Berberine Slows the Progression of Prediabetes to Diabetes in Zucker Diabetic Fatty Rats by Enhancing Intestinal Secretion of Glucagon-Like Peptide-2 and Improving the Gut Microbiota

A total of 20 male Zucker diabetic fatty (ZDF; fa/fa) rats and 5 male Zucker lean (ZL; fa/+) rats aged 6 weeks (190–210 g) were purchased from Beijing Vital River Laboratory Animal Technology Co. Ltd. [Beijing, China; animal license no: SCXK (Beijing) 2012-0001]. The animals were housed and maintained under standard laboratory conditions (12:12 h light-dark cycle; 20 ± 2°C temperature; 40–60% humidity; noise < 50 dB) at the animal house of the Institute of Biomedical Engineering, Chinese Academy of Medical Sciences, Tianjin, China. All experiments in the current study followed the ethical principles approved by the Animal Ethics Committee of Tianjin Medical University (approval # DXBYY-IACUC-2020023). After acclimatization, rats were fed a high-energy rodent diet (Purina 5008 rat chow; Charles River Laboratories, Wilmington, MA, USA) and water ad libitum. An oral glucose tolerance test (OGTT) was performed weekly to determine the onset and stage of experimental T2DM in the rats so that fasting blood samples from the medial canthus and fecal samples could be taken at the stages of normal glucose tolerance (NGT), IGT and T2DM.

The ZDF rats were randomly divided (using a random number table) into an IGT group (n = 10) and a berberine group (n = 10). The ZL rats were defined as the control group (n = 5). Rats in the berberine group were administered berberine hydrochloride (Jinke Pharmaceutical Company, Yunnan, China) by gavage at a dose of 100 mg/kg/d for 3 weeks (32), whereas rats in the IGT and control groups were given the same volume of vehicle (distilled water).

Rats were fasted for 12 hours (from 20:00 to 8:00 the next day) once weekly so that biochemistry investigations and an OGTT could be performed. First, blood was drawn from the tail vein (after disinfection of the tail with 70% isopropyl alcohol) using sterile blood collection tubes for the measurement of fasting blood glucose (FBG; One Touch Ultra blood glucose meter; Johnson & Johnson, New Brunswick, NJ, USA) and other biochemical parameters (see below). Then, an OGTT was performed to stage the development of T2DM and identify when all rats in the IGT group had progressed to T2DM. The rat was administered 50% glucose (2 g/kg), and the glucose levels in tail vein blood were measured at 30 min, 60 min, and 120 min. The following diagnostic criteria were used to define IGT (one criterion met) and T2DM (both criteria met): a blood glucose level at 2 hours after glucose loading > 11.1 mmol/L; and peak blood glucose level > 16.7 mmol/L (33). Glucose tolerance was evaluated by calculation of the area under the curve (AUC) using the trapezoidal area method: AUC = ¼ × FBG value + ½ × 0.5-hour glucose value + ¾ × 1-hour glucose value + ½ × 2-hour glucose value.

The fasting plasma levels of GLP-1, GLP-2, lipopolysaccharide (LPS), insulin, and glucagon were assessed using enzyme-linked immunosorbent assay (ELISA) kits (Hermes Criterion Biotechnology, Vancouver, BC, Canada) and a monochromator-based multi-mode microplate reader (SynergyMx; BioTek, Winooski, VT, USA) according to the manufacturer’s protocol. The insulin resistance index (Homeostatic Model Assessment of Insulin Resistance, HOMA-IR) was calculated as: HOMA-IR = (FBG level × fasting plasma insulin level)/22.5. The homeostatic model assessment of insulin secretion (HOMA-β) was calculated as HOMA-β = 20×fasting plasma insulin level/(FBG level-3.5).

An automatic biochemical analyzer (Modular DP800; Roche, Basel, Switzerland) was used for the determination of blood urea nitrogen (BUN) level and the plasma levels of alanine aminotransferase (ALT), aspartate aminotransferase (AST), creatinine (Cr), uric acid (UA), triglycerides (TG), total cholesterol (CHO), and high-density lipoprotein (HDL). Low-density lipoprotein (LDL) levels were determined using the Friedewald equation: LDL=CHO-HDL-TG/5.

After the last OGTT had been performed, the rats were fasted for 24 hours and then sacrificed. Segments of the distal ileum (1–2 cm above the junction with the cecum) were removed and flushed in ice-cold phosphate-buffered saline (PBS). Some of the small intestinal segments were fixed with 4% formaldehyde (in buffer) for use in immunohistochemistry experiments, some of the small intestinal segments of distal ileal were embedded and frozen in OCT compound, then snap-frozen in liquid nitrogen, and stored at −80°C, and the other segments of ileal tissue were used for cell culture.

The distal ileum tissues from the rats were fixed with formalin, embedded in paraffin, and sectioned into 4-μm-thick sections. Hematoxylin and eosin (H&E) staining was performed following standard methods. The slides were observed under a microscope with a digital camera. The morphometric analysis of villi length and goblet cells was undertaken using Image-Pro Plus (Media Cybernetics, Inc., Rockville, MD, USA) by a blinded investigator.

Segments of distal ileum were fixed with formalin, embedded in paraffin, sectioned into 4-μm-thick slices, and subjected to immunohistochemistry (IHC) to detect ZO-1 and occludin. The sections were deparaffinized, rehydrated, and incubated with hydrogen peroxide for 10 min at room temperature to block endogenous peroxidase activity. After incubation for 20 min in normal goat serum, sections were immunostained using goat anti-ZO-1 polyclonal primary antibody (ab190085; Abcam, Cambridge, UK) or rabbit anti-occludin polyclonal primary antibody (ab31721; Abcam, Cambridge, UK) together with a streptavidin-biotin-peroxidase detection system and a DAB reagent kit (Zhongshan Jinqiao Biotechnology Company, Beijing, China). The sections were then stained by hematoxylin, thoroughly washed, dehydrated, and mounted using coverslips and DPX mounting medium. The slides were observed under a microscope (#BX51T-PHD-J11, Olympus Corporation, Tokyo, Japan). The number of immunopositive cells in 5 fields of view was counted independently by two researchers who were blinded to the grouping. The number of positive cells (value A) was graded as 0 (0–1%), 1 (1–10%), 2 (10–50%), 3 (50–80%), or 4 (80–100%). The staining intensity (value B) was graded as 0 (negative), 1 (weakly positive), 2 (positive), or 3 (strongly positive). The IHC score was then calculated as IHC = A × B (34).

Cryostat sections cut at 6–7 microns. After 4% paraformaldehyde fixation and antigen repair, the tissues were treated with 3% H₂O₂ for 10 min, blocked with 1% bovine serum albumin for 1 h, and then incubated with specific primary antibodies against TLR-4 (ab22048; Abcam, Cambridge, UK), NF-κB (ab16502; Abcam, Cambridge, UK), TNF-α (ab6671; Abcam, Cambridge, UK), and Mucin 2 (ab272692; Abcam, Cambridge, UK) at 4°C overnight. After washing, the sections were incubated with the secondary antibody for 1 h at 37°C. Finally, diaminobenzidine was added to the slides, which were then counterstained with hematoxylin. Pathological changes in tissue and protein expression were viewed under a light microscope with a digital camera. For western blots, proteins from tissues were extracted with RIPA buffer containing a protease inhibitor cocktail, and the protein concentration was determined by a BCA protein assay kit (Solarbio, China). Then, 40 μg of protein was separated by SDS/PAGE, transferred to nitrocellulose filter membranes, and blocked with 5% nonfat milk for 1 h at room temperature, and then the membranes were incubated with the following primary antibodies against mucin (ab272692; Abcam, Cambridge, UK) at 4°C overnight. After the blots were washed, the secondary antibody was added and incubated for 1 h at room temperature. The immunoblots were detected with an ECL kit (Advansta, United States). Band intensity was analyzed using ImageJ software and normalized to the expression of GAPDH.

Three-centimeter lengths of distal ileum were kept in PBS for the determination of glutamine-stimulated GLP-1 and GLP-2 secretion. The ileum was cleared of connective tissue, cut open, and rinsed with PBS and cell culture medium, and the outer muscle layers were carefully stripped off. Subsequently, the ileum was cut into three one-centimeter-long segments. Each segment was chopped using eye scissors and transferred to a 24-well plate, and 1000 mL of pre-warmed Roswell Park Memorial Institute (RPMI) 1640 cell culture medium (containing 2.05 mM glutamine) was added. The tissues were incubated for 60 min at 37°C in a humidified incubator at 5% CO₂, and then 4 mM glutamine was added to each well to stimulate the secretion of GLP-2. The tissue pieces were removed after 30 min, and the conditioned medium was stored at -80°C until use. The concentrations of GLP-1 and GLP-2 in the medium were assayed using ELISA kits (Hermes Criterion Biotechnology) (31).

The rats were placed on a horizontal test bench, the tail was lifted, and the abdomen was gently massaged to stimulate defecation. A clean 4-ml centrifuge tube was placed near the anus to take the feces. The feces were placed in liquid nitrogen immediately after the tube was sealed, and then stored in the refrigerator at -80°C for fecal DNA extraction and analysis. The extraction of fecal DNA from stool samples (180–220 mg) was carried out using the TIANamp Stool DNA kit (Tiangen Biotech, Beijing, China). DNA yields were measured from the ratio of the absorbances at 260 nm and 280 nm (pure DNA has a 260/280 nm absorbance ratio of 1.7–1.9) using NanoDrop spectrophotometry (Maestrogen, Hsinchu City, Taiwan). The DNA extracted from each stool sample was used as a template for the amplification of the V4 region of 16S rDNA genes. A dual-index sequencing approach was used to allow the generation of a large number of high-quality sequences while minimizing the cost of long and customized primers. Polymerase chain reaction (PCR) was performed using the V4 Dual-index Fusion PCR Primer Cocktail (515F:5’-GTGCCAGCMGCCGCGGTAA-3’, 806R:5’-GGACTACHVGGGTWTCTAAT-3’ and Phusion High-Fidelity PCR Master Mix (New England Biolabs, Ipswich, MA, USA) with a melting temperature of 56°C and 30 cycles. The PCR products were purified with Agencourt AmpureXP beads (Beckman Coulter, Brea, CA, USA) to remove unspecific products. The final library was quantified by determination of the average molecule length (Agilent 2100 Bioanalyzer Instrument and Agilent DNA 1000 Reagents; Agilent, Santa Clara, CA, USA) and quantitative PCR (EvaGreen dye, Biotium, Hayward, CA, USA). Paired-end sequencing was performed on a MiSeq platform (Illumina, San Diego, CA, USA) with the PE250 (PE251+8+8+251) sequencing strategy (MiSeq Reagent Kit). All procedures for fecal microbe analysis were performed under sterile conditions.

Raw pyrosequencing reads generated from the sequencer were quality filtered into Clean Data, and paired-end reads were assembled using FLASH (Fast Length Adjustment of SHort reads) v1.2.11. High-quality valid reads were clustered into operational taxonomic units (OTUs) with 97% similarity cutoff using USEARCH v7.0.1090. The phylogenetic affiliation of each OTU was analyzed by RDP Classifier v2.2 against the Greengene (V201305), Silva (V119), and UNITE (Version 6 20140910) databases using a confidence threshold of 80%. The representative sequences of the OTUs were used to analyze the alpha diversity (rarefaction curve analysis, observed species index, and Shannon diversity index) on the basis of their relative abundance (Mothur v1.31.2). Statistical comparisons among groups were made using the Kruskal-Wallis Test. A heatmap was generated according to the relative abundance of the OTUs using R v3.1.1. Weighted and unweighted UniFrac distance metric analysis was performed using OTUs for each sample, and principal coordinates analysis (PCoA) was conducted according to the matrix of distance.

Continuous measurement data are expressed as the mean ± standard error of the mean (S.E.M.) and compared between groups using one-way analysis of variance (ANOVA) and the Least Significance Difference (LSD) post hoc test. Count data are expressed as the median (four-point interval) and percentage and compared between groups using the rank-sum test. A P-value < 0.05 was taken to indicate a significant difference.

All ZDF rats had NGT at 6 weeks of age but exhibited IGT at 9 weeks of age. After 3 weeks of the intervention (i.e., berberine or vehicle, depending on the group), all rats in the IGT group (100%) developed T2DM, whereas only 3 rats (30%) in the berberine group progressed to T2DM, with 3 rats (30%) remaining in the IGT stage and 4 rats (40%) recovering to NGT. No rats died during the study.

FBG level was significantly higher in the IGT group than in the control group (9.38 ± 1.66 mmol/L vs. 3.52 ± 1.13 mmol/L, P < 0.05). The FBG level in the berberine group (5.03 ± 1.26 mmol/L) was significantly lower than that in the IGT group (P < 0.05) but still higher than that in the control group (P < 0.05). Oral glucose tolerance, measured by the AUC, was significantly worse in the IGT group than in the control group (39.82 ± 6.07 vs. 12.54 ± 0.86, P < 0.05; Figure 1A). Furthermore, oral glucose tolerance in the berberine group (24.68 ± 5.24) was significantly better than that in the IGT group (P < 0.05), although it remained significantly worse than that in the control group (P < 0.05; Figure 1A).

There were no differences between groups in the fasting plasma insulin level (Figure 1B). However, plasma glucagon level was significantly higher in the IGT group than in the control group (P < 0.05; Figure 1C). Treatment with berberine significantly reduced the plasma glucagon level (P < 0.05), although it remained at an elevated level in comparison to the control group (P < 0.05; Figure 1C). HOMA-IR was markedly increased in the IGT group as compared with the control group (P < 0.05; Figure 1D). Moreover, HOMA-IR was significantly decreased following treatment with berberine (P < 0.05), although it was higher than that in the control group (P < 0.05; Figure 1D). HOMA-β was markedly reduced in the IGT group compared with the control group(16.06 ± 2.94 vs 241.98 ± 95.07, P < 0.05), but there were no differences between the IGT and berberine groups (Figure 1E).

As can be observed in, the berberine group showed a decreased food intake by 2 weeks compared with the IGT group, but body weight was almost the same in the two groups. In addition, the berberine group had markedly improved blood glucose levels compared with the IGT group. 1

The IGT group had a higher ALT level and a lower AST level than the control group (P < 0.05), while treatment with berberine was associated with significant reductions in both ALT and AST (P < 0.05; Figure 1F and Supplementary Table 1). ZDF rats had a higher BUN level and a lower Cr level than ZL rats (P < 0.05), with no significant differences between the IGT and berberine groups (Figure 1G and Supplementary Table 1). The levels of CHO, TG, and HDL were higher in the IGT group than in the control group (P < 0.05), and treatment with berberine was associated with improvements in all three parameters (P < 0.05; Figure 1H and Supplementary Table 1), but there was no improvement in LDL.

The IGT group exhibited evidence of damage to the intestinal mucosa, including a disorderly arrangement of irregular villi, broadening and fusion of villi, reduced villi length, reduced numbers of goblet cells and inflammatory cell infiltration into the lamina propria (Figures 2A, B). However, treatment with berberine improved the intestinal mucosal morphology (Figure 2A).

Positive staining for occludin and ZO-1 was observed in the intestinal epithelium and glandular cells (Figure 2C). The percentage of cells staining positive for occludin differed significantly (P < 0.05 for all pairwise comparisons) between the control group (43.0 ± 4.1%), IGT group (15.6 ± 2.8%), and berberine group (23.4 ± 6.3%) (Figure 2D). Similarly, the percentage of cells staining positive for ZO-1 differed significantly (P < 0.05 for all pairwise comparisons) between the control group (38.8 ± 2.9%), IGT group (14.7 ± 2.9%), and berberine group (27.3 ± 4.5%) (Figure 2D). The IHC score was higher in the control group than in the IGT group (P < 0.05; Supplementary Table 2), and treatment with berberine was associated with improvements in IHC score (P < 0.05; Supplementary Table 2).

The positive staining for mucin was observed in intestinal mucus. Figures 2C, E, F showed that mucin expression was lower in the IGT group compared with the control group (P < 0.05) and that berberine increased mucin expression in IGT animals (P < 0.05).

Intestinal tissues were examined for the inflammatory markers TNF-α, NF-κB, and TLR-4 (Figures 3A–D). The IGT group displayed higher levels of TNF-α, NF-κB, and TLR-4 compared with the Control group. Berberine decreased inflammation compared with the IGT group, but without reaching the values of the control group.

Fasting plasma GLP-2 level was reduced in the IGT group in comparison to the control group (P < 0.05) but was partially restored by treatment with berberine (P < 0.05; Figure 4A). Glutamine-induced GLP-2 secretion from ileal tissue was not significantly different between the control and IGT groups but was enhanced after treatment with berberine (P < 0.05; Figure 4B). Fasting plasma GLP-1 level and glutamine-induced GLP-1 secretion from ileal tissue were both higher in the IGT group than in the control group (P < 0.05) but were not affected by treatment with berberine (Figures 4C, D).

Fasting plasma LPS level after 3 weeks of the intervention was significantly higher in the IGT group than in the control group (P < 0.05; Figure 4E). However, treatment with berberine resulted in a reduction in LPS (P < 0.05) to a level comparable with that in the control group (Figure 4E).

High-quality pyrosequencing technology was used to determine the structure of the fecal microbiota. A total of 528,887 usable raw sequences (an average of 31,111 sequences per sample) were generated from 17 samples. The high-quality sequences were delineated into 1,664 OTUs at a similarity cutoff of 97%. Shannon and observed species diversity curves demonstrated that most of the microbial diversity in each sample was captured with the current sequencing depth (Supplementary Figure 1). Calculation of the Shannon diversity index and observed species index revealed that species richness (as indicated by the observed species index) and diversity (as shown by the Shannon diversity index) were lower in the IGT group than in the control group (Figures 5A, B). Furthermore, treatment with berberine partially restored the richness and diversity of the gut microbiota (Figures 5A, B).

PCoA was applied to visualize the overall structural changes in the gut microbiota. Weighted and unweighted UniFrac significance tests demonstrated that there were distinct structural differences between groups after 3 weeks of intervention (Figures 5C–E).

The dominant phyla were Bacteroidetes, and Firmicutes followed by Proteobacteria and Verrucomicrobia. Bacteroidetes were significantly increased, and Firmicutes distinctly decreased in rats with IGT (in comparison to the control group), but berberine reversed these changes (Figure 6A). At the genus level, treatment with berberine was associated with enrichment of Bacteroides, Oscillospira, Akkermansia, Aggregatibacter, Clostridium, Roseburia, and Eubacterium as well as inhibition of Prevotella, which was the main component of the IGT group (Figure 6B). At the species level, the IGT group exhibited an increased abundance of Prevotella copri, a common Gram-negative bacterium in humans, and a decreased abundance of Akkermansia muciniphila, Lactobacillus reuteri, and Bacteroides caccae in comparison to controls (Figure 6C). Berberine improved the structure of the gut microbiota, as well as the abundance of Gram-negative bacteria (Figure 6C).

The original contributions presented in the study are publicly available. This data can be found here: NCBI PRJNA681546.