Melatonin-Mediated Colonic Microbiota Metabolite Butyrate Prevents Acute Sleep Deprivation-Induced Colitis in Mice

Radical cure colitis is a severe public health threat worldwide. Our previous studies have confirmed that melatonin can effectively improve gut microbiota disorder and mucosal injury caused by sleep deprivation (SD). The present study further explored the mechanism whereby exogenous melatonin prevented SD-induced colitis. 16S rRNA high-throughput sequencing and metabolomics analysis were used to explore the correlation between SD-induced colitis and intestinal microbiota and metabolite composition in mice. Fecal microbiota transplantation (FMT) and melatonin or butyrate supplementation tests verified the core role of gut microbiota in melatonin-alleviating SD-induced colitis. Further, in vitro tests studied the modulatory mechanism of metabolite butyrate. The results demonstrated that SD leads to reductions in plasma melatonin levels and colonic Card9 expression and consequent occurrence of colitis and gut microbiota disorder, especially the downregulation of Faecalibacterium and butyrate levels. The FMT from SD-mice to normal mice could restore SD-like colitis, while butyrate supplementation to SD-mice inhibited the occurrence of colitis, but with no change in the plasma melatonin level in both treatments. However, melatonin supplementation reversed all inductions in SD-mice. In intestinal epithelial cells, the inflammatory ameliorative effect of butyrate was blocked with pretreatments of HDAC3 agonist and HIF-1α antagonist but was mimicked by GSK-3β and p-P65 antagonists. Therefore, the administration of MLT may be a better therapy for SD-induced colitis relative to butyrate. A feasible mechanism would involve that melatonin up-regulated the Faecalibacterium population and production of its metabolite butyrate and MCT1 expression and inhibited HDAC3 in the colon, which would allow p-GSK-3β/β-catenin/HIF-1α activation and NF-κB/NLRP3 suppression to up-regulate Card9 expression and suppress inflammation response.


Introduction
Inflammatory bowel disease (IBD), which is an umbrella term for ulcerative colitis (UC) and Crohn's disease (CD), is characterized by chronic and recurrent inflammation in the gut. Its incidence and prevalence have increased worldwide in the last 50 years and IBD affects about 3 million people in the United States [1]. To radically cure IBD is a considerable challenge because IBD is defined as a multifactorial disease, and its pathogenesis is still unclear. IBD is caused by a complex interplay between genetic, immunologic, microbial, and environmental factors, making the development of a subtype-specific treatment challenging. Thus, efforts are ongoing to develop personalized therapies for the various types of IBD and improve the quality of life of patients [2][3][4]. Our previous studies have demonstrated that melatonin (MLT) can effectively improve sleep deprivation (SD)induced gut microbiota disturbances and intestinal mucosal damage [5]. Sleep impairment in IBD patients is a common problem that deserves attention in everyday clinical practice [6]. In addition, sleep is also closely related to many disorders including depression, etc. [7]. Psychological stress causes an increase in intestinal permeability and destruction HIF-1α antagonist; Abcam, Cambridge, MA, USA; LPS + Butyrate + Chetomin-cells). Each plate of treated cells was incubated for 24 h.
The cells from the 96-well culture plates were assessed for proliferation activity using an MTT (3-(4,5)-dimethylthiahiazo (-z-y1)-3,5-di-phenytetrazoliumromide; Sigma, St. Louis, MO, USA) assay. The optical density was determined using a microplate reader (Model 680, Bio-Rad, St. Louis, MO, USA) equipped with a 570 nm wavelength filter. The cells from the 12-well culture plates were collected for LDH assessment and Western blotting. Each assay used a repeat of 8 wells.

LDH Assessment
The cell supernatants were detected using an LDH assay kit (Solarbio Ltd., Beijing, China) according to the manufacturer's instructions. The LDH data were measured at 450 nm wavelength using a microplate reader (Model 680, Bio-Rad, St. Louis, MO, USA) and were expressed as U/10 4 cells. Each sample was assayed three times.

Animal Model Establishment
A total of 132 male SPF ICR mice (8 weeks old; Vital River Laboratory Animal Technology Co. Ltd., Beijing, China) were housed in 22 cages (6 mice/cage) under conventional conditions. After acclimatization for one week, the mice were randomly divided into eleven groups: sleep deprivation (SD), SD + antibiotics (SD + ABs), SD + melatonin supplementation (SD + MLT), SD + antibiotics + melatonin supplementation (SD + Abs + MLT), SD + antibiotics + butyrate supplementation (SD + Abs + butyrate), only antibiotic supplementation (ABs), non-sleep-deprived control (CON) groups, fecal microbiota transplantation (FMT) from mice of CON group (F-CON), FMT from mice of SD group (F-SD), FMT from mice of SD + MLT group (F-SM) and FMT from mice of vehicle (F-R). For more details see the Supplementary Materials.

Gut Microbiota Analysis
The colonic contents were collected and stored at −80 • C. Bacterial genomic DNA was extracted from frozen intestinal contents using the QIAamp DNA Stool Mini Kit from Qiagen (Hilden, Germany) according to the manufacturer's instructions. The 16S rRNA gene, comprising V3 and V4 regions, was amplified using PCR and composite specific bacterial primers as described previously [5]. All effective reads from each sample were clustered into operational taxonomic units based on a 97% sequence similarity according to UCLUST.

Metabolomics Profiling
We performed LC-MS analyses on a quadrupole-time-of-flight (Q-TOF) 6510 mass spectrometer (Agilent Technologies, Santa Clara, CA, USA) with an electrospray ionization source. For more details see the Supplementary Material.

Faecal Occult Blood Test
A small quantity of faces was applied to the center of the slide, and three drops of 10 g/L methylaminophenol sulphate solution and three drops of 3% hydrogen peroxide solution were added dropwise. The results were observed and images were taken immediately. The judging criterion was negative: (−) No rosy red or cherry red after 3 min; Positive: (+) Rose red or cherry red appears within 30-60 s; strong positive: (++) Rose red or cherry red appears immediately; strongest positive: (+++) Deep rose red or deep cherry red appears immediately.

Intestinal Permeability to FITC-Dextran
Before the experiment ended at 6:00 am, all mice were deprived of food for 2 h and orally gavaged with 0.6 mg/g body weight 4 kDa fluorescein isothiocyanate (FITC)dextran at a concentration of 80 mg/mL 1 h before euthanasia. Blood (approximately 1 mL) was collected by retro-orbital eye bleed and centrifuged (500× g, 10 min) to collect serum (around 500 µL). The fluorescence in the serum was measured using a fluorescent spectrophotometer with 485 nm excitation and 535 nm emission. A standard curve was created by diluting FITC-dextran in PBS. The concentration of FITC-dextran in the serum was calculated using the standard curve.

Enzyme-Linked Immunosorbent Assay (ELISA)
Plasma samples were collected for the detection of MLT, and colon samples were collected for the detection of inflammatory factors (IL-10, IFN-γ, IL-1β, IL-6, TNF-α, and IL-17) using a competitive ELISA assay (Uscn Life Science, Inc., Wuhan, China). All tests were performed according to the manufacturer's instructions. Eight samples were used in each group. Each sample was tested in triplicate. The intra-assay coefficient of variation (CV) was <10%, and the inter-assay CV was <12%. The data were measured using a microplate reader (Model 680, Bio-Rad, St. Louis, MO, USA) equipped with a 450 nm filter. The data were expressed as pg/mL for the plasma MLT level and pg/mg protein for the IL-10, IFN-γ, IL-1β, IL-6, TNF-α, and IL-17 levels of the colonic tissue.

Histological Staining
Forty-two colonic segments were immediately fixed in 4% paraformaldehyde in 0.1 M phosphate-buffered saline (pH 7.4, 4 • C) for 48 h and embedded in paraffin. Tissue crosssections (5 µm) were stained with haematoxylin and eosin (H&E). At least 30 random fields in six sections of each sample were photographed at × 400 magnification with a microscope (BX51; Olympus, Tokyo, Japan), and a total of at least 180 fields were analyzed per treatment. The scoring criterion was: (a) 0: No evidence of inflammation; (b) 2: Low level of inflammation, with scattered infiltrating mononuclear cells (1-2 foci); (c) 4: Moderate inflammation with multiple foci; (d) 6: High level of inflammation, with increased vascular density and marked wall thickening; (e) 8: Maximal severity of inflammation, with transmural leukocyte infiltration.

Immunohistochemical Staining
Paraffin sections were incubated overnight at 4 • C with a rabbit anti-mouse monoclonal primary antibody (Card9, 1:500; Abcam, Cambridge, MA, USA). For more details see the Supplementary Material.

Western Blotting
Total protein (obtained from colon samples and IECs) was extracted using lysis buffer (62.5 mmol/L Tris-HCl, 2% SDS, and 10% glycerol; pH 6.8). For more details see the Supplementary Material.

Statistical Analysis of Data
Data were analyzed using SPSS 10.0 statistical software (SPSS, Inc., Chicago, IL, USA). Data is first analyzed for normality. Data conforming to normality analysis are expressed as the mean ± SD. Differences between groups were statistically analyzed using multiway ANOVA followed by one-way ANOVA analysis, which was used to determine the significance of differences among groups and denoted as follows: different lowercase letters or different uppercase letters: p < 0.05; different letters and one contains uppercase letters and the other is lowercase letters: p < 0.01; same letter: p > 0.05.

MLT Improved the Gut Microbiota Population in SD Mice
To investigate the improvement effect of MLT on SD-induced gut microbiota disorder, we compared the changes in colonic microbiota components among the CON, SD, and SD + MLT groups. High-throughput pyrosequencing of 16S rRNA showed the distribution of colonic microbiota. Co-inertia analysis showed dysregulation of intestinal microbiota in SD mice. Compared with the CON group, the proportion of Firmicutes and Proteobacteria in SD mice increased significantly by 21.9 ± 0.12% (p = 0.002) and 39.8 ± 0.026% (p = 0.007), while the proportion of Bacteroidetes and Verrucomicrobia decreased significantly by 30.5 ± 0.032% (p = 0.022) and 59.8 ± 0.006% (p = 0.001), respectively, ( Figure 1A,B). In particular, the content of colonic Faecalibacterium reduced significantly by 60.3 ± 0.034% (p < 0.001) in the SD group compared with that in the CON group ( Figure 2D). However, supplementation with MLT restored these changes in flora populations and resulted in no significant difference between the CON and SD + MLT groups (p > 0.209; Figure 1C). Meanwhile, we correlated the most remarkable changes in intestinal microbes and metabolites of the three groups using correlation network analysis ( Figure 1D-F, Tables S1-S3). After SD, the levels of metabolites associated with Firmicutes and Proteobacteria were significantly upregulated by 15.4 ± 0.029% (p = 0.003) and 27.9 ± 0.065% (p = 0.013), respectively, compared with those in the CON group, but were downregulated in Bacteroidetes by 15.9 ± 0.018% (p = 0.032). Similarly, this process was dramatically suppressed by MLT supplementation. As shown in Figure 1G, the most prominent alteration associated with intestinal microbiota and its metabolites involved 12 metabolites and 83 intestinal microbiota. The content of butyrate was directly proportional to the abundance of Faecalibactium, Lachnospiraceae, Erysipelatoclostridium, and Butyricicoccus, but was inversely proportional to the abundance of Alistipes, Mollicutes, and Rikenellaceae.

MLT Improved Gut Microbiota Metabolite Disproportionality in SD Mice
We further analyzed the changes in metabolites that responded to the exposure of SD mice to MLT. The analysis of the metabolites showed that there were 1023 metabolites in the colon. The β-diversity analysis showed an obvious clustering of microbiota metabolite composition of all CON, SD, and SD+MLT groups ( Figure 3A). SD led to a significant increase in colonic microbiota metabolite dispersion, which suggested a decrease in microbiota metabolite homogeneity. The Venn diagram indicated that different treatments resulted in different metabolite changes ( Figure 3B). Compared with the CON group, 51 metabolite contents were changed in the SD group, while MLT supplementation restored the content of 27 of these metabolites. The 27 metabolites included butyrate, L-Proline, and Deoxyinosine, which are closely related to treatments such as SD and MLT supplementation. We then performed a statistical analysis of the differences in metabolite expression levels among the CON, SD, and SD + MLT groups (Figures 3C,D and S3). Specifically, compared with the CON group, 25 metabolites were upregulated and 26 were downregulated after SD, while MLT supplementation reversed these processes ( Figure 3C). After MLT supplementation, there were 103 metabolites increased and 28 metabolites decreased, compared with the SD group. Furthermore, we screened the 49 most changed metabolites in the three groups ( Figure 3E). Of these, 19 metabolites were significantly reduced and 17 metabolites increased in the SD group compared with those in the CON group. Figure 3F-K summarizes the six metabolites with the most significant changes caused by SD. There was a remarkable decrease in the content of butyrate (75.9 ± 0.039%, p = 0.007, Figure 3F), acetylcarnitine (41.7 ± 0.045%, p = 0.009, Figure 3G), deoxyadenosine (45.7 ± 0.051%, p = 0.002, Figure 3H), and acetylcholine (46.4 ± 0.021%, p < 0.001, Figure 3I) as well as an increase in the content of L-proline (108.1 ± 0.09%, p < 0.001, Figure 3H) and dopamine (114.2 ± 0.010%, p = 0.001, Figure 3K) in the SD group compared to those in the CON group. However, MLT supplementation reversed these changes to the control levels (p > 0.55), indicating there was no significant difference between the control group and the MLT supplementation group.   Relative Faecalibacterium level (D) was measured using a metabolomics test in the CON, SD, and SD + MLT groups. Values are presented as the mean ± SE. Differences were assessed using ANOVA and are denoted as follows: different lowercase letters: p < 0.05; different uppercase letters: p < 0.01; same letter: p > 0.05.

MLT Improved Gut Microbiota Metabolite Disproportionality in SD Mice
We further analyzed the changes in metabolites that responded to the exposure of SD mice to MLT. The analysis of the metabolites showed that there were 1023 metabolites in the colon. The β-diversity analysis showed an obvious clustering of microbiota metabolite composition of all CON, SD, and SD+MLT groups ( Figure 3A). SD led to a significant increase in colonic microbiota metabolite dispersion, which suggested a decrease in microbiota metabolite homogeneity. The Venn diagram indicated that different treatments resulted in different metabolite changes ( Figure 3B). Compared with the CON group, 51 metabolite contents were changed in the SD group, while MLT supplementation restored the content of 27 of these metabolites. The 27 metabolites included butyrate, L-Proline, and Deoxyinosine, which are closely related to treatments such as SD and MLT supplementation. We then performed a statistical analysis of the differences in metabolite expression levels among the CON, SD, and SD + MLT groups ( Figures 3C,D and S3). Specifically, compared with the CON group, 25 metabolites were upregulated and 26 were downregulated after SD, while MLT supplementation reversed these processes ( Figure  3C). After MLT supplementation, there were 103 metabolites increased and 28 metabolites decreased, compared with the SD group. Furthermore, we screened the 49 most changed metabolites in the three groups ( Figure 3E). Of these, 19 metabolites were significantly reduced and 17 metabolites increased in the SD group compared with those in the CON group. Figure 3F-K summarizes the six metabolites with the most significant changes Figure 2. KEGG analysis. KEGG difference analysis (A) and KEGG pathway enrichment (B), CON group compared to SD group; Links between signal pathways and functions based on KEGG analysis (C) (n = 8). Relative Faecalibacterium level (D) was measured using a metabolomics test in the CON, SD, and SD + MLT groups. Values are presented as the mean ± SE. Differences were assessed using ANOVA and are denoted as follows: different lowercase letters or different uppercase letters: p < 0.05; different letters and one contains uppercase letters and the other is lowercase letters: p < 0.01; same letter: p > 0.05.
To identify how the changes in intestinal microbiota metabolites affect host signaling pathways, we classified the annotation results of differential metabolites according to the pathway types in KEGG (Figure 2A,B and Figure S4A-D). SD mainly affects the following signaling pathways: metabolic pathways, cAMP-PKG signaling pathway, neuroactive ligand-receptor interaction, gap junction, cocaine addiction, bile secretion, tight junction, Parkinson's disease, and FOXO signaling pathway. Meanwhile, the SD-induced pathway changes were closely linked to the production of reactive oxygen species, tight junction proteins, inflammatory factors, HDAC, and the NF-κB pathway ( Figure 2C). It can be speculated that SD leads to changes in metabolites, especially a decrease in butyrate, which in turn affects these pathways and functions, ultimately resulting in colitis. p = 0.007, Figure 3F), acetylcarnitine (41.7 ± 0.045%, p = 0.009, Figure 3G), deoxyadenosine (45.7 ± 0.051%, p = 0.002, Figure 3H), and acetylcholine (46.4 ± 0.021%, p < 0.001, Figure 3I) as well as an increase in the content of L-proline (108.1 ± 0.09%, p < 0.001, Figure 3H) and dopamine (114.2 ± 0.010%, p = 0.001, Figure 3K) in the SD group compared to those in the CON group. However, MLT supplementation reversed these changes to the control levels (p > 0.55), indicating there was no significant difference between the control group and the MLT supplementation group. The β-diversity of principal component analysis (PCA) (A) and Venn based on the microbiota metabolites (B) in the CON, SD, and SD + MLT groups; volcano plot based on the differential metabolite screening (C), compared with the CON and SD groups; volcano plot based on the differential metabolite screening (D), compared with the SD and SD + MLT groups; heatmap showing the relative abundance of the key identified 49 metabolites that were significantly altered by melatonin in SD mice (E). The relative abundance of butyrate (F), acetylcarnitine (G), deoxyadenosine (H), acetylcholine (I), L-proline (J), and dopamine (K) in the CON, SD, and SD + MLT groups in the colon microbiota based on the heatmap results. CON: non-sleep-deprived control group; SD: sleep deprivation group; SD + MLT: SD + melatonin supplement group.
To identify how the changes in intestinal microbiota metabolites affect host signaling pathways, we classified the annotation results of differential metabolites according to the pathway types in KEGG (Figures 2A,B and S4A-D). SD mainly affects the following signaling pathways: metabolic pathways, cAMP-PKG signaling pathway, neuroactive ligand-receptor interaction, gap junction, cocaine addiction, bile secretion, tight junction, Parkinson's disease, and FOXO signaling pathway. Meanwhile, the SD-induced pathway changes were closely linked to the production of reactive oxygen species, tight junction proteins, inflammatory factors, HDAC, and the NF-κB pathway ( Figure 2C). It can be The β-diversity of principal component analysis (PCA) (A) and Venn based on the microbiota metabolites (B) in the CON, SD, and SD + MLT groups; volcano plot based on the differential metabolite screening (C), compared with the CON and SD groups; volcano plot based on the differential metabolite screening (D), compared with the SD and SD + MLT groups; heatmap showing the relative abundance of the key identified 49 metabolites that were significantly altered by melatonin in SD mice (E). The relative abundance of butyrate (F), acetylcarnitine (G), deoxyadenosine (H), acetylcholine (I), L-proline (J), and dopamine (K) in the CON, SD, and SD + MLT groups in the colon microbiota based on the heatmap results. CON: non-sleep-deprived control group; SD: sleep deprivation group; SD + MLT: SD + melatonin supplement group. Differences were assessed using ANOVA and are denoted as follows: different lowercase letters or different uppercase letters: p < 0.05; different letters and one contains uppercase letters and the other is lowercase letters: p < 0.01; same letter: p > 0.05.

FMT Promote Re-Establishment of the Intestinal Microecology in Mice
As illustrated in Figure 4, FMT from SD mice induced a colitis, inflammatory response and intestinal microbiota dysbiosis. Compared with the F-CON group, body weight and colonic length decreased by 4.3 ± 0.021% (p = 0.026, Figure 4A) and 54.3 ± 0.027% (p = 0.004, Figure 4D,E), while colonic permeability and histopathological score increased by 38.1 ± 0.003% (p = 0.002, Figure 4B,C) and 51.5 ± 0.038% (p < 0.001, Figure 4F,G) in the F-SD group. However, the stimulatory effects of F-SD on changes in colitis were reversed in the colon by F-SM supplementation.
Further, we examined whether F-SD caused colonic inflammation and intestinal microbiota imbalance. The IL-10 and IFN-γ levels were significantly reduced by 36.3 ± 0.014% (p = 0.010, Figure 4H) and 41.4 ± 0.041% (p = 0.002, Figure 4I) in the F-SD group, respectively, versus the F-CON group. By contrast, IL-1β and IL-6 content was significantly elevated by 32.5 ± 0.005% (p = 0.032, Figure 4J) and 46.3 ± 0.049% (p = 0.010, Figure 4K) in the colon of the F-SD group versus the control group. However, in the F-SM group, all indexes were restored to the level of the F-CON group, resulting in no statistically significant differences between the control group and the F-SM and F-MT groups (p > 0.783).
Moreover, high-throughput pyrosequencing of 16S rRNA demonstrated that the relative abundance of Firmicutes and Proteobacteria increased significantly and the content of Bacteroidetes and Faecalibacterium decreased markedly in the F-SD-treated mice compared with the F-CON group. However, in the F-SM group, all indexes were restored to the level of the F-CON group ( Figure S5A,B).
Collectively, these results indicated that FMT promotes the re-establishment of the intestinal microecology.
Further, we examined whether F-SD caused colonic inflammation and intestinal microbiota imbalance. The IL-10 and IFN-γ levels were significantly reduced by 36.3 ± 0.014% (p = 0.010, Figure 4H) and 41.4 ± 0.041% (p = 0.002, Figure 4I) in the F-SD group, respectively, versus the F-CON group. By contrast, IL-1β and IL-6 content was significantly elevated by 32.5 ± 0.005% (p = 0.032, Figure 4J) and 46.3 ± 0.049% (p = 0.010, Figure  4K) in the colon of the F-SD group versus the control group. However, in the F-SM group, all indexes were restored to the level of the F-CON group, resulting in no statistically significant differences between the control group and the F-SM and F-MT groups (p > 0.783).
Moreover, high-throughput pyrosequencing of 16S rRNA demonstrated that the relative abundance of Firmicutes and Proteobacteria increased significantly and the content of Bacteroidetes and Faecalibacterium decreased markedly in the F-SD-treated mice compared with the F-CON group. However, in the F-SM group, all indexes were restored to the level of the F-CON group ( Figure S5A,B).
Collectively, these results indicated that FMT promotes the re-establishment of the intestinal microecology.

MLT Supplementation Improved Plasma MLT Levels and Colonic Butyrate Production in SD Mice
As shown in Table 1, compared with the CON group, plasma MLT levels decreased by 62.6 ± 0.037% (p = 0.026) and 62.2 ± 0.004% (p = 0.040) in the SD and SD + ABs groups, but no change was observed in the ABs group. However, after MLT supplementation, the MLT content increased by 73.6 ± 0.004-117.8 ± 0.007% (p = 0.026-0.040) and 72.7 ± 0.014-141.5 ± 0.071% (p = 0.015-0.024) in the SD + MLT and SD + Abs + MLT groups, respectively, relative to that in the SD and SD + ABs groups, but there were no significant differences among that in the CON, SD + MLT, and SD + Abs + MLT groups. In contrast, butyrate supplementation did not alter the MLT level in the SD + Abs + Butyrate group compared to that in the SD + ABs group (p = 0.206). Similar to the change in plasma MLT levels, the colonic butyrate content decreased markedly in the SD (20.4 ± 0.079%, p = 0.003), ABs (20.5 ± 0.009%, p = 0.002), and SD + ABs (21.9 ± 0.035%, p = 0.002) groups compared with that in the CON group, whereas it increased in the SD + MLT (22.8 ± 0.052%, p = 0.012), SD + Abs + MLT (23.7 ± 0.058%, p = 0.009), and SD + Abs + Butyrate (42.3 ± 0.064%, p = 0.002) groups compared with that in the SD group. However, there were no significant differences in the butyrate content among the SD, ABs, and SD + ABs groups (p > 0.126) or among the CON, SD + MLT, SD + Abs + MLT, and SD + Abs + Butyrate groups (p > 0.626). In contrast to butyrate, colonic acetate and propionate contents did not change in all treatment groups (p > 0.734).
In contrast, MLT supplementation reversed the SD-induced changes in colitis and inflammatory response along with an increase in MLT and butyrate; no significant difference was observed in fecal occult blood, body weight (p > 0.052), colon length (p > 0.099), colonic permeability (p > 0.335), histopathological score (p > 0.243), anti-inflammatory cytokines (IL-10 and IFN-γ, p > 0.283), proinflammatory cytokines (TNF-α and IL-17, p > 0.215), and enteritis sensitive protein Card9 (p > 0.220) among the SD + MLT, SD + Abs + MLT, and CON groups. Similar to MLT supplementation, butyrate supplementation improved the SD-induced colitis and inflammatory response along with an increase in butyrate levels; no difference was observed in all the parameters between the SD + Abs + Butyrate and CON groups. However, SD + Abs + Butyrate treatment did not alter the reduction in MLT compared to that in the SD + MLT and SD + Abs + MLT treatments. tokines (IL-10 and IFN-γ, p > 0.283), proinflammatory cytokines (TNF-α and IL-17, p > 0.215), and enteritis sensitive protein Card9 (p > 0.220) among the SD + MLT, SD + Abs + MLT, and CON groups. Similar to MLT supplementation, butyrate supplementation improved the SD-induced colitis and inflammatory response along with an increase in butyrate levels; no difference was observed in all the parameters between the SD + Abs + Butyrate and CON groups. However, SD + Abs + Butyrate treatment did not alter the reduction in MLT compared to that in the SD + MLT and SD + Abs + MLT treatments.

Discussions
Our previous studies showed that the administration of exogenous MLT can effectively alleviate intestinal barrier dysfunction caused by SD [5,30]. The present study further explored the role of MLT in acute 72 h SD-induced gut microbiota metabolite disproportionality. Consistent with a decrease in plasma MLT levels, we observed an imbalance in microbiota composition in the acute SD-mice, which showed an increase in abundance of Firmicutes and Proteobacteria and a decrease in the abundance of Bacteroidetes and Verrucomicrobia, especially the decline in Faecalibacterium. Our results are supported by studies that found an increase in the content of Firmicutes and Proteobacteria, and a decrease in the content of Verrucomicrobia in the faces of obese mice fed a high-fat diet [31]. However, unlike our observation, Zhang found that SD did not have any significant effect on changes in the composition of the gut microbiota, as reported in humans and rats [32]. This discrepancy could be due to differences in model organisms (mice versus humans and rats) and samples for microbiota composition analysis (colonic content versus faces). A previous study demonstrated that the imbalance in the microbial community towards a dysbiotic state is more pronounced in colonic samples than in fecal samples in patients with Crohn's disease [33], which indicates that the microbiota composition of the colonic contents can better reflect the changes in intestinal microbiota in colitis.
Faecalibactium is one of the most abundant butyrate-producing bacteria in the gastrointestinal tract [34]. In this study, 51 metabolite levels were affected in SD mice. Compared with the control group, 25 metabolites increased significantly, and 26 decreased. Of these, butyrate decreased the most. Liquid phase analysis also further demonstrated that the butyrate content decreased significantly in the colon of SD mice, but no changes were observed in acetate and propionate contents. Previous studies have also shown that physiological stressors could lead to a significant decrease in butyrate [35].
Moreover, the mice, receiving feces microbiota from SD mice also suffered colitis phenotype and intestinal microbiota disorder, but no changes in plasma MLT were observed. However, MLT supplementation reversed all inductions in SD-mice, and transplanting feces microbiota from SD + MLT mice significantly restored SD-induced colitis and intestinal microbiota imbalance, including the reduction of Faecalibactium and butyrate. Similarly, MLT prevents obesity in mice through modulation of gut microbiota, including increasing the content of Faecalibactium [31]. Consequently, our results indicate that MLT effectively improved acute SD-induced colitis via restoring the intestinal microbiota homeostasis especially the increase in Faecalibactium and butyrate production.
Faecalibacterium, a butyrate producer, can be considered a sensor and a marker of human health [36] because its abundance is associated with intestinal disorders such as IBD, irritable bowel syndrome, and colorectal cancer [37]. Butyrate plays a major role in gut physiology: intrarectal delivery of butyrate-producing bacteria can prevent colitis [19]. In this study, metabolic function and pathway enrichment analysis showed that the decrease in Faecalibactium and butyrate content in the colon induced inflammatory response in SD mice. In vivo experiments further showed that acute SD induced weight loss, shorter colon length, fecal occult blood, higher histopathological score as well as increased intestinal permeability and pro-inflammatory factor (TNF-α and IL-17) levels, and decreased Card9 protein and anti-inflammatory factor (IL-10 and IFN-γ) levels. Our results corroborate previous studies in mice that were exposed to three SD-induced typical characteristics of colitis: weight loss, shorter colon length, and fecal occult blood [38]. Genome-wide association studies found strong correlations between loss-of-function Card9 mutations and an increased likelihood of developing inflammatory diseases [39,40]. Specifically, one of the numerous colitis susceptibility factors, Card9-signaling has been implicated in intestinal immune responses and the maintenance of homeostasis after epithelial injury and bacterial infection in mice [41]. Therefore, our studies have confirmed that SD leads to the occurrence of colitis along with a decrease in MLT and butyrate levels.
Interestingly, even though either supplementation with MLT (SD + MLT and SD + Abs + MLT group) or butyrate (SD + Abs + Butyrate group) to SD mice significantly increased butyrate levels and improved SD-induced colitis, the former can increase MLT levels, whereas the latter cannot remove the suppression of MLT secretion, which is the fundamental cause of SD-induced colitis. In particular, a single treatment of mice with ABs showed a reduction in butyrate and the occurrence of colitis, similar to SD treatment, but there was no significant change in MLT levels compared with the control group. MLT suppression plays a crucial role in SD-induced intestinal barrier dysfunction [5]. Consequently, these results suggest that SD inhibits MLT secretion, drives a selective reduction in colonic Faecalibactium population and butyrate production, and ultimately induces colitis. Furthermore, the scientific literature has demonstrated that glutamine is one of the main beneficial amino acids. It plays an important role in gut microbiota and immunity [42]. Therefore, the combined supplementation of exogenous MLT and glutamine may be a better choice to treat SD-induced colitis relative to administration of butyrate, although MLT prevents SD-induced colitis via butyrate mediation.
Butyrate exerts anti-inflammatory properties through the downregulation of proinflammatory cytokines [18]. Butyrate may permeate the cell membrane using two different solute transporters, the proton-coupled MCT1 and SMCT1 [20,21], or through the activation of metabolite-sensing G-protein coupled receptors (GPR41, GPR43, and GPR109) [13,22,43]. Once inside the cell, butyrate can regulate cell functions by HDAC inhibition [16,25,26]. This study explored the mechanism of butyrate-mediated MLT-improved SD-induced colitis using in vivo and in vitro tests. In vivo, we found that SD upregulated the expression of HDAC3, p-P65, p-IκB, and NLRP3 and downregulated the expression of MCT1, p-GSK-3β, β-catenin, HIF-1α, and Card9. However, supplementation of MLT or butyrate to SD mice effectively suppressed this process. Similarly, in vitro, butyrate reversed the LPS-induced increase in HDAC3, p-P65, p-IκB, NLRP3, and LDH index and the decrease in MCT1, p-GSK-3β, β-catenin, HIF-1α, Card9, and proliferation activity in IECs. Furthermore, the improvement in butyrate could be blocked by ITSA-1 and Chetomin. However, ITSA-1 did not change the expression of MCT1 and HDAC3, and Chetomin did not affect the expression levels of MCT1, HDAC3, p-GSK-3β, β-catenin, and HIF-1α after butyrate pre-treatment. A previous study confirmed that HDAC3 mediates cardioprotection of remifentanil postconditioning by targeting GSK-3β in H9c2 cardiomyocytes in hypoxia/reoxygenation injury [44]. HIF-1α is involved in the regulation of GSK-3β signaling in H9c2 cells in myocardial I/R injury [45]. Similarly, TWS119 and PDTC mimicked the effect of butyrate in the LPS-treated IECs, including the increase in Card9 and proliferative activity as well as the reduction in p-P65, p-IκB, NLRP3, and LDH index. TWS119 did not affect the expression levels of MCT1, HDAC3, and p-GSK-3β, while PDTC did not affect the expression levels of MCT1, HDAC3, p-GSK-3β, β-catenin, and HIF-1α. Moreover, previous studies revealed physical and functional interactions between HIF and NF-κB [46]. HIF-1α has been shown to restrict NF-κB in vivo and in vitro under inflammatory conditions [46,47]. Our study showed that MLT mediated butyrate-improved SD-induced colitis through the MCT1/HDAC3/P-GSK-3β/HIF-1α/NF-KB loop.
This study provides a deeper understanding of how MLT improved SD-induced colitis by increasing Faecalibacterium and its metabolite butyrate production ( Figure 7A,B). This finding demonstrates the effectiveness of MLT in the treatment of SD-induced colitis relative to butyrate and may facilitate the development of new treatments using MLT for various gastrointestinal disorders.