Multiomics reveals glutathione metabolism as a driver of bimodality during stem cell aging

Abstract

With age, skeletal muscle stem cells (MuSCs) activate out of quiescence more slowly and with increased death, leading to defective muscle repair. To explore the molecular underpinnings of these defects, we combined multiomics, single-cell measurements, and functional testing of MuSCs from young and old mice. The multiomics approach allowed us to assess which changes are causal, which are compensatory, and which are simply correlative. We identified glutathione (GSH) metabolism as perturbed in old MuSCs, with both causal and compensatory components. Contrary to young MuSCs, old MuSCs exhibit a population dichotomy composed of GSH^high cells (comparable with young MuSCs) and GSH^low cells with impaired functionality. Mechanistically, we show that antagonism between NRF2 and NF-κB maintains this bimodality. Experimental manipulation of GSH levels altered the functional dichotomy of aged MuSCs. These findings identify a novel mechanism of stem cell aging and highlight glutathione metabolism as an accessible target for reversing MuSC aging.

Keywords: GSH; MuSC; NAC; aging; bimodality; multiomics; satellite cells; stem cells.

Figure 1.. Multiomics identifies glutathione metabolism as altered with age in quiescent MuSCs.

A) PCA of RNA-seq of quiescent MuSCs from young (n=4) and old (n=4) mice. Each replicate, shown as a single triangle, represents MuSCs from one mouse. B) Volcano plot of genes profiled by RNA-seq. Genes involved in glutathione metabolism and significantly different at FDR 5% are indicated in red. C) PCA of shotgun proteomics of quiescent MuSCs of young (n=3) and old (n=2) biological replicates. Each replicate, shown as a single triangle, represents MuSCs pooled from 16–20 mice. D) Heatmap of proteins profiled by shotgun proteomics. To facilitate visualization of genes of interest, shown are the proteins for which replicates showed reproducible behavior. Proteins involved in glutathione metabolism are labelled. E) Hierarchical clustering of metabolomics of quiescent MuSCs of young (n=8) and old (n=12) biological replicates. Each replicate, shown as a single leaf, represents MuSCs pooled from 2–6 mice. F) M-A plot of metabolites profiled. Metabolites that are significantly different at FDR 30% are colored blue; metabolites involved in glutathione metabolism that are significantly different at this cutoff are highlighted in red. G) Kernel density plot of methylation levels of ~33 million individual CpGs profiled in young and old MuSCs (n=4 individual mice in each age group). CpGs significantly different at FDR 5% (~80 thousand) are plotted in red. H) Each DMR is shown as a point displaying young and old mean methylation levels. DMRs associated with genes are blue; DMRs associated with glutathione metabolism genes are red. I) Simplified diagram of mammalian glutathione metabolism depicting changes that occur with age. Boxes are genes products; circles are metabolites. Each element is colored by the old vs. young change that occurs with age; RNA changes are on the left, and protein changes are on the right. Genes with DNA methylation changes are bolded. Gray indicates no data. Only genes that are expressed in the RNA or protein datasets are shown. J) Consensus ranking of KEGG Pathways generated by combining the ranked pathway lists of individual datasets through rank aggregation. Pathways, ordered by consensus rank, have individual ranks across datasets shown in different colors. The p-value represents the significance of the rank vector (not the consensus rank). The left shows all the pathways for which consensus ranks were generated; the right shows the top ten.

Figure 2.. The bulk population of old MuSCs is comprised of GSH^high and GSH^low populations with differential functionalities.

A) Representative flow cytometry contour plots (top) and histograms (bottom) from freshly isolated MuSCs from young (4-month) and old (24-month) mice incubated for 30 minutes with ThiolTracker Violet. The GSH^high population is shaded red, and the GSH^low population is shaded blue. B) Percentage of GSH^high and GSH^low cells in freshly isolated MuSCs from young and old mice (n=5–6). C) Percentage of GSH^high and GSH^low cells in freshly isolated MuSCs from mice over an age gradient (n=4–6). D) EdU incorporation in GSH^high and GSH^low MuSCs isolated from 24-month-old mice and maintained in culture for 48 hours in the continuous presence of EdU (n=4). E) Cell death (PI staining) in GSH^high and GSH^low MuSCs isolated from 24-month-old mice and maintained in culture for 48 hours (n=4). F) Seahorse assay comparing 2×10⁵ GSH^high and GSH^low MuSCs isolated from 24-month-old mice. Each replicate represents a pool of 4–5 mice (n=2 replicates). G) Representative images (left) of YFP-expressing GSH^high and GSH^low MuSCs that were transplanted into pre-injured and pre-irradiated TA muscles. Cryosections, harvested 28 days after transplantation, were analyzed for YFP expression, and the number of YFP-positive fibers was quantified (right). (n=5). Error bars represent SEM. *p < 0.05; **p < 0.01.

Figure 3.. GSH levels causally determine MuSC function.

A) Representative images from freshly isolated GSH^high MuSCs, GSH^low MuSCs, and GSH^low MuSCs treated for 48 hours with either 2 mM NAC or the combination of 2 mM NAC with 1 mM BSO (an inhibitor of GCS). All cells were grown in culture for 48 hours in the continuous presence of EdU. B-C), Quantification of EdU-positive (b) and TUNEL-positive (c) cells from the experiment as described in (a) (n=4). D EdU incorporation in MuSCs isolated from young (4-month-old) mice treated with either vehicle, 10 μM 6-AN, or the combination of 10 μM 6-AN plus 2 mM NAC for 48 hours. All cells were grown in culture for 48 hours in the continuous presence of EdU. E) Representative flow cytometry plot displaying ThiolTracker intensity in young MuSCs treated with either vehicle or 10 μM 6-AN for 48 hours as well as old MuSCs treated with either vehicle or with 2 mM NAC for 48 hours. F) In vivo MuSC EdU incorporation in mice injured with BaCl₂ and subsequently injected intramuscularly with either 150 mg/kg NAC or vehicle. 2 days after injury, mice were injected with 50 mg/kg EdU i.p., and MuSCs were isolated 12 hours after EdU administration (n=6–8). Error bars represent SEM. *p < 0.05; **p < 0.01; ***p < 0.001; NS, not significant.

Figure 4.. Transcriptomic profiling of GSH^high and GSH^low MuSCs reveals an NF-κB-mediated failure of GSH^low cells to engage compensatory GSH biosynthetic pathways.

A) PCA of RNA-seq profiles of freshly isolated GSH^high, GSH^low, and total young MuSCs. Each biological replicate (triangle) represents MuSCs pooled from 3–4 mice (n=2 biological replicates). B) GSEA signature plots for inflammation and cell-cycle gene sets with genes ranked by PC1 (“biological age axis”) loading. ES: enrichment score, NES: normalized enrichment score. C) GSEA signature plot for a gene set representing GSH metabolism and the general xenobiotic response with genes ranked by PC2 (“glutathione turnover axis”) loading. D) RT-qPCR for G6PDx (left) and GCLM (right) in GSH^high, GSH^low, and total young MuSCs. MuSCs were derived from mice independent from those used for the RNA-seq experiment. Ct values were normalized first to GAPDH and then to the mean of the young levels (n=4). E) Western blot analysis (top) and quantification (bottom) of GCLM protein in GSH^high and GSH^low MuSCs isolated from 27-month-old mice. Total Histone H3 was used as a loading control. Each replicate represents a pool of 3–4 mice (n=3 replicates) F-G) GSEA results using predicted conserved transcription factor target (TFT) gene sets in the MSigDB. Predicted Nrf2 targets are enriched among GSH^high-upregulated genes, and predicted NF-ᴋB targets are enriched among GSH^low-upregulated genes. H) EdU incorporation in MuSCs isolated from mice treated for 5 days with either 2 mg/kg/day of the NF-κB inhibitor JSH-23 or vehicle. MuSCs were maintained in culture for 48 hours in the continuous presence of EdU (n=4). I) Cell death as the percentage of cells that are PI-positive from the MuSCs described in (g) following 48 hours in culture. J) Percentage of GSH^high and GSH^low MuSCs using the ThiolTracker probe from the experiment described in (g). Error bars represent SEM. *p < 0.05; **p < 0.01; NS, not significant.

Figure 5.. NF-κB AMO treatment and old serum transfer can alter the proportion of GSH^high and GSH ^low MuSCs

A) Representative FACS plots measuring ThiolTracker content of MuSCs isolated from 27-month-old mice treated with control AMO (top) or RelA AMO (bottom). B) Percentage of GSH^high and GSH^low MuSCs isolated from 27-month-old mice treated with control AMO or RelA AMO (n=4). C) Percentage of GSH^high and GSH^low MuSCs isolated from young mice injected with young serum or old serum (n=4). Error bars represent SEM. *p < 0.05