Butyrate ameliorates DSS-induced ulcerative colitis in mice by facilitating autophagy in intestinal epithelial cells and modulating the gut microbiota through blocking the PI3K-AKT-mTOR pathway

DSS was purchased from MP Biomedicals (0216011080, California, USA). BA was obtained from Sigma-Aldrich (B5887, Missouri, USA). 5-ASA was sourced from Aladdin (A129982, Shanghai, China). Antibodies for AKT, phospho-AKT (p-AKT), LC3II/I, and Atg5 were acquired from Cell Signaling Technology (#4691, #4060, #12741, and #12994, Boston, USA). mTOR antibody was purchased from Proteintech (81670–1-RR, Chicago, USA), and phospho-mTOR (p-mTOR) from Hu'an (HA600094, Hangzhou, China). PI3K, P62, and Beclin-1 antibodies were obtained from ABclonal (A19742, A19700, and A21191, Wuhan, China), and phospho-PI3K (p-PI3K) from Bioss (bs-5570R, Beijing, China). Mouse tumor necrosis factor-α (TNF-α), mouse interleukin-1β (IL-1β), and mouse interleukin-6 (IL-6) ELISA kits were procured from Wuhan Elabscience Biotechnology Co., Ltd. (E-EL-M3063, E-EL-M0037c, and E-EL-M0044c, Wuhan, China).

Thirty-six 5-week-old male SPF-grade C57BL/6 mice (weight: 20 ± 2 g) were acquired from Hangzhou Medical College (Production license: SCXK (Zhe) 2019–0002, Hangzhou, China). These mice were accommodated at the Animal Experiment Center of Hangzhou Medical College (Use license: SYXK (Zhe) 2019–0011). Environmental conditions were meticulously controlled, maintaining a relative humidity of 45–55%, a temperature of 24 °C (± 2 °C), and a 12-hour light/dark cycle. The mice had free access to water and standard lab feed throughout the study. The experimental protocols involving these animals were reviewed and approved by the Zhejiang Province Experimental Animal Center's Animal Ethics Committee (Approval number: ZJCLA-IACUC-20010192).

Thirty-six C57BL/6 mice were randomized into six groups: control group (Control), model group (DSS), low-dose butyrate group (DSS + BA-L, 300 mg/kg), medium-dose butyrate group (DSS + BA-M, 600 mg/kg), high-dose butyrate group (DSS + BA-H, 1200 mg/kg), and a positive control group treated with 5-ASA (100 mg/kg). All groups except the control group were administered 3% DSS in drinking water for seven days to establish the UC model. From day 3 of model establishment, the DSS BA-L, DSS BA-M, DSS BA-H, and 5-ASA groups were treated via oral gavage with the corresponding drugs for 7 consecutive days. During the experiment, daily records of body weight and changes in fecal characteristics were maintained. The severity of UC in mice was assessed utilizing a Disease Activity Index (DAI) scoring chart (Table 1). On the 11th day of the experiment, mice were anesthetized with sodium pentobarbital (50 mg/kg), followed by blood sample collection. Subsequently, the mice were euthanized to harvest the colorectal tissues and fecal samples (Fig 1A).

The colon tissue was fixed with 4% paraformaldehyde, embedded and sectioned using paraffin. The sections (5 µm) were then deparaffinized with xylene and stained with hematoxylin and eosin stain (H&E). Changes in colonic tissue structure, glands and cup cells were assessed using microscopic observation. The degree of tissue damage was assessed using the Histological Injury Score (HI) (Table 2).

Mouse blood samples collected from the orbits were allowed to stand at 4°C for 2–3 hours before being centrifuged at 12,000 rpm for 15 minutes to obtain the supernatant. The levels of inflammatory markers TNF-α, IL-1β, and IL-6 in the serum were quantitatively assessed employing ELISA kits.

Genomic DNA was isolated from fecal samples of mice, and its structural integrity was verified through 1% agarose gel electrophoresis. Both the concentration and purity of the isolated DNA were quantified before using it as a template for the PCR amplification of the V3-V4 regions of the 16S rRNA gene, which were identified by barcode-tagged primers. Subsequently, the PCR products were retrieved, purified, and utilized to prepare a sequencing library. This library was then sequenced on an Illumina Miseq PE300 system. Quality control of the raw paired-end reads was conducted using fastp software, and the reads were merged using FLASH software. Operational taxonomic units (OTUs) were clustered with a 97% similarity threshold using UPARSE software, and chimeric sequences were excised. Taxonomic classification of the OTUs was carried out against the Silva 16S rRNA gene database version 138, applying an RDP classifier with a 70% confidence level. The composition and variations of the microbial communities at different taxonomic levels were examined utilizing the cloud-based bioinformatics platform provided by Majorbio, Shanghai.

Approximately 20 mg of colon tissue was placed in a 2 mL centrifuge tube containing RIPA lysis buffer, protease inhibitors, phosphatase inhibitors, and grinding beads. The tissue was homogenized using a tissue homogenizer and then incubated on ice for 30 minutes before being centrifuged at 12,000 rpm for 15 minutes to separate the supernatant. Protein levels in the supernatant were quantified utilizing a BCA Protein Assay Kit. The proteins were then resolved through SDS-PAGE and electrotransferred onto a PVDF membrane via a wet transfer technique. This membrane was blocked with non-fat milk at ambient temperature for an hour and then incubated overnight at 4°C with primary antibodies targeting PI3K, p-PI3K, AKT, p-AKT, mTOR, p-mTOR, Beclin-1, P62, LC3II/I, and Atg5. Following this, the membrane was incubated with horseradish peroxidase-labeled secondary antibodies at ambient temperature on a shaker for an hour. Visualization of the protein bands was achieved using ECL reagent, and Image J software was employed for analysis.

Paraffin-embedded mouse colon tissue sections were first deparaffinized and rehydrated, followed by antigen retrieval using EDTA buffer in a microwave. The sections were then blocked with 10% goat serum at ambient temperature for 30 minutes. Overnight incubation at 4°C was carried out with primary antibodies specific for LC3, Beclin-1, and P62. Subsequent to TBST washing, the sections were treated with secondary antibodies for an hour at ambient temperature and then stained with DAPI to visualize nuclei. The prepared slides were analyzed under a fluorescence microscope to capture detailed images.

A 1 mm³ piece of colon tissue was fixed with electron microscopy fixative and rinsed with phosphate buffer. The tissue was then fixed again at room temperature for 2 hours in 1% osmium tetroxide in the dark. Following room temperature dehydration, the tissue was embedded in Epon 812 resin in a 37°C oven overnight. The following day, the resin block was further polymerized at 60°C for 48 hours. Ultra-thin sections were then precisely cut with an ultramicrotome, collected on 200-mesh copper grids, and sequentially stained—first with 2% uranyl acetate in ethanol for 15 minutes in darkness, followed by lead citrate under carbon dioxide-free conditions for 10 minutes. Once stained, the sections were left to dry overnight at ambient temperature in a grid box. The prepared sections were finally examined and imaged employing a transmission electron microscope.

Data analysis and plotting of statistical graphs were done using GraphPad Prism 9.5 and data were expressed as mean ± standard error (Mean ± SEM). Comparisons between groups were analyzed using one-way ANOVA and the least significant difference (LSD) test. Spearman correlation analysis was used to assess the correlation between intestinal flora and autophagy. p < 0.05 indicated statistical significance.

In the control group, the mice had smooth and shiny fur, normal food intake and feces. After three days of continuous intake of 3% DSS, the mice showed a gradual roughening of fur, decreased vigor, reduced feeding, and thin feces with occult blood. On the fourth day, some of the mice had worsened symptoms and showed obvious bloody stools. Compared with the control group, mice in the DSS group showed weight loss (Fig 1B), significantly shorter colon length (9.63 ± 0.16 cm vs 5.60 ± 0.37 cm), and colorectal bleeding. Compared with the mice in the DSS group, the shortening of the colon in the mice treated with BA intervention and 5-ASA was significantly improved (Fig 1C and D), and the mice showed a trend of recovery in body weight (Fig 1B), with the most significant effect in the DSS + BA-H group. In addition, the DAI scores of mice in the DSS group (4.9 ± 1.07) were significantly higher than those in the DSS + BA-L group (4.24 ± 0.89), DSS + BA-M group (3.58 ± 0.86), DSS + BA-H group (3.80 ± 0.86), and 5-ASA group (3.76 ± 0.94) (Fig 1E).

Histological examination revealed that in the control group, the mucosal layer of the mouse colon was intact, with well-organized and regular glands, and no signs of submucosal congestion or edema were observed. In contrast, the DSS group exhibited disrupted mucosa and glands, marked submucosal congestion and edema, extensive inflammatory cell infiltration, crypt loss, and reduced goblet cell numbers. Relative to the DSS group, the BA intervention groups and the 5-ASA group showed relatively intact colonic mucosa, reduced submucosal congestion and edema, more regular gland arrangement, and significantly fewer inflammatory cells, with the DSS + BA-H demonstrating notably fewer inflammatory cells than the DSS + BA-M and DSS + BA-L groups (Fig 2A). The HI scores of DSS-induced mice were significantly higher compared with those of the control group. Meanwhile, we found that the HI scores of mice gradually decreased with the increase of BA dose (P < 0.01), and the scores of the DSS + BA-H group were comparable to those of the 5-ASA group (Fig 2B).

ELISA analysis showed that the expression levels of inflammatory factors TNF-α (Fig 2C), IL-1β (Fig 2D) and IL-6 (Fig 2E) in the colonic tissue of DSS group mice were significantly higher than those in the Control group. However, after BA intervention, the expression levels of these inflammatory cytokines in the colonic tissue were markedly reduced. Among them, the DSS-BA-H group exhibited the most pronounced decrease in inflammatory cytokines.

We examined the composition of mouse gut microbes by 16S rRNA high-throughput gene sequencing. The Shannon index rose and then leveled off indicating that our sequencing data were reasonable enough to reflect the diversity of most microorganisms in the samples (Fig 3A). Simpson index analysis revealed that the diversity of intestinal microorganisms in mice in the DSS group was significantly lower than that in the Control group, and the BA drug intervention group (Fig 3B). Beta diversity analysis showed that Control and BA drug intervention groups showed significant clustering in principal component analysis (PCA) (Fig 3C) and principal coordinate analysis (PCoA) (Fig 3D), whereas the composition of gut microorganisms of mice in the DSS group was significantly different from these two groups. This suggests that BA altered the abundance and diversity of gut flora affected by DSS.

We analyzed species variation in the gut microbiota at both phylum and genus levels. The results at the phylum level showed that the five most abundant phyla were Bacteroidota, Firmicutes, Verrucomicrobiota, Proteobacteria, and Actinobacteriota. Compared to the control group, Bacteroidota, Firmicutes were the dominant DSS group Bacteroidota. The relative abundance of Firmicutes and Verrucomicrobiota increased in the BA drug intervention group compared to the DSS group, while the relative abundance of Bacteroidota decreased significantly (Fig 4A). The results of species composition of gut microorganisms at genus level showed higher abundance of norank_f__Muribaculaceae and lower relative abundance of Akkermansia in the DSS group compared to the control group. The abundance of Akkermansia, Turicibacter and norank__f__norank____Clostridia_UCG-014 was significantly higher in the BA drug intervention group compared to the DSS group (Fig 4B). We also performed LefSe (LDA Effect Size, LDA > 2.0) to analyze the significantly different gut microbial species in each subgroup. The results showed that Muribaculaceae and Bacteroidota were more significant in the DSS model group, and the LDA scores were > 5. In contrast to the DSS group, the significance of Muribaculaceae was significantly reduced in the BA drug intervention group, whereas the significance of Peptostreptococcales species under the phylum Firmicutes and Clostridia was increased. However, unexpectedly the significance of species under the phylum Bacteroidota such as Enterobacteriaceae and Escherichia-Shigella did not decrease in the BA drug intervention group (Fig 4C).

Autophagy is a cellular activity that maintains cellular homeostasis, and in the intestinal environment, the process of autophagy removes invading microorganisms and assists the intestinal immune system. The structure and species of intestinal flora can influence autophagic activity, and changes in intestinal microbiota due to dysbiosis can stimulate or hinder autophagy. To investigate whether the effect of BA on the DSS-induced UC model is related to autophagy, we found a close relationship between the abundance of 19 bacterial genera and the expression of autophagy-associated proteins by Spearman correlation analysis between the fecal gut microbiota and autophagy-associated proteins in mice. Notably, Parabacteroides, Turicibacter and Romboutsia were negatively correlated with the expression of autophagy-critical proteins, whereas Muribaculum and Bifidobacterium were positively correlated with the expression of autophagy-critical proteins (Fig 4D).

To understand whether the effect of BA on DSS-induced UC is through whether it involves the PI3K/AKT/mTOR pathway, a key pathway of autophagy, firstly, we observed the microstructure of mouse colon tissues by electron microscope. The mitochondrial membranes were intact, and mitochondrial inner cristae were clear and structurally intact in the supracolonic tissues of mice in Control group. The structure of mitochondrial inner cristae in the colonic tissues of mice in the DSS group was reduced or even disappeared, the mitochondrial swelling volume increased, and the mitochondrial cristae had a fracture condition. In the BA drug intervention group and 5-ASA group, not only mitochondrial damage of different conditions but also the formation of autophagic vesicles and autophagic lysosomes were observed in the colonic epithelial cells of mice (Fig 5A).

Then we found by immunofluorescence staining of autophagy key proteins Beclin-1, P62 and LC3II/I proteins (Fig 5B) that the expression of Beclin-1 (Fig 5C) and LC3II/I (Fig 5D) was reduced and the expression of P62 proteins (Fig 5E) was significantly higher in the colon tissues of mice in the DSS group. Compared with the DSS model group, the expression levels of Beclin-1 and LC3II/I were significantly increased and the expression level of P62 protein was significantly decreased in the BA drug intervention group (Fig 5C-E).

Finally, the expression levels of autophagy key proteins Beclin-1, Atg5, P62 and LC3II/I proteins (Fig 6A and B) and autophagy negative regulatory pathway-related PI3K, AKT and mTOR proteins in mouse colon tissues were further detected by Western blot (Fig 6C). The results revealed that the expression levels of Beclin-1 (Fig 6D) and Atg5 proteins (Fig 6E) and the ratio of LC3II/I (Fig 6F) were significantly lower, the expression of P62 (Fig 6G) was significantly higher, and the expression ratios of the pathway proteins P-PI3K/PI3K (Fig 6H), P-AKT/AKT (Fig 6I), and P-mTOR/mTOR (Fig 6J) were significantly increased in the colonic tissues of mice in the DSS group. Compared with the DSS model group, the BA drug intervention group up-regulated the expression of autophagy-related proteins Beclin-1 and Atg5 and the ratio of LC3II/I (Fig 6D-F), and suppressed the expression of P62 proteins and pathway proteins P-PI3K/PI3K, P-AKT/AKT and P-mTOR/mTOR (Fig 6G-J). The trend of DSS-BA-H histone changes was more obvious. This suggests that BA can activate autophagy shipment is occurring by inhibiting the PI3K/AKT/mTOR signaling pathway.