Back
Diabetic kidney disease (DKD) is a serious healthcare dilemma, associated with significant microbiota architecture and function alterations. Yet, the functional capacity of the DKD microbiome and its interplay with the host metabolism remain incompletely understood.
We conducted full-length 16S rRNA gene sequencing of fecal samples and targeted lipidomic profiling in a large cohort of diabetic patients with diverse renal function and control subjects.
A total of 990 subjects (455 control, 204 diabetes mellitus, 182 DKD and 149 chronic kidney disease, CKD, patients) were enrolled. The mean age was 62 years and 48.4% of patients were men. The mean serum creatinine was 0.8 mg/dL and the estimated glomerular filtration rate was 84 mL/min/1.73 m2. We carried out machine learning (ML) methods to identify the top distinguishing features that can discriminate DKD, DM, CKD patients and controls. The top 30 features selected by ML yield an accuracy rate of 0.74 and an area under curve (AUC) of 0.82 to identify diabetes mellitus and Control. The top 40 features selected by ML yield an accuracy rate of 0.76 and an AUC of 0.82 to identify DKD and diabetes mellitus. The top 30 features selected by AI yield an accuracy rate of 0.78 and an AUC of 0.86 to identify DKD and CKD. The top 20 features selected by ML yield an accuracy rate of 0.72 and an AUC of 0.79 to identify CKD and control. Using the integration of abundance analysis and machine-learning algorithm, we found that levels of 13 discriminatory microbial species and 4 circulating branched-chain amino acids (BCAA, leucine, isoleucine, valine and methionine) changed among disease groups. The gene functions of a DKD-specific biomarker, Gemmiger, were enriched in distinctive carbohydrate and BCAA metabolism pathways. Coincidently, circulating levels of many BCAA precursors, such as L-glutamate and L-aspartate, were also increased in diabetic and DKD patients, suggesting alteration of pyruvate fermentation and Krebs cycle in the presence of disturbed glycolysis and hyperglycemia.
Our findings reveal the connections between intestinal microbes and circulating metabolites perturbed in DKD and highlight the imbalance of energy utilization and BCAA metabolism in hyperglycemia.