Electroacupuncture Mimics Exercise-Induced Changes in Skeletal Muscle Gene Expression in Women With Polycystic Ovary Syndrome

Abstract

Context: Autonomic nervous system activation mediates the increase in whole-body glucose uptake in response to electroacupuncture but the mechanisms are largely unknown.

Objective: To identify the molecular mechanisms underlying electroacupuncture-induced glucose uptake in skeletal muscle in insulin-resistant overweight/obese women with and without polycystic ovary syndrome (PCOS).

Design/participants: In a case-control study, skeletal muscle biopsies were collected from 15 women with PCOS and 14 controls before and after electroacupuncture. Gene expression and methylation was analyzed using Illumina BeadChips arrays.

Results: A single bout of electroacupuncture restores metabolic and transcriptional alterations and induces epigenetic changes in skeletal muscle. Transcriptomic analysis revealed 180 unique genes (q < 0.05) whose expression was changed by electroacupuncture, with 95% of the changes towards a healthier phenotype. We identified DNA methylation changes at 304 unique sites (q < 0.20), and these changes correlated with altered expression of 101 genes (P < 0.05). Among the 50 most upregulated genes in response to electroacupuncture, 38% were also upregulated in response to exercise. We identified a subset of genes that were selectively altered by electroacupuncture in women with PCOS. For example, MSX1 and SRNX1 were decreased in muscle tissue of women with PCOS and were increased by electroacupuncture and exercise. siRNA-mediated silencing of these 2 genes in cultured myotubes decreased glycogen synthesis, supporting a role for these genes in glucose homeostasis.

Conclusion: Our findings provide evidence that electroacupuncture normalizes gene expression in skeletal muscle in a manner similar to acute exercise. Electroacupuncture might therefore be a useful way of assisting those who have difficulties performing exercise.

Trial registration: ClinicalTrials.gov NCT01457209.

Keywords: PCOS; epigenetics; transcriptomics.

© Endocrine Society 2020.

Figures

Figure 1.

Electroacupuncture induces similar gene expression response in human and rat skeletal muscle. (a) Overview of the human study design and procedures: single bout of electroacupuncture in women with and without PCOS. (b) The 6 genes with the largest increase in mRNA expression and the 5 selected genes involved in muscle function and metabolism in skeletal muscle in response to a single bout of electroacupuncture (EA) in 15 women with PCOS, based on linear regression analyses (q < 0.05); data presented as mean ± SD. All differentially expressed genes identified by genome-wide arrays are presented in Supplementary Table S2 (30). (c) Skeletal muscle (EDL) gene expression in rat after a single bout of electroacupuncture (n = 9) and nonstimulated controls (n = 5); data presented as mean ± SEM of fold change (fc) and analyzed with Mann-Whitney U test; ##P < 0.01, ###P < 0.001.

Figure 2.

More than 50% of the unique transcripts with a change in expression in response to electroacupuncture had one or several CpG sites annotated to the gene with a change in DNA methylation. CpG sites (P < 0.05) with a change in DNA methylation annotated to genes with altered expression (q < 0.05). Values are change before vs after electroacupuncture presented as % points for CpG site methylation and % change in gene expression presented next to each gene name. All altered CpG methylation sites of genes with changes in mRNA expression in response to electroacupuncture were calculated based on linear regression analyses and are presented in Supplementary Table S4 (30).

Figure 3.

Transcription factors could be direct regulators of gene sets altered in women with PCOS or in response to electroacupuncture or exercise. (a) Motif enrichment analysis of the top 100 genes with the largest fold increase in each condition, and Supplementary Table S6 (30). Fold change (Fc) of transcription factor gene expression in response to (b) electroacupuncture, (c) exercise, and (d) PCOS (q < 0.05) based on linear regression analyses.

Figure 4.

One bout of electroacupuncture or exercise shows similarities in gene expression changes and signaling pathways. Comparison of genes and pathways responding to electroacupuncture, exercise and PCOS. (a) Number of genes significantly modified by electroacupuncture (q < 0.05), exercise (q < 0.05), and PCOS (P < 0.05) were overlapped in a Venn diagram. (b) Correlation of fold-changes of genes significantly modified by electroacupuncture (q < 0.05) with genes changed in women with PCOS. Spearman r = -0.79. (c) Correlation of fold-changes of genes significantly modified by exercise (meta-analysis (37), q < 0.05) with genes modified by electroacupuncture in the present study. Spearman r = 0.59. (d) Overrepresentation analysis using gene ontology on genes significantly modified by electroacupuncture (q < 0.05), exercise (q < 0.05) or PCOS (P < 0.05).

Figure 5.

Many genes regulated in response to electroacupuncture changed toward a healthier phenotype. Genes differentially expressed in muscle from PCOS vs controls (controls are set to 1) before electroacupuncture (EA)(*P < 0.05), and genes changing after a single bout of acupuncture (#q < 0.05). Data presented as fold change mean ± SD and analyzed with Student t test. Overlap between all differentially expressed genes are presented in Supplementary Table S7 (30).

Figure 6.

Twenty-five percent of validated genes were upregulated in response to electrical stimulation in women with and without PCOS, and in myotubes following EPS. (a-d) Effect of electrical stimulation on genes with the largest changes in mRNA expression in PCOS, controls and myotubes (targeted approach); CCL2, GADD45B, SLC2A3, LDLR. (e-i) Genes regulated in opposite directions in response to electrical stimulation in PCOS (n = 15) and controls (n = 14); PRINS, AKAP13, HOMOX1, MSX1, SRNX1. *P < 0.05, ***P < 0.001 vs control before EA, #P < 0.05, ##P < 0.01, ###P < 0.001 vs PCOS before EA, ∧P < 0.05, ∧∧P < 0.01, ∧∧∧P < 0.001 vs unstimulated myotubes, data were analyzed by Student t test and presented as mean ± SEM. All investigated genes in this targeted approach are presented in Supplementary Table S1 (30). (j) Correlation of fold-changes of genes significantly modified in myotubes following EPS (P < 0.05) with genes changed in women by electroacupuncture. Spearman r = 0.46, P < 0.001. (k-n) Insulin stimulated glycogen incorporation in myotubes from healthy controls after gene silencing using siRNA for AKAP13, HMOX1, MSX1, SRXN1 (n = 5-7). Data were analyzed by two-way ANOVA and presented as individual values, *P < 0.05 vs scramble control at baseline, and after 10 nM and 100 nM insulin stimulation.