Alterations in the gut microbiome and metabolome profiles of septic rats treated with aminophylline

The experimental animals were Sprague–Dawley rats (6–8 weeks old, all males) from Beijing Viton Lever Experimental Co., Ltd. The rats were acclimatized to the environment and provided with sufficient food and water before being randomly divided into 3 groups (n = 20/group): the sham operation (SC) group, the sepsis model (CLP) group, and the sepsis + aminophylline administration (Amino) group. Sepsis was induced by cecal ligation and puncture method, and which is based on our previous research [19]. Briefly, the rat abdomen was shaved and thoroughly cleaned with complex iodine following intraperitoneal injection of 10% chloral hydrate (350 mg/kg), and surgeries were conducted on a sterilized board. Approximately 2.0 incisions were made at the midline of the abdomen. The cecum was squeezed until a small amount of fecal material was excreted and then returned to the peritoneal cavity. The abdominal cavity was then sutured with aseptic 5 + 0 surgical sutures, and the skin was closed using a septic 3 ± 0 surgical sutures. After the operation, all rats were immediately placed in a warm environment and subcutaneously injected with preheated physiological saline (50 mL/kg) for fluid resuscitation. Rats in the Amino group were additionally injected with 50 mg/kg aminophylline intraperitoneally at 1 h after surgery, which were based on our pre-experimental survival results. Rats in the SC group underwent the same surgical operation as the CLP and Amino groups, except that the caecum was not ligated or punctured. The sepsis model for rats in the CLP and Amino groups was constructed with reference to previous literature [20]. The survival status of the rats in each group was recorded at 24 h postoperatively, after which the faeces were collected and stored at − 80 °C for freezing.

Faecal DNA was extracted by the CTAB method; the concentration was detected by agarose gel electrophoresis and diluted to 1 ng/µl with sterile water. Subsequently, PCR amplification of the V3-V4 region of the 16S rRNA gene was performed using the diluted DNA as a template. Finally, 16S rRNA sequencing was performed using NovaSeq6000.

Raw tags were obtained by stitching together the reads of each sample using FLASH software [21] and filtered using QIIME (V1.9.1) to obtain clean tags [22, 23]. Afterwards, the final effective tags were obtained by (https://​github.​com/​torognes/​vsearch/​) matching with the species annotation database and removing chimeric sequences. All effective tags were clustered using Uparse (v7.0.1001), and sequences with a similarity threshold above 97% were assigned to operational taxonomic units (OTUs) by default. Species annotation was performed with the SSUrRNA database [24] using the Mothur method (set threshold of 0.8–1). The alpha and beta diversity indices were calculated using QIIME software, and LEfSe analysis was performed using LEfSe software with the default setting of a screening value of 4 for the LDA score. Other plots, such as rarefaction curves, PCoA plots, and plots of significant differences in species between groups, were created in R (v2.15.3).

Twenty-five milligrams of stool was placed into an Eppendorf (EP) tube, and 500 µl of extraction solution (methanol: acetonitrile: water = 2:2:1 (V/V), containing isotopically labelled internal standard mixture) was added. The solution was mixed well and allowed to stand at − 40 °C for 1 h before centrifugation for 15 min (4 °C, 12,000 rpm). The supernatant was collected in an injection vial for the assay.

The target compounds were chromatographically separated on a Waters ACQUITY UPLC BEH Amide (2.1 mm × 100 mm, 1.7 μm) liquid chromatographic column using a Vanquish (Thermo Fisher Scientific) ultra-performance liquid chromatograph. The A phase of the liquid chromatography was aqueous, containing 25 mmol/L ammonium acetate and 25 mmol/L ammonia, and the B phase was acetonitrile. The sample tray temperature was 4 °C, with an injection volume of 2 μL. Finally, primary and secondary mass spectrometry data acquisition was performed by a Q Exactive HF-X (Thermo Fisher Scientific) mass spectrometer.

The raw data were first converted to mzXML format by Proteo Wizard software, and peak identification, peak extraction, peak alignment and integration were performed using an in-house R package (kernel XCMS) [25] before data were matched with the BiotreeDB (V2.1) self-built secondary mass spectrometry database for substance annotation (cut-off value = 0.3). After data management [26] of the raw data, univariate and multivariate statistical analyses of the qualitative and quantitative results of the metabolic groups were performed. The permutation test was used to evaluate the robustness of the OPLS-DA model, and differential metabolites were defined as metabolites with a variable importance in the projection (VIP) > 1 and P < 0.05. KEGG annotation and pathway enrichment analysis of differential metabolites were conducted in combination with the Kyoto Encyclopedia of Genes and Genomes (KEGG) Pathway database (http://​www.​kegg.​jp/​kegg/​pathway.​html). R software was used to visualize the differential metabolites and pathways.

T tests, Wilcox rank sum tests and Tukey tests were used to analyse whether the species differences between groups were significant. MetaStat analysis was performed at each taxonomic level to obtain p values for the permutation test between groups; p values were corrected using the Benjamini and Hochberg false discovery rate [27] method to obtain q-values. Species with significant differences were screened according to the q-values. Correlation analysis was performed using Pearson correlations and Spearman correlations.

To determine whether aminophylline has a protective effect on septic rats, we first examined mortality rates across the three groups (Additional file 3: Table S1). The results showed a significant increase in 24-h mortality (55%) in CLP-induced septic rats compared to the sham-operated group (P = 0.002); the mortality rate decreased to 30% in the Amino group, although the difference between the CLP and Amino groups was not statistically significant (P = 0.110).

To verify whether aminophylline exerts beneficial effects at least partially through the gut microbiome, we performed 16S rRNA sequencing of the rat faeces. In addition, the sequencing depth was evaluated. Figure 1A shows that the curve flattened as the number of sequences increased, indicating that the results of this sequencing were reasonably plausible. We also generated a rank abundance curve (Fig. 1B), which further confirmed the above point. Then, to compare the overall structural characteristics of the intestinal flora of the three groups, we compared the Shannon index of each group (Fig. 1C) and performed PCoA (Fig. 1D). The CLP group was distant from the SC group, but the Amino and SC groups were clustered together, indicating that the intestinal flora of rats was significantly altered by sepsis, but the flora structure became similar to that of healthy rats after aminophylline administration.

We then focused on bacterial abundance at each taxonomic level, particularly the dominant groups at the phylum and genus levels. Figure 2A–C shows that the Amino and SC groups had similar bacteria as the top ten most abundant phyla, but the CLP group was more variable. For example, Firmicutes and unidentified_Bacteria decreased in abundance in the CLP group, but Proteobacteria and Spirochaeta increased in abundance; interestingly, these changes were restored after aminophylline administration. We then analysed the top30 most abundant bacteria at the genus level (Fig. 2B–D). Similar to the distribution characteristics at the phylum level, the abundance of bacteria such as Lactobacillus and Romboutsia decreased in the CLP group compared to the SC group and increased in the Amino group; Escherichia-Shigella and Allobaculum increased in abundance in the CLP group but decreased to levels similar to those in healthy rats after aminophylline administration. MetaStat analysis was then used to further screen for statistically significant differences in bacteria, and the results showed that the above differences between the above Firmicutes, unidentified_Bacteria, Proteobacteria phylum and lactobacillus, Escherichia-Shigella genera were statistically significant, which empirically demonstrates the beneficial effect of aminophylline on the abnormal intestinal flora of septic rats (Fig. 3A, B). In addition, we performed LEfSe analysis and plotted cladogram (Fig. 3C), which allowed us to determine that CLP group-specific bacteria included Enterobacteriaceae and Bacteroidaceae; these CLP group-specific bacteria were significantly reduced or even disappeared after aminophylline administration.

In general, bacteria usually affect certain specific metabolites and thus the immune status of the host [28, 29]. Therefore, we further analysed the faecal metabolome to look for alterations in septic rats after treatment with aminophylline. First, we determined the total ion chromatograms of the groups (Additional file 1: Fig. S1), and we found that the peak pattern of the Amino group was more similar to that of healthy rats than the peak pattern of the CLP group. We then created a 3D PCA score plot for the three metabolite groups to visualize overall changes in metabolites (Fig. 4A, B). The results showed significant differences between the CLP and SC groups both in positive and negative ion mode, while the Amino group had a significant tendency to retrace, which suggested that aminophylline modulated the metabolites of septic rats to approach those of healthy rats. We also performed OPLS-DA, and similar to the PCA results, both the CLP and SC groups differed significantly under positive and negative ions (Additional file 2: Fig. S2A, B), indicating that the metabolic profile of rats was significantly altered in sepsis. The metabolites after aminophylline administration were also significantly different from those of the CLP group (Fig. 4C, D), suggesting that aminophylline altered the abnormal metabolic profile of septic rats. In addition, we used a volcano plot containing all substances measured in this experiment to visualize the overall distribution of metabolite differences between groups under positive and negative ions (Fig. 4E, F; Additional file 2: Fig. S2C, D). In conclusion, these results suggest that aminophylline alters the metabolites in septic rats to levels similar to those in healthy rats.

For further screening of aminophylline-related metabolites, the value of VIP of the first principal component in OPLS-DA was obtained, which summarized the contribution of each variable to the model. Metabolites with VIP > 1 and P < 0.05 (Student’s t test) were considered to be significantly changed metabolites. Finally, a total of 26 differential metabolites were screened between the Amino and CLP groups (Additional file 4: Table S2). Figure 5A and B show the content of differential metabolites in each sample under positive and negative ions. Maleic acid, isolithocholicacid, oleoyl glycine, leukotriene B4, by ssochlamicacid, 5a-tetrahydrocorticosterone, [10]-dehydrogingerdione, prolyl-hydroxyproline, cortisone, linamarin, 3b-hydroxy-5-cholenoic acid, chenodeoxycholic acid, andtheophyllinewere increased by aminophylline treatment. Pantothenic acid, lysoPE (18:1(9Z)/0:0), pseudouridine, xanthine, beta-D-galactose, uracil, methylsuccinicacid, indole-3-propionic acid, geranylgeranyl-PP, 1-methylguanine, palmitoylethanolamide, and lysoPE (0:0/14:0) were decreased.

In addition, commercial databases, including KEGG (http://​www.​genome.​jp/​kegg/​) and MetaboAnalyst (http://​www.​metaboanalyst.​ca/​), were used for pathway enrichment analysis. The results showed that aminophylline mainly affected metabolic pathways, pyrimidine metabolism, beta-alanine metabolism, pantothenate and CoA biosynthesis, as well as other pathways (Fig. 5C, D).

To investigate whether the altered metabolite abundance induced by aminophylline treatment was correlated with intestinal flora, we performed a Spearman correlation between differential flora and differential metabolites at the genus level in the Amino and CLP groups. We found a significant correlation between 22 differential genera and 26differential metabolites (Fig. 6). Some bacteria enriched in amino acids, such as Lactobacillus, were found to be associated with 5a-tetrahydrocorticosterone and [10]-dehydrogingerdione, while Escherichia-Shigella decreased in abundance after aminophylline treatment and was associated with isolithocholic acid, chenodeoxycholic acid, 3b-hydroxy-5-cholenoic acid, leukotriene B4, xanthine, 5a-tetrahydrocorticosterone, maleic acid, linamarin, [10]-dehydrogingerdione, prolyl-hydroxyproline, pseudouridine, cortisone and pseudouridine.