All mouse experimental procedures were approved by The Guangdong Pharmaceutical University Experimental Animal Ethics Committee (Guangzhou, China; approval no. gdpulacspf2017030-1) and are reported according to the ARRIVE guidelines. A total of 50 male C57BL/6 mice (age, 7 weeks; body weight, ~24 g) were purchased from Hunan Lex Jingda Laboratory Animal Co., Ltd. The mice were housed in a specific pathogen-free animal facility with a 12-h light/dark cycle and 60-65% humidity, at 25˚C, and with free access to food and water. After 1 week of acclimatization, 40 mice were fed an HCD (Dyets, Inc.; cat. no. ASHF3; containing 22.60% protein, 45.20% carbohydrate, 20.10% fat and 1.25% cholesterol) and the remaining 10 mice continued to be fed a normal food diet (NFD; Beijing Keao Xieli Feed Co., Ltd.; cat. no. 2212; containing 23.07% protein, 65.08% carbohydrate and 11.85% fat). After 4 weeks, the mice fed the HCD were divided into four groups (n=10/group), based on the TC concentration observed in the serum from blood collected from the tail vein, such that the TC level was similar in each group (Fig. S1A and B). The dosage of sodium sulphate used in the present study was based on the sodium sulphate content in mirabilite administered to mice and rats in a previous study (6). In three of the HCD groups, mice were intragastrically treated with sodium sulphate (Damao Chemical Regent Factory; cat. no. 7757-82-6) at a low dose (LSS; 158.5 mg/kg/day; aqueous solute, 0.01 ml/g/mouse), middle dose (MSS; 317.0 mg/kg/day; aqueous solute, 0.01 ml/g/mouse) and high dose (HSS; 634.0 mg/kg/day; aqueous solute, 0.01 ml/g/mouse), along with the HCD. As such, these three groups were designated the HCD + LSS, HCD + MSS and HCD + HSS groups, respectively. The fourth HCD group was fed the HCD only and was designated the HCD group. The mice fed the NFD were used as the control group (CON group). Mice from the CON and HCD groups were intragastrically administered an equal volume of water (0.01 ml/g/mouse). The body weight of the mice was measured once a week. After 3 weeks of sodium sulphate administration, blood was collected from the orbital vein of the mice following anesthesia with 3% isoflurane. After blood collection, the mice were sacrificed via cervical dislocation. Animal death was confirmed through the stoppage of breathing. The hepatic tissues and intestines of the mice were then collected for biochemical, histological and molecular analyses.

The concentrations of TC and triglycerides (TG) in the serum and hepatic tissues, and the concentrations of high-density lipoprotein cholesterol (HDL-C), LDL-C, total bile acid (TBA), alanine aminotransferase (ALT) and aspartate aminotransferase (AST) in the serum were measured according to the manufacturer's protocols for each kit. The kits for measuring the concentrations of TC (cat. no. A111-1-1), TG (cat. no. A110-1-1), HDL-C (cat. no. A112-1-1), LDL-C (cat. no. A113-1-1), TBA (cat. no. E003-2-1), ALT (cat. no. C009-2-1) and AST (cat. no. C010-2-1) were purchased from Nanjing Jiancheng Bioengineering Institute.

Mouse hepatic tissues were fixed in 4% paraformaldehyde at 4˚C overnight and then embedded in paraffin for H&E staining. The 4-µm paraffin sections were stained with haematoxylin (cat. no. H9627; Sigma-Aldrich; Merck KGaA) for 3 min followed by eosin (cat. no. E4009; Sigma-Aldrich; Merck KGaA) for 20 sec, both at room temperature. The images were captured with a PerkinElmer Automated Quantitative Pathology System (PerkinElmer, Inc.).

Hepatic tissues from each group of mice were fresh frozen in liquid nitrogen and then stored at -80˚C. RNA extraction, library construction, sequencing and transcriptome analysis were all conducted by Guangzhou Gene Denovo Biotechnology Co., Ltd. Briefly, total RNA from each sample was extracted using TRIzol^® reagent (Invitrogen; Thermo Fisher Scientific, Inc.). The RNA quality was evaluated using an Agilent 2100 Bioanalyzer (Agilent Technologies, Inc.) and then checked via RNase-free agarose gel electrophoresis. Following extraction, eukaryotic mRNA was enriched using oligo(dT) beads (New England BioLabs, Inc.). The enriched mRNA molecules were fragmented in fragmentation buffer, and then reverse transcribed into cDNA using random primers (New England BioLabs, Inc.). The cDNA fragments were subsequently purified using a QiaQuick PCR extraction kit (Qiagen China Co., Ltd.), followed by end-repaired poly(A) addition and then ligation to Illumina sequencing adapters. The HiSeq Rapid Cluster Kit v2 (PE-402-4002) and HiSeq Rapid SBS Kit v2 (FC-402-4023) were used. The ligation products were separated by size via agarose gel electrophoresis, then PCR amplified and finally sequenced using an Illumina HiSeq 2500 System (Illumina, Inc.). Then, an ABI StepOnePlus Real Time PCR System (hermo Fisher Scientific, Inc.) was used to detect library concentration. The concentration was >5 ng/µl. The reads were then filtered using fastp (version 0.18.0) (21) and differential expression analysis was performed using DESeq2 software (version 1.20.0) (22). Differentially expressed mRNAs with a false discovery rate (FDR) <0.05 and absolute fold change ≥2 were considered included in subsequent analyses.

Kyoto Encyclopedia of Genes and Genomes (KEGG) is a public pathway-related database used to analyze identified significantly enriched metabolic pathways or signal transduction pathways of differentially expressed genes (DEGs) compared with the whole genome background (23). The calculated P-values underwent FDR correction, and KEGG pathways with FDR ≤0.05 were defined as significantly enriched pathways of the DEGs.

Total RNA was extracted from each hepatic tissue and ileal sample harvested from the experimental mice using TRIzol reagent, and then subjected to RT using the PrimeScript™ RT Reagent kit (Takara Bio, Inc.) at 37˚C for 15 min followed by 85˚C for 5 sec. qPCR was performed using the SYBR Premix Ex Taq kit (Takara Bio, Inc.) and the LightCycler 480II System (Roche Diagnostics). The thermocycling conditions were as follows: 95˚C for 30 sec; and then 40 cycles of 95˚C for 5 sec, 60˚C for 20 sec and 65˚C for 15 sec. GAPDH was used as the internal reference. All primers used for qPCR in the present study are listed in Table SI. The 2^-ΔΔCq method was used to quantify the relative transcriptional level of each gene (24).

Hepatic tissues were lysed in Radio-Immunoprecipitation Assay lysis buffer (Dalian Meilun Biology Technology Co., Ltd.), and then centrifuged at 13,680 x g at 4˚C for 30 min to harvest the supernatant. The protein concentration in the supernatant was measured using a BCA kit (cat. no. P0011; Beyotime Institute of Biotechnology). Then, equal amounts of protein (32 µg/lane) were separated using SDS-PAGE (10% gel) and subsequently transferred to a PVDF membrane. The PVDF membrane was blocked with 5% skimmed milk in Tris-buffered saline Tween 20 (0.1%) buffer for 1 h at room temperature, then incubated with primary antibodies at 4^˚C overnight. The primary antibodies comprised: Rabbit anti-GAPDH (1:1,000; cat. no. 2118S; CST Biological Reagents Co., Ltd.), mouse anti-CYP7A1 (1:1,000; cat. no. 2683295; MilliporeSigma), rabbit anti-hydroxy-δ-5-steroid dehydrogenase, 3 β- and steroid δ-isomerase 7 (HSD3B7; 1:1,000; cat. no. ab190223; Abcam), rabbit anti-aldo-keto reductase family 1 member D1 (AKR1D1; 1:1,000; cat. no. ab101393; Abcam), rabbit anti-acyl-CoA oxidase 2 (1:1,000; cat. no. ab197808; Abcam), rabbit anti-hydroxysteroid 17-β dehydrogenase 4 (HSD17B4; 1:1,000; cat. no. ab97971; Abcam), rabbit anti-CYP27A1 (1:2,000; cat. no. ab126785; Abcam), rabbit anti-CYP39A1 (1:1,000; cat. no. ab129334; Abcam), rabbit anti-isopentenyl-diphosphate δ isomerase 1 (IDI1; 1:3,000; cat. no. ab97448; Abcam), rabbit anti-lanosterol synthase [anti-oxidosqualene-lanosterol cyclase (OSC); 1:1,000; cat. no. ab80364; Abcam], rabbit anti-farnesyl diphosphate farnesyltransferase 1 (FDFT1; 1:1,000; cat. no. ab195046; Abcam), rabbit anti-farnesyl diphosphate synthase (FDPS; 1:2,000; cat. no. ab153805; Abcam), rabbit anti-mevalonate diphosphate decarboxylase (MVD; 1:1,000; cat. no. ab96226; Abcam), rabbit anti-mevalonate kinase (MVK; 1:1,000; cat. no. ab154515; Abcam), rabbit anti-3-hydroxy-3-methylglutaryl-CoA reductase (HMGCR; 1:1,000; cat. no. ab174830; Abcam), rabbit anti-low density lipoprotein receptor (LDLR; 1:1,000; cat. no. ab52818; Abcam), rabbit anti-c-Jun N-terminal kinase (JNK; 1:1,000; cat. no. 9252s; CST Biological Reagents Co., Ltd.), mouse anti-phosphorylated (p)-JNK (1:2,000; cat. no. 9255s; CST Biological Reagents Co., Ltd.), rabbit anti-p-insulin receptor substrate 1 (p-IRS1; Tyr608) mouse (1:1,000; cat. no. 09-432; MilliporeSigma), rabbit anti-IRS-1 (1:1,000; cat. no. 2382S; CST Biological Reagents Co., Ltd.), rabbit anti-p-AKT (Ser473) (1:1,000; cat. no. 4060S; CST Biological Reagents Co., Ltd.), rabbit anti-AKT (1:1,000; cat. no. 4685S; CST Biological Reagents Co., Ltd.), mouse anti-tribbles pseudokinase 3 (TRB3; 1:100; cat. no. sc-390242; Santa Cruz Biotechnology, Inc.), rabbit anti-c-Jun (1:1,000; cat. no. ab32137; Abcam), mouse anti-p-c-Jun (1:200; cat. no. sc-822; Santa Cruz Biotechnology, Inc.) and goat anti-KLB (0.5 µg/ml; cat. no. AF5889; R&D Systems China Co., Ltd.). Then, the PVDF membrane was washed in Tris-buffered saline Tween 20 (0.1%) buffer for 1 h at room temperature and subsequently incubated with horseradish peroxidase (HRP)-labeled secondary antibodies, including HRP-goat anti-rabbit IgG (1:5,000; cat. no. os0701; Earthox Life Sciences), HRP-donkey anti-goat IgG (1:2,000; cat. no. ab6885; Abcam) and HRP-goat anti-mouse IgG (1:2,000; cat. no. ab6789; Abcam) at room temperature for 1 h. Finally, the signals were detected using Enhanced Chemiluminescence ECL solution (cat. no. MA0186-2; Meilunbio), and quantification of the bands was conducted using ImageJ software (version 1.53a; National Institutes of Health).

Fecal samples collected the day before tissue harvesting were flash frozen in liquid nitrogen after collection from each mouse and stored at -80˚C until use. The extraction of bacterial DNA from the fecal samples, the PCR amplification of bacterial 16S rDNA genes, sequencing and analysis were conducted by Guangzhou Gene Denovo Biotechnology Co., Ltd. All experimental procedures were performed as previously described (25).

Statistical differences were determined using SPSS software (version 23.0; IBM Corp.). Data are presented as the mean ± SEM. One-way ANOVA followed by Tukey's post hoc test was conducted to analyze differences among the groups. P<0.05 was considered to indicate a statistically significant difference.

To evaluate the effects of sodium sulphate on hypercholesterolemia, 8-week-old male C57BL/6 mice were fed an HCD for 4 weeks, and then the TC concentration in the mice serum was determined (Fig. S1A). After 4 weeks of feeding with the HCD, the serum TC concentration was significantly higher compared with that of the mice in the CON group (Fig. S1B). The mice fed an HCD were then divided into four groups, with matching of the serum TC concentration in each group. Then, mice in the HCD + LSS, HCD + MSS and HCD + HSS groups were treated with sodium sulphate for 3 weeks (Fig. S1A). The body weight and food intake of mice in the CON, HCD, HCD + LSS, HCD + MSS and HCD + HSS groups exhibited no significant differences during these 3 weeks (Fig. S1C and D). Although the average daily defecation mass of mice fed an HCD was lower than that of the CON group, it was higher in the sodium sulphate treated mice compared with the CON and HCD groups (Fig. S1E). However, these differences in defecation mass were not found to be significant, and the fecal pellets of the mice in the sodium sulphate groups were normal.

Although the serum TC concentration of the mice in the HCD group was significantly higher than that in the CON group, administration of sodium sulphate significantly reduced the serum TC concentration in mice fed an HCD (Fig. 1A). However, the serum TG concentration of the mice in each group was similar (Fig. 1B). The TC level in the liver was significantly increased in the mice from the HCD group compared with the CON group (Fig. 1C). In addition, the TC level in the liver exhibited a downward trend in the mice from the HCD + LSS, HCD + MSS and HCD + HSS groups compared with the HCD group, but this reduction was not significant (Fig. 1C). The TG level in the liver was significantly higher in mice fed an HCD compared with those fed an NFD, and the administration of sodium sulphate did not reduce it (Fig. 1D). The serum LDL-C concentration did not significant differ among the mice of the five groups (Fig. 1E). However, the serum HDL-C concentrations in the mice from the HCD + LSS, HCD + MSS and HCD + HSS groups were significantly higher than those in the mice from the HCD group (Fig. 1F). In addition, the LDL-C/HDL-C ratio was significantly reduced in the serum of the mice from the HCD + LSS and HCD + HSS groups compared with the HCD group (Fig. 1G). The serum TBA concentration in the mice from the HCD group was significantly reduced compared that in the mice from the CON group, and the serum TBA concentrations in the HCD + LSS, HCD + MSS and HCD + HSS groups were slightly but not significantly lower than the serum TBA concentration in the HCD group (Fig. 1H). In addition, when the serum ALT and AST concentrations were compared among the mice in all groups, no significant differences were detected (Fig. 1I and J). These results indicate that sodium sulphate may ameliorate the hypercholesterolemia in mice induced by an HCD.

Since the liver is the major organ involved in cholesterol metabolism (26), the morphology and histology of the livers of the mice in the five groups were assessed. The morphology of the livers in mice fed with an HCD with or without sodium sulphate appeared slightly paler compared with those in the CON group (Fig. S2A). The gallbladders from mice in the HCD group were dark red, and the administration of sodium sulphate improved the colour and increased the size of the gallbladders from these mice (Fig. S2A). The liver weight of mice in the HCD group was significantly reduced compared with that in the CON group, although the reduction was slight, and the liver weights of mice in the sodium sulphate administration groups exhibited no significant differences compared with the HCD group (Fig. S2B). The liver/body weight ratio of mice fed an HCD was significantly lower than that of mice fed an NCD, and the administration of sodium sulphate did not increase the liver/body weight ratio of these mice (Fig. S2C). The results of H&E staining demonstrated that excessive quantities of lipid droplets accumulated in the liver tissues of mice from the HCD group, and the haematoxylin staining intensity of the nuclei in certain hepatocytes of this group was weaker (Fig. S2D). The weak nuclear staining indicated that the hepatocytes were damaged (27). Following the administration of sodium sulphate to mice fed an HCD, excessive lipid accumulation was still detected in the hepatocytes, but the nuclear staining in most hepatocytes of these mice was normal (Fig. S2D). These results demonstrated that, although sodium sulphate did not mitigate non-alcoholic fatty liver disease (NAFLD) in mice fed an HCD, it protected hepatocytes against HCD-induced damage.

RNA-sequencing (seq) technology was used to compare the transcriptional profiles of hepatic tissues from mice in the CON, HCD and HCD + MSS groups. There were 148 upregulated and 230 downregulated genes in the hepatic tissues from the HCD group compared with the CON group (Fig. S3A and B). The mRNA expression levels of 116 genes were significantly increased and those of 127 genes were significantly decreased in the hepatic tissues from the HCD + MSS group compared with the HCD group (Fig. S3A and C). The expression levels of 203 genes were higher and those of 286 genes were lower in the hepatic tissues from the HCD + MSS group compared with the CON group (Fig. S3A and D). The RNA-seq results demonstrated that genes associated with bile acid synthesis, such as Cyp7a1, were upregulated in the livers of the HCD + MSS group compared with the HCD group (Fig. 2A). Since bile acid transporters play important roles in the regulation of bile acid metabolism (28), the transcriptional levels of genes encoding bile acid transporters in the liver were also analysed. The RNA-seq results demonstrated that the mRNA level of solute carrier organic anion transporter family member 1b2 (Slco1b2) was notably reduced in the hepatic tissues from the HCD group compared with the CON group, but there was no significant change in Slco1b2 in the HCD + MSS group compared with the HCD group (Fig. S3E).

The expression levels of genes associated with bile acid synthesis were further tested via RT-qPCR and western blotting. Cyp7a1 expression was increased at the mRNA level in the hepatic tissues from the HCD group compared with the CON group (Fig. 2B). However, the CYP7A1 protein expression level was slightly lower in the livers from the HCD group compared with the CON group, although this reduction was not significant (Fig. 2C and D). The Cyp7a1 mRNA levels were significantly higher in the livers from the HCD + MSS and HCD + HSS groups compared with the HCD group (Fig. 2B), and the CYP7A1 protein levels were significantly increased in the liver tissues from the HCD + LSS and HCD + MSS groups compared with the HCD group (Fig. 2C and D). The mRNA levels of Akr1d1 in the HCD + MSS group and Cyp27a1 in the HCD + LSS group were significantly upregulated compared with those in the HCD group (Fig. 2B), but no marked change in the AKR1D1 and CYP27A1 protein levels was observed among the five groups (Fig. 2C). The mRNA and protein expression levels of CYP39A1 were significantly reduced in the hepatic tissues from the HCD group compared with the CON group (Fig. 2B-D). In addition, the hepatic Cyp39a1 mRNA expression levels were significantly upregulated in the HCD + LSS group compared with the HCD group (Fig. 2B), and the CYP39A1 protein expression level was significantly higher in the hepatic tissues from mice in the HCD + MSS group compared with the HCD group (Fig. 2C and D). No significant difference in the Ldlr mRNA level was detected between the hepatic tissues of the HCD and CON groups (Fig. 2E). However, the Ldlr transcriptional level was significantly higher in the hepatic tissues from mice in the HCD + MSS group compared with the HCD group (Fig. 2E). The Ldlr mRNA levels in the livers of mice from the HCD + LSS and HCD + HSS groups were also increased compared with the Ldlr mRNA level in the CON group, but this increase was not significant (Fig. 2E). The western blotting results demonstrated that the LDLR protein level in the hepatic tissues of mice from the HCD + LSS group was also significantly increased compared with that in the HCD group (Fig. 2F). These results suggest that sodium sulphate might alleviate hypercholesterolemia in mice fed an HCD via upregulation of the expression of genes that encode enzymes that catalyze bile acid production, including Cyp7a1 and Cyp39a1, and upregulation of the expression of Ldlr, which takes up excessive LDL-C from the blood into the hepatocytes of mice.

The hepatic expression levels of genes associated with cholesterol synthesis were also assessed. The mRNA expression levels of 3-hydroxy-3-methylglutaryl-CoA synthase 1, Hmgcr, Mvk, Mvd, Idi1, Fdps, Fdft1, squalene epoxidase, Osc, 7-dehydrocholesterol reductase and Cyp51a1 in hepatic tissues from the HCD group were significantly reduced 2 to >10-fold compared with those in the CON group (Fig. S4A), and the levels of most of these mRNAs were slightly increased after sodium sulphate administration, some significantly, namely Mvd, Idi1, Fdps and Cyp51a1 (Fig. S4A). However, the Fdft1 mRNA levels in the hepatic tissues of mice from the HCD + MSS and HCD + HSS groups were markedly lower compared with those in the HCD group (Fig. S4A). At the protein level, the expression of HMGCR, the rate-limiting enzyme of cholesterol biosynthesis (29,30), was not significantly different among the five groups. However, the MVK, MVD, IDI1, FDPS, FDFT1 and LSS levels were all notably downregulated in the hepatic tissues from the HCD group compared with the CON group, and were not affected by sodium sulphate administration (Fig. S4B).

The RNA-seq results demonstrated that the expression of Trib3, the mRNA that encodes TRB3, was significantly reduced in the livers of mice from the HCD + MSS group compared with the HCD group (Fig. 2A). Increased Trib3 expression is known to cause insulin resistance in hepatocytes in vivo and in vitro (31-34). KEGG analysis of the mRNAs in the livers of the mice that were differentially expressed between the HCD and the HCD + MSS groups indicated that ‘Insulin signaling pathway’ and ‘Insulin resistance’ were among the top 10 pathways that affected the HCD group following sodium sulphate administration (Fig. 3A). The results of RT-qPCR and western blotting analysis demonstrated that Trib3 expression was significantly increased in the hepatic tissues of mice from the HCD group compared with the CON group, at the mRNA and protein levels (Fig. 3B). The Trib3 mRNA levels were significantly reduced in the hepatic tissues of mice fed an HCD following the administration of sodium sulphate (Fig. 3B). In addition, the TRB3 protein levels were significantly lower in the hepatic tissues of mice from the HCD + LSS and HCD + MSS groups compared with the HCD group (Fig. 3B).

Furthermore, the activation and protein levels of components of the insulin signal pathways were evaluated by the western blotting. The p-IRS1 (Tyr608) protein levels in the hepatic tissues from the HCD + LSS, HCD + MSS and HCD + HSS groups were higher compared with those in the CON and HCD groups (Fig. 3C). The p-AKT levels in the hepatic tissues of mice from the HCD group were lower compared with those in the CON group, and the administration of sodium sulphate to HCD-fed mice increased these levels (Fig. 3C). The ratio of p-AKT/AKT was significantly lower in liver tissue from the HCD group compared with the CON group, and was significantly increased in the HCD + MSS and HCD + HSS groups compared with the HCD group (Fig. 3C). These results indicate that sodium sulphate attenuated the hepatic insulin resistance in mice fed an HCD by inhibiting the expression of Trib3 in hepatocytes.

A number of studies have confirmed that the reduction of FGF15 expression in the ileum can induce the expression of Cyp7a1 in hepatocytes to increase the conversion of cholesterol to bile acid (13,35,36). Therefore, the Fgf15 mRNA expression levels in the ileum of mice from the five groups were assessed using RT-qPCR. The Fgf15 transcriptional level in the HCD group was slightly lower than that in the CON group, but the reduction was not significant (Fig. 4A). The Fgf15 mRNA levels were lower in the HCD + LSS, HCD + MSS and HCD + HSS groups compared with the HCD group, but these changes were also not significant (Fig. 4A). However, it is speculated that sodium sulphate may inhibit the expression of FGF15 in the ileum of mice fed an HCD. In the hepatic tissues, the mRNA expression level of Fgfr4, which encodes the receptor for FGF15 (15,37), was significantly reduced in the HCD group compared with the CON group, and was unaffected by the administration of sodium sulphate (Fig. 4B). No significant differences in the mRNA and protein expression levels of KLB, which is the co-receptor of FGFR4(14), were detected in the hepatic tissues from mice in the CON and HCD groups (Fig. 4C and D). However, the Klb mRNA levels were significantly downregulated in the hepatic tissues from mice fed an HCD following the administration of sodium sulphate, and the KLB protein levels were also significantly lower in the hepatic tissues of mice from the HCD + MSS and HCD + HSS groups compared with the HCD group (Fig. 4C and D). These results demonstrate that the FGF15 signal in the hepatocytes of HCD fed mice treated with sodium sulphate might weaken due to the downregulation of FGF15 in the ileum and the reduction of KLB expression in the liver of these mice.

JNK is an important kinase that is activated by FGF15 signaling to inhibit the transcription of Cyp7a1 in hepatocytes (38,39). Activated JNK suppresses the expression of Cyp7a1 via the activation of c-Jun (40). Therefore, the activation and expression levels of JNK and its target, c-Jun, were further assessed using western blotting. The p-JNK/JNK ratio was slightly higher in the hepatic tissues of mice from the HCD group compared with the CON group, but this increase was not significant, and the administration of sodium sulphate reduced this ratio, although this reduction was also not significant (Fig. 4E). The p-c-Jun level was notably increased in the liver tissue of the HCD group compared with the CON group, and was decreased following administration of sodium sulphate (Fig. 4F). These results indicate that sodium sulphate may upregulate Cyp7a1 expression via inhibition of the FGF15/FGFR4-KLB/JNK/c-Jun signaling pathway to promote the conversion of cholesterol to bile acid.

The gut microbiota has been demonstrated to have an important role in regulating the lipid metabolism of the host (41). Therefore, 16S rDNA sequencing was performed to analyze the composition and function of the gut microbiota of mice from the CON, HCD, HCD + LSS, HCD + MSS and HCD + HSS groups. The Shannon rarefaction curves of each group reached the saturation plateau (Fig. 5A), indicating that the sequence coverage of the fecal samples was sufficient to describe the composition of the gut microbiota of all groups. Principal coordinate analysis demonstrated that the five groups were significantly distinguished (Fig. 5B). At the phylum level, the proportional abundance of Bacteroidetes was significantly lower in the HCD group compared with the CON group, and was significantly reduced in the HCD + LSS and HCD + MSS groups compared with the HCD group (Fig. 5C and D). The proportional abundance of Proteobacteria was significantly higher in the HCD group compared with the CON group, and was significantly reduced in the HCD + LSS group compared with the HCD group (Fig. 5C and D). Furthermore, the proportional abundance of Firmicutes was significantly increased in the sodium sulphate treated groups compared with the HCD group (Fig. 5C and D).

The results of the LEFse analysis demonstrated that there were 147 bacterial taxa that differed in abundance between the CON and HCD groups, with 76 predominant in the CON group and 71 predominant in the HCD group (Fig. 6). There were 115 bacterial taxa that differed in abundance between the HCD and HCD + LSS groups, with 41 predominant in the HCD group and 74 predominant in the HCD + LSS group (Fig. 7). There were 91 bacterial taxa that differed in abundance between the HCD and HCD + MSS groups, with 31 predominant in the HCD group and 60 predominant in the HCD + MSS group (Fig. 8). There were 43 bacterial taxa that differed in abundance between the HCD and HCD + HSS groups, with 18 predominant in the HCD group and 25 predominant in the HCD+HSS group (Fig. 9). Compared with the HCD group, there were 13 predominant bacterial taxa in all three groups treated with sodium sulphate, including: Phylum, Firmicutes; order, Clostridiales; class, Clostridia; family, Peptostreptococcaceae; genus, Ruminococcaceae_NK4A214_group; family, Family_XIII; genus, Eubacterium_nodatum_group; species, Ileibacterium_valens; genus, Ileibacterium; genus, Erysipelatoclostridium; genus, Eubacterium_brachy_group; genus, Sphingomonas; and genus, Faecalibaculum (Fig. 7, Fig. 8 and Fig. 9).

The top 20 altered pathways in the KEGG pathway analysis are shown in Fig. 10. These included 19 pathways that were significantly increased in all three sodium sulphate treated groups compared with the CON and HCD groups, including ‘Biosynthesis of ansamycins’, ‘Valine, leucine and isoleucine biosynthesis’, ‘Biosynthesis of vancomycin group antibiotics’, ‘D-Glutamine and D-glutamate metabolism’, ‘Peptidoglycan biosynthesis’, ‘Alanine, aspartate and glutamate metabolism’, ‘One carbon pool by folate’, ‘Pantothenate and CoA biosynthesis’, ‘Streptomycin biosynthesis’, ‘Pentose phosphate pathway’, ‘Carbon fixation in photosynthetic organism’, ‘Aminoacyl-tRNA biosynthesis’, ‘D-Alanine metabolism’, ‘C5-Branched dibasic acid metabolism’, ‘Mismatch repair’, ‘Lysine biosynthesis’, ‘Ribosome’, ‘Fatty acid biosynthesis’ and ‘Cell cycle-Caulobacter’. Most of these pathways were also increased in the HCD group compared to the CON group, with the exception of ‘D-Glutamine and D-glutamate metabolism’, ‘One carbon pool by folate’, ‘Carbon fixation in photosynthetic organism’ and ‘Mismatch repair’. The ‘Other glycan degradation’ pathway was significantly decreased in all three of the sodium sulphate treated groups compared with the CON and HCD groups, and also decreased in the HCD group compared with the CON group.