The Animal Care Committee of Binzhou Medical University approved the protocol of animal experiments, with approval codes BAC18-115 and BAC19-056. Male Wistar rats and C57BL/6J mice were purchased from Shandong Laboratory Animal Center (Shandong, China) and housed on a 12-h light–dark cycle, at a temperature of 24°C. Male DIO rats and mice were fed a diabetogenic diet, which is high-fat diet with 60% kcal from fat [high-fat diet (60%) diet D12492], for a minimum of 4 months, and the final experimental animals used were between the ages of 7 and 8 months. Wistar rats were randomly assigned to one of four groups: (1) Saline; (2) 10% (wt/wt) OFS (Solarbio, Beijing, China); (3) MET (Sigma, 150 mg/kg); or (4) 10% OFS+MET, for 6 weeks. The dose of OFS and MET was selected based on previous reports (24, 27). ATM Skirrow (including 5 mg/ml neomycin, 5 mg/ml bacitracin, and 2 mg/ml natamycin) was brought from TOKU-E (USA). Food intake and body weight were measured daily. One day prior to sacrifice, rats were lightly anesthetized with isoflurane and body composition was measured via DXA scan with software for small animals. Using blood samples collected from DIO rats, blood biochemical indexes including HbA1c were measured with an automatic biochemical analyzer according to the manufacturer's instructions.

After overnight fasting (12 h), 2 g/kg glucose was orally administered and blood was collected at 0, 15, 30, 60, 90, and 120 min. Blood glucose was measured immediately with a OneTouch blood glucose meter (Johnson & Johnson, America). The insulin level was detected by the rat insulin ELISA kit (Millipore, Germany).

Plasma amylin, ghrelin, PYY, CCK, glucose-dependent insulinotropic polypeptide (GIP), and plasma inflammation-associated indexes, including endotoxin, TNF-α, IL-2, and IL-6, were measured using rat ELISA kits (Millipore, Germany). In mice, the levels of leptin, resistin, non-esterified fatty acid (NEFA), tumor necrosis factor-α (TNF-α), adiponectin, and IL-6 were also detected using ELISA kits purchased from Millipore.

Stool samples from DIO rats were collected after 8 weeks of treatment, quickly frozen using liquid nitrogen, and stored in a freezer (−80°C). The analysis of 16S rRNA sequence was conducted at BGI Genomics (Shenzhen, China). The extraction and quality of microbial genome DNA were conducted using the conventional method of TACB/SDS and measured on agarose gel (1%). V3–V4 region 16S rRNA genes were amplified using previous reported primers and methods (28). Libraries were constructed using the Ion Fragment Library Kit (Thermo, America) according to the manufacturer's protocols and sequenced on an Illumina Miseq PE300 system with a generation of single-end 400-bp reads. All sequences were classified using the NCBI BLAST and SILVA databases. Distance calculation, operational taxonomic units (OTU) cluster, rarefaction analysis, and estimator calculation (α-diversity and β-diversity) were performed by qiime2.

Fecal samples were obtained and then stored immediately after defecation at −80°C for further transplantation. One gram of fecal sample was diluted in 10 ml of PBS and then suspended to obtain supernatant including fecal microbiota. Transfer of fecal microbiota was conducted on both DIO rats and mice treated with antibiotics. As shown in Figure 6, from day −56 to −1, DIO rats were randomly assigned to the four groups. At day 0, OGTT and other metabolic profile baseline of rats were measured. Then, the rats were conducted with gavage of antibiotic solution during days 0–7. Further, intragastric administration of 0.2 ml of fecal material was conducted from day 7 to day 21 (14 consecutive days) to pseudo-germ-free rats. The OGTT and hematological index detection was carefully performed at days 22 and 23, respectively. As shown in Figure 8A, after a 7-day acclimatization, the pseudo-germ-free mice model was successfully established by administrating a large dose of antibiotic for 10 consecutive days. In the next 14 days (day 1 to day 14), 0.2 ml of the obtained fecal microbiota from the rats was orally transplanted to pseudo-germ-free mice and OGTT tests were subsequently measured.

All data are presented as the mean ± SEM. The comparison between three treatment groups to the control and normal groups were performed by one-way ANOVA test. P < 0.05 was considered significant.

In order to investigate if the combination of OFS and MET affects the metabolism of animals, we treated the DIO rats with placebo, 150 mg/kg of MET, 10% (wt/wt) OFS, and 10% (wt/wt) OFS+150 mg/kg of MET, respectively, for 8 consecutive weeks, using the healthy rats treated with saline as normal group. As a result, the effects of OFS, MET, and OFS+MET treatment on metabolic parameters in DIO rats are shown in Table 1. Compared to the control group, weight gain and final body weight of DIO rats were significantly decreased in both MET (P < 0.05) and OFS+MET (P < 0.01) groups; however, the effect was not significant, though the result had the same situation in the OFS group. Both MET- and MET+OFS-treated groups showed decrease in food intake compared to the control and OFS-only group. After the 8-week period of the experiment, significant differences in fat weight and fat/body (%) were observed in OFS+MET (40.1 ± 7.0 g, 8.3 ± 1.5%, P < 0.02) compared to the control (59.1 ± 9.1 g, 11.9 ± 1.8%) and similar to the normal ones (Figures 1A,B). Administration of both MET and OFS+MET significantly lowers the %HbA1c levels (Figure 1C). The other blood biochemical indexes were also carefully detected at the end of the experiment. As a result, all the sample-treated groups showed a down trending for the levels of triglyceride, cholesterol, and LDL-cholesterol in DIO rats compared to the control. HDL-cholesterol level was also determined and a significant difference was observed only in the OFS+MET-treated group, which was not far from normal values (Table 1).

The effects of OFS and MET treatment on fasting blood glucose, insulin levels, and HOMA value for insulin resistance (IR) in DIO rats are presented in Table 2. Except for the normal group, all the other groups showed significantly increased fasting blood glucose levels and blood insulin level, while the smallest increase was shown in the OFS+MET-treated group. Furthermore, index of insulin resistance was calculated using the obtained concentration of BGL and insulin level accordingly. Not surprisingly, significantly decreased HOMA-IR indices were similarly shown in all treatment groups compared with the control, especially for the OFS+MET-treated group showing the lowest HOMA-IR value and close to the normal.

To evaluate whether an 8-week treatment of OFS, MET, and OFS+MET helped to improve the glucose metabolism in DIO rat, OGTT tests were conducted before and during the chronic treatment (week 0, week 4, and week 8). As is presented in Figure 2, all of the MET, OFS, and MET+OFS treatment exert an obvious improvement on the glucose tolerance of the DIO rats, which were closer to the normal group compared with the control. DIO rats chronically treated with OFS+MET exhibited a best increased glucose stabilizing ability compared with all the groups that received placebo, individual OFS, and MET. There is no significant difference between the MET-treated group and the OFS+MET-treated group in glucose lowering at weeks 4 and 8. However, there is a tendency for OFS+MET treatment to exert a slightly better glucose-stabilizing effect than individual treatment of MET.

We further evaluated whether and how the plasma levels of appetite-related hormones were affected by the administration of OFS, MET, and OFS+MET during the OGTT at week 8. As shown in Figure 3 and Table 3, OFS+MET treatment significantly reduced the fasting concentration and AUC of insulin, amylin, and ghrelin (Figures 3B,C; Table 3), and increased the fasting levels and AUC of peptide YY (PYY), cholecystokinin (CCK), and GIP, which were close to the normal ones. These changes explained an obvious reduction of food intake and body weight due to appetite suppression lead (Figures 3D–F; Table 3).

To explore whether the 8-week treatments of OFS, MET, and OFS+MET affect the gut microbial composition of DIO rats or not, we detected 40 fecal samples from each group using the 16S rRNA gene sequencing-based method. As a result, the 16S rRNA sequencing yielded probably 83,000 clean reads (duplicate reads > 95%) in each sample after a strict process including quality control, denoising, and the removal of chimera. Then, the clean reads were clustered into operational taxonomic units (OTUs) and then carefully assigned to taxa from phylum to species level.

Beta diversity analysis was performed by principal coordinates analysis (PCoA) of the unweighted UniFrac distances of gut microbiota samples to explore the similarity of bacterial community patterns among all the DIO rat groups. As a result, a clear separation of the microbial community is shown in the placebo-treated control group, while the OFS and OFS+MET group tended to be closer to the normal group as compared with the MET group (Figure 4A). Alpha diversity is the mean diversity of bacteria or species within a community or habitat. Observed species', Shannon's, and Chao1's diversity indexes were also used to evaluate the community diversity of gut microbiota among all the treatment groups. Compared with the normal group, a significant decrease was observed in all the three diversity indexes among the control and Met group, while the OFS treatment exerts a non-obvious increase in Observed species and Chao1's diversity. Interestingly, the combination of OFS and MET could reverse the downward effect from MET on all three indexes, indicating that the supplement of OFS helps to reverse the decrease in gut microbiome diversity induced by MET administration (Figures 4B–D).

As shown in Figure 4E, the gut microbiota of the DIO rats were mainly composed of Bacteroidetes, Firmicutes, Proteobacteria, Actinobacteria, and Verrucomicrobia at the phylum level, especially for Bacteroidetes and Firmicutes. Compared with the normal rats, abundance of Firmicutes in placebo-treated DIO rats was obviously increased, while the Bacteriodetes and Actinobacteria were decreased relatively. In addition, the ratio of Bacteriodetes/Firmicutes, a measure associated with obesity and increased host inflammation, was reduced in the control group. In contrast, OFS+MET treatment increased the abundance of Bacteriodetes and Actinobacteria, which not only partially rescue the declining ratio of Bacteriodetes/Firmicutes, but also substantially correct the loss in obesity, respectively. Notably, abundance of Verrucomicrobia, which was a kind of beneficial bacterium helping to improve insulin resistance and reduce weight, in the OFS group and OFS+MET group was similar to the normal ones, while these were significantly decreased in placebo-treated DIO rats.

Compared with abundance of B. pseudolongum in normal rats, obesity induced an obvious decline in this species, while the treatment of OFS and MET led to a reverse trend. In addition, combination treatment of OFS and MET in DIO rats was associated with a more significant improvement. A similar pattern is shown in the abundance of Bifdobacterium species (Figure 5A). Since B. pseudolongum and Bifdobacterium are reportedly related to positive health effects, including improved gut barrier integrity, stimulated immune response, and reduced inflammation caused by diet-induced obesity, the abovementioned shifts indicate the potentially enhanced protection from OFS+MET against obesity-induced intestinal and systemic effects, compared with the treatment of individual OFS or MET only (29, 30). In addition, MET treatment induced increase in the abundance of Shewanella and Allobaculum associated (Figures 5G,H) with improved metabolic endotoxemia, while there are no obvious changes in the OFS group. Interestingly, both of the microbes were obviously increased with OFS+MET treatment.

Interestingly, two members of the Peptostreptococca species (Peptococcaceae rc4-4 sp. and Peptostreptococcaceae sp.), which were associated with intestinal inflammation and obesity, were both significantly reduced under OFS or OFS+MET treatment (Figure 5I). However, treatment of individual MET increased and maintained the abundance of Peptococcaceae rc4-4 sp. and Peptostreptococcaceae sp., respectively, indicating that OFS may be an effective complement to MET on the improvement of inflammation and obesity reduction (29, 31). Similar improvement of OFS on the function of MET was also shown in the alternation of the abundance of microbes related with side effects of MET treatment. Both on a compositional and functional level, we that found significant microbiome, including enrichment of virulence factors and gas metabolism genes, could be attributed to the significantly increased abundance of Escherichia species (Figure 5J), which was obviously reduced by the treatment of OFS (32–34).

Not only that, Intestinibacter was previously reported to improve the resistance to oxidative stress, be able to degrade fucose, and possess the genetic potential for sulfite reduction, including part of an assimilatory sulfate reduction pathway, and, in this research, has been shown to obviously increase in abundance under OFS treatment (33, 34). Furthermore, the addition of OFS also reversed the down-regulatory effect on the abundance of Intestinibacter under individual MET treatment (Figure 5K).

Our results suggest partial gut microbial mediation of both therapeutic and adverse effects of the OFS and MET, especially since the addition of OFS effectively improved the side effects related to microbes under MET treatment, although further in vivo validation is required to conclude causality and to verify our finding on the whole animal level.

To investigate whether the OFS+MET-altered microbiota could effectively contribute to the improvement of side effects and other metabolic profiles of MET, we exchanged and transferred fecal samples from each group to antibiotics-induced germ-free rats, and the design of this experiment is shown in Figure 6.

The serum levels change of endotoxin, TNF-α, IL-2, and IL-6 are shown in Figure 7. The MET group (transplantation of MET fecal microbiota, green histogram in Figure 7) showed a continuously deteriorating trend compared with the MET group (transplantation of OFS+MET fecal microbiota, blue histogram in Figure 7). The opposite trend is shown in two OFS+MET groups (transplantation of MET or OFS+MET fecal microbiota, white or red histogram, respectively, in Figure 7), indicating that transplantation of OFS+MET fecal microbiota has positive effects on inflammation-related factors that were reportedly induced by the administration of MET (21, 32). We further investigate the effects of fecal microbiota transplantation on glucose tolerance of DIO rats. Results are shown in Figure 8; fecal transfer to germ-free rats resulted in improved glucose tolerance in recipients of MET+OFS-altered microbiota, with overall substantial improvement of glucose metabolism, thus indicating that MET+OFS-adapted microbiota could contribute to more beneficial effects on glucose homeostasis compared with individual treatment of MET.

After intake of antibiotics through oral administration for 2 weeks, fecal microbiota from rats treated with placebo, OFS, MET, or MET+OFS were transplanted into pseudo-germ-free mice for 4 consecutive weeks (Figure 9A). The 4- to 6-week-old germ-free DIO mice were fed a 60% high-fat diet during the entire experiment for 45 days. After the 7-day antibiotics treatment, all the mice, with similar body weight and fasting blood glucose levels, were divided into four groups and then received an OGTT as baseline (Figure 9B). Then, the next OGTT were conducted at day 43, and results are presented in Figure 9C, indicating that the transplantation of fecal microbiota from OFS+MET significantly improved the glucose tolerance of the mice compared with the DIO control. Significant differences were also observed in body weight and fat mass between the mice that received placebo and OFS+MET-altered microbiota (Figures 9D,E).

Not only that, we also found that OFS+MET-altered fecal microbiota transplantation significantly decreased mesenteric and epididymal fat and total fat coefficient compared with the DIO control group (Table 4). In addition, especially for the results, Lee's index, an effective parameter to evaluate the obesity in adult rats, in the OFS+MET group was obviously improved compared to those in the DIO control group, which had a slightly better effect than the normal group.

We further evaluated whether the blood biochemical indexes were also affected by the fecal microbiota transplantation at day 44. As shown in Table 5, transplantation of fecal microbiota obtained from the OFS+MET group exerted a down trending in the CHOL level and significantly reduced levels of ALT, TG, and LDL-C levels compared with the DIO control. As shown in Figure 10, the serum levels of resistin, free fatty acid (FFA), and adiponectin were not significantly changed but exerted a more obvious downtrend in the germ-free mice that received fecal microbiota from OFS+MET-treated rats, whereas leptin and several inflammation-related factors, including TNF-α and IL-6, were all relatively declined due to the fecal transfer of MET+OFS-altered microbiota to germ-free mice.

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.