Human bone marrow was collected from patients undergoing hip replacement surgery after obtaining their informed consent. The study protocol was approved by the Human Ethics Committee of Guizhou Provincial People's Hospital (Guiyang, China). MSCs were isolated and cultured as previously reported (9).

MSCs were cultured with human recombinant leptin (50 ng/ml; R&D Systems, Inc.) in Dulbecco's modified Eagle's medium (DMEM; Corning, Inc.) with 10% fetal bovine serum (FBS; Thermo Fisher Scientific, Inc.) under normoxic conditions for 24 h, and were then exposed to DMEM without glucose or FBS in a hypoxic chamber (0.5% O₂/5% CO₂; BioSpherix, Ltd.) for an additional 24 h.

A siRNA silencing the expression of human OPA1 (siOPA1) and a scramble siRNA (NC) were purchased from GenePharma. siRNA transfection was coordinated with Lipofectamine RNAi MAX (Invitrogen; Thermo Fisher Scientific, Inc.) according to the manufacturer's instructions. Transfection of 50 nM siRNA and the specific sequences of OPA1 (sense, CCA UGU GGC CCU AUU UAA ATT; antisense, UUU AAA UAG GGC CAC AUG GTT) were conducted as described previously (9).

The harvested cells (2×10⁵ cells per sample) were incubated with a mixture of Annexin V and PI for 30 min according to the manufacturer's instructions (Dojindo Molecular Technologies, Inc.). Apoptotic cells were quantified by flow cytometry (BD Biosciences).

The oxygen consumption of the intact cells was measured as described previously (16,17). Briefly, the harvested cells (1×10⁶ cells) were resuspended with 1 ml MiR05 [containing 0.5 mM EGTA, 3 mM MgCl₂•6H₂O, 60 mM potassium lactobionate, 20 mM taurine, 10 mM KH₂PO₄, 20 mM HEPES, 110 mM sucrose and 1 g/l fatty acid-free bovine serum albumin (BSA) at pH 7.1], and were then transferred into the experimental chamber. Subsequently, using Oroboros Oxygraph-2k, the OCR was analyzed by titrating various substrates/inhibitors of the respiratory chain in a stepwise manner at 30°C (Oroboros Instruments GmbH).

ECAR was measured by FLUOstar Omega (BMG LABTECH GmbH) according to the manufacturer's instructions. Briefly, MSCs were plated in a 96-well plate (1×10⁴ cells/well) overnight. The next day, after replacing with fresh culture media, 10 μl reconstituted pH-Xtra reagent was added into each well (a well with 10 μl fresh culture media without cells was used as blank control). Subsequently, 100 μl mineral oil was promptly added into each well, and fluorescence intensity was measured by a fluorescence plate reader (380 nm excitation and 615 nm emission).

Western blotting was conducted as described previously (7). Proteins were extracted from MSCs using RIPA solution (Beyotime Institute of Biotechnology) with protease inhibitor cocktail (Roche Diagnostics). Protein concentrations were quantified using BCA Protein Assay Kit (Thermo Fisher Scientific, Inc.) and separated by sodium dodecyl sulfate-polyacrylamide gel electrophoresis (8-10%), followed by transfer to polyvinylidene difluoride membranes (EMD Millipore). After blocking with 5% BSA for 1 h at room temperature, the proteins of interest were incubated with primary antibodies overnight at 4°C: Cleaved caspase-3 (1:500, Cell Signaling Technology, Inc.); voltage-dependent anion channel (VDAC) (1:500, Santa Cruz Biotechnology, Inc.); poly-OPA1 (1:1,000, BD Biosciences); glucose transporter (GLUT)4 (1:500, Santa Cruz Biotechnology, Inc.); GLUT1 (1:1,000, Abcam); sodium-glucose symporter (SGLT)1 (1:1,000, Abcam); β-actin (1:3,000, KangChen Biotech. Co., Ltd.). On the following day, the membranes were incubated with anti-rabbit and anti-mouse secondary antibody (1:5,000, Abcam) for 1 h at room temperature, and then detected using the Gel Doc EZ Imaging System (Bio-Rad Laboratories, Inc.) with an ECL kit (Merck KGaA), and analyzed using Image Lab software (Bio-Rad Laboratories, Inc.).

The gene expression levels were measured by extracting total RNA using TRIzol reagent (Invitrogen; Thermo Fisher Scientific, Inc.), according to the manufacturer's instructions. RT-qPCR was performed as previously described (18). Briefly, the RNA concentration was measured by a NanoDrop spectrophotometer (Thermo Fisher Scientific, Inc.). Total RNA (2 μg) was reverse–transcribed to cDNA using a Reverse Transcription kit (Takara). RT-qPCR was performed using SYBR Green PCR Master Mix (Takara). The cycling conditions were as follows: 95°C for 10 min, followed by 40 cycles of 15 sec at 95°C and 30 sec at 60°C. mRNA levels were normalized to β-actin and determined using the 2^−ΔΔCq method. The gene primers are listed in Table I.

Mitochondrial ultrastructure was examined by TEM. Briefly, the cells were fixed with 2.5% glutaraldehyde for at least 4 h. After washing with phosphate–buffered saline, the cells were post–fixed with 1% OsO₄ for 1-2 h. Then, following dehydration by the ethanol gradient method, the cells were infiltrated by acetone overnight. Subsequently, the specimens were embedded in Spurr resin, heated at 70°C for 9 h and sectioned with Leica EM UC7 Ultramicrotome (Leica Microsystems GmbH). Uranyl acetate and alkaline lead citrate were used to stain the sections. Images were captured randomly using Hitachi H-7650 TEM (Hitachi, Ltd.).

Glycogen content was measured by PAS staining according to the manufacturer's instructions (SenBeiJia Biotechnology, Co.). Briefly, the harvested cells were fixed by PAS fixing liquid for 10 min. After washing and drying, the cells were incubated in periodic acid solution at room temperature for 15 min and stained by Schiff reagent for 30 min in the dark. The images were captured randomly at a magnification of ×200. Glucose content was detected using the glucose oxidase method according to the manufacturer's instructions (Applygen Technologies, Inc), and absorbance was measured at 570 nm by a microplate reader (Bio-Rad Laboratories, Inc.).

Cellular ATP content was analyzed by a luciferin/luciferase-based kit according to the manufacturer's instructions (Beyotime Institute of Biotechnology).

The lactate content was measured using a lactate assay kit (Jiancheng Biotechnology) according to the manufacturer's instructions.

MSCs were incubated with leptin (50 ng/ml) under normal culture conditions for 24 h. Subsequently, these MSCs were cultured with leptin in the presence or absence of 36 nM sotagliflozin (LX4211; Selleck Chemicals), an inhibitor of SGLT1, and subjected to hypoxia and glucose + serum deprivation for an additional 24 h.

A MI mouse model was induced by ligation of the left anterior descending coronary artery in 10-12-week-old male C57BL/6J mice (20-25 g), as previously described (7). MSCs (1.5×10⁵ cells in 20 μl DMEM per mouse) were transplanted to the border zone of the infarct.

The mice were euthanized, and the hearts were harvested and embedded in paraffin on day 28 post–MI. The tissue samples were cut into 3-μm sections and stained with Sirius Red (Beijing Solarbio Science & Technology Co., Ltd.), as described previously (19). Infarct areas were analyzed according to the sum of the endocardial and epicardial length of the infarct zone in proportion to the total length of the endocardial and epicardial left ventricle using Image-Pro-Plus software (Media Cybernetics, Inc.).

Echocardiography was performed on day 28 post-MI, as described previously (20). Cardiac morphology and function were analyzed via a Vevo 2100 system (VisualSonics, Inc.). Left ventricular ejection fraction was measured for at least three cardiac cycle intervals.

All experiments were performed at least three times. All data are presented as mean ± standard error of the mean. Student's t–test was used to determine significance when comparing two sets of data. One-way analysis of variance followed by Tukey's post hoc test was used for comparison of more than three sets of data. All data were analyzed using SPSS 17.0 (SPSS Inc.). The statistics were presented by GraphPad Prism 5 (GraphPad Software, Inc.). P<0.05 was considered to indicate statistically significant differences.

In order to examine the protective effect of leptin on MSCs in response to the microenvironment after transplantation, MSCs were incubated with recombinant leptin (MSCs-Lep) for 24 h under normal culture conditions, and then exposed to hypoxia and glucose + serum deprivation for an additional 24 h in vitro. The solvent of leptin, Tris buffer, was used as control (MSCs-Con).

To distinguish the apoptotic cells affected by hypoxia, the percentage of Annexin V⁺ and PI^−/+ staining was calculated by flow cytometry. Decreased apoptotic ratio was detected in MSCs-Lep, while a higher percentage of Annexin V⁺ and PI^−/+ staining was detected in MSCs-Con (Fig. 1A and B). Similarly, lower expression of cleaved caspase-3 was observed in MSCs-Lep compared with MSCs-Con (Fig. 1C and D). These data suggest that leptin exerted a cytoprotective effect on MSCs against hypoxic insults.

Mitochondrial morphology is closely associated with cell fate; an abundance of pro-apoptotic factors is stored in mitochondrial cristae, and these pro-apoptotic proteins are released from the mitochondria into the cytoplasm to activate the apoptosis signaling pathway (21-23). First, mitochondrial morphology was evaluated by TEM. Elongated tubular mitochondria were prominent in MSCs-Lep under hypoxic stress, while the mitochondria observed in MSCs-Con were fewer and sparsely distributed (Fig. 2A). In contrast to MSCs-Con, increased mitochondrial length and area were observed in MSCs-Lep (Fig. 2B). Moreover, the increased protein level of VDAC indicated that the mitochondrial content was higher in MSCs-Lep compared with MSCs-Con (Fig. 2C and D).

The different isoforms of OPA1 modulate mitochondrial inner membrane dynamics, among which L-OPA1 (isoforms a-b, ~85-100 kDa) are implicated in fusion, whereas S-OPA1 (isoforms c-e, ~75-85 kDa) are associated with fission (12,24). It was previously reported that leptin enhanced the accumulation of L-OPA1 via OMA1 ubiquitination, thereby inducing mitochondrial fusion, to prevent apoptosis of MSCs (9). Similarly, in the present study, significant accumulation of L-OPA1 was detected in MSCs-Lep, with higher protein expression of S-OPA1 in MSCs-Con (Fig. 2E and F), indicating that leptin promotes mitochondrial fusion through accumulation of L-OPA1.

Glucose uptake and glycolysis are crucial for MSC survival, particularly in hypoxia (6). Despite the fact that leptin was found to prevent MSC apoptosis, whether glucose metabolic alterations were involved in the mechanism underlying leptin-induced MSC survival remained largely unclear. In order to determine the changes in glycolysis of MSCs, the glycogen assay was performed, which revealed that leptin induced an increase in the glycogen content of MSCs-Lep compared with MSCs-Con, as determined by PAS staining (Fig. 3A). Furthermore, cellular glucose was examined using the glucose oxidase method. In contrast to MSCs-Con, increased glucose uptake, elicited by leptin treatment, was found at the time point of hypoxic insults (Fig. 3B). Moreover, several essential genes linked to glucose metabolism were analyzed, including glyceraldehyde-3-phosphate dehydrogenase (GAPDH), hexo-kinase (HK)-1, lactate dehydrogenase (LDH)-α, and pyruvate dehydrogenase (PDH)-α, as well as key genes of fatty acid oxidation, such as carnitine palmitoyltransferase (cpt)-1α (25). In contrast to MSCs-Con, increased gene expression of GAPDH, HK-1, LDH-α and PDH-α, and decreased gene expression of cpt-1α, were observed in MSCs-Lep (Fig. 3C). Subsequently, the increased lactate content of the supernatant was assessed in MSCs-Lep compared with MSCs-Con (Fig. 3D) and was found to be decreased in MSCs-Lep (Fig. 3D). Furthermore, the level of adenosine triphosphate (ATP) production was higher in MSCs-Lep compared with that in MSCs-Con (Fig. 3E), and increased ECAR was measured in MSCs-Lep compared with MSCs-Con (Fig. 3F), indicating that leptin activated glycolysis to produce energy. However, a markedly lower OCR was observed in MSCs-Lep compared with MSCs-Con (Fig. 3G), demonstrating that mitochondrial OXPHOS was insufficient following leptin administration. Therefore, increased glycolysis of MSCs was observed in association with improved survival following leptin treatment under hypoxic conditions.

Improved glycolysis regulated by leptin has been confirmed based on the increased insulin sensitivity and reduction of hepatic gluconeogenesis (8,26). In addition, insulin regulates glucose uptake through the Akt-mTOR-NFκB-OPA1 signaling pathway in cardiomyocytes (27). These results prompted us to investigate whether the anti-apoptotic effect of leptin was mediated via OPA1-induced glycolysis. OPA1 gene expression was silenced by small interfering RNA (siOPA1) treatment. The siOPA1 transfection efficiency was verified by western blotting (Fig. S1). In the siOPA1 group, no significant effect of leptin on glycogen staining (Fig. 4A) or the glucose content (Fig. 4B) was observed. The increase in the gene expression levels of GAPDH, HK-1, LDH-α and PDH-α following leptin treatment was blunted by OPA1 silencing, and there was no significant difference in cpt–1α gene expression between the siOPA1 + Con and siOPA1 + Lep groups (Fig. 4C). Similarly, we also observed that the increased lactate levels in the supernatant and ECAR in MSCs-Lep were blunted by siOPA1 treatment, whereas the reduced cellular lactate in MSCs-Lep started to accumulate again (Fig. 4D and F). Following siOPA1 transfection, ATP production was not altered in the siOPA1 + Con or siOPA1 + Lep groups, while a slight reduction in ATP was detected when comparing the NC + Con and siOPA1 + Con groups (Fig. 4E). These results indicate that improvement of leptin-induced glycolysis is primarily dependent on OPA1. Following siOPA1 transfection, the levels of OCR did not change in the siOPA1 + Con and siOPA1 + Lep groups, while they were decreased compared with the NC + Con group (Fig. 4G), which was in agreement with an earlier study reporting that OPA1 can regulate mitochondrial metabolism (28).

Glucose uptake is essential for cell growth, metabolism and proliferation. Glucose diffusion across the cell membrane is accomplished by glucose transporters, including Na⁺-independent facilitated diffusion glucose transporters (GLUTs), sodium-glucose symporters 1 (SGLT1) and SWEETs; GLUTs and SGLT1 are responsible for glucose transport in mammals, whereas SWEETs is mainly found in plants (29). To explore how leptin regulates glucose transport, several major glucose transporters were investigated to identify the obviously increased gene expression of GLUT1, GLUT4 and SGLT1 (Fig. 5A). GLUT1 acts as 'housekeeping' gene, as it is ubiquitous in all tissues, whereas GLUT4 is principally found in adipocytes and muscle (30). Consequently, the protein levels of GLUT1, GLUT4 and SGLT1 were measured; we observed activation of SGLT1 at the protein level in MSCs–Lep compared with MSCs–Con, while no significant impact on the expression levels of GLUT1 and GLUT4 was noted following leptin treatment (Fig. 5B and C). Subsequently, to confirm that leptin modulates glycolysis through the SGLT1 protein, leptin–treated MSCs were administered sotagliflozin (LX4211), which is as an inhibitor of SGLT1 (31). We observed that the increase in glucose content mediated by leptin was reversed after LX4211 administration (Fig. 5D and E). Moreover, the upregulated gene expression of GAPDH, HK-1, LDH-α and PDH-α by leptin treatment was blunted by LX4211 administration, in contrast with the unchanged gene expression of cpt-1α in the Con + LX4211 and Lep + LX4211 groups (Fig. 5F). Additionally, the increased lactate content of the supernatant and the increased ECAR level mediated by leptin were blocked by LX4211 administration, while the reduction in cellular lactate accumulation following leptin treatment was hindered by LX4211 administration (Fig. 5G and H), suggesting that leptin modulates glycolysis in MSCs primarily through SGLT1 activation. However, a slightly higher lactate content of the supernatant and ECAR level were observed in the Lep + LX4211 group compared with the Con + LX4211 group (Fig. 5G and H), indicating that there may be another SGLT1-independent pathway involved in leptin-mediated glycolysis, at least in part.

To evaluate the leptin-mediated MSC survival in vivo, MSCs were transplanted into the border zone of the infarcted area in an MI mouse model. On the 28th day after MI, cardiac function was analyzed by two-dimensional echo-cardiography examination, and the infarct size was measured by Sirius Red staining. Notably, in contrast to MSCs-Con, improved cardiac function (Fig. 6A and B) and decreased infarct size (Fig. 6C and D) were observed in MSCs-Lep. Taken together, these data suggest that leptin endowed MSCs with apoptotic resistance through a switch to glycolysis, mediated at least in part via SGLT1, which was highly dependent on OPA1 (Fig. 6E).