Fecal microbiota transplantation alters the susceptibility of obese rats to type 2 diabetes mellitus

Abstract

Obesity is one of the susceptibility factors for type 2 diabetes (T2DM), both of which could accelerate the aging of the body and bring many hazards. A causal relationship is present between intestinal microbiota and body metabolism, but how the microbiota play a role in the progression of obesity to T2DM has not been elucidated. In this study, we transplanted healthy or obese-T2DM intestinal microbiota to ZDF and LZ rats, and used 16S rRNA and targeted metabonomics to evaluate the directional effect of the microbiota on the susceptibility of obese rats to T2DM. The glycolipid metabolism phenotype could be changed bidirectionally in obese rats instead of in lean ones. One possible mechanism is that the microbiota and metabolites alter the structure of the intestinal tract, and improve insulin and leptin resistance through JAK2 / IRS / Akt pathway. It is worth noting that 7 genera, such as Lactobacillus, Clostridium and Roche, can regulate 15 metabolites, such as 3-indolpropionic acid, acetic acid and docosahexaenoic acid, and have a significant improvement on glycolipid metabolism phenotype. Attention to intestinal homeostasis may be the key to controlling obesity and preventing T2DM.

Keywords: intestinal microbiota and metabolites; leptin receptor; obesity; susceptibility; type 2 diabetes mellitus.

Figure 1

Changes in glycolipid metabolism phenotypes in recipient rats before and after transplantation. (A) Detailed information. LZ rats were fed a normal diet, and ZDF rats were fed an induced diet #5008. After adaptive feeding, the four groups were given an antibiotic mixture for 10 days, and then the corresponding supernatant from the LZ group and ZDF group was given to LZ and ZDF recipient rats, whereas the control group was given PBS. The course of T2DM was judged by OGTT, ITT, RBG, and FSI. After antibiotic administration and FMT, feces were collected for 16S rRNA sequencing and metabolomic analysis of intestinal contents at the end of the experiment; (B) Weight gain at different stages (g; Time: F_{3, 115} = 90.60, P < 0.0001; Group: F_{4, 45} = 110.2, P < 0.0001; Interaction: F_{20, 190} = 1.844, P < 0.05; n = 7-10); (C) Abdominal circumference at different stages (cm; Time: F_{5, 270} = 318.0, P < 0.0001; Group: F_{4, 270} = 67.39, P < 0.0001; Interaction: F_{20, 270} = 16.09, P < 0.0001; n = 10); (D) Random blood glucose at different stages (mM; Time: F_{2, 90} = 131.3, P < 0.0001; Group: F_{4, 45} = 55.78, P < 0.0001; Interaction: F_{20, 214} = 12.67, P < 0.0001; n = 8-10); (E) Glycosylated hemoglobin after FMT (%; F_{5, 54}=2396, P < 0.0001; n = 10); (F) Comparison of OGTT (mM; Time: F_{4, 216} = 190.3, P < 0.0001; Group: F_{5, 54} = 55.93, P < 0.0001; Interaction: F_{20, 216} = 20.59, P < 0.0001; n = 10); (G) Comparison of ITT (mM; Time: F_{5, 270} = 134.4, P < 0.0001; Group: F_{5, 54} = 11.58, P < 0.0001; Interaction: F_{25, 270} = 1.942, P = 0.0056; n = 10); (H) The levels of TG (mM; F_{5, 45} = 84.27, P < 0.0001); (I) TC (mM; F_{5, 54} = 20.55, P < 0.0001); (J) LDL-C (mM; F_{5, 49} = 131.0, P < 0.0001), and (K) HDL-C (mM; F_{5, 48} = 68.74, P < 0.0001) after FMT (mM, n = 7-10). *P < 0.05, **P < 0.01, and ***P < 0.001 vs. L-P, ^#P < 0.05, ^##P < 0.01, and ^###P < 0.001 vs. Z-P, ^&P < 0.05 vs. Z-Lg in (B–D, F, G). *P < 0.05, **P < 0.01, ***P < 0.001, and ****P < 0.0001 indicated inter-group changes in (E) and (H–K). Statistical analysis was performed with two-way ANOVA in (B–D), one-way ANOVA in (E) and (H–K) and repeated ANOVA in (F, G). The data were expressed as the mean ± SD.

Figure 2

Establishment of the pseudoaseptic rat model and evaluation of intestinal microbiota structure after FMT. (A) Shannon index and Simpson index among six groups after intragastric administration of antibiotics and a three-dimensional sequence plot of unweighted UniFrac PCoA analysis corresponding to LZ and ZDF rats after antibiotics (Shannon: F_{5, 52} = 10.03, P < 0.0001; Simpson: F_{5, 50} = 12.94, P < 0.0001; n = 10); (B) Shannon index and Simpson index among six groups after FMT (Shannon: F_{5, 53} = 13.48, P < 0.0001; Simpson: F_{5, 53} = 14.69, P < 0.0001; n = 10) and a three-dimensional sequence plot of unweighted UniFrac PCoA analysis corresponding to LZ and ZDF rats after FMT (n = 10). The percentage in parentheses of coordinate axes represented the proportion of differences in the original data that the corresponding principal coordinates could explain. Statistical analysis was performed with one-way ANOVA in (A, B). *P < 0.05, **P < 0.01, and ***P < 0.001. The data were expressed as the mean ± SD; (C) Unweighted UniFrac distance box plots. Horizontal coordinates corresponded to statistical comparisons between groups and within groups, and longitudinal coordinates indicated the corresponding distance values. Borders of boxes represented the interquartile range (IQR), horizontal lines represented the median value, and upper and lower whiskers represented 1.5 outside the upper and lower quartiles. In the IQR range, the symbol “+” denoted potential outliers that exceed the range. Statistical analysis was performed with Student’s t-test and Monte Carlo permutation test; (D) PLS-DA discriminant analysis graph. Each point represented a sample. The same color points belonged to the same grouping, and the same grouping points were marked with ellipses (n = 10). Yellow: L-P; Green: L-Lg; Purple: L-Zg; Orange: Z-P; Red: Z-Lg; Blue: Z-Zg.

Figure 3

Specific phyla and genera in each group after FMT. (A) Relative abundance of bacteria at the phylum level (n = 10); (B) Relative abundance of bacteria at genus level (n = 10); (C) The petal diagram revealed common and unique genera associated with different groups. Different colors represented different modules; (D) Firmicutes / Bacteroidetes ratio (F_{5, 45} = 6.511, P = 0.0001; n = 8-9). Statistical analysis was performed with two-way ANOVA. *P < 0.05, **P < 0.01. The data were expressed as the mean ± SD; (E) Violin maps of abundance distribution of seven OTUs with the most significant difference among sample groups. The abscissa represented the group, and the ordinate represented the number of sequences of each taxon in each sample (group) (n = 10). Using Mothur software, the statistical algorithm of Metastats was invoked to test the difference in sequence quantity (absolute abundance) between the samples (groups) of each taxon at the phylum and genus levels; (F) The venn diagram of common functional groups predicted by PICRUSt. Each ellipse represented a sample (group). The overlapping regions between ellipses indicated common functional groups among the samples (groups). The number in each block indicated the number of common or unique functional groups of the samples (groups) included in the block; (G) KEGG third-level pathway heat map predicted by PICRUSt. The abscissa was the third level functional group of KEGG, and the ordinate was the sample number. The color markers were the number of macrogenomes constructed from biom files. The intensity of the colors represented the degree of association (red, higher number of corresponding samples; green, lower number of corresponding samples).

Figure 4

The effects of FMT on the intestinal pathological structure, IR, and LR in rats. (A) The colon surface of rats was magnified 2000 times (upper), 5000 times (middle), and 10,000 times (lower). IV: mucosal layer, microvilli on cell surface. mit: mitochondria. N: nucleus. BC: goblet cells. White vesicle structures were secretory vesicles. ER: endoplasmic reticulum; (B) Fasting serum insulin (mIU / L; F_{5, 36} = 351.5, P < 0.0001); (C) HOMA-IR Index (F_{5, 39} = 85.58, P < 0.0001); (D) Leptin in 6 groups (pg / mL; F_{5, 43} = 141.7, P < 0.0001; n = 7-9); (E) Western blotting analysis of IR and LR signaling pathway molecules in liver tissues was performed after FMT; (F) Quantification of western blotting analysis in (D) (p-IRS2 / IRS2: F_{3, 15} = 2.175, P = 0.0132; p-JAK2 / JAK2: F_{3, 15} = 0.5387, P = 0.6630; p-Akt / Akt: F_{3, 15} = 7.221, P = 0.0032; FoxO1 / β-actin: F_{3, 15} = 6.224, P = 0.0059; n = 3-4). Statistical analysis was performed with two-way ANOVA. *P < 0.05, **P < 0.01. The data were expressed as the mean ± SD.

Figure 5

Metabolite composition of intestinal microbiota in ZDF rats after FMT. (A) The composition of metabolite types in each sample; (B) Score plot of 2D and 3D PLS-DA (n = 10). The green dots indicated L-P, the blue dots indicated Z-P, the red dots indicated Z-Lg, and the orange dots indicated Z-Zg; (C) Z-score heat map of differential metabolites. In the figure, the horizontal direction represented samples, and the longitudinal direction represents metabolites. The intensity of the colors represented the degree of association (red, higher content in the corresponding samples; blue, content in the corresponding samples. The relative numerical values represented by the colors were shown in the ribbon on the right.); (D) According to the results of single-dimensional statistics, the P-value was statistically significant for 15 groups of different metabolites as shown in box plots (n = 7-10); (E) Venn diagram of different metabolites. The number of shared and unique different metabolites screened by each group was shown; (F) Bubble map of the P-value of the metabolic pathway involved in the different metabolites. When the bubble was larger or the color was darker, the corresponding P value was smaller. Gray bubble, 0.05 < P < 0.1, Colored bubble, P < 0.05.

Figure 6

Association map of the three-tiered analyses integrating the gut microbiome, phenotypes, and metabolome. The left side of the panel showed associations between gut microbiota and phenotypes. The right side of the panel showed associations between metabolites and phenotypes. The intensity of the colors represented the degree of association (red, positive correlation; blue, negative correlation). *P < 0.05, **P < 0.01, ***P < 0.001.