As described in our previous study (Ma et al., 2021), all the fresh GE samples were obtained from Xiaocaoba planting base in Zhaotong (Yunnan province, China), and were authenticated by Prof. Qin Meng (Yunnan institute of food and drug control, Kunming, China). The voucher sample (Authentication No. TM20190135) were stored at Institution of Yunnan Traditional Chinese Medicine and Materia Medica, Kunming, China. GAS, HBA, and PA (purity >99%) were purchased from National Institutes for Drug Control (Beijing, China). Methanol and acetonitrile (HPLC grade) were purchased from Tedia Company Inc. (Fairfield, United States), while the MS grade solvent was obtained from Merck Company Inc. (Darmstadt, Germany). Estazolam was purchased from Huazhong Pharmaceutical Co., LTD. (Xiangyang, China). Ultrapure water was prepared by a Milli-Q purification system (Bedford, United States). Anti-tryptophan hydroxylase 2 (TPH2), Anti-Indoleamine 1(IDO1), and Anti-Brain derived neurotrophic factor (BDNF) were purchased from Abcam Company Inc. (Cambridge, United Kingdom).

The fresh GE samples were driedin 60°C desiccators for 48 h. In order to prepare the alcohol-steamed GE samples, the fresh GE was firstly cooked in alcohol vapor at 80°C for 20 min and then dried at 60°C for 48 h (Figure 1A).

The chemical ingredients of fresh GE and AGE were detected in accordance with previous studies (Li et al., 2019; Ma et al., 2021).

All the experimental protocols on animals were approved by the Ethical Committee of China Pharmaceutical University, Nanjing, and the Laboratory Animal Management Committee of Jiangsu Province (SYXK (SU) 2018-0019). Adult male Sprague- Dawley (SD) rats (200 ± 20 g body weight) and C57BL/6 mice (20 g ± 2 g body weight) were obtained from the SLAC Laboratory Animal Co., Ltd. (Shanghai, China). During the study, all the animals were kept in a 12-h light-dark cycle at room temperature (22◦C ± 1°C) and humidity (60% ± 10%) with free access to water and a rodent chow diet.

Power analysis for sample size: The ARRIVE Guidelines (Animal Research: Reporting of In Vivo Experiments) were used for power analysis. We used the sleep duration index for mice experiment, and open-filed test index for rat experiment between the Model and AGE group as primary outcomes to calculate the number of animals needed in each group by “Animal Sample Size Calculations for Anticipated Means or from Pilot Study/Publication Data” (https://intraweb.hku.hk/local/launit/content/forms/Animal_Sample_Size_Calculations_for_Anticipated_Means_or_from_Pilot_Study_or_Publication_Data.xlsx). A power analysis indicated that to detect an effect size of 15 min in mean sleep duration between groups with 90% power at alpha = 0.05 in a one-way ANOVA with four levels, a sample of eight per group was necessary in the mice experiment. An effect size of 24.5 s in mean open-filed test between groups with 90% power at alpha = 0.05 in a one-way ANOVA with five levels, a sample of six per group was necessary, therefore, we used eight animals per group in rat experiment.

The phenobarbital-induce sleep mice model was performed according to our previous study (Ma et al., 2021). The rats were randomly divided into four groups: control, estazolam, fresh GE, and alcohol-steamed GE (AGE). The mice were respectively intraperitoneally injected with an equal volume of 0.5% CMC-Na solution, estazolam (1 mg/kg), fresh GE (1.1 g/kg), and AGE (1.1 g/kg) for 5 days. Then, a sub-hypnotic dose of phenobarbital (32 mg/kg/L) was administrated via peritoneal injection. The sleep latency and sleep duration time were observed and recorded after 45 min later (Figure 2A).

Previous studies have demonstrated that both fresh GE and water-steamed GE have anti-anxiety-like activity. To further compare the anxiolytic effect of AGE, 40 animals were divided into five groups (8 mice per group): control, model, estazolam, fresh GE, and AGE. At the beginning of the experiment, all animals were accordingly treated with an equal volume of 0.5% CMC-Na solution, estazolam (1 mg/kg), fresh GE (1.1 g/kg), and AGE (1.1 g/kg) by gavage for 7 days. The chronic restraint stress model was established as described previously (Ma et al., 2021), and the group without CRS stimulation was set as a control. The open field test (OFT) and elevated plus maze test (EPM) were used to assess the anxiety behavior of animals on the last day (Figure 2D).

All animals were sacrificed after the behavioral test and their serum and whole brain tissue were collected for stepping detection. Serum samples (eight samples per group) were prepared and detected by the method mentioned in our previous study (Ma et al., 2021).

Non-targeted-metabonomic was carried out on a 1260 HPLC system coupled to 6500 quadrupoles of flight (Q-TOF) MS (Agilent, Santa Clara, CA) with a poroshell 120 SB-C18 column (2.1 × 100 mm, 2.7 µm). A flow rate of 0.4 mL/min was used, and the injection volume was 5 μL. The mobile phase consisted of A (water with 0.1% formic acid) and B (acetonitrile) in the positive mode analysis, while A (1 mM ammonium fluoride in the water) and B (acetonitrile) in the negative mode analysis. The following gradient elution conditions were used: 0–2 min, 2% B; 2–20 min, 2%–100% B; 20–25 min, 100% B.

The electrospray ionization (ESI) source was operated both in positive and negative ion modes and MS parameters were optimized as follows: Gas Temp, 325°C; Gas Flow, 8 mL/min; Nebulizer, 40 psig; Sheath Gas Temp, 350°C; Sheath Gas Flow, 12 mL/min; V Cap, 4000; Nozzle Voltage, 0 V (positive mode), and 500 V (negative mode); Fragmentor, 150 V; Skimmer, 65; Octopole RF peak, 750. And mass scan range, m/z 50–1000.

Additionally, Amixture of each sample was considered as the quality control (QC) sample to evaluate instrumental precision, and the QC sample was detected after every eight samples during the whole sample batch.

Principal component analysis (PCA) and orthogonal partial least squares-discriminant analysis (OPLS-DA) was performed using the online MetaboAnalyst platform (https://www.metaboanalyst.ca). Ions with a VIP >1 and a significant t-test result (p < 0.05) were selected for identification. The HMDB database was used to match ions, and the KEGG database was used for pathway analysis after the secondary identification of matched ions by LC-MS/MS.

The brain tissue samples (50 mg) were homogenized with 1 mL 4°C precooling methanol for 15 min, then vortexed for 3 min to precipitate the protein. The samples were then centrifugated at 14,800 g for 15 min at 4°C. The supernatant was transferred to a clean EP tube (1.5 mL) and dried. Residues were reconstituted in 50 μL pre-cooled acetonitrile aqueous (50%) and continued to be centrifugated at 14,000 rpm for 15 min at 4°C. Finally, the supernatant was then transferred into the vials for the stepping analysis. The relative content of L-tryptophan, serotonin, kynurenine, kynurenic acid, and Quinolinic acid was detected on LC/MS standard reference material purchased from Sigma.

The total protein in the brain tissue was extracted using the RIPA lysis buffer, and protein concentrations determination was performed using a Pierce BCA protein assay kit. Subsequently, the samples were separated by 10% SDS-PAGE electrophoresis (Ye P. et al., 2019). The following antibodies were used: anti-TPH2 (1:1000), anti-IDO1 (1:1000), and anti-BDNF (1:1000). Protein expression difference was quantified using Image J Software and normalized to the expression of GAPDH.

Data were represented as the mean ± SD. Statistical analysis of the results was determined using one-way ANOVA with Tukey’s correction for multiple comparisons by GraphPad Prism Software (United States). A value of p < 0.05 was considered statistically significant. Non-targeted metabolomics data analysis was processed following our previous study (Ma et al., 2021).

A total of 17 compounds were identified in fresh GE and AGE (Figure 1B) and compared with fresh GE, the relative content of Parishin A (PA) and Parishin E (PE) decreased significantly after alcohol steaming, whereas Gastrodin (GAS), P-hydroxy benzyl alcohol (HBA), Parishin B (PB), and Parishin C (PC) were increased gradually (Table 1). The alcohol steaming process was shown to effectively separate the active components from GE, which is consistent with previous research on water-steamed GE (Li et al., 2019; Ma et al., 2021). Notably, the relative content of GAS, PA, and HBA changed more significantly after alcohol steaming.

Compared with the control group, both fresh GE and AGE could significantly shorten the sleep latency time and prolong the duration of sleep, while there was no marked difference between them (Figures 2B, C).

OFT and EPM were commonly used to evaluate the locomotion and exploration of animals, which were applied to assess the anti-anxiety effect of animals. Interestingly, we found that the central mobile time of OFT and the keep time in open arms of EMP significantly increased in both fresh GE and AGE groups, among them the effects of AGE even nearly to positive group (Figures 2E, F). It was observed that AGE showed a strongeranxiolytic effect than fresh GE.

Non-targeted metabolomics analysis was performed on the serum samples in fresh GE and AGE-treated groups, and a total of 2119 ions were obtained by non-targeted metabolic profiling in ESI positive and negative modes. Firstly, an unsupervised principal components analysis (PCA) was conducted for dimensionality reduction analysis, the result showed that the fresh GE and AGE groups overlap each other but distinguished from the other two groups (Figure 3A), which indicated that both fresh GE and AGE affect serum metabolic profiling. Furthermore, supervised OPLS-DA analysis was applied to find the differential variables between groups (Figure 3B). The OPLS-DA results showed that four groups could be well distinguished in serum samples. Both the fresh GE and AGE groups were close to the control group, which indicated fresh GE and AGE can ameliorate the abnormal metabolic changes in the model group, among them AGE group was more effective, which is consistent with previous anti-anxiety behavior test results.

The metabolites with a VIP value of OPLS-DA >1 and p-value < 0.05 will be regarded as important differential metabolites between the control and the model group. A total of 46 ions were identified as potential biomarkers, and hierarchical cluster analysis of these potential biomarkers in plasma showed that the metabolites in two drug-treated groups were separated from the control and model groups (Figure 3C). These results indicated that both fresh GE and AGE have intervention effects on anxiety in the mice.

Moreover, OPLS-DA analysis was applied to find the differential variables between AGE and model groups (Figure 3D), and these significant metabolites were imported into MetaboAnalyst 5.0 for pathway enrichment (Figure 3E). The pathway with a p-value < 0.05 was considered the potential target pathway associated with the anti-anxiety effect of AGE. Interestingly, we noticed that the TOP2 pathway was the mood disorder-related tryptophan metabolism pathway, which should be further investigated.

According to the enriched metabolic pathway, tryptophan metabolism might be the potential metabolic pathway that is related to psychological diseases, as mentioned in the previous study (Roager and Licht 2018). We further detected the relative content change of tryptophan-related metabolites in whole brain samples by LC-MS. As expected, CRS significantly decreased the brain 5-HT and KYNA levels while increasing the QA and KYN levels, which were improved by AGE intervention (Figure 4A).

To validate the tryptophan-related metabolites changes, we detected the key enzyme for the tryptophan metabolic pathway in the brain (IDO1 and TPH2) by western blot. The expression of TPH2 significantly decreased in the CRS model group while both fresh GE and AGE could significantly improve the expression (Figure 4B). In contrast, IDO1 significantly increased in the model group and only AGE could reduce IDO1 expression induced by CRS stimulation. Moreover, BDNF is known to modulate behavioral stress response and to mediate the therapeutic effects of antidepressant agents, which interact with the 5-HT system at several levels (Homberg et al., 2014). We found that BDNF significantly decreased in the CRS group, while improved by fresh GE and AGE.