Transcriptomic Profiling of Rectus Abdominis Muscle in Women with Gestational Diabetes-Induced Myopathy: Characterization of Pathophysiology and Potential Muscle Biomarkers of Pregnancy-Specific Urinary Incontinence

Abstract

Gestational diabetes mellitus (GDM) is recognized as a "window of opportunity" for the future prediction of such complications as type 2 diabetes mellitus and pelvic floor muscle disorders, including urinary incontinence and genitourinary dysfunction. Translational studies have reported that pelvic floor muscle disorders are due to a GDM-induced-myopathy (GDiM) of the pelvic floor muscle and rectus abdominis muscle (RAM). We now describe the transcriptome profiling of the RAM obtained by Cesarean section from GDM and non-GDM women with and without pregnancy-specific urinary incontinence (PSUI). We identified 650 genes in total, and the differentially expressed genes were defined by comparing three control groups to the GDM with PSUI group (GDiM). Enrichment analysis showed that GDM with PSUI was associated with decreased gene expression related to muscle structure and muscle protein synthesis, the reduced ability of muscle fibers to ameliorate muscle damage, and the altered the maintenance and generation of energy through glycogenesis. Potential genetic muscle biomarkers were validated by RT-PCR, and their relationship to the pathophysiology of the disease was verified. These findings help elucidate the molecular mechanisms of GDiM and will promote the development of innovative interventions to prevent and treat complications such as post-GDM urinary incontinence.

Keywords: gestational diabetes mellitus (GDM); gestational diabetic-induced myopathy (GDiM); pregnancy-specific urinary incontinence (PSUI); rectus abdominis muscle (RAM); transcriptomic profile.

Figure 1

The workflow of the present study.

Figure 2

Heatmap showing relative gene expressions. The heat map depicts the relative expressions of each probe set (row) for each woman. The intensity of each block, either yellow (higher expression) or green (lower expression), represents the magnitude of difference from the mean.

Figure 3

Profiling of differentially expressed genes (DEGs) of non-diabetic continent women (ND-C) in comparison with gestational diabetic and incontinent women (GDM-I). (a) Biological processes identified for up- and down regulated genes. Enrichment was defined as the 10 most significant terms according to the highest scores and p-values (<0.05). (b) Protein–protein interaction network representing DEGs. The interaction map was generated using STRING with clusters, a high confidence of 0.7, and all criteria for linkage (co-occurrence, co-expression, experiments, neighborhood, databases, text-mining, and homology). The genes used in this network are listed in the Supplemental Table S1.

Figure 4

Profiling of differentially expressed genes (DEGs) of non-diabetic incontinent women (ND-I) in comparison with gestational diabetic and incontinent women (GDM-I). (a) Biological processes were identified for up- and downregulated genes. Enrichment was defined as the 10 most significant terms according to the highest scores and p-values (<0.05). (b) Protein–protein interaction network representing DEGs. The interaction map was generated using STRING with clusters, a high confidence of 0.7, and all criteria for linkage (co-occurrence, co-expression, experiments, neighborhood, databases, text-mining, and homology). The genes used in this network are listed in Supplemental Table S1.

Figure 5

The profiling of differentially expressed genes (DEGs) of gestational diabetic continent women (GDM-C) in comparison with gestational diabetic and incontinent women (GDM-I). (a) Biological processes were identified for up- and downregulated genes. Enrichment was defined as the 10 most significant terms according to the highest scores and p-values (<0.05). (b) Protein–protein interaction network representing DEGs. The interaction map was generated using STRING with clusters, a high confidence of 0.7, and all criteria for linkage (co-occurrence, co-expression, experiments, neighborhood, databases, text-mining, and homology). The genes used in this network are listed in Supplemental Table S1.

Figure 6

Prognostic values of potential muscle biomarker genes for gestational diabetic-induced myopathy. (a) Venn diagram showing the intersection of eight genes expressed differentially in the GDM-I group (GDM-I): ATP2A2, EEF1A1, EIF1, G0S2, MYH7, NACA, TPM3, and UBC. (b) Balloon plot showing the expression profiles of these genes, using the count/percentage of RNA-seq. ND-C: non-GDM continent women (no PSUI); ND-I: non-GDM incontinent women (with PSUI); GDM-C: GDM continent women (no PSUI); GDM-I: GDM incontinent women (with PSUI).

Figure 7

Real-time RT-PCR analysis of eight differentially expressed genes in RAM. Real-time RT-PCR was carried out on six independent biological replicates per group containing three replicates. The relative quantification of each transcript was normalized against β-actin and GAPDH. In the featured line, there are genes that were differentially expressed by RT-PCR in a larger group of analyzed women. * ND-C: non-GDM continent women (no PSUI); ND-I: non-GDM incontinent women (with PSUI); GDM-C: GDM continent women (no PSUI); GDM-I: GDM incontinent women (with PSUI).

Figure 8

Relative heatmaps of RNA-seq (a) and qRT-PCR (b) used to validate and visualize biomarker candidates. Gene expressions from the two platforms were normalized by the quantile normalization method. ND-C: non-GDM continent women (no PSUI); ND-I: non-GDM incontinent women (with PSUI); GDM-C: GDM continent women (no PSUI); GDM-U: GDM incontinent women (with PSUI).