Integrative analyses of 16S rDNA sequencing and serum metabolomics demonstrate significant roles for the oral microbiota and serum metabolites in post-kidney transplant diabetes mellitus

Abstract

Oral microbiota and serum metabolites play crucial roles in diabetes, but their relationship with post-transplant diabetes mellitus (PTDM), a common complication post-kidney transplantation, is not well characterized. This study investigated the relationship of oral microbiota and serum metabolites with PTDM using integrative analysis of 16S rDNA sequencing and serum metabolomics. We recruited 61 kidney transplant recipients, including 30 in the PTDM group and 31 in normal glucose tolerance controls. Oral samples and serum samples were collected from all the kidney transplant patients to perform 16S rDNA sequencing and serum metabolomics analysis. We annotated 689 oral microbial species, including 134 species unique to the PTDM group and 157 species unique to the control group. PTDM group showed upregulation of 36 metabolites and downregulation of 19 metabolites. Based on the random forest machine learning algorithm, genera such as UCG-005 (AUC = 0.9355), Succinivibrio (AUC = 0.8108); Akkermansia (AUC = 0.7742), Anaerovibrio (AUC = 0.2667), and Schwartzia (AUC = 0.2667), and serum metabolites such as LPI 18:0 (AUC: 0.8086), methylglyoxal (AUC: 0.7946), Vulgarin (AUC: 0.7828), 2-mercaptobenzothiazole (AUC: 0.7591), and PI(18:0/20:3(5Z,8Z,11Z)) (AUC: 0.7419) showed high diagnostic potential and may serve as clinical biomarkers. Furthermore, clinical indicators in PTDM patients, such as creatinine, cystatin C, and urea, showed a significant association with the differential oral microbiota and serum metabolites. Dysbiosis in the oral microbiota of the PTDM patients was associated with changes in the serum metabolites and alterations in their functions. These findings provide new insights toward identifying mechanisms by which oral microbiota and serum metabolites contribute to the development of PTDM.IMPORTANCEThis study reveals an imbalance in oral microbiota in patients with post-transplant diabetes and uncovers the potential relationship between oral microbiota and serum metabolites. These findings provide new insights into the role of oral microbiota and serum metabolites in the treatment of post-transplant diabetes, offering relevant biomarkers for clinicians and future research.

Keywords: kidney transplantation; oral microbiota; post-transplant diabetes mellitus; serum metabolites.

Fig 1

Compositional characteristics of oral microorganisms in the diabetic recipients after renal transplantation. (A) The species rank abundance curve depicts the species richness and species evenness in the control and PTDM groups. (B) Stacked bar charts demonstrate the differences in the composition and abundance of the oral microbes at the genus level in the control and PTDM groups. (C) Venn diagrams show the number of common and unique microbial species in the control and PTDM groups. (D) Histograms show the relative abundances of oral microorganisms in different subgroups. (E) Evolutionary branching plots show significant differences in the oral microorganisms between the PTDM and control groups. (F) Histograms show the distribution of the significantly distinct species between the two groups with LDA scores > 2.5 and P < 0.05 as threshold parameters. (G) Identification of the top five oral microbial traits related to PTDM using the random forest algorithm. (H) ROC curve analysis of the top five oral microbial species predicted by the random forest algorithm.

Fig 2

Characteristics of serum metabolites in diabetic patients after renal transplantation. (A) PLS-DA score plot shows the differences in the serum metabolites between the PTDM and the normal groups. (B) The displacement test plot shows the discriminating ability of the PLS-DA model and whether it is overfitted. (C) The volcano plot demonstrates the overall distribution of the differential metabolites in the PTDM group compared to the control group. (D) The bar graph shows the KEGG annotation and classification of the differential metabolites. (E) The bubble plot demonstrates the KEGG enrichment results of the differential metabolites between the two groups. (F) The top 10 metabolites predicted by the random forest algorithm to cause PTDM. (G) ROC curve analyses of the top five metabolites predicted by the random forest algorithm to cause PTDM. (H) ROC curve analysis using the logistic regression model for the top five metabolites and the total AUC value for the target metabolites.

Fig 3

Mantel test network heatmap of the differential oral microbiota and differential metabolites. The squares in the heatmap on the right represent the strength of correlations between the metabolomes, with the red squares indicating stronger positive correlations (coefficients closer to 1) and the blue squares indicating stronger negative correlations (coefficients closer to −1). The network diagram in the lower left panel displays the correlation analysis results between the 10 differential microbial species and the 10 differential metabolites. The color represents the P-values. The thickness of the lines represents the correlation coefficients (r value), with a thicker line indicating a stronger correlation.

Fig 4

Analysis of the relationships between oral microbiota, serum metabolites, and clinical indicators in PTDM. (A) Relationship of the PTDM-related oral microbiota with the clinical indicators. (B) Relationship of the PTDM-related serum metabolites with the clinical indicators. *P ≤ 0.05, **P ≤ 0.001. The red color indicates positive correlation, and the blue color indicates negative correlation. A darker color indicates a stronger correlation. eGFR: estimated glomerular filtration rate; TG: triglyceride; TC: total cholesterol; HDL: high-density lipoprotein; LDL: low-density lipoprotein; WBC: white blood cells; PLT: platelet count.