Aspirin enteric-coated tablets were purchased from Bayer (Shanghai, China). Formaldehyde was purchased from Longhai Dandong Reagent Factory (Dandong, China). Potassium hydroxide was purchased from Tianjin Kaixin Chemical Industry Co., Ltd. (Tianjin, China). High-performance liquid chromatography (HPLC)-grade acetonitrile was purchased from J. T. Baker (Thermo Fisher Scientific, Inc., Waltham, MA, USA). HPLC-grade formic acid was purchased from Fluka (Merck & Co., Inc., Whitehouse Station, NJ, USA). Ultrapure water (18.2 MSZ) was prepared with a Milli-Q water purification system (Merck KGaA, Darmstadt, Germany).

Licorice was collected in Tongliao, the Inner Mongolia Autonomous Region in China at locations between 42°15′N-45°41′N and 119°15′E-123°43′E. In the Chinese Pharmacopoeia (2015 edition) the liquiritigenin and glycyrrhizic acid content is required to be >0.50 and >2.0%, respectively (22). The licorice used in this experiment contained 1.84 and 2.57% of liquiritigenin and glycyrrhizic acid, respectively, in accordance with the relevant provision of Chinese Pharmacopoeia. Licorice flavonoids were prepared as previously described (23). The average paste rate was 5.40%; the purity of total flavonoids reached up to 90%, as detected by HPLC method (24), primarily contained 1.1695% liquiritin apioside, 1.8402% liquiritin, 0.2087% isoliquiritin apioside, 0.2042% ononin, 0.4023% isoliquiritin, 0.6443% liquiritigenin, 0.1548% calycosin and 0.0726% isoliquiritigenin.

A total of 54 male Kunming mice weighing 18–22 g were provided by Liaoning Biological Technology Co., Ltd. (Liaoning, China). The mice were bred in a specific pathogen-free environment and had free access to a standard diet and tap water. The animals were housed in stainless steel metabolic cages under standard humidity (50±10%), temperature (25±2°C) and light (12/12 h light/dark cycle). All animal treatments were conducted in strict accordance with the National Institutes of Health Guide to the Care and Use of Liaoning University of Traditional Chinese Medicine, and the present study was approved by Attitude of the Animal Core and Welfare Committee of Liaoning University of Traditional Chinese Medicine.

Mice were divided randomly into 6 groups (n=9). These groups were: The control group (mice injected with 0.05 ml saline into the sub-plantar region of the right hind paw), the model group (mice injected with 0.05 ml 2.5% formaldehyde solution into the sub-plantar region of the right hind paw), the licorice flavonoid high-dose group (treated with 0.65 g/kg of licorice flavonoids), the middle-dose group (0.21 g/kg of licorice flavonoids), the low-dose group (0.07 g/kg of licorice flavonoids) and the aspirin group (0.04 g/kg of aspirin). All mice were orally administered the active group solution twice daily (control and model groups were administered saline) for 3 days. The mice were prohibited any food for 24 h before the final day of the experiment, but were allowed ad libitum access to water. At 1 h after the last administration, the paw volume of mice in each group prior to inducing inflammation was measured with a self-made foot volume-measuring device. Subsequently, mice were injected with 0.05 ml of 2.5% formaldehyde solution into the sub-plantar region of the right hind paw except control group (25). Following inducing inflammation 24 h, the paw volume was measured with a self-made foot volume-measuring device again; the ratio of paw edema (%) and the inhibitory rate of paw edema (%) were taken as indexes of detection. Mice were then anesthetized and sacrificed, blood was collected into Na-heparin tubes and plasma was obtained following centrifugation (1,006.2 × g, 4°C for 15 min), and stored at −80°C until metabolomics analysis were performed. The inflammatory paws of the mice from their right ankle were removed, weighed, and soaked for 1 h with 3 ml normal saline (precooled to <4°C). Following centrifugation (1,006.2 × g, 4°C for 15 min) of the solution, 2 ml supernatant was absorbed and mixed with 2 ml of 0.15 mol l⁻¹ KOH-CH₄O solution, samples were incubated in water bath for 20 min at 50°C, then the content of prostaglandin E2 (PGE2) was detected with a UV spectrophotometer at a wavelength of 278 nm once the liquid had cooled.

The blood samples were thawed prior to analysis and 100-µl aliquots of bloods were added to 400 µl methanol. The mixture was vortexed for 3 min and centrifuged at 1,006.2 × g for 15 min at 4°C. The supernatant was then transferred to auto-sampler vials. A pooled quality control (QC) sample was made by mixing 20-µl aliquots of each sample, using the same method as that used for each of the other samples. The pooled QC sample was analyzed randomly through the analytical run to monitor instrument stability. In addition, a random sample was divided into 6 and was treated in the same way. These 6 samples were continuously analyzed to validate the repeatability of the sample preparation method.

Metabolic profiling of the blood serum was conducted using an Agilent-1260 LC system coupled with an Agilent-6530 Q-TOF mass spectrometer (Agilent Technologies, Inc., Santa Clara, CA, USA). Plasma chromatographic separation was performed on an Agilent Poroshell 120 SB-C₁₈ column (2.7 µm, 4.6×100 mm; Agilent Technologies, Inc.) with the temperature of the column set at 45°C. The flow rate was 1.0 ml/min and the mobile phase was ultrapure water with 1‰ formic acid and 1‰ acetonitrile The gradient elution of acetonitrile was performed as follows: 5–30% acetonitrile at 0–4 min; 30–67% acetonitrile at 4–6 min; 67–98% acetonitrile at 6–11 min; 98–100% acetonitrile at 11–17 min; followed by a 7-min re-equilibration step. The volume of sample injected was 1 µl. The parameters of mass detection were set as followed: In positive mode, the flow rate of the dry gas of MS mode was 9 l/min, Vcap 4000 V; the nebulizer pressure was 45 psig; the fragmentation voltage was 175 V; the Sheath Gas Temp was 350°C; the Sheath gas velocity was 12 l/min; the acquisition rate was 25 Spectra/sec; MS data were acquired in full-scan mode from m/z 50 to 1,050 amu; and the collision energy of MS/MS date acquisition was set at 20 eV. The corrected mixed solution (Agilent Technologies, Inc.; m/z=121.050873, 149.02332, 922.009798) was used as the lock mass.

The raw MS data were exported by Agilent Mass Hunter Qualitative Analysis software (vB.04.00; Agilent Technologies, Inc.) for peak detection, alignment and filtering (Agilent Technologies, Inc.). Integrated information of the retention time and molecular mass were then outputted from the Agilent Mass Profiler software (vB.02.00; Agilent Technologies, Inc.). Finally, principal components analysis (PCA), the cluster analysis used for multivariate analysis and one-way analysis of variance (ANOVA) used for expressing the significance of data were performed with Mass Profiler Professional (MPP) 12.6 software (Agilent Technologies, Inc.), as described previously (26). The metabolites were preliminarily identified in METLIN (https://metlin.scripps.edu/landing_page.php?pgcontent=mainPage) and PUBCHEM (https://www.ncbi.nlm.nih.gov/pccompound/) databases, and biochemical reactions involving the identified metabolites were obtained through KEGG (http://www.kegg.jp/), HMDB (http://www.hmdb.ca/) and LIPID MAPS (http://www.lipidmaps.org/) databases.

SPSS 19.0 statistical software (IBM SPSS, Armonk, NY, USA) was used for statistical analysis, and one-way ANOVA was used to design the group data followed by Student-Newman-Keuls test. Student's t-test was used to analyze the different metabolites between the groups with MPP software. All quantitative data are expressed as the mean ± standard deviation, as indicated. P<0.01 was considered to indicate a statistically significant difference.

The inhibitory effects of different doses of groups on formalin-induced paw edema in mice are shown in Figs. 1 and 2. The degree of paw edema in the model group was substantially higher compared with that in control group (no inflammation), indicating that the acute inflammation model was successfully established (data not shown). Compared with the model group, the ratio of paw edema [(the volume of paw after inflammation-the volume of paw before inflammation)/the volume of paw after inflammation)] in each administered group was significantly lower (P<0.01), among which aspirin group, the inhibition rate of the high- and middle-dose licorice flavonoid groups reached >50% (Fig. 2), indicating that licorice flavonoids inhibited the increase of inflammatory mediators caused by formalin stimuli. In addition, the content of PGE2 in the model group significantly increased when compared with the control group (Fig. 3), which indicated that the inflammation in the paws of the mice caused by formalin is able to raise the PGE2 level in vivo. The content of PGE2 in the licorice flavonoid dose groups was significantly decreased compare to that of model group (P<0.01), suggesting that licorice flavonoid can inhibit inflammation though regulating the level of PGE2.

PCA and hierarchical clustering analysis were used to classify the metabolic phenotypes and identify the different metabolites. On the observation of 3-dimensional plots, samples in the same group were evidently clustered and the model group differed significantly from others (Fig. 4), indicating that the endogenous metabolites in the formalin-induced paw edema in the model group differed significantly from the drug-treated groups. The spatial position of the licorice flavonoid groups was close to that of the control group, demonstrating that the plasma metabolite composition of inflammation mice has a tendency to return to normal following treatment. Clustering analysis of metabolomics data also revealed distinct segregation between the control, model and licorice flavonoid dose groups (Fig. 5).

Small molecule metabolites whose levels appeared to differ significantly (P<0.05) between licorice flavonoids, model and control groups were analyzed using MPP software. According to the MS/MS fragments of differing endogenous metabolites, compared with METLIN, PUBCHEM and other associated databases, 9 metabolites associated with inflammation were preliminary identified (Table I).

The pathways associated with endogenous metabolites and anti-inflammatory effects of licorice flavonoids are listed in Table I and Fig. 6. These metabolites were identified to be involved in pathways, including in arachidonic acid (AA) metabolism, tryptophan metabolism, sphingolipid metabolism, according to HMDB, KEGG, LIPID MAPS and other backend knowledge databases.