The herbal constituents of CKF were purchased from the Guangdong Province Traditional Chinese Medical Hospital (Guangzhou, China), were morphologically authenticated by Prof. H. Y. Ma following the standards of China Pharmacopoeia (Part I, 2020 Version). A voucher number was assigned to each specimen and preserved in the School of Traditional Chinese Medicine, Guangdong Pharmaceutical University. The procedure for synthesizing the CKF extract has been described previously (Mao et al., 2017). Briefly, the seven herbal components of CKF (150 g Paeoniae Alba Radix, 60 g Saposhnikoviae Radix, 50 g Rehmanniae Radix, 50 g Cuscutae Semen, 30 g Coptidis Rhizoma, 80 g Fagopyri Dibotryis Rhizoma, and 30 g Periostracum cicadae) were blended for 2 h in 3-volumes of water (1,350 ml). The mixture was filtered and the filtrate was collected (Mao et al., 2017). Finally, the liquid was condensed and concentrated to 450 ml (1 g/ml) and stored at 4°C until subsequent use. For animal treatment, the solution was filtered and concentrated in a vacuum at 50°C. The final concentration was 5 g crude drug per ml. The resulting extract was stored at −20°C. The CKF was performed in previous study to simultaneously identify the major chemical components and the UPLC-Q-TOF-MS chromatogram of CKF obtained from its water extract is shown in Supplementary Figure S1 (Supplementary Table S1) (Mao et al., 2017).

A total of 32 (6-weeks old, 180–200 g body weight) male Sprague-Dawley (SD) rats were obtained from the Centre for Experimental Animals, Guangdong Province [Guangzhou, China approval number SCXK (Yue) 2017-0002]. Rats were housed in controlled conditions (room temperature: 23 ± 2°C; relative humidity: 55 ± 5%, and 12-h dark/12-h light cycle) for 1 week with ad libitum access to food and water. The animal studies were conducted following the guidelines for the Care and Use of Experimental Animals. The designs of all animal experiments were approved by the Guangdong Pharmaceutical University Animal Care and Use Committee, China [SCXK (Yue) 2017-0125].

CACS-treated rats were subjected to chronic unpredictable stress for 21 days, as illustrated in Figure 1 (Zou et al., 2008; Yu et al., 2015). On day 22, these rats were subjected to restraint stress for 3 h. The stool samples of the rats were collected within a 4 h period and the fecal water content and total fecal output were calculated. After weighing (wet weight), the stool samples were dried under vacuum overnight and the dry weight was recorded (Nelson et al., 2011). To verify the IBS rat model, visceral hypersensitivity was measured using the abdominal withdrawal reflex (AWR) scoring system; depression-like behaviors were measured using the sucrose preference and forced swimming tests; upper GI transit was used to evaluate gastrointestinal motility, and the number of fecal pellets and fecal water content was used to evaluate the severity of diarrhea in each group as described previously (Li et al., 2018; Xu et al., 2018).

After acclimatization, the rats were randomly divided into the following four groups (n = 8): Normal, IBS, CKF, and trimebutine maleate (TM) (Figure 1). TM was purchased from Shanghai yuanye Bio-Technology Co., Ltd. (Shanghai, China). The rats in the normal and IBS groups were intragastrically administered 10 ml/kg normal saline, while those in the CKF group rats were intragastrically administered 5.0 g/kg CKF, while in the TM treated group (10 ml/kg body weight), 17.0 mg/kg TM was administered. These treated rats were regularly examined and changes in the characteristics of food intake and body weights were recorded.

Rats in each group were subjected to fasting for 24 h, followed by an oral administration of 2 ml suspension of the charcoal meal (10% charcoal in 5% gum arabic) (Lv et al., 2011). These rats were sacrificed after 20 min. Next, the abdomen was opened up and the intestines from the pyloric junction to the caecal end were quickly excised (Dong et al., 2014). After dissection, the entire length of the intestines from the pyloric junction to the caecal end and the distance traveled by the charcoal meal were measured. The percentage of the distance traveled by the charcoal meal to the total length of the pyloric junction to the caecal end was the measure of the upper GI transit (Dong et al., 2014).

The visceral sensitivity to colorectal distention (CRD) was evaluated by measuring the AWR following a previously described protocol (Al-Chaer et al., 2000; Winston et al., 2007). Before the procedure, rats were allowed free access for drinking but were food-deprived for 24 h. Next, the rats were lightly anesthetized with 2% isoflurane, followed by insertion of an 8F catheter through the anus after lubrication with paraffin oil. Throughout, we ensured, by taping the associated tubing onto the rat tail, that 1 cm of a tube was left between the end of the gasbag and anus (Yu et al., 2015; Gong et al., 2015; Xu et al., 2019). Rats were subsequently placed in Lucite cubicles (20 cm × 8 cm × 8 cm) and allowed to acclimatize to the environment for 20 min (after awakening) before the test (Babygirija et al., 2017). After acclimatization, the rats were subjected to distention pressures (10, 20, 30, and 40 mmHg) for 20 s with a 5-min interval between different pressures. Finally, the AWR scores were determined according to as detailed in Table 1 (Yu et al., 2019; Li et al., 2020).

Sucrose preference test was performed by providing the rats with both 1% sucrose solution and double distilled water for 2 days (Wang et al., 2015). Briefly, rats were individually housed and were food- and water-deprived for 24 h (Nadal et al., 1996). On the following day, rats were free to choose between water and 1% sucrose solution. The sucrose preference was determined according to the ratio of the sugar solution consumed to that of the water (water and sucrose solution).

Rats were subjected to 15 min of training for forced swimming in a glass cylinder 24 h before the formal test (height = 45 cm and diameter = 25 cm) at a water depth of 30 cm and temperature of 24 ± 1°C. After administration, rats were again placed in the glass cylinder for swimming for 5 min, and the immobility time within 5 min was observed and recorded (Rao et al., 2021). Immobility time was defined as the time when the rat stopped struggling, floated in the water, and remained motionless, or only made necessary light movements to keep its head afloat in the water (Porsolt et al., 1977; Pan et al., 2021).

The distal colon and hippocampal tissues were isolated and homogenized in cold PBS solution. Next, the frozen colonic and hippocampal tissues were homogenized and lysed in tissue lysis buffer, followed by centrifugation at 12,000 g for 10 min at 4°C. The supernatant was collected and the levels of 5-HT and BDNF in the colonic and hippocampal tissues were measured using ELISA kits following the manufacturer’s instructions (Shanghai MLBIO Biotechnology Co., Ltd.).

Total RNA was extracted from the distal colon and hippocampal samples using the Trizol reagent (Accuratse Biology Co., Ltd., China), followed by cDNA synthesis using the PrimeScriptTM RT reagent Kit with a gDNA Eraser (Accuratse Biology Co., Ltd., China) according to the manufacturer’s instructions. Next, the mRNA levels of individual genes were determined through real-time PCR using the 2 × SYBR Green Pro Taq HS premix (Accuratse Biology Co., Ltd., China) on the CFX Connect fluorescence quantitative PCR detection platform (BIO-RAD, United States). The 20 µl PCR reaction mixture comprised of the following components: 10 µl 2 × SYBR Green Pro Taq HS Premix, 0.4 µl forward primer (10 µM), 0.4 µl reverse primer (10 µM), 2 µl reaction solution (cDNA), and 7.2 µl RNase free water. The standard protocol for a shuttle PCR was as follows: stage 1 included one cycle of initial denaturation at 95°C for 30 s; stage 2 included 40 cycles of the PCR at 95°C for 5 s and 60°C for 30 s, and stage 3 was the dissociation step. The data were evaluated according to the comparative threshold cycle (Cq) value. Finally, the relative levels of mRNA expression of 5-HT1A, PKA, CREB, and BDNF were normalized against that of glyceraldehyde-3-phosphate dehydrogenase (GAPDH) mRNA and calculated using the 2^−∆∆CT method. The primers used in this experiment as listed in Table 2.

Rats were sacrificed at the end of the experiments. Brain and colon tissues were then isolated, embedded on paraffinized blocks, and cut into 7 and 4 µm thick sections, respectively, using a microtome (Leica, Germany). Next, the sections were incubated with the anti-5-HT1A rabbit polyclonal antibody (Cat. No.D260012), anti PKA rabbit polyclonal antibody (Cat. NO.D164496), anti-CREB1 (Phospho-Ser133) rabbit polyclonal antibody (Cat. No.D155349), and anti-BDNF polyclonal antibody (Cat. No.D221057) (1:50) (Sangon Biotech, China), overnight at 4°C in a dilution ratio of 1:100 using the BondTM Primary Antibody Diluent (Servicebio, China). On the next day, the sections were incubated for 1 h with horseradish peroxidase 4-layered goat anti-rabbit secondary antibodies at 37°C (Sangon Biotech, China) according to the manufacturer’s instructions. Finally, the sections were treated with diaminobenzidine (DAB) solution (Servicebio, China) and visualized under a microscope (NIKON, Eclipse, Ci). We measured the integrated optical density (IOD) from at least three fields of each slice using the Image pro-plus 6.0 software, which could accurately reflect the complete expression of the proteins in immunohistochemical staining.

Three fecal samples were randomly collected from rats in the Normal, IBS, and CKF groups at the end of the therapy cycle for further 16S rDNA gene sequencing. Rats were sacrificed and the fresh stool samples from the colon were collected and immediately stored at −80°C until further analyses. Subsequently, bacterial genomic DNA was extracted from the frozen stool samples using the Qiagen QIAamp DNA stool Mini Kit (Hilden, Germany) according to the manufacturer’s instructions. Next, conventional PCR was performed to amplify the 16S rRNA in the V3-V4 region (341F-805R, F: GAT​CCT​ACG​GGA​GGC​AGC​A; R: GCT​TAC​CGC​GGC​TGC​TGG​C). Following were the PCR conditions: initial denaturation at 98°C for 1 min, followed by 30 cycles of denaturation at 98°C for 10 s, annealing at 50°C for 30 s, and elongation at 72°C for 60 s with a final hold at 72°C for 5 min. Samples were then purified using the MinElute Gel Extraction Kit (Qiagen, China), and samples with lengths of 400–450 bps were chosen for further experiments. Next, sequencing libraries were generated using the NEB Next Ultra DNA Library Prep Kit from Illumina (NEB, United States) following the manufacturer’s protocol. After the addition of the index codes, high-throughput pyrosequencing was performed for the PCR products on the Illumina MiSeq/HiSeq2500 platform (Biomarker Technologies Co., Ltd., Beijing, China). The QIIME software package (18.0 version) was employed to identify the operational taxonomic units (OTUs) and the RDP classifier was used to annotate the taxonomic information for each representative sequence using the Greengenes database (13.5 version). The weighted UniFrac distance-based principal coordinate analysis (PCoA) and binary Jaccard pair group method with arithmetic mean (UPGMA) clustering were performed using the QIIME software package. Finally, to identify the biomarker species in different groups, the linear discriminant analysis effect size (LEfSe) was estimated using the Metastats software. PICRUSt (Phylogenetic Investigation of Communities by Reconstruction of Unobserved States, http://picrust.github.io/picrust/) is a bioinformatic software package designed to predict the functional profiling of microbial communities based on the 16S rDNA sequences (Langille et al., 2013). STAMP (http://kiwi.cs.dal.ca/Software/STAMP) was used for functional profiling (Park et al., 2013).

Methanol (400 µl) was added to 100 µl of the serum sample, followed by vortexing for 60 s and centrifugation at 12,000 rpm for 10 min at 4°C. Subsequently, the supernatant was transferred into a fresh 20 µl vial for UPLC-MS analysis (Liu et al., 2018). The protocol followed is described in Supplementary Methods.

CKF comprises the following seven botanical drugs: Paeoniae alba Radix., Fagopyri dibotryis Rhizoma., Saposhnikoviae radix, Cuscutae semen, Rehmanniae radix, Coptidis rhizoma, and Periostracum cicadae (Fan et al., 2021). First, we performed an extensive literature review for these botanical drugs using the Traditional Chinese Medicine Systems Pharmacology Database and Analysis Platform (TCMSP) and retrieved the information on all chemical ingredients and their protein targets (https://sea.bkslab.org/). The molecules that met the criteria of bioavailability (OB) ≥30% and drug-likeness (DL) ≥0.1 (Xu et al., 2012) were used for further analysis. Second, the targets were identified after ascertaining the molecular targets associated with IBS and depression and were utilized for further analysis. Third, an interaction network was constructed for the identified targets using the STRING database (https://string-db.org/) to systematically elucidate the mechanism underlying the effects of CKF(3). Fourth, to explain the potential role of the active components in CKF in gene function and signaling pathways in the rat model, gene ontology (GO) enrichment and Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway analyses were performed for the hub genes using the Database for Annotation, Visualization and Integrated Discovery (DAVID, https://david.ncifcrf.gov/summary.jsp) tools for functional enrichment analysis (Chen et al., 2021). Species and background were limited to Rattus norvegicus (rat). The results were then plotted using the tools in Bioinformatics Web Server (http://www.bioinformatics.com.cn), an online platform for data analysis and visualization. Moreover, the gene-pathway network was constructed to screen the key genes targeted by CKF and relevant in the treatment of IBS and depression. To show the relationship among CKF active components, the hub targets, and KEGG pathways, a Sankey diagram was constructed using the tools of Bioinformatics Web Server (http://www.bioinformatics.com.cn).

All statistical analyses were performed using the SPSS Statistics 17.0 software, and all data were presented as mean ± SD. One-way analysis of variance (ANOVA) was used for multiple comparisons. p < 0.05 was considered statistically significant. To demonstrate the relationships between parameters, Spearman’s correlation was measured. The correlation coefficient was always in the range of +1 to −1. The closer was the absolute value of the correlation coefficient to 1, the closer was the relationship to being perfect linear.

After modeling IBS in rats, the body weights in the IBS group were found to be substantially lower than in those in the normal group (p < 0.001). The initial body weight and food intake were not statistically different between the IBS, CKF, and TM groups on day 1. After administration of CKF and TM for 22 days, the body weights in the CKF and TM groups increased significantly relative to the IBS group (p < 0.05, p < 0.001, Figure 2A).

We also measured the total fecal output and the water content in the stool for 4 h (Figure 3). The results showed that there was an increase in total fecal output and the stool water content in the IBS as compared to the normal rats (p < 0.001, Figures 2B,C). Moreover, the total fecal output and the water content in stool samples of rats in the CKF and TM groups decreased markedly as compared to the rats in the IBS group (p < 0.001, p < 0.001, Figures 2B,C).

The time for GI transit was used to evaluate intestinal mobility in rats. The results suggested that there was an increase in the GI transit rate in rats in the IBS group as compared to those in the normal group (p < 0.001, Figure 2D). This indicated that the frequency of GI transit increased significantly after the modeling of IBS. Moreover, rats in the CKF and TM groups exhibited decreased upper GI transit as compared to those in the IBS group (p < 0.001, p < 0.001).

There was a significant increase in the AWR scores in the IBS group as compared to the normal group, which indicated increased visceral sensitivity after the modeling of IBS. However, the AWR scores decreased significantly after CKF and TM treatments as compared to that in the IBS rats (Figure 2E), which suggested an improved visceral sensitivity after treatment with CKF and TM.

The antidepressant-like effects of CKF were assessed by sugar preference and forced swimming tests. The IBS group showed a significant decrease in the sucrose preference ratio (p < 0.001, Figure 2F). Treatment with CKF and TM significantly increased the sucrose preference ratio as compared to the IBS group (p < 0.001, p < 0.01, Figure 2F). Results of the forced swimming test showed that rats with IBS for 3 weeks had a longer period of immobility as compared to those in the control group (p < 0.001, Figure 2G). In addition, rats treated with CKF (5 mg/kg) and TM showed decreased immobility time as compared to those in the IBS group (p < 0.01, p < 0.01, Figure 2E). Collectively, these results suggested that CKF exhibited anti-depressant-like effects in the rat model of IBS.

To examine the efficacy of CKF in IBS rats, the expression levels of 5-HT and BDNF in the colon and hippocampal tissues were determined, and the findings are shown in Figures 4A,B,E,F. The results showed that the concentration of 5-HT and BDNF in colon tissues decreased significantly after CKF treatment as compared to the IBS group, whereas 5-HT and BDNF levels were markedly high in the hippocampal tissues after CKF treatment. To determine the levels of mRNA expression of 5-HT1A and BDNF, qPCR analysis was performed. The results showed that the levels of expression of both genes were higher in colon tissues, while lower in hippocampus tissues in the IBS group as compared to the normal group (Figures 4C,D,G,H). However, CKF treatment caused a significant decrease in the expressions of 5-HT1A and BDNF in the colon (Figures 4C,G), however, those in the hippocampus were upregulated (Figures 4D,H). Next, the immunohistological changes in the protein levels of 5-HT1A and BDNF were examined in the colon. The results suggested that stress caused an increase in the levels of 5-HT1A and BDNF on the membrane surface of the intestinal tissues, which decreased substantially after CKF treatment (Figures 5A,C). Furthermore, potential immunohistological changes in the protein expressions of 5-HT1A and BDNF in the dentate gyrus (DG) of the hippocampus were evaluated. The results showed a significant increase in 5-HT1A and BDNF-positive protein expression in the DG of the hippocampus in the CKF group relative to the IBS group (Figures 5B,D).

Using qPCR analysis, we examined the gene expressions of PKA and CREB. The results showed higher levels of PKA and CREB expression in colon tissues; the expression of both genes in the hippocampal tissues of rats in the IBS group reduced as compared to those in the normal group (Figure 4). However, CKF treatment caused a decline in the expression of PKA and CREB in colon tissues (Figures 4I,K), and an increase in the hippocampi (Figures 4J,L). Next, we examined the immunohistological changes in protein levels of PKA and CREB in the colon. The results showed that stress caused an increase in PKA and CREB density on the membrane surface of the intestinal tissues, which decreased after CKF treatment (Figures 6A,C). Furthermore, potential immunohistological changes in the protein levels of PKA and CREB in the DG of the hippocampus were examined. Results showed a significant increase in PKA and CREB-positive protein in the DG of hippocampus tissues obtained from rats in the CKF group as compared to those in the IBS group (Figures 6B,D).

In our study, 16S rDNA sequencing was conducted to examine whether CKF modulated gut microbiota and characterized these modifications in the composition of gut microbiota. Therefore, we sequenced the V3 + V4 region of bacterial 16S rRNA to evaluate the effects of CKF on the composition of gut microbiota. High-throughput pyrosequencing of the samples yielded 720,355 pairs of raw reads. After alignment of the pair-end reads and filtering, we obtained 718,081 clean tags for subsequent analyses. The effective reads were then clustered into OTUs based on the 97% similarity cut-off. The Shannon diversity index was higher in the IBS group related to the Normal and CKF groups (Figure 7E). Next, structural changes in the gut microbiota were evaluated using unsupervised multivariate statistical methods and principal coordinates analysis (PCoA). Results of PCoA showed that the three groups had distinct clustering for microbiota composition. Rats in the CKF group showed a structure similar to those in the Normal group (Figures 7A–D). In addition, analysis of the UPGMA clustering tree (Figure 7F) showed that the community structure of the CKF group was more similar to that of the normal group than the IBS group, indicating that CKF could significantly modulate the composition of the gut microbiota.

Further analysis of the structure of the microbial community showed a significant positive effect of CKF at both phylum, order, and genus levels. Results obtained after analysis at the phylum level showed a significant decrease in the relative abundance of Bacteroidetes in IBS rats and CKF treatment could restore these altered levels (Figure 7G). Moreover, the F-B ratio was higher in IBS rats as compared to normal rats, and treatment with CKF decreased the F-B ratio in IBS rats. To further examine these differences among samples, analysis was performed at the order and genus level and the results were similar to those obtained at the phylum level.

In IBS rats, analysis at the order level showed a significant increase in the relative abundances of Clostridiales and Corynebacteriales belonging to the Firmicutes and Actinobacteria phyla (Figure 7H). Notably, CKF treatment could significantly restore the relative abundances of Clostridiales and Corynebacteriales.

Analysis at the genus level showed a significant increase in the relative abundance of Christensenellaceae_R-7_group, uncultured_bacterium_f_Ruminococcaceae, uncultured_bacterium_f_Lachnospiraceae belonging to the Firmicutes phylum in IBS rats (Figure 7I). However, a decrease in the relative abundance of Lactobacillus and Dubosiella, belonging to the Firmicutes phylum, was observed in IBS rats. Notably, CKF treatment could significantly restore the relative abundance of Christensenellaceae_R-7_group,uncultured_bacterium_f_Ruminococcaceae, uncultured_bacterium_f_Lachnospiraceae, Dubosiella, and Lactobacillus in IBS rats.

Furthermore, the profile of gut microbiota indicated that CKF could significantly modulate the structure of the gut microbiota (Supplementary Table S2). To identify key phylotypes that were significantly altered after CKF treatment, all effective sequences were analyzed by the LEfSe method. Results of the LEfSe analysis showed the presence of high-dimensional biomarkers in the gut microbiota in each group.

As shown in Figure 8A, on comparing the significant changes of gut microbiota in IBS vs. Normal groups, the dominant kinds in the IBS group were as follows: s__uncultured_bacterium_g_Christensenellaceae, _R_7_group, g__Christensenellaceae_R_7_group, f__ Christensenellaceae, s__uncultured_bacterium_f_Lachnospiraceae, g__uncultured_bacterium_f_Lachnospiraceae, s__uncultured_bacterium_f_Ruminococcaceae, g__uncultured_bacterium_f_Ruminococcaceae, f__Ruminococcaceae, s_uncultured_bacterium_f_Desulfovibrionaceae, g__uncultured_bacterium_f_Desulfovibrionaceae, c__ Clostridia, and o__Clostridiales.

The dominant types in the Normal group were s__uncultured_bacterium_g_Lactobacillus, g__Lactobacillus, f__Lactobacillaceae, o__Lactobacillales, s__uncultured_bacterium_g_Dubosiella, g__Dubosiella, s__uncultured_bacterium_g_Dubosiella, g__Dubosiella, c__Bacilli, f__Lactobacillaceae, g__Lactobacillu, s__uncultured_bacterium_g_Lactobacillus, and o__Lactobacillales.

As shown in Figure 8B on comparing the significant changes in gut microbiota in CKF vs. IBS groups, the dominant types in the CKF group were identified as follows: s__uncultured_bacterium_f_Muribaculaceae, g__uncultured_bacterium_f_Muribaculaceae, f__Muribaculaceae, s__uncultured_bacterium_g_Lactobacillus, g__Lactobacillus, f__Lactobacillaceae, o__Lactobacillales, s__uncultured_bacterium_g_Dubosiella, and g__Dubosiella.

The dominant types in IBS group were s__uncultured_bacterium_g_Adlercreutzia, s__uncultured_bacterium_g_Christensenellaceae_R_7_group, g__Christensenellaceae_R_7_group, f__Christensenellaceae, s__uncultured_bacterium_g_Agathobacter, g__Agathobacter, s__uncultured_bacterium_f_Lachnospiraceae, g__uncultured_bacterium_f_Lachnospiraceae, f__Lachnospiraceae, s__uncultured_bacterium_g_Ruminococcaceae_NK4A214_group, g__Ruminococcaceae_NK4A214_group, s__uncultured_bacterium_g__Eubacterium__coprostanoligenes, f__Ruminococcaceae, s__uncultured_bacterium_f_Desulfovibrionaceae, g__uncultured_bacterium_f_Desulfovibrionaceae, s__uncultured_bacterium_o_Mollicutes_RF39, g__uncultured_bacterium_o_Mollicutes_RF39, f__uncultured_bacterium_o_Mollicutes_RF39, o__Mollicutes_RF39, c__ clostridia, and o__clostridiales.

As shown in Supplementary Table S3 on comparing the significant changes in gut microbiota between the CKF and Normal groups, the dominant types were identified, including f__Lachnospiraceae, g__Lachnospiraceae_NK4A136, g__Eubacterium__coprostanoligenes, and s__uncultured_bacterium_g__Eubacterium__coprostanoligee.

Supplementary Table S3 also suggest that organisms in the Lactobacillus and Dubosiella genus were mainly enriched in the Normal and CKF groups, whereas Christensenellaceae_R_7_group, uncultured_bacterium_f_Lachnospiraceae, uncultured_bacterium_f_Ruminococcaceae, and uncultured_bacterium_f_Desulfovibrionaceae were mainly enriched in the IBS group. Collectively, these findings suggested that CKF treatment could reverse IBS-induced dysbiosis of the gut microbiota.

Metabolomics is an effective method to identify the interactions between the host and gut microbiota. Consequently, UPLC-MS/MS was used to examine serum metabolites. Serum metabolic profiling (PCA) showed that the Normal, IBS, and CKF groups could be distinguished, and the CKF group was closer to the Normal group than the IBS group (Figure 9A).

The data obtained from the serum samples were analyzed using the OPLS-DA score. The OPLS-DA score plot presented a distinct clustering of metabolites in serum samples between IBS and CKF groups as well as between the normal and IBS groups, suggesting that both IBS and CKF substantially influenced the metabolic profiles (Figures 9B,C). Following the OPLS-DA test, the prospective differential metabolites were evaluated as the integral, with a varied significance (VIP > 1.0 and p < 0.05) using the Student t-test. In total, 118 metabolites were significantly differentially expressed in the serum of IBS rats as compared to the normal rats (Supplementary Table S4). Treatment with CKF could reverse the levels of nine metabolites towards the normal, including L-glutamic acid, adenosine, indoleacetaldehyde, serotonin, scopoletin, acetylcholine, alanyltryptophan, perillyl aldehyde, and taurine (Figure 9D). The altered metabolites and differential metabolic pathways were identified by KEGG analysis (http://www.genome.jp/kegg/pathway.html). This analysis suggested their involvement in 50 metabolic pathways (Supplementary Table S5). Among them, metabolomics showed that amino acids (tryptophan, glutamate, taurine, hypotaurine, histidine, alanine, aspartate, glutamate, arginine, and proline), SCFAs-related (Butanoate), and 5-HT-related metabolic pathways (neuroactive ligand-receptor interaction, cAMP signaling pathway, serotonergic synapses, and inflammatory mediator regulation of TRP channels) induced by IBS were attenuated after CKF treatment. KEGG pathway enrichment analysis demonstrated that nine differentially expressed metabolites were enriched in 50 different signaling pathways; among them, 38 were significantly enriched (p < 0.05) (Figure 9E). To examine the possible metabolic pathways influenced by CKF, pathway analysis was performed following metabolomics using MetPA (http://www.metaboanalyst.ca/faces/home.xhtml). As shown in Figure 9F, 18 metabolic pathways were identified as important for understanding the effects of CKF on CACS-induced IBS. These results suggested that the differentially expressed metabolites in the serum of CKF and IBS rats may be associated with amino acids, SCFAs metabolism, and 5-HT related metabolic pathways. Tryptophan and butanoate metabolism were found to be the key regulatory pathways, thereby suggesting that the metabolites enriched in tryptophan and butanoate metabolism may substantially regulate the gut microbiota in CKF and IBS groups. Taken together, these findings suggested that serum metabolites effectively responded to both IBS and CKF administration and that IBS-induced serum metabolism aberrations were improved by CKF treatment.

Moreover, 36 metabolites were significantly differentially expressed in the sera of rats in the CKF group as compared to the normal group (Supplementary Table S6). The altered metabolites and differential metabolic pathways were identified by KEGG analysis. This analysis suggested their involvement in 17 metabolic pathways (Supplementary Table S7). Among them, the metabolite-related pathways were enriched in tyrosine and lactic acid metabolism pathways (Supplementary Figure S2 and Supplementary Table S7).

A complete literature search of the CKF mix resulted in the identification of 55 chemical constituents and 767 putative protein targets, of which, 92 targets were ascertained to be associated with IBS and depression-like behaviors (Figure 10A). Next, a PPI network was constructed to examine the interaction between CKF-related hub genes in IBS and CKF-known therapeutic hub genes for IBS. The parameters for network construction were set at high confidence (interaction score ≥0.7) and the disconnected nodes were excluded from the network. The PPI network consisted of 92 nodes and 324 edges, and the two-fold median of node degree was 7.04 (Figure 10B). The major hub genes included HTR1A and CREB1. The major hub genes in the PPI network were then input into clusterProfiler for GO and KEGG analyses. A total of 436 GO terms were ascertained, including 328 biological processes (BP), 48 molecular functions (MF), and 60 cellular components (CC). The top ten terms in each category are represented in a bar chart (Figure 10C). In total, 90 KEGG pathways were identified. Among them, neuroactive ligand-receptor interaction, cAMP signaling, calcium signaling, long-term depression, and serotonergic synapses were the major pathways with the highest gene ratios and the lowest p-values. This suggested that they may play important roles in the CKF treatment for IBS. The top 15 KEGG pathway terms are shown in an advanced bubble chart (Figure 10D). The relationship between CKF botanical drugs, hub genes, and 5-HT related metabolic pathways (neuroactive ligand-receptor interaction, cAMP signaling pathway, calcium signaling pathway, long-term depression, and serotonergic synapses) are presented in the Sankey diagram (Figure 11A). Collectively, these results suggested that CKF may exert its therapeutic effects on IBS through multiple biological processes and pathways.

PICRUSt analysis based on the 16S rRNA gene sequence data was used to investigate the gut microbiome functions related to CKF treatment. The potential metabolic functions of the gut microbiota may be attributed to the alterations in bacterial taxa. The gut microbiota functions were analyzed using the KEGG database. As shown in Figure 12A, 15 different functional pathways were identified between the Normal and IBS groups. Those involved in phenylalanine, tyrosine, and tryptophan biosynthesis were upregulated, while the pathways involved in taurine and hypotaurine metabolism were downregulated in the IBS rats. CKF treatment downregulated phenylalanine, tyrosine, and tryptophan biosynthesis pathways, whereas upregulated D-glutamine and D-glutamate metabolism, along with the cAMP signaling pathway (Figure 12B). Additionally, we structured the metabolism-protein networks and identified the serum metabolites and CKF key protein targets involved in the cAMP signaling pathway (Figure 11B). The 5-HT-PKA-CREB-BDNF axis was involved in the cAMP signaling pathway. The serum metabolites and CKF key protein targets were also involved in the neuroactive ligand-receptor interaction and serotonergic synapses (Supplementary Figures S3A,B). The CKF key protein target (5-HT1A) was also involved in serotonergic synapses. The key serum metabolites (l-glutamate, taurine) were also found to participate in neuroactive ligand-receptor interactions.

The relationships among host phenotypes, serum metabolites, and the gut microbiota were comprehensively analyzed. The correlation matrix was generated by calculating Spearman’s correlation coefficient (Supplementary Figure S4 and Supplementary Table S8). In the correlation between serum metabolites and gut microbiota, uncultured_bacterium_f_Lachnospiraceae were found to be negatively correlated with tryptophan-related metabolites (indoleacetaldehyde), neuroactive ligand-receptor interaction-related metabolite (taurine), and butanoate-related metabolite (L-glutamic acid), whereas two bacterial genera, including Dubosiella and Lactobacillus, showed a positive correlation with the corresponding aforementioned serum metabolites. Clostridiales was found to be negatively correlated with indoleacetaldehyde and taurine and Corynebacteriales was found to be positively correlated with serotonin.

For gut microbiota and phenotypes, 20 mmg AWR Score and GI transit were found to be negatively correlated with Dubosiella and Lactobacillus, whereas positively correlated with Clostridiales and uncultured_bacterium_f_Lachnospiraceae. Sucrose Preference were found to be positively correlated with Dubosiella and Lactobacillus, whereas negatively correlated with Clostridiales, uncultured_bacterium_f_Ruminococcaceae and uncultured_bacterium_f_Lachnospiraceae. Clostridiales and Corynebacteriales were found to be positively correlated with FST immobility time.

For gut microbiota and indicators of molecular biology, including the concentrations of 5-HT and BDNF in the colon were found to be negatively correlated with Lactobacillus, whereas positively correlated with 5-HT and BDNF concentrations, the mRNA expressions of 5-HT1A, BDNF, PKA, and CREB in the hippocampus. 5-HT and BDNF concentrations in the colon were positively correlated with Clostridiales and uncultured_bacterium_f_Lachnospiraceae, whereas negatively correlated with 5-HT and BDNF concentrations, the mRNA expressions of 5-HT1A, BDNF, PKA, and CREB in the hippocampus. The gut microbiota could thus significantly affect the host phenotypes, molecular biology indicators, and metabolite concentrations.