Peanut skin extract ameliorates the symptoms of type 2 diabetes mellitus in mice by alleviating inflammation and maintaining gut microbiota homeostasis

Alleviation of the clinical symptoms and energy metabolism disorder of mice with T2DM via oral PSE administration

The change in fasting blood glucose, body weight, food intake, and water consumption of mice with T2DM after they were orally administered with PSE are shown in Figures 1A–1D. The fasting blood glucose (measured with a test paper; Figure 1A) and body weight (Figure 1B) of the mice in the diabetic model (DM) group were significantly higher than those of the normal mice in all the experimental periods (p < 0.001 and p < 0.001). However, the food intake (Figure 1C) and water consumption (Figure 1D) of the mice with T2DM increased on the 6^th week (p < 0.001 and p < 0.001). Meanwhile, these parameters in the mice with T2DM significantly decreased after 6 weeks of administration of 80 and 160 mg/kg PSE and 140 mg/kg MET (Figures 1A–1D; p < 0.001). At the end of the experiment, fasting blood glucose was measured again with the biochemical analyzer, and the results revealed that this parameter obviously decreased. The effect of 80 mg/kg PSE was similar to that of 140 mg/kg MET (Figure 1E; p < 0.001). These results suggested that PSE improved the clinical symptoms of mice with T2DM and elicited the same effects as those of MET.

The weights of the liver, epididymal fat, heart, spleen, pancreas, and kidney of the mice with T2DM after PSE treatment are shown in Supplementary Figure 1. HFD significantly increased the weights of the liver, epididymal fat, heart, pancreas, and kidney of the mice with T2DM (p < 0.001, p < 0.01, p < 0.01, p < 0.01, and p < 0.01, respectively). Furthermore, 80 and 160 mg/kg PSE inhibited obesity in these organs, and their weights returned to normal levels (Supplementary Figure 1; p < 0.001, p < 0.01, p < 0.05, p < 0.001, and p < 0.001). However, the weights of these organs and the adipose tissue of the mice with T2DM were not affected by 10 mg/kg PSE. The blood biochemical markers were also examined. The plasma TG and TC of the mice with T2DM significantly increased compared with those of the normal control group (Supplementary Table 2; p < 0.01 and p < 0.05). The plasma TG and TC of the groups treated with 80 mg/kg PSE and 140 mg/kg MET significantly decreased (Supplementary Table 2; p < 0.01 and p < 0.05). The plasma AST and ALT showed no significant change compared with those of the normal and pathological control groups (Supplementary Table 2). These results indicated that PSE treatment alleviated the symptoms such as hyperphagia, polydipsia, energy metabolism disorder, and hyperglycemia in the mice with HFD-induced T2DM. PSE in the dose range that caused a significant effect did not trigger a toxic response from the mice. The effect of 80 mg/kg PSE was similar to that of 160 mg/kg PSE, so 80 mg/kg PSE was selected to analyze the mechanism and fecal microflora via 16S RNA sequencing.

PSE-induced improvement of the glucose and insulin tolerance of mice with T2DM

Hyperglycemia, hyperinsulinemia, and glucose tolerance are associated with obesity and metabolic syndrome [25]. After the oral administration of PSE for 5 weeks, an oral glucose tolerance test (OGTT) and an insulin tolerance test (ITT) were performed to detect the glucose and insulin tolerance of the mice with T2DM. The OGTT results are shown in Figures 2A and 2B. The blood glucose levels of the groups with diabetes at time points 0, 15, 60, 90, and 120 min and the area under the curve (AUC) were higher than those of the normal control group (Figures 2A and 2B; p < 0.001). These indices were significantly decreased by the PSE at all the tested doses and MET at the indicated dose (Figures 2A and 2B; p < 0.01 and p < 0.001). The ITT experiment also revealed that the blood glucose, AUC, and insulin of the DM group were higher than those of the normal control group (Figures 2C–2E; p < 0.001). However, these characteristics of the mice with T2DM were obviously improved by PSE and MET (Figures 2C–2E; p < 0.05 and p < 0.01). These results suggested that PSE treatment alleviated the clinical symptoms of mice with T2DM by increasing glucose tolerance and enhancing insulin sensitivity.

Sequences and operational taxonomic units of the stool samples of normal, diabetic, PSE-treated, and MET-treated mice

The disorder of gut microflora composition is associated with obesity and diabetes [1, 7]. In our study, Illumina high-throughput sequencing was conducted to read the 16S RNA sequences of the V4 region of 35 fecal microflora samples of the control, DM, 140 mg/kg MET treatment, and 10 and 80 mg/kg PSE treatment groups and to understand whether PSE affected the gut microbiota of mice. Each group had seven samples. Operational taxonomic units (OTUs) were classified on the basis of 97% similarity, and OTU1 with few sequences was removed (i.e., the number of sequences in OTUs was less than 2). The tag numbers of the control, DM, MET-treated, and PSE-treated groups were 72,644 ± 466.3, 74,290 ± 603.1, 71,807 ± 563.8, 72,780 ± 583.9, and 71,990 ± 458.3, respectively (Supplementary Table 3). The averages of the number of OTUs in each group were 512 ± 19.4, 511 ± 16.7, 524 ± 13.8, 521 ± 36.6, and 438 ± 32.3, respectively (Supplementary Table 3). The number of tags and OTUs of each group did not have significant changes.

The results of the PCA analysis based on OTU abundance and intergroup alpha diversity box map are given in Figures 3A and 3B. Letters a, b, c, d, and e in the graph represent the 10 mg/kg PSE-treated, 80 mg/kg PSE-treated, normal control, DM, and 140 mg/kg MET-treated groups, respectively. The results showed that the normal control and DM groups could be distinguished clearly on the basis of the gut microbiota composition. Similar to MET, 80 mg/kg PSE significantly affected the gut microbiota composition of the mice with T2DM (Figure 3A). The results of the multivariate ANOVA of the PCoA matrix scores indicated a statistically significant separation between the microbiota of the control group and the DM group. Significant separations were also observed in the DM, MET, and 80 mg/kg PSE groups (Figure 3C). However, the gut microbiota of the mice in the 10 mg/kg PSE-treated group did not significantly differ from that of the mice with diabetes (Figure 3C). This outcome was consistent with the weak effects observed in the mice treated with 10 mg/kg PSE (Figures 1 and 2). These results suggested that PSE significantly regulated the intestinal microflora.

Regulatory effects of PSE on gut microbiota

Species and abundance analysis was conducted at phylum, class, order, family, genus, and species levels to further understand how PSE regulated gut microbiota. For 35 samples, 9 phyla, 13 classes, 14 orders, 20 families, 22 genera, and 10 species were obtained. Nine phyla, namely, Acidobacteria, Actinobacteria, Bacteroidetes, Cyanobacteria, Deferribacteres, Firmicutes, Proteobacteria, TM7, and Tenericutes, were examined (Figures 4A and 4B). The abundance of Actinobacteria, Proteobacteria, and Deferribacteres of the DM group significantly increased compared with that of the control group (p < 0.001, p < 0.05, and p < 0.001, respectively). The abundance of Bacteroidetes and Cyanobacteria of the DM group clearly decreased compared with that of the control group (p < 0.05 and p < 0.001, respectively). After 80 mg/kg PSE and 140 mg/kg MET were orally administered, Actinobacteria was not detected in the PSE- and MET-treated groups (Figure 4B). Moreover, the ratio of Firmicutes to Bacteroidetes of the DM group significantly increased. Although no effect was observed in 10 mg/kg PSE-treated group, 80 mg/kg PSE and 140 mg/kg MET significantly decreased the ratio of Firmicutes to Bacteroidetes and allowed the gut microbiota of the mice with T2DM to recover to normal level.

Thirteen classes, namely, 4C0d-2, Actinobacteria, Bacilli, Bacteroidia, Betaproteobacteria, Clostridia, Deferribacteres, Deltaproteobacteria, Epsilonproteo-bacteria, Erysipelotrichi, Gammaproteobacteria, Mollicutes, and TM7-3, were investigated (Supplementary Figures 2A and 2B). The abundance of Actinobacteria, Betaproteobacteria, Deltaproteo-bacteria, Erysipelotrichi, and Deferribacteres of the DM group significantly increased compared with that of the control group (p < 0.0001, p < 0.01, p < 0.01, p < 0.01, and p < 0.01, respectively). By contrast, the abundance of Bacilli, Bacteroidia, Chloroplast, Gammaproteobacteria, and 4C0d-2 decreased (p < 0.001, p < 0.05, p < 0.001, p < 0.01, and p < 0.01, respectively). These effects were reversed by treating the mice with 80 mg/kg PSE and 140 mg/kg MET. However, 10 mg/kg PSE-treated group did not have this effect.

Fourteen orders, namely, Bacteroidales, Bifidobacteriales, Burkholderiales, CW040, Campylobacterales, Clostridiales, Deferribacterales, Desulfovibrionales, Enterobacrerides, Erysipelotrichales, Lactobacillales, RF39, Turicibacterales, and YS2, were analyzed (Supplementary Figures 3A and 3B). The abundance of Bacillales, Turicibacterales, Burkholderiales, Bifidobacteriales, Deferribacterales, Desulfovibrionales, and Erysipelotrichales of the DM group significantly increased compared with that of the control group (p < 0.05, p < 0.0001, p < 0.0001, p < 0.0001, p < 0.0001, p < 0.0001, and p < 0.0001, respectively). By contrast, the abundance of Bacteroidales, Enterobacteriales, Streptophyta, YS2, and Lactobacillales of the DM group significantly decreased compared with that of the control group (p < 0.05, p < 0.01, p < 0.01, and p < 0.0001, respectively). The normal level of the gut microflora of the mice with T2DM was recovered after they were treated with 80 mg/kg PSE and 140 mg/kg MET. However, the abundance of Lactobacillales significantly increased in the 10 mg/kg PSE-treated group.

Twenty families, namely, Alcaligenaceae, Bacteroidaceae, Bifidobacteriaceae, Deferribacteraceae, Desulfovibrionaceae, Enterobacteriaceae, Erysipelotri-chaceae, F16, Helicobacteraceae, Lachnospiraceae, Lactobacillaceae, Mogibacteriaceae, Odoribacteraceae, S24-7, Turicibacteraceae, Paraprevolellaceae, Porphy-romobbnadaceae, Prevoellaceae, Rikenellaceae, and Ruminococcaceae, were analyzed (Supplementary Figures 4A and 4B). The abundance of Alcaligenaceae, Deferribacteraceae, Bifidobacteriaceae, Desulfovibrio-naceae, Erysipelotrichaceae, Eubacteriaceae, Mogi-bacteriaceae, Peptococcaceae, and Peptostrep-tococcaceae of the DM group significantly increased (p < 0.0001, p < 0.0001, p < 0.001, p < 0.0001, p < 0.001, p < 0.001, p < 0.001, p < 0.001, and p < 0.001, respectively). By contrast, the abundance of Clostridiaceae, Enterobacteriaceae, Lactobacillaceae, Rikenellaceae, Odoribacteraceae, and Prevotellaceae of the DM group significantly decreased (p < 0.0001, p < 0.0001, p < 0.001, p < 0.0001, p < 0.0001, and p < 0.001). These changes in the abundance of the microflora of the mice in the DM group were reversed by the treatment with 80 mg/kg PSE and 140 mg/kg MET. Furthermore, Turicibacteraceae was found in the DM group. However, this effect was not observed in the 10 mg/kg PSE-treated group.

Twenty-two genera, namely, AF12, Allobaculum, Bacteroides, Bifidobacterium, Clostridium, Coprococcus, Desulfovibrio, Escherichia, Flexispira, Helicobacter, Lactobacillus, Mucispirillum, Odoribacter, Oscillospira, Parabacteroides, Paraprevotella, Prevotella, Rikenella, Roseburia, Ruminococcus, Sutterella, and Turicibacter, were checked (Supplementary Figures 5A and 5B). The abundance of AF12, Allobaculum, Bifidobacterium, Christensenella, Coprobacillus, Flexispira, Mucispirillum, Odoribacter, Turicibacter, Anaerofustis, and Sutterella of the DM group significantly increased compared with that of the control group (p < 0.05, p < 0.0001, p < 0.001, p < 0.0001, p < 0.0001, p < 0.01, p < 0.001, p < 0.0001, p < 0.0001, p < 0.0001, and p < 0.0001, respectively). The abundance of Escherichia, Lactobacillus, Lactococcus, Odoribacter, Paraprevotella, Prevotella, and Rikenella of the DM group was lower than that of the control group (p < 0.05, p < 0.001, p < 0.05, p < 0.0001, p < 0.0001, p < 0.05, and p < 0.01, respectively). These changes in abundance of gut microbiota at genera were amended after the mice were treated with 80 mg/kg PSE and 140 mg/kg MET. However, 10 mg/kg PSE did not have the same effect.

Ten species, namely, Bacteroides acidifaciens, Bifidobacterium pseudolc, Clostridium methylpentos, Desulfovibrio C21 c2, Escherichia coli, Lactobacillus reuteri, Mucispirillum schaedleri, Parabacteroides distasor, Ruminococcus flavefacie, and Ruminococcus gnavus, were investigated (Supplementary Figures 6A and 6B). The abundance of B. acidifaciens, Conterminous flavefaciens, Ruminococcus gnavus, Lactobacillus reuteri, and Lactococcus garvieae of the DM group decreased (p < 0.0001, p < 0.05, p < 0.01, p < 0.01, and p < 0.05, respectively), whereas the abundance of Bifidobacterium pseudolc, Clostridium colinum, Clostridium hungatei, Mucispirillum schaedleri (p < 0.001, p < 0.01, p < 0.05, and p < 0.001, respectively) of the DM group increased. After 80 mg/kg PSE and 140 mg/kg MET were administered, the abundance of these gut microbiota recovered to normal levels similar to that of the control group. These results suggested that the increase in B. pseudolc and Parabacteroides distasor in gut microflora might be the causative factors of T2DM. The growth of these microorganisms was inhibited by 80 mg/kg PSE, and this effect was similar to that of 140 mg/kg MET. Furthermore, 80 mg/kg PSE improved the clinical symptoms of the mice with T2DM induced by long-term HFD. However, 10 mg/kg PSE did not exhibit such effects.

Beta diversity and cluster analysis of the species composition of the fecal samples of mice in the normal, diabetes, MET, and PSE treatment groups

The Bray Curtis heatmap of the results of the beta diversity and cluster analysis of the species composition is shown in Figures 5A and 5B. The similarity of microbial flora among the control group, 80 mg/kg PSE treatment group, and MET treatment group was very high. The gut microflora in the DM and 10 mg/kg PSE-treated groups were similar. Cluster analysis revealed that the microbiota of 80 mg/kg PSE-treated, control, and 140 mg/kg MET-treated groups significantly differed from those of 10 mg/kg PSE-treated and DM groups. These results implied that the imbalance of gut microbiota was associated with T2DM. Therefore, 80 mg/kg PSE and 140 mg/kg MET might significantly affect intestinal microorganisms. However, 10 mg/kg PSE was inefficient.

LDA effect size analysis of the 16S RNA sequence of intestinal microflora after oral PSE administration

The linear discriminant analysis (LDA) score and profiling of LDA effect size (LEFSE) analysis are displayed in Figures 6A and 6B. Erysipelotrichi, Allobaculum, Erysipelotrichaceae, Erysipelotrichales, Deferribacteraceae, Deferribacterales, Deferribacteres, Mucispirillum, and Flexispira were the main microorganisms in the DM group. The abundance of Firmicutes, Bacilli, Lactobacillaceae, Lactobacillales, Actinobacteria, Bifidobacteriaceae, Bifidobacterium, Bifidobacterialeacs, Sutterella, Turicibacteraceae, Turicibacter, Turicibacterales, Alcaligenaceae, Burkholderiales, Betaproteobacteria, and Bataproteo-bacteria of the 10 mg/kg PSE-treated group significantly increased compared with that of the control group. Conversely, Bacteroidetes, Bacteroidia, Bacteroidales, Prevotella, Prevotellaceae, Enterobacteriales, Gamma-proteobacteria, Enterobacteriaceae, and Escherichia in the 80 mg/kg PSE-treated group were few. By contrast, the Rikenellaceae, Rikenella, Odoribacteraceae, Paraprevotellaceae, and Paraprevotella in the 140 mg/kg MET-treated group were abundant, and they occupied a high proportion among all the microorganisms. These results suggested that 80 mg/kg PSE and 140 mg/kg MET might regulate the gut microbiota of the mice with T2DM and elicit an antidiabetic effect on mice with T2DM depending on the type of microorganisms.

PICRUST function analysis of the gut microbiota of mice with T2DM after PSE treatment

PICRUST was applied to determine the 16S RNA sequence and predict the abundance of gut microbiota at levels 2 and 3 of the KEGG pathway. The results of PICRUST function analysis of the gut microbiota of the mice with T2DM are shown in Figure 7, Supplementary Figures 7 and 8. The abundance of the gut microbiota in cellular process signaling and lipid metabolism at level 2 of the KEGG pathway of the DM group significantly decreased compared with that of the normal control group (Figure 7A; p < 0.05 and p < 0.01, respectively). After the treatment with 80 mg/kg PSE, the abundance of cellular process signaling and lipid metabolism of the diabetic model mice significantly increased compared with that of the DM group (Figure 7A; both with p < 0.001). Furthermore, the abundance of the gut microbiota in fatty acid biosynthesis, lipid biosynthesis, proteins, transcription machinery, other ion-coupled transporters, and starch and sucrose metabolism at level 3 of the KEGG pathway in the DM group significantly decreased (Figures 7B and 7C; p < 0.001, p < 0.01, p < 0.01, p < 0.05, and p < 0.001, respectively). The abundance of the gut microflora in aminoacyl-tRNA biosynthesis and an unknown function at the same level in the DM group increased (Figures 7B and 7C; p < 0.05 and p < 0.001, respectively). All these parameters or factors in the mice with T2DM recovered or mostly recovered to normal levels after they were treated with 80 mg/kg PSE and MET at the indicated dose. These results suggested that the gut microorganisms involved in lipid, sugar, and protein metabolism were affected by PSE and MET treatments that caused antidiabetic effects.

PSE-induced reduction of LPS and pro-inflammatory cytokines in the blood of mice with T2DM

Inflammation is one of the important factors that induce T2DM [7–9]. As such, this factor was investigated in our study. The changes in LPS and pro-inflammatory cytokines in the blood of experimental mice are displayed in Figure 8. The LPS (Figure 8A) and pro-inflammatory cytokines (Figure 8B) of DM significantly increased after the long-term feeding of HFD. Simultaneously, the LPS of the mice with T2DM obviously decreased after they were treated with 80 and 160 mg/kg PSE (Figure 8A; both with p < 0.001). However, 10 mg/kg PSE did not significantly affect the LPS in the blood of the mice with T2DM (Figure 8A). Pro-inflammatory cytokines, such as TNF-α, IL-6, and IL-1α, did not significantly change in the mice treated with 10 mg/kg PSE. However, they significantly decreased after the treatment with 80 mg/kg and 160 mg/kg PSE and with the indicated dose of MET (p < 0.001, p < 0.001, p < 0.001, p < 0.05, p < 0.001, and p < 0.01, respectively). These results indicated that PSE and MET could significantly prevent the increase in LPS and pro-inflammatory cytokines in the blood of the mice with T2DM induced by HFD. However, low-dose PSE (10 mg/kg) was not effective against these factors.

PSE-induced reduction of inflammation-related gene expression and protein levels in the adipose tissues and liver of mice with T2DM

The changes in inflammation-related gene expression and protein levels in adipose tissues are displayed in Figures 9A and 9B. The gene expression levels of TNF-α, IL-1β, IL-6, and PAI-1 in the adipose tissues of the DM group were higher than those of the normal control mice (Figure 9A; p < 0.05, p < 0.01, p < 0.001, and p < 0.01, respectively). These gene expression levels in the mice with T2DM decreased after they were treated with 80 mg/kg PSE and 140 mg/kg MET (Figure 9A; p < 0.05, p < 0.01, p < 0.001, p < 0.05, and p < 0.001, respectively). Furthermore, the protein levels of TNF-α, IL-6, and IL-1β were examined (Figure 9B). These results strongly supported the conclusion from the mRNA analysis. The effect of PSE on inflammation in the liver was also investigated. The changes in inflammation-related gene expression and protein levels are shown in Figures 9C and 9D. The significant increase and decrease in the mRNA expression and protein levels of these factors in the liver were similar to those in the adipose tissue (p < 0.05, p < 0.01, and p < 0.001). Interestingly, the IL-10 gene expression in the tissues in the DM group increased after the mice were treated with 80 mg/kg PSE and 140 mg/kg MET (Figures 9A and 9C; p < 0.01 and p < 0.001, respectively). These results indicated that PSE alleviated inflammation in the adipose tissue and liver of mice with T2DM induced by HFD.

PSE-induced alleviation of the inflammation of the hypothalamus of mice with T2DM

Hypothalamic immunity is mediated by IkappaB kinase-β (IKK-β) and nuclear factor kappaB (NF-κB). Hypothalamus combatant aging-related health problem is restored via immune-neuroendocrine integration and immune inhibition or gonadotropin-releasing hormones (GnRH) [26]. Thus, changes in GnRH, IKK-β, and NF–Kβ gene expression in the hypothalamus were examined in our study. The results are displayed in Figure 10. The mRNA abundance of IKK-β and NF-Kβ (p50) in the hypothalamus of the mice with T2DM was higher than that of the normal control mice (p < 0.05 and p < 0.05, respectively). The gene expression level of GnRH was lower than that of the normal mice (p < 0.05). No significant difference in the NF-Kβ (p65) gene expression between the normal and diabetic mice was observed. Importantly, the gene expression levels of IKK-β and NF-Kβ (p50) in the hypothalamus of the mice with T2DM significantly decreased after they were treated with 80 mg/kg PSE (p < 0.05 and p < 0.05) and 140 mg/kg MET (p < 0.05). Interestingly, the expression levels of GnRH in the hypothalamus of the mice with T2DM obviously increased after the treatment with 10 and 80 mg/kg PSE and 140 mg/kg MET (p < 0.01, p < 0.01, and p < 0.01, respectively). These results revealed that PSE prevented the inflammation of the hypothalamus by regulating the IKK-β/NF-Kβ/GnRH signaling pathway.

Preparation of PSE

Dried peanut skin (500 g; Anji County, Huzhou City, Zhejiang Province, China) was powdered and then extracted with a mixture of ethanol and water at a ratio of 40:60 three times. The extracts were collected, and ethanol was evaporated to give a water solution applied to HP-20 resin. The sample was washed with water and eluted with 40% aqueous ethanol. The obtained fractions were combined, concentrated, and dried by freeze-drying to give a total of 57.5 g of PSE powder (11.5% of dry weight). PSE components were analyzed through high-performance liquid chromatography [15]. The stability of PSE was also examined by comparing the properties of the ingredients before and after the experiment. When the animal tests were completed, the remaining PSE was evaluated through ¹H NMR and HR-ESI-MS. No changes were detected.

Construction of T2DM mouse model and animal experimental design

Five-week-old male mice from Institute for Cancer Research were purchased from the Zhejiang Academy of Medical Sciences (Hangzhou, China). They were divided into five mice per cage and kept under controlled conditions: temperature of 22 ± 1 °C, humidity of 50% ± 5%, and under 12 h light/dark cycle. They were fed with a commercial diet (normal diet [ND] or HFD with 22.3% crude protein, 19.8% fat, and 44.6% carbohydrate from SLAC Experimental Animal Co., Ltd., Shanghai, China) for 4 months and allowed to take food and water ad libitum. During the animal experiments, water consumption, food intake, and body weight were monitored weekly. After the mice were fed with HFD for 4 months, their fasting blood glucose level was measured. The T2DM mouse model was considered to be successfully established when the following symptoms were observed or measured simultaneously: >7 mmol/L blood glucose concentration, obesity, polydipsia, polyphagia, and polyuria.

Subsequently, 10 normal mice and 50 mice with diet-induced diabetes were randomly divided into six groups: normal mouse group (control), DM control group, diabetic model with MET treatment as a positive control, and DM with PSE treatment. The sample was dissolved in water and administered orally. The control group was fed with ND, whereas the other groups were given HFD during experiments. The control and DM groups were given water as a vehicle every day. The positive control group received MET (Aladdin Bio-chem Technology Company, Shanghai, China) at a dose of 140 mg/kg per day. The PSE-treated groups were given doses of 10, 80, and 160 mg/kg per day. MET and PSE were administered daily for 6 weeks. Fasting blood glucose, body weight, food intake, and water consumption were measured at the end of each week. At the end of the experiment, the mice were starved overnight and sacrificed via neck dislocation under deep anesthesia by using ether. The blood, heart, liver, pancreas, spleen, kidney, adipose tissue, and hypothalamus were removed, weighed, and stored at −30 °C for future analysis. This experiment was repeated twice and performed in accordance with international ethical standards with guidance and approval from the Committee of Experimental Animal Care, Zhejiang University (Permit Number: ZJU20160436).

Enzyme-linked immunosorbent analysis

Blood samples were collected from the orbital sinus. The samples were allowed to stand at room temperature for 2 h, centrifuged at 12,000 × g for 15 min at 4 °C to obtain the plasma, and stored at −30 °C until use for biochemical analysis. TNF-α, IL-1β, and IL-6 protein levels in the plasma, liver, and adipose tissue were measured using commercially available mouse TNF-α, IL-1β, and IL-6 ELISA kits (4A Bio-Tech, Beijing, China) in accordance with the manufacturer’s instructions.

Detection of fasting blood glucose and plasma biochemical indicators at the end of the experiment

At the end of each week, the mice were fasted, and the fasting blood glucose level was measured from the blood collected from the tip of the tail vein by using a glucometer (Andon Health, Tianjin, China). The plasma samples at the end of the experiment were sent to Hangzhou Dian Medical Test Center (Hangzhou City, Zhejiang Province, China) to evaluate the levels of alanine aminotransferase (ALT), aspartate aminotransferase (AST), triglycerides (TGs), total cholesterol (TC), and glucose in the plasma with a biochemical analyzer (COBAS 8000; Roche, Basel, Switzerland) and assay kits (Roche, Basel, Switzerland).

Oral glucose and insulin tolerance tests

In this experiment, glucose tolerance was measured after PSE was administered for 5 weeks. Overnight-fasted mice were given glucose orally at a dose of 1.5 g/kg. The glucose concentrations in the blood taken from the tail vein at 0, 15, 30, 60, 90, and 120 min were measured with a glucometer.

Subsequently, an ITT was performed. The mice were fasted for 4 h. Afterward, insulin (Yeason, Shanghai, China) was intraperitoneally administered at a dose of 0.4 units/kg. Then, glucose concentration was determined using the blood samples from the tail vein at 0, 15, 30, 45, and 60 min.

Analysis through real-time polymerase chain reaction

The liver and epididymal fat were removed surgically, weighed, quickly frozen, and kept at −30 °C until RNA extraction. RNA from the white adipose tissue, liver, and hypothalamus was extracted, and their cDNA was synthesized [36]. Approximately 100 mg of epididymal fat, 50 mg of liver tissue, and one hypothalamus were used for RNA extraction with a Trizol reagent (Invitrogen, California, USA). Afterward, 2.5 μg of the total RNA of each sample was used for reverse transcription by using a HiFi-MMLV cDNA kit (Fermentas, Shenzhen, China). Real-time PCR analysis was performed using CFX96-Touch (Bio-rad, Hercules, CA, USA) and SYBR Premix EX Taq^TM (Takara, Otsu, Japan). The primer sequences (Sangon Biotech Co., Ltd., Shanghai, China) used in PCR are given in Supplementary Table 1. Glyceraldehyde-3-phosphate dehydrogenase was used as an internal control for normalization.

Analysis of gut microbiota

The feces of the mice in each group was taken at the end of the animal experiments to detect the change in gut microflora. Stool samples were quickly frozen in liquid nitrogen and sent to BGI Company (Wuhan, China) for the analysis of gut microbiota in accordance with a previously described method [1]. The analysis method is also presented in Supplementary Information.

Statistical analysis

Animal experiments were repeated twice. Data were expressed as the mean ± standard error of the mean and statistically analyzed through two-tailed multiple t tests with GraphPad Prism 5 and a Kruskal test or a Wilcoxon test with Metastats. Data about gut microflora were examined using the R software. Results with p < 0.05 were considered significantly different between groups.

Statement of ethics

This experiment was repeated twice and performed according to international ethical standards with guidance and approval from the Committee of Experimental Animal Care, Zhejiang University (Permit Number: ZJU20160436).