Glutamine-driven Reductive TCA cycle metabolism supports aged muscle stem cell function via de novo lipogenesis

Abstract

Sarcopenia and the age-related decline in muscular strength and regenerative capacity contribute directly to loss of autonomy, greater risk for hospitalization and healthcare burden. One contributing cellular phenotype associated with skeletal muscle aging is a loss in the function and number of resident muscle stem cells (MuSCs) or satellite cells. MuSC activation leads to dramatic changes in cellular architecture and metabolic reprogramming including both mitochondrial biogenesis and increased glycolysis. Despite these changes to increase energy production, high energy demands may not be fully met during periods of MuSC activation. In this study, we utilize in vitro and in vivo approaches in mice to demonstrate the function of glutaminase (GLS) for age-related changes in MuSC function. By combining FACS-isolation with metabolomics and stable-isotope tracing, we show an age-related decline in reductive (“counterclock-wise”) flux of glutamine through the TCA-cycle, a pathway by which MuSCs build cellular fatty acid stores as necessary biomass for MuSC cell function.

Introduction

Sarcopenia and the age-related decline in muscle strength and regenerative capacity contribute directly to loss of autonomy, greater risk for hospitalization and healthcare burden^1,2. Improving resilience could allow for improved mobility and independence. One contributing cellular phenotype associated with skeletal muscle aging is a loss in the function and number of resident muscle stem cells (MuSCs) or satellite cells. MuSCs directly participate in the replacement of myonuclei following muscle injury; however, more recent evidence shows a contribution of MuSCs to myonuclear maintenance even in the absence of damage³. The molecular processes governing MuSC quiescence, activation and differentiation to myoblasts have garnered much attention as potential therapeutic targets to enhance muscle resilience. However, the metabolic circuits which support rapid cellular remodeling from a dormant state are not well defined in adult stem cell populations.

Remodeling of the cellular metabolic landscape is a hallmark of stem cell activation⁴. Upon activation by muscle injury or in vitro culture, muscle stem cells show robust mitochondrial biogenesis and increased capacity for mitochondrial oxygen consumption^5,6. The energetic demands required for cellular remodeling from quiescence to a rapidly replicating state are thought to drive an increase in mitochondrial energy production. However, activated muscle stem cells also upregulate glycolysis suggesting that oxidative phosphorylation may not be fully meeting the energy demands of the cell^5,7. Consistent with an overall energy deficit, phosphorylation of AMPK is increased in activated stem cells despite greater glycolytic and mitochondrial energy production^6,8,9. This suggests that changes in metabolism during MuSC activation may not be directly related to ATP production. Ryall et al.⁷ provided evidence to support this concept by showing increased glycolytic flux supports cytoplasmic acetyl-CoA production used in protein acetylation during MuSC differentiation. Glutamine is a versatile metabolic substrate that can be used in pathways related to conversion to other amino acids, synthesis of nucleic acids, and energy production¹⁰. Other stem cell populations including skeletal stem cells have been shown to be reliant on glutamine metabolism for proliferation and bone formation¹¹. Recent data from Soro-Arnaiz et al.¹², demonstrate greater anaplerotic metabolism of glutamine in proliferating myoblasts compared to differentiated myotubes. Moreover, some rapidly dividing or differentiating cells, including cancer cells, utilize a glutamine-dependent reductive carboxylation pathway to form citrate for processes such as de novo lipogenesis¹³. The reliance of glutamine metabolism, and more specifically the utility of the reductive carboxylation and downstream biosynthetic pathways, has not been investigated in the MuSC in the context of stem cell aging.

A holistic understanding of metabolic remodeling in response to stem cell activation, and how such responses become impaired with aging, is needed to better understand how to revive stem cells in aging muscle. In this study, we utilize targeted metabolomics and flux analysis datasets to identify a crucial metabolic pathway for MuSC activation – glutamine metabolism. Using in vitro and in vivo approaches, we demonstrate the function of glutaminase (GLS) for age-related changes to MuSC function. By combining FACS-isolation with stable-isotope tracing, we identify an age-related decline in reductive “counterclock-wise” flux of glutamine through the TCA-cycle to form citrate and contribute to de novo lipogenesis, thereby providing lipids needed for rapid cellular growth and replication.

Results

Glutaminase supports MuSC metabolism for successful activation in vitro.

We tested the necessity of pyruvate, fatty acids and glutamine for MuSC activation by monitoring incorporation of EdU in freshly isolated MuSCs (Supplemental Figure 1) with specific inhibitors for each fuel substrate (Fig 1 A,B, Supplemental Figure 2A–C). While there was reduced EdU incorporation observed with inhibition of pyruvate and fatty acid metabolism at the 24 hour time point, these groups regained proliferative rates similar to control MuSCs at the 48 and 72 hour time points (Fig 1A,B). In contrast, only MuSCs treated with the GLS inhibitor, BPTES, had a marked reduction in proliferation throughout all time points (Fig 1 A,B). To assess whether glutaminase inhibition would affect MuSC survival, we measured apoptosis in vitro via TUNEL. We found a 30% increase in TUNEL staining (p < 0.001) in BPTES-treated MuSCs compared to vehicle (Fig 1 C,D).

Figure 1. MuSCs require glutamine for successful activation in vitro.

A) EdU was added to freshly isolated MuSCs to measure rate of activation as a percent of EdU+ cells to total nuclei over time in vehicle-treated (control), 1μM BPTES-treated (glutaminase), 3μM Etomoxir-treated (Fatty acid oxidation), and 2μM UK5099-treated (Pyruvate) conditions. n=9/group/timepoint biological replicates; data are mean +/− SEM; two-sided t-test was used to compare each treatment to control at each timepoint, without correction for multiple comparisons. B) Representative micrographs of EdU (green) and DAPI (blue) at 48hr timepoint. C) Quantified TUNEL assay of MuSCs 48 hours following isolation -- results shown are TUNEL+ cells as a percent of total nuclei. N = 3 (control) and 4 (BPTES) biological replicates; data are mean +/− SEM; p value indicated two-tailed student t-test. D) Representative micrographs of TUNEL (green) and DAPI (blue) from C. E) Individual myofibers were isolated and cultured for 72 hours before fixation and immunostaining of PAX7 and Ki67 to analyze MuSC proliferation in situ. Each point represents an individual isolated fiber (n = 29 control, 22 BPTES; data are mean +/− SEM; p value indicates two-tailed t-test). F) Representative micrographs of (E). G) Isolated MuSCs were cultured for 48 hours on Seahorse XFe 24 well plates and analyzed by Seahorse Extracellular Flux Analysis. Oxygen Consumption rates were normalized to total cell number (quantified immediately after analysis). n=3 biological replicates/group; data are mean +/− SEM; p value indicates two-tailed t-test at specified timepoint without correction for multiple comparisons. H) MuSCs were activated in vitro for 40 hours before 8-hour treatment with BPTES and subsequent extraction for untargeted metabolomics by LC/MS. Abundances were log2 transformed and z-scaled; n =3/group asterisk represents p < 0.05 Student’s two-tailed t-test.

To better understand the changes in MuSC activation during GLS inhibition, we performed ex vivo experiments using intact myofibers (Fig 1 E,F). Seventy-two hours following myofiber isolation, BPTES-treated myofibers had 45% fewer PAX7+/Ki67+ nuclei compared to vehicle control (p < 0.001, Fig 1 E,F) once again suggesting a reduced proliferative rate with GLS inhibition. Because treatment with BPTES directly prevents glutamine anaplerosis, we tested the oxygen consumption rate (OCR) of control and BPTES treated MuSCs. At all time points, basal OCR was significantly reduced in BPTES treated wells (by 33%, p = 0.038); however, these differences were lost upon injection of the electron transport inhibitor oligomycin and ATP synthase inhibitor rotenone/antimycin A, suggesting the differences in OCR were directly linked to substrate oxidation (Fig 1G). For a more global view of the metabolic landscape of the activating MuSCs (beyond just extracellular measurements of OCR), we performed nontargeted metabolomic phenotyping to characterize the glutaminase-supported metabolic signature in MuSCs. For this experiment, we analyzed two groups of MuSCs that were isolated and cultured in vitro for 24 hours with a 1 hour treatment period by either vehicle (CON) or BPTES before collecting metabolite extracts for UHPC-LC/MS metabolomics. The lack of glutaminase activity (confirmed by glutamine and glutamate abundances) resulted in deficiencies in several metabolic intermediates across multiple metabolic pathways including a significant reduction in ATP level and increase in AMP levels (Fig 1H). These initial results support a critical role for the glutamine pathway in MuSC proliferation, survival and metabolism.

Glutaminase is supportive of successful MuSC activation in vivo

While experiments in Figure 1 show a clear in vitro phenotype of GLS-inhibition and glutamine deprivation in isolated MuSC, we next sought to determine the contribution of glutaminase function to MuSC activation in an in vivo regenerative context. We generated a MuSC-specific, inducible glutaminase1 knock-out mouse line by crossbreeding Pax7^CreERT2 (Gaka strain) mice with a Gls1^fl/fl mouse strain¹¹ to allow tamoxifen-inducible, MuSC-specific depletion of the Gls1 gene (Pax7^Gls−/−Fig 2A, see methods for specific strain details). Pax7^Gls−/− mice were treated with tamoxifen no earlier than 4 months of age to ensure full development. Fourteen days following tamoxifen administration (or corn oil for controls), MuSCs were isolated from uninjured muscle for immunoblot analysis of GLS1, demonstrating a 53% reduction in GLS1 protein abundance (P < 0.001; Fig 2B). Flow cytometry analysis of isolated MuSC shows a decline in the percent of MuSC from uninjured muscles (~0.8% lower VCAM1+ cells, 8% relative reduction, in % of VCAM1+ cells to non-heme cells, p = 0.061; Fig 2C) as well as a blunted response to BaCl₂ injury in the tibialis anterior (TA) muscle (25% reduction in MuSC relative to non-heme cells, p = 0.015; Fig 2D). To substantiate the flow cytometry analysis, we performed immunofluorescence of PAX7 on muscle sections from uninjured and 4 days post-BaCl₂ muscles. Quantification of the PAX7+ nuclei shows no differences in quiescent (subsarcolemmal) satellite cells in the uninjured state although this is only 2 weeks following GLS ablation (p = 0.174, Fig 2E,F). However, once activated by BaCl₂ injection, the increased number of PAX7+ cells at the 4-day timepoint is ablated in the tamoxifen treated Pax7^Gls−/− muscles (p = 0.038, Fig 2G,H). Using the same experimental design, we collected the tibialis anterior (TA) muscles 4 days post-BaCl₂ for histological analysis of embryonic myosin heavy chain (eMHC), a marker of early muscle regeneration (Fig 2I). Tamoxifen treated Pax7^Gls−/− mice showed significantly reduced fluorescence of eMHC and regenerating fibers were 19% smaller (p = 0.039; Fig 2J,K). Next, we used a 14-day recovery time point after BaCl₂ treatment in the TA to determine the degree of regeneration between groups (Fig 2L). Analysis of hematoxylin and eosin staining showed a reduction in TA muscle fiber cross-sectional area by 29.2% in tamoxifen-treated Pax7^Gls−/− mice compared to TA from vehicle-treated controls (p = 0.032; Fig 2M,N). Finally, we reanalyzed the 14-day injury timepoint to determine the proportion of PAX7+ cells that were interstitial or subsarcolemma during the resolution of regeneration (Fig 2O). Similar to findings at the earlier timepoint, we saw tamoxifen treated mice had a reduction in the total number of PAX7+ cells (Fig 2P, p < 0.001) and the proportion of those subsarcolemmal, PAX7+ cells was also reduced (Fig 2Q, p = 0.024). While our experiments do not include tamoxifen treated, Cre-controls, evidence suggests minimal Pax7^CreERT2 driven regeneration defects in the strain used here (see mouse strain details in methods)¹⁴. These results further support the necessity for GLS in the MuSC during muscle injury.

Figure 2. MuSC glutaminase is required for successful activation and regeneration in vivo.

A) Pax7^CreERT2XGls1^flfl mice were bred to produce Pax7^Gls−/− experimental cohorts where mice received either corn oil vehicle (Control) or Tamoxifen for 5 days; graphic generated in BioRender. Two weeks following final tamoxifen dose, mice were used for experiments. B) Immunoblot analysis of GLS1 to validate genetic model. Protein was isolated from FACS-isolated muscle stem cells; n = 9 control, 12 tamoxifen biological replicates; data are mean +/− SEM. Flow cytometry of MuSCs from Pax7^Gls−/− uninjured limb muscles, n=6/group; data are mean +/− SEM (C) or 2day BaCl2 injected TAs, n = 5/group; data are mean +/− SEM (D). Immunofluorescence staining by anti-PAX7 antibody and quantification of PAX7+ nuclei in uninjured, n = 3 biological replicates/group; data are mean +/− SEM (E,F) and 4 days post BaCl2, n=3 biological replicates/group; data are mean +/− SEM (G,H) tibialis anterior muscles from control or tamoxifen treated-Pax7^Gls−/−. I) eMHC immunofluorescence of 4 day BaCl2 injected TA muscles from Pax7^Gls−/−. J) Mean cross-sectional area (CSA; data are mean +/− SEM) and K) fiber size frequency distribution from eMHC 5day histology, n =3 biological replicates/group; center dash line is mean, thin dotted lines are upper and lower quartiles bounds. L) 14-day post BaCl2 injected TA histology by hematoxylin and eosin staining. M) Mean CSA (n=3 Con,5 Tam; data are mean +/− SEM) and N) fiber size distribution (n=3 Con,5 Tam; center dash line is mean, thin dotted lines are upper and lower quartiles bounds). O) PAX7 and laminin immunofluorescence of 14DPI TA with PAX7+ cells/myofiber quantified in (P) (n = 3 con, 5 tam; data are mean +/− SEM) and proportion of subsarcolemmal PAX7 cells in (Q) (n = 3 con, 5 tam; data are mean +/− SEM). Yellow arrows indicate PAX7+ nuclei under basal lamina. p-values indicate 2-tailed, t-test.

GLS is deficient in old MuSCs but sufficient to enhance MuSC transplantation

Many of our findings observed with glutaminase loss of function in MuSCs (altered metabolism, activation, regeneration) are reminiscent of deficiencies seen in the context of aged MuSCs. We first wanted to determine if aged MuSCs had altered expression of GLS. We measured protein expression of GLS in FACS isolated MuSCs from uninjured mouse muscle (Fig 3A), and found a significant, 51% reduction in GLS abundance in MuSC from 24m (old) compared to 6m (young) samples (Fig 3B, Supplemental Figure 3A; p = 0.007). We queried a publicly available scRNA-seq dataset made recently available by Lai et al¹⁵, which showed a similar downregulate of the Gls1 transcript in aged human donors compared to young (Supplemental Figure 3B). We then isolated MuSCs from the two age groups and analyzed proliferation by incorporation of EdU during the initial phase of MuSC in vitro activation in the presence and absence of the GLS inhibitor BPTES. Aged MuSCs had an inherent reduction in proliferation (14%, p < 0.001) 48 hours after isolation compared to young MuSCs. Proliferation was decreased by an additional 40% when old MuSCs were treated with BPTES (p < 0.001; Fig 3C). A similar pattern of results was seen when we analyzed young and old MuSCs oxygen consumption rate (OCR). Old MuSCs had 48% lower basal OCR compared to young MuSCs (p < 0.001) which was further lowered in old MuSCs treated with BPTES (68% lower in Old BPTES vs Old, p = 0.005; Fig 3D). Based on our observation of decreased EdU incorporation and OCR in BPTES-treated Old MuSCs compared to Old controls, we hypothesized that residual, but reduced, GLS activity continues to contribute to the activation of Old MuSCs, and that increasing GLS expression might enhance Old MuSC function. To test this hypothesis, we performed a transplantation experiment using lentiviral driven expression of GLS1 in MuSCs from 24-month-old mice (Fig 3E). We validated reporter expression, glutaminase protein abundance and glutaminase conversion of glutamine to glutamate in MuSC transduced with lenti-gls (Fig 3F, Supplemental figure 4A–E). We transplanted these MuSCs (from 24m old donor mice) into the TA muscle of 24m old recipients 1 day after BaCl₂ injection, using one limb for lenti-gfp cells and the contralateral limb for lenti-gls cells. Thirteen days following transplantation, we excised TA muscle and performed immunofluorescence using anti-GFP and anti-PAX7 antibodies to assess contribution of donor cells towards regenerating myofibers and the remaining PAX7+ stem cell pool (Fig 3G). We found a significantly greater proportion of regenerated fibers as GFP+ in the lenti-gls-treated TA muscle (73% greater, p = 0.002; Fig 3H), and those GFP+ fibers in the lenti-gls limb were larger than the GFP+ fibers in the lenti-gfp transplanted limb (mean difference of 422μm², Fig 3I,J). When we counted the proportion of PAX7+/GFP+ nuclei, we did not see a difference in the number of GFP-labelled donor cells replenishing the PAX7+ quiescent pool. These results indicate that in the pathological context of aging, glutaminase is deficient in the aged MuSC yet still supportive for activation and metabolism. Moreover, the restoration of GLS1 expression is a potential therapeutic target to enhance myogenic function in aged MuSCs.

Figure 3. GLS is deficient in old MuSCs and is sufficient to enhance MuSC transplantation.

A,B) Freshly isolated MuSCs were used for immunoblot analysis of GLS1 from C57/B6 young (6m) or Old (24m) limb muscles, n =5 biological replicates/group; data are mean +/− SEM; p represents 2-tailed t-test C) EdU incorporation assay of MuSC activation (n=3/group/timepoint; data are mean +/− SEM) and D) Seahorse OCR analysis (n=3 biological replicates/group; data are mean +/− SEM) from young (6m) or old (24m) under control media or with 1uM BPTES to inhibit GLS activity. P indicates 2-tailed t test between indicated groups. E) Experimental design of transplantation studies. Donor MuSCs were isolated from 24m old mice, transfected ex vivo with Lentivirus-GFP control or Lentivirus-GFP-Gls1 for 24 hours before being injected into recipient (24m old, 1day post-BaCl2 injection) TA muscles; graphic generated in BioRender. Recipient mice received Lenti-Gls MuSCs in one TA and contralateral limb received lenti-control MuSCs, n = 6 recipient mice. F) Micrographs of MuSC lentiviral GFP reporter expression in vitro; representative micrograph from 5 repeated experiments. G) Immunofluorescence staining with anti-PAX7 and anti-GFP antibodies in transplanted muscle sections from 14 days post BaCl2 injection (13 days post-transplant) with DAPI and phalloidin staining of nuclei and actin, respectively; n = 3mice. H) Percent engraftment (GFP fibers/total fibers)(n = 6/group; data are mean +/− SEM). I) Average CSA of GFP+ fibers (n = 6/group; data are mean +/− SEM), J) Percent PAX7+/GFP+ nuclei (n=4/group; data are mean +/− SEM) and K) CSA fiber distribution of GFP+ fibers (n=3/group, >30 fibers/sample; center dash line is mean, thin dotted lines are upper and lower quartiles bounds). P-values indicate 1-tailed t-tests. Connected datapoints in H-J are contralateral-matched samples from the same mouse.

Stable Isotope Flux Analysis of glutamine during aged MuSC activation.

In order to directly assess the utilization of glutamine in the TCA cycle, we activated FACS-isolated MuSC for 24 hours followed by incubation with 200μM [¹³C₅] glutamine. Following 2 hours of incubation, cell extracts were prepared in ice cold methanol and subjected to GC/MS analysis of ¹³C-labeled mass isotopomers of glutamine-derived metabolites (e.g., M0,M1, …M_n ¹³C-labelled carbons) Fig 4A). We verified that the cellular glutamine pool was comprised entirely of M5-glutamine in both Old and Young MuSCs (Fig 4B). Glutamate and α-ketoglutarate were effectively labeled as well, with predominance of the M5 isotopomer, but labeling of both of these metabolites was reduced in MuSC from 24 month old mice compared to 6 month old young mice (Fig 4C and 4D; p = 0.009 and p = 0.018, respectively), consistent with the lower levels of GLS activity in aged MuSCs. Citrate can be derived from glutamine by M4-Oxaloacetate, and/or M2-acetyl CoA (generated through pyruvate anaplerosis) to generate M2, M4 or M6 isotopomers of citrate (See Fig 4A, orange box). All of these labeled species of citrate were present at higher levels in young compared to old MuSC (Fig 4E). Interestingly, we also saw clear labelling of M5-citrate, which is generated by the reductive TCA cycle metabolism involving isocitrate dehydrogenase 2 (IDH2) (Fig 4A purple box)¹⁶. We found that M5-citrate was reduced by ~80% in old compared to young MuSC (p < 0.001, Fig 4E). Taken together, these data demonstrate reduced glutamine metabolism through the TCA cycle in aged MuSCs with a clear decrease in oxidative metabolism (reflected in decreases in M2, M4, and M6 citrate) which is accompanied by a proportionate decrease in reductive flux of glutamine (reflected in decreased M5 citrate). Similar levels of reductive carboxylation of glutamine to citrate by proliferating MuSCs was recently published by Soro-Arnaiz¹²; however, the downstream consequences of this process and the impact of age-related MuSC function have yet to be explored so we chose to investigate reductive MuSC glutamine metabolism.

Figure 4. Aged MuSCs have proportionately lower reductive glutamine flux during activation.

A) Schematic of labeling strategy with U-¹³C5-glutamine through the TCA cycle with possible oxidative labeling of citrate in orange box and possible reductive labeling of citrate in purple box; graphic generated in BioRender. Resulting labeling percentage of indicated isotopomers for (B) Glutamine M+5, (C) glutamate M+5, (D) alphaketoglutarate M+5 and (E) oxidative and reductive citrate isotopomers; n=3 biological replicates/group; data are mean +/− SEM. F) Seahorse Oxygen Consumption Rate (OCR) analysis; n=3 biological replicates/group, and (G) EdU incorporation of in vitro activating MuSCs with treatment of AGI6780 to inhibit IDH2 reductive TCA flux; n=3 biological replicates/group/timepoint; data are mean +/− SEM; P-values indicate 2-tailed t-tests.

We next wanted to determine if restoring Gls1 expression in aged MuSCs would enhance energy production, and if concurrent inhibition of the reductive glutamine pathway would negate such benefits. To test this, we measured OCR from 6-month (young), and 24-month (old) MuSCs at 48 hours following FACS isolation. During the first 24 hours, we transduced a group of old MuSCs with a lentiviral-Gls1 vector to overexpress glutaminase (Old Gls) and subsequently treated the cells with 10μM AGI6780, a selective IDH2 inhibitor¹⁶. Basal OCR was 16% lower in Old MuSC compared to Young (Fig 4F; p = 0.019) but was restored to the level of young MuSC upon overexpression of Gls1 (p = 0.009 Old vs Old Gls1). Interestingly, Old MuSCs treated with the IDH2 inhibitor AGI6780 exhibited a further 21% induction of basal OCR (p = 0.045) and maintained high levels of oxygen consumption following oligomycin treatment (Fig 4F). To understand the connection between energy production (OCR) and MuSC activation, we measured in vitro EdU incorporation in the same treatment groups throughout 72 hours in culture. The first 24 hours were used to transduce either lentiviral-Gfp (control) or lenti-Gls before addition of EdU and fixation during the subsequent 48 hours (Fig 4G). We again observed a greater rate of EdU incorporation in young MuSC compared to Old (p < 0.001 at each time point; Fig 4G). Matching the OCR data, the old Gls-lenti MuSCs activated as rapidly as the young, and had 36–43% more EdU+ cells than Old MuSCs treated with the control Gfp-lenti (p < 0.001 at each time point; Fig 4G). Although the Gls-lenti treatment improved EdU incorporation throughout the 48 hour labeling period, concurrent treatment with AGI6780 prevented the rescuing effects, reduced proliferation by 32% (p < 0.001, Fig 4G), returning proliferation levels back to Old MuSCs. Importantly, the effects of the AGI6780 were not due to drug toxicity at the selected doses (Supp. Fig. 5) and the drugs inhibitory effects on metabolism were specific to reductive IDH2 inhibition (Supp. Fig 6A). Together, these results show that forced expression of GLS1 in aged MuSCs can restore proliferation and metabolism, which is dependent on the reductive glutamine pathway.

IDH2 is required by MuSCs to support cellular fatty acid levels during activation

Although our data highlights the importance of the reductive glutamine pathway in the MuSC, the metabolic pathways and cellular function supported by reductive flux in MuSCs remain undefined. We hypothesized that reductive glutamine flux (via IDH2) may serve to generate biomolecules needed for cell transformation and division during the initial phase of activation from quiescence. To test this hypothesis, we isolated MuSCs from 6-month-old mice, cultured them in the presence or absence of the selective IDH2 inhibitor AGI6780 for 24 hours, and performed targeted metabolomics analyses on fatty acids (Fig 5A), amino acids (Fig 5B) and nucleotides (Fig 5C). AGI6780-treated cells had significantly lower levels of several common fatty acids (C6, C8, C10, C12, and C16 fatty acids by ~1.5 standard deviations, p = 0.05; Fig 5D). Nucleotide abundances shifted from an energized state (more nucleotide triphosphates) to a less energetic state (increased diphosphates and monophosphates) when treated with AGI6780, although total nucleotide levels remained unchanged p = 0.45, Fig 5C,D). These shifts in the metabolite abundances were accompanied by a doubling of basal OCR in AGI6780-treated MuSCs compared to controls (p < 0.001; Fig 5E) and a resistance to oligomycin-induced decrease in oxygen consumption, similar to what was observed in Fig 4I. We hypothesized that IDH2 inhibition was forcing glutamine oxidation in the TCA cycle which accounted for greater basal oxygen consumption rates and reductions in fatty acid levels by reducing citrate-supported de novo lipogenesis (Fig 5F). To directly monitor the extent to which reductive glutamine flux supports de novo lipogenesis in MuSCs, we used deuterated water (D₂O) to label palmitate synthesis (Fig 5G,H)⁸. In MuSCs isolated from 6-month-old mice we observed a 15% decrease in de novo lipogenesis in AGI6780-treated MuSCs versus vehicle-treated matched controls (p = 0.013, Fig 5H). To test this in vivo, we performed a MuSC transplantation study using young whole body GFP mouse (6 months old) where we isolated MuSCs (Supplemental Figure 7) before treatment with an inhibitor of fatty acid synthetase (GSK2194069; termed FASNi) to determine the necessity for this pathway to support MuSC activation (Fig 5I, Supp. Fig 6B). We used anti-GFP immunofluorescence of muscle sections taken 14days following injury and observed decreases in fiber size and engraftment in FASNi-treated MuSC donors (Fig 5J–L). These data suggest glutamine reductive metabolism supports MuSC de novo lipogenesis and energy metabolism. Furthermore, inhibition of the downstream lipogenic pathway in young MuSCs reduces myogenic function.

Figure 5. IDH2 supports reductive glutamine metabolism to support cellular fatty acid pool.

Metabolomics analysis of 6 month old MuSCs activated for 24 hours in vitro treated with IDH2 inhibitor AGI6780, n=3 biological replicates/group; data are mean +/− SEM: A) Fatty acids, B) Amino acids, C) nucleotides and D) combined total levels of each metabolite category normalized to control. Asterisk indicates p < 0.05 Student’s one tailed T-test. E) Seahorse OCR analysis of 6 month MuSC treated with AGI6780 during activation in vitro; n=3 biological replicates/group; data are mean +/− SEM. F) Conceptual illustration of reductive glutamine flux influencing both oxygen consumption rates and fatty acid abundances; graphic generated in BioRender. G) Schematic diagram of deuterium labeling of palmitate to directly assess de novo lipogenesis; graphic generated in BioRender. H) percent de novo lipogenesis in isolated MuSC from 6-month-old mice treated with vehicle or AGI6780; n=3/group; data are mean +/− SEM. I) Experimental design, J) anti-GFP immunofluorescence micrographs, and (K,L) quantification of MuSC transplant experiment in 6 month young mice using MuSCs treated with the fatty acid synthase inhibitor GSK2194069 (FASNi); n=9 recipient mice; data are mean +/− SEM. Connected datapoints from H,K, L are contralateral-matched samples from the same mouse. P-values indicate two-tailed t-tests

Reductive glutamine flux supports MuSC activation via de novo lipogenesis

With the findings that reductive glutamine flux through IDH2 supports de novo lipogenesis, we hypothesized that reduced function of aged MuSCs could be rejuvenated by replacing the fatty acid end products of reductive glutamine flux. Measuring in vitro proliferation via Edu incorporation, we overexpressed Gls1 in Old MuSCs in combination with AGI6780 to inhibit IDH2. We also included treatment with the dietary lipids palmitate and oleate as a cell permeable end product of DNL. We used a combination of palmitate with oleate because it offers a better approximation of physiological lipid conditions as published by our group^17,18. Similar to our previous experiments (Fig 4J), Old MuSCs with lentiviral driven GLS expression had reduced activation when treated with AGI6780 (by 18-31%, p < 0.001) which was restored to the level of the controls when media was supplemented with 200 μM palmitate/oleate (p < 0.001, Fig 6A). To test the rejuvenating effect of exogenous fatty acids in the context of myogenesis, we performed a transplantation experiment in which we exposed Old (24 month) MuSCs to 200 μM palmitate/oleate following isolation to support activation before transplantation into injured, 24-month-old recipient (Fig 6B). The palmitate/oleate-treated MuSCs formed myofibers that were on average 417um² larger (~45%) than contralateral limb BSA (vehicle)-treated fibers (p = 0.019; Fig 6C,D) and had 220% greater engraftment rates (p = 0.011; Fig 6E). Furthermore, we tested to see if these results extended to a physiological, functional recovery of the hindlimbs using this metabolite replacement approach. We performed three assessments of mouse physical function at baseline and 14 days following the same injury/transplantation shown in Fig 6B with outcomes similar to Kang et al.¹⁹ including RotaRod performance test, open field activity monitoring and treadmill gate analysis of hindlimbs. 24-month-old mice receiving palmitate/oleate-treated MuSCs had physical function closer to the uninjured state when assessed 14 days post injury by RotaRod (Fig 6F, p = 0.024), open field activity monitoring (Fig 6G, p = 0.044) and hindlimb gait analysis (Fig 6H, p = 0.017 for brake phase, 0.011 for propel phase). These results supported our hypothesis that aged MuSCs have a reduced capacity for reductive glutamine metabolism and downstream de novo lipogenesis. By restoring the fatty acyl end product, aged MuSC returns proliferative, myogenic function and hindlimb functional recovery.

Figure 6. Reductive glutamine flux supports MuSC activation via de novo lipogenesis.

A) EdU Incorporation time course of MuSCs from 6 month and 24 month wildtype mice transduced with GFP or GLS lentivirus and combining AGI678-inhibitor treatment with exogenous palmitate/oleate (P:O); n=3 biological replicates/group; data are mean +/− SEM. B) Experimental design; graphic generated in BioRender, C) micrographs, and (D,E) quantification of MuSC transplant experiment in 24 month old mice using MuSCs treated with palmitate/oleate (P:O); data are mean +/− SEM; p value indicates one tailed t-test. Connected datapoints in D,E are matching samples from the same mouse; n=4 recipient mice. Functional recovery of mouse hindlimbs from experiments in panel B was measured by F) RotaRod, G) Open field activity, and H) Hindlimb Gait Analysis. F-H, p-values indicate one-tailed t-tests, numbers are shown as percent of the uninjured measurement (indicated by dashed line at 100); n=6/group; data are mean +/− SEM.

Discussion.

Muscle stem cell function is recognized as a critical process for successful maintenance and regeneration of skeletal muscle⁵. Understanding metabolic adaptations that occur in this cellular remodeling and proliferation process could present targets for therapeutic intervention to sustain MuSC regenerative function with age. In this study, we identify glutamine metabolism by glutaminase, to be necessary for MuSC function. We extend these observations by showing that glutaminase is required for both oxidative energy production and reductive “counter-clockwise” TCA metabolism. Aged MuSCs have reduced capacity to metabolize glutamine due to the reduction in glutaminase abundance. When glutaminase function is restored in the aged MuSC, proliferative function returns, which is dependent on reductive flux through the TCA cycle. Lastly, we used targeted metabolomics and metabolic flux analyses to demonstrate that reductive glutamine metabolism in the TCA cycle supports de novo lipogenesis as a key pathway for MuSC activation and cellular proliferation. Collectively, these data identify a supportive role of glutamine metabolism in the MuSC and potential therapeutic targets to restore stem cell function with age including a reductive pathway required for MuSC energy production, biosynthetic pathways and function.

Traditional glutamine metabolic circuits lead to production of glutamate and α-ketoglutarate for oxidation of the latter in the TCA cycle, as well as other metabolic pathways including serine biosynthesis^20,21. Reductive metabolism of glutamine in the TCA cycle has been shown to be active in cancer models²² as well as metabolic control of insulin secretion from pancreatic islet β-cells, demonstrating that this metabolic pathway is not unique to cancer cells and can contribute to non-pathological cellular functions¹⁶. Supporting this idea, we were able to demonstrate that glutaminase activity is required for normal MuSC activation in vitro, in situ and in vivo. We also found that glutamine is metabolized in MuSC by both oxidative and reductive pathways, each contributing to stem cell activation and proliferation. Our data agree with those recently published by Soro-Arnaiz et al.¹² which also measured reductive TCA carboxylation of glutamine in proliferating myoblasts which drops of significantly in differentiated myotubes. We sought to further elucidate the downstream metabolic outcomes of this cell-state specific metabolic activity especially in the context of aging.

Satellite cells exhibit unique cycles of quiescence and activation in order to contribute to myonuclear accretion without exhausting the muscle stem cell pool. In the quiescent (or G0) state, MuSCs have relatively small cytoplasm and limited stores of amino acids, lipids and other metabolic substrates. Once a satellite cell is stimulated to enter the cell cycle, a conserved stress response begins to coordinate transcriptional changes to drive protein synthesis, nucleotide synthesis and ATP production in preparation for the initial cell replication event (typically within 40 hours of activation)^23,24. Additionally, the satellite cell niche maintains MuSCs in a state of relative hypoxia to limit oxidation of metabolic fuels^25–27. These unique characteristics of the muscle stem cell underscore the need for metabolic flexibility to support successful activation. Examples of this include the uptake of macrophage-derived glutamine²⁸ and the generation of ATP via autophagy stimulation⁶. Our results define a non-canonical metabolic pathway supporting stem cell activation, the IDH2-driven reductive metabolism of glutamine in the TCA cycle. As originally described in oncogenic hepatocytes, the reverse shunting of glutamine through the TCA cycle yields NADPH and citrate -- both of which support de novo lipogenesis (DNL)²². Synthesis of fatty acids by DNL in rapidly proliferating or differentiating cells provides precursors for synthesis of triglyceride, phospholipids and other complex lipids necessary for membrane biogenesis. Our data clearly demonstrate the activation of these pathways in activating muscle stem cells using metabolic flux analysis to measure glutamine flux and DNL.

Natural aging can present with a plethora of muscle-related ailments including decreased ambulation, muscle mass, and regenerative capacity, as well as muscle stem cell exhaustion². Maintaining the pool of muscle stem cells and their ability to successfully activate may help prevent sarcopenia, while enhancing physical capacity and socio/economic productivity in aging populations²⁹. We identified glutaminase as an enzyme contributing to the loss of oxidative capacity during MuSC activation in aging and potentially implicating glutaminase activity in the maintenance of the PAX7 satellite cell pool via self-renewal (Fig 2O–Q). Furthermore, we showed that restoration of GLS in MuSC can rescue oxidative capacity and MuSC activation, while also enhancing regeneration as shown by our transplantation studies in old, sarcopenic mice. Data from other groups does support the age-related downregulation of Gls1 at the transcriptional level¹⁵ despite little change to the chromatin accessibility surrounding the Gls1 coding sequence³⁰. Evidence from caloric restrictions³¹, AMPK^9,32, rapamycin³³, FoxO³⁴, and SIRT1⁷ transgenic studies show that improving metabolic flexibility in age MuSC can improve regenerative function. These studies highlighted changes in mitochondrial dynamics during MuSC activation including markers of mitochondrial biogenesis and increased oxygen consumption rates. Our results build upon and extend this foundational knowledge by demonstrating that MuSC exhibit age-dependent differences in reductive glutamine flux. Moreover, when we blocked reductive TCA flux using AGI6780, we saw an increase in oxygen consumption rates (glutamine flux is forced towards oxidation) concomitant with signs of mitochondrial uncoupling (oligomycin resistant oxygen consumption -- Fig 4I, 5E -- and a shift away from an “energized” nucleotide profile in Fig 5C). There is precedent for transient UCP2 expression in myoblasts during proliferation and differentiation^35–37. Taken together, this suggests that MuSC activation requires controlled oxidative TCA cycle flux for cellular remodeling and mitigating excessive redox stress. Our experiments did not test the necessity of glutamine metabolism for rapamycin or calorie restriction-mediated rejuvenation effects on MuSCs; however, we do provide evidence that directly replacing the downstream metabolic products of glutamine-fueled reductive flux and DNL (palmitate/oleate in this case) is sufficient to produce similar functional benefits (in vitro activation, transplantation) as the glutaminase enzyme itself (Fig 5). Future studies should identify terminal metabolic products which might be replenished to offer metabolic support during MuSC activation, analyze complex lipids to understand changes in the lipidomic landscape during activation and supported by glutamine reductive TCA flux, and further understanding the small molecule metabolite dynamics in and around the MuSC niche especially during regeneration.

In summary, we identify glutaminase-driven reductive TCA flux in the MuSC as a key metabolic pathway supporting DNL during MuSC activation. Leveraging static and flux approaches to stem cell metabolomics, we identified several unique metabolic characteristics of the activating MuSC at the level of specific metabolic pathways and small molecule metabolite levels rather than the more global transcriptomic or protein level. Age-related detriments to these pathways may account for lost function and exhaustion of the MuSC pool. Moreover, replenishing GLS, or downstream metabolites such as palmitate may serve as new strategies for achieving functional rejuvenation of MuSC function.

Methods.

Animals

All animal experiments and care followed the guidelines and were approved by the Institutional Animal Care and Use Committee at Duke University Medical Center (Protocol #A027-24-01). For experiments with wildtype mice, male C57/BL6J mice were acquired through the National Institute of Aging animal colony and used for experiments at 6-8 months of age (labeled as ‘Young’) or at age 24-26 months (‘Old’). The Pax7^CreERT2 mouse was acquired through Jax laboratories (Gaka strain, Jax ID: 017763). Gls1^fl/fl mice were a kind gift by Dr. Courtney Karner and are available through Jax Laboratories (JAX id:017894). The GFP mice used were acquired from Jax Laboratories (β-actin-GFP, strain id: 006567). All mice were housed in a climate-controlled facility with temperature maintained at 22°C +/− 1°C with humidity maintained between 30-70%, and light/dark cycle of 12h/12h with water and chow (PMI 5053, Picolab rodent diet 20) available ad libidum. Once experimental endpoints were reached, mice were euthanized by CO2 asphyxiation for >5 minutes followed by decapitation.

Muscle injury

Barium chloride (BaCl2) was used to induce muscle damage in tibialis anterior muscle via intramuscular injection of 30uL of 1.2%(vol/vol) BaCl2 in PBS using a 30 guage needle. Control muscles injected with an equal volume of PBS.

Primary mouse MuSC isolation by FACS

Mouse muscles were processed into a single cell suspension for flow cytometry or fluorescence activated cell sorting (FACS) using methods adapted from Liu et al³⁸. Triceps, quadriceps, hamstrings, tibialis anterior, gastrocnemius, soleus and plantaris muscles were collected from uninjured mice or tibialis anterior muscles were used in mouse injury experiments. Muscles were dissected, minced using scissors, digested in 10mL of F10 + 10%(v/v) horse serum and 1000u/mL Collagenase, Type 2 (Worthington) for 60 minutes at 37C in a shaking water bath, washed in F10 media, digested again for 30 minutes at 37C in F10+10%(v/v) horse serum + 1.1U/mL dispase (Gibco) and 100U/mL type 2 collagenase, washed, passed through 40uM cell strainer and stained with antibodies for flow cytometry. Cells were identified on flow cytometers by forward scatter and side scatter parameters. For MuSC isolation, cells were negatively gated for APC-CD45, APC-CD11b, APC-Ter119, and APC-CD31, and FITC-SCA1 populations before selecting VCAM+ cells. Flow cytometry and FACS were performed on Sony SH800 cell sorter equipped with 4 lasers and appropriate detector filter sets. Cell sorting was performed in biosafety cabinet at 4C using 100uM microfluidic chip at a pressure near 20PSI. Cells were collected in media used for culturing but washed once prior to seeding to remove FACS sorting sheath fluid. Post sorting analysis was routinely used to confirm successful sorter calibration. For isolation of VCAM1+ MuSCs from GFP mouse muscle, identical methods were used with the exception of the FITC-SCA1 antibody which was replaced with APC-SCA1 (see supplementary figure 7).

MuSC culture and in vitro analysis

Isolated MuSC were collected by FACS and cultured on tissue culture treated plates (384-well, 48-well, 12-well and Seahorse XFe24) which were pre-treated with 1ug/mL Collagen,5ug/mL Laminin in PBS for at least 2 hours. Culture medium for proliferation, TUNEL and Seahorse assays was F10 supplemented with 10% horse serum, 10% fetal bovine serum, 1% penicillin/streptomycin and 10ng/mL bFGF (Gibco #13256-029). For isotope labeling studies, culture media was changed to DMEM without glucose, pyruvate, glutamine or phenol red (Gibco #A1443001) and supplemented 10% horse serum, 10% fetal bovine serum, 1% penicillin/streptomycin and 10ng/mL bFGF for a 10 minute washout before replacing with the same DMEM mixture with 1M glucose, and 200μM glutamine (either ¹²C5 for negative control or ¹³C5 for isotope labeling [Sigma 605166]). For lenti-viral isotope labeling study (supplemental figure 4), the labeling media was modified to contain no glucose for maximal stimulation of TCA metabolism of glutamine. For EdU labeling, 10uM EdU was added to culture media at the time of seeding isolated MuSCs and fixed using 10% formalin for 10 minutes at the times indicated. EdU was visualized using Click Chemistry Tools (Scottsdale, AZ, USA) Click-&-Go Imaging Kit (Cat#1314). TUNEL analysis was performed on MuSCs fixed 24 hours following isolation using Abcam TUNEL assay kit (ab66108). Seahorse Extracellular Flux analysis was performed on Seahorse XFe24 platform according to manufacturer protocols and using drug concentrations previously published and optimized^39,40. Rate data were normalized to total cell number or protein mass which was quantified immediately following the analysis. For studies where MuSCs were supplemented with exogenous fatty acids, we supplemented media with 200μM palmitate/oleate-conjugated to BSA at a 1:1 ratio as previously optimized¹⁷. Our previous studies support the inclusion of both oleate alongside palmitate to provide a physiological blend of fatty acyls to the cells^17,18.

Single myofiber isolation and culture

Following euthanasia, the extensor digitorum longus (EDL) muscle was carefully dissected to avoid shearing or tearing and placed in 0.5% collagenase I (Worthington Biochemical) in DMEM at 37C for 90 minutes. Following three washes in DMEM, individual fibers were triturated with flame polished glass pipettes for ~ 3-5minutes to release myofibers. After 60 minutes equilibration period, intact, individual fibers were selected and transferred to ECM-coated 24-well culture dish and cultured until experimental endpoints in DMEM containing 1% penicillin/streptomycin, 10% fetal bovine serum, 1% chicken embryo extract. After 72hrs, myofibers were gently rinsed in 2 changes of 10% formalin for 10 minutes to fix samples before immunofluorescence staining.

LC/MS-based metabolomics of Isolated MuSCs –

For metabolomics analyses summarized in Figure 1, MuSCs were isolated by FACS and seeded at ~100,000 cells/well of a 12-well dish. Cells were cultured/activated for 24 hours, rinsed with deionized water, covered in ice cold 80% Methanol/20% water (v/v) for 10 minutes on dry ice to extract metabolites, centrifuged at 18,000g for 10 minutes to remove debris, and dried under vacuum centrifugation. Extracts were reconstituted in liquid chromatography solvent for hydrophilic interaction chromatography (HILIC) and analyzed with Agilent Infinity ii 1290 UHPLC coupled to Agilent 6546 Q-TOF mass spectrometer. Chromatography gradient parameters were set to flow rate of 0.350mL/min progressing from 10% to 60% mobile phase B. Mass scanning was performed from 50 to 1200 m/z. Metabolite peaks were matched and quantified to in house standard library using retention time and m/z matching using Agilent Profinder software V10.0. Compound isotope peaks were extracted, integrated and corrected to background abundances. Values were normalized as a percent abundance of the total abundance of all isotopes of that metabolite.

Muscle histology –

Muscle architecture was analyzed by freezing tibialis anterior muscle embedded OCT by dipping in liquid nitrogen cooled isopentane. 10uM cross sections were cut using a cryotome and stained using hematoxylin and eosin, or immunofluorescence staining of eMHC. Images were taken using Olympus microscope and analyzed using ImageJ software to determine area of individual fibers and number of GFP fibers. Cross sectional area was calculated by determining area of >100 fibers across three images for each sample and using this result to average with each sample from that experimental group. For immunofluorescence imaging (myofibers, eMHC, GFP transplants), slides were incubated in 10% formalin for 20 minutes, 0.1M glycine for 10 minutes and blocked/permeabilized in 5% goat serum, 2%BSA, 0.1%triton x-100 in PBS for 1 hour at room temperature before antibody staining. All images were taken using identical camera and light settings. All image quantification was done blinded to experimental group. To analyze interstitial vs subsarcolemmal muscle stem cells in PAX7/Laminin co-stained sections, we used reference (wildtype uninjured) images to exemplify quiescent (subsarcolemmal) nuclei which were typically nearer the spaces between fibers and showed clear overlap with laminin while interstitial nuclei were identified by less overlap with laminin or overlapping with fibers regions.

MuSC transplantation –

For MuSC transplantation, donor mouse muscle was dissected, and MuSCs isolated as described above. When MuSCs were cultured, 1x10^7 lentiviral ORF mGFP or lenti-ORF Gls1 particles were added to 20,000 cells in culture. After 24 hours, GFP fluorescence was confirmed, cells were trypsinized, rinsed in PBS with 5% FBS and injected into recipient mouse tibialis anterior muscles. Recipient mice were syngeneic young (6-8 months) or old (24 months) mice were injected with BaCl2 to induce muscle damage 24 hours prior to transplantation. Recipient mice were euthanized, and tibialis anterior muscles snap frozen for histology of GFP+ fibers 14 days following BaCl2-induced injury (that is, 13 days following transplantation). Histology images were analyzed for GFP+ fiber number and GFP fiber cross-sectional area. Where indicated, fatty acid synthase inhibitor and palmitic acid solution were included in the culture media of donor MuSCs during the 24 hour lenti-viral labelling period. Fatty acid synthase inhibitor (Sigma GSK2194069 referred to as FASNi) was used at a concentration of 20nM with DMSO vehicle as control. 100uM palmitate/100uM oleate was conjugated to fatty acid depleted BSA at 7% BSA and supplemented to culture media with unconjugated BSA as control.

Stable Isotope tracer flux analysis –

MuSCs were isolated by FACS from young and old mouse muscles. 200,000 cells were cultured in 12 well dishes for 16 hours before media was rinsed and replaced by media lacking unlabeled glutamine and with 200μM L-glutamine-¹³C5 (U-¹³C glutamine; Sigma 605166) substituted (see MuSC culture section for details). Following 2 hours of labeling, cells were washed with ice cold PBS and scraped in 1mL of ice-cold methanol. Debris was pelleted and cell extracts were dried by vacuum centrifugation. Dried residues were derivatized by resuspending in 25μL methoxylamine hydrochloride (2% w/v in pyridine) and incubated for 90minutes at 40°C followed by addition of 25μL of MTBSTFA+1% TBDMS and incubation at 60°C for 30minutes. Samples were transferred to GC-MS autosampler vials and analyzed using an Agilent 7890B GC system equipped with HP-5MS capillary column (30m, 0.25mm i.d., 0.25μM-phase thickness; Agilent J7W scientific, Santa Clara, CA) operating at 80°C for 2 minutes and increasing to 280°C at 7°C/min for a total run time of 40 minutes. This was connected to an Agilent 5977A mass spectrometer operating under ionization by electron impact (EI) at 70 eV. Helium flow was maintained at 1mL/min. Source temperature was maintained at 230°C, the MS quad temperature at 150°C, interface temperature at 280°C and inlet at 250°C. Mass spectra were recorded in mass scan mode with m/z range 50-700. Following identification of individual metabolites, the distribution of isotopologue labeling was determined and quantified in MassProfiler software. For each experiment, three biological replicates per group (cells isolated from separate animals) were pooled and analyzed across three separate isolation experiments. All samples from separate isolations were analyzed by GC/MS in the same batch.

For De Novo Lipogenesis studies, ~5e6 freshly isolated MuSCs were cultured in 10cm dishes and ~12 hours following isolation, media was replaced with labeling medium consisting of culture medium supplemented with 6% ²H₂O for 24 hours to allow labeling of newly synthesized fatty acids. Methanol extracts of cells were harvested, dried, hydrolyzed at 85°C for 30 minutes and acidified with 6M HCl before chloroform extraction of fatty acids for derivatization. GC-MS analysis was performed on the same instrumentation setup as detailed above. M0-palmitate (m/z 313) was used as unlabeled reference standard with other isotopomers quantified and normalized to total palmitate and cell number to quantify de novo lipogenesis^16,41,42.

For validation studies of inhibitors (Supplemental figure 2, 6), identical labeling protocols were used for glutamine flux analysis applied to proliferating C2C12 myoblasts. Similar proportions of TCA intermediates were labeled by ¹³C5-glutamine in C2C12 myoblasts as primary mouse MuSC (see figure 4B–E and supplementary figure 2A) supporting their use for metabolic inhibitor validation studies. C2C12 cells were acquired through ATCC (line CRL-1772).

Targeted metabolomics analysis of MuSCs –

For experiments performed in figure 5, MuSCs were FACS isolated, cultured for 24 hours with DMSO or 10uM AGI6780 (Sigma SML0895) before extracts were collected as described for nontargeted metabolomics. For analysis of fatty acids and amino acids, samples were derivatized and analyzed by the same GC/MS approach as detailed for isotope tracer studies with the inclusion of deuterated myristic acid as internal standard. Using the standard, spectra were annotated to metabolites using an in house, retention time-locked, standard library and quantified by peak area^43–45. For nucleotide analysis, samples were analyzed using Waters Xevo TQ-XS quadrupole mass spectrometry coupled to Acquity UPLC system (Milford, MA) and Chromolity FastGradient RP-18e 50-2mm column (EMD Millipore, Billerica, MA, USA). Nucleotides were detected in negative ion multiple reaction monitoring mode based on characteristic fragmentation reactions⁴⁶.

Antibodies –

Antibodies used included: APC anti-mouse CD45 (1:200, Biolegend 157605), APC anti-mouse CD11b (1:200, Biolegend, 101211), PE anti-mouse CD106 (1:50, VCAM, Biolegend 105713), FITC anti-mouse Ly6A/E (SCA1, 1:50, Biolegend, 108105), APC anti-mouse Ly6A/E (SCA1, 1:50, Biolegend, 163806), Pacific Blue AnnexinV (Biolegend 640926), rabbit anti-Glutaminase1 (1:500, Cell Signaling #88964). For immunofluorescence, mouse-PAX7 (supernatant used at 1:50, DSHB AB_528428), Rabbit-Ki67 (1:1000, Abcam #ab15580), rabbit-laminin (1:500, Sigma #L9393) and mouse-eMHC (1:500, DSHB F1.652) were used with DAPI and Phalloidin-TRITC (Sigma P1951) used to visualize nuclei and actin filaments.

Hindlimb Functional Analysis –

Mice were subjected to three assessments to gauge recovery of hindlimb functional capacity following BaCl₂ injury: RotaRod (Scantox, Denmark), Open field cage activity, and hindlimb gait analysis. Mice were acclimated/familiarized with testing procedures 3 days prior to a baseline assessment. Following baseline assessment, mice were subjected to BaCl₂ muscle injury and MuSC transplantation as detailed above. Fourteen days following BaCl₂ injury, follow-up assessments were performed to gauge recovery. Results were calculated as a percent of baseline measurement. For RotaRod, mice are balanced on rotating rod which begins at 4 RPM and accelerates to 40 RPM over 300 seconds and assessed as latency to fall in seconds. Five repeated tests were performed with the best and worst performance excluded. For open field cage activity, mice were allowed free exploration of a 20x20x30cm open field arena for 15 minutes with video tracking to analyze voluntary movement measured as distance traveled in centimeters. For TreadScan Gait Analysis (Clever Sys, Reston, VA, USA) mice are placed on transparent treadmill moving at 10 cm/sec with video recording to determine foot swing and stance times. Only rear feet were analyzed (due to hindlimb injury protocol) and side with the greatest change over the uninjured state (indicating decreased performance and therefore successful injury) was used to determine paw swing time, propel time and brake time averaged across 20 seconds of walking. Illustration of gait phases is presented in Fig 6I of Kang et al.¹⁹.

Statistical analysis –

Statistical analyses performed included Student’s t-test (unpaired and paired, 1- and 2-tailed based on experimental design and hypothesis) between pre-planned comparisons of two groups. All bar graphs presented are means +/− SEM and α was set at p < 0.05. Hierarchical clustering heatmaps were generated in Graphpad Prism V9. The number of replicates, statistical approach used as well as defined markers for significant comparisons or p values for each experiment are indicated in respective figure legends. For all cell culture experiments, technical replicates are not included in n’s rather biological replicates (separate mice/isolations) are indicated and shown. For mass spectrometry studies (metabolomics), repeated measurements/mass spectrometer injections were averaged and used as resulting data for each biological sample which is indicated by n’s. Data distribution was assumed to be normal but this was not formally tested. No statistical methods were used to pre-determine sample sizes but our sample sizes are similar to those reported in previous publications^47–50. For all data quantifications (peak picking, micrograph counting/quantification) data processes were automated irrespective of sample grouping or individuals performing processing were blinded to experimental group/condition. Animals were not randomly assigned to condition (young vs old; wildtype vs transgenic) but animals were randomly selected from within those predetermined groups for experiments and each experiment was blocked such that all groups were represented.

Data Availability –

All source data will be provided as source data file.