDD: Gehua Jiejiu Dizhi decoction.

3.3. Identification results of targeted proteome quantitative technology based on mass spectrometry

The final expressions of the 15 differentially co-expressed proteins were identified by mass spectrometry-based targeted proteome quantitative technique. The results showed that 11 proteins such as Aox3, H1-5, Fabp5, Ces3a, Nudt7, Serpinb1a, Fkbp11, Rpl22l1, Keg1, Acss2 and Slco1a1 were all detected. Among them, compared with the control group, the expression levels of Aox3, Fabp5, Ces3a, Nudt7, Serpinb1a, Keg1, Acss2 and Slco1a1 in the AFLD group decreased significantly by 0.56, 0.21, 0.15, 0.56, 0.21, 0.51, 0.21 and 0.44-fold respectively, and significantly increased by 1.6-23.69-fold after GJDD intervention. On the contrary, GJDD treatment could significantly down regulate the up-regulated H1-5 (8.06-fold), Fkbp11 (6.74-fold) and Rpl22l1 (12.31-fold) protein expression levels in AFLD mice by 0.26-0.35-fold. These detected protein expression patterns were consistent with the above 4DLabel-free proteomic results (Table 3).

Table 3.

Differentially co-expressed proteins confirmed by targeted quantitative proteomics based on mass spectrometry

Proteins accession	Proteins description	Gene	Peptides	Ratio
				M/C	P/C	P/M
Q63880	Carboxylesterase 3A	Ces3a	LGIFGFLSTGDK	0.15	3.08	19.87
Q05816	Fatty acid-binding protein 5	Fabp5	ELGVGLALR	0.21	0.54	2.58
Q9D154	Leukocyte elastase inhibitor A	Serpinb1a	FQSLNAEVSK	0.21	0.37	1.78
Q9QXG4	Acetyl-coenzyme A synthetase	Acss2	TACPGPFLQYNFDVTK	0.21	0.33	1.60
Q9QXZ6	Solute carrier organicanion transporter family member	Slco1a1	GVQHPLYGEK	0.44	10.53	23.69
Q9DCY0	Glycine N-acyltransferase-like protein	Keg1	VIESLGATNLGK	0.51	3.16	6.20
G3X982	Aldehyde oxidase 3	Aox3	TTWIAPGTLNDLLELK	0.56	3.80	6.79
Q99P30	Peroxisomal coenzyme A	Nudt7	EVFFVPLDYFLHPQVYYQK	0.56	3.62	6.45
Q9D7S7	60S ribosomal protein L22-like 1	Rpl22l1	TGNLGNVVHIER	12.31	3.16	0.26
P43276	Histone H1.5	H1-5	GGVSLPALK	8.06	2.71	0.34
Q9D1M7	Peptidyl-prolyl cis-trans isomerase	Fkbp11	DPLVIELGQK	6.74	2.34	0.35

Notes: Ces3a: carboxylesterase 3a; Fabp5: fatty acid-binding protein 5; Serpinb1a: seukocyte elastase inhibitor 1a; Acss2: acetyl-coenzyme a synthetase2; Slco1a1: solute carrier organic anion transporter family member 1a1; Keg1: glycine n-acyltransferase-like protein1; Aox3: aldehyde oxidase 3; Nudt7: nudix hydrolase 7; Rpl22l1: ribosomal protein l22 like 1; H1-5: histone h1.5; Fkbp11: peptidyl-prolyl cis-trans isomerase11. M/C: AFLD group/Control group; P/C: GJDD group/Control group; P/M: GJDD group/AFLD group; Control group (C): treated only with free drinking water; AFLD model group (M): treated with ethanol and drinking water (2.6 g/kg); GJDD intervention group (P): treated with GJDD of 4.9 g/kg; AFLD: alcoholic fatty live disease; GJDD: Gehua Jiejiu Dizhi decoction.

4. DISCUSSION

TCM has played an extensive and profoundly important role in many disease fields since ancient times.^31,32 It also shows the unique advantages of TCM in the treatment of liver injury caused by alcohol intake, which makes up to a certain extent for the lack of specific therapeutic drugs in ALD in modern medicine. However, TCM still lacks standards and norms in syndrome differentiation, judgment criteria and objective index evaluation of ALD. There are still differing opinions on drug selection and compatibility, and there are still great limitations in practical selection and application. Therefore, it is very necessary to further screen Traditional Chinese Medicine prescriptions with strong operability and good curative effect.^33,34 Based on the TCM dialectical understanding and clinical practice of "alcohol poison loves the liver and phlegm turbidity condenses " in AFLD, and under the guidance of the treatment principles and methods of "treating both symptoms and signs, detoxifying and removing turbidity", GJDD was prepared. As shown in Figure 1, GJDD could significantly improve liver steatosis and reduce fat deposition in AFLD mice. Similarly, as reported in the literature, resveratrol has a significant inhibitory effect on hepatocyte steatosis caused by alcohol intake. However, the problems of poor oral bioavailability, poor chemical stability and low water solubility limit the clinical application of resveratrol.²³ Combining the previous findings, at present, there is an urgent need to understand the mechanism of GJDD against AFLD.

Proteomics tools are widely considered to be valuable for exploring the pharmacology of complex Traditional Chinese Medicine systems. In this paper, the results of 293 differentially expressed proteins in the AFLD group showed that chronic alcohol intake could cause significant changes in the protein expression profile of liver tissue, suggesting that these differential proteins could be involved in the formation of AFLD. Interestingly, after the intervention of GJDD, the expression levels of several proteins also changed significantly, which were either up-regulated or down-regulated. The opposite expression regulation results after the intervention suggested that these proteins were closely related to the AFLD treatment of GJDD. Therefore, the results of protein expression profile changes in liver tissue after chronic alcohol intake provided preliminary research results for the follow-up discussion of drug intervention mechanism and the further improvement of AFLD occurrence mechanism. KEGG enrichment analysis showed that the three most enriched pathways in the M/C comparison group were the regulation of steroid hormone biosynthesis, retinol metabolism and pyruvate metabolism. Ding et al ³⁵ conducted serum metabolite experiments on NAFLD patients by using UPIC-Q-TOF-MSE analysis technology, which showed that regulating metabolic pathways such as retinol steroid hormone biosynthesis and combined supplementation of n-3 polyunsaturated fatty acids had a potential role in reducing hepatic steatosis. Some studies have also shown that the regulation of retinol metabolism plays an important role in inducing adipocyte differentiation, increasing energy consumption and reducing lipid accumulation.³⁶ Errafi et al ³⁷ used oleic acid to induce transcriptome analysis of HepG2 cells and the results showed that glucagon-like peptide-1 receptor agonist EX-4 reduced lipid accumulation by regulating many key pathways of lncRNA, including fatty acid and pyruvate metabolism. It is suggested that these altered proteins may directly or indirectly affect the liver pathology of mice after alcohol intake through a variety of ways. In order to further explore the correlation between these protein and AFLD and the intervention effect of GJDD, through Venn diagram analysis, there were 15 differentially co-expressed proteins in the three comparison pairs, named M/C, P/M, P/C (Table 2). The expression differences of these 15 proteins were verified by targeted proteomic quantitative technology, and finally, of which 11 proteins (Aox3, H1-5, Fabp5, Ces3a, Nudt7, Serpinb1a, Fkbp11, Rpl22l1, Keg1, Acss2, and Slco1a1) had the same expression pattern as the results of 4D Label-free proteomic quantitative analysis. Therefore, we reviewed and sorted out the literature on the biological characteristics of these 11 differentially co-expressed proteins and their relationships with diseases and put forward the proteins that should be first concerned and explored in the follow-up research.

Ces3a is a member of the carboxylesterase 3 (Ces3) gene family, and recent studies have shown that carboxylesterase family genes are associated with metabolic diseases. Ces3 activity increased significantly during adipocyte differentiation by using protein mass spectrometry analysis, and Ces3 inhibitors could improve metabolic syndrome in obese and diabetic mice.^38,39 Contrary to this view, Mukherjee et al ⁴⁰ used the 3T3-L1 cell model to discuss the positive effect of Ces3 in lipid metabolism, and its bright prospect in the treatment of obesity was pointed out. Yang et al ⁴¹ found and identified that the blockage of Ces3 led not only to decreased lipolysis but also to impaired β-adrenergic-stimulated thermogenesis. Therefore, it is very important to explore whether this protein prevents liver lipid accumulation after alcohol intake by maintaining fat metabolic balance. Our 4D Label-free proteomic and targeted protein quantitative data consistently showed that the protein was significantly decreased at least 0.2 fold in the model group and significantly increased at least 13 fold after GJDD intervention. These results showed that the regulation of Ces3a expression could be involved in the AFLD intervention of GJDD, which needs to be further discussed. Studies have shown that the regulation of bile acid metabolism is an important factor to maintain the balance of sugar, lipid and energy metabolism in liver, intestine and adipose tissue. However, these functions are inhibited in many metabolic diseases such as obesity.^42,43 Slcola1 is a transporter of bile acids between blood and liver, which has a protective effect on a series of liver damage caused by cholestasis. In addition, under the clear understanding of the relationship between intestinal microorganisms and ALD,^43,44 it was found that the stable expression of Slcola1 was closely related to the maintenance of intestinal flora and the normal regulation of bile acid metabolism in mice.^43,45 Both metagenomic and RNA sequence analysis also revealed that the decrease of Slco1a1 expression resulted in increased plasma bile acid levels.^46,47 These studies suggest an important relationship between the decrease of Slcola1 level and liver injury after alcohol intake to a certain extent. As a protein with significant differential expression, the functional enrichment analysis of Go classification and KEGG pathway also suggested that Slcola1 played a role through bile secretion pathway. Whether the results suggest that GJDD may affect the metabolism of bile acids to achieve the protective effect of AFLD by regulating the changes of intestinal microbiota also is very necessary to explore further.

As an intracellular lipid transporter, FABP5 gene knockout promoted the anti-inflammatory response of mouse liver after endotoxin exposure.⁴⁸ It has also been found that Fabp5 deletion activates the PPARγ pathway in macrophages protecting them against pro-inflammatory and pro-atherogenic changes.⁴⁹ Keg1 has been found to be one of the members that specifically expresses glycine-N-acyltransferase in the liver and kidney, which is down-regulated in hepatoma cells and becomes a good hepatoma marker.⁵⁰ Unfortunately, no study has been found on the role of Fabp5 and Keg1 in alcohol-related diseases, while our results of 4D Label-free quantitative proteomic analysis showed that Fabp5 and Keg1 were significantly down-regulated in the AFLD mice and up-regulated after treatment with GJDD, respectively, and targeted proteomic quantitative technique analysis also further verified this expression pattern. Results of this paper preliminarily showed that decreased expression of Fabp5 in liver tissue could promote the occurrence of AFLD, and also indicated that Fabp5 protein expression level was different in metabolic pathological injuries caused by different etiologies. Combined with the results of functional enrichment analysis, it was suggested that GJDD could participate in the AFLD intervention effect by regulating Fabp5 protein expression level and via PPAR signaling pathway, but how to exert the regulatory effect of this pathway needs to be further discussed.

As a special hydrolase of coenzyme A (CoA), studies have shown that reduced Nudt7 synthesis will lead to fasting hypoglycemia and liver triglyceride accumulation.^51,⇓-53 In addition, Nudt7 promotes fatty acid oxidation by scavenging oxidative CoA in peroxisomes, and increased liver Nudt7 expression decreased bile acid levels.^52,54 In addition, activated PPARα will inhibit the expression of Nudt7, which may be related to fatty liver.^55,56 Consistently, we also found that Nudt7 was down-regulated in the AFLD mice and up-regulated after the intervention of GJDD. The results of functional enrichment analysis preliminarily proved that Nudt7 played a role through the peroxisome pathway, further providing a reliable basis for GJDD as playing a protective role through the regulation of the Nudt7-CoA-peroxisome pathway, but it is necessary to determine how CoA and peroxisome communicate in drug intervention. As a functional homolog of human SerpinB1, mouse SerpinB1a is mainly synthesized in the liver.⁵⁷ It was found that the inhibition of its expression would lower the level of FOXO1, a key regulator of autophagy, thereby weakening the antioxidant stress effect of FOXO1 and aggravating the oxidative stress response.^58,59 It is known that FOXO1, in addition to participating in the regulation of autophagy and lipid catabolism, can also reduce the transcription activity of PPARγ, reduce the uptake of free fatty acids, and inhibit the expression of fatty acid synthase and acetyl CoA carboxylase 1, ultimately realizing the inhibition of fat production and reducing the accumulation of triglycerides in the liver.^60,61 Similarly, our proteomic analysis also showed that the protein was significantly down-regulated in AFLD mice and significantly up-regulated after treatment with GJDD or resveratrol. It is suggested that SerpinB1a may participate in lipid metabolism and anti-oxidative stress through positive regulation of FOXO1 expression level, but whether this effect is involved in the occurrence of AFLD and drug intervention has not been reported. Acss2, an alternative pathway enzyme catalyzing the synthesis of acetyl-CoA, its defect will reduce the accumulation of liver fat, and it is considered that the selectively acts on Acss2 may contribute to the treatment of hepatic steatosis and obesity.^62,63 It has also been found that AMPK can reduce the level of malonyl-CoA converted from acetyl-CoA catalyzed by ACC, increase fatty acid oxidation and improve AFLD.⁶⁴ Unfortunately, up to now, there is no research on Acss2 in alcoholic liver injury. Our results showed the significant down-regulation of Acss2 expression in the AFLD group and the significant up-regulation after the intervention of GJDD. Combined with the results of functional enrichment analysis, it suggested that this protein could be involved in the occurrence of AFLD and the intervention effect of GJDD on anti-AFLD.

Studies have shown that aldehyde oxidases is an important enzyme for the synthesis of retinoic acid.^65,66 Zarei et al ⁶⁷ found that all trans-rentinoic acid (atRA) effectively improved high-fat induced hepatic steatosis in rabbits through antioxidant effect. It has also been shown that RA can significantly reduce mice nonalcoholic fatty liver disease by up-regulating the expression of leptin receptor in liver.⁶⁸ Aox3 is a subtype of aldehyde oxidase, which is abundant in mouse liver and has similar function to human Aox1. In our study, Aox3 expression was down-regulated in the liver tissue of AFLD mice and significantly up-regulated about 7-fold after intervention of GJDD. It was suggested that regulating the expression of Aox3 and promoting the biosynthesis of atRA could also be an important mechanism for GJDD to regulate liver fat metabolism and exert anti-AFLD role after alcohol intake. It is very necessary to pay attention to Aox3 protein in the future research of GJDD intervention in AFLD.

In addition to the above eight proteins significantly down-regulated in the liver tissue of AFLD mice, there were also three proteins worthy of attention, namely Fkbp11, Rpl22l1 and H1-5. They were up-regulated 4.91-fold, 5.08-fold and 6.48-fold respectively in liver tissue of AFLD mice, and down-regulated 0.38, 0.45 and 0.37-fold respectively after treatment with GJDD. Unfortunately, except for the research in the field of oncological diseases, the research of these three proteins in the field of alcoholic liver disease has hardly been reported. For example, Fkbp11 expression was found to gradually increase during hepatocellular carcinoma development, Rpl22l1 and H1-5 proteins were also found to be highly expressed in colorectal cancer and ovarian granulosa cell tumor, respectively, and served as prognostic markers.^69,⇓-71 But, how these three proteins are involved in the occurrence of AFLD and the AFLD intervention effect of GJDD need to be elucidated in further studies.

In summary, we investigated the protein expression changes in liver tissue of AFLD mice and their correlation with the intervention effect of GJDD. The results showed that at least 293 proteins were involved in the occurrence of AFLD, of which at least 227 proteins were related to the intervention effect of GJDD on AFLD. In addition, we further identified 11 proteins, which could be involved in the occurrence of AFLD and the intervention of GJDD, especially Aox3, Nudt7, Ces3a, Slco1a1, Acss2, Serpinb1a and Fabp5. They may participate in GJDD mediated AFLD intervention at least through the modulation of lipid metabolism, bile acid metabolism and with exertion of antioxidant stress. Although we screened and identified proteins that may mediate the anti-AFLD effect of GJDD, Unfortunately, we did not conduct research on their functional validation and interaction. Therefore, subsequent relevant studies should be carried out as far as possible.

5. ACKNOWLEDGEMENTS

We would like to thank Fred Bogott, M.D., Ph.D., a former Medical Director of Austin Medical Center, the Mayo Clinic in Austin, Minnesota, USA, for his excellent English revision of our manuscript in writing and grammar.

6. SUPPORTING INFORMATION

Supporting data to this article can be found online at http://journaltcm.cn.