Polygalae Radix Oligosaccharide Esters May Relieve Depressive-like Behavior in Rats with Chronic Unpredictable Mild Stress via Modulation of Gut Microbiota

Polygalae radix (PR) is a well-known traditional Chinese medicine that is used to treat depression, and polygalae radix oligosaccharide esters (PROEs) are the main active ingredient. Although gut microbiota are now believed to play key role in depression, the effects of PROEs on depression via modulation of gut microbiota remain unknown. In this article, we investigate the effect of PROEs on the gut microbiota of a depression rat and the possible mechanism responsible. The depression rat model was induced by solitary rearing combined with chronic unpredictable mild stress (CUMS). The depression-like behavior, the influence on the hypothalamic–pituitary–adrenal (HPA) axis, the contents of monoamine neurotransmitter in the hippocampus, and the quantity of short-chain fatty acids (SCFAs) in the feces were each assessed, and the serum levels of lipopolysaccharide (LPS) and interleukin-6 (IL-6) were measured by ELISA. Additionally, ultrastructural changes of the duodenal and colonic epithelium were observed under transmission electron microscope, and the gut microbiota were profiled by using 16S rRNA sequencing. The results show that PROEs alleviated the depression-like behavior of the depression model rats, increased the level of monoamine neurotransmitters in the brain, and reduced the hyperfunction of the HPA axis. Furthermore, PROEs regulated the imbalance of the gut microbiota in the rats, relieving intestinal mucosal damage by increasing the relative abundance of gut microbiota with intestinal barrier protective functions, and adjusting the level of SCFAs in the feces, as well as the serum levels of LPS and IL-6. Thus, we find that PROEs had an antidepressant effect through the restructuring of gut microbiota that restored the function of the intestinal barrier, reduced the release of intestinal endotoxin, and constrained the inflammatory response.


Introduction
As a mood disorder disease, depression is characterized by anhedonia, unresponsiveness, poor appetite, hypokinesia, insomnia, and suicidal tendencies [1], and according to the World Health Organization, depression will be the largest contributor to worldwide disease burden by 2030 [2]. The pathogenesis of depression is complex and is usually associated with dysregulation of the hypothalamic-pituitary-adrenal (HPA) axis, monoaminergic deficiencies, immune system dysfunction, and gut microbiota disorders [3]. Recent studies have demonstrated that gut microbiota and their metabolites in particular play an important role in the pathogenesis and treatment of depression.
The gut microbiome includes all the microorganisms that inhabit the gut and their respective genomes [4]. Gut microbiota regulate a variety of host metabolic pathways, Figure 1. The effects of PROEs on CUMS-induced depression-like behaviors in rats. (A) The bo weights of rats; (B) the sucrose preference rate; (C) the immobility time in the FST; (D-F) the to movement distance, the number of rearings, and the time spent in the central area of the OFT; da represent mean ± SEM (n = 8); ## p < 0.01, and ### p < 0.001 compared to the NC group; ** p < 0.01 a *** p < 0.001 compared to the CUMS group.

The Effect of PROEs on Monoamine Neurotransmitters and Their Metabolites in the Hippocampus
As shown in Table 1, the content of 5-HT, NE, DA, DOPAC, 5-HIAA, and HVA we significantly lower in the hippocampus of CUMS rats compared to the NC group (p < 0.0 but the content of 5-HT, NE, DOPAC, and 5-HIAA were significantly higher in the FX and PROE groups compared to the CUMS group (p < 0.05). Furthermore, compared to t CUMS group, the content of DA was significantly higher in the FXT and PROE-H grou (p < 0.01), and the content of HVA was significantly higher in the FXT, PROE-H, an PROE-M groups (p < 0.01).  (D-F) the total movement distance, the number of rearings, and the time spent in the central area of the OFT; data represent mean ± SEM (n = 8); ## p < 0.01, and ### p < 0.001 compared to the NC group; ** p < 0.01 and *** p < 0.001 compared to the CUMS group.

The Effect of PROEs on Monoamine Neurotransmitters and Their Metabolites in the Hippocampus
As shown in Table 1, the content of 5-HT, NE, DA, DOPAC, 5-HIAA, and HVA were significantly lower in the hippocampus of CUMS rats compared to the NC group (p < 0.01), but the content of 5-HT, NE, DOPAC, and 5-HIAA were significantly higher in the FXT and PROE groups compared to the CUMS group (p < 0.05). Furthermore, compared to the CUMS group, the content of DA was significantly higher in the FXT and PROE-H groups (p < 0.01), and the content of HVA was significantly higher in the FXT, PROE-H, and PROE-M groups (p < 0.01).

The Effect of PROEs on HPA Axis
As shown in Figure 2A-C, the levels of CORT, ACTH, and CRF were significantly higher in the CUMS group compared to the NC group (p < 0.001), and the serum levels of CORT and ACTH were significantly lower in the FXT and PROE groups compared to the (A−C) The levels of CORT, ACTH, and CRF in Serum; (D−F) the levels of TRP and KYN in plasma and the ratio of TRP/KYN; (G,H) the levels of IL-6 and LPS in serum; data represent mean ± SEM (n = 6); ## p < 0.01 and ### p < 0.001 compared to the NC group; * p < 0.05, ** p < 0.01 and *** p < 0.001 compared to the CUMS group.

Changes in Gut Microbiota Diversity
As shown in Figure 3A,B, the Shannon and Chao 1 index suggests that the CUMS challenge led to a significant increase in gut microbiota diversity, but the PROE-H groups had significant reductions in gut microbiota diversity compared to the CUMS group (p < 0.01). In addition, our alpha diversity analysis indicated that there was a significantly higher amount of richness and diversity in the gut microbiota of the CUMS group compared to the NC and PROE groups. Beta diversity analysis, however, usually begins by calculating the distance matrix to determine differences in overall microbial composition between samples. To this end, principal co-ordinate analysis (PCoA) showed that the structure of the gut microbiota in the CUMS group was significantly separated from that in the NC and NC-PROE groups, whereas the structure of the gut microbiota in the PROE group deviated from the CUMS group and approached that of the NC group ( Figure 3C). Furthermore, the unweighted pair group method with arithmetic mean (UPGMA) also yielded the same results, indicating that PROEs changed the gut microbiota structure of the host ( Figure 3D). (A-C) The levels of CORT, ACTH, and CRF in Serum; (D-F) the levels of TRP and KYN in plasma and the ratio of TRP/KYN; (G,H) the levels of IL-6 and LPS in serum; data represent mean ± SEM (n = 6); ## p < 0.01 and ### p < 0.001 compared to the NC group; * p < 0.05, ** p < 0.01 and *** p < 0.001 compared to the CUMS group.
2.5. The Effects of PROEs on Gut Microbiota Composition 2.5.1. Changes in Gut Microbiota Diversity As shown in Figure 3A,B, the Shannon and Chao 1 index suggests that the CUMS challenge led to a significant increase in gut microbiota diversity, but the PROE-H groups had significant reductions in gut microbiota diversity compared to the CUMS group (p < 0.01). In addition, our alpha diversity analysis indicated that there was a significantly higher amount of richness and diversity in the gut microbiota of the CUMS group compared to the NC and PROE groups. Beta diversity analysis, however, usually begins by calculating the distance matrix to determine differences in overall microbial composition between samples. To this end, principal co-ordinate analysis (PCoA) showed that the structure of the gut microbiota in the CUMS group was significantly separated from that in the NC and NC-PROE groups, whereas the structure of the gut microbiota in the PROE group deviated from the CUMS group and approached that of the NC group ( Figure 3C). Furthermore, the unweighted pair group method with arithmetic mean (UPGMA) also yielded the same results, indicating that PROEs changed the gut microbiota structure of the host ( Figure 3D).

Changes in the Composition of Gut Microbiota
Taxonomical analysis of microbial structure at the phylum level showed that Firmicutes and Bacteroidetes were the main phyla in all groups. Compared to the NC group, the relative abundance of the Firmicutes and actinobacteria of the CUMS group was significantly decreased (p < 0.01), and the relative abundances of Bacteroidetes and proteobacteria were significantly elevated, but PROE treatment restored the relative abundance of the microbiota (Figure 4).

Changes in the Composition of Gut Microbiota
Taxonomical analysis of microbial structure at the phylum level showed that Firmicutes and Bacteroidetes were the main phyla in all groups. Compared to the NC group, the relative abundance of the Firmicutes and actinobacteria of the CUMS group was significantly decreased (p < 0.01), and the relative abundances of Bacteroidetes and proteobacteria were significantly elevated, but PROE treatment restored the relative abundance of the microbiota ( Figure 4). . The effects of PROEs on the relative abundance of gut microbiota (A) and significantly changed gut microbiota (B) at the phylum level in depression-model rats (n = 8); # p < 0.05, ## p < 0.01 compared to the NC group; * p < 0.05, ** p < 0.01 and *** p < 0.001 compared to the CUMS group.

Changes in the Composition of Gut Microbiota
Taxonomical analysis of microbial structure at the phylum level showed that Firmicutes and Bacteroidetes were the main phyla in all groups. Compared to the NC group, the relative abundance of the Firmicutes and actinobacteria of the CUMS group was significantly decreased (p < 0.01), and the relative abundances of Bacteroidetes and proteobacteria were significantly elevated, but PROE treatment restored the relative abundance of the microbiota ( Figure 4).  At the family level ( Figure 5), redundancy analysis (RDA) indicated that the CUMS group was clearly separated from the NC group, and the PROE-H group was close to the NC group. Bacteroidaceae, Muribaculaceae, and Lachnospiraceae were most affected by the CUMS protocol, whereas Peptostreptococcaceae, Ruminococcaceae, Lachnospiraceae, and Prevotellaceae were the dominant families in the NC and PROE-H groups. Additionally, the relative abundances of Bacteroidaceae and Muribaculaceae were significantly improved (p < 0.05), and those of Peptostreptococcaceae, Ruminococcaceae, and Lachnospiraceae were strikingly decreased in the CUMS group compared to the NC group (p < 0.05). Interestingly, the PROE treatment restored the relative abundances of the above gut microbiota at the family level.
At the family level ( Figure 5), redundancy analysis (RDA) indicated that the CUMS group was clearly separated from the NC group, and the PROE-H group was close to the NC group. Bacteroidaceae, Muribaculaceae, and Lachnospiraceae were most affected by the CUMS protocol, whereas Peptostreptococcaceae, Ruminococcaceae, Lachnospiraceae, and Prevotellaceae were the dominant families in the NC and PROE-H groups. Additionally, the relative abundances of Bacteroidaceae and Muribaculaceae were significantly improved (p < 0.05), and those of Peptostreptococcaceae, Ruminococcaceae, and Lachnospiraceae were strikingly decreased in the CUMS group compared to the NC group (p < 0.05). Interestingly, the PROE treatment restored the relative abundances of the above gut microbiota at the family level. At the genus level ( Figure 6), the relative abundances of Bacteroides, Oscillibacter, Parasutterella, and Intestinimonas were significantly higher here (p < 0.05), but those of Romboutsia, Roseburia, Lachnospiraceae_NK4A136_group, Prevotella_9, and Eubacterium_copros-tanoligenes_group were significantly lower in the CUMS group compared to the NC group (p < 0.05). However, once again, the PROE-H treatment restored the composition of the above gut microbiota at the genus level. At the genus level ( Figure 6), the relative abundances of Bacteroides, Oscillibacter, Parasutterella, and Intestinimonas were significantly higher here (p < 0.05), but those of Romboutsia, Roseburia, Lachnospiraceae_NK4A136_group, Prevotella_9, and Eubacterium_coprostanoli-genes_group were significantly lower in the CUMS group compared to the NC group (p < 0.05). However, once again, the PROE-H treatment restored the composition of the above gut microbiota at the genus level.

Taxonomic Biomarkers in Rat Gut Microbiota
Linear discriminant analysis of effect size (Lefse) showed that there were 16 identified taxonomic biomarkers found with p-values < 0.01 and LDA scores (log 10) > 4.0, among which four were enriched in the CUMS group and three were enriched in the PROE-H group ( Figure 7A). The four biomarkers in the CUMS group were all from Bacteroidetes and included Bacteroidaceae, Muribaculaceae, Bacteroides_acidifaciens, and Bacteroides. The three biomarkers enriched in the PROE-H group were from Firmicutes and Bacteroidetes and included Prevotella_9, Peptostreptococcaceae, and Romboutsia.

Taxonomic Biomarkers in Rat Gut Microbiota
Linear discriminant analysis of effect size (Lefse) showed that there were 16 identified taxonomic biomarkers found with p-values < 0.01 and LDA scores (log 10) > 4.0, among which four were enriched in the CUMS group and three were enriched in the PROE-H group ( Figure 7A). The four biomarkers in the CUMS group were all from Bacteroidetes and included Bacteroidaceae, Muribaculaceae, Bacteroides_acidifaciens, and Bacteroides. The three biomarkers enriched in the PROE-H group were from Firmicutes and Bacteroidetes and included Prevotella_9, Peptostreptococcaceae, and Romboutsia.   Heat map analysis of the effects of PROEs on gut microbiota (A) and significantly changed gut microbiota (B) at the genus level in depression-model rats (n = 8); # p < 0.05, ## p < 0.01 and ### p < 0.001 compared to the NC group; * p < 0.05, ** p < 0.01 and *** p < 0.001 compared to the CUMS group.

Taxonomic Biomarkers in Rat Gut Microbiota
Linear discriminant analysis of effect size (Lefse) showed that there were 16 identified taxonomic biomarkers found with p-values < 0.01 and LDA scores (log 10) > 4.0, among which four were enriched in the CUMS group and three were enriched in the PROE-H group ( Figure 7A). The four biomarkers in the CUMS group were all from Bacteroidetes and included Bacteroidaceae, Muribaculaceae, Bacteroides_acidifaciens, and Bacteroides. The three biomarkers enriched in the PROE-H group were from Firmicutes and Bacteroidetes and included Prevotella_9, Peptostreptococcaceae, and Romboutsia.

Prediction of Metagenomic Functions
The KEGG pathways at the level-2 results showed that there were significant differences in terms of the nervous system, amino acid metabolism, immune system diseases, and neurodegenerative disease pathways between the CUMS group and NC group (p < 0.05), which illustrates that gut microbiota may reflect the physiological state of their host. Com-pared to the CUMS group, PROEs had significant differences in nervous system, amino acid metabolism, lipid metabolism, immune system diseases, and neurodegenerative disease pathways (p < 0.05). In addition, KEGG pathways at level 3 showed that the PROE-H group regulated depression-like pathophysiology by affecting amino acid metabolism and lipid metabolism, including tryptophan metabolism, valine/leucine/isoleucine degradation, glycine/serine/threonine metabolism, and sphingolipid metabolism ( Figure 8).

Prediction of Metagenomic Functions
The KEGG pathways at the level-2 results showed that there were significant differences in terms of the nervous system, amino acid metabolism, immune system diseases, and neurodegenerative disease pathways between the CUMS group and NC group (p < 0.05), which illustrates that gut microbiota may reflect the physiological state of their host. Compared to the CUMS group, PROEs had significant differences in nervous system, amino acid metabolism, lipid metabolism, immune system diseases, and neurodegenerative disease pathways (p < 0.05). In addition, KEGG pathways at level 3 showed that the PROE-H group regulated depression-like pathophysiology by affecting amino acid metabolism and lipid metabolism, including tryptophan metabolism, valine/leucine/isoleucine degradation, glycine/serine/threonine metabolism, and sphingolipid metabolism ( Figure 8).

The Effect of PROEs on Plasma Tryptophan and Kynurenine Levels
Compared to the NC group, the levels of TRP were significantly lower (p < 0.001), and the levels of KYN and the TRP/KYN ratio were significantly higher in the CUMS group (p < 0.01, Figure 2D-F). In the FXT and PROE groups, the levels of TRP were significantly higher than in the CUMS group (p < 0.01), but the levels of KYN and the TRP/KYN ratio were significantly lower than in the CUMS group (p < 0.05). These results demonstrate that PROEs may regulate the TRP-KYN metabolic pathway. The results of our histopathological examination show that the duodenum and colon had intact mucosa, clear crypt (C), and villus (V) structure with adequate goblet cells in the NC group ( Figure 9). The duodenal villi were severely shedding in depressionmodel rats, however, and the duodenal and colonic crypts were distorted, with massive inflammatory cell infiltration in the mucosa. However, PROE administration reversed these histopathological changes. Compared to the NC group, the length of the duodenal villi was significantly lower in the CUMS group (p < 0.001), the colonic villi had a downward trend, the depth of duodenal and colonic crypts were significantly higher (p < 0.001), and the ratio of V/C was significantly lower (p < 0.01). Furthermore, compared to the CUMS group, the length of the duodenal villi was significantly lower in the FXT and PROE-H groups (p < 0.001); these colonic villi had an upward trend, the depth of duodenal and colonic crypts were significantly lower (p < 0.001), and the ratio of V/C was significantly higher (p < 0.01). ella and Oscillibacter.

The Effect of PROEs on Plasma Tryptophan and Kynurenine Levels
Compared to the NC group, the levels of TRP were significantly lower (p < 0.001), and the levels of KYN and the TRP/KYN ratio were significantly higher in the CUMS group (p < 0.01, Figure 2D-F). In the FXT and PROE groups, the levels of TRP were significantly higher than in the CUMS group (p < 0.01), but the levels of KYN and the TRP/KYN ratio were significantly lower than in the CUMS group (p < 0.05). These results demonstrate that PROEs may regulate the TRP-KYN metabolic pathway.

The Effects of PROEs on Duodenum and Colon Histopathological Changes
The results of our histopathological examination show that the duodenum and colon had intact mucosa, clear crypt (C), and villus (V) structure with adequate goblet cells in the NC group ( Figure 9). The duodenal villi were severely shedding in depression-model rats, however, and the duodenal and colonic crypts were distorted, with massive inflammatory cell infiltration in the mucosa. However, PROE administration reversed these histopathological changes. Compared to the NC group, the length of the duodenal villi was significantly lower in the CUMS group (p < 0.001), the colonic villi had a downward trend, the depth of duodenal and colonic crypts were significantly higher (p < 0.001), and the ratio of V/C was significantly lower (p < 0.01). Furthermore, compared to the CUMS group, the length of the duodenal villi was significantly lower in the FXT and PROE-H groups (p < 0.001); these colonic villi had an upward trend, the depth of duodenal and colonic crypts were significantly lower (p < 0.001), and the ratio of V/C was significantly higher (p < 0.01). data represent mean ± SEM (n = 10); ## p < 0.01 and ### p < 0.001 compared to the NC group; *** p < 0.001 compared to the CUMS group.

Effects of PROEs on Duodenum and Colon Epithelium Ultrastructure
In the NC group, the microvilli in the duodenum and colon were rich and orderly, and the structure was complete, with normal intercellular tight junctions, as observed by electron microscope. However, in the CUMS group, the duodenum and colon mucosa were seriously damaged, the microvilli were atrophic and even deficient; gaps in intercellular tight junctions became significantly larger, and the goblet cells had disappeared almost entirely. The intestinal mucosa injury of the FXT and PROE-H groups was improved, and the microvilli of the duodenum and colon were dense and arranged in an orderly manner. Moreover, the structure was complete, and the tight junctions returned to normal after treatment ( Figure 10).  ,E) The effects of PROEs on duodenum and colon histopathological changes; (B,F) the length of duodenal and colonic villi; (C,G) the depth of duodenal and colonic crypts; (D,H) the ratio of V/C in the duodenum and colon; data represent mean ± SEM (n = 10); ## p < 0.01 and ### p < 0.001 compared to the NC group; *** p < 0.001 compared to the CUMS group.

Effects of PROEs on Duodenum and Colon Epithelium Ultrastructure
In the NC group, the microvilli in the duodenum and colon were rich and orderly, and the structure was complete, with normal intercellular tight junctions, as observed by electron microscope. However, in the CUMS group, the duodenum and colon mucosa were seriously damaged, the microvilli were atrophic and even deficient; gaps in intercellular tight junctions became significantly larger, and the goblet cells had disappeared almost entirely. The intestinal mucosa injury of the FXT and PROE-H groups was improved, and the microvilli of the duodenum and colon were dense and arranged in an orderly manner. Moreover, the structure was complete, and the tight junctions returned to normal after treatment ( Figure 10).

Effects of PROEs on the Protein Expression of Occludin in the Colon
The protein expression of occludin in the colon of the CUMS group was significantly lower compared to the NC group (p < 0.05). Compared to the CUMS group, the protein

Effects of PROEs on the Protein Expression of Occludin in the Colon
The protein expression of occludin in the colon of the CUMS group was significantly lower compared to the NC group (p < 0.05). Compared to the CUMS group, the protein expression of occludin was significantly higher in the PROE groups (p < 0.05), and this expression in the FXT group had an upward trend ( Figure 11).

The Effect of PROEs on Serum IL-6 and LPS Levels
As presented in Figure 2G,H, the levels of IL-6 and LPS were significantly higher in the CUMS group compared to the NC group (p < 0.001) but were significantly lower in the FXT and PROE groups as compared to the CUMS group (p < 0.05). expression of occludin was significantly higher in the PROE groups (p < 0.05), and expression in the FXT group had an upward trend ( Figure 11). Figure 11. The effects of PROEs on the protein expression of occludin in the colon; data repre mean ± SEM (n = 3); # p < 0.05 compared to the NC group; * p < 0.05, ** p < 0.01 compared to CUMS group.

The Effect of PROEs on Serum IL-6 and LPS Levels
As presented in Figure 2G,H, the levels of IL-6 and LPS were significantly highe the CUMS group compared to the NC group (p < 0.001) but were significantly lower in FXT and PROE groups as compared to the CUMS group (p < 0.05).

The Effect of PROEs on the Expression of 5-HT1A, 5-HT2A, IDO1, and TNF-α mRNA
As shown in Figure 12A-F, compared to the CUMS group, the expression of 5-H mRNA was significantly lower in the FXT and PROE groups (p < 0.01), and the express of 5-HT1A mRNA was significantly elevated in the FXT, PROE-M, and PROE-L group < 0.05). The expression of IDO1 and TNF-α mRNA in the cerebral cortex and duoden was significantly higher in the CUMS group compared to the NC group (p < 0.01), h ever, and the expression of IDO1 and TNF-α mRNA in the cerebral cortex and duoden was significantly lower in the PROE groups compared to the CUMS group (p < 0.05).

The Effect of PROEs on the Concentrations of SCFAs in Feces
As shown in Figure 12G-I, the concentrations of acetic acid, propionic acid, and tyric acid in feces were significantly lower in the CUMS group than in the NC group 0.01). Compared to the CUMS group, the concentrations of acetic acid and propionic a were significantly higher in the FXT, PROE-H, and PROE-L groups (p < 0.05), and the c centrations of butyric acid were significantly higher in the PROE-H group (p < 0.05).  As shown in Figure 12A-F, compared to the CUMS group, the expression of 5-HT 2A mRNA was significantly lower in the FXT and PROE groups (p < 0.01), and the expression of 5-HT 1A mRNA was significantly elevated in the FXT, PROE-M, and PROE-L groups (p < 0.05). The expression of IDO1 and TNF-α mRNA in the cerebral cortex and duodenum was significantly higher in the CUMS group compared to the NC group (p < 0.01), however, and the expression of IDO1 and TNF-α mRNA in the cerebral cortex and duodenum was significantly lower in the PROE groups compared to the CUMS group (p < 0.05).

The Effect of PROEs on the Concentrations of SCFAs in Feces
As shown in Figure 12G-I, the concentrations of acetic acid, propionic acid, and butyric acid in feces were significantly lower in the CUMS group than in the NC group (p < 0.01). Compared to the CUMS group, the concentrations of acetic acid and propionic acid were significantly higher in the FXT, PROE-H, and PROE-L groups (p < 0.05), and the concentrations of butyric acid were significantly higher in the PROE-H group (p < 0.05). acetic acid, propionic acid, and butyric acid in feces; data represent mean ± SEM (n = 6); ## p < 0.01 and ### p < 0.001 compared to the NC group; * p < 0.05, ** p < 0.01 and *** p < 0.001 compared to the CUMS group.

Discussion
Depression is a common and recurrent mood disorder illness. Although its pathogenesis is not completely clear, many studies have suggested that stress factors play a key role and that long-term stress is the main cause of depression. In this study, a depression rat model was established using solitary feeding combined with CUMS. CUMS rats were randomly given the stimulation of variable and unpredictable stress factors, which is similar to the hypothesized pathogenic process of human depression [23]. The body weights of the rats in the CUMS group decreased, and they showed depression-like behaviors such as loss of pleasure, increased demonstration of hopeless behavior, decreased autonomous activity, and decreased desire to explore new environments. Furthermore, compared to the NC group, the levels of DA, 5-HT, NE, DOPAC, and 5-HIAA in the hippocampi of the CUMS group were significantly lower, and the levels of CORT, ACTH, and CRF in serum propionic acid, and butyric acid in feces; data represent mean ± SEM (n = 6); ## p < 0.01 and ### p < 0.001 compared to the NC group; * p < 0.05, ** p < 0.01 and *** p < 0.001 compared to the CUMS group.

Discussion
Depression is a common and recurrent mood disorder illness. Although its pathogenesis is not completely clear, many studies have suggested that stress factors play a key role and that long-term stress is the main cause of depression. In this study, a depression rat model was established using solitary feeding combined with CUMS. CUMS rats were randomly given the stimulation of variable and unpredictable stress factors, which is similar to the hypothesized pathogenic process of human depression [23]. The body weights of the rats in the CUMS group decreased, and they showed depression-like behaviors such as loss of pleasure, increased demonstration of hopeless behavior, decreased autonomous activity, and decreased desire to explore new environments. Furthermore, compared to the NC group, the levels of DA, 5-HT, NE, DOPAC, and 5-HIAA in the hippocampi of the CUMS group were significantly lower, and the levels of CORT, ACTH, and CRF in serum were significantly higher. PROEs can evidently improve depression-like behaviors in model rats and the hyperfunction of the HPA axis, increasing the levels of monoamine neurotransmitters in the hippocampus, which indicates that PROEs may have an antidepressant effect.
In addition, we found that PROEs can regulate the gut microbiota of depressionmodel rats. In the analysis of gut microbiota structure, alpha diversity is a comprehensive indicator of richness and evenness that is commonly used to characterize the presence of different species in community ecology. Some previous studies have suggested that depression is associated with an increased richness and diversity of gut microbiota [24]. In this study, the difference of gut microbiota among different groups was directly observed by visualization methods in beta diversity analysis. This analysis showed that the structure of gut microbiota in the CUMS group was significantly changed but that PROEs made the structure of the gut microbiota closer to that of the NC group, which is consistent with the literature.
Studies have shown that there are significant changes in the species composition of gut microbiota in patients with depression and depression-model animals. At the phylum level, we found that PROEs significantly increased the abundance of Actinobacteria and Firmicutes and decreased that of Bacteroidetes, an important result given that the members of the phylum Actinobacteria might be potential biomarkers for ketamine's antidepressant efficacy [25,26]. At the family and genus levels, we found that the gut microbiota regulated by PROE were mainly related to the protection of the intestinal barrier, and the production of SCFAs and endotoxins. Researchers have reported that various harmful bacteria, toxins, and some inflammatory factors induced by gut microbiota disorder can alter the integrity and permeability of tight connections and destroy the function of the intestinal mucosa barrier in different ways [27]. Researchers have also found that the structure of intestinal microvilli and epithelial in CUMS depression-model rats is seriously damaged, that their intestinal glands are reduced or even gone completely, and that there is inflammatory cell infiltration in the lamina propria mucosa [28]. Tight junctions are an important structure of the intestinal mucosal barrier, and decreased tight junction protein occludin expression leads to intestinal mucosal barrier dysfunction and alters intestinal permeability [29]. SCFAs such as acetic acid, propionic acid, and butyric acid, which are important metabolites of gut microbiota, can directly affect the permeability of intestinal mucosa. SCFAs are an important energy source for epithelial cells, which can stimulate cell proliferation [30]. Butyric acid, for example, has been shown to reduce diarrhea caused by intestinal barrier dysfunction, such as inflammatory bowel disease, as well as to increase the levels of occludin and cingulin proteins in HeLa cells [31].
In this study, PROEs significantly increased the abundance of both Prevotellaceae and Peptostreptococcaceae, which can ferment carbohydrates in food to produce SCFAs. A decrease in the abundance of these organisms has been shown to be related to increased intestinal permeability [32,33] in depression-model rats. At the genus level, the abundance of Romboutsia, Roseburia, Lachnospiraceae_NK4A136_group, Prevotella_9, and Eubac-terium_coprostanoligenes_group was significantly lower in the CUMS group compared to the NC group, which has already been reported to be related to the maintenance of intestinal barrier function and the alleviation of inflammatory response. Romboutsia, Roseburia, Lach-nospiraceae_NK4A136_group, and Prevotella_9 can produce SCFAs [34][35][36][37], and the intestinal epithelial oxidative and inflammatory damage of mice rich in these four bacteria has been found to be lower than that of control mice [38]. Lachnospiraceae_NK4A136_group has also been found to be positively correlated with factors related to the maintenance of intestinal barrier integrity and negatively correlated with pro-inflammatory factors (LPS and IL-6) and neurotoxic quinolone [39,40]. In addition, the increased abundance of Romboutsia is associated with decreased levels of proinflammatory cytokines in plasma [34], and the lower the abundance of Eubacterium_coprostanoligenes_group, the more easily the intestinal barrier will be destroyed [41]. Our results also show that PROEs not only increased the abundance of beneficial bacteria but also reduced the relative abundance of harmful bacteria such as Bacteroides Oscillibacter, Parasutterella, and Intestinimonas in the depression-model rats. Studies have reported that the relative abundance of Oscillibacter and Parasutterella is higher in cases of major depressive and bipolar disorder [42,43]. Bacteroides are a kind of Gram-negative bacteria that increase the production of harmful metabolites such as amyloids, LPS, enterotoxins, and neurotoxins [44]. In addition, LPS and other bacterial endotoxins are more easily absorbed into the blood after the intestinal barrier function is damaged, and this can induce the release of peripheral and central proinflammatory cytokines. Proinflammatory cytokines play an important role in the pathogenesis of depression, and the levels of proinflammatory cytokines in depression patients are generally higher than those of a normal person [45].
In addition to mediating the microbiota, we also observed that PROEs relieved intestinal mucosal damage, increased the expression of occludin protein in the colon, adjusted the level of SCFAs in the feces and serum levels of LPS and IL-6, and reduced the expression of TNF-α m RNA in the cerebral cortex and duodenum. One study has reported that RP can inhibit the activation of NLRP3 inflammasome and the production of pro-inflammatory cytokines such as TNF-α in the prefrontal cortex of rats, thereby exerting antidepressant effects by promoting autophagy and inhibiting neuroinflammation [46]. Our results indicate that PROEs may regulate the metabolism of SCFAs and help to improve intestinal mucosal function, thereby attenuating the systemic inflammatory response by restructuring the gut microbiota.
The results of our KEGG pathway analysis showed that PROEs affected the tryptophan metabolic pathway, which plays an important role in the development of depression. The gut microbiota may affect neurotransmitters in the brain by regulating the metabolism of tryptophan and downstream metabolites such as kynurenic acid and quinolinic acid [47,48]. Studies have reported that mice that received fecal bacteria transplantation from chronic-stress mice had reduced adult hippocampal neurogenesis and damaged tryptophan metabolic pathways but that supplementing with 5-HTP, the direct precursor of 5-HT, alleviated depression-like behavior and restored hippocampal neurogenesis [49]. Tryptophan is an essential amino acid in the human body whose decomposition pathways include the 5-HT and kynurenine pathways. Indoleamine 2,3-dioxygenase-1 (IDO1) is the key rate-limiting enzyme of tryptophan metabolism and is distributed in the intestine and brain. When IDO1 increases or inflammation induces hyperactivation, more tryptophan becomes metabolized into kynurenine, leading to tryptophan depletion and insufficient synthesis of 5-HT, which affects the function of both the cerebral cortex and hippocampus, inducing depression-like behaviors such as loss of interest and insomnia [50]. The ratio of kynurenine to tryptophan is therefore considered to be a sensitive indicator of IDO1 activity and cellular immune status [51]. Tryptophan-kynurenine metabolism has even been reported to be directly or indirectly regulated by gut microbiota [52]. Butyric acid, the main metabolite of gut microbiota, inhibits IDO1 expression in intestinal epithelial cells [53]. The systemic inflammatory response caused by an imbalance of gut microbiota can induce the inflammatory inducible enzyme IDO1, which affects the metabolism of tryptophan-kynurenine [54]. In addition, 5-HT 1A and 5-HT 2A receptors in the 5-HT system are most closely related to emotional disorders [55,56]. The 5-HT 1A receptor is the key target of 5-HT antianxiety drugs, which can inhibit anxiety-like behavior after activation [57,58]. The 5-HT 2A receptor antagonists have antianxiety effects, and 5-HT 2A knockout mice exhibited attenuated depression-like behavior [59,60]. In this study, PROEs restored the levels of tryptophan and kynurenine in plasma, increased the expression of 5-HT 1A mRNA in the cortex, reduced the expression of 5-HT 2A mRNA in the cortex, and reduced the expression of IDO1 mRNA in the cortex and duodenum of depression-model rats.
Gut microbiota are also closely related to the function of the HPA axis and neurotransmitter system. Hyperfunction of the HPA axis leads to increased secretion of cortisol, increased permeability of intestinal mucosa, changes in the composition of gut microbiota, and increased levels of LPS in serum, though transplantation of fecal bacteria from normal animals into depression-model animals can improve the HPA axis function of the recipients [61]. In addition, the gut microbiota can directly synthesize neurotransmitters such as 5-HT, NE, and DA, which may affect the function of central neurotransmitters through the circulatory system and vagus nerve [27,62]. Our Spearman correlation analysis showed that the levels of ACTH, CORT, and CRF in the HPA axis were negatively correlated with the relative abundance of gut microbiota protective bacteria and SCFAproducing bacteria (Romboutsia, Roseburia, and Prevotella_9) but were positively correlated with LPS-producing bacteria (Bacteroides). Additionally, the levels of 5-HT, DA, NE, and other neurotransmitters in the brain were positively correlated with the relative abundance of gut microbiota protective bacteria and SCFA-producing bacteria but were negatively correlated with LPS-producing bacteria. These results indicate that the improvements to the HPA axis and neurotransmitter function from PROEs may be related to their regulation of gut microbiota. However, the relationship between PROEs that antagonize the HPA axis and thereby enhance the function of neurotransmitter and gut microbiota still needs further study.

PROEs Preparation
The PR was refluxed three times with 8-fold 60% EtOH for 1.5 h, and then the solvent was removed to obtain the extracts of PR. These extracts further underwent column chromatography on MCI GEL CHP20P resin eluted with 20% EtOH, 50% EtOH, and 90% EtOH. The fraction of 50% EtOH was then gathered and concentrated to obtain the PROEs, resulting in an 8.4% extraction rate.

PROEs Ingredient Identification
A Waters Acquity UPLC system, coupled with a Q-TOF SYNAPT G2-Si high-definition mass spectrometer (Waters, Milford, MA, USA), was used to carry out LC-MS. First, the PROEs were separated on an Acquity UPLC BEH C18 column (1.7 µm, 2.1 × 100 mm) that was kept at 40 • C and at a flow rate of 0.4 mL/min. Acetonitrile (A) and 0.1% aqueous formic acid (B) were used as the mobile phase, and the gradient programs were as follows: 5-12% A (0-3 min), 12-20% A (3-6 min), 20-30% A (6-10 min), 30-35% A (10-10.5 min), 35-45% A (10.5-15 min), 45-50%A (15-18 min), and 50-95%A (18-20 min). The mass spectrometry instrument was equipped with an ESI-Ion source and was used to detect the spectrum in the m/z 50-2000 range with MS n as the scan mode in negative mode. Here, the ion source temperature was 100 • C, and the desolvation gas temperature was 500 • C, and the flow rates of cone and desolvation gas were set at 50 L/h and 800 L/h, respectively. The voltages of the capillary and cone in negative ion mode were set at 2.5 kV and 40 V, respectively, and leucine enkephalin (m/z 554.2615 in negative ion mode) was used as a reference mass.

Animals and Treatments
All the animal studies in this paper were authorized by the Experimental Animal Management Committee of Capital Medical University (No: AEEI-2016-023). The 230 ± 20 g Sprague-Dawley rats (male, n = 56) were supplied by Beijing Vital River Laboratory Animal Technology Co., Ltd. (Beijing, China, SCXK (jing) 2016-0006). The rats were housed in a standard SPF environment at 24 ± 1 • C and 60 ± 5% humidity where a standard diet and water were freely available.
After acclimatization for one week, the rats were divided randomly into seven groups (each group = 8): normal control (NC), normal control + PROE (NC-PROE, 84 mg/kg), fluoxetine (FXT, 2 mg/kg), CUMS, CUMS + high-dose PROE (PROE-H, 126 mg/kg), CUMS + medium-dose PROE (PROE-M, 84 mg/kg), and CUMS + low-dose PROE (PROE-L, 42 mg/kg). The NC and NC-PROE groups were placed in a separate undisturbed area (three rats/cage), and the other groups (one rat/cage) were subjected to CUMS stressors according to methods previously described in the literature [63]. There were 9 stressors, namely, 4 • C cold-water swimming (5 min), food deprivation (24 h), water deprivation (24 h), tail clipping (1 min), flash stimulation (150 flashes per min, 24 h), white-noise exposure (24 h), binding (2 h), continuous illumination (24 h), and damp bedding (200 mL of water was put into 100 g of sawdust bedding for 24 h). One or two kinds of stimulation were randomly arranged every day for eight weeks. The NC and CUMS groups were given 5 mL/kg of distilled water, and the other groups were given corresponding drugs according to their body weights via intragastric administration for eight weeks, once a day. The weight of each rat was measured every two weeks.

Behavioral Tests
The rats' behavioral characteristics were measured using the sucrose preference test (SPT), open-field test (OFT), and forced swimming test (FST), as determined according to methods described in the references [63].

ELISA Measurement
At the end of behavioral tests during the eighth week after the drug administrations, rats were anesthetized via 0.56% sodium pentobarbital, and blood was collected from the abdominal aorta for serum and plasma separation. Subsequently, the brain of each rat was immediately dissected on an ice plate and stored at −80 • C. The serum and hippocampus were then thawed at 4 • C, and the concentrations of CORT, CRF, ACTH, IL-6, and LPS in the serum were measured using ELISA kits following the manufacturer's instructions.

Detection of Monoamine Neurotransmitters and Their Metabolites
In order to detect monamine neurotransmitters and their metabolites, the hippocampus from each rat was weighed and quickly homogenized in 120 µL of pretreatment solution A (0.4 mol/L of perchloric acid) on ice prior to being centrifuged at 12,000 rpm/min for 20 min at 4 • C after standing at room temperature for 30 min. Next, 90 µL of supernatant was collected, and 45 µL of pretreatment solution B (20 mmol/L of potassium citrate, 0.3 mol/L of dipotassium hydrogen phosphate, and 2 mmol/L of EDTA·2Na) was added. The sample was next vortexed and then centrifuged at 12,000 rpm/min for 20 min at 4 • C after also standing at room temperature for 30 min. The supernatants were collected and analyzed by high-performance liquid chromatography (HPLC) coupled with an electrochemical detector (Waters ECD2465, Milford, MA, USA) [64]. Here, the chromatographic column Waters symmetry shield RP 18 (150 × 3.9 mm, 5 µm, Waters Atlantis) was maintained at 30 • C, and the flow rate was 0.8 mL/min. The mobile phase (methanol-water, 8:92, v/v) was mixed with sodium acetate, 1-octanesulfonate, citric acid, and EDTA•2Na, and the detector potential was +0.6 V, with an injection volume of 40 µL.

Detection of Tryptophan and Kynurenine in Plasma
To detect the tryptophan and kynurenine, the rat plasma (100 µL) was mixed with 5% perchloric acid (100 µL) by vortex for 5 s in a centrifuge tube then centrifuged at 14,000 rpm/min for 10 min. The supernatants were then collected and analyzed by HPLC coupled with a UV detector (1200, Agilent, Santa Clara, CA, USA). The chromatographic column Diamonsil C18 (250 mm × 4.6 mm, 5 µm, DIKMA, Beijing, China) was maintained at 30 • C; the mobile phase A was 15 mmol/L sodium acetate buffer (pH 4.0), and mobile phase B was acetonitrile (92:8, v/v). The flow rate was 1.0 mL/min, with an injection volume of 50 µL, and the UV detection wavelengths of tryptophan and kynurenine were 280 and 360 nm, respectively.

Determination of Fecal Short-Chain Fatty Acids
The fecal concentrations of short-chain fatty acids (SCFAs), including acetic acid, propionic acid, and butyric acid, were analyzed by GC-MS (7000C, Agilent, Santa Clara, CA, USA), along with an HP-INNOWAX column (30 m × 0.32 mm × 0.25 µm, Agilent, Santa Clara, CA, USA). Standard solutions of acetic acid, propionic acid, and butyric acid were prepared at 500, 200, 100, 50, 20, 10, 5, 2, 1, and 0.5 µg/mL, respectively, and the internal standard substance of 4-methylvaleric acid was prepared at 10 µg/mL. Each fecal sample (80.0 mg) was soaked in 0.9 mL 0.5% phosphoric acid solution by vortexing and then centrifuging at 12,000 rpm/min for 5 min at 4 • C. Afterward, 700 µL liquid of supernatant was extracted with 800 µL of ethyl acetate and centrifuged at 12,000 rpm/min for 10 min at 4 • C. Next, 15 µL internal standard solution was added to 585 µL liquid of supernatant, which was vortexed for 15 sec and centrifuged at 12,000 rpm/min for 5 min. Finally, the supernatants were collected and measured by GC-MS.

Analysis of Intestinal Morphology
To analyze intestinal morphology, segments of the duodenum and colon from three rats in each group were fixed in freshly prepared 4% paraformaldehyde solution. These tissues were then dehydrated in a series of increasingly concentrated ethanol solutions, hyalinized in xylene, and embedded in paraffin wax. The samples were next cut into 5 µm sections and stained with hematoxylin and eosin (H&E). Finally, the sections were examined under a Panoramic SCAN (3DHistech, Budapest, Hungary), and the histopathology of duodenum and colon were analyzed with Image-Pro-Plus 6.0 image analyzer software (Media Cybernetics Inc., Rockville, MD, USA).
After the above analysis, segments of the duodenum and colon were collected and sectioned into pieces about 1 mm × 1 mm × 2 mm, fixed in 4 • C 2.5% glutaraldehyde buffer for 2 h, and washed with 0.1 M PB buffer 3 times for 15 min each time. These tissues were then fixed with 4 • C 1% osmic acid for 1 h, dehydrated using an ethanol gradient, embedded in epoxy resin, and sliced with an ultra-thin slicing machine at a thickness of 50 to 70 nm. Ultrastructural changes of the duodenum and colon epithelium were then observed under a transmission electron microscope (HT7700, HITACHI, Tokyo, Japan).

Western Blot Analysis
For Western blot analysis, colon tissues were cut into pieces and lysed in RIPA buffer with a protein phosphatase inhibitor, and the total proteins were extracted according to the manufacturer's protocols. Then, the protein c concentration was detected by a BCA protein assay kit. Here, 50 µg of total protein was separated by 10% SDS-polyacrylamide gel electrophoresis and then transferred onto a 0.22 µm polyvinylidene fluoride membrane. After washing four times in 1× TBS buffer, the membranes were then blocked with 1× TBST containing 5% nonfat milk for 1 h at room temperature. Next, the membranes were incubated with primary antibodies overnight at 4 • C. The primary antibodies and their dilution concentrations were occludin (1:1000) and β-actin (1:1000). Subsequently, the membranes were washed with 1× TBST and incubated with secondary antibodies of goat anti-mouse or anti-rabbit IgG (1:20,000) for 1 h at room temperature. After the membranes were washed again using 1× TBST buffer, the proteins were visualized using enhanced chemiluminescence.

Quantitative Reverse-Transcription Polymerase Chain Reaction
Total RNA from the cerebral cortex and duodenum samples was extracted using RNA extraction reagent, and the RNA concentration was measured using the NanoDrop 2000 UV-vis spectrophotometer (Thermo, Waltham, MA, USA). The RNA was subsequently reverse transcribed to cDNA using a Servicebio ® RT First Strand cDNA Synthesis Kit according to the manufacturer's instructions. Amplification and quantitative detection were then performed in a real-time PCR machine (Bio-Rad, Hercules, CA, USA), where the RT-PCR protocol was as follows: initial denaturation at 95 • C for 10 min, then 40 cycles at 95 • C for 30 s and 60 • C for 30 s, and the 2 −∆∆Ct method was used for relative quantitative analysis of the results. Primer sequences for the genes of interest are listed in Table 2. 4.13. DNA Extraction and 16S rRNA Gene Sequencing of Fecal Samples DNA from fecal samples was extracted using a QIAamp DNA stool mini-kit. The V3-V4 region of the bacterial 16S rRNA gene was then PCR amplified with primers 338F and 806R, and the sequencing library was prepared by a Truseq nano DNALT library prep kit (Illumina, San Diego, CA, USA). The effective sequences were merged and divided into operational taxonomic units (OTUs) with a 97% similarity cutoff via QIIME 2 software (version 2019.4), and the representative sequences of OTUs were then compared to the template sequences in the Greengenes database (Release 13.8) to be analyzed. Next, the fecal gut microbiome was analyzed by taxonomic composition analysis, alpha and beta diversity analysis, redundancy analysis, heatmap analysis, linear discriminant analysis of effect size, KEGG pathways, and PICRUSt analysis.

Statistical Analysis
GraphPad Prism 8.0.1, QIIME 2, and R software (v3.2.0) were used for all statistical analysis, and prior to this analysis, all data were expressed as mean ± standard error of the mean (SEM). One-way ANOVA was used for comparison between multiple groups, two-way ANOVA was used for changes in rat bodyweight, and the abundance of gut microbiota was compared between multiple groups using the Kruskal-Wallis rank-sum test.

Conclusions
PROEs play an antidepressant role by regulating the diversity and structure of gut microbiota, restoring intestinal barrier function, reducing the release of intestinal endotoxin, and lowering the level of inflammation. In addition, PROEs may also affect the level of neurotransmitters in the brain by regulating the tryptophan-kynurenine metabolic pathway to improve depression symptoms. However, further research is still necessary to study whether the antidepressant effects of PROEs depend on the existence of gut microbiota in combination with fecal bacteria transplantation in germ-free animals. Although the underlying effects of PROE on gut microbiota remain elusive, our present findings provide new insight and approaches for understanding the systemic mechanisms of the antidepressant effects of natural medicines and other products with good efficacy, poor absorption, and unclear mechanisms of action.
Supplementary Materials: The supporting information can be downloaded at https://www.mdpi. com/article/10.3390/ijms241813877/s1. Author Contributions: Y.B. and X.W. conceived the study; Y.B., Q.C. and T.J. prepared the animals, samples, conducted various detection experiments, analyzed and interpreted the data, and wrote the manuscript. X.C., J.W. and X.W. contributed to revising the manuscript. All authors have read and agreed to the published version of the manuscript.