Development and validation of diagnostic models for immunoglobulin A nephropathy based on gut microbes

Abstract

Background: Immunoglobulin A nephropathy (IgAN) is a highly prevalent glomerular disease. The diagnosis potential of the gut microbiome in IgAN has not been fully evaluated. Gut microbiota, serum metabolites, and clinical phenotype help to further deepen the understanding of IgAN.

Patients and methods: Cohort studies were conducted in healthy controls (HC), patients of IgA nephropathy (IgAN) and non-IgA nephropathy (n_IgAN). We used 16S rRNA to measure bacterial flora and non-targeted analysis methods to measure metabolomics; we then compared the differences in the gut microbiota between each group. The random forest method was used to explore the non-invasive diagnostic value of the gut microbiome in IgAN. We also compared serum metabolites and analyzed their correlation with the gut microbiome.

Results: The richness and diversity of gut microbiota were significantly different among IgAN, n_IgAN and HC patients. Using a random approach, we constructed the diagnosis model and analysed the differentiation between IgAN and n_IgAN based on gut microbiota. The area under the receiver operating characteristic curve for the diagnosis was 0.9899. The metabolic analysis showed that IgAN patients had significant metabolic differences compared with HCs. In IgAN, catechol, l-tryptophan, (1H-Indol-3-yl)-N-methylmethanamine, and pimelic acid were found to be enriched. In the correlation analysis, l-tryptophan, blood urea nitrogen and Eubacterium coprostanoligenes were positively correlated with each other.

Conclusion: Our study demonstrated changes in the gut microbiota and established models for the non-invasive diagnosis of IgAN from HC and n_IgAN. We further demonstrated a close correlation between the gut flora, metabolites, and clinical phenotypes of IgAN. These findings provide further directions and clues in the study of the mechanism of IgAN.

Keywords: IgA; chronic kidney disease; glomerulonephritis; gut microbiome; metabolic networks; non‐invasive diagnostic tools.