Gut microbiota-metabolite interactions meditate the effect of dietary patterns on precocious puberty

Abstract

Precocious puberty, a pediatric endocrine disorder classified as central precocious puberty (CPP) or peripheral precocious puberty (PPP), is influenced by diet, gut microbiota, and metabolites, but the specific mechanisms remain unclear. Our study found that increased alpha-diversity and abundance of short-chain fatty acid-producing bacteria led to elevated levels of luteinizing hormone and follicle-stimulating hormone, contributing to precocious puberty. The integration of specific microbiota and metabolites has potential diagnostic value for precocious puberty. The Prevotella genus-controlled interaction factor, influenced by complex carbohydrate consumption, mediated a reduction in estradiol levels. Interactions between obesity-related bacteria and metabolites mediated the beneficial effect of seafood in reducing luteinizing hormone levels, reducing the risk of obesity-induced precocious puberty, and preventing progression from PPP to CPP. This study provides valuable insights into the complex interplay between diet, gut microbiota and metabolites in the onset, development and clinical classification of precocious puberty and warrants further investigation.

Keywords: Developmental biology; Microbiology; Microbiome.

Graphical abstract

Figure 1

Gut microbiome dysbiosis in girls suffering from CPP and PPP (A) Stack bar plot showing the bacterial composition at the species level in the PPP, CPP, and normal groups. Relative abundance of the top 10 most abundant bacterial species within each group were displayed. The category “Others” represents the combined abundance of all remaining species. (B) Alpha diversity of the gut microbiota (measured by the observed metric) between groups. (C–G) PCoA of beta diversity based on the Bray-Curtis distance between groups with the PERMANOVA test. Linear discriminant analysis Effect Size analysis results showed the differential abundance of species between PPP and normal groups (D) as well as CPP and normal groups (E). The significant differential KEGG pathways between CPP and normal groups (F) as well as PPP and normal groups (G). Statistical significance was indicated by asterisks (∗: p < 0.05, ∗∗: p < 0.01). CPP, central precocious puberty; PPP, peripheral precocious puberty; PCoA, principal co-ordinates analysis; PERMANOVA, permutational multivariate analysis of variance; LDA, Linear discriminant analysis; KEGG, Kyoto Encyclopedia of Genes and Genomes. See also Tables S1 and S2.

Figure 2

Microbiome clusters based on dietary patterns correlated with PPP and CPP (A) Forest plot displayed the significance and OR value of latent dietary patterns associated with CPP and PPP. (B) Venn diagram of common latent dietary patterns associated with CPP and PPP. (C) Hierarchical clustering map of the first two Sparse partial least squares (sPLS) dimensions, displaying pairwise correlations between 287 amplicon sequence variants (ASVs) and 11 dietary patterns and 4 clinical variables. Correlations greater than 0.25 or less than −0.25 were shown in red and blue, respectively. (D) The correlation circle plot for the first two sPLS dimensions displayed correlations greater than 0.25 or less than −0.25. The two black circles indicate correlation coefficient radii at 0.5 and 1.0. Bacterial ASVs are displayed as circles, and are colored based on their affiliated cluster (cluster 1: blue; cluster 2: red). (E) Canonical correspondence analysis (CCpnA) of gut microbiota (data not shown), dietary patterns and clinical variables. It showed dimension 1 and 2 of the CCpnA, which included continuous clinical variables (arrows), categorical variables (+) and samples (triangles). The samples in the CCpnA plot were colored based on their groups, and the ellipses presented an 80% confidence interval. CPP, central precocious puberty; PPP, peripheral precocious puberty; OR, odd ratios; LH, luteinizing hormone; FSH, follicle-stimulating hormone; E2, estradiol; Mg, magnesium. See also Figure S1, Tables S3 and S4.

Figure 3

Effect and diagnostic potential of microbiota-metabolite factors for PPP and CPP (A) Volcano plots were used to visualize the up-regulated (represented by red dots) and down-regulated metabolites (represented by blue dots) in CPP and PPP compared to the normal groups, with statistical significance represented by p < 0.05 and |log2FoldChange|>1. Metabolites from different sources were labeled with different colors. (B) The proportion of variance explained by microbiome and metabolome in microbiota-metabolite factors including Factor 2, Factor 4, and Factor 5. (C–E) Boxplots were used to display the factor values of Factor 2, Factor 4, and Factor 5 in CPP, PPP, and normal groups. The macro- and micro-average area under the receiver operating characteristic curve (AUC) for values of Factor 2 (D), Factor 4 (E), and Factor 5 (F) demonstrated excellent diagnostic value for PPP and CPP. Statistical significance was indicated by asterisks (∗: p < 0.05, ∗∗: p < 0.01, ∗∗∗: p < 0.001). CPP, central precocious puberty; PPP, peripheral precocious puberty; Non-Diff, non-differential; AUC, area under the curve. See also Figures S2 and S3, Table S5. Differentially abundant metabolites in CPP and PPP compared to normal controls, related to Figure 3, Table S6. The weights of microbiota in factors, related to Figure 3, Table S7. The weights of metabolites in factors, related to Figure 3.

Figure 4

Dietary pattern-microbiota-metabolite interaction associated with precocious puberty (A) Sankey plot was used to illustrate the Spearman correlation between dietary patterns, microbiota-metabolite factors, and clinical variables associated with precocious puberty. In this plot, the nodes represent dietary patterns, microbiota-metabolite factors, and clinical variables related to precocious puberty. The red and blue links indicate positive and negative correlations between the nodes, respectively. (B) Results of the meditation analysis: Nodes represented by ellipses, diamonds and rectangles correspond to dietary patterns, microbiota-metabolite factors and clinical variables, respectively. Edges between nodes are color coded in red and blue to represent positive and negative effects. BA, bone age; LH, luteinizing hormone; FSH, follicle-stimulating hormone; E2, estradiol; TES, testosterone. See also Table S8.