All animal experiments complied with the Guide for the Care and Use of Laboratory Animals in China and were approved by the Animal Care and Use Committee of Wenzhou Medical University (wydw 2017-0046).

According to previous studies (16, 17), we cultivated Candida albicans strain NBRC1385 in the C-limiting medium (270 rpm, 27°C, 48h). Then, an equivalent volume of ethanol was added to the medium and the mixture was kept at 4°C overnight. Next, we centrifuged the cultures and collected the pellets. Dissolve the pellets with water and centrifuge again to obtain the soluble part. Add an equivalent volume of ethanol before being kept at 4°C overnight. Repeat centrifugation to collect the pellets. Last, we added acetone to refine the pellets and obtained the final CAWS product when the acetone vaporized. At the time of preparing the CAWS solution, we dissolved the powder with phosphate-buffered saline (PBS) and it will be autoclaved before use.

C57BL/6J WT male mice at the age of 3-4 weeks were randomly divided into four groups (n=10 for each group): PBS group, CAWS group, and CAWS + C. butyricum group, CAWS + antibiotics (ABX) group. For the CAWS-treated group, mice were injected intraperitoneally with 4mg/kg of CAWS solution for 5 days. The PBS group was equally injected intraperitoneally with PBS buffer. C. butyricum was prepared for daily oral administration (5×10⁶ CFU/g) that started from the first day of CAWS modeling and lasted for 4 weeks. In the CAWS + ABX group, mice were treated with 5-day antibiotic cocktails of 1mg/ml streptomycin, neomycin, and bacitracin dissolved in PBS at the time of the CAWS modeling. In the addition experiment of butyrate, we added 100mM sodium butyrate in the daily drinking water for mice since the day of CAWS modeling. After 4 weeks of the final CAWS injection, mice were anesthetized and sacrificed for the heart and colon tissues.

Mice fecal samples were collected and frozen at −80°C for storage. We used the 16S ribosomal ribonucleic acid (16S rRNA) sequencing to detect the V3 and V4 regions. The 16S rRNA community profile was evaluated by Novogene’s Illumina MiSeq Technology (Suzhou, China) and analyzed on NovoMagic (https://magic.novogene.com/) to assign operational taxonomic units (OTUs). Representative sequences of each OTU were used to examine the taxonomy assignment at different levels. We generated the principal component analysis (PCA) plots for visualization of UniFrac dissimilarity using the QIIME pipeline. The characteristics of differences were assessed through linear discriminant analysis effect size (LEfSe).

Stool samples were homogenized in phosphoric acid and isohexanoic acid solution at 70 Hz and centrifuged at 4 °C to collect the supernatant for GC-MS analysis. The GC was fitted with a capillary column Agilent HP-INNOWAX (Agilent Technologies, Santa Clara, CA, USA). The temperature was initially set at 90 °C, then was increased to 120 °C at 10 °C/min, to 150 °C at 5 °C/min, and finally to 250 °C at 25 °C/min and kept at this temperature for 2 min (total running time:15min). The detector was operated using the electron impact ionization mode (electron energy: 70 eV), the full scan mode, and the selective ion monitoring (SIM) mode.

Colon and heart tissues were fixed in 4% paraformaldehyde and embedded with paraffin. Five-µm-thick sections of tissues were dewaxed and hydrated and then stained with H&E solution. All slides were observed under a light microscope and images were obtained using an SCN400 slide scanner and Digital Image Hub software (Nikon, Tokyo, Japan).

The colon and heart tissues were de-paraffinized and rehydrated before incubation within 3% H₂O₂ and 5% BSA for the block. Afterwards, the heart sections were incubated with primary antibodies against CD31(Abcam, ab24950, 1:1000), ICAM-1 (Affinity Biosciences. AF6088, 1:100) and CD68 (Proteintech, 28058-1-AP, 1:100), and primary antibodies against Claudin-1 (Proteintech, 13050-1-AP, 1:100), Jam-1 (Affinity Biosciences. DF6373, 1:100), Occludin (Proteintech, 27260-1-AP, 1:100), ZO-1 (Proteintech, 21773-1-AP, 1:300), MCT-1 (Proteintech, 20139-1-AP, 1:50) and SMCT-1 (SLC5A8, Proteintech, 21433-1-AP, 1:50) were used for colon sections. LFA-1 (ITGAL, Affinity Biosciences. DF5625, 1:50) was used for the RAW264.7. Next, sections were washed and cultured with fluorescent secondary antibodies. Images were examined under a laser scanning confocal microscope (Nikon, A1 PLUS, Tokyo, Japan).

The mouse monocytic leukemia cell line RAW264.7 (American Type Culture Collection, Manassas, VA, USA) was plated in 6-,12-well plates (Becton Dickinson Lab-ware, Franklin Lakes, NJ) and cultured in high glucose Dulbecco’s modified Eagle’s medium (DMEM) supplemented with 10% fetal bovine serum (FBS) at 37°C with 5% CO₂/95% air. No mycoplasma contamination was found. CAWS (0.5mg/ml) or an equivalent volume of PBS was added into the medium and treated RAW264.7 cells for 24h. Butyrate (1mM) was used an hour before CAWS administration.

Tissues and cells were obtained and lysed in TRIzol. Markers for RT-qPCR were selected: Claudin-1, Occludin-1, Jam-1, ZO-1, MCT-1, SMCT-1, IL-1β, IL-6, IL-8, TNF-α, LFA-1, MCP-1, MKP-1, PP1, PP2, PP2A, PTP1B and SHP2. All mRNA samples were extracted and transcribed into cDNA according to standard protocols (TAKARA, PrimeScript RT Master Mix (Perfect Real Time), RR036A). 5 ng/µL cDNA was prepared with 1.6 µL each of forward and reverse primers (5µM) and 10 µL SYBR Green PCR Master Mix (TAKARA, TB GreenTM Premix Ex TaqTM II (Tli RNaseH Plus, RR820A)) before performing on the LightCycler R 480 Real-Time PCR System (Roche Molecular Systems, Inc. Indiana, USA). The expression of all the target genes was normalized to GAPDH. The relative changes in gene expressions were calculated using the 2^−ΔΔCt method. The primers used in the study were shown in Supplementary Table 1 .

Total proteins were extracted from heart tissues or RAW264.7 cell lysates and supplemented with a cocktail of protease inhibitors and phosphatase inhibitors. Proteins were quantified to an equivalent amount, separated by SDS-PAGE gel, and electro-transferred to a PVDF membrane. Having been blocked in 5% skimmed milk, membranes were incubated with antibodies: anti-LFA-1 (ITGAL, Affinity Biosciences. DF5625, 1:1000), anti-p-JNK (Affinity Biosciences. AF3318, 1:1000), anti-total JNK (Affinity Biosciences. AF6138, 1:1000), anti-p-ERK (Affinity Biosciences. AF1015, 1:1000), anti-total ERK (Affinity Biosciences. AF0155, 1:1000), anti-p-p38(Affinity Biosciences. AF4001, 1:1000), anti-total p38 (Affinity Biosciences. AF6456, 1:1000) and GAPDH. The loading control was GAPDH (Proteintech, 60004-1-lg, 1:10000). Western blot bands were analyzed with the ChemiDicTM XRS + Imaging System (Bio-Rad Laboratories, Hercules, CA, USA), and the densities were quantified with Multi Gauge Software of Science Lab 2006 (FUJIFILM Corporation, Tokyo, Japan).

Concentrations of IL-1β, IL-6, TNF-α, MCP-1, and D-lactate in the plasma were determined by the corresponding ELISA kits (Shanghai Westang Bio-Tech CO.LTD, Shanghai, China) following the standard protocols. For examining the concentration of inflammatory cytokines in RAW264.7 cell supernatants, RAW264.7 cells were stimulated with CAWS and then washed with PBS for 24h. After that, the levels of inflammatory cytokines were determined following the manufacturer’s instructions.

All data were analyzed using SPSS version 19.0 (IBM, Armonk, NY, USA). Tests of normality were performed and data were normally distributed. A two-tailed unpaired Student’s t-test was used to compare two experimental groups, and one-way analysis of variance (ANOVA) followed by Duncan’s multiple-range test was performed for the comparison among more than two groups. Values are expressed as the mean ± SEM.

We observed significant changes in fecal microbiota compositions of the CAWS group using principal components analysis and abundance clustering heatmap ( Figures 1A, B ; Supplementary Figures 1A, B ). In the CAWS group, the relative abundance of Bacteroidetes was approximately twice as much as that in the PBS group while Firmicutes was dwindled by one third. Moreover, another manifested feature of gut microbiota in CAWS group was the decrease of Proteobacteria and Actinobacteria on phylum level ( Figure 1C ). Figures 1D, E show increased level of plasma inflammatory markers, IL-1β, IL-6, TNF-α, and monocyte chemoattractant protein-1 (MCP-1) and aggravated severity of coronary lesions in CAWS group. Interestingly, when using probiotic C. butyricum, the decreased IL-1β and IL-6 level in the plasma were observed with the attenuated vascular lesions, compared to more severe coronary lesions and increased chemokine MCP-1 in the CAWS+ABX group. Moreover, we found increased expression of the adhesion molecules CD31, ICAM-1 and CD68, indicating the activated neutrophil and macrophages recruitment and infiltration in the coronary artery wall ( Figures 1F, G ). Different gut microbiota status was associated with inflammatory conditions, manifested by the less relative fluorescence intensity of adhesion molecules in CAWS + C. butyricum group while higher intensity in ABX group. Together, we supposed the gut dysbiosis plays a crucial role in inflammatory responses of the KD mouse model during the acute phase and the use of antibiotics may have a worsening effect.

According to the LEfSe differential analysis, we identified 42 significant discriminative features (LDA>3) at the class (n=12), order (n=15) and family (n=16) levels ( Figures 2A, B ). Lactobacillus was the dominant phylum in the PBS group while Muribaculaceae and Rikenellaceae in the Bacteroidetes phylum were enriched in CAWS groups. Moreover, antibiotic use increased the abundance of Bacteroidia and some of the mucin-degrading family Akkermansiaceae, while the addition of probiotic C. butyricum significantly improved the population of Firmicutes. In Figures 2C, D , heat maps illustrated the differences in the relative abundance of phylum level and genus level. The heatmaps demonstrated that the gut microbial community was significantly changed due to the administration of CAWS, and microbial interventions further modulated the gut microbiota composition leading to different gut microbiota phenotypes. Next, we portrayed the relative abundance of gut microbiota at the phylum level ( Figure 2E ) and genus level ( Figure 2F ). Significantly, we observed a lower total abundance of gut microbiota at the genus level in the CAWS group, especially SCFAs-producing bacteria such as Lactobacillus. While with the treatment of C. butyricum, the increased total abundance and high abundance of Firmicutes implied the successful implantation and suggested its probiotic effect in promoting the gut microbial community. Interestingly, when the mice of CAWS group received the antibiotic treatment, the total abundance of gut microbiota came oppositely larger. It’s no surprise that the antibiotics will kill bacteria while potentially boosting the proliferation of harmful opportunistic pathogens and leaving more vacant niches for antibiotic-resistant bacteria (3). In CAWS + ABX group, Lachnoclostridium was detected increased in CAWS + ABX group, which could promote the production of trimethylamine (TMA) harmful to cardiovascular health and the abundance of some SCFAs-producing bacteria was detected decreased including Ruminococcaceae and Lachnospiraceae. Moreover, mice in the CAWS + ABX group were treated with 5-day antibiotic cocktails at the time of the CAWS modeling. Thus, before the final day when we collected the fecal samples, there have been some time interval for gut microbiota to restore, causing the increased abundance in CAWS + ABX group. In Figure 2G , it shows six significant SCFAs-producing gut microbes between the PBS group and CAWS group: Ruminococcus, Clostridium, Roseburia, Blautia, Faecalibacteriu, and Bacteroides, five of which were increased with the C. butyricum supplement. Taken together, we characterized the gut dysbiosis in CAWS group as the shrinking of the total gut microbiota community, particularly the decrease of SCFAs-producing bacteria which could thereby influence the inflammation and coronary artery injuries.

Since the excessive inflammatory response could consequently cause injury to tissues and organs, we aimed to evaluate the function of the gut barrier in the KD mouse model. According to Figures 3A–C , it exhibited decreased expression of tight junction proteins Claudin-1, Jam-1, Occludin, and ZO-1 in the KD mouse model. And with the probiotic supplement, we observed the restoration of tight junction proteins in the intestinal barrier, where the antibiotics depleting gut bacteria oppositely deteriorated its integrity. For evaluating the intestinal permeability, we measured the plasma level of D-lactate produced by the intestinal bacteria and physiologically limited in gut lumen. In Figure 3D , it showed higher level of D-lactate in the CAWS group than the PBS group, and interestingly, when C. butyricum was supplemented, the leaky gut condition was alleviated with decreased D-lactate levels. In this regard, these findings suggested the gut barrier impairment in KD mouse model to aggravate inflammation response of KD.

SCFAs are one of the most significant metabolites of gut microbes to maintain the intestinal barrier integrity and inhibit inflammation, especially butyrate, acetate and propionate. Presumably, the shrink of SCFAs production and decreased absorption could undermine their protective effects to contribute the inflammation of KD. In our study, we first detected the reduction in the SCFAs-producing bacteria by 16S rRNA sequencing and then employed gas chromatography-mass spectrometry (GC-MS) quantification to measure fecal SCFAs level, including acetate, propionate, butyrate, valerate, isobutyrate, and isovalerate. In Figure 4A , it shows a notable decrease in all detected SCFAs in the CAWS group in line with the decreased abundance of SCFAs-producing bacteria. The transportation of SCFAs into the circulation is facilitated by their transporters, mainly MCT-1 and SMCT-1 (18). Going through Figures 4B–D , the expression of MCT-1 and SMCT-1 was reduced in the KD mouse model, which was further decreased by antibiotics administration, maybe leading to a lower SCFAs absorption, and was partially recovered by C. butyricum supplement.

In order to determine the effects of SCFAs on the inflammation of KD, acetate, propionate and butyrate, as the main component parts of fecal SCFAs, were chosen to observe the effects on the CAWS-induced inflammation in the mice-derived macrophages RAW264.7 cells. From Figures 5A, B , butyrate significantly decreased the inflammation-related markers, including IL-1β, IL-6, IL-8, TNF-α, LFA-1 and MCP-1, superior to the acetate or propionate. In further studies, we obtained the significant anti-inflammatory effect of butyrate on KD-sera-stimulated human-derived THP-1 and PBMC cells with the reduced IL-1β, IL-6, IL-8, TNF-α, LFA-1 and MCP-1 ( Supplementary Figure 2 ). As a crucial chemokine and ligand for ICAM-1, LFA-1 serves as an important player in mediating the inflammation-related vascular injuries. In Figures 5C, D , CAWS stimulated the expression of LFA-1 in RAW264.7 cells, suggesting the importance of macrophages in inflammation-related vascular injuries, which was significantly downregulated by butyrate. Moreover, in view of the anti-inflammatory effect of butyrate in vitro, we have performed an experiment adding butyrate to the drinking water for CAWS-treated mice, and it also shown the therapeutic effect of butyrate ( Figures 5E, F ). It demonstrated that butyrate reduced the inflammatory cytokines IL-6 and TNF-α, and ameliorated the coronary lesions featured by decreased infiltration of inflammatory cells. Seen above, butyrate shows superior anti-inflammatory effect than acetate and propionate in macrophages of KD models.

In studies to date, JNK, ERK1/2, and p38 MAPK signaling pathways are suggested in initiating inflammatory reactions in KD (19, 20). Thus, we first used SP600125, SCH772984, and SB20358 which are inhibitors to JNK, ERK1/2 and p38, respectively, in CAWS-stimulated RAWS264.7 cells to validate the inflammatory role of them in KD (21). Hence, we examined a series of inflammatory cytokines including IL-1β, IL-6, IL-8, TNF-α, and MCP-1, which have been found significantly increased in KD mice. From Figure 6A , it suggested these kinase inhibitors significantly reduced the expression of different inflammatory markers, which implied the participation of JNK, ERK1/2 and p38 MAPK signaling pathways in KD. Since butyrate significantly mitigated the inflammation in CAWS-stimulated RAW264.7, we next investigated whether butyrate could act on the MAPK signaling pathways to fulfill its protective effect on RAWS264.7 cells. In Figure 6B , butyrate inhibited the activation of JNK, ERK1/2 and p38 in CAWS-stimulated RAW264.7 cells. Therefore, we speculated whether its protective effect could be related to the expression of phosphoprotein phosphatases (PPs) to dephosphorylate activated JNK, ERK1/2 and p38. Thus, among the known MPK-1, PP1, PP2, PP2A, PTP1B, and SHP2 detection for MAPK pathways inhibition, butyrate enhanced the MPK-1 expression in the CAWS-stimulated RAW264.7 cells, while no significant difference was observed in other PPs ( Figures 6C, D ). Furthermore, MKP-1 knockdown was found to promote the phosphorylation of JNK, ERK1/2, and p38, and increase inflammatory cytokines IL-6 and TNF-α, even higher than that in the CAWS group ( Figure 6E ).

As seen above, our results suggested butyrate increased expression of phosphatase MKP-1 to dephosphorylate activated JNK, ERK1/2 and p38 MAPK, consequently decreasing the inflammation in CAWS-stimulated RAW264.7 cells.