Dynamic changes of gut microbiota between the first and second trimester for women with gestational diabetes mellitus and their correlations with BMI: a nested cohort study in China

Abstract

Introduction: Gut microbiota (GM) has been implicated in gestational diabetes mellitus (GDM), yet longitudinal changes across trimesters remain insufficiently explored.

Methods: This nested cohort study aimed to investigate GM alterations before 24 weeks of gestation and their association with GDM. Ninety-three Chinese participants provided fecal samples during the first and second trimesters. Based on oral glucose tolerance tests, 11 participants were classified as GDM, and 82 as non-diabetic (ND). Using 16S rRNA sequencing, we analyzed both cross-sectional and longitudinal differences in GM structure between those two groups.

Results: In the first trimester, GDM group exhibited lower levels of Bacteroides_H and Acetatifactor compared to ND group (p < 0.05). In the second trimester, GDM individuals showed increased abundance of Fusobacteriota and Firmicutes_D, and genera including Fusobacterium_A and Fournierella, while Anaerotruncus and others decreased (P<0.05). Inflammation-associated genera like Gemmiger_A_73129 and Enterocloster increased, while Megamonas decreased in overweight or obese GDM women, which was not identified in normal-weight women. The ratios of relative abundance of genera Streptococcus, Enterocloster, and Collinsella exceeded 1.5 in the GDM group, particularly in overweight or obese individuals. Inflammatory pathways related to African trypanosomiasis and Staphylococcus aureus infection were predicted to be up-regulated in overweight or obese GDM individuals but not in normal-weight GDM women.

Discussion: This study suggests that GM of women with GDM undergoes significant alterations between the first and second trimesters, potentially linked to inflammation, with more pronounced changes observed in overweight or obese individuals.

Keywords: body mass index; gestational diabetes mellitus; gut microbiota; inflammation; obesity.

Figure 1

The flow chart of this study. GDM, gestational diabetes mellitus; ND, non-diabetics; HbA1c, glycated hemoglobin A1c; OGTT, oral glucose tolerance test; TBA, total bile acid; LEfSe, Linear discriminant analysis Effect Size; PICRUSt, Phylogenetic Investigation of Communities by Reconstruction of Unobserved States; KEGG, Kyoto Encyclopedia of Genes and Genomes.

Figure 2

Comparison of the structure and diversity of gut microbiota between the GDM and ND groups in the first trimester. The dominant phyla (A) and genera (B) differed between ND and GDM groups in FT. Bacteroides_H (C), Acetatifactor (D), and Megasphaera_A_38685 (E) were significantly lower in the GDM group (p < 0.05). No significant differences were found in the Chao (F) or Shannon indices (G). NMDS (H) and PCoA (I) showed no distinct microbiota separation. The volcano plot (J) highlighted six genera that increased and one that decreased significantly in the GDM group. LEfSe analysis (K) identified one biomarker for GDM and five for ND in FT. The Circos plot (L) showed a closer relationship between Acetatifactor, Scybalousia, Megasphaera_A_38685, and Burkholderia with ND. Significant differential genus related to GDM were identified in participants with normal weight (M), and those overweight or obese (O). LEfSe analysis also pinpointed biomarkers for both groups among participants who were normal weight (N) and those overweight or obese (P). * p < 0.05.

Figure 3

Comparison of the structure and diversity of gut microbiota between the GDM and ND groups in the second trimester. The proportion of dominant phyla (A) and genera (B) varied between the ND and GDM groups in the second trimester (ST). At the phylum level, Fusobacteriota (C) and Firmicutes_D (D) significantly decreased in GDM group. At the genus level, Fusobacterium_A (E), Scatomonas (F), and Fournierella (G) significantly increased, whereas Anaerotruncus (H), Coprobacter (I), and Angelakisella (J) significantly decreased in GDM group. No significant differences were noted in the Chao (K) or Shannon indices (L). NMDS (M) and PCoA (N) showed no microbiota separation between groups. The volcano plot (O) showed two genera increased and twelve decreased in GDM. LEfSe analysis (P) identified thirteen GDM biomarkers and fifteen ND biomarkers. The Circos plot (Q) highlighted close relationships between Butyribacter, Duncaniella, and Allisonella with GDM. The volcano plot showed no significant differential genera related to GDM in normal-weight participants (R), but overweight or obese women had increases in Enterocloster, Gemmiger_A_73129, and nine other genera, and decreases in Megamonas, Bacteroides_H, and three others (T). LEfSe analysis identified biomarker for the GDM group and ND group in the participants with normal weight (S), overweight or obese (U). *p < 0.05.

Figure 4

Variations in the relative abundance of dominant phyla and genera between the GDM and ND groups. The VRAs of Eubacterium_R (A), Allisonella (B), Fournierella (C), Intestinibacter (D) and CAG_41 (E) increased, while Angelakisella (F) decreased in GDM group. Overweight or obese GDM women showed increased VRA of Megamonas (G), whereas normal-weight GDM women had higher VRA of Faecalibacterium (H) and lower VRA of Escherichia_710834 (I). Phyla Fusobacteriota, Verrucomicrobiota, and Desulfobacterota_I had RRAs over 1.5, while Bacteroidota was below 0.8 (J). Genera Staphylococcus, Streptococcus, Enterocloster, Collinsella, and Faecalibacillus exceeded 1.5, with Phocaeicola_A_858004 and Bifidobacterium_388775 below 0.8 (K). In normal-weight women, Staphylococcus, Enterocloster, and Collinsella had RRAs less than 0.8, with Streptococcus, Mediterraneibacter_A_155507, and Faecalibacillus between 0.8 and 1.5, and only CAG_269 and Megamonas above 1.5 (L). In overweight or obese women, RRAs of CAG_269, Streptococcus, Enterocloster, Mediterraneibacter_A_155507, and Collinsella exceeded 1.5 (M). RA, relative abundance; VRA, variance of the relative abundance from the first to the second trimester; RRA, ratio of the relative abundance. * p < 0.05.

Figure 5

The relationship of variances of relative abundance with the blood glucose and HblAc. The significantly positive relationship (red) and negative relationship (dark turquoise) were label by asterisk. OGTT, oral glucose tolerance test; HblAc, glycated hemoglobin A1c; FPG, fasting plasma glucose; * p < 0.05, **p < 0.01.

Figure 6

The KEGG pathways predicted by PICRUSt2 differed between the GDM and ND groups. The up-regulated KEGG pathways (red) and down-regulated pathways (dark turquoise) were predicted in the total participants of the first trimester (A), second trimester (B) and the overweight or obese participants in the second trimester (C). ND, non-diabetic; GDM, gestational diabetes mellitus; FT, first trimester; ST, second trimester; ECM, extracellular matrix. * p < 0.05, **p < 0.01.