All procedures in this study received ethical approval from the Experimental Animal Welfare and Ethical Committee of Institute of Animal Science of Chinese Academy of Agricultural Sciences (IAS2019-37). A total of 24 21-day-old pigs (Yorkshire × Landrace, 6.77 ± 0.92 kg), half male and female piglets in each group, were randomly distributed into three groups with eight replicates per treatment: (1) Control group (CON), (2) LPS-challenged group (LPS), (3) Pectin-LPS group (PECL). The initial body weight and health status of the piglets were no significant difference among the groups in this study. Each group of piglets was fed with corn-soybean meal diets containing 5% citrus pectin (PEC, [with a purity of >81.4%] purchased from Yuzhong Biotech Corporation, Henan, China) or microcrystalline cellulose (MCC [99.5% purity], purchased from Engineering research center of cellulose and its derivatives, Beijing, China) respectively.

The experiment lasted for 28 days which consisted of a pre-starter period (7 days) and a starter period (21 days). During the whole test period, the diets had been formulated to meet the nutritional requirements suggested by NRC (2012) for pigs within the corresponding weight range (Table 1). On day 14 and day 21, piglets from LPS group and PECL group received a simulated bacterial challenge by an intraperitoneal injection of LPS solution (80 μg per kg BW, E. coli 0111: B4, Sigma). And meanwhile, piglets in the CON group received an intraperitoneal injection of 200 μl of saline. Previous studies proved that LPS causes serious damage to the intestinal structure and barrier function within 2 to 6 h after injection (Xu et al., 2020). Hence, 3 h after the 2nd injection of LPS or Saline, blood samples were collected from the anterior vena cava before being anaesthetized. After slaughtering, samples of cecal contents and mucosa were obtained and immediately frozen in liquid nitrogen and then transferred to −80°C for further analysis.

The cecum segment was fixed in paraformaldehyde, dehydrated, and embedded in paraffin to prepare paraffin sections (section thickness of 5um), and then stained with PAS solution. The number of goblet cells was assessed by random measurement of 10 crypts per section using DS-U3 (Nikon, Japan). For more specific details about this process, we refer to (Tang et al., 2022).

We extract the total RNA from cecal mucosa using the RNeasy Mini Kit (GeneBetter, Beijing, China). The concentration of each RNA sample was quantified using the NanoDrop 2000 (Nanodrop Technologies, Wilmington, DE, USA). The cDNA was transcribed by using the High-Capacity cDNA Archive kit (Takara, Takara Biomedical Technology in Beijing, China). qRT-PCR was conducted with a commercial kit (PerfectStart Green Qpcr SuperMix, Transgen in Beijing, China). The mRNA level of β-actin was used as an internal control. The relative genes expression of mRNA of tight junction protein (Claudin-1, Claudin-4, ZO-1), mucin (Muc-2) inflammatory cytokine genes (IL-1β, IL-6, IL-8, IL-18, TNF-α, NF-κB, IL-10, and IL-22), and immune-related receptors (GPR41, GPR43, GPR109, AHR) was detected by qRT-PCR. A single peak was checked to confirm the specificity of each primer (Table 2) set in the melting curves after 40 PCR amplification cycles. (Primer efficiency was checked by using different cDNA concentrations and only primer with mathematical efficiency between 90 and 110% were used). Relative expression of each primer between the control group and treatment group was calculated by 2^-△△Ct method, and the values were normalized to the reference house-keeping genes β-actin.

Total bacterial DNA was extracted from the intestinal chyme and mucosa using the EZNATM Soil DNA kit (D5625-02, Omega Bio- Tek Inc., Norcross, GA, USA) according to the instructions of the manufacturer. The V3-V4 hypervariable regions of the bacterial 16S rDNA were amplified by a two-step PCR method using primers 338F (5′-ACTCCTRCGGGAGGCAGCAG-3′) and 806R (5′-GGACTACCVGGGTATCTAAT-3′) with unique 8-bp barcodes to facilitate multiplexing, and sequencing was carried out with an Illumina sequencing platform using Miseq PE300 (Illumina, San Diego, USA) according to the standard protocols by Majorbio Bio-Pharm Technology Co. Ltd. (Shanghai, China).

The raw 16S rRNA gene sequencing reads were demultiplexed, quality-filtered by fastp version 0.20.0 (Chen et al., 2018), and merged by FLASH version 1.2.7 (Magoc and Salzberg, 2011). Operational taxonomic units (OTUs) with 97% similarity cutoff (Edgar, 2013) were clustered using UPARSE version 7.1 [3], and chimeric sequences were identified and removed. The taxonomy of each OTU representative sequence was analyzed by RDP Classifier version 2.2 (Wang et al., 2007) against the 16S rRNA database using a confidence threshold of 70% (PRJNA889391).

Tissue total protein was extracted in a RIPA lysis buffer on ice. Protein content was quantified using the BCA protein assay kit (Cat# 23225, Thermo, Waltham, MA, USA). A total of 30 μg protein was loaded per lane and boiled for 15 min before separation by 10% SDS-PAGE gel. The SDS-PAGE gel result was transferred onto a polyvinylidene difluoride membrane under 90 V for 1.5 h using the wet transfer method. Then the membranes were incubated in 5% skimmed milk for 2 h at room temperature for blotting. After incubation with a primary antibody of Occlaudin (Thermo Fisher Scientific Inc., MA, USA, #40–4,700, 1:500)), Claudin-1 (Thermo Fisher Scientific Inc., MA, USA, #51–9,000, 1:500), and β-actin (Proteintech, Chicago, USA, #20536-1-AP, 1:1,000) overnight at 4°C, the membrane was washed with TBST buffer and incubated with the HRP-labeled goat anti-rabbit/mouse second antibody (Abcam, Cambridge, UK, 10#ab6721, 1:5,000) for 40 min at room temperature. Protein blots were visualized using SuperSignal® West Femto Maximum Sensitivity Substrate (Cat # 34094, Thermo) and a gel imaging system (Tanon Science & Technology Co., Ltd., China). Band density was quantified by the Image J 10.0 software and normalized to β-Actin.

Quantification of Cecum Short-Chain Fatty Acids (SCFAs) in pig cecal contents were measured as described in our previous report (Tang et al., 2021). Briefly, cecal contents (200.0 mg) were thoroughly mixed with ultrapure water at a ratio of 1:9, then shocked for 30 min to mix evenly, and then incubated at 4°C overnight and then centrifuged at 10000 g for 10 min. After this, 0.9 milliliters of the supernatant were mixed with 0.1 ml of metaphosphoric acid (25% (v/v)) and kept for 3 h. The sample was then cleared by centrifugation at 10,000 g for 10 min and passed through a 0.45-μm Milled-LG filter (Jinteng, Tianjin, China) and subjected to SCFA analysis.

Data on cytokines, bacterial α-diversity indices (Chao1, and Shannon), microbial metabolites (SCFAs), and gene expression were analyzed by the Tukey–Kramer test and the Duncan multiple comparison method (JMP version 10.0, SAS Institute, Inc., Cary, NC, USA). Significance is presented as *p < 0.05, **p < 0.01, and ***p < 0.001. In addition, the correlation analysis was performed using the Mental test and Spearman’s correlation (R package “pheatmap”).

The piglets challenged with LPS showed signs of diarrhea, fever, and cough. To observe morphological and histological changes in the cecum, we performed PAS staining. It was apparent from Figure 1 that LPS decreased the number of goblet cells compared with that in the CON group (p < 0.05). Addition of pectin significantly increased the number of goblet numbers when compared with the LPS group (Figures 1A,B; p < 0.01). Furthermore, LPS significantly decreased the mRNA expression levels of tight junction protein [Claudin 1, Figure 1C, (p < 0.001), Claudin 4, Figure 1D, (p < 0.001)], pectin supplementation notably increased the mRNA levels of Claudin 1 (p < 0.01), Claudin 4 (p < 0.01), compared with LPS group (Figure 1C,D). Similarly, piglets challenged with LPS significantly reduced the mRNA expression of Muc2, and it was restored after the addition of pectin in PEC group (p < 0.05). Although WB data did not reach the significant levels (Figures 1G,H). Accordingly, pectin supplementation markedly restored the intestine damage in piglets due to LPS stress.

As shown in Figure 2, piglets in the LPS group had higher expression levels of IL-1β (p < 0.01), IL-6 (p < 0.01), IL-18 (p < 0.01), TNF-α (p < 0.001), and NF-κB (Figures 2A,B,D,E), and lower expression levels of IL-10 (p < 0.001), IL-22 (p < 0.001; Figures 2G,H) than piglets in CON group. After adding pectin, the mRNA expression levels of pro-inflammatory was significantly reduced, including IL-1β (Figure 2A; p < 0.01), IL-6 (Figure 2B; p < 0.01), IL-18 (Figure 2D; p < 0.001), TNF-α (Figure 2E; p < 0.001), and NF-κB (Figure 2F; p < 0.001), and restored the levels of the anti-inflammatory IL-10 (Figure 2G; p < 0.001) and IL-22 (Figure 2H; p < 0.01).

We profiled the composition and structure of the microbial communities in both mucosa and chyme of cecum using 16S rRNA amplicon sequencing. Rarefaction curves approached asymptotes across all samples, implying that the sequencing depth was sufficient to cover almost all microbes in the samples (Figures 3A,B). The Venn diagram in the cecal mucosa showed that piglets in the CON, LPS, and PECL groups contained 255 same OTUs and 90, 26, and 159 unique OTUs, respectively (Figure 3C). When it comes to cecal chyme, there were 163 common OTUs between these three groups. Meantime, the CON, LPS, and PECL groups contained individual 72, 45, and 197 OTUs, respectively (Figure 3D).

And after quality-filtered and merged, 723 OTUs in mucosa and 704 OTUs in chyme were obtained, respectively. In mucosa, the α-diversity was significantly higher in PECL than in CON and LPS group, while without any difference between CON and LPS group (Figures 4A–D). Similarly, the α-diversity of chyme in PECL was higher than that in other two groups (Figures 4B–D). As for beta diversity result clearly showed dissimilarity among the communities in mucosal and chyme bacteria (Figures 4E,F).

As for the microbial composition in mucosa between these three groups. At the phylum, the results showed that Firmicutes, Bacteroides, Actinobacteria, as well as Proteobactere were the four main bacterial genera phylum levels in cecum chyme (Figure 5A). Meanwhile, the abundance of Firmicutes was higher than that in other groups and decreased the abundance of Actinobacteria. As for bacteria genera in genus level of the mucosa, the composition was more complex. The main bacteria genera in the CON group and LPS group were Streptococcus Olsenella and Bacteroides. These three dominant bacteria accounted for more than 50 and 70% of the intestinal flora, respectively, in the CON group and LPS group, whereas it decreased in the PECL group (Figure 5C).

In cecal chyme, the predominant bacterial communities in CON groups and LPS were Firmicutes, Bacteroides, and Actinobacteria in cecal chyme (Figure 5B). Piglets treated with LPS which fed pectin led to a significant increase in Firmicutes as well as a clear decrease in that of Bacteroidetes and Actinobacteria. Moreover, the ratio of Firmicutes / Bacteroidetes was significantly higher in the PECL group compared with the CON and LPS group (p < 0.01). As for the community abundance on genus level, Streptococcus, Enterococcus, and Prevotella_9 were important bacterial genera in the CON group, while the LPS group showed enhanced relative abundance of streptococcus and decreased relative abundance of Enterococcus and Prevotella. And in the PECL group, the abundance of Streptococcus was dramatically decreased.

Oppositely, the abundance of Clostridium and Terrisporobacter was increased (Figure 5D). The differences in abundance between CON, LPS, and PECL groups at the genus level were calculated using the Kruskal-Wallis with FDR correlation. Figures 5E,F listed the top 9 different species (Figures 5E,F). In mucosa, LPS challenge, compared with the CON group, appeared to have some effect on the relative abundance of Olsenella, Bacteroides, Lactobacillus, Alistipes, unclassified_o_Bacteroidales, Blautia, Clostridium_sensu_stricto_1, but the difference did not reach statistical significance. However, in PECL group, the relative abundance of Olsenella, Bacteroides, and Alistipes was decreased, Lactobacillus, unclassified_o_Bacteroidales, Blautia, Subdoligranulum, Clostridium_sensu_stricto_1, and Collinsella were increased compared with LPS group (Figure 5E; p < 0.05). And in chyme, the relative abundance of Streptococcus was higher due to LPS challenged (p < 0.05), whereas Enterococcus, Prevotella_9, Alistipes, and Atopobium were decreased, compared with the CON group. After being treated with pectin, the relative abundance of Clostridium_sensu_stricto_1, Terrisporobacter and Subdoligranulum were increased (p < 0.05), Streptococcus, Enterococcus, Prevotella_9, Bacteroides, Alistipes, and Atopobium were decreased (Figure 5F; p < 0.05).

For the expression of metabolite-associated receptors, the data showed in Figure 6. Piglets challenged with LPS decreased the mRNA expression of GPR41 (p < 0.05; Figure 6A), GPR43 (Figure 6B), GPR109 (p < 0.05; Figure 6C) and AhR (p < 0.05; Figure 6D). In PECL group, the decreased mRNA expression levels of GPR43 (p < 0.05; Figure 6B), GPR109 (p < 0.001; Figure 6C) and AhR (p < 0.001; Figure 6D) were restored by pectin supplementation. Even more, the expression of GPR109 and AhR were notably higher than that in the CON group.

Among the groups, LPS challenge reduced the concentration of Acetic acid, Propionic acid, Butyric acid, and total SCFA compared with the CON group (Figures 7A; p < 0.05). Pectin supplementation markedly restored the levels of Acetic acid, Propionic acid, Isobutyric acid, Butyric acid, and total SCFA (p < 0.01), and the concentration of Isovaleric acid tended to increase.

Beyond this, the SCFA composition also varied greatly (Figure 7B). LPS decreased the percentage of propionic acid and butyric acid, whereas increased the percentage of isobutyric acid, isovaleric acid, and Valeric acid. Pectin supplementation restored those shifts, especially in butyric acid. These results demonstrated that supplementation of pectin changed not only the levels of SCFA after being challenged with LPS but also changed the composition ratio of SCFA.

To find the correlation between the chyme intestinal microbiota and cecal parameters in piglets, spearman’s correlation analysis was carried out based on experimental parameters. As shown in Figure 8A, g_Terrisporobacter was positively correlated with Goblet numbers, Muc-2, GPR109, and AhR, but negatively correlated with TNF-α, and NF-κB (p < 0.05). For g_Subdoligranulum, Goblet numbers and AhR were positively but NF-κB was negatively related with it (p < 0.05). g_Streptococcus was positively correlated with TNF-α and negatively correlated with GPR109, and AhR (p < 0.05). g_Prevotella_9 was notably positively with NF-κB (p < 0.05). g_Enterococcus and g_Bacteroides had a negative correlation with ZO-1 (p < 0.05). g_Clostridium_sensu_stricto_1 was positively correlated with the number of Goblet cells, Muc-2, IL-10, GPR43, GPR109, and AhR (p < 0.05), but negatively correlated with IL-6, TNF-α, NF-κB (p < 0.05). g_Alistipes was positively correlated with TNF-α and negatively related with ZO-1, Muc-2, GPR109, and AhR (p < 0.05).

Mantel tests were performed to detect the correlation between SCFAs and cytokines in cecal mucosa (Figure 8B). The mantel correlation analysis demonstrated that a significant correlation was observed between Claudin-4, Muc-2, IL-1β, IL-6, IL-18, TNF-α, NF-κB, IL-10, IL-22, and Acetic acid (Mental’s r > 0.25, p < 0.05). We also found Propionic acid had a powerful relationship with Goblet number, ZO-1, Muc-2, IL-6, TNF-α, GPR41, GPR109, and AhR; Isoburic acid had a significant relationship with IL-6, and NF-κB; Butyric acid had dramatically correlations with Claudin-4, Muc-2, TNF-α, NF-κB, IL-10, IL-22, GPR109, and AhR; Isovaleric acid was associated with NF-κB. Thus, SCFAs were closely associated with the goblet cells, tight junction protein, inflammatory cytokines, mucus, and metabolite-related receptors.