With the condition that constant temperature is 24 to 26 °C, humidity is 50% to 60%, and light/dark alternates each 12 h, 8-wk-old Wistar rats, provided by the Laboratory Animal Center of Shandong University, were separately housed in independently ventilated cages. Then, we picked 7 from 30 male rats as the control group, which were given free access to standard chow (14% calories from fat; Shandong University Laboratory Animal Center) till the study ended. In order to induce insulin resistance, we fed the remaining rats for 4 wk with a HFD (Huafukang Biotech Company, China), which contained 40% fat as calories. A 12-h fasting period was succeeded by intraperitoneal injection with 2% streptozotocin (STZ) (Sigma, United States), at the dose of 35 mg/kg weight, so the rats reached a diabetic state. After 2 wk, 21 rats met the diabetic criterion - a blood glucose level of at least 16.7 mmol/L during oral glucose tolerance test (OGTT) - and were randomly divided into two groups receiving sham (n = 7) or DJB operation (n = 14), in which rats were fed HFD postoperatively. Twelve wk after surgery, the DJB group was subdivided into the DJB-recurrence group (DJB-RC, 5 rats, defined as having blood glucose ≥ 16.7 mmol/L during OGTT) and the DJB-remission group (DJB-RM group, 6 rats).

Body weight and calorie content in the food intake (calorie intake) were measured at baseline, 4, 8 and 12 wk postoperatively. Finally, all rats were euthanized by chloral hydrate overdose (intraperitoneal injection, 15 mL/kg, 10% chloral hydrate) for tissue collection. All animal experimental procedures involved in our study were approved by the Animal Care and Utilization Committee of Qilu Hospital of Shandong University, Jinan, China.

Fifteen days after induction of diabetes, a low-residue diet was administered to the rats in the sham group and DJB group from 48 h before surgery to 72 h after surgery. We performed DJB or sham surgery on rats under anesthesia (10% chloral hydrate at the dose of 3 mL/kg). All of the surgeries were completed within 3 d. Seventy-two postoperative hours, all rats in both groups were allowed access to the HFD and water ad libitum.

DJB: We transected the duodenum just at the site of connection with the pylorus, and 7-0 silk sutures (Ningbo Medical Needle, China) were applied in the closure of the duodenum distal end, followed by transecting jejunum 15 cm at the site of distal to the Treitz’s ligament. Then, we proceeded anastomose of the proximal end of the duodenum and distal jejunum (duodenojejunal anastomosis). Moreover, we anastomosed jejunum proximal end to the antimesenteric border of the alimentary limb 15 cm distal to the duodenojejunal anastomosis, to form an end-to-side anastomosis.

Sham: We transected the intestines at the same sites with those in enterotomies of DJB, and then carried out re-anastomosis in situ. Because of similar stress from surgery and anesthesia, the duration of the sham operation was similar to DJB.

We conducted the oral glucose tolerance test (OGTT) at several time points, including baseline and postoperative wk 4 and 12. After completion of 8-hr fasting, all rats received 1 g/kg glucose by oral gavage. Then, we estimated the levels of blood glucose at six time points (baseline, 15, 30, 60, 90 and 120 min administration) respectively.

During the OGTT at 4 and 12 wk postoperatively, we respectively gathered blood samples by the retrobulbar venous plexus approach at time points of baseline, 15, 30, 60 and 120 min after gavage with glucose, to collect serum by centrifugation (1006 ×g, 4 °C, 15 min) and stored at -80 °C for further measurement. Concentrations of insulin and GLP-1 in serum and LPS in fasting serum were tested by enzyme-linked immune sorbent assay (ELISA) kits from Millipore (United States), kits from Uscn Life Science Inc. (China), and a rat LPS ELISA kit from Bio-Swamp (China) respectively. Levels of TBAs in fasting serum were detected by a Hitachi automatic biochemical analyzer (Japan).

The calculations of homeostasis model assessment of insulin resistance (HOMA-IR) at 4 and 12 wk postoperatively were performed in order to evaluate insulin resistance, using this formula: HOMA-IR=fasting insulin (mIU/L) × fasting glucose (mmol/L)/22.5[24].

Because host metabolism can be influenced by the biological activity of microbiota in the colon, we focused our study on the colonic microbiota. At 12 wk after surgery, all rats were narcotized with 10% chloral hydrate (3 mL/kg). A 3-cm segment of colon was ligatured and removed from the enterocoelia. The colon was incised longitudinally, and then we gathered the colonic contents and stored them at -80 °C. All operations were performed under aseptic conditions. Genomic DNA of colonic microbiota was isolated based on the protocol of an E.Z.N.A. Soil DNA kit (Omega, United States). The method to amplify the V4 region of microbial 16S rDNA g by PCR has been reported previously[25]. We used an Illumina MiSeq platform (BGI Technology, China) to sequence the amplified V4 region. The raw data were filtered to eliminate the adapter pollution and low quality reads so as to obtain clean reads. Then, paired-end reads with overlap were merged to tags, which were clustered, at 97% sequence similarity, to operational taxonomic units (OTUs). By using the Ribosomal Database Project (RDP) Na, Bayesian Classifier v.2.2, we assigned taxonomic ranks to OTU representative sequence. At last, the different species screening tests were analyzed based on OTU and taxonomic ranks.

All quantitative data were presented as mean ± SD. By the use of trapezoidal integration, we calculated the area under the curves for OGTT (AUC_OGTT). Body weight, calorie intake, AUC_OGTT, HOMA-IR, TBAs and LPS were evaluated with the use of one-way analysis of variance (ANOVA) followed by Bonferroni post hoc comparison. All the data of the insulin and total GLP-1 concentrations were analyzed by mixed model ANOVA followed by Bonferroni post hoc comparison after glucose administered by gavage to rats. All statistical calculations were processed with the use of SPSS version 19.0 (IBM, United States), at an alpha level of 0.05.

The information of the 16S rDNA-based study involving gut microbiota was analyzed by BGI Technology. Package “ade4” of software R (v3.0.3) was used in principal component analysis (PCA). Heatmap was generated using the package “gplots” of software R (v3.0.3). The relative abundance of specific microbes between groups was analyzed by the Wilcoxon test.

Compared with rats in the other three groups at any time points, the rats in the control group had markedly lower body weight and calorie intake (P < 0.05) (Figure 1A and B). The differences of other groups were not significant.

Open in New Tab   Full Size Figure   Download Figure

Figure 1 Body weight and calorie content in the food intake (calorie intake). Shown are body weight (A) and calorie intake (B) of rats at baseline, 4, 8 and 12 wk after surgery. ^aP < 0.05 vs control group.

As Figure 2C shows, compared with the other three groups, preoperative AUC_OGTT of the control group (797.04 ± 81.01) was statistically lower (P < 0.05 vs sham 1913.46 ± 184.55; P < 0.05 vs DJB-RC 1869.90 ± 216.37; P < 0.05 vs DJB-RM 1935.88 ± 250.73), with no differences among the other three groups. The postoperative AUC_OGTT of the control group remained unchanged. Compared with the sham group (4 wk after surgery: 1935.88 ± 250.73; 12 wk after surgery: 1985.57 ± 152.56), both the DJB-RM (4 wk after surgery: 920.88 ± 62.90, P < 0.05 vs sham; 12 wk after surgery: 1009.86 ± 119.90, P < 0.05 vs sham) and control (4 wk after surgery: 827.86 ± 63.83,P < 0.05 vs sham; 12 wk after surgery: 895.39 ± 44.80, P < 0.05 vs sham) groups showed lower AUC_OGTT at all postoperative time points, indicating a deterioration in diabetes in sham rats. Notably, the AUC_OGTT in the DJB-RC group (890.85 ± 114.41) was comparable to that in the DJB-RM group (920.88 ± 62.90, P > 0.05 vs DJB-RC) 4 wk postoperatively, which was higher when compared with the DJB-RM group 12 wk postoperatively (P < 0.05 DJB-RC vs DJB-RM, 1876.20 ± 178.48 vs 1009.86 ± 119.90), demonstrating that the improved glucose tolerance in the DJB-RC group was reversed.

Open in New Tab   Full Size Figure   Download Figure

Figure 2 Serum insulin after oral glucose administration. Shown are levels of serum insulin after oral glucose gavage (1 g/kg) at 4 wk (A) and 12 wk (B) after surgery, between which there are no significant differences with the use of mixed model one-way analysis of variance followed by Bonferroni post hoc comparisons. Area under the curves for oral glucose tolerance test and homeostasis model assessment of insulin resistance are demonstrated in C and D respectively. ^aP < 0.05 vs control group; ^cP < 0.05 vs sham group; ^eP < 0.05 vs DJB-RC group. DJB-RC: Duodenal-jejunal bypass recurrence group; DJB-RM: Duodenal-jejunal bypass-remission group.

Figure 2A and B shows the postoperative curves of serum insulin levels during the OGTT. It can be seen that no statistically difference was observed in serum insulin between the groups 4 wk and 12 wk postoperatively.

Postoperative HOMA-IR is shown in Figure 2D. Four weeks after surgery, compared with the other three groups, HOMA-IR in the sham group (5.18 ± 0.98) was higher (P < 0.05 vs control 2.08 ± 0.38; P < 0.05 vs DJB-RC 2.30 ± 0.46; P < 0.05 vs DJB-RM, 2.17 ± 0.60), and no difference was observed among the control group, DJB-RC group and DJB-RM group. Consistent with the results of OGTT, 12 wk after surgery, the DJB-RC group (4.23 ± 0.72) exhibited higher HOMA-IR values than the DJB-RM group (2.82 ± 0.73, P < 0.05 vs DJB-RC), indicating re-impaired insulin sensitivity in the DJB-RC group 12 wk after surgery.

At both 4 wk (Figure 3A) and 12 wk (Figure 3B) after surgery, total serum GLP-1 secretion was comparable among the control group, DJB-RM group and DJB-RC group (P > 0.05), and rats in the sham group showed a lower total serum GLP-1 secretion (P < 0.05 vs control; P < 0.05 vs DJB-RC; P < 0.05 vs DJB-RM).

Open in New Tab   Full Size Figure   Download Figure

Figure 3 Level of glucagon-like peptide 1 after administration of oral glucose. Shown are glucagon-like peptide 1 (GLP-1) level after oral glucose gavage (1 g/kg) at 4 wk (panel A) and 12 wk (panel B) after surgery. In terms of global GLP-1 concentration, the illustrations in the rectangles show that there are significant differences between the groups with the use of mixed model one-way analysis of variance followed by Bonferroni post hoc comparisons. DJB-RC: Duodenal-jejunal bypass recurrence group; DJB-RM: Duodenal-jejunal bypass-remission group.

As shown in Figure 4A, at both 4 wk and 12 wk after surgery, TBAs levels were higher in the DJB-RC (4 wk after surgery: 21161.60 ± 2550.10 ng/mL, P < 0.05 vs control, P < 0.05 vs sham; 12 wk after surgery: 18744.00 ± 3047.26 ng/mL, P < 0.05 vs control, P < 0.05 vs sham) and DJB-RM (4 wk after surgery: 19543.00 ± 2639.35 ng/mL, P < 0.05 vs control, P < 0.05 vs sham; 12 wk after surgery: 27803.17 ± 4673.42 ng/mL, P < 0.05 vs control, P < 0.05 vs sham) groups than in the control (4 wk after surgery: 11608.00 ± 1248.10 ng/mL; 12 wk after surgery: 13001.86 ± 1613.55 ng/mL) and sham (4 wk after surgery: 13190.14 ± 1237.38 ng/mL; 12 wk after surgery: 12064.86 ± 1809.36 ng/mL) groups. Notably, at 12 wk postoperatively, rats in the DJB-RM group showed higher TBAs than rats in the DJB-RC group (P < 0.05).

Open in New Tab   Full Size Figure   Download Figure

Figure 4 Fasting serum total bile acids and lipopolysaccharide. Fasting serum TBAs (A) and LPS (B) were measured at 4 and 12 wk after surgery. ^aP < 0.05 vs control group; ^cP < 0.05 vs sham group; ^eP < 0.05 vs DJB-RC group. DJB-RC: Duodenal-jejunal bypass recurrence group; DJB-RM: Duodenal-jejunal bypass-remission group; TBAs: total bile acids; LPS: Lipopolysaccharide.

As shown in Figure 4B, 4 wk after surgery, levels of fasting serum LPS in the sham group (0.20 ± 0.03 ng/mL) were higher compared with the other three groups (P < 0.05 vs control 0.10 ± 0.02 ng/mL; P < 0.05 vs DJB-RC 0.13 ± 0.02 ng/mL; P < 0.05 vs DJB-RM, 0.12 ± 0.04 ng/mL). Twelve weeks after surgery, the sham (0.22 ± 0.04 ng/mL) and DJB-RC (0.19 ± 0.03 ng/mL) groups exhibited similar levels of LPS, but higher levels than the control group (0.11 ± 0.02 ng/mL; P < 0.05 vs sham; P < 0.05 vs DJB-RC) and the DJB-RM (0.12 ± 0.04 ng/mL; P < 0.05 vs sham; P < 0.05 vs DJB-RC) group.

As demonstrated by PCA (Figure 5A), the gut microbiota showed marked changes in the four groups, which indicated a relative-centralized tendency intra-group and a relative-dispersed distribution inter-group. These tendencies were also observed in the Heatmap analysis at the phylum level (Figure 5B).

Open in New Tab   Full Size Figure   Download Figure

Figure 5 Principal component analysis and heatmap analysis. A: Principal component analysis was used to construct a 2-D graph to summarize the factors mainly responsible for this difference. Similarity was high when two samples were closely located. A number in brackets represents contributions of principal components to differences among samples; B: Heat map analysis: The longitudinal clustering indicates the similarity of all species among different samples, and horizontal clustering indicates the similarity of certain species among different samples, the closer the distance and the shorter the branch length, the more similar the species composition is between the samples.

The taxonomic composition distribution histograms, at the phylum level, in each group of rats are shown in Figure 6A. Bacteroidetes were the predominant gut microbes in the sham (46.88% ± 2.34%) and DJB-RC (47.41% ± 5.67%) groups (Figure 6B), while Firmicutes were predominant in the control (58.06% ± 11.12%) and DJB-RM groups (55.58% ± 6.16%; Figure 6C). The relative abundance of Proteobacteria was higher in the sham (24.14% ± 2.89%) and DJB-RC (21.83% ± 5.09%) groups than in the control group (6.86% ± 3.79%; P < 0.05 vs sham; P < 0.05 vs DJB-RC) and the DJB-RM group (8.31% ± 4.39%; P < 0.05 vs sham; P < 0.05 vs DJB-RC; Figure 6D).

Open in New Tab   Full Size Figure   Download Figure

Figure 6 Species annotation. Species annotation (A) is the taxonomic composition distribution histogram of each sample at the phylum level. The ratios of each phylum in certain samples are displayed. The relative abundance of Bacteroidetes (B), Firmicutes (C), Proteobacteria (D) and Escherichia coli (E) between the groups was analyzed by the Wilcoxon test. ^aP < 0.05 vs control group; ^cP < 0.05 vs sham group; ^eP < 0.05 vs DJB-RC group. DJB-RC: Duodenal-jejunal bypass recurrence group; DJB-RM: Duodenal-jejunal bypass-remission group.

It is noteworthy that although resolution at the species level was low, the relative abundance of Escherichia coli (E. coli) in the sham group (15.72% ± 1.67%) and DJB-RC group (16.42% ± 3.00%) was remarkably higher than that in the control group (3.58% ± 3.67%; P < 0.05 vs sham; P < 0.05 vs DJB-RC) and the DJB-RM group (4.15% ± 2.76%; P < 0.05 vs sham; P < 0.05 vs DJB-RC; Figure 6E).

Bariatric surgery can contribute to remission of type 2 diabetes mellitus. Some patients, however, experience diabetes again postoperatively. Several factors have been found to be closely related to diabetes recurrence, including percentage of excess body weight loss, body mass index before operation, age, duration and severity of diabetes, weight regain, postoperative diet and lifestyle. The intrinsic mechanism of diabetes postoperative recurrence, however, remains unclear.

Not only can gut microbiota modulate host metabolism, but it also exerts an influence on diabetes postoperative remission. Alterations in gut microbiota after bariatric surgery may be responsible for diabetes remission possibly by influencing serum levels of lipopolysaccharide (LPS) and total bile acids (TBAs).

This study created the high-fat diet-induced rat models of diabetes recurrence after duodenal-jejunal bypass (DJB) and determined the alterations in gut microbiota during diabetes recurrence after DJB. The rats with diabetes recurrence presented the phenomena with a reduced relative abundance of Firmicutes and an increased Bacteroidetes and Escherichia coli. Alterations in gut microbiota may be responsible for the diabetes postoperative recurrence, possibly by influencing serum LPS and TBAs.

The findings in this study could enable investigators to focus more on the link between gut microbiota and diabetes postoperative recurrence. A new design of therapeutic interventions aiming to target gut microbiota may prevent diabetes postoperative recurrence.

DJB, serving as an experimental procedure, was devised to estimate the anti-diabetic effects independent from weight related to Roux-en-Y gastric surgery, which is the most effective treatment for diabetes.

This is a very well designed, performed and written experimental study for investigation of the role of alterations in gut microbiota in the pathogenesis and remission of type 2 diabetes after bariatric surgery. For investigation of this aim the authors created and used a rat model of diabetes recurrence after DJB and compared the results obtained in diabetes rats induced by high-fat diet after bypass and in the group with sham operation.