Chnghui Du, Zhenjie Li, Jing Zhng, Ni Yin, Lirong Tng, Jie Li, Jingyin Sun,Xioqing Yu, Wei Chen, Hng Xio, Xin Wu*, Xuexing Chen,*

a School of Public Health, Guangzhou Medical University, Guangzhou 510642, China

b Department of Information Systems and Analytics, Miami University, Oxford 45056, USA

c Department of Food Science, University of Massachusetts, Amherst 01003, USA

d Department of Kinesiology, Nutrition, and Health, Miami University, Oxford 45056, USA

Keywords:Carnosic acid Inf lammatory bowel diseases Colitis Gut microbiota Metabolites

A B S T R A C T Accumulating evidence suggests that the gut microbiota plays an important role in the pathogenesis of inflammatory bowel disease (IBD). Carnosic acid (CA) is a major antioxidant component of rosemary and sage. Herein, we investigated the protective effects of dietary CA on dextran sodium sulfate (DSS)-induced colitis mouse model with an emphasis on its impact on the composition and metabolic function of gut microbiota. We found that CA effectively attenuated DSS-stimulated colitis in mice, as evidenced by reduced disease activity index (DAI), and systemic and colonic inflammation. Additionally, CA restored microbial diversity and improved the composition of gut microbiota in DSS-treated mice. Moreover,Spearman’s correlation coeff icient showed a signif icant correlation between the fecal metabolites and the gut microbiota species. Changes in gut microbiota and the correlated metabolites might partially explain CA’s anti-inf lammatory effects against colitis. Future clinical trials are needed to determine the therapeutic effects and mechanisms of CA on IBD in humans.

1. Introduction

Inf lammatory bowel diseases (IBD), including relapsing-remitting ulcerative colitis and Crohn’s disease, is a non-specific chronic intestinal inf lammatory disease [1]. The precise pathophysiology of IBD remains unclear, and some possible inf luencing factors include persistent intestinal infection, intestinal mucosal barrier defects,immune disorders, genetic factors and environmental factors [2,3].The involvement of the gut microbiota in the pathophysiology of IBD has recently been highlighted [4]. At present, there have been a number of reports on the relationship between the gut microbiota and IBD. Studies have shown that the composition and structure of the gut microbiota of IBD patients have undergone major changes,comparing with that of the healthy controls [4]. Specifically, the diversity of gut microbiota in patients with ulcerative colitis was reduced, and the composition was also signif icantly altered, e.g., the abundance of Firmicutes andClostridiumdecreased significantly,and the abundance of Proteobacteria increased significantly [5,6].The number of Bifidobacteria in the intestine was significantly lower, whileProteuswas significantly higher in IBD patients, as compared with the healthy controls [7,8]. The intestinal bacterial ecology of IBD patients was disordered, and the risk of infection withClostridium diff icilewas thus greatly elevated [9]. The gut microbiota plays an important role in protecting and maintaining the normal metabolism and physiological functions of the human body [10].Studies have shown that IBD susceptible genes might be associated with the recognition and processing of microbes [11]. Therefore, the gut microbiota has become an important target for the prevention and treatment of IBD.

Metabolomics techniques are used to comprehensively characterize the changes of small molecule metabolites during biological metabolism by exploiting cutting-edge analytical platforms [12]. Metabolomic profiling has been employed to study the influence of diseases on the human metabolic spectrum, interpret the process of human diseases, and search for disease-related genes and biomarkers [13]. Moreover, some gut microbiota families, such as Verrucomicrobiaceae and Bacteroidaceae, significantly correlate with alterations in fecal metabolites. Whether an improved gut microenvironment could prevent IBD-related complications and delay the progress of IBD is a propitious objective for future study.

Carnosic acid (CA) is a phenolic diterpenoid found inSalvia officinalisL. (sage) andRosmarinus officinalisL. (rosemary), both of which are commonly used in culinary herbs and spices [14]. CA is heat-resistant and shows potent antioxidant activities [15,16].CA has also been reported to possess biological activities such as antibacterial [17,18], anti-tumor [19], anti-inflammatory [20], liver protection [21] and neuroprotection [22], suggesting it could be used as a preventive and/or therapeutic agent for human diseases.Recently, CA has been reported as a potential treatment for IBD in DSS-induced colitis mouse model [23]. The anti-colitis effects of CA and CA-rich rosemary extract are partially through the regulation of the Keap1/Nrf2 pathway and the improvement in the intestinal barrier integrity [24,25]. Based on prior findings, we hypothesized that CA would prevent the development of colitis by improving the composition (by 16S rRNA sequencing) and metabolic function (by untargeted metabolomics) of the gut microbiota. Our study aims to provide more insights into the efficacy and mechanism of action of CA in ameliorating IBDin vivo.

2. Materials and methods

2.1 Animals and experimental design

Ninety male ICR mice (4-5 weeks old) were purchased from GemPharmatech (Nanjing, China), and raised in the Laboratory Animal Center of Guangzhou Medical University. The protocol was approved by the Institutional Animal Care and Use Committee of Guangzhou Medical University (S2020-032). Animals were randomly assigned to 6 groups (15 animals/group) and housed at 5 mice per cage in a temperature-controlled environment of (20 ± 4) °C with 35%-55% relative humidity and fixed 12 h light/dark cycle.After 2-week of acclimation, the 6 groups, each sizedn= 15, were designated as follows: the control group (DSS-/CA-); the model group (DSS+/CA-) treated with 1.5% DSS water (salt form); the CA group (DSS-/CA+) supplemented with 50 mg/kg CA daily by oral gavage (≥ 95% purity); the Mod-CA group (DSS+/CA+)received different doses of CA daily by oral gavage (50, 100 and 200 mg/kg) and 1.5% DSS drinking water. CA (purity ≥ 95%)was purchased from Hunan Xianwei Sunshine Biotechnology Co.,Ltd. (Luxi, China). DSS was added to sterilized drinking water and offeredad libitumfor 4 days, followed by 7 days of tap water for recovery. This cycle was repeated 4 times. At the end of the 4thcycle,all mice were sacrificed with CO2asphyxiation. The spleen was quickly excised and weighted. The colon from individual mouse was dissected, weighted and measured, and cut open longitudinally. Fecal samples were collected from the colon, immediately snap-frozen, and then stored at -80 °C for later analyses.

2.2 Disease activity index (DAI) and histological assessment

DAI was scored based on the extent of rectal bleeding: 0 (none),1 (hemoccult+), 2 (visible blood traces in stool), 3 (gross bleeding);stool consistency: 0 (normal), 1 (soft but still formed), 2 (soft),3 (diarrhea); and weight loss: 0 (0%), 1 (1%-5%), 2 (5%-10%),3 (10%-20%), 4 (＞ 20%) [26]. The final macroscopic score for each animal is the sum of these three separate scores. Formalin-fixed colonic tissues were processed for paraffin embedding, sectioning, and hematoxylin and eosin (H&E) staining as we previously described.Based on H&E staining, histological alterations such as mucosal ulceration, dysplasia, and infiltration of immune cells were evaluated under a microscope.

2.3 Detection of inflammatory factors in serum and feces

For fecal calprotectin measures, frozen fecal samples obtained in the colon upon euthanasia were lysed in ice-cold PBS containing 1% bovine serum albumin, 0.05% Tween 20 and protease inhibitor.Protein concentrations were determined using the BCA method.Fecal calprotectin levels were measured by Mouse S100A8/S100A9 Heterodimer DuoSet ELISA according to manufacturer’s instructions(R&D Systems, Minneapolis, MN, USA) and normalized by protein concentrations. The mice blood was obtained by eyeballs extraction,and was left standing at room temperature for 30 min, centrifuged at 4 °C for 10 min at 3 500 r/min. The supernatant was taken, and the serum IL-6 content in each group was determined by ELISA according to manufacturer’s instructions (R&D Systems).

2.4 Fecal DNA extraction, 16S rRNA analysis and Illumina MiSeq sequencing

Fecal DNA extraction and 16S rRNA analysis were performed according to the references [27-29]. Five mice from each group were randomly selected for 16S rRNA sequencing. Briefly, fecal bacterial DNA was extracted using the E.Z.N.A.?Soil DNA Kit(Omega Bio-tek, Norcross, GA, USA) following manufacturer’s protocol. The DNA was quantified, and quality was checked by NanoDrop 2000 spectrophotometer (Thermo Fisher Scientific,Wilmington, DE, USA). The V3-V4 region of the 16S rRNA gene was targeted for amplification by PCR using upstream primer 338F ACTCCTACGGGAGGCAGCAG and downstream primer 806R GGACTACHVGGGTWTCTAAT. PCR products were then purified with AxyPrep DNA Gel Extraction Kit (Axygen Biosciences, Union City, CA, USA), and the PCR products were detected and quantified by Quantus? Fluorometer (Promega, Madison, WI, USA). After combining the amplicon library, the samples were sequenced by Illumina’s Miseq PE300/NovaSeq PE250 platform (San Diego, CA,USA). Data optimization, OTU clustering, and taxonomy analysis were completed by the MG Sunshine Cloud Analysis Platform(Shanghai Majorbio Bio-pharm Technology, Shanghai, China).

2.5 Untargeted metabolomics

Untargeted metabolomics was performed according to the references [30,31]. Three mice from each group were randomly selected for untargeted metabolomics analysis. Briefly, mouse fecal samples were analyzed using the LC-MS platform (Thermo,210 Ultimate 3000LC, Q Exactive). LC-MS was performed on a Waters UPLC-class system equipped with a binary solvent delivery manager and a sample manager coupled with a Waters VION IMS Q-TOF mass spectrometer equipped with an electrospray ionization source operating in either positive or negative ion mode (Waters,Milford, USA). Source and desolvation temperatures were set to 120 and 500 °C, respectively, with a desolvation gas flow of 900 L/h. Centroid data were collected fromm/z50 tom/z1 000at a scan time of 0.1 s and an interscan delay of 0.02 s over a 13 min analysis time. Before pattern recognition, the original data were subjected to data preprocessing via baseline filtering, peak identification, integration, retention time correction, peak alignment,and normalization by using the instrument’s own metabolomics processing software Progenesis QI (Waters). Positive and negative data were imported into the SIMCA-P software package. The metabolites were compared by orthogonal partial least squares discriminant analysis (OPLS-DA).

2.6 Statistical analysis

Data are presented as means ± standard error (SE) or standard deviation (SD) for the indicated number of independently performed experiments. The Origin 2018 software (OriginLab, Northampton,MA, USA) was used for graphing. All statistical analyses were performed using SPSS 26.0 software (SPSS Inc., Chicago, IL,USA). The statistical significance of differences among groups was performed using one-way ANOVA followed by one-way nonparametric ANOVA Kruskal-Wallis test.P＜ 0.05 is considered a significant difference, andP＜ 0.01 is considered a very significant difference.

Mass spectra of metabolites were identified by using the accurate mass, MS/MS fragments spectra and isotope ratio difference with searching in reliable biochemical databases as Human metabolome database (HMDB, http://www.hmdb.ca/) and Metlin database (https://metlin.scripps.edu/). The selection of metabolites with significant differences depends on the variable weight value (VIP) obtained by PLS-DA model and thePvalue of Student’st-test. Metabolites with VIP ＞ 1.2,P＜ 0.05 were labeled as significantly different metabolites.The pathways involved in differential metabolites were obtained through functional annotation of metabolic pathways conducted in KEGG database (https://www.kegg.jp/kegg/pathway.html). Pathway enrichment analysis was performed through the Python software package scipy.stats (https://docs.scipy.org/doc/scipy/). Fisher’s exact test was used to identify the biological pathways most relevant to the experimental treatment.

3. Results and discussion

3.1 CA alleviated the symptoms of colitis in DSS-treated mice

An acute toxicity study found that the LD50of CA was 7 100 mg/kg(95% CI, 6 060-8 940 mg/kg) in mice [32]. In the sub-chronic toxicity study, no difference in hematology and clinical chemistry parameters was seen between the control rats and the rats fed 150 mg/(kg·day)of CA, which is equal to ~300 mg/(kg·day) in mice [33]. Heart and liver damage were observed only with higher dose supplementation of CA in rats (600 mg/(kg·day), which is equal to ~1 200 mg/(kg·day)in mice) [32]. In the present study, mice received dosages of CA well below the reported LD50, and no apparent behavioral or appearance difference was observed with any CA supplemented groups,suggesting that the dosages of CA used in this study were generally safe and well-tolerated in mice.

Body weight loss is a characteristic hallmark of the DSS-induced model of colitis [34]. Nonetheless, we did not observe a downward trend in body weight in DSS-treated mice in this study (Fig. 1A).In rodent models, 1.5%-5% of DSS water with recovery periods in between DSS cycles is widely used to induce colitis [35]. In the present study, we employed four 4-day cycles of a low dose of DSS(1.5%) with 7-day recovery periods in between, which might explain the non-significant changes in body weight. Interestingly, mice received 100 and 200 mg/kg CA had a lower average body weight compared to other groups, but the cause was not evident, which requires further investigation.

Fig. 1 (Continued)

Fig. 1 Dietary CA attenuated DSS-induced colitis in mice. (A) Body weight during the entire experiment; (B) DAI; (C) Spleen weight and spleen coefficient;(D) Colon length. (E) Representative images of H&E stained colon: yellow arrows represent inflammatory cell infiltration, red arrows represent connective tissue proliferation, black arrows represent exfoliation of colonic epithelial cells, and blue arrows represent atrophy of muscle layer muscle cells. Data were presented as mean ± SE. Significant differences were indicated as *P ＜ 0.05, **P ＜ 0.01 vs Con; #P ＜ 0.05, and ##P ＜ 0.01 vs Mod.

The DAI score is commonly used to evaluate the severity of colitis in mice, which consist of the fecal consistency, hematochezia,and body weight loss [36]. As shown in Fig. 1B, the DAI score rated as high as 4.14 in the model group (Mod), which was drastically higher than the control group (Con) (P＜ 0.01). Supplementation of CA significantly reduced the DAI compared to the model group,although no dose-dependent manner was observed among the three CA-supplemented groups. Clinically, splenomegaly (enlarged spleen)is common in patients with colitis [37]. The organ coefficient is defined as the ratio of organ weight to body weight. An increase in organ coefficient indicates that the organ may have inflammatory lesions such as hyperemia, edema, and hypertrophy [37]. As shown in Fig. 1C, spleen weight and coefficient of the model group were both significantly higher compared to the control group (P＜ 0.01). Both metrics showed a downward trend after dietary CA intervention, and supplementation of CA at 200 mg/kg reached statistical significance(P＜ 0.05). Colon length is another metric correlated with the progression of intestinal inflammation [38]. As shown in Fig. 1D, DSS significantly diminished the length of the colon by 20%, as compared to the control group (P＜ 0.05), while the shortening of the colon was effectively prevented by 200 mg/kg of CA supplementation (P＜ 0.05).

Histopathology result (Fig. 1E) showed that the control group had no apparent inflammatory changes, and the mucosal epithelium was relatively intact; whereas the model group exhibited distortion of colonic crypts, loss of goblet cells, severe epithelial injury,and inflammatory cell infiltration in the mucosa and submucosa.Supplementation of CA showed a tendency to alleviate these changes caused by DSS treatment. Collectively, these observations indicate that CA effectively suppressed DSS-induced colitis in mice.

Fig. 2 Dietary CA inhibited DSS-induced colonic inflammation as evidenced by fecal calprotectin concentration (A) and systemic inflammation as evidenced by serum IL-6 (B). Data were presented as mean ± SE. Significant differences were indicated as *P ＜ 0.05; **P ＜ 0.01.

3.2 CA attenuated DSS-induced intestinal and systemic inflammation in mice

In inflammatory diseases, especially ulcerative colitis,calprotectin, a neutrophil-driven calcium-binding protein, is usually overexpressed [39]. Therefore, fecal level of calprotectin has long been used as a biomarker of intestinal inflammation in clinical and pre-clinical settings [40,41]. As shown in Fig. 2A, DSS treatment caused a 1.68-fold increase in fecal calprotectin content as compared to the control group (P＜ 0.05), whereas supplementation of CA at 50-200 mg/kg led to 12%-36% inhibition in comparison to the model group (P＜ 0.05). The serum level of IL-6 of mice in the model group was also significantly elevated compared to the control group(105.9 vs. 78.9 pg/mL,P＜ 0.05). CA supplementations significantly reduced the elevated serum IL-6 level to 85.0-87.2 pg/mL (P＜ 0.05).Serum level of IL-6 rises concurrently with elevating inflammatory responses and is strongly associated with the development of intestinal inflammation in IBD [42]. Altogether, these findings further confirm that CA is capable of suppressing DSS-induced intestinal and systemic inflammation.

3.3 Effects of CA on DSS-induced gut microbiota dysbiosis

3.3.1 Dietary CA increased gut microbiota diversity in mice

Next, we sought to determine whether CA modulated the structure and composition of the gut microbiota in mice by 16S rRNA sequence analysis. Most clinical studies have shown that the diversity of the gut microbiota of patients with IBD was reduced [4,43]. Although there were no significant differences between the control and model groups with respect to theα-diversity of the microbiome as indicated by the Simpson and coverage indices (Fig. 3A), supplementation of CA resulted in a modest increase inα-diversity, in comparison to the model group, suggesting that dietary CA is capable of boosting the species abundance, richness, and diversity of gut microbiota regardless of disease state. Next, the relative similarity of gut microbiota was visualized by PCoA followingβ-diversity analysis(Fig. 3B). An apparent separation between control, model and CA-supplemented groups was observed, which suggests that DSS treatment shifted the structure of the microbial community, while CA intervention partially reversed the alteration. Overall, these results suggests that the observed anti-colitis effects of CA were elicited, at least partially, through its modulatory effects on the structure of gut microbiota.

3.3.2 Dietary CA modulated the composition of gut microbiota

To further understand the impact of dietary CA on the composition of gut microbiota, we examined the relative abundance of the predominant taxa within and among groups. At the phylum level, Bacteroidota and Firmicutes were the dominant bacteria(Fig. 3C). The phyla of the gut microbiota in each group were roughly the same, but the relative ratio was very different. In the model group,the proportion of Bacteroidetes drastically increased and the proportion of Firmicutes greatly decreased, which was consistent with prior findings [44]. A decreased ratio of Firmicutes/Bacteroidetes (F/B)is a key indicator of microbiota imbalance and has been used for evaluating the degree of inflammation in IBD [45]. Compared with the control group, the F/B ratio (Fig. 3C) of the model group was significantly decreased (P＜ 0.05), whereas the F/B ratio of the CA intervention groups was rebounded to varying degrees, and elevation produced by 200 mg/kg of CA reached statistical significance (P＜ 0.05).

At the family level, the comparison of the fungi families with a relative abundance greater than 1% was shown in Fig. 3D. The relative abundance of Muribaculaceae in the model group was the highest,and the proportion of this microbiota in the CA intervention group was lower than that in the model control group. A study found that the abundance of Muribaculaceae was correlated with inflammationassociated parameters, suggesting that its elevation might contribute to colitis development [46]. In addition, DSS treatment significantly reduced the relative abundance of Lactobacillaceae compared with the control group. The proportion of Lactobacillaceae in the control,model, CA control, 50, 100, and 200 mg/kg CA intervention groups were 15.8%, 10.8%, 14.7%, 7.34%, 5.52%, 17.0%, respectively.Supplementation of CA significantly enhanced the relative abundance of Lactobacillaceae, whereas highest concentration of CA markedly raised that by 1.6-fold compared with the model group. Interestingly,there was almost no Bifidobacteriaceae detected in the control and model groups; however, dietary CA substantially boosted the proportion of Bifidobacteriaceae. These probiotics might help to rebalance the gut microbiota in an anti-inflammatory direction [47].

At the genus level (Fig. 3E), compared with the control group,the proportions of Rikenellaceae_RC9_gut_group,Colidextribacter,Odoribacter, andEnterorhabdusin the model control group significantly increased, and the proportions ofBacteroides,Alloprevotella, andLactobacillusdecreased. Compared with the model group, the proportions of Lachnospiraceae NK4A136 group,Odoribacter, andAlloprevotellain CA intervention groups significantly reduced, and the proportions ofBacteroides,Rikenellaceae_RC9_gut_group,Lactobacillus,Odoribacter,Akkermansia, andBifidobacteriumincreased, and these changes were generally in a dose-response manner.BifidobacteriumandLactobacillusare common beneficial bacteria in the intestines, which play an important role in maintaining the body’s micro-ecological balance [48]. The proportion ofLactobacillusin the control, model,CA control, 50, 100, and 200 mg/kg CA intervention groups were 13.1%, 1.01%, 14.1%, 1.35%, 0.50%, 4.82%, respectively. Similar to the results at the family level, there was almost noBifidobacteriumin the control and model groups, but dietary CA at 100 and 200 mg/kg markedly promoted the proportion ofBifidobacterium.A decrease inLactobacilluswas observed in DSS-treated mice,compared with healthy mice [49]. Transplant ofLactobacillusto DSS-treated mice could reduce inflammation, reverse colon shortening and reduction of colon stem cells, thereby improving the symptoms and pathology of IBD mice [50]. In addition, IBD patients after treatment had significantly higher levels ofBifidobacterium[51].The growth ofBifidobacteriumwas stimulated in IBD rats treated with anti-colitis drug [52]. As intestinal probiotics,LactobacillusandBifidobacteriumcan restore the balance of gut microbiota, regulate the intestinal immune system, protect the intestinal barrier, thus preventing and improving IBD [53].

Linear discriminant analysis effective size (LEfSe) analysis is comparative method used to identify biomarkers that may explain differentiating phenotypes [54]. LEfSe results show that significantly different species appeared among the 6 groups (Fig. 3F). Thereinto,the differentiating species at the genus level of the control group wereRikenellaandEubacterium_ventriosum_group; the differentiating species of the model group wereFlavonifractorandEubacterium_siraeum_group; the differentiating species of the CA control group wereOdoribacterandNegativibacillus; the differentiating species of the low-dose CA group were Rikenellaceae_RC9_gut_group andDesulfovibrio; the differentiating species of the medium-dose CA group wereAkkermansia,Turicibacter,Christensenellaceae_R_7_ group andParvibacter; and the differentiating species of the high-dose CA group wereEubacterium_fissicatena_group and unclassified_f_Prevotellaceae. Studies has found that the intestinal microecological balance of UC patients was disturbed, the number of beneficial bacteria in the intestine was reduced, and the number of pathogenic bacteria was significantly increased [55,56]. For example,it was found thatFlavonifractor plautiiwas one of the potentially opportunistic pathogens and its growth was unrestricted in the UC patients [57]. Another study found that the changes in the gut microbiota of the acute colitis model were closer to that observed in UC, and the relative abundance ofRikenella, one of the main SCFA producing-bacteria, was significantly decreased in mice after acute DSS induction [58]. Similarly, the relative abundance ofAkkermansiawere decreased in the UC patients [59]. which could be promoted to proliferate after dietary intervention [60].

Taken together, gut microbiota results support our hypothesis that dietary CA alleviates IBD symptoms via improving the structure of microbial community in the gut and selectively restoring the abundance of probiotics and suppressing the growth or otherwise resisting out-competition of potential pathogens in DSS-treated colitic mice.

3.4 CA improved the fecal metabolomic profiles during colitis

Partial least squares discriminant analysis (PLS-DA) was utilized to screen out differential metabolites between groups [61].As shown in Fig. 4A, there were significant differences between groups and good repeatability within groups. Significantly altered metabolites were identified using databases HMDB and Metlin,based on the following criteria: 1) VIP ＞ 1.2 by OPLS-DA model and 2)P＜ 0.05 by Student’st-test, with red dots representing increased expression and blue dots representing decreased expression(Fig. 4B). Consequently, 25 fecal metabolites were identified,including 5 phenylpropanoids and polyketides, 4 heterocyclic compounds, 7 organic acids and derivatives, 7 lipids and lipid molecules, and 2 benzene ring compounds (Table 1). Specifically, in the comparison of differential metabolites and the involved metabolic pathways between the model and low-dose CA intervention groups,we found that many of the differential metabolites were involved in amino acid metabolism, followed by phospholipid-related compounds and neurotransmitter-related compounds, such as histidinyltryptophan,prolylglycine, glutamylphenylalanine, phosphatidylcholine,lysophosphatidylethanolamine [62], and phenylpropanolamine [63].

Table 1Differential metabolites in mouse feces between groups.

Fig. 3 Dietary CA modulated the gut microbiota diversity and composition in DSS-treated mice. (A) Simpson and coverage indices; (B) PCoA analysis at the genus level; Bacterial taxonomic profiling of the (C) phylum, (D) family, and (E) genus levels of fecal microbiota from different treatment groups; (F) LEfSe analysis on genus levels. Significantly different species among groups with LDA values greater than 3 are displayed. Data presented as mean ± SD. Significant differences were indicated as *P ＜ 0.05 vs Con; #P ＜ 0.05 vs Mod. The red dot indicated significant differences as P ＜ 0.05.

Fig. 3 (Continued)

Next, we sought to determine the metabolic pathways that are involved in the anti-colitis effects of CA. Pathway enrichment analysis is able to significantly enrich metabolic pathways or signal transduction pathways in differentially expressed genes (DEGs)compared with the whole genome background [64]. As shown in Fig. 4C, a total of 1 771 DEGs were mapped to 485 KEGG pathways.We found that several main metabolic pathways were represented,including those involved in nucleotide metabolism, lipid metabolism,amino acid metabolism and secondary compounds metabolism,membrane transport and signal transduction, the digestive system and nervous system-related channel compounds, and cancer-related compounds. Furthermore, we identified 4 significantly altered pathways by combining thePvalue and pathway impact value,and they are those related to 1)D-glutamine andD-glutamic acid metabolism, 2) tryptophan metabolism, 3) alanine, aspartate and glutamate metabolism, and 4) glutathione (GSH) metabolism(Fig. 4D,P＜ 0.05 by Fisher’s exact test). GSH plays an important role in body’s antioxidant defense, nutrient metabolism, and regulation of gene expression, DNA and protein synthesis, cell proliferation and apoptosis, signal transduction, cytokine production,and immune response [65]. It has been shown that the level of GSH decreased significantly after DSS treatment and the oxidative stress of GSH was elevated, which might evoke the mucosal inflammation of IBD [66]. Another study also found that CA significantly increased the level of GSH in DSS-treated mice [67], suggesting that CA could attenuate oxidative stress to inhibit DSS-induced colitis. In addition, the composition of gut microbiota was also changed, with the population of Firmicutes decreasing [66], which is consistent with our finding on Firmicutes. Moreover, tryptophan metabolism could directly or indirectly influence the gut microbiota composition to control intestinal inflammation in IBD patients [68]. Lipid metabolism was also significantly different between IBD patients and healthy controls [69], and lipid metabolism interacted with gut microbiota [70].Additionally, during the metabolic process of gut microbiota, there are some harmful compounds that may play a role in gut diseases such as colon cancer or IBD [71].

Fig. 4 Identification of significant fecal metabolites. (A) PLS-DA score plot of metabolomic features between groups. (B) Volcanic map of differential metabolites between groups. Red dots represent up-regulated metabolites and blue dots represent down-regulated metabolites. (C) KEGG classification of differentially expressed metabolites. Percentage of annotated metabolites (X-axis) versus KEGG categories (Y-axis); (D) KEGG pathway analysis of differentially expressed metabolites. The circles represent different KEGG pathway. The size of circle represents the number of differentially expressed metabolites in this pathway, and the color represents the P value, which is the significance level in enrichment analysis statistics.

Fig. 5 The heatmap of Spearman’s correlation analysis between candidate metabolites and microbiota. Different colors represent the value of correlation coefficient, red indicates positive correlation and blue indicated negative correlation. *P ＜ 0.05, **P ＜ 0.01, ***P ＜ 0.001.

Collectively, these observations show that CA might exert its protective effects through enhancing glutamine-glutamate metabolism and tryptophan metabolism, increasing the synthesis of GSH and other anti-inflammatory components, reducing the accumulation of lipid metabolites and the production of inflammatory mediators,thereby suppressing DSS-induced colitis in mice.

3.5 Correlation analysis between gut microbiota and metabolites

In order to explore the relationship between the distribution of intestinal microbiota and metabolites, Spearman correlation analysis was performed (Fig. 5). The results showed that there was a significant correlation between the gut bacterial strains and fecal metabolites. Among these, g_norank_f_UCG-010 negatively correlated with phosphatidylethanolamine/PS(16:0/18:2(9Z,12Z)),but positively correlated with lysophosphatidylcholine/LysoPC(20:2(11Z,14Z)), monoacylglyceride MG(24:0/0:0/0:0),lysophospholipid LysoPC(P-16:0) and phospho-ether lipid PE(P-16:0e/0:0). Three bacterial strains, including g_Blautia,g_Citrobacterand g_Clostridium_innocuum_group,had negative correlations with mesobilirubinogen and 3α,7αdihydroxycoprostanic acid, but positive correlations withL-3-cyanoalanine,L-glutamate, hypoxanthine (except g_Blautia).g_Candidatus_Stoquefichuswas negatively correlated withL-tryptophan, 4-formyl indole, deoxyinosine, 6-methylquinoline hypoxanthine and serotonin significantly. g_Ileibacteriumwas positively correlated with monoacylglyceride MG(24:0/0:0/0:0).

Phosphatidylserine was significantly lower in the CA group,compared to the model group. Lipid classes of phosphatidylserine were significantly increased in IBD patients compared with healthy volunteers [72]. Glutamate is an essential amino acid for protein synthesis and plays a key role in cellular metabolism, protecting cells from oxidative stress [73]. The level ofL-glutamate was low in the model group, but the intervention of CA was able to restore that, which could partially explain its anti-inflammatory activity. Neurotransmitter serotonin (5-HT) is also involved in the development of inflammation in IBD, causing typical symptoms of IBD, such as diarrhea, abdominal cramping and pain. After the treatment of CA, the level of 5-HT was lowered [74]. The results suggested that gut microbiota might influence host metabolism by altering these metabolic pathways, thereby promoting or inhibiting the development of IBD. These interactions between intestinal microbiota and metabolites could be potential therapeutic targets of drugs or nutraceuticals in IBD prevention and/or treatment.

This study has limitations. Although our data are all consistent with the concept that CA supplementation inhibited DSS-induced colitis in mice by improving the composition and metabolic function of the gut microbiota, the causal relationship was not established.Future studies using germ-free, antibiotics-treated mice, and/or fecal transplantation are needed to prove the causal role of the microbiota in CA’s anti-inflammation activity. Prior observations have suggested that the regulation of the Keap1/Nrf2 antioxidant defense system [24] and tight junction protein zonula occludens-1 (ZO-1) [25]might contribute to CA’s protection against DSS-induced colitis.However, the signaling pathways involved in the anti-colitis effect of CA were not examined in this study. Lastly, we did not distinguish the origins of the altered fecal metabolites— it was not clear that which metabolites were from the modification of host molecules, which were derived directly from the improved gut bacteria after CA intervention.Metabolic products of CA by either the host or the microbiota also warrant further investigation.

4. Conclusion

In summary, the present study demonstrated that oral administration of CA at reasonably achievable doses attenuated DSS-induced colitis in mice. The beneficial effects of CA against colonic and systemic inflammation were, at least partially, through modulating the composition and metabolic function of the gut microbiota.Collectively, our results demonstrated the protective role of CA as a potential prebiotic to alleviate inflammation in IBD in humans.Future studies are warranted to further determine the molecular and metabolic mechanisms underlying the anti-colitis effect of CA.

Conflict of interest

The authors have declared no conflict of interest.

Acknowledgement

This study was supported by Natural Science Foundation of Guangdong basic and applied basic research foundation(2021A1515010965), General project of Basic and applied basic Research in Guangzhou (202102080241), Laboratory opening project of Guangzhou Medical University (PX-1020423), Natural Science Foundation of Guangdong basic and applied basic research foundation([2018]105), Guangdong Provincial Department of Education(S202010570042) and Communist Youth League Committee of Guangzhou Medical University (2019A060).