Summary of sequencing results

High-quality sequences ranging from 735,027 to 3,159,798 were retrieved from intestinal stool samples of all the rats and had an average length of 413 to 422 bp. In the alpha diversity of this study, rarefaction curves reached a stable point and Coverage surpassed 99.4%, indicating that the microbial community was near saturation, and most species could be detected in the sequencing amount. Alpha diversity measures include the rarefaction curve, Coverage, Shannon and Chao indexes, and those for β-diversity include the hierarchical clustering tree and PCoA.

Successful establishment of pseudo-germ-free rats

Alpha diversity measurement (Fig. 2A–D) showed that the abundance and diversity of gut microbiota in the ABX group were considerably lower than those in the NS group (P < 0.01), and the β-diversity results (Fig. 2E–F) demonstrated distinct clustering for each group at the OTU level. Circos, heatmap and barplot analysis (Fig. 2G–I) all indicated that the bacterial genera of the ABX group was significantly less diverse than that of the NS group as well. In conclusion, the biological abundance and biodiversity of gut microbiota in the ABX group were substantially reduced, and the pseudo-germ-free rats were successfully established.

Fig. 2

Diversity of fecal microbiota in the NS and ABX groups. A Rarefaction curve. B–D Bacteria that were different between the NS and ABX groups in the B Coverage, C Shannon, and D Chao indexes. Differences were assessed by the Wilcoxon rank-sum test. *P < 0.05, **P < 0.01, in comparison to the NS group. E Hierarchical clustering tree at the operational taxonomic unit (OTU) level. F Principal co-ordinate analysis (PCoA) at the OTU level. G Circos sample–species relation map. H Community heatmap analysis at the genus level. I Community barplot analysis

We used Wilcoxon rank-sum tests to investigate the differences between the above two groups' fecal bacterial communities (selecting species with the top 15 mean sums, P < 0.05, Additional file 2: Fig. S1 A-E), and compared to the NS group, p__Proteobacteria, c__Gammaproteobacteria, o__Enterobacterales, f__Enterobacteriaceae and f__Morganellaceae were evidently more numerous in the ABX group. At the genus level, g__Klebsiella was also evidently more abundant, but the other species of this genus were evidently less. So we then determined the altered specific bacterial taxa between the two groups (Additional file 2: Fig. S1F) by utilizing a linear discriminant analysis (LDA) effect size (LEfSe) algorithm (LDA values of > 2 with P < 0.05). The NS group had 16 species with proportions exceeding 1%, which are listed in Additional file 1: Table S3, but the ABX group only had 2.

Dynamic changes in gut microbiota after the first FMT

Alpha diversity measurements (Fig. 3A–D) also showed that the abundance and diversity of the gut microbiota in the ABX group were evidently lower than in the NS group (P < 0.01). The richness and diversity of species in the FMT-Diab, FMT-Non, and FMT-Con groups were significantly higher than the ABX group (P < 0.01), which demonstrated that the abundance and diversity of gut microbiota in the pseudo-germ-free rats were distinctly enhanced after FMT. The β-diversity results showed that the groups were strongly clustered, and that the NS and FMT-Con groups had a certain similarity at the OTU level. Furthermore, circos, heatmap and barplot (Fig. 3E–I) showed that the groups were distinctly different from each other at the genus level. The NS and FMT-Con groups had some similarity. Based on the above analysis, fecal microbiota did successfully colonize the intestines of the pseudo-germ-free rats after FMT.

Fig. 3

Diversity of fecal microbiota in the NS, ABX, FMT-Diab, FMT-Non and FMT-Con groups. A Rarefaction curve. B–D Bacteria that were different among the groups in the B Coverage, C Shannon index and D Chao index. Differences were assessed by the Wilcoxon rank-sum test. *P < 0.05, **P < 0.01. E Hierarchical clustering tree at the operational taxonomic unit (OTU) level. F Principal co-ordinate analysis (PCoA) at the OTU level. G Circos sample–species relation map. H Community heatmap analysis at the genus level. I Community barplot analysis

Dynamic changes in gut microbiota structure after 2 weeks of HFD

Alpha diversity measurements (Additional file 3: Fig. S2A-D) indicated that the richness of the gut microbiota in the ABX-ord group was obviously lower than that in the NS-ord group (P < 0.01), but the diversity of the gut microbiota showed no remarkable difference, implying that with the extension of feeding time, the gut microbiota in the pseudo-germ-free rats became progressively revived. The abundance and diversity of the gut microbiota in the ABX-fat group were both evidently lower than in the ABX-ord group (P < 0.01), which suggests that HFD was not conducive to the self-recovery of the gut microbiota. At the OTU and genus level, the results showed that the NS-fat, ABX-fat, FMT-Diab, FMT-Non, and FMT-Con groups had close similarity, and each group had certain clustering that was readily observed (Additional file 3: Fig. S2 E-I). In addition, the differences in the gut microbiota among the groups were lower than those after the first FMT.

Dynamic changes in gut microbiota structure after 4 weeks of HFD

The 16S rRNA sequencing results showed that the clustering of the groups was poor, and the differences in the gut microbiota among groups became smaller than after two weeks of HFD (Additional file 4: Fig. S3 A-I). Since the gut microbiota of the rats had self-recovery ability, the difference among the groups steadily diminished with time, and we therefore executed a second FMT.

Dynamic changes in gut microbiota after the second FMT

The 16S rRNA results after the second FMT (Additional file 5: Fig. S4 A-I) showed that the abundance and diversity of the gut microbiota communities in the NS-fat, FMT-Diab, FMT-Non, and FMT-Con groups were signally higher compared to the ABX-fat group (P < 0.01). The ABX-ord group was similar in depth to the ABX-fat group, with g__Klebsiella being the most abundant bacterial genus. In conclusion, the second FMT effectively improved the composition of the gut microbiota of rats in every group.

The effects of FMT on glucose and lipid metabolism and IR

OGTT assay was used to evaluate the rats’ glucose tolerance (Fig. 4A–B). The NS-fat, FMT-Diab, FMT-Non, and FMT-Con groups showed significantly elevated glucose excursions following glucose challenge compared to the ABX-ord group (P < 0.01), and the FMT-Diab group increased its glucose excursions more than the ABX-fat group (P < 0.01). Compared to the ABX-fat group, the molding rate (fasting blood glucose ≥ 11.1 mmol/L) of the FMT-Non group was lower, and the mortality rate of the FMT-Diab was higher. For blood markers, the levels of blood glucose, serum insulin, HOMA-IR, TC, TG, and LDL-C in the FMT-Diab group were clearly higher than those in the ABX-fat group (P < 0.05), and the level of HbA1c showed an upward trend. However, the level of HbA1c in the FMT-Diab group was distinctly higher than in the ABX-ord group (P < 0.05). These results indicate that T2DM-susceptible flora transplantation could increase the level of blood glucose, decrease the level of serum insulin, and promote IR, slowing down lipid metabolism in rats.

Fig. 4

The effects of HFD and STZ on glycolipid metabolism. A Blood glucose levels were measured before 0 min and at 15, 30, 60, 90, and 120 min after glucose loading. B AUC of the OGTT. C molding rate. D mortality rate. E FBG. F HbA1c. G insulin. H HOMA-IR. I TC. J TG. K LDL-C. L HDL-C. Data are shown as mean ± SD.*P < 0.05, **P < 0.01 vs. ABX-ord group; #P < 0.05, ##P < 0.01 vs. ABX-fat group; &P < 0.05, &&P < 0.01 vs. FMT-Diab group

The effect of FMT on pancreatic islet histopathological alterations

As shown in Fig. 5, the rats’ pancreatic islets were plump and elliptical, and the exocrine acinar cells were around the islets, which themselves showed no abnormal pathological changes in the NS-ord and ABX-ord groups. A diminished number and volume of islets, islet cell necrosis, vacuolar degeneration, and fibrous tissue hyperplasia were observed in the entire field of vision in the NS-fat group. Furthermore, the volume of islets in the ABX-fat group was markedly reduced, and their structure was disordered and accompanied by fibrous tissue hyperplasia. In the FMT-Diab group, the islet structure was also disordered, and the hemosiderin was deposited, accompanied by fibrous tissue hyperplasia and severe inflammation. For the FMT-Non group, there was likewise a decreased number of islets, reduced volume, disordered structure, and hyperplasia of fibrous tissue. Finally, in the FMT-Con group, the number of islets  was decreased, the volume was reduced, the structure was disordered, and a small amount of vacuolar degeneration occurred.

Fig. 5

Micrographs of rat pancreas specimens by H&E staining in the A NS-ord, B ABX-ord, C NS-fat, D ABX-fat, E FMT-Diab, F FMT-Non, and G FMT-Con (Magnification: × 200). Scale bar: 100 μm

The effects of FMT on the composition of the colonic microbiota

Alpha diversity measurements (Fig. 6A–D) showed that the richness of gut microbiota was not substantially different between the ABX-ord and ABX-fat groups but that the diversity in the ABX-fat group was distinctly lower than in the ABX-ord group (P < 0.01). The richness and diversity of communities in the FMT-Diab, FMT-Non, and FMT-Con groups were prominently higher compared to the ABX-fat group as well (P < 0.01). Neither the diversity nor richness of the FMT-Con group and NS-fat group were clearly different. Similarly, the abundance and diversity of communities in the FMT-Diab group were not evidently different compared to the FMT-Non group.

Fig. 6

Diversity of fecal microbiota in the NS-ord, ABX-ord, NS-fat, ABX-fat, FMT-Diab, FMT-Non and FMT-Con groups. A Rarefaction curve. B–D Bacteria that were different among the groups in the B Coverage, C Shannon index and D Chao index. Differences were assessed by the Wilcoxon rank-sum test. *P < 0.05, **P < 0.01. E Hierarchical clustering tree at the operational taxonomic unit (OTU) level. F Principal co-ordinate analysis (PCoA) at the OTU level. G Circos sample—species relation map. H Community heatmap analysis at the genus level. I Community barplot analysis

The β-diversity results (Fig. 6E–F) showed that the NS-ord group had distinct clustering and was obviously different from other groups at the OTU level. At the genus level, the ABX-ord group was similar to the ABX-fat group, and the NS-fat, FMT-Diab, FMT-Non, and FMT-Con groups were also all similar to one another (Fig. 3G–I), because they were obviously affected by the STZ, which exceeded the effect of FMT and resulted in a convergence of the flora in each group.

To understand the differences in every group further and to obtain the bacterial genera that may be linked to T2DM development and occurrence, LEfSe multi-level discriminant analysis was conducted. The genera proportions exceeding 1% can be found in Additional file 1: Table S4. In Fig. 7, compared to the ABX-fat group, g__Klebsiella and g__Escherichia-Shigella were lower in the FMT-Diab and FMT-Non group, and g__Blautia, g__Lactobacillus, g__Ruminococcus_torques_group, g__unclassified_f__Lachnospiraceae, g__Lachnoclostridium, g__Bifidobacterium, g__Ruminococcus_gauvreauii_group, g__Bacteroides, g__Fusicatenibacter, g__Ruminococcus_gnavus_group, g__Coriobacteriaceae_UCG-002, g__norank_f__Butyricicoccaceae, g__norank_f__Lachnospiraceae, g__Anaerostipes and g__Romboutsia were higher in the FMT-Diab group. Among them, g__Klebsiella reached 92.32% in the ABX-fat group, and g__Lactobacillus reached 23.23%, and g__Blautia reached 18.93% in the FMT-Diab group. In addition, compared to the FMT-Diab group, the g__Lactobacillus, g__norank_f__Butyricicoccaceae and g__Ruminococcus_gnavus_group were noticeably lower in the FMT-Non group. Our team's previous study showed that g__Ruminococcus_gnavus_group was markedly different in the donor T2DM and Non-T2DM rats [15], and it was still markedly different after FMT. Therefore, we consider g__Ruminococcus_gnavus_group to be a specific genus that affects T2DM outcomes.

Fig. 7

LDA scores of gut microbiota for the A ABX-fat and FMT-Diab groups; B ABX-fat and FMT-Diab groups; C FMT-Diab and FMT-Non groups at the genus level

Mechanism validation

Researchers have found that T2DM is associated with SCFAs, primarily acetic acid, propionic acid, and butyric acid, which play a crucial role in regulating glycolipid metabolic disorders, improving IR, and treating obesity and other metabolic diseases. In Fig. 8A–C, we can see that compared with the ABX-ord group, the contents of acetic and butyric acid were distinctly lower in the ABX-fat group (P < 0.01). The contents of acetic, propionic and butyric acid in the FMT-Diab and FMT-Non groups were significantly increased compared to the ABX-fat group as well (P < 0.01). In addition, compared to the FMT-Diab group, the FMT-Non group had an increasing trend.

Fig. 8

The mechanism was investigated using gas chromatography, Western blot, and qRT-PCR. A–D SCFA content in the ABX-ord, ABX-fat, FMT-Diab, and FMT-Non groups. A Acetate. B Propionate. C Butyrate. D–E GPR41 and GPR43 mRNA expression (n = 3). F–H GPR41 and GPR43 protein expression. Data were shown as mean ± SD. *P < 0.05, **P < 0.01 vs. ABX-ord group; #P < 0.05, ##P < 0.01 vs. ABX-fat group; &P < 0.05, &&P < 0.01 vs. FMT-Diab group

QRT-PCR and WB assay was performed to examine the connection between SCFAs and T2DM by detecting GPR41 and GPR43 mRNA and protein levels. In Fig. 8D–H, the results from the WB agreed with those obtained from qRT-PCR. Compared to the ABX-ord group, the expression of GPR41/43 in the ABX-fat group were remarkably lower. Compared to the ABX-fat group, the expression of GPR41/43 were higher in the FMT-Diab and FMT-Non groups, and compared to the FMT-Diab group, the expression of GPR41/43 in the FMT-Non group were higher. These results indicate that T2DM-susceptible flora transplantation can reduce the production of SCFAs and the expression of GPR41/43 in the intestines of rats.