Hypoglycemic mechanism of intestinal bypass surgery in type 2 diabetic rats

To investigate the effect of duodenal-jejunal bypass (DJB) surgery on postoperative blood glucose in type 2 diabetic rats, and further explore possible mechanisms for the effect of surgical treatment of type 2 diabetes. Forty rats with type 2 diabetes were randomly assigned to 4 groups (n = 10 rats per group), which subsequently underwent DJB, new biliopancreatic diversion (NBPD) or duodenal-jejunal exclusion (DJE) surgery or a sham operation (SHAM). Fasting glucose, 2-h postprandial glucose and blood lipids were measured, and the mRNA in liver and intestinal tissue for bile acid receptor (FXR), as well as the FXR protein expression in the liver tissues were determined. Postprandial blood glucose and fasting TG and FFA in the DJB and NBPD groups were significantly lower than those in the SHAM group and preoperative (p < 0.05) at 8 weeks postoperation. Liver FXR protein was expressed at significantly higher in the DJB and NBPD groups than in the other two (p < 0.05), and the intestinal FXR mRNA in the DJE group were highest. DJB up-regulates the expression of bile acid receptors in the liver and down-regulates those receptors in the intestinal tract via biliopancreatic diversion. This process reduces TG levels, and subsequently any lipotoxicity to islet cells to produce a hypoglycemic effect.

Grouping of type 2 diabetes rat models. The model rats were randomly assigned to 4 groups (10 rats per group).
Surgical procedures. Surgical procedure for NBPD. (1) A 1-cm intestinal segment was removed at the bile-intestine confluence. (2) The proximal end was anastomosed to the distal duodenum. (3) One end of the intestine at the bile-intestine confluence was closed, and the other end was anastomosed to the jejunum at 20 cm below the ligament of Treitz. (4) The abdominal cavity was flushed with normal saline, and the wound was closed 5 (Fig. 5).
Surgical procedure for DJE. (1) The duodenum was transected 0.5 cm below the pylorus, and its distal end was closed. (2) The jejunum was cut 10 cm below the ligament of Treitz, and the distal jejunum was anastomosed to  www.nature.com/scientificreports/ the proximal duodenum. (3) An incision was made at 10 cm below the duodenal-jejunal anastomosis site, and the proximal jejunum was anastomosed to this anastomosis site. (4) The intestine was transected 0.5 cm below the bile intestine confluence and the distal end was closed. (5) A lateral incision was made at 5 cm below the duodenal-jejunal anastomosis site, followed by anastomosis with the intestine at the bile intestine confluence. (6) The abdominal cavity was flushed with normal saline, and the wound was closed 5 (Fig. 6).
Blood glucose measurements. Rats were fasted overnight for 12 h. Next, samples of tail vein blood were obtained and analyzed for their glucose levels. The rats were then fed a normal diet for 2 h; after which, samples of blood were obtained and analyzed for their 2-h postprandial glucose levels. These blood glucose measurements were made at 1 week preoperation, and again at 8 and 24 weeks postoperation, respectively, with the use of a Roche Accu-Chek blood glucose meter and test strips. Tissue sampling. Rats received an intragastric administration of 20% glucose saline solution (1 g/kg) for 2 h on week 24 postoperation, and then were sacrificed. Samples of liver and intestinal tract tissue were collected and stored in liquid nitrogen (rat samples were coded as the group number + abbreviation for the surgical method; e.g., for the 7th rat in group A using the DJB procedure, the code was A7B).
Western blot detection of FXR expression in the liver tissues of animals treated with different surgical procedures. Livers were homogenized in lysis buffer containing protease and phosphatase inhibitors (BestBio, Shanghai, China) and centrifuged (12,000 rpm, 10 min, 4 °C). The supernatants were extracted and the protein concentration was determined by BCA Kit (Beyotime, Shanghai, China). Equivalent amounts of samples were loaded on 8% SDS-PAGE gels (Beyotime, China) and separated by electrophoresis. Then proteins were transferred onto PVDF membrane (Millipore, USA). After blocked in 5% fat-free milk for 2 h, the membrane was divided into two parts by the help of Marker (Thermo Fisher, USA), and incubated with primary antibodies to farnesoid X receptor (FXR, 56 kDa) (ABclonal Biotech, Wuhan, China) and glyceraldehyde-3-phosphate dehydrogenase (GAPDH, 37 kDa) (CST, USA) overnight respectively, followed by incubation in horseradish peroxidase-conjugated secondary antibodies (CST, USA) for 60 min. The protein bands were visualized by ECL solution (Beyotime, Shanghai, China), and the band intensity was quantified with the ImageJ software.
Primer information. FXR primers: F: TCT ACA GCG AAA GTG CTG GG R: GGA GGG CCC AAT CAG ATT CA SREBP-lc primers:  Detection of islet cell apoptosis. Rats were sacrificed at 24 weeks postoperation and their pancreatic tissues were removed. Apoptosis of islet cells was detected by the TUNEL assay: We counted all islets identified in the tissue section, and count insulin-positive cells that are positive for TUNEL and divide by total number of beta cells 6 .
Statistical analysis. Data for each group are presented as the mean ± standard deviation (x ± s). Inter-and intra-group statistical analyses were performed using repeated-measures multivariate analysis of variance, the F test, t test (T test for two independent samples), and one-way ANOVA of independent samples for inter-group and intra-group statistics. Relative levels of gene expression are expressed as the 2 − ΔΔCt value. All data were analyzed using IBM SPSS Statistics for Windows, Version 22 (IBM Corp., Armonk, NY, USA) and p-values < 0.05 were considered to be statistically significant.
Ethic approval and consent to participate. All applicable institutional and/or national guidelines for the care and use of animals were followed. A statement confirming the study was carried out in compliance with the ARRIVE guidelines.

Results
Blood glucose. There were no significant differences in the fasting blood glucose levels of rats in the 4 different groups either preoperation or postoperation (Fig. 7A). However, the 2-h postprandial blood glucose levels in the DJB and NBPD groups were significantly lower than those in the SHAM and DJE groups at 8 and 24 weeks postoperation and also at 1 week preoperation (p < 0.05) (Fig. 7B).
Fasting serum cholesterol, FFAs, and TGs. While the fasting serum cholesterol levels in the DJB and NBPD groups were significantly lower than those in the SHAM and DJE groups at 8 weeks postoperation (p < 0.05), no significant difference in fasting serum cholesterol levels was found among the 4 groups at 24 weeks postoperation (Fig. 7C). However, the fasting serum FFA levels in the DJB and NBPD groups were significantly lower than those in the SHAM and DJE groups at 8 and 24 weeks postoperation (p < 0.05) (Fig. 7D). The fasting serum TG levels in the DJB and NBPD groups were significantly lower than those in the SHAM and DJE groups at 8 and 24 weeks postoperation and also at 1 week preoperation (p < 0.05). The fasting TG levels in the DJE group were significantly higher than those in the NBPD group at 8 weeks postoperation (p < 0.05). No differences in other measurements were found when compared with the SHAM group (Fig. 7E).
Fasting serum insulin and HOMA-IR. Serum insulin levels in the DJB and NBPD groups were significantly lower than those in the SHAM and DJE groups at 8 and 24 weeks postoperation and also at 1 week preoperation (p < 0.05) (Fig. 8A). The HOMA-IR values in the DJB and NBPD groups were significantly lower than those in the SHAM group at 8 and 24 weeks postoperation and also at 1 week preoperation. Furthermore, the HOMA-IR values in the DJB group were significantly lower than those in the DJE group at both 8 and 24 weeks postoperation (p < 0.05) (Fig. 8B).

Apoptosis rate of islet cells.
The apoptosis rates of islet cells in the DJB and NBPD groups were significantly lower than that in the SHAM group (p < 0.05), and no significant difference was found between the apoptosis rates in the DJE group and SHAM group (Table 1) (Fig. 9). PCR, qPCR and western-blot results. 1. RNA was not degraded, and its quality was high (Fig. 10A). 2. A comparison of the relative expression of liver FXR mRNA(Relative expression is presented as the difference in gene expression (expressed by multiples) between the experimental group and the control group i.e., the 2−ΔΔCt value.) in animals treated via different operations (Fig. 10C) showed that the levels of FXR mRNA in the DJB group were significantly higher than those in the DJE group (p < 0.05), but did not significantly differ from those in the NBPD and SHAM groups; furthermore, FXR expression in the NBPD group was significantly higher than that in the DJE group (p < 0.01). 3. A comparison of the relative expression of intestinal FXR mRNA in animals treated via different operations (Fig. 10E) showed that the levels of FXR mRNA in the DJE group were significantly higher than those in the DJB, NBPD, and SHAM groups (p < 0.01).   www.nature.com/scientificreports/ 4. Results of electrophoresis of qPCR products produced by the FXR and GAPDH genes in animals treated via different operations (Fig. 10B) were consistent with the original 2 − ΔΔCt values of the FXR gene qPCR products for each corresponding operation (Fig. 10 D). 5. An electrophoresis of liver FXR protein (Fig. 10 F) and an analysis of liver FXR protein expression in animals treated via different operations (Fig. 10G) showed that the levels of liver FXR protein in the NBPD and DJB groups were significantly higher than those in the DJE and SHAM groups (p < 0.01). (The blots were cut prior  www.nature.com/scientificreports/ to hybridisation with antibodies, the 37 kDa and 56 kDa were closely, so the membrane should have been a whole membrane before we divided it, and it was area of blank above them) 6. A comparison of the relative expression of liver SREBP-1c mRNA in animals treated via different operations (Fig. 11) showed no significant difference in SREBP-1c mRNA expression among the different operation groups (p > 0.05). 7. A comparison of relative expression of liver CYP7A1 mRNA in animals treated via different operations (Fig. 12) showed that the levels of CYP7A1 mRNA in the NBPD group were significantly higher than those in the DJE group (p < 0.05).   www.nature.com/scientificreports/ 8. A comparison of fasting serum bile acid concentrations in rats treated via different operations (Fig. 13) showed no significant difference among the groups (p > 0.05).

Discussion
DJB is a surgical procedure commonly used to investigate the mechanism of how surgery affects diabetes. This procedure preserves the original biliopancreatic diversion of a gastric bypass and also the duodenal-lower jejunal exclusion. NBPD with biliopancreatic diversion alone and DJE with duodenal-jejunum exclusion alone were used in the present study to further investigate the hypoglycemic mechanism of DJB. Our previous studies showed that blood glucose levels were improved in both the DJB and NBPD treatment groups, but not in a DJE treatment group. Our present study showed that 2-h postprandial blood glucose levels in the DJB and NBPD groups as determined at 8 and 24 weeks postoperation were lower than those in the DJE and SHAM groups (p < 0.05) and significantly lower than those at 1 week preoperation (p < 0.05). The 2-h postprandial blood glucose levels in the DJE and SHAM groups at 8 and 24 weeks postoperation were not significantly different from those detected at 1 week preoperation. These results showed that the hypoglycemic effects of DJB and NBPD were stable and reliable, and that DJE or a sham operation could not reduce blood glucose levels.
Bile and pancreatic fluid are important digestive juices. Bile acid is the principal component of bile, and also an important ligand for the bile acid pathway. Our previous studies found that the postoperative blood bile acid levels in the DJB and NBPD groups were higher than those in SHAM and DJE groups 3 . Our present study showed that the levels of serum total bile acids at 24 weeks postoperation in the DJB, NBPD, DJE, and SHAM groups were 78.32 ± 26.51, 50.45 ± 18.75, 33.63 ± 13.58, and 18.13 ± 5.51 μmol/L, respectively, and the trend was consistent with findings in our previous studies; however, the differences among these groups were not statistically significant (p > 0.05). Reactive components of the bile acid signaling pathway, including the farnesoid X receptor (FXR, also known as the bile acid receptor) and the small heterodimer partner (SHP), are distributed throughout the body. These components mediate the expression of lipid metabolism-related genes to inhibit the de novo synthesis of fatty acids in the liver, which can be affected by an altered bile acid concentration. Previous studies 5 reported that secretion of fibroblast growth factor 15 (FGF15) was reduced if intestinal FXR was not activated or was inhibited. Insufficient FXR activity resulted in the reduced inhibition of CYP7A1 (a rate-limiting enzyme for bile acid synthesis) and a subsequent expansion of the bile acid pool to effectively reduce body weight and blood glucose levels. However, activation of lipid metabolism pathways in the liver, such as the FXR-SHP-Srebp-1c pathway 7,8 , depends on an expansion of serum and liver bile acid pools, and plays a major role in inhibiting the synthesis of cholesterol lipids, which can alleviate insulin resistance in liver and islet cells 9 . Moreover, CYP7A1 activity in the liver is regulated by negative feedback provided by intestinal FXR, and this regulation constitutes a bile acid "liver-intestinal negative feedback mechanism." Changes in any component of the feedback loop may alter the other components. (Bile acid feedback loop: Intestinal FXR activation↑ → FGF15 (fibroblast growth factor 15,with the help of DEIT1 gene) is produced↑ → The expression of liver CYP7A1 down-regulates↓ → Serum bile acid down-regulates↓ → Bile acid pool decreases↓ → Liver FXR/ Intestinal FXR inhibition↓ → The level of FGF15 shrink↓ → CYP7A1 activation↑ → The bile acid pool is enlarged↑ → Intestinal FXR activation↑).
Our present study revealed that FXR mRNA was expressed at significantly higher levels in the livers of rats in the DJB and NBPD groups than in the livers of rats in the DJE group. Furthermore, the levels of FXR mRNA in the DJB and NBPD groups were higher than those in the SHAM group; however, that difference was not statistically significant. A subsequent test for FXR protein expression revealed that the FXR levels in the DJB and NBPD groups were significantly higher than those in the DJE and SHAM groups (p < 0.05). In the intestine, the levels of FXR mRNA expression in the DJE group were significantly higher than those in the DJB, NBPD, and SHAM groups, which supported the findings mentioned above. The mean level of CYP7A1 mRNA expression in the livers of rats in the NBPD group (1.426 ± 0.159) was significantly higher than that in the DJE group (0.510 ± 0.098, p < 0.05); however, there was no significant difference in CYP7A1 mRNA expression between the DJB group (0.957 ± 0.259) and SHAM group (1.000 ± 0.180) (p > 0.05), indicating that the amount of secreted of bile acids had increased in the NBPD group. Therefore, we believe that elevated serum bile acid levels had upregulated the expression of liver FXR and down-regulated the expression of intestinal FXR. An up-regulation of FXR expression in the liver could reduce lipid deposition and lipotoxicity in the liver and exert a hypoglycemic effect by activating lipid metabolism pathways.
Studies have suggested that the FXR-SHP-Srebp-1c pathway 7,8 may play a major role in the hypoglycemic effect of a surgical procedure. Our present study detected expression of Srebp-1c downstream of FXR in the liver, and our results showed no statistically significant difference in Srebp-1c mRNA expression among the different groups. Furthermore, the fasting serum cholesterol levels before and after surgery in all surgical rats (Fig. 7C) were not significantly different at 24 weeks postoperation when compared with those at 1 week preoperation, and also did not significantly vary among the different groups (p > 0.05).
Recent studies have focused on interactions that occur between the targets of hypoglycemic drugs; i.e., transcription factor YY1 and liver FXR 10 . It is believed that YY1 can inhibit transcription of the liver FXR gene and activate Srebp-1c. In addition, YY1 may also up-regulate the expression of lipogenesis-related genes (such as the fatty acid synthase gene) 10,11 , resulting in excessive deposition of TGs and lipotoxicity in the liver. The results of our present study (Fig. 7E) showed that the fasting TG levels in the DJB and NBPD groups at 24 weeks postoperation were significantly lower than those at 1 week preoperation (p < 0.05), and the fasting TG levels in the DJB and NBPD groups at 24 weeks postoperation were significantly lower than those in the DJE and SHAM groups (p < 0.05). These finding indicated that the operation group with high liver FXR expression, however, has a low TG level. The ectopic deposition of TGs and FFAs plays an important role in the process of lipotoxicity leading to www.nature.com/scientificreports/ T2DM. Studies have shown that TGs and FFAs are closely related to the development and progression of diabetes, and may induce insulin resistance in organs and tissues [12][13][14] , and apoptosis in islet cells 15 ,while a decreased level of FFAs may help to alleviate insulin resistance 16 . In the present study, the HOMA-IR values decreased in both the intestinal bypass and modified biliopancreatic diversion groups, while no significant change was observed in the intestinal exclusion group and sham operation group. These results suggest that the intestinal bypass surgery had reduced serum FFA levels and lipotoxicity via biliopancreatic diversion, and thereby alleviated insulin resistance and inhibited islet cell apoptosis. In summary, biliopancreatic diversion primarily accounts for the hypoglycemic mechanism of intestinal bypass surgery. Biliopancreatic diversion can reduce lipotoxicity and improve the sensitivity of cells to insulin in T2DM rats by affecting bile acid pathways. These effects serve to alleviate insulin resistance and inhibit islet cell apoptosis, which eventually produces a hypoglycemic effect.

Conclusion
DJB postoperatively up-regulates the expression of bile acid receptors in the liver and down-regulates those receptors in the intestinal tract via biliopancreatic diversion. These changes reduce TG levels, and subsequently the lipotoxic effects of TGs and FFAs on islet cells to produce a hypoglycemic effect.

Data availability
They have been attached to the submission.