Analytical grade sodium chloride, methanol, ethanol, acetic acid, K₂HPO₄·3H₂O and NaH₂PO₄·2H₂O were purchased from Nanjing Chemical Reagent Co., Ltd. (Nanjing, China). NaN₃ was obtained from Sinopharm Chemical Reagent Co., Ltd. (Shanghai, China). D₂O (99.9% D) and trimethylsilyl propionate (TSP) were obtained from Sigma-Aldrich (St. Louis, MO, USA). Phosphate buffer (K₂HPO₄/NaH₂PO₄, 1.5 M, pH 7.4) prepared in D₂O containing 0.1% (w/v) TSP and 0.1% NaN₃ (w/v) was used for NMR sample preparation.

All crude herbs were authenticated by Prof. Ping Li (Department of Medicinal Plants, China Pharmaceutical University, Nanjing, China). A mixture of 9 g Gardenia jasminoides Ellis, 3 g Rheum officinale Baill., 12 g Citrus aurantium L. and 24 g Semen Sojae Preparatum were extracted in 480 ml of distilled water and boiled for 30 min. The mixture was strained through a five-layer bandage. This procedure was repeated twice and the filtrate was freeze-dried for experimental usage. High-performance liquid chromatography analysis of freeze-dried ZZDHD powder is presented in Fig. 1.

A total of 24 clean-grade, male Sprague-Dawley rats (age, 6 weeks; weight, 200±20 g) were obtained from the Animal Multiplication Centre of Qinglong Mountain (Nanjing, China). All animal experiments were implemented strictly in accordance with the guidance for Experimental Animal Welfare of the National Guidelines at the Centre for SPF-grade Animal Experiments at the Jiangsu Institute for Food and Drug Control (Nanjing, China). The animals were housed in an animal facility with the following parameters: Temperature, 25±2°C; humidity, 60±5% and artificial 12-h light/dark cycle. The animals were acclimatized for 7 days prior to experiments with access to the certified standard chow and water. Subsequently, the rats were randomly divided into three groups (n=8), the control group (CG), the alcohol group (AG) and the ZZDHD group (ZG). Water was orally administered to the control rats between 5:00 pm and 6:00 pm from day 1–8. The AG rats received water orally between 5:00 pm and 6:00 pm from day 1–2 and then 50% alcohol at a dose of 5 g/kg/day from day 3–8. The ZG rats were orally treated with freeze-dried ZZDHD powder at a dose of 12 g/kg/day between 3:00 pm and 4:00 pm from day 1–8. Simultaneously, the ZG rats were also orally administered with 50% alcohol at a dose of 5 g/kg/day between 5:00 pm and 6:00 pm from day 3–8. All overnight urine samples for each rat were collected (from 8:00 pm to 8:00 am) manually to prevent contamination on days 0, 3, 4, 5, 6, 7 and 8. Urine collection tubes were stored at low temperatures (0–4°C) using a cooling bath and NaN₃ solution (0.1% w/v) was added to inhibit bacterial growth. The urine samples were stored at −80°C prior to ¹H NMR analysis. At 8:00 pm on day 8, all animals were fasted and then euthanized after 12 h using 4% isoflurane anesthesia (Shandong Keyuan Pharmaceutical Co., Ltd., Jinan, China). The liver and serum samples from all groups were collected immediately on day 9 (8:00 am). The experimental procedures were approved by the Animal Care and Ethics Committee at the Jiangsu Institute for Food and Drug Control. All surgeries were performed under isoflurane anesthesia and all efforts were made to ameliorate the suffering of the animals.

Alanine aminotransferase (ALT) and aspartate aminotransferase (AST) activities, liver superoxide dismutase (SOD) activity, and glutathione (GSH) and malondialdehyde (MDA) concentration in the liver homogenate were analyzed with commercial kits according to the manufacturer's protocol in the Experiment Centre at the Jiangsu Institute for Food and Drug Control (Jiangsu, China). The assay kits for AST (cat. no. 20121007), ALT (cat. no. 20121010), SOD (cat. no. 20121209), GSH (cat. no. 20121208) and MDA (cat. no. 20121211) were purchased from the Nanjing Jiancheng Bioengineering Institute (Nanjing, China). The biochemical parameters were calculated and expressed as the mean ± standard deviation and P<0.05 was considered to indicate a statistically significant difference. For hematoxylin and eosin staining, a portion of the same liver lobe in each rat was immediately fixed via immersion in 10% neutral-buffered formalin, embedded in paraffin and sectioned into slices of 4–5 μm. For the remaining liver tissue, Oil red O staining was performed on 10–15 μm frozen liver sections.

Frozen urine samples were thawed at room temperature. A total of 500 μl urine was analyzed following the addition of 50 μl D₂O solution, this addition provided a lock signal containing 5 mM TSP as a reference for the chemical shift. The mixture was centrifuged at 16,099 × g for 10 min at 4°C. These solutions were then transferred to 5 mm NMR tubes. For each sample, a ¹H NMR spectrum was acquired on a Bruker AV 500 MHz spectrometer at 300 K and recorded using a standard NOESYPR1D pulse sequence (recycle delay-90°-t₁-90°-t_m-90°-acquisition). The ¹H NMR spectra were determined using 160 scans spectra and a Fourier transform was used, following an exponential line-broadening function of 0.25 Hz. All urinary spectra were then manually phased, baseline corrected and referenced to TSP (CH₃, δ 0.0) using TOPSPIN software (version 2.1, Bruker Biospin GmbH, Rheinstetten, Germany).

All ¹H NMR spectra were processed using the AMIX software package (Bruker Biospin GmbH). Regions from 0.5–4.5 and 5.98–9.5 ppm were included in the integration, as regions ≤0.5 and ≥9.5 ppm contained only noise, and the spectral region 4.50–5.98 ppm contained the suppression of water resonances and cross-relaxation effects. To account for the presence of ethanol and ethyl glucuronide (EtG) in the samples from the AG, the spectra were integrated into bins with a width of 0.04 ppm, following the removal of the bins that contained the ethanol and EtG peaks (21) and then normalized via probabilistic quotient normalization. A multivariate data analysis was performed using the software package SIMCA-P (version 13.0, Umetrics, Malmö, Sweden). Orthogonal projection to latent structure discriminant analysis (OPLS-DA), a supervised multivariate data analysis tool, was used with 7-fold cross-validation and the pareto-scaled data as the X-matrix and the class information as the Y-matrix in order to identify metabolites with significant intergroup differences. The fit of the models were evaluated using R²X and Q², which respectively represent the explained variations and predictability of the models. Values of these parameters near 1.0 indicated an effective and stable model, with predictive reliability. All models were additionally tested using the cross-validated analysis of variance approach. P<0.05 was considered to indicate a statistically significant difference. A univariate analysis was also applied to verify statistically significant metabolites using parametric (Student's t-test) or nonparametric (Mann-Whitney test) tests for potential characteristic biomarkers.

To obtain the dynamic changes of potential characteristic biomarkers following ZZDHD intervention, the ratios of the metabolites were calculated for the corresponding time points in the forms of [F_A - F_C]/F_C and [F_Z - F_C]/F_C, where F_A, F_C and F_Z stood for the average metabolite concentrations of the corresponding metabolites following probabilistic quotient normalization in the AG, CG and ZG, and these results were expressed as a line graph.

To determine the hepatoprotective effects of ZZDHD, the levels of serum AST and ALT and liver SOD, GSH and MDA were detected. The biochemical analysis indicated that the levels of AST, ALT, SOD, GSH and MDA were significantly changed in the AG compared with the CG and ZG (P<0.05; Table I). AG rats had significantly higher levels of AST and ALT in the serum compared with the CG (P<0.05). AST and ALT activity in the serum are typically low, however, the levels of ALT and AST increase in the event of liver injury. The levels of SOD and GSH were significantly decreased (P<0.05), and MDA levels were significantly increased, in the AG compared with the CG (P<0.05). However, the biochemical indicators, including serum AST, ALT, liver SOD, GSH and MDA, were markedly improved in the ZG compared with the AG (P<0.05; Table I). Thus, when combined with alcohol treatment, the ZZDHD alleviated alcohol-induced liver damage.

In order to investigate the protective effects of ZZDHD on hepatic tissue damage, histological analysis was conducted on rat liver tissues (Fig. 2). The CG rats exhibited normal liver tissue and no pathological changes were observed (Fig. 2A), with a similar phenomenon observed following Oil red O staining (Fig. 2B). However, the AG rats exhibited fat particles of varying sizes (Fig. 2C and D), thus presenting a degree of hepatic injury. Histological examination of the ZG did not indicate any observable fat particles (Fig. 2E and F); thus, ZZDHD may ameliorate the hepatic steatosis of rats with alcohol-induced liver damage.

The typical ¹H NMR spectra of rat urine are presented in Fig. 3 for the CG, AG and ZG, with major metabolites labeled. With respect to the identification of differential metabolites, the corresponding chemical shifts of the metabolites were previously described (22–25) and are publicly accessible in metabolomic databases, including Madison (www.mmcd.nmrfam.wisc.edu), Kyoto Encyclopedia of Genes and Genomes (www.genome.jp/kegg/) and the Human Metabolome Database (www.hmdb.ca), and using Chenomx NMR Suite software (version 7.5, Chenomx, Inc., Edmonton, Canada). The detectable metabolites in these spectra included ethanol, EtG, acetate, taurine, glycine, lactate, creatinine, creatine, N-acetyl glycoproteins, dimethylamine (DMA), dimethylglycine (DMG), methylamine and tricarboxylic acid cycle intermediates (citrate, α-ketoglutarate and succinate), hippurate, trigonelline and formate. Compared with the CG, the primary differences in the AG were the appearance of ethanol and EtG, an increase in creatine, lactate and glycine, and a decrease in hippurate, DMG, DMA and tricarboxylic acid cycle intermediates. In the ZG compared with the AG, hippurate and DMG levels were increased, whereas creatinine and lactate levels were decreased. Additionally, hippurate levels in the ZG were significantly increased compared with the AG, and were even higher compared with the CG.

To obtain further details regarding metabolic changes and to identify potential characteristic biomarkers, which represent the effect of ZZDHD on alcohol-induced liver damage, multivariate and univariate analyses were conducted on the NMR data on day 8 (Table II). OPLS-DA was performed to determine if there were significant differences in the metabolites of the different treatment groups. A clear separation between the AG and CG (Fig. 4A) was demonstrated and significant metabolomic differences with good model fit (R²X=0.845, Q²=0.907, P=3.54×10⁻⁴) were observed. The corresponding S-plot (Fig. 4B) distinguished nine metabolites. Lactate, creatine and glycine were increased in the AG compared with CG, and were located in the upper right quadrant. The decreased metabolites, including hippurate, TCA cycle intermediates, DMG and DMA, were located in the lower-left quadrant. The score plot of OPLS-DA for AG and ZG (Fig. 4C) indicated a clear separation between the two groups (R²X=0.523, Q²=0.936, P=1.68×10⁻⁶) and four metabolites exhibited changes in the S-plot (Fig. 4D), including hippurate and DMG increased in the ZG, while creatine and lactate levels were decreased.

The altered urine metabolites described above from the AG compared with CG and ZG were validated via variable importance in the projection (VIP) using a VIP ≥1. The P-values for the detected metabolites in the different groups were calculated using Student's t-test or Mann-Whitney test. The results demonstrated that the levels of metabolites, including lactate, creatine, hippurate and DMG, were significantly altered following ZZDHD intervention. Thus, these four metabolites may be potential characteristic biomarkers of the intervention effects of ZZDHD on alcohol-induced liver damage. The present study demonstrated that alcohol administration disrupted energy, amino acid, methionine and gut bacterial metabolism. The results indicated that ZZDHD partially regulated the altered metabolic changes induced by alcohol to promote a return to basal levels, which was also validated by the assessment of biochemical indicators and histopathology.

The ratios of the four potential characteristic biomarkers, including lactate, hippurate, DMG and creatine were calculated as AG or ZG relative to CG (Fig. 5), in order to investigate the changes and dynamic effects in the occurrence, development and ZZDHD intervention of early-stage alcoholic liver injury. All four metabolites exhibited time-dependent changes. The ratio of lactate in the AG compared with the CG varied between 0.8 and 1 from day 3–6. Lactate exhibited a nearly 2-fold increase, with a maximum level observed on day 5 following ethanol administration. Lactate in the ZG, relative to the CG, ranged between 0.6 and 1 from day 3–6; however, this ratio was approximately 5-fold lower compared with the AG after day 7 and 8. The ratio of DMG in the AG (relative to CG) decreased between 0.22 and 0.31 on day 3 and 4, and it was 0.6 from day 5–8, which suggested that DMG was reduced compared with the controls during ethanol treatment from day 5–8. The DMG level in the ZG was maintained below 0.22 compared with the CG during ethanol administration. Creatine increased to 0.27 on day 7 and to 0.52 on day 8 in the AG compared with the CG. The creatine in the ZG remained constant compared with the CG rats. The hippurate in the AG was decreased to ~0.3 on day 3 and 4, and was lower than 0.4 from day 5–8 in the corresponding period. The hippurate level in the ZG exhibited similar changes to the AG compared with the CG rats on day 3 and 4, that remained unchanged compared with CG rats from day 5 to day 7 and increased to >0.4 on day 8. This indicated that ZZDHD altered the gut bacteria to restore the normal metabolism of hippuric acid, with increased production of hippurate due to the polyphenols of ZZDHD. Metabolite changes were observed in the AG and the ZG compared with the CG for lactate, DMG, creatine, and hippurate.