Lacticaseibacillus rhamnosus LRa05 (purchased from Wecare Probiotics Co., Ltd.) was cultured in De Man Rogosa Sharpe (MRS) medium (Qingdao Hi-Tech Industrial Park Haibo Biotechnology Co., Ltd. Qingdao, China) and harvested after 18 h of culture at 37°C. A suspension of LRa05 was then obtained by suspending it in sterile water, and the concentration was adjusted to 5 × 10⁹ CFU/mL. Every day, we administered 200 μL of LRa05 suspension by gavage to the mice in the LRa05 group to investigate the effect of LRa05 on chronic low-grade inflammation.

We obtained 30 specific pathogen-free male C57BL/6 mice (6–8 weeks old) and obtained permission from the Shanghai Laboratory Animal Center. The study was approved by the Animal Care and Use Committee of the Shanghai Laboratory Animal Center (No.: 2023033012). These mice were kept in an environment with humidity (65% ± 5%), constant temperature (25 ± 2°C), and a 12 h light–dark cycle. In the experiment, we divided the mice into three groups: control (CTL, n = 10), dextran sulfate sodium (DSS, MW: 36,000–50,000 Da, YEASEN Biotechnology, Shanghai, China) at low concentration (DSS, n = 10) and DSS at low concentration combined with LRa05 (LRa05, n = 10).

To establish a model of chronic low-grade inflammation, 0.2% DSS (w/v, g/mL) was dissolved in drinking water and gave it ad libitum to the mice in the DSS and LRa05 groups, while the control group drank normal water. The experimental procedure with a 12-week experimental cycle was shown in Figure 1A, and body weight, food intake and health status of the mice were monitored regularly, and blood samples were collected via the orbital vein periodically. At the end of the experiment, the mice were fasted for 12 h and then killed by dislocating the neck. The colon and liver tissues were collected carefully, and fixed in 4% formaldehyde solution until histological analysis.

The blood samples were centrifuged at 4500 g for 10 min to obtain serum, then stored at −80°C until use. The serum was analyzed for lipopolysaccharides (LPS), and inflammatory cytokines, including, tumor necrosis factor α (TNFα), interleukin 4 (IL4), IL6, IL8 and IL10, using enzyme-linked immunosorbent assay (ELISA) kits (Wuhan Chundu Co. Ltd., Wuhan, China), under the guidelines of the manufacturer’s instructions (Funderburg et al., 2013).

Serum alanine aminotransferase (ALT) and aspartate aminotransferase (AST) levels were monitored using an automated biochemical analyzer (Hitachi 7,020, Hitachi, Tokyo, Japan) to assess the liver function of the mice. The colon and liver tissues fixed in formaldehyde solution were dehydrated, embedded with paraffin, cut into 4 μm thick sections and incubated at 37°C for 30 min in the dark (Han et al., 2023). The colon tissue was stained with hematoxylin and eosin (H&E) (Dong et al., 2022), and liver tissue was stained with Oil Red O (Riva et al., 2018), H&E and Masson (Matsumoto et al., 2013) to determine the degree of liver inflammation and fibrosis. In addition, Ethidine fluorescence was observed with 5 μmol/L dihydroethidine (DHE) to detect reactive oxygen species (ROS) in the liver (Chen et al., 2021).

We extracted total DNA from mice feces with the QIAamp DNA Stool Mini Kit (Qiagen, Hilden, NRW, Germany). The V3 – V4 region of the 16S rRNA gene was amplified with specific primers (341F: 5 ‘-CCTAYGGGRBGCASCAG-3’, 806R: 5 ‘-GGACTACNGGGTATCTAAT-3’). We then performed double-end sequencing (2 × 300 bp) of the amplicons using the Illumina MiSeq platform. The raw sequence data were analyzed with USEARCH software¹ to obtain amplicon sequence variants (ASVs) (Edgar, 2013). We assessed alpha diversity using the vegan (Simpson et al., 2022) package, while beta diversity analysis utilized principal coordinate analysis (PCoA) and PERMANOVA significance testing using the adonis2 function of the vegan package. Finally, we used the STAMP software to analyze the differences between groups.

The metadata significant differences between the experimental groups were examined using a t-test or analysis of variance. For the gut microbiota data, we used a non-parametric test with the Wilcoxon rank sum test to detect significant differences in microbial composition from different groups. All plots were created using the ggplot2 package (Wickham, 2011), correlation analyses were performed using the psych package, and heatmaps were created using the pheatmap package. We performed the analyses under the R language environment version 4.3. The significance level is p < 0.05.

The sequence data used in this article were stored in the NCBI database (accession number, SRA: PRJNA*).

Normal growth of the mice was observed throughout the experiment and no hematochezia was observed. In addition, the body weights of the mice showed no significant differences among the treatment groups (Figure 1B). Contrary to the effects observed with high concentrations of DSS (2–3.5%), which are known to significantly shorten colon length, the low concentrations used in this study (0.5–1%) did not lead to any notable reductions in colon length (Figure 1C). This observation supports our use of a milder inflammation model and is consistent with the absence of significant statistical variations, as detailed in the figure legend. The results of H&E staining of the colon (Figure 1D) also showed no inflammatory infiltrates in the tissue, indicating that low concentrations of DSS did not cause colonic inflammation.

According to the results in Figure 2A, serum LPS levels in DSS and LRa05 mice gradually increased with increasing duration of DSS administration. Serum LPS was significantly increased from week 4 in DSS mice compared to CTL mice. However, intervention with LRa05 led to a significant diminution in serum LPS levels, and it was not until week 12 that serum LPS was significantly higher in mice in the LRa05 group than in the CTL group. Of note, mice in the LRa05 group had significantly lower serum LPS levels at week 12 than mice in the DSS group. Similar to LPS, serum TNFα gradually increased in DSS and LRa05 mice with increasing duration of DSS administration (Figure 2B). At week 8, serum TNFα was significantly higher in the DSS group than that in the CTL group. However, intervention with LRa05 significantly reduced serum TNFα levels in the mice, which were not significantly different from the CTL group throughout the experiment, although serum TNFα levels were slightly higher in the mice in the LRa05 group compared with the CTL group. Serum IL6 levels in DSS and LRa05 mice showed a similar increasing trend as TNFα (Figure 2C). Serum IL6 levels were significantly increased from week 4 in DSS mice compared to CTL mice. Intervention with LRa05 significantly reduced serum IL6 levels, especially after week 8, which was significantly lower in the LRa05 group compared to the DSS group. Remarkably, no significant differences in serum IL4 (Figure 2D), IL8 (Figure 2E) and IL10 (Figure 2F) of the mice were observed between the different treatment groups throughout the experimental period.

Considering that LPS could enter the liver from the intestine via the portal vein, we examined serum ALT and AST levels to assess the effect of DSS on the liver of mice. As shown in Figure 3A, serum ALT levels in DSS mice increased with increasing treatment duration. Intervention with LRa05 significantly inhibited serum ALT levels in the mice, and we observed that at week 8, serum ALT levels in LRa05-treated mice were significantly lower than those in DSS-treated mice and decreased to levels similar to those in CTL mice at week 12. In contrast to ALT, although fluctuations in serum AST levels were observed in the mice throughout the experiment, there were no significant differences between groups (Figure 3B). These results suggested that sustained low concentrations of DSS resulted in impaired liver function in mice, whereas LRa05 protected the liver.

To gain a better insight into the effects of low concentrations of DSS on liver tissue, we performed tissue staining analysis. As shown in Figure 4A, low concentrations of DSS for 12 weeks did not lead to the accumulation of liver fat in mice. In addition, the results of hepatic H&E staining showed that DSS did not cause inflammatory infiltration and cell necrosis in the liver of mice (Figure 4B). In the results of Masson staining, we could not observe fibrosis in the liver of mice, and there was no significant difference between the groups of mice (Figure 4C). Notably, the results of liver DHE showed that sustained low concentrations of DSS led to a significant increase in reactive oxygen species in the liver of mice, and the intervention with LRa05 significantly reduced the level of reactive oxygen species in liver tissue compared with the DSS group (Figure 4D).

The results of the PCoA analysis showed (Figure 5A; Supplementary Table S1) that DSS caused significant changes in the beta diversity of the gut microbiota of mice, while the intervention with LRa05 significantly modulated the gut microbiota. The results of species accumulation curves showed that 284 ASVs were present in the gut microbiota of mice in the three groups, and 2, 1 and 1 unique ASVs were present in the CTL, DSS and LRa05 groups, respectively (Figure 5B). This indicated that DSS did not cause significant changes in the types of gut microbiota, while the main difference was manifested in the relative abundance of the microbiota. This was also confirmed by the analysis of alpha diversity, as shown in Figure 5C, where the intervention with DSS resulted in a significant increase in the richness (ACE index) and diversity (Shannon index) of the mice gut microbiota. However, the results of Chao1, ACE and Shannon indices showed that LRa05 was able to significantly decrease the DSS-induced changes in the alpha diversity of the gut microbiota.

We also analyzed the differences in the abundance of the gut microbiota in each group of mice at the phylum level (Figure 5D). In each group of mice, the gut microbiota consisted mainly of Bacteroidetes, Firmicutes, Proteobacteria, Campilobacterota and Actinobacteria, which accounted for more than 99%. The use of DSS resulted in a significant increase in the relative abundance of Bacteroidetes in the gut microbiota of mice, while Firmicutes and Campilobacterota showed a significant decrease in relative abundance. The relative abundance of Actinobacteria did not differ significantly among the three groups. The intervention with LRa05 reversed the DSS-induced changes in the gut microbiota to some extent, as shown by the increase in the relative abundance of the Firmicutes and the decrease in the relative abundance of Bacteroidetes. In addition, interestingly, the relative abundance of Proteobacteria in the gut microbiota of mice in the LRa05 group was significantly reduced compared to the CTL and DSS groups.

The differences in the gut microbiota of mice from each group at the end of the experiment were analyzed at the genus level. Compared to the CTL group, treatment with low concentrations of DSS for 12 weeks resulted in a significant decrease in the relative abundance of the beneficial bacteria Lactobacillus, Clostridium_XlVa and Bifidobacterium, and the pathogenic bacteria Duncaniella, Allobaculum, Parabacteroides and Parasutterella in the gut of the mice (Figure 6A). Treatment with LRa05 reversed this trend and significantly improved the DSS-induced changes in the gut microbiota. Compared to the DSS group, intervention with LRa05 significantly decreased the relative abundance of Duncaniella, Allobaculum and Parasutterella, and increased the relative abundance of beneficial bacteria such as Lactobacillus, Clostridium_XlVa, Bifidobacterium and Limosilactobacillus (Figure 6B).

To further analyze the effect of DSS on the gut microbiota of mice, we investigated the changes in the gut microbiota of mice over time. We collected feces from DSS and LRa05 mice after 1, 4, 8 and 12 weeks of intervention and performed gut microbiota analysis. As shown in the results of the PCoA analysis, the beta diversity of the gut microbiota of DSS mice (Figure 7A; Supplementary Table S2) and LRa05 mice (Figure 7B; Supplementary Table S3) differed significantly at different time points. In addition, the beta gut microbiota was observed to be more distant in DSS mice compared to LRa05 mice at different time points. This suggested that DSS led to disruption of the gut microbiota, while intervention with LRa05 can effectively reduce the disruption of the gut microbiota and reduce the differences in the gut microbiota at different time points. Similar to beta diversity, alpha diversity also changed over time. The abundance (Chao1 and ACE index) and diversity (Shannon index) of the gut microbiota gradually increased in the DSS mice, especially the alpha diversity of the gut microbiota increased significantly at week 12 compared to week 1 (Figure 7C). In contrast to the DSS group, the richness and diversity of the gut microbiota of the mice in the LRa05 group fluctuated but were relatively stable overall, and only the ACE index changed significantly (Figure 7D).

As shown by the analysis of the species accumulation curve (Supplementary Figure S1), no significant new bacteria emerged in the gut microbiota of DSS and LRa05 mice over time. These results suggested that the differences in the gut microbiota of mice were mainly associated with changes in the relative abundance of the gut microbiota. DSS led to a disruption of the gut microbiota, whereas LRa05 can effectively reduce the DSS-induced changes in the gut microbiota. At the phylum level, the gut microbiota was mainly composed of Bacteroidetes, Firmicutes, Proteobacteria, Campilobacterota and Actinobacteria (Supplementary Figures S2A,B). We observed DSS-induced fluctuations in the abundance of the gut microbiota at the phyla level, which was mainly manifested by a decrease in the relative abundance of Firmicutes and an increase in the relative abundance of Bacteroidetes and Proteobacteria (Supplementary Figure S2C). However, the intervention of LRa05 was able to effectively reduce the changes in the gut microbiota caused by DSS (Supplementary Figure S2D).

To investigate the relationship between the gut microbiota and inflammatory cytokines and liver parameters, we performed a Spearman correlation analysis (Figure 8). We found that microbial genera enriched in the gut of DSS mice, such as Duncaniella, Allobaculum and Parabacteroides, were positively correlated with LPS, TNFα, IL6 and ALT. However, microbial genera that decreased in DSS mice, such as Clostridium_XlVa, Limosilactobacillus, Lactobacillus and Bifidobacterium, were negatively correlated with LPS, TNFα, IL6 and ALT.

We further investigated the evolution of these microbial genera over time. We observed no significant change in the relative abundance of the beneficial bacterium Clostridium_XlVa over time in the DSS group (Figure 9A), but the intervention with LRa05 significantly increased the relative abundance of Clostridium_XlVa (Figure 9B). In addition, the relative abundance of Limosilactobacillus, Lactobacillus and Bifidobacterium in the DSS group decreased significantly over time (Figure 9A), while the intervention with LRa05 significantly increased the relative abundance of Limosilactobacillus and Lactobacillus (Figure 9B). We observed that the relative abundances of the opportunistic pathogens Duncaniella, Allobaculum, Parabacteroides and Parasutterella gradually increased over time in the DSS group (Figure 9C). In particular, the relative abundance of these microbial genera increased significantly at week 12 compared to week 1 (Figure 9C). While the relative abundances of Duncaniella, Allobaculum, and Parasutterella initially increased and then decreased after treatment with LRa05, the relative abundances of these microbial genera eventually remained stable (Figure 9D).