Obesity-driven mitochondrial dysfunction in human adipose tissue-derived mesenchymal stem/stromal cells involves epigenetic changes

Abstract

Obesity exacerbates tissue degeneration and compromises the integrity and reparative potential of mesenchymal stem/stromal cells (MSCs), but the underlying mechanisms have not been sufficiently elucidated. Mitochondria modulate the viability, plasticity, proliferative capacity, and differentiation potential of MSCs. We hypothesized that alterations in the 5-hydroxymethylcytosine (5hmC) profile of mitochondria-related genes may mediate obesity-driven dysfunction of human adipose-derived MSCs. MSCs were harvested from abdominal subcutaneous fat of obese and age/sex-matched non-obese subjects (n = 5 each). The 5hmC profile and expression of nuclear-encoded mitochondrial genes were examined by hydroxymethylated DNA immunoprecipitation sequencing (h MeDIP-seq) and mRNA-seq, respectively. MSC mitochondrial structure (electron microscopy) and function, metabolomics, proliferation, and neurogenic differentiation were evaluated in vitro, before and after epigenetic modulation. hMeDIP-seq identified 99 peaks of hyper-hydroxymethylation and 150 peaks of hypo-hydroxymethylation in nuclear-encoded mitochondrial genes from Obese- versus Non-obese-MSCs. Integrated hMeDIP-seq/mRNA-seq analysis identified a select group of overlapping (altered levels of both 5hmC and mRNA) nuclear-encoded mitochondrial genes involved in ATP production, redox activity, cell proliferation, migration, fatty acid metabolism, and neuronal development. Furthermore, Obese-MSCs exhibited decreased mitochondrial matrix density, membrane potential, and levels of fatty acid metabolites, increased superoxide production, and impaired neuronal differentiation, which improved with epigenetic modulation. Obesity elicits epigenetic changes in mitochondria-related genes in human adipose-derived MSCs, accompanied by structural and functional changes in their mitochondria and impaired fatty acid metabolism and neurogenic differentiation capacity. These observations may assist in developing novel therapies to preserve the potential of MSCs for tissue repair and regeneration in obese individuals.

Fig. 1. Obesity alters 5-hydroxymethylcytosine (5hmC) levels of mitochondria-associated genes in human adipose tissue MSCs.

A Volcano plot of mitochondria-related genes with significant changes in 5hmC levels between Non-obese- and Obese-MSCs (n = 5 each). The y-axis corresponds to −log₂ (p value), whereas the x-axis displays the log₂ fold-change (Obese-MSCs/Non-obese-MSCs) value. Hyper-hydroxymethylated (n = 99; 89 genes) and hypo-hydroxymethylated (n = 150; 134 genes) peaks are indicated with red and blue dots, respectively. Cutoff values of p ≤ 0.05 and log₂ fold-changes ≥0.5 or ≤−0.5 are indicated by gray dashed lines. B Genomic location annotations of hyper- and hypo-hydroxymethylated peaks. C Distribution across the gene body relative to the transcription start site (TSS).

Fig. 2. mRNA-seq and integrated hMeDIP-seq/mRNA-seq analysis.

A Heat maps of mitochondria-associated genes upregulated (left) or downregulated (right) in Obese-MSCs compared to Non-obese-MSCs (n = 5 each). B Venn diagrams showing 4 genes (SLC25A4, SLC22A4, COQ10B, and LAMC1) with hyper-hydroxymethylated peaks that were upregulated (red) and 3 genes (COASY, ECI1, and MECR) with hypo-hydroxymethylated peaks that were downregulated (blue) in Obese-MSCs versus Non-obese-MSCs (n = 5 each).

Fig. 3. Visualization of 5hmC peaks.

Representative integrative genomics viewer (IGV) tracks showing hyper-hydroxymethylated and hypo-hydroxymethylated peaks of SLC25A4 (A), COQ10B (B), and COASY (C) in Obese-MSCs (purple) versus Non-obese-MSCs (green) (n = 5 each). All IGV tracks in a given comparison have the same scaling factor for the y-axis, and the scale of the x-axis is indicated in the upper right-hand region of each set of tracks. The region of the genome identified as differentially hyper- or hypo-hydroxymethylated is indicated by a red rectangle through the tracks. The RefSeq gene map is presented in blue at the bottom of each panel showing the overall gene structure.

Fig. 4. Obesity impairs mitochondrial structure in human MSCs.

Representative transmission electron microscopy images (A) and quantification of mitochondrial area (B) and matrix density (C) in Non-obese- and Obese-MSCs untreated or treated with Bobcat339 (n = 5 each). *p value < 0.05 vs. Non-obese-MSCs (untreated).

Fig. 5. Obesity impairs mitochondrial function in human MSCs.

Representative MitoSOX and tetramethylrhodamine ethyl ester (TMRE) staining (A) and quantification of mitochondrial reactive oxygen species (ROS) production (B), membrane potential (C), and ATP generation (D) in Non-obese- and Obese-MSCs untreated or treated with Bobcat339 or dimethyl alpha-ketoglutarate (DMαKG) (n = 5 each). *p value < 0.05 vs. non-obese-MSCs (untreated).

Fig. 6. Obesity impairs fatty acid metabolism of human MSCs.

Proliferation (A percent phase object confluence unit) and migration (B Colorimetric Cell Assay) of Non-obese and Obese-MSCs (n = 5 each). C Liquid chromatography-tandem mass spectrometry (LC-MS/MS) metabolomic analysis of the fatty acid (FA) metabolites eicosapentaenoic acid (EPA), docosahexaenoic acid (DHA), and linolenic, myristic, palmitoleic, arachidonic, linoleic, palmitic, oleic, elaidic, and stearic acids in non-obese- and Obese-MSCs untreated or treated with Bobcat339 or DMαKG (n = 5 each). *p value < 0.05 vs. Non-obese-MSCs (untreated). NS: non-significant (p > 0.05).

Fig. 7. Obesity impairs neuronal differentiation of human MSCs.

A Representative immunofluorescence β-III tubulin staining (green) and Nissl body staining (dark black-violet), and their quantification (B-C, respectively) in Non-obese- and Obese-MSCs untreated or treated with Bobcat339 or dimethyl alpha-ketoglutarate (DMαKG) (n = 5 each). *p-value < 0.05 vs. Non-obese-MSCs (untreated).